diff options
279 files changed, 9244 insertions, 3004 deletions
diff --git a/Documentation/cgroups/cgroups.txt b/Documentation/cgroups/cgroups.txt index d9014aa0eb68..e33ee74eee77 100644 --- a/Documentation/cgroups/cgroups.txt +++ b/Documentation/cgroups/cgroups.txt | |||
@@ -227,7 +227,6 @@ Each cgroup is represented by a directory in the cgroup file system | |||
227 | containing the following files describing that cgroup: | 227 | containing the following files describing that cgroup: |
228 | 228 | ||
229 | - tasks: list of tasks (by pid) attached to that cgroup | 229 | - tasks: list of tasks (by pid) attached to that cgroup |
230 | - releasable flag: cgroup currently removeable? | ||
231 | - notify_on_release flag: run the release agent on exit? | 230 | - notify_on_release flag: run the release agent on exit? |
232 | - release_agent: the path to use for release notifications (this file | 231 | - release_agent: the path to use for release notifications (this file |
233 | exists in the top cgroup only) | 232 | exists in the top cgroup only) |
@@ -360,7 +359,7 @@ Now you want to do something with this cgroup. | |||
360 | 359 | ||
361 | In this directory you can find several files: | 360 | In this directory you can find several files: |
362 | # ls | 361 | # ls |
363 | notify_on_release releasable tasks | 362 | notify_on_release tasks |
364 | (plus whatever files added by the attached subsystems) | 363 | (plus whatever files added by the attached subsystems) |
365 | 364 | ||
366 | Now attach your shell to this cgroup: | 365 | Now attach your shell to this cgroup: |
@@ -479,7 +478,6 @@ newly-created cgroup if an error occurs after this subsystem's | |||
479 | create() method has been called for the new cgroup). | 478 | create() method has been called for the new cgroup). |
480 | 479 | ||
481 | void pre_destroy(struct cgroup_subsys *ss, struct cgroup *cgrp); | 480 | void pre_destroy(struct cgroup_subsys *ss, struct cgroup *cgrp); |
482 | (cgroup_mutex held by caller) | ||
483 | 481 | ||
484 | Called before checking the reference count on each subsystem. This may | 482 | Called before checking the reference count on each subsystem. This may |
485 | be useful for subsystems which have some extra references even if | 483 | be useful for subsystems which have some extra references even if |
@@ -498,6 +496,7 @@ remain valid while the caller holds cgroup_mutex. | |||
498 | 496 | ||
499 | void attach(struct cgroup_subsys *ss, struct cgroup *cgrp, | 497 | void attach(struct cgroup_subsys *ss, struct cgroup *cgrp, |
500 | struct cgroup *old_cgrp, struct task_struct *task) | 498 | struct cgroup *old_cgrp, struct task_struct *task) |
499 | (cgroup_mutex held by caller) | ||
501 | 500 | ||
502 | Called after the task has been attached to the cgroup, to allow any | 501 | Called after the task has been attached to the cgroup, to allow any |
503 | post-attachment activity that requires memory allocations or blocking. | 502 | post-attachment activity that requires memory allocations or blocking. |
@@ -511,6 +510,7 @@ void exit(struct cgroup_subsys *ss, struct task_struct *task) | |||
511 | Called during task exit. | 510 | Called during task exit. |
512 | 511 | ||
513 | int populate(struct cgroup_subsys *ss, struct cgroup *cgrp) | 512 | int populate(struct cgroup_subsys *ss, struct cgroup *cgrp) |
513 | (cgroup_mutex held by caller) | ||
514 | 514 | ||
515 | Called after creation of a cgroup to allow a subsystem to populate | 515 | Called after creation of a cgroup to allow a subsystem to populate |
516 | the cgroup directory with file entries. The subsystem should make | 516 | the cgroup directory with file entries. The subsystem should make |
@@ -520,6 +520,7 @@ method can return an error code, the error code is currently not | |||
520 | always handled well. | 520 | always handled well. |
521 | 521 | ||
522 | void post_clone(struct cgroup_subsys *ss, struct cgroup *cgrp) | 522 | void post_clone(struct cgroup_subsys *ss, struct cgroup *cgrp) |
523 | (cgroup_mutex held by caller) | ||
523 | 524 | ||
524 | Called at the end of cgroup_clone() to do any paramater | 525 | Called at the end of cgroup_clone() to do any paramater |
525 | initialization which might be required before a task could attach. For | 526 | initialization which might be required before a task could attach. For |
@@ -527,7 +528,7 @@ example in cpusets, no task may attach before 'cpus' and 'mems' are set | |||
527 | up. | 528 | up. |
528 | 529 | ||
529 | void bind(struct cgroup_subsys *ss, struct cgroup *root) | 530 | void bind(struct cgroup_subsys *ss, struct cgroup *root) |
530 | (cgroup_mutex held by caller) | 531 | (cgroup_mutex and ss->hierarchy_mutex held by caller) |
531 | 532 | ||
532 | Called when a cgroup subsystem is rebound to a different hierarchy | 533 | Called when a cgroup subsystem is rebound to a different hierarchy |
533 | and root cgroup. Currently this will only involve movement between | 534 | and root cgroup. Currently this will only involve movement between |
diff --git a/Documentation/controllers/memcg_test.txt b/Documentation/controllers/memcg_test.txt new file mode 100644 index 000000000000..08d4d3ea0d79 --- /dev/null +++ b/Documentation/controllers/memcg_test.txt | |||
@@ -0,0 +1,342 @@ | |||
1 | Memory Resource Controller(Memcg) Implementation Memo. | ||
2 | Last Updated: 2008/12/15 | ||
3 | Base Kernel Version: based on 2.6.28-rc8-mm. | ||
4 | |||
5 | Because VM is getting complex (one of reasons is memcg...), memcg's behavior | ||
6 | is complex. This is a document for memcg's internal behavior. | ||
7 | Please note that implementation details can be changed. | ||
8 | |||
9 | (*) Topics on API should be in Documentation/controllers/memory.txt) | ||
10 | |||
11 | 0. How to record usage ? | ||
12 | 2 objects are used. | ||
13 | |||
14 | page_cgroup ....an object per page. | ||
15 | Allocated at boot or memory hotplug. Freed at memory hot removal. | ||
16 | |||
17 | swap_cgroup ... an entry per swp_entry. | ||
18 | Allocated at swapon(). Freed at swapoff(). | ||
19 | |||
20 | The page_cgroup has USED bit and double count against a page_cgroup never | ||
21 | occurs. swap_cgroup is used only when a charged page is swapped-out. | ||
22 | |||
23 | 1. Charge | ||
24 | |||
25 | a page/swp_entry may be charged (usage += PAGE_SIZE) at | ||
26 | |||
27 | mem_cgroup_newpage_charge() | ||
28 | Called at new page fault and Copy-On-Write. | ||
29 | |||
30 | mem_cgroup_try_charge_swapin() | ||
31 | Called at do_swap_page() (page fault on swap entry) and swapoff. | ||
32 | Followed by charge-commit-cancel protocol. (With swap accounting) | ||
33 | At commit, a charge recorded in swap_cgroup is removed. | ||
34 | |||
35 | mem_cgroup_cache_charge() | ||
36 | Called at add_to_page_cache() | ||
37 | |||
38 | mem_cgroup_cache_charge_swapin() | ||
39 | Called at shmem's swapin. | ||
40 | |||
41 | mem_cgroup_prepare_migration() | ||
42 | Called before migration. "extra" charge is done and followed by | ||
43 | charge-commit-cancel protocol. | ||
44 | At commit, charge against oldpage or newpage will be committed. | ||
45 | |||
46 | 2. Uncharge | ||
47 | a page/swp_entry may be uncharged (usage -= PAGE_SIZE) by | ||
48 | |||
49 | mem_cgroup_uncharge_page() | ||
50 | Called when an anonymous page is fully unmapped. I.e., mapcount goes | ||
51 | to 0. If the page is SwapCache, uncharge is delayed until | ||
52 | mem_cgroup_uncharge_swapcache(). | ||
53 | |||
54 | mem_cgroup_uncharge_cache_page() | ||
55 | Called when a page-cache is deleted from radix-tree. If the page is | ||
56 | SwapCache, uncharge is delayed until mem_cgroup_uncharge_swapcache(). | ||
57 | |||
58 | mem_cgroup_uncharge_swapcache() | ||
59 | Called when SwapCache is removed from radix-tree. The charge itself | ||
60 | is moved to swap_cgroup. (If mem+swap controller is disabled, no | ||
61 | charge to swap occurs.) | ||
62 | |||
63 | mem_cgroup_uncharge_swap() | ||
64 | Called when swp_entry's refcnt goes down to 0. A charge against swap | ||
65 | disappears. | ||
66 | |||
67 | mem_cgroup_end_migration(old, new) | ||
68 | At success of migration old is uncharged (if necessary), a charge | ||
69 | to new page is committed. At failure, charge to old page is committed. | ||
70 | |||
71 | 3. charge-commit-cancel | ||
72 | In some case, we can't know this "charge" is valid or not at charging | ||
73 | (because of races). | ||
74 | To handle such case, there are charge-commit-cancel functions. | ||
75 | mem_cgroup_try_charge_XXX | ||
76 | mem_cgroup_commit_charge_XXX | ||
77 | mem_cgroup_cancel_charge_XXX | ||
78 | these are used in swap-in and migration. | ||
79 | |||
80 | At try_charge(), there are no flags to say "this page is charged". | ||
81 | at this point, usage += PAGE_SIZE. | ||
82 | |||
83 | At commit(), the function checks the page should be charged or not | ||
84 | and set flags or avoid charging.(usage -= PAGE_SIZE) | ||
85 | |||
86 | At cancel(), simply usage -= PAGE_SIZE. | ||
87 | |||
88 | Under below explanation, we assume CONFIG_MEM_RES_CTRL_SWAP=y. | ||
89 | |||
90 | 4. Anonymous | ||
91 | Anonymous page is newly allocated at | ||
92 | - page fault into MAP_ANONYMOUS mapping. | ||
93 | - Copy-On-Write. | ||
94 | It is charged right after it's allocated before doing any page table | ||
95 | related operations. Of course, it's uncharged when another page is used | ||
96 | for the fault address. | ||
97 | |||
98 | At freeing anonymous page (by exit() or munmap()), zap_pte() is called | ||
99 | and pages for ptes are freed one by one.(see mm/memory.c). Uncharges | ||
100 | are done at page_remove_rmap() when page_mapcount() goes down to 0. | ||
101 | |||
102 | Another page freeing is by page-reclaim (vmscan.c) and anonymous | ||
103 | pages are swapped out. In this case, the page is marked as | ||
104 | PageSwapCache(). uncharge() routine doesn't uncharge the page marked | ||
105 | as SwapCache(). It's delayed until __delete_from_swap_cache(). | ||
106 | |||
107 | 4.1 Swap-in. | ||
108 | At swap-in, the page is taken from swap-cache. There are 2 cases. | ||
109 | |||
110 | (a) If the SwapCache is newly allocated and read, it has no charges. | ||
111 | (b) If the SwapCache has been mapped by processes, it has been | ||
112 | charged already. | ||
113 | |||
114 | This swap-in is one of the most complicated work. In do_swap_page(), | ||
115 | following events occur when pte is unchanged. | ||
116 | |||
117 | (1) the page (SwapCache) is looked up. | ||
118 | (2) lock_page() | ||
119 | (3) try_charge_swapin() | ||
120 | (4) reuse_swap_page() (may call delete_swap_cache()) | ||
121 | (5) commit_charge_swapin() | ||
122 | (6) swap_free(). | ||
123 | |||
124 | Considering following situation for example. | ||
125 | |||
126 | (A) The page has not been charged before (2) and reuse_swap_page() | ||
127 | doesn't call delete_from_swap_cache(). | ||
128 | (B) The page has not been charged before (2) and reuse_swap_page() | ||
129 | calls delete_from_swap_cache(). | ||
130 | (C) The page has been charged before (2) and reuse_swap_page() doesn't | ||
131 | call delete_from_swap_cache(). | ||
132 | (D) The page has been charged before (2) and reuse_swap_page() calls | ||
133 | delete_from_swap_cache(). | ||
134 | |||
135 | memory.usage/memsw.usage changes to this page/swp_entry will be | ||
136 | Case (A) (B) (C) (D) | ||
137 | Event | ||
138 | Before (2) 0/ 1 0/ 1 1/ 1 1/ 1 | ||
139 | =========================================== | ||
140 | (3) +1/+1 +1/+1 +1/+1 +1/+1 | ||
141 | (4) - 0/ 0 - -1/ 0 | ||
142 | (5) 0/-1 0/ 0 -1/-1 0/ 0 | ||
143 | (6) - 0/-1 - 0/-1 | ||
144 | =========================================== | ||
145 | Result 1/ 1 1/ 1 1/ 1 1/ 1 | ||
146 | |||
147 | In any cases, charges to this page should be 1/ 1. | ||
148 | |||
149 | 4.2 Swap-out. | ||
150 | At swap-out, typical state transition is below. | ||
151 | |||
152 | (a) add to swap cache. (marked as SwapCache) | ||
153 | swp_entry's refcnt += 1. | ||
154 | (b) fully unmapped. | ||
155 | swp_entry's refcnt += # of ptes. | ||
156 | (c) write back to swap. | ||
157 | (d) delete from swap cache. (remove from SwapCache) | ||
158 | swp_entry's refcnt -= 1. | ||
159 | |||
160 | |||
161 | At (b), the page is marked as SwapCache and not uncharged. | ||
162 | At (d), the page is removed from SwapCache and a charge in page_cgroup | ||
163 | is moved to swap_cgroup. | ||
164 | |||
165 | Finally, at task exit, | ||
166 | (e) zap_pte() is called and swp_entry's refcnt -=1 -> 0. | ||
167 | Here, a charge in swap_cgroup disappears. | ||
168 | |||
169 | 5. Page Cache | ||
170 | Page Cache is charged at | ||
171 | - add_to_page_cache_locked(). | ||
172 | |||
173 | uncharged at | ||
174 | - __remove_from_page_cache(). | ||
175 | |||
176 | The logic is very clear. (About migration, see below) | ||
177 | Note: __remove_from_page_cache() is called by remove_from_page_cache() | ||
178 | and __remove_mapping(). | ||
179 | |||
180 | 6. Shmem(tmpfs) Page Cache | ||
181 | Memcg's charge/uncharge have special handlers of shmem. The best way | ||
182 | to understand shmem's page state transition is to read mm/shmem.c. | ||
183 | But brief explanation of the behavior of memcg around shmem will be | ||
184 | helpful to understand the logic. | ||
185 | |||
186 | Shmem's page (just leaf page, not direct/indirect block) can be on | ||
187 | - radix-tree of shmem's inode. | ||
188 | - SwapCache. | ||
189 | - Both on radix-tree and SwapCache. This happens at swap-in | ||
190 | and swap-out, | ||
191 | |||
192 | It's charged when... | ||
193 | - A new page is added to shmem's radix-tree. | ||
194 | - A swp page is read. (move a charge from swap_cgroup to page_cgroup) | ||
195 | It's uncharged when | ||
196 | - A page is removed from radix-tree and not SwapCache. | ||
197 | - When SwapCache is removed, a charge is moved to swap_cgroup. | ||
198 | - When swp_entry's refcnt goes down to 0, a charge in swap_cgroup | ||
199 | disappears. | ||
200 | |||
201 | 7. Page Migration | ||
202 | One of the most complicated functions is page-migration-handler. | ||
203 | Memcg has 2 routines. Assume that we are migrating a page's contents | ||
204 | from OLDPAGE to NEWPAGE. | ||
205 | |||
206 | Usual migration logic is.. | ||
207 | (a) remove the page from LRU. | ||
208 | (b) allocate NEWPAGE (migration target) | ||
209 | (c) lock by lock_page(). | ||
210 | (d) unmap all mappings. | ||
211 | (e-1) If necessary, replace entry in radix-tree. | ||
212 | (e-2) move contents of a page. | ||
213 | (f) map all mappings again. | ||
214 | (g) pushback the page to LRU. | ||
215 | (-) OLDPAGE will be freed. | ||
216 | |||
217 | Before (g), memcg should complete all necessary charge/uncharge to | ||
218 | NEWPAGE/OLDPAGE. | ||
219 | |||
220 | The point is.... | ||
221 | - If OLDPAGE is anonymous, all charges will be dropped at (d) because | ||
222 | try_to_unmap() drops all mapcount and the page will not be | ||
223 | SwapCache. | ||
224 | |||
225 | - If OLDPAGE is SwapCache, charges will be kept at (g) because | ||
226 | __delete_from_swap_cache() isn't called at (e-1) | ||
227 | |||
228 | - If OLDPAGE is page-cache, charges will be kept at (g) because | ||
229 | remove_from_swap_cache() isn't called at (e-1) | ||
230 | |||
231 | memcg provides following hooks. | ||
232 | |||
233 | - mem_cgroup_prepare_migration(OLDPAGE) | ||
234 | Called after (b) to account a charge (usage += PAGE_SIZE) against | ||
235 | memcg which OLDPAGE belongs to. | ||
236 | |||
237 | - mem_cgroup_end_migration(OLDPAGE, NEWPAGE) | ||
238 | Called after (f) before (g). | ||
239 | If OLDPAGE is used, commit OLDPAGE again. If OLDPAGE is already | ||
240 | charged, a charge by prepare_migration() is automatically canceled. | ||
241 | If NEWPAGE is used, commit NEWPAGE and uncharge OLDPAGE. | ||
242 | |||
243 | But zap_pte() (by exit or munmap) can be called while migration, | ||
244 | we have to check if OLDPAGE/NEWPAGE is a valid page after commit(). | ||
245 | |||
246 | 8. LRU | ||
247 | Each memcg has its own private LRU. Now, it's handling is under global | ||
248 | VM's control (means that it's handled under global zone->lru_lock). | ||
249 | Almost all routines around memcg's LRU is called by global LRU's | ||
250 | list management functions under zone->lru_lock(). | ||
251 | |||
252 | A special function is mem_cgroup_isolate_pages(). This scans | ||
253 | memcg's private LRU and call __isolate_lru_page() to extract a page | ||
254 | from LRU. | ||
255 | (By __isolate_lru_page(), the page is removed from both of global and | ||
256 | private LRU.) | ||
257 | |||
258 | |||
259 | 9. Typical Tests. | ||
260 | |||
261 | Tests for racy cases. | ||
262 | |||
263 | 9.1 Small limit to memcg. | ||
264 | When you do test to do racy case, it's good test to set memcg's limit | ||
265 | to be very small rather than GB. Many races found in the test under | ||
266 | xKB or xxMB limits. | ||
267 | (Memory behavior under GB and Memory behavior under MB shows very | ||
268 | different situation.) | ||
269 | |||
270 | 9.2 Shmem | ||
271 | Historically, memcg's shmem handling was poor and we saw some amount | ||
272 | of troubles here. This is because shmem is page-cache but can be | ||
273 | SwapCache. Test with shmem/tmpfs is always good test. | ||
274 | |||
275 | 9.3 Migration | ||
276 | For NUMA, migration is an another special case. To do easy test, cpuset | ||
277 | is useful. Following is a sample script to do migration. | ||
278 | |||
279 | mount -t cgroup -o cpuset none /opt/cpuset | ||
280 | |||
281 | mkdir /opt/cpuset/01 | ||
282 | echo 1 > /opt/cpuset/01/cpuset.cpus | ||
283 | echo 0 > /opt/cpuset/01/cpuset.mems | ||
284 | echo 1 > /opt/cpuset/01/cpuset.memory_migrate | ||
285 | mkdir /opt/cpuset/02 | ||
286 | echo 1 > /opt/cpuset/02/cpuset.cpus | ||
287 | echo 1 > /opt/cpuset/02/cpuset.mems | ||
288 | echo 1 > /opt/cpuset/02/cpuset.memory_migrate | ||
289 | |||
290 | In above set, when you moves a task from 01 to 02, page migration to | ||
291 | node 0 to node 1 will occur. Following is a script to migrate all | ||
292 | under cpuset. | ||
293 | -- | ||
294 | move_task() | ||
295 | { | ||
296 | for pid in $1 | ||
297 | do | ||
298 | /bin/echo $pid >$2/tasks 2>/dev/null | ||
299 | echo -n $pid | ||
300 | echo -n " " | ||
301 | done | ||
302 | echo END | ||
303 | } | ||
304 | |||
305 | G1_TASK=`cat ${G1}/tasks` | ||
306 | G2_TASK=`cat ${G2}/tasks` | ||
307 | move_task "${G1_TASK}" ${G2} & | ||
308 | -- | ||
309 | 9.4 Memory hotplug. | ||
310 | memory hotplug test is one of good test. | ||
311 | to offline memory, do following. | ||
312 | # echo offline > /sys/devices/system/memory/memoryXXX/state | ||
313 | (XXX is the place of memory) | ||
314 | This is an easy way to test page migration, too. | ||
315 | |||
316 | 9.5 mkdir/rmdir | ||
317 | When using hierarchy, mkdir/rmdir test should be done. | ||
318 | Use tests like the following. | ||
319 | |||
320 | echo 1 >/opt/cgroup/01/memory/use_hierarchy | ||
321 | mkdir /opt/cgroup/01/child_a | ||
322 | mkdir /opt/cgroup/01/child_b | ||
323 | |||
324 | set limit to 01. | ||
325 | add limit to 01/child_b | ||
326 | run jobs under child_a and child_b | ||
327 | |||
328 | create/delete following groups at random while jobs are running. | ||
329 | /opt/cgroup/01/child_a/child_aa | ||
330 | /opt/cgroup/01/child_b/child_bb | ||
331 | /opt/cgroup/01/child_c | ||
332 | |||
333 | running new jobs in new group is also good. | ||
334 | |||
335 | 9.6 Mount with other subsystems. | ||
336 | Mounting with other subsystems is a good test because there is a | ||
337 | race and lock dependency with other cgroup subsystems. | ||
338 | |||
339 | example) | ||
340 | # mount -t cgroup none /cgroup -t cpuset,memory,cpu,devices | ||
341 | |||
342 | and do task move, mkdir, rmdir etc...under this. | ||
diff --git a/Documentation/controllers/memory.txt b/Documentation/controllers/memory.txt index 1c07547d3f81..e1501964df1e 100644 --- a/Documentation/controllers/memory.txt +++ b/Documentation/controllers/memory.txt | |||
@@ -137,7 +137,32 @@ behind this approach is that a cgroup that aggressively uses a shared | |||
137 | page will eventually get charged for it (once it is uncharged from | 137 | page will eventually get charged for it (once it is uncharged from |
138 | the cgroup that brought it in -- this will happen on memory pressure). | 138 | the cgroup that brought it in -- this will happen on memory pressure). |
139 | 139 | ||
140 | 2.4 Reclaim | 140 | Exception: If CONFIG_CGROUP_CGROUP_MEM_RES_CTLR_SWAP is not used.. |
141 | When you do swapoff and make swapped-out pages of shmem(tmpfs) to | ||
142 | be backed into memory in force, charges for pages are accounted against the | ||
143 | caller of swapoff rather than the users of shmem. | ||
144 | |||
145 | |||
146 | 2.4 Swap Extension (CONFIG_CGROUP_MEM_RES_CTLR_SWAP) | ||
147 | Swap Extension allows you to record charge for swap. A swapped-in page is | ||
148 | charged back to original page allocator if possible. | ||
149 | |||
150 | When swap is accounted, following files are added. | ||
151 | - memory.memsw.usage_in_bytes. | ||
152 | - memory.memsw.limit_in_bytes. | ||
153 | |||
154 | usage of mem+swap is limited by memsw.limit_in_bytes. | ||
155 | |||
156 | Note: why 'mem+swap' rather than swap. | ||
157 | The global LRU(kswapd) can swap out arbitrary pages. Swap-out means | ||
158 | to move account from memory to swap...there is no change in usage of | ||
159 | mem+swap. | ||
160 | |||
161 | In other words, when we want to limit the usage of swap without affecting | ||
162 | global LRU, mem+swap limit is better than just limiting swap from OS point | ||
163 | of view. | ||
164 | |||
165 | 2.5 Reclaim | ||
141 | 166 | ||
142 | Each cgroup maintains a per cgroup LRU that consists of an active | 167 | Each cgroup maintains a per cgroup LRU that consists of an active |
143 | and inactive list. When a cgroup goes over its limit, we first try | 168 | and inactive list. When a cgroup goes over its limit, we first try |
@@ -207,12 +232,6 @@ exceeded. | |||
207 | The memory.stat file gives accounting information. Now, the number of | 232 | The memory.stat file gives accounting information. Now, the number of |
208 | caches, RSS and Active pages/Inactive pages are shown. | 233 | caches, RSS and Active pages/Inactive pages are shown. |
209 | 234 | ||
210 | The memory.force_empty gives an interface to drop *all* charges by force. | ||
211 | |||
212 | # echo 1 > memory.force_empty | ||
213 | |||
214 | will drop all charges in cgroup. Currently, this is maintained for test. | ||
215 | |||
216 | 4. Testing | 235 | 4. Testing |
217 | 236 | ||
218 | Balbir posted lmbench, AIM9, LTP and vmmstress results [10] and [11]. | 237 | Balbir posted lmbench, AIM9, LTP and vmmstress results [10] and [11]. |
@@ -242,10 +261,106 @@ reclaimed. | |||
242 | 261 | ||
243 | A cgroup can be removed by rmdir, but as discussed in sections 4.1 and 4.2, a | 262 | A cgroup can be removed by rmdir, but as discussed in sections 4.1 and 4.2, a |
244 | cgroup might have some charge associated with it, even though all | 263 | cgroup might have some charge associated with it, even though all |
245 | tasks have migrated away from it. Such charges are automatically dropped at | 264 | tasks have migrated away from it. |
246 | rmdir() if there are no tasks. | 265 | Such charges are freed(at default) or moved to its parent. When moved, |
266 | both of RSS and CACHES are moved to parent. | ||
267 | If both of them are busy, rmdir() returns -EBUSY. See 5.1 Also. | ||
268 | |||
269 | Charges recorded in swap information is not updated at removal of cgroup. | ||
270 | Recorded information is discarded and a cgroup which uses swap (swapcache) | ||
271 | will be charged as a new owner of it. | ||
272 | |||
273 | |||
274 | 5. Misc. interfaces. | ||
275 | |||
276 | 5.1 force_empty | ||
277 | memory.force_empty interface is provided to make cgroup's memory usage empty. | ||
278 | You can use this interface only when the cgroup has no tasks. | ||
279 | When writing anything to this | ||
280 | |||
281 | # echo 0 > memory.force_empty | ||
282 | |||
283 | Almost all pages tracked by this memcg will be unmapped and freed. Some of | ||
284 | pages cannot be freed because it's locked or in-use. Such pages are moved | ||
285 | to parent and this cgroup will be empty. But this may return -EBUSY in | ||
286 | some too busy case. | ||
287 | |||
288 | Typical use case of this interface is that calling this before rmdir(). | ||
289 | Because rmdir() moves all pages to parent, some out-of-use page caches can be | ||
290 | moved to the parent. If you want to avoid that, force_empty will be useful. | ||
291 | |||
292 | 5.2 stat file | ||
293 | memory.stat file includes following statistics (now) | ||
294 | cache - # of pages from page-cache and shmem. | ||
295 | rss - # of pages from anonymous memory. | ||
296 | pgpgin - # of event of charging | ||
297 | pgpgout - # of event of uncharging | ||
298 | active_anon - # of pages on active lru of anon, shmem. | ||
299 | inactive_anon - # of pages on active lru of anon, shmem | ||
300 | active_file - # of pages on active lru of file-cache | ||
301 | inactive_file - # of pages on inactive lru of file cache | ||
302 | unevictable - # of pages cannot be reclaimed.(mlocked etc) | ||
303 | |||
304 | Below is depend on CONFIG_DEBUG_VM. | ||
305 | inactive_ratio - VM inernal parameter. (see mm/page_alloc.c) | ||
306 | recent_rotated_anon - VM internal parameter. (see mm/vmscan.c) | ||
307 | recent_rotated_file - VM internal parameter. (see mm/vmscan.c) | ||
308 | recent_scanned_anon - VM internal parameter. (see mm/vmscan.c) | ||
309 | recent_scanned_file - VM internal parameter. (see mm/vmscan.c) | ||
310 | |||
311 | Memo: | ||
312 | recent_rotated means recent frequency of lru rotation. | ||
313 | recent_scanned means recent # of scans to lru. | ||
314 | showing for better debug please see the code for meanings. | ||
315 | |||
316 | |||
317 | 5.3 swappiness | ||
318 | Similar to /proc/sys/vm/swappiness, but affecting a hierarchy of groups only. | ||
319 | |||
320 | Following cgroup's swapiness can't be changed. | ||
321 | - root cgroup (uses /proc/sys/vm/swappiness). | ||
322 | - a cgroup which uses hierarchy and it has child cgroup. | ||
323 | - a cgroup which uses hierarchy and not the root of hierarchy. | ||
324 | |||
325 | |||
326 | 6. Hierarchy support | ||
327 | |||
328 | The memory controller supports a deep hierarchy and hierarchical accounting. | ||
329 | The hierarchy is created by creating the appropriate cgroups in the | ||
330 | cgroup filesystem. Consider for example, the following cgroup filesystem | ||
331 | hierarchy | ||
332 | |||
333 | root | ||
334 | / | \ | ||
335 | / | \ | ||
336 | a b c | ||
337 | | \ | ||
338 | | \ | ||
339 | d e | ||
340 | |||
341 | In the diagram above, with hierarchical accounting enabled, all memory | ||
342 | usage of e, is accounted to its ancestors up until the root (i.e, c and root), | ||
343 | that has memory.use_hierarchy enabled. If one of the ancestors goes over its | ||
344 | limit, the reclaim algorithm reclaims from the tasks in the ancestor and the | ||
345 | children of the ancestor. | ||
346 | |||
347 | 6.1 Enabling hierarchical accounting and reclaim | ||
348 | |||
349 | The memory controller by default disables the hierarchy feature. Support | ||
350 | can be enabled by writing 1 to memory.use_hierarchy file of the root cgroup | ||
351 | |||
352 | # echo 1 > memory.use_hierarchy | ||
353 | |||
354 | The feature can be disabled by | ||
355 | |||
356 | # echo 0 > memory.use_hierarchy | ||
357 | |||
358 | NOTE1: Enabling/disabling will fail if the cgroup already has other | ||
359 | cgroups created below it. | ||
360 | |||
361 | NOTE2: This feature can be enabled/disabled per subtree. | ||
247 | 362 | ||
248 | 5. TODO | 363 | 7. TODO |
249 | 364 | ||
250 | 1. Add support for accounting huge pages (as a separate controller) | 365 | 1. Add support for accounting huge pages (as a separate controller) |
251 | 2. Make per-cgroup scanner reclaim not-shared pages first | 366 | 2. Make per-cgroup scanner reclaim not-shared pages first |
diff --git a/Documentation/hwmon/abituguru-datasheet b/Documentation/hwmon/abituguru-datasheet index 4d184f2db0ea..d9251efdcec7 100644 --- a/Documentation/hwmon/abituguru-datasheet +++ b/Documentation/hwmon/abituguru-datasheet | |||
@@ -121,7 +121,7 @@ Once all bytes have been read data will hold 0x09, but there is no reason to | |||
121 | test for this. Notice that the number of bytes is bank address dependent see | 121 | test for this. Notice that the number of bytes is bank address dependent see |
122 | above and below. | 122 | above and below. |
123 | 123 | ||
124 | After completing a successfull read it is advised to put the uGuru back in | 124 | After completing a successful read it is advised to put the uGuru back in |
125 | ready mode, so that it is ready for the next read / write cycle. This way | 125 | ready mode, so that it is ready for the next read / write cycle. This way |
126 | if your program / driver is unloaded and later loaded again the detection | 126 | if your program / driver is unloaded and later loaded again the detection |
127 | algorithm described above will still work. | 127 | algorithm described above will still work. |
@@ -141,7 +141,7 @@ don't ask why this is the way it is. | |||
141 | 141 | ||
142 | Once DATA holds 0x01 read CMD it should hold 0xAC now. | 142 | Once DATA holds 0x01 read CMD it should hold 0xAC now. |
143 | 143 | ||
144 | After completing a successfull write it is advised to put the uGuru back in | 144 | After completing a successful write it is advised to put the uGuru back in |
145 | ready mode, so that it is ready for the next read / write cycle. This way | 145 | ready mode, so that it is ready for the next read / write cycle. This way |
146 | if your program / driver is unloaded and later loaded again the detection | 146 | if your program / driver is unloaded and later loaded again the detection |
147 | algorithm described above will still work. | 147 | algorithm described above will still work. |
diff --git a/Documentation/kernel-parameters.txt b/Documentation/kernel-parameters.txt index 532eacbbed62..fb849020aea9 100644 --- a/Documentation/kernel-parameters.txt +++ b/Documentation/kernel-parameters.txt | |||
@@ -1562,6 +1562,9 @@ and is between 256 and 4096 characters. It is defined in the file | |||
1562 | 1562 | ||
1563 | nosoftlockup [KNL] Disable the soft-lockup detector. | 1563 | nosoftlockup [KNL] Disable the soft-lockup detector. |
1564 | 1564 | ||
1565 | noswapaccount [KNL] Disable accounting of swap in memory resource | ||
1566 | controller. (See Documentation/controllers/memory.txt) | ||
1567 | |||
1565 | nosync [HW,M68K] Disables sync negotiation for all devices. | 1568 | nosync [HW,M68K] Disables sync negotiation for all devices. |
1566 | 1569 | ||
1567 | notsc [BUGS=X86-32] Disable Time Stamp Counter | 1570 | notsc [BUGS=X86-32] Disable Time Stamp Counter |
diff --git a/Documentation/powerpc/dts-bindings/fsl/board.txt b/Documentation/powerpc/dts-bindings/fsl/board.txt index 81a917ef96e9..6c974d28eeb4 100644 --- a/Documentation/powerpc/dts-bindings/fsl/board.txt +++ b/Documentation/powerpc/dts-bindings/fsl/board.txt | |||
@@ -18,7 +18,7 @@ This is the memory-mapped registers for on board FPGA. | |||
18 | 18 | ||
19 | Required properities: | 19 | Required properities: |
20 | - compatible : should be "fsl,fpga-pixis". | 20 | - compatible : should be "fsl,fpga-pixis". |
21 | - reg : should contain the address and the lenght of the FPPGA register | 21 | - reg : should contain the address and the length of the FPPGA register |
22 | set. | 22 | set. |
23 | 23 | ||
24 | Example (MPC8610HPCD): | 24 | Example (MPC8610HPCD): |
@@ -27,3 +27,33 @@ Example (MPC8610HPCD): | |||
27 | compatible = "fsl,fpga-pixis"; | 27 | compatible = "fsl,fpga-pixis"; |
28 | reg = <0xe8000000 32>; | 28 | reg = <0xe8000000 32>; |
29 | }; | 29 | }; |
30 | |||
31 | * Freescale BCSR GPIO banks | ||
32 | |||
33 | Some BCSR registers act as simple GPIO controllers, each such | ||
34 | register can be represented by the gpio-controller node. | ||
35 | |||
36 | Required properities: | ||
37 | - compatible : Should be "fsl,<board>-bcsr-gpio". | ||
38 | - reg : Should contain the address and the length of the GPIO bank | ||
39 | register. | ||
40 | - #gpio-cells : Should be two. The first cell is the pin number and the | ||
41 | second cell is used to specify optional paramters (currently unused). | ||
42 | - gpio-controller : Marks the port as GPIO controller. | ||
43 | |||
44 | Example: | ||
45 | |||
46 | bcsr@1,0 { | ||
47 | #address-cells = <1>; | ||
48 | #size-cells = <1>; | ||
49 | compatible = "fsl,mpc8360mds-bcsr"; | ||
50 | reg = <1 0 0x8000>; | ||
51 | ranges = <0 1 0 0x8000>; | ||
52 | |||
53 | bcsr13: gpio-controller@d { | ||
54 | #gpio-cells = <2>; | ||
55 | compatible = "fsl,mpc8360mds-bcsr-gpio"; | ||
56 | reg = <0xd 1>; | ||
57 | gpio-controller; | ||
58 | }; | ||
59 | }; | ||
diff --git a/Documentation/scsi/scsi_fc_transport.txt b/Documentation/scsi/scsi_fc_transport.txt index 38d324d62b25..e5b071d46619 100644 --- a/Documentation/scsi/scsi_fc_transport.txt +++ b/Documentation/scsi/scsi_fc_transport.txt | |||
@@ -191,7 +191,7 @@ Vport States: | |||
191 | This is equivalent to a driver "attach" on an adapter, which is | 191 | This is equivalent to a driver "attach" on an adapter, which is |
192 | independent of the adapter's link state. | 192 | independent of the adapter's link state. |
193 | - Instantiation of the vport on the FC link via ELS traffic, etc. | 193 | - Instantiation of the vport on the FC link via ELS traffic, etc. |
194 | This is equivalent to a "link up" and successfull link initialization. | 194 | This is equivalent to a "link up" and successful link initialization. |
195 | Further information can be found in the interfaces section below for | 195 | Further information can be found in the interfaces section below for |
196 | Vport Creation. | 196 | Vport Creation. |
197 | 197 | ||
@@ -320,7 +320,7 @@ Vport Creation: | |||
320 | This is equivalent to a driver "attach" on an adapter, which is | 320 | This is equivalent to a driver "attach" on an adapter, which is |
321 | independent of the adapter's link state. | 321 | independent of the adapter's link state. |
322 | - Instantiation of the vport on the FC link via ELS traffic, etc. | 322 | - Instantiation of the vport on the FC link via ELS traffic, etc. |
323 | This is equivalent to a "link up" and successfull link initialization. | 323 | This is equivalent to a "link up" and successful link initialization. |
324 | 324 | ||
325 | The LLDD's vport_create() function will not synchronously wait for both | 325 | The LLDD's vport_create() function will not synchronously wait for both |
326 | parts to be fully completed before returning. It must validate that the | 326 | parts to be fully completed before returning. It must validate that the |
diff --git a/Documentation/w1/masters/00-INDEX b/Documentation/w1/masters/00-INDEX index 7b0ceaaad7af..d63fa024ac05 100644 --- a/Documentation/w1/masters/00-INDEX +++ b/Documentation/w1/masters/00-INDEX | |||
@@ -4,5 +4,7 @@ ds2482 | |||
4 | - The Maxim/Dallas Semiconductor DS2482 provides 1-wire busses. | 4 | - The Maxim/Dallas Semiconductor DS2482 provides 1-wire busses. |
5 | ds2490 | 5 | ds2490 |
6 | - The Maxim/Dallas Semiconductor DS2490 builds USB <-> W1 bridges. | 6 | - The Maxim/Dallas Semiconductor DS2490 builds USB <-> W1 bridges. |
7 | mxc_w1 | ||
8 | - W1 master controller driver found on Freescale MX2/MX3 SoCs | ||
7 | w1-gpio | 9 | w1-gpio |
8 | - GPIO 1-wire bus master driver. | 10 | - GPIO 1-wire bus master driver. |
diff --git a/Documentation/w1/masters/mxc-w1 b/Documentation/w1/masters/mxc-w1 new file mode 100644 index 000000000000..97f6199a7f39 --- /dev/null +++ b/Documentation/w1/masters/mxc-w1 | |||
@@ -0,0 +1,11 @@ | |||
1 | Kernel driver mxc_w1 | ||
2 | ==================== | ||
3 | |||
4 | Supported chips: | ||
5 | * Freescale MX27, MX31 and probably other i.MX SoCs | ||
6 | Datasheets: | ||
7 | http://www.freescale.com/files/32bit/doc/data_sheet/MCIMX31.pdf?fpsp=1 | ||
8 | http://www.freescale.com/files/dsp/MCIMX27.pdf?fpsp=1 | ||
9 | |||
10 | Author: Originally based on Freescale code, prepared for mainline by | ||
11 | Sascha Hauer <s.hauer@pengutronix.de> | ||
diff --git a/Documentation/w1/w1.netlink b/Documentation/w1/w1.netlink index 3640c7c87d45..804445f745ed 100644 --- a/Documentation/w1/w1.netlink +++ b/Documentation/w1/w1.netlink | |||
@@ -5,69 +5,157 @@ Message types. | |||
5 | ============= | 5 | ============= |
6 | 6 | ||
7 | There are three types of messages between w1 core and userspace: | 7 | There are three types of messages between w1 core and userspace: |
8 | 1. Events. They are generated each time new master or slave device found | 8 | 1. Events. They are generated each time new master or slave device |
9 | either due to automatic or requested search. | 9 | found either due to automatic or requested search. |
10 | 2. Userspace commands. Includes read/write and search/alarm search comamnds. | 10 | 2. Userspace commands. |
11 | 3. Replies to userspace commands. | 11 | 3. Replies to userspace commands. |
12 | 12 | ||
13 | 13 | ||
14 | Protocol. | 14 | Protocol. |
15 | ======== | 15 | ======== |
16 | 16 | ||
17 | [struct cn_msg] - connector header. It's length field is equal to size of the attached data. | 17 | [struct cn_msg] - connector header. |
18 | Its length field is equal to size of the attached data | ||
18 | [struct w1_netlink_msg] - w1 netlink header. | 19 | [struct w1_netlink_msg] - w1 netlink header. |
19 | __u8 type - message type. | 20 | __u8 type - message type. |
20 | W1_SLAVE_ADD/W1_SLAVE_REMOVE - slave add/remove events. | 21 | W1_LIST_MASTERS |
21 | W1_MASTER_ADD/W1_MASTER_REMOVE - master add/remove events. | 22 | list current bus masters |
22 | W1_MASTER_CMD - userspace command for bus master device (search/alarm search). | 23 | W1_SLAVE_ADD/W1_SLAVE_REMOVE |
23 | W1_SLAVE_CMD - userspace command for slave device (read/write/ search/alarm search | 24 | slave add/remove events |
24 | for bus master device where given slave device found). | 25 | W1_MASTER_ADD/W1_MASTER_REMOVE |
26 | master add/remove events | ||
27 | W1_MASTER_CMD | ||
28 | userspace command for bus master | ||
29 | device (search/alarm search) | ||
30 | W1_SLAVE_CMD | ||
31 | userspace command for slave device | ||
32 | (read/write/touch) | ||
25 | __u8 res - reserved | 33 | __u8 res - reserved |
26 | __u16 len - size of attached to this header data. | 34 | __u16 len - size of data attached to this header data |
27 | union { | 35 | union { |
28 | __u8 id; - slave unique device id | 36 | __u8 id[8]; - slave unique device id |
29 | struct w1_mst { | 37 | struct w1_mst { |
30 | __u32 id; - master's id. | 38 | __u32 id; - master's id |
31 | __u32 res; - reserved | 39 | __u32 res; - reserved |
32 | } mst; | 40 | } mst; |
33 | } id; | 41 | } id; |
34 | 42 | ||
35 | [strucrt w1_netlink_cmd] - command for gived master or slave device. | 43 | [struct w1_netlink_cmd] - command for given master or slave device. |
36 | __u8 cmd - command opcode. | 44 | __u8 cmd - command opcode. |
37 | W1_CMD_READ - read command. | 45 | W1_CMD_READ - read command |
38 | W1_CMD_WRITE - write command. | 46 | W1_CMD_WRITE - write command |
39 | W1_CMD_SEARCH - search command. | 47 | W1_CMD_TOUCH - touch command |
40 | W1_CMD_ALARM_SEARCH - alarm search command. | 48 | (write and sample data back to userspace) |
49 | W1_CMD_SEARCH - search command | ||
50 | W1_CMD_ALARM_SEARCH - alarm search command | ||
41 | __u8 res - reserved | 51 | __u8 res - reserved |
42 | __u16 len - length of data for this command. | 52 | __u16 len - length of data for this command |
43 | For read command data must be allocated like for write command. | 53 | For read command data must be allocated like for write command |
44 | __u8 data[0] - data for this command. | 54 | __u8 data[0] - data for this command |
45 | 55 | ||
46 | 56 | ||
47 | Each connector message can include one or more w1_netlink_msg with zero of more attached w1_netlink_cmd messages. | 57 | Each connector message can include one or more w1_netlink_msg with |
58 | zero or more attached w1_netlink_cmd messages. | ||
48 | 59 | ||
49 | For event messages there are no w1_netlink_cmd embedded structures, only connector header | 60 | For event messages there are no w1_netlink_cmd embedded structures, |
50 | and w1_netlink_msg strucutre with "len" field being zero and filled type (one of event types) | 61 | only connector header and w1_netlink_msg strucutre with "len" field |
51 | and id - either 8 bytes of slave unique id in host order, or master's id, which is assigned | 62 | being zero and filled type (one of event types) and id: |
52 | to bus master device when it is added to w1 core. | 63 | either 8 bytes of slave unique id in host order, |
64 | or master's id, which is assigned to bus master device | ||
65 | when it is added to w1 core. | ||
66 | |||
67 | Currently replies to userspace commands are only generated for read | ||
68 | command request. One reply is generated exactly for one w1_netlink_cmd | ||
69 | read request. Replies are not combined when sent - i.e. typical reply | ||
70 | messages looks like the following: | ||
53 | 71 | ||
54 | Currently replies to userspace commands are only generated for read command request. | ||
55 | One reply is generated exactly for one w1_netlink_cmd read request. | ||
56 | Replies are not combined when sent - i.e. typical reply messages looks like the following: | ||
57 | [cn_msg][w1_netlink_msg][w1_netlink_cmd] | 72 | [cn_msg][w1_netlink_msg][w1_netlink_cmd] |
58 | cn_msg.len = sizeof(struct w1_netlink_msg) + sizeof(struct w1_netlink_cmd) + cmd->len; | 73 | cn_msg.len = sizeof(struct w1_netlink_msg) + |
74 | sizeof(struct w1_netlink_cmd) + | ||
75 | cmd->len; | ||
59 | w1_netlink_msg.len = sizeof(struct w1_netlink_cmd) + cmd->len; | 76 | w1_netlink_msg.len = sizeof(struct w1_netlink_cmd) + cmd->len; |
60 | w1_netlink_cmd.len = cmd->len; | 77 | w1_netlink_cmd.len = cmd->len; |
61 | 78 | ||
79 | Replies to W1_LIST_MASTERS should send a message back to the userspace | ||
80 | which will contain list of all registered master ids in the following | ||
81 | format: | ||
82 | |||
83 | cn_msg (CN_W1_IDX.CN_W1_VAL as id, len is equal to sizeof(struct | ||
84 | w1_netlink_msg) plus number of masters multipled by 4) | ||
85 | w1_netlink_msg (type: W1_LIST_MASTERS, len is equal to | ||
86 | number of masters multiplied by 4 (u32 size)) | ||
87 | id0 ... idN | ||
88 | |||
89 | Each message is at most 4k in size, so if number of master devices | ||
90 | exceeds this, it will be split into several messages, | ||
91 | cn.seq will be increased for each one. | ||
92 | |||
93 | W1 search and alarm search commands. | ||
94 | request: | ||
95 | [cn_msg] | ||
96 | [w1_netlink_msg type = W1_MASTER_CMD | ||
97 | id is equal to the bus master id to use for searching] | ||
98 | [w1_netlink_cmd cmd = W1_CMD_SEARCH or W1_CMD_ALARM_SEARCH] | ||
99 | |||
100 | reply: | ||
101 | [cn_msg, ack = 1 and increasing, 0 means the last message, | ||
102 | seq is equal to the request seq] | ||
103 | [w1_netlink_msg type = W1_MASTER_CMD] | ||
104 | [w1_netlink_cmd cmd = W1_CMD_SEARCH or W1_CMD_ALARM_SEARCH | ||
105 | len is equal to number of IDs multiplied by 8] | ||
106 | [64bit-id0 ... 64bit-idN] | ||
107 | Length in each header corresponds to the size of the data behind it, so | ||
108 | w1_netlink_cmd->len = N * 8; where N is number of IDs in this message. | ||
109 | Can be zero. | ||
110 | w1_netlink_msg->len = sizeof(struct w1_netlink_cmd) + N * 8; | ||
111 | cn_msg->len = sizeof(struct w1_netlink_msg) + | ||
112 | sizeof(struct w1_netlink_cmd) + | ||
113 | N*8; | ||
114 | |||
115 | W1 reset command. | ||
116 | [cn_msg] | ||
117 | [w1_netlink_msg type = W1_MASTER_CMD | ||
118 | id is equal to the bus master id to use for searching] | ||
119 | [w1_netlink_cmd cmd = W1_CMD_RESET] | ||
120 | |||
121 | |||
122 | Command status replies. | ||
123 | ====================== | ||
124 | |||
125 | Each command (either root, master or slave with or without w1_netlink_cmd | ||
126 | structure) will be 'acked' by the w1 core. Format of the reply is the same | ||
127 | as request message except that length parameters do not account for data | ||
128 | requested by the user, i.e. read/write/touch IO requests will not contain | ||
129 | data, so w1_netlink_cmd.len will be 0, w1_netlink_msg.len will be size | ||
130 | of the w1_netlink_cmd structure and cn_msg.len will be equal to the sum | ||
131 | of the sizeof(struct w1_netlink_msg) and sizeof(struct w1_netlink_cmd). | ||
132 | If reply is generated for master or root command (which do not have | ||
133 | w1_netlink_cmd attached), reply will contain only cn_msg and w1_netlink_msg | ||
134 | structires. | ||
135 | |||
136 | w1_netlink_msg.status field will carry positive error value | ||
137 | (EINVAL for example) or zero in case of success. | ||
138 | |||
139 | All other fields in every structure will mirror the same parameters in the | ||
140 | request message (except lengths as described above). | ||
141 | |||
142 | Status reply is generated for every w1_netlink_cmd embedded in the | ||
143 | w1_netlink_msg, if there are no w1_netlink_cmd structures, | ||
144 | reply will be generated for the w1_netlink_msg. | ||
145 | |||
146 | All w1_netlink_cmd command structures are handled in every w1_netlink_msg, | ||
147 | even if there were errors, only length mismatch interrupts message processing. | ||
148 | |||
62 | 149 | ||
63 | Operation steps in w1 core when new command is received. | 150 | Operation steps in w1 core when new command is received. |
64 | ======================================================= | 151 | ======================================================= |
65 | 152 | ||
66 | When new message (w1_netlink_msg) is received w1 core detects if it is master of slave request, | 153 | When new message (w1_netlink_msg) is received w1 core detects if it is |
67 | according to w1_netlink_msg.type field. | 154 | master or slave request, according to w1_netlink_msg.type field. |
68 | Then master or slave device is searched for. | 155 | Then master or slave device is searched for. |
69 | When found, master device (requested or those one on where slave device is found) is locked. | 156 | When found, master device (requested or those one on where slave device |
70 | If slave command is requested, then reset/select procedure is started to select given device. | 157 | is found) is locked. If slave command is requested, then reset/select |
158 | procedure is started to select given device. | ||
71 | 159 | ||
72 | Then all requested in w1_netlink_msg operations are performed one by one. | 160 | Then all requested in w1_netlink_msg operations are performed one by one. |
73 | If command requires reply (like read command) it is sent on command completion. | 161 | If command requires reply (like read command) it is sent on command completion. |
@@ -82,8 +170,8 @@ Connector [1] specific documentation. | |||
82 | Each connector message includes two u32 fields as "address". | 170 | Each connector message includes two u32 fields as "address". |
83 | w1 uses CN_W1_IDX and CN_W1_VAL defined in include/linux/connector.h header. | 171 | w1 uses CN_W1_IDX and CN_W1_VAL defined in include/linux/connector.h header. |
84 | Each message also includes sequence and acknowledge numbers. | 172 | Each message also includes sequence and acknowledge numbers. |
85 | Sequence number for event messages is appropriate bus master sequence number increased with | 173 | Sequence number for event messages is appropriate bus master sequence number |
86 | each event message sent "through" this master. | 174 | increased with each event message sent "through" this master. |
87 | Sequence number for userspace requests is set by userspace application. | 175 | Sequence number for userspace requests is set by userspace application. |
88 | Sequence number for reply is the same as was in request, and | 176 | Sequence number for reply is the same as was in request, and |
89 | acknowledge number is set to seq+1. | 177 | acknowledge number is set to seq+1. |
@@ -93,6 +181,6 @@ Additional documantion, source code examples. | |||
93 | ============================================ | 181 | ============================================ |
94 | 182 | ||
95 | 1. Documentation/connector | 183 | 1. Documentation/connector |
96 | 2. http://tservice.net.ru/~s0mbre/archive/w1 | 184 | 2. http://www.ioremap.net/archive/w1 |
97 | This archive includes userspace application w1d.c which | 185 | This archive includes userspace application w1d.c which uses |
98 | uses read/write/search commands for all master/slave devices found on the bus. | 186 | read/write/search commands for all master/slave devices found on the bus. |
diff --git a/MAINTAINERS b/MAINTAINERS index a01884407fe4..57e0309243cc 100644 --- a/MAINTAINERS +++ b/MAINTAINERS | |||
@@ -1360,6 +1360,11 @@ P: Maciej W. Rozycki | |||
1360 | M: macro@linux-mips.org | 1360 | M: macro@linux-mips.org |
1361 | S: Maintained | 1361 | S: Maintained |
1362 | 1362 | ||
1363 | DELL LAPTOP DRIVER | ||
1364 | P: Matthew Garrett | ||
1365 | M: mjg59@srcf.ucam.org | ||
1366 | S: Maintained | ||
1367 | |||
1363 | DELL LAPTOP SMM DRIVER | 1368 | DELL LAPTOP SMM DRIVER |
1364 | P: Massimo Dal Zotto | 1369 | P: Massimo Dal Zotto |
1365 | M: dz@debian.org | 1370 | M: dz@debian.org |
@@ -3484,6 +3489,12 @@ L: linuxppc-dev@ozlabs.org | |||
3484 | L: cbe-oss-dev@ozlabs.org | 3489 | L: cbe-oss-dev@ozlabs.org |
3485 | S: Supported | 3490 | S: Supported |
3486 | 3491 | ||
3492 | PS3VRAM DRIVER | ||
3493 | P: Jim Paris | ||
3494 | M: jim@jtan.com | ||
3495 | L: cbe-oss-dev@ozlabs.org | ||
3496 | S: Maintained | ||
3497 | |||
3487 | PVRUSB2 VIDEO4LINUX DRIVER | 3498 | PVRUSB2 VIDEO4LINUX DRIVER |
3488 | P: Mike Isely | 3499 | P: Mike Isely |
3489 | M: isely@pobox.com | 3500 | M: isely@pobox.com |
diff --git a/arch/arm/configs/clps7500_defconfig b/arch/arm/configs/clps7500_defconfig deleted file mode 100644 index 49e9f9d8b3d1..000000000000 --- a/arch/arm/configs/clps7500_defconfig +++ /dev/null | |||
@@ -1,801 +0,0 @@ | |||
1 | # | ||
2 | # Automatically generated make config: don't edit | ||
3 | # Linux kernel version: 2.6.12-rc1-bk2 | ||
4 | # Sun Mar 27 17:20:48 2005 | ||
5 | # | ||
6 | CONFIG_ARM=y | ||
7 | CONFIG_MMU=y | ||
8 | CONFIG_UID16=y | ||
9 | CONFIG_RWSEM_GENERIC_SPINLOCK=y | ||
10 | CONFIG_GENERIC_CALIBRATE_DELAY=y | ||
11 | CONFIG_GENERIC_IOMAP=y | ||
12 | |||
13 | # | ||
14 | # Code maturity level options | ||
15 | # | ||
16 | CONFIG_EXPERIMENTAL=y | ||
17 | CONFIG_CLEAN_COMPILE=y | ||
18 | CONFIG_BROKEN_ON_SMP=y | ||
19 | |||
20 | # | ||
21 | # General setup | ||
22 | # | ||
23 | CONFIG_LOCALVERSION="" | ||
24 | CONFIG_SWAP=y | ||
25 | CONFIG_SYSVIPC=y | ||
26 | # CONFIG_POSIX_MQUEUE is not set | ||
27 | # CONFIG_BSD_PROCESS_ACCT is not set | ||
28 | # CONFIG_SYSCTL is not set | ||
29 | # CONFIG_AUDIT is not set | ||
30 | # CONFIG_HOTPLUG is not set | ||
31 | CONFIG_KOBJECT_UEVENT=y | ||
32 | # CONFIG_IKCONFIG is not set | ||
33 | CONFIG_EMBEDDED=y | ||
34 | CONFIG_KALLSYMS=y | ||
35 | # CONFIG_KALLSYMS_EXTRA_PASS is not set | ||
36 | CONFIG_BASE_FULL=y | ||
37 | CONFIG_FUTEX=y | ||
38 | CONFIG_EPOLL=y | ||
39 | CONFIG_CC_OPTIMIZE_FOR_SIZE=y | ||
40 | CONFIG_SHMEM=y | ||
41 | CONFIG_CC_ALIGN_FUNCTIONS=0 | ||
42 | CONFIG_CC_ALIGN_LABELS=0 | ||
43 | CONFIG_CC_ALIGN_LOOPS=0 | ||
44 | CONFIG_CC_ALIGN_JUMPS=0 | ||
45 | # CONFIG_TINY_SHMEM is not set | ||
46 | CONFIG_BASE_SMALL=0 | ||
47 | |||
48 | # | ||
49 | # Loadable module support | ||
50 | # | ||
51 | # CONFIG_MODULES is not set | ||
52 | |||
53 | # | ||
54 | # System Type | ||
55 | # | ||
56 | CONFIG_ARCH_CLPS7500=y | ||
57 | # CONFIG_ARCH_CLPS711X is not set | ||
58 | # CONFIG_ARCH_CO285 is not set | ||
59 | # CONFIG_ARCH_EBSA110 is not set | ||
60 | # CONFIG_ARCH_FOOTBRIDGE is not set | ||
61 | # CONFIG_ARCH_INTEGRATOR is not set | ||
62 | # CONFIG_ARCH_IOP3XX is not set | ||
63 | # CONFIG_ARCH_IXP4XX is not set | ||
64 | # CONFIG_ARCH_IXP2000 is not set | ||
65 | # CONFIG_ARCH_L7200 is not set | ||
66 | # CONFIG_ARCH_PXA is not set | ||
67 | # CONFIG_ARCH_RPC is not set | ||
68 | # CONFIG_ARCH_SA1100 is not set | ||
69 | # CONFIG_ARCH_S3C2410 is not set | ||
70 | # CONFIG_ARCH_SHARK is not set | ||
71 | # CONFIG_ARCH_LH7A40X is not set | ||
72 | # CONFIG_ARCH_OMAP is not set | ||
73 | # CONFIG_ARCH_VERSATILE is not set | ||
74 | # CONFIG_ARCH_IMX is not set | ||
75 | # CONFIG_ARCH_H720X is not set | ||
76 | |||
77 | # | ||
78 | # Processor Type | ||
79 | # | ||
80 | CONFIG_CPU_32=y | ||
81 | CONFIG_CPU_ARM710=y | ||
82 | CONFIG_CPU_32v3=y | ||
83 | CONFIG_CPU_CACHE_V3=y | ||
84 | CONFIG_CPU_CACHE_VIVT=y | ||
85 | CONFIG_CPU_COPY_V3=y | ||
86 | CONFIG_CPU_TLB_V3=y | ||
87 | |||
88 | # | ||
89 | # Processor Features | ||
90 | # | ||
91 | CONFIG_TIMER_ACORN=y | ||
92 | |||
93 | # | ||
94 | # Bus support | ||
95 | # | ||
96 | CONFIG_ISA=y | ||
97 | |||
98 | # | ||
99 | # PCCARD (PCMCIA/CardBus) support | ||
100 | # | ||
101 | # CONFIG_PCCARD is not set | ||
102 | |||
103 | # | ||
104 | # Kernel Features | ||
105 | # | ||
106 | # CONFIG_PREEMPT is not set | ||
107 | CONFIG_ALIGNMENT_TRAP=y | ||
108 | |||
109 | # | ||
110 | # Boot options | ||
111 | # | ||
112 | CONFIG_ZBOOT_ROM_TEXT=0x0 | ||
113 | CONFIG_ZBOOT_ROM_BSS=0x0 | ||
114 | CONFIG_CMDLINE="mem=16M root=nfs" | ||
115 | # CONFIG_XIP_KERNEL is not set | ||
116 | |||
117 | # | ||
118 | # Floating point emulation | ||
119 | # | ||
120 | |||
121 | # | ||
122 | # At least one emulation must be selected | ||
123 | # | ||
124 | # CONFIG_FPE_NWFPE is not set | ||
125 | |||
126 | # | ||
127 | # Userspace binary formats | ||
128 | # | ||
129 | CONFIG_BINFMT_ELF=y | ||
130 | # CONFIG_BINFMT_AOUT is not set | ||
131 | # CONFIG_BINFMT_MISC is not set | ||
132 | # CONFIG_ARTHUR is not set | ||
133 | |||
134 | # | ||
135 | # Power management options | ||
136 | # | ||
137 | # CONFIG_PM is not set | ||
138 | |||
139 | # | ||
140 | # Device Drivers | ||
141 | # | ||
142 | |||
143 | # | ||
144 | # Generic Driver Options | ||
145 | # | ||
146 | CONFIG_STANDALONE=y | ||
147 | CONFIG_PREVENT_FIRMWARE_BUILD=y | ||
148 | # CONFIG_FW_LOADER is not set | ||
149 | |||
150 | # | ||
151 | # Memory Technology Devices (MTD) | ||
152 | # | ||
153 | CONFIG_MTD=y | ||
154 | # CONFIG_MTD_DEBUG is not set | ||
155 | # CONFIG_MTD_CONCAT is not set | ||
156 | # CONFIG_MTD_PARTITIONS is not set | ||
157 | |||
158 | # | ||
159 | # User Modules And Translation Layers | ||
160 | # | ||
161 | # CONFIG_MTD_CHAR is not set | ||
162 | # CONFIG_MTD_BLOCK is not set | ||
163 | # CONFIG_MTD_BLOCK_RO is not set | ||
164 | # CONFIG_FTL is not set | ||
165 | # CONFIG_NFTL is not set | ||
166 | # CONFIG_INFTL is not set | ||
167 | |||
168 | # | ||
169 | # RAM/ROM/Flash chip drivers | ||
170 | # | ||
171 | # CONFIG_MTD_CFI is not set | ||
172 | # CONFIG_MTD_JEDECPROBE is not set | ||
173 | CONFIG_MTD_MAP_BANK_WIDTH_1=y | ||
174 | CONFIG_MTD_MAP_BANK_WIDTH_2=y | ||
175 | CONFIG_MTD_MAP_BANK_WIDTH_4=y | ||
176 | # CONFIG_MTD_MAP_BANK_WIDTH_8 is not set | ||
177 | # CONFIG_MTD_MAP_BANK_WIDTH_16 is not set | ||
178 | # CONFIG_MTD_MAP_BANK_WIDTH_32 is not set | ||
179 | CONFIG_MTD_CFI_I1=y | ||
180 | CONFIG_MTD_CFI_I2=y | ||
181 | # CONFIG_MTD_CFI_I4 is not set | ||
182 | # CONFIG_MTD_CFI_I8 is not set | ||
183 | # CONFIG_MTD_RAM is not set | ||
184 | # CONFIG_MTD_ROM is not set | ||
185 | # CONFIG_MTD_ABSENT is not set | ||
186 | |||
187 | # | ||
188 | # Mapping drivers for chip access | ||
189 | # | ||
190 | # CONFIG_MTD_COMPLEX_MAPPINGS is not set | ||
191 | |||
192 | # | ||
193 | # Self-contained MTD device drivers | ||
194 | # | ||
195 | # CONFIG_MTD_SLRAM is not set | ||
196 | # CONFIG_MTD_PHRAM is not set | ||
197 | # CONFIG_MTD_MTDRAM is not set | ||
198 | # CONFIG_MTD_BLKMTD is not set | ||
199 | # CONFIG_MTD_BLOCK2MTD is not set | ||
200 | |||
201 | # | ||
202 | # Disk-On-Chip Device Drivers | ||
203 | # | ||
204 | # CONFIG_MTD_DOC2000 is not set | ||
205 | # CONFIG_MTD_DOC2001 is not set | ||
206 | # CONFIG_MTD_DOC2001PLUS is not set | ||
207 | |||
208 | # | ||
209 | # NAND Flash Device Drivers | ||
210 | # | ||
211 | # CONFIG_MTD_NAND is not set | ||
212 | |||
213 | # | ||
214 | # Parallel port support | ||
215 | # | ||
216 | CONFIG_PARPORT=y | ||
217 | CONFIG_PARPORT_PC=y | ||
218 | CONFIG_PARPORT_PC_FIFO=y | ||
219 | # CONFIG_PARPORT_PC_SUPERIO is not set | ||
220 | # CONFIG_PARPORT_ARC is not set | ||
221 | # CONFIG_PARPORT_GSC is not set | ||
222 | CONFIG_PARPORT_1284=y | ||
223 | |||
224 | # | ||
225 | # Plug and Play support | ||
226 | # | ||
227 | # CONFIG_PNP is not set | ||
228 | |||
229 | # | ||
230 | # Block devices | ||
231 | # | ||
232 | # CONFIG_BLK_DEV_FD is not set | ||
233 | # CONFIG_BLK_DEV_XD is not set | ||
234 | # CONFIG_PARIDE is not set | ||
235 | # CONFIG_BLK_DEV_COW_COMMON is not set | ||
236 | # CONFIG_BLK_DEV_LOOP is not set | ||
237 | CONFIG_BLK_DEV_NBD=y | ||
238 | CONFIG_BLK_DEV_RAM=y | ||
239 | CONFIG_BLK_DEV_RAM_COUNT=16 | ||
240 | CONFIG_BLK_DEV_RAM_SIZE=4096 | ||
241 | # CONFIG_BLK_DEV_INITRD is not set | ||
242 | CONFIG_INITRAMFS_SOURCE="" | ||
243 | # CONFIG_CDROM_PKTCDVD is not set | ||
244 | |||
245 | # | ||
246 | # IO Schedulers | ||
247 | # | ||
248 | CONFIG_IOSCHED_NOOP=y | ||
249 | CONFIG_IOSCHED_AS=y | ||
250 | CONFIG_IOSCHED_DEADLINE=y | ||
251 | CONFIG_IOSCHED_CFQ=y | ||
252 | # CONFIG_ATA_OVER_ETH is not set | ||
253 | |||
254 | # | ||
255 | # ATA/ATAPI/MFM/RLL support | ||
256 | # | ||
257 | # CONFIG_IDE is not set | ||
258 | |||
259 | # | ||
260 | # SCSI device support | ||
261 | # | ||
262 | # CONFIG_SCSI is not set | ||
263 | |||
264 | # | ||
265 | # Multi-device support (RAID and LVM) | ||
266 | # | ||
267 | # CONFIG_MD is not set | ||
268 | |||
269 | # | ||
270 | # Fusion MPT device support | ||
271 | # | ||
272 | |||
273 | # | ||
274 | # IEEE 1394 (FireWire) support | ||
275 | # | ||
276 | |||
277 | # | ||
278 | # I2O device support | ||
279 | # | ||
280 | |||
281 | # | ||
282 | # Networking support | ||
283 | # | ||
284 | CONFIG_NET=y | ||
285 | |||
286 | # | ||
287 | # Networking options | ||
288 | # | ||
289 | # CONFIG_PACKET is not set | ||
290 | # CONFIG_NETLINK_DEV is not set | ||
291 | CONFIG_UNIX=y | ||
292 | # CONFIG_NET_KEY is not set | ||
293 | CONFIG_INET=y | ||
294 | # CONFIG_IP_MULTICAST is not set | ||
295 | # CONFIG_IP_ADVANCED_ROUTER is not set | ||
296 | CONFIG_IP_PNP=y | ||
297 | # CONFIG_IP_PNP_DHCP is not set | ||
298 | CONFIG_IP_PNP_BOOTP=y | ||
299 | # CONFIG_IP_PNP_RARP is not set | ||
300 | # CONFIG_NET_IPIP is not set | ||
301 | # CONFIG_NET_IPGRE is not set | ||
302 | # CONFIG_ARPD is not set | ||
303 | # CONFIG_SYN_COOKIES is not set | ||
304 | # CONFIG_INET_AH is not set | ||
305 | # CONFIG_INET_ESP is not set | ||
306 | # CONFIG_INET_IPCOMP is not set | ||
307 | # CONFIG_INET_TUNNEL is not set | ||
308 | CONFIG_IP_TCPDIAG=y | ||
309 | # CONFIG_IP_TCPDIAG_IPV6 is not set | ||
310 | # CONFIG_IPV6 is not set | ||
311 | # CONFIG_NETFILTER is not set | ||
312 | |||
313 | # | ||
314 | # SCTP Configuration (EXPERIMENTAL) | ||
315 | # | ||
316 | # CONFIG_IP_SCTP is not set | ||
317 | # CONFIG_ATM is not set | ||
318 | # CONFIG_BRIDGE is not set | ||
319 | # CONFIG_VLAN_8021Q is not set | ||
320 | # CONFIG_DECNET is not set | ||
321 | # CONFIG_LLC2 is not set | ||
322 | # CONFIG_IPX is not set | ||
323 | # CONFIG_ATALK is not set | ||
324 | # CONFIG_X25 is not set | ||
325 | # CONFIG_LAPB is not set | ||
326 | # CONFIG_NET_DIVERT is not set | ||
327 | # CONFIG_ECONET is not set | ||
328 | # CONFIG_WAN_ROUTER is not set | ||
329 | |||
330 | # | ||
331 | # QoS and/or fair queueing | ||
332 | # | ||
333 | # CONFIG_NET_SCHED is not set | ||
334 | # CONFIG_NET_CLS_ROUTE is not set | ||
335 | |||
336 | # | ||
337 | # Network testing | ||
338 | # | ||
339 | # CONFIG_NET_PKTGEN is not set | ||
340 | # CONFIG_NETPOLL is not set | ||
341 | # CONFIG_NET_POLL_CONTROLLER is not set | ||
342 | # CONFIG_HAMRADIO is not set | ||
343 | # CONFIG_IRDA is not set | ||
344 | # CONFIG_BT is not set | ||
345 | CONFIG_NETDEVICES=y | ||
346 | CONFIG_DUMMY=y | ||
347 | # CONFIG_BONDING is not set | ||
348 | # CONFIG_EQUALIZER is not set | ||
349 | # CONFIG_TUN is not set | ||
350 | |||
351 | # | ||
352 | # ARCnet devices | ||
353 | # | ||
354 | # CONFIG_ARCNET is not set | ||
355 | |||
356 | # | ||
357 | # Ethernet (10 or 100Mbit) | ||
358 | # | ||
359 | CONFIG_NET_ETHERNET=y | ||
360 | # CONFIG_MII is not set | ||
361 | # CONFIG_NET_VENDOR_3COM is not set | ||
362 | # CONFIG_LANCE is not set | ||
363 | # CONFIG_NET_VENDOR_SMC is not set | ||
364 | # CONFIG_SMC91X is not set | ||
365 | # CONFIG_NET_VENDOR_RACAL is not set | ||
366 | # CONFIG_AT1700 is not set | ||
367 | # CONFIG_DEPCA is not set | ||
368 | # CONFIG_HP100 is not set | ||
369 | # CONFIG_NET_ISA is not set | ||
370 | CONFIG_NET_PCI=y | ||
371 | # CONFIG_AC3200 is not set | ||
372 | # CONFIG_APRICOT is not set | ||
373 | CONFIG_CS89x0=y | ||
374 | # CONFIG_NET_POCKET is not set | ||
375 | |||
376 | # | ||
377 | # Ethernet (1000 Mbit) | ||
378 | # | ||
379 | |||
380 | # | ||
381 | # Ethernet (10000 Mbit) | ||
382 | # | ||
383 | |||
384 | # | ||
385 | # Token Ring devices | ||
386 | # | ||
387 | # CONFIG_TR is not set | ||
388 | |||
389 | # | ||
390 | # Wireless LAN (non-hamradio) | ||
391 | # | ||
392 | # CONFIG_NET_RADIO is not set | ||
393 | |||
394 | # | ||
395 | # Wan interfaces | ||
396 | # | ||
397 | # CONFIG_WAN is not set | ||
398 | # CONFIG_PLIP is not set | ||
399 | CONFIG_PPP=y | ||
400 | # CONFIG_PPP_MULTILINK is not set | ||
401 | # CONFIG_PPP_FILTER is not set | ||
402 | # CONFIG_PPP_ASYNC is not set | ||
403 | # CONFIG_PPP_SYNC_TTY is not set | ||
404 | # CONFIG_PPP_DEFLATE is not set | ||
405 | # CONFIG_PPP_BSDCOMP is not set | ||
406 | # CONFIG_PPPOE is not set | ||
407 | CONFIG_SLIP=y | ||
408 | CONFIG_SLIP_COMPRESSED=y | ||
409 | # CONFIG_SLIP_SMART is not set | ||
410 | # CONFIG_SLIP_MODE_SLIP6 is not set | ||
411 | # CONFIG_SHAPER is not set | ||
412 | # CONFIG_NETCONSOLE is not set | ||
413 | |||
414 | # | ||
415 | # ISDN subsystem | ||
416 | # | ||
417 | # CONFIG_ISDN is not set | ||
418 | |||
419 | # | ||
420 | # Input device support | ||
421 | # | ||
422 | CONFIG_INPUT=y | ||
423 | |||
424 | # | ||
425 | # Userland interfaces | ||
426 | # | ||
427 | CONFIG_INPUT_MOUSEDEV=y | ||
428 | CONFIG_INPUT_MOUSEDEV_PSAUX=y | ||
429 | CONFIG_INPUT_MOUSEDEV_SCREEN_X=1024 | ||
430 | CONFIG_INPUT_MOUSEDEV_SCREEN_Y=768 | ||
431 | # CONFIG_INPUT_JOYDEV is not set | ||
432 | # CONFIG_INPUT_TSDEV is not set | ||
433 | # CONFIG_INPUT_EVDEV is not set | ||
434 | # CONFIG_INPUT_EVBUG is not set | ||
435 | |||
436 | # | ||
437 | # Input Device Drivers | ||
438 | # | ||
439 | CONFIG_INPUT_KEYBOARD=y | ||
440 | CONFIG_KEYBOARD_ATKBD=y | ||
441 | # CONFIG_KEYBOARD_SUNKBD is not set | ||
442 | # CONFIG_KEYBOARD_LKKBD is not set | ||
443 | # CONFIG_KEYBOARD_XTKBD is not set | ||
444 | # CONFIG_KEYBOARD_NEWTON is not set | ||
445 | CONFIG_INPUT_MOUSE=y | ||
446 | CONFIG_MOUSE_PS2=y | ||
447 | # CONFIG_MOUSE_SERIAL is not set | ||
448 | # CONFIG_MOUSE_INPORT is not set | ||
449 | # CONFIG_MOUSE_LOGIBM is not set | ||
450 | # CONFIG_MOUSE_PC110PAD is not set | ||
451 | # CONFIG_MOUSE_VSXXXAA is not set | ||
452 | # CONFIG_INPUT_JOYSTICK is not set | ||
453 | # CONFIG_INPUT_TOUCHSCREEN is not set | ||
454 | # CONFIG_INPUT_MISC is not set | ||
455 | |||
456 | # | ||
457 | # Hardware I/O ports | ||
458 | # | ||
459 | CONFIG_SERIO=y | ||
460 | # CONFIG_SERIO_SERPORT is not set | ||
461 | # CONFIG_SERIO_PARKBD is not set | ||
462 | CONFIG_SERIO_RPCKBD=y | ||
463 | CONFIG_SERIO_LIBPS2=y | ||
464 | # CONFIG_SERIO_RAW is not set | ||
465 | # CONFIG_GAMEPORT is not set | ||
466 | CONFIG_SOUND_GAMEPORT=y | ||
467 | |||
468 | # | ||
469 | # Character devices | ||
470 | # | ||
471 | CONFIG_VT=y | ||
472 | CONFIG_VT_CONSOLE=y | ||
473 | CONFIG_HW_CONSOLE=y | ||
474 | # CONFIG_SERIAL_NONSTANDARD is not set | ||
475 | |||
476 | # | ||
477 | # Serial drivers | ||
478 | # | ||
479 | CONFIG_SERIAL_8250=y | ||
480 | CONFIG_SERIAL_8250_CONSOLE=y | ||
481 | CONFIG_SERIAL_8250_NR_UARTS=4 | ||
482 | # CONFIG_SERIAL_8250_EXTENDED is not set | ||
483 | |||
484 | # | ||
485 | # Non-8250 serial port support | ||
486 | # | ||
487 | CONFIG_SERIAL_CORE=y | ||
488 | CONFIG_SERIAL_CORE_CONSOLE=y | ||
489 | CONFIG_UNIX98_PTYS=y | ||
490 | CONFIG_LEGACY_PTYS=y | ||
491 | CONFIG_LEGACY_PTY_COUNT=256 | ||
492 | CONFIG_PRINTER=y | ||
493 | # CONFIG_LP_CONSOLE is not set | ||
494 | # CONFIG_PPDEV is not set | ||
495 | # CONFIG_TIPAR is not set | ||
496 | |||
497 | # | ||
498 | # IPMI | ||
499 | # | ||
500 | # CONFIG_IPMI_HANDLER is not set | ||
501 | |||
502 | # | ||
503 | # Watchdog Cards | ||
504 | # | ||
505 | # CONFIG_WATCHDOG is not set | ||
506 | # CONFIG_NVRAM is not set | ||
507 | # CONFIG_RTC is not set | ||
508 | # CONFIG_DTLK is not set | ||
509 | # CONFIG_R3964 is not set | ||
510 | |||
511 | # | ||
512 | # Ftape, the floppy tape device driver | ||
513 | # | ||
514 | # CONFIG_DRM is not set | ||
515 | # CONFIG_RAW_DRIVER is not set | ||
516 | |||
517 | # | ||
518 | # TPM devices | ||
519 | # | ||
520 | # CONFIG_TCG_TPM is not set | ||
521 | |||
522 | # | ||
523 | # I2C support | ||
524 | # | ||
525 | CONFIG_I2C=y | ||
526 | # CONFIG_I2C_CHARDEV is not set | ||
527 | |||
528 | # | ||
529 | # I2C Algorithms | ||
530 | # | ||
531 | CONFIG_I2C_ALGOBIT=y | ||
532 | # CONFIG_I2C_ALGOPCF is not set | ||
533 | # CONFIG_I2C_ALGOPCA is not set | ||
534 | |||
535 | # | ||
536 | # I2C Hardware Bus support | ||
537 | # | ||
538 | # CONFIG_I2C_ELEKTOR is not set | ||
539 | # CONFIG_I2C_PARPORT is not set | ||
540 | # CONFIG_I2C_PARPORT_LIGHT is not set | ||
541 | # CONFIG_I2C_PCA_ISA is not set | ||
542 | |||
543 | # | ||
544 | # Hardware Sensors Chip support | ||
545 | # | ||
546 | # CONFIG_I2C_SENSOR is not set | ||
547 | # CONFIG_SENSORS_ADM1021 is not set | ||
548 | # CONFIG_SENSORS_ADM1025 is not set | ||
549 | # CONFIG_SENSORS_ADM1026 is not set | ||
550 | # CONFIG_SENSORS_ADM1031 is not set | ||
551 | # CONFIG_SENSORS_ASB100 is not set | ||
552 | # CONFIG_SENSORS_DS1621 is not set | ||
553 | # CONFIG_SENSORS_FSCHER is not set | ||
554 | # CONFIG_SENSORS_FSCPOS is not set | ||
555 | # CONFIG_SENSORS_GL518SM is not set | ||
556 | # CONFIG_SENSORS_GL520SM is not set | ||
557 | # CONFIG_SENSORS_IT87 is not set | ||
558 | # CONFIG_SENSORS_LM63 is not set | ||
559 | # CONFIG_SENSORS_LM75 is not set | ||
560 | # CONFIG_SENSORS_LM77 is not set | ||
561 | # CONFIG_SENSORS_LM78 is not set | ||
562 | # CONFIG_SENSORS_LM80 is not set | ||
563 | # CONFIG_SENSORS_LM83 is not set | ||
564 | # CONFIG_SENSORS_LM85 is not set | ||
565 | # CONFIG_SENSORS_LM87 is not set | ||
566 | # CONFIG_SENSORS_LM90 is not set | ||
567 | # CONFIG_SENSORS_MAX1619 is not set | ||
568 | # CONFIG_SENSORS_PC87360 is not set | ||
569 | # CONFIG_SENSORS_SMSC47B397 is not set | ||
570 | # CONFIG_SENSORS_SMSC47M1 is not set | ||
571 | # CONFIG_SENSORS_W83781D is not set | ||
572 | # CONFIG_SENSORS_W83L785TS is not set | ||
573 | # CONFIG_SENSORS_W83627HF is not set | ||
574 | |||
575 | # | ||
576 | # Other I2C Chip support | ||
577 | # | ||
578 | # CONFIG_SENSORS_EEPROM is not set | ||
579 | # CONFIG_SENSORS_PCF8574 is not set | ||
580 | # CONFIG_SENSORS_PCF8591 is not set | ||
581 | # CONFIG_SENSORS_RTC8564 is not set | ||
582 | # CONFIG_I2C_DEBUG_CORE is not set | ||
583 | # CONFIG_I2C_DEBUG_ALGO is not set | ||
584 | # CONFIG_I2C_DEBUG_BUS is not set | ||
585 | # CONFIG_I2C_DEBUG_CHIP is not set | ||
586 | |||
587 | # | ||
588 | # Misc devices | ||
589 | # | ||
590 | |||
591 | # | ||
592 | # Multimedia devices | ||
593 | # | ||
594 | # CONFIG_VIDEO_DEV is not set | ||
595 | |||
596 | # | ||
597 | # Digital Video Broadcasting Devices | ||
598 | # | ||
599 | # CONFIG_DVB is not set | ||
600 | |||
601 | # | ||
602 | # Graphics support | ||
603 | # | ||
604 | CONFIG_FB=y | ||
605 | CONFIG_FB_CFB_FILLRECT=y | ||
606 | CONFIG_FB_CFB_COPYAREA=y | ||
607 | CONFIG_FB_CFB_IMAGEBLIT=y | ||
608 | CONFIG_FB_SOFT_CURSOR=y | ||
609 | # CONFIG_FB_MODE_HELPERS is not set | ||
610 | # CONFIG_FB_TILEBLITTING is not set | ||
611 | CONFIG_FB_ACORN=y | ||
612 | # CONFIG_FB_VIRTUAL is not set | ||
613 | |||
614 | # | ||
615 | # Console display driver support | ||
616 | # | ||
617 | # CONFIG_VGA_CONSOLE is not set | ||
618 | # CONFIG_MDA_CONSOLE is not set | ||
619 | CONFIG_DUMMY_CONSOLE=y | ||
620 | CONFIG_FRAMEBUFFER_CONSOLE=y | ||
621 | CONFIG_FONTS=y | ||
622 | CONFIG_FONT_8x8=y | ||
623 | CONFIG_FONT_8x16=y | ||
624 | # CONFIG_FONT_6x11 is not set | ||
625 | # CONFIG_FONT_PEARL_8x8 is not set | ||
626 | # CONFIG_FONT_ACORN_8x8 is not set | ||
627 | # CONFIG_FONT_MINI_4x6 is not set | ||
628 | # CONFIG_FONT_SUN8x16 is not set | ||
629 | # CONFIG_FONT_SUN12x22 is not set | ||
630 | |||
631 | # | ||
632 | # Logo configuration | ||
633 | # | ||
634 | CONFIG_LOGO=y | ||
635 | CONFIG_LOGO_LINUX_MONO=y | ||
636 | CONFIG_LOGO_LINUX_VGA16=y | ||
637 | CONFIG_LOGO_LINUX_CLUT224=y | ||
638 | # CONFIG_BACKLIGHT_LCD_SUPPORT is not set | ||
639 | |||
640 | # | ||
641 | # Sound | ||
642 | # | ||
643 | # CONFIG_SOUND is not set | ||
644 | |||
645 | # | ||
646 | # USB support | ||
647 | # | ||
648 | CONFIG_USB_ARCH_HAS_HCD=y | ||
649 | # CONFIG_USB_ARCH_HAS_OHCI is not set | ||
650 | # CONFIG_USB is not set | ||
651 | |||
652 | # | ||
653 | # USB Gadget Support | ||
654 | # | ||
655 | # CONFIG_USB_GADGET is not set | ||
656 | |||
657 | # | ||
658 | # MMC/SD Card support | ||
659 | # | ||
660 | # CONFIG_MMC is not set | ||
661 | |||
662 | # | ||
663 | # File systems | ||
664 | # | ||
665 | CONFIG_EXT2_FS=y | ||
666 | # CONFIG_EXT2_FS_XATTR is not set | ||
667 | # CONFIG_EXT3_FS is not set | ||
668 | # CONFIG_JBD is not set | ||
669 | # CONFIG_REISERFS_FS is not set | ||
670 | # CONFIG_JFS_FS is not set | ||
671 | |||
672 | # | ||
673 | # XFS support | ||
674 | # | ||
675 | # CONFIG_XFS_FS is not set | ||
676 | CONFIG_MINIX_FS=y | ||
677 | # CONFIG_ROMFS_FS is not set | ||
678 | # CONFIG_QUOTA is not set | ||
679 | CONFIG_DNOTIFY=y | ||
680 | # CONFIG_AUTOFS_FS is not set | ||
681 | # CONFIG_AUTOFS4_FS is not set | ||
682 | |||
683 | # | ||
684 | # CD-ROM/DVD Filesystems | ||
685 | # | ||
686 | # CONFIG_ISO9660_FS is not set | ||
687 | # CONFIG_UDF_FS is not set | ||
688 | |||
689 | # | ||
690 | # DOS/FAT/NT Filesystems | ||
691 | # | ||
692 | # CONFIG_MSDOS_FS is not set | ||
693 | # CONFIG_VFAT_FS is not set | ||
694 | # CONFIG_NTFS_FS is not set | ||
695 | |||
696 | # | ||
697 | # Pseudo filesystems | ||
698 | # | ||
699 | CONFIG_PROC_FS=y | ||
700 | CONFIG_SYSFS=y | ||
701 | # CONFIG_DEVFS_FS is not set | ||
702 | # CONFIG_DEVPTS_FS_XATTR is not set | ||
703 | # CONFIG_TMPFS is not set | ||
704 | # CONFIG_HUGETLB_PAGE is not set | ||
705 | CONFIG_RAMFS=y | ||
706 | |||
707 | # | ||
708 | # Miscellaneous filesystems | ||
709 | # | ||
710 | # CONFIG_ADFS_FS is not set | ||
711 | # CONFIG_AFFS_FS is not set | ||
712 | # CONFIG_HFS_FS is not set | ||
713 | # CONFIG_HFSPLUS_FS is not set | ||
714 | # CONFIG_BEFS_FS is not set | ||
715 | # CONFIG_BFS_FS is not set | ||
716 | # CONFIG_EFS_FS is not set | ||
717 | # CONFIG_JFFS_FS is not set | ||
718 | # CONFIG_JFFS2_FS is not set | ||
719 | # CONFIG_CRAMFS is not set | ||
720 | # CONFIG_VXFS_FS is not set | ||
721 | # CONFIG_HPFS_FS is not set | ||
722 | # CONFIG_QNX4FS_FS is not set | ||
723 | # CONFIG_SYSV_FS is not set | ||
724 | # CONFIG_UFS_FS is not set | ||
725 | |||
726 | # | ||
727 | # Network File Systems | ||
728 | # | ||
729 | CONFIG_NFS_FS=y | ||
730 | # CONFIG_NFS_V3 is not set | ||
731 | # CONFIG_NFS_V4 is not set | ||
732 | # CONFIG_NFS_DIRECTIO is not set | ||
733 | # CONFIG_NFSD is not set | ||
734 | CONFIG_ROOT_NFS=y | ||
735 | CONFIG_LOCKD=y | ||
736 | CONFIG_SUNRPC=y | ||
737 | # CONFIG_RPCSEC_GSS_KRB5 is not set | ||
738 | # CONFIG_RPCSEC_GSS_SPKM3 is not set | ||
739 | # CONFIG_SMB_FS is not set | ||
740 | # CONFIG_CIFS is not set | ||
741 | # CONFIG_NCP_FS is not set | ||
742 | # CONFIG_CODA_FS is not set | ||
743 | # CONFIG_AFS_FS is not set | ||
744 | |||
745 | # | ||
746 | # Partition Types | ||
747 | # | ||
748 | CONFIG_PARTITION_ADVANCED=y | ||
749 | # CONFIG_ACORN_PARTITION is not set | ||
750 | # CONFIG_OSF_PARTITION is not set | ||
751 | # CONFIG_AMIGA_PARTITION is not set | ||
752 | # CONFIG_ATARI_PARTITION is not set | ||
753 | # CONFIG_MAC_PARTITION is not set | ||
754 | # CONFIG_MSDOS_PARTITION is not set | ||
755 | # CONFIG_LDM_PARTITION is not set | ||
756 | # CONFIG_SGI_PARTITION is not set | ||
757 | # CONFIG_ULTRIX_PARTITION is not set | ||
758 | # CONFIG_SUN_PARTITION is not set | ||
759 | # CONFIG_EFI_PARTITION is not set | ||
760 | |||
761 | # | ||
762 | # Native Language Support | ||
763 | # | ||
764 | # CONFIG_NLS is not set | ||
765 | |||
766 | # | ||
767 | # Profiling support | ||
768 | # | ||
769 | # CONFIG_PROFILING is not set | ||
770 | |||
771 | # | ||
772 | # Kernel hacking | ||
773 | # | ||
774 | # CONFIG_PRINTK_TIME is not set | ||
775 | # CONFIG_DEBUG_KERNEL is not set | ||
776 | CONFIG_LOG_BUF_SHIFT=14 | ||
777 | # CONFIG_DEBUG_BUGVERBOSE is not set | ||
778 | CONFIG_FRAME_POINTER=y | ||
779 | # CONFIG_DEBUG_USER is not set | ||
780 | |||
781 | # | ||
782 | # Security options | ||
783 | # | ||
784 | # CONFIG_KEYS is not set | ||
785 | # CONFIG_SECURITY is not set | ||
786 | |||
787 | # | ||
788 | # Cryptographic options | ||
789 | # | ||
790 | # CONFIG_CRYPTO is not set | ||
791 | |||
792 | # | ||
793 | # Hardware crypto devices | ||
794 | # | ||
795 | |||
796 | # | ||
797 | # Library routines | ||
798 | # | ||
799 | # CONFIG_CRC_CCITT is not set | ||
800 | CONFIG_CRC32=y | ||
801 | # CONFIG_LIBCRC32C is not set | ||
diff --git a/arch/arm/kernel/isa.c b/arch/arm/kernel/isa.c index 50a30bc91872..8ac9b8424007 100644 --- a/arch/arm/kernel/isa.c +++ b/arch/arm/kernel/isa.c | |||
@@ -16,6 +16,7 @@ | |||
16 | #include <linux/fs.h> | 16 | #include <linux/fs.h> |
17 | #include <linux/sysctl.h> | 17 | #include <linux/sysctl.h> |
18 | #include <linux/init.h> | 18 | #include <linux/init.h> |
19 | #include <linux/io.h> | ||
19 | 20 | ||
20 | static unsigned int isa_membase, isa_portbase, isa_portshift; | 21 | static unsigned int isa_membase, isa_portbase, isa_portshift; |
21 | 22 | ||
diff --git a/arch/arm/mach-at91/at91cap9.c b/arch/arm/mach-at91/at91cap9.c index 0a38c69fdbc4..73376170fb91 100644 --- a/arch/arm/mach-at91/at91cap9.c +++ b/arch/arm/mach-at91/at91cap9.c | |||
@@ -15,6 +15,7 @@ | |||
15 | #include <linux/module.h> | 15 | #include <linux/module.h> |
16 | #include <linux/pm.h> | 16 | #include <linux/pm.h> |
17 | 17 | ||
18 | #include <asm/irq.h> | ||
18 | #include <asm/mach/arch.h> | 19 | #include <asm/mach/arch.h> |
19 | #include <asm/mach/map.h> | 20 | #include <asm/mach/map.h> |
20 | 21 | ||
diff --git a/arch/arm/mach-at91/at91rm9200.c b/arch/arm/mach-at91/at91rm9200.c index 28594fcc88e3..2e9ecad97f3d 100644 --- a/arch/arm/mach-at91/at91rm9200.c +++ b/arch/arm/mach-at91/at91rm9200.c | |||
@@ -12,6 +12,7 @@ | |||
12 | 12 | ||
13 | #include <linux/module.h> | 13 | #include <linux/module.h> |
14 | 14 | ||
15 | #include <asm/irq.h> | ||
15 | #include <asm/mach/arch.h> | 16 | #include <asm/mach/arch.h> |
16 | #include <asm/mach/map.h> | 17 | #include <asm/mach/map.h> |
17 | #include <mach/at91rm9200.h> | 18 | #include <mach/at91rm9200.h> |
diff --git a/arch/arm/mach-at91/at91sam9260.c b/arch/arm/mach-at91/at91sam9260.c index accb69ec478e..0894f1077be7 100644 --- a/arch/arm/mach-at91/at91sam9260.c +++ b/arch/arm/mach-at91/at91sam9260.c | |||
@@ -13,6 +13,7 @@ | |||
13 | #include <linux/module.h> | 13 | #include <linux/module.h> |
14 | #include <linux/pm.h> | 14 | #include <linux/pm.h> |
15 | 15 | ||
16 | #include <asm/irq.h> | ||
16 | #include <asm/mach/arch.h> | 17 | #include <asm/mach/arch.h> |
17 | #include <asm/mach/map.h> | 18 | #include <asm/mach/map.h> |
18 | #include <mach/cpu.h> | 19 | #include <mach/cpu.h> |
diff --git a/arch/arm/mach-at91/at91sam9261.c b/arch/arm/mach-at91/at91sam9261.c index 7b51a59ae8b3..3acd7d7e6a42 100644 --- a/arch/arm/mach-at91/at91sam9261.c +++ b/arch/arm/mach-at91/at91sam9261.c | |||
@@ -13,6 +13,7 @@ | |||
13 | #include <linux/module.h> | 13 | #include <linux/module.h> |
14 | #include <linux/pm.h> | 14 | #include <linux/pm.h> |
15 | 15 | ||
16 | #include <asm/irq.h> | ||
16 | #include <asm/mach/arch.h> | 17 | #include <asm/mach/arch.h> |
17 | #include <asm/mach/map.h> | 18 | #include <asm/mach/map.h> |
18 | #include <mach/at91sam9261.h> | 19 | #include <mach/at91sam9261.h> |
diff --git a/arch/arm/mach-at91/at91sam9263.c b/arch/arm/mach-at91/at91sam9263.c index ada4b6769107..942792d630d8 100644 --- a/arch/arm/mach-at91/at91sam9263.c +++ b/arch/arm/mach-at91/at91sam9263.c | |||
@@ -13,6 +13,7 @@ | |||
13 | #include <linux/module.h> | 13 | #include <linux/module.h> |
14 | #include <linux/pm.h> | 14 | #include <linux/pm.h> |
15 | 15 | ||
16 | #include <asm/irq.h> | ||
16 | #include <asm/mach/arch.h> | 17 | #include <asm/mach/arch.h> |
17 | #include <asm/mach/map.h> | 18 | #include <asm/mach/map.h> |
18 | #include <mach/at91sam9263.h> | 19 | #include <mach/at91sam9263.h> |
diff --git a/arch/arm/mach-at91/at91sam9rl.c b/arch/arm/mach-at91/at91sam9rl.c index 252e954b49fd..211c5c14a1e6 100644 --- a/arch/arm/mach-at91/at91sam9rl.c +++ b/arch/arm/mach-at91/at91sam9rl.c | |||
@@ -12,6 +12,7 @@ | |||
12 | #include <linux/module.h> | 12 | #include <linux/module.h> |
13 | #include <linux/pm.h> | 13 | #include <linux/pm.h> |
14 | 14 | ||
15 | #include <asm/irq.h> | ||
15 | #include <asm/mach/arch.h> | 16 | #include <asm/mach/arch.h> |
16 | #include <asm/mach/map.h> | 17 | #include <asm/mach/map.h> |
17 | #include <mach/cpu.h> | 18 | #include <mach/cpu.h> |
diff --git a/arch/arm/mach-at91/board-sam9rlek.c b/arch/arm/mach-at91/board-sam9rlek.c index 9b937ee4815a..35e12a49d1a6 100644 --- a/arch/arm/mach-at91/board-sam9rlek.c +++ b/arch/arm/mach-at91/board-sam9rlek.c | |||
@@ -29,6 +29,7 @@ | |||
29 | #include <mach/hardware.h> | 29 | #include <mach/hardware.h> |
30 | #include <mach/board.h> | 30 | #include <mach/board.h> |
31 | #include <mach/gpio.h> | 31 | #include <mach/gpio.h> |
32 | #include <mach/at91sam9_smc.h> | ||
32 | #include <mach/at91_shdwc.h> | 33 | #include <mach/at91_shdwc.h> |
33 | 34 | ||
34 | #include "sam9_smc.h" | 35 | #include "sam9_smc.h" |
diff --git a/arch/arm/mach-clps711x/edb7211-mm.c b/arch/arm/mach-clps711x/edb7211-mm.c index c58e32ec4c5d..0bea1454ae03 100644 --- a/arch/arm/mach-clps711x/edb7211-mm.c +++ b/arch/arm/mach-clps711x/edb7211-mm.c | |||
@@ -24,7 +24,6 @@ | |||
24 | 24 | ||
25 | #include <mach/hardware.h> | 25 | #include <mach/hardware.h> |
26 | #include <asm/page.h> | 26 | #include <asm/page.h> |
27 | #include <asm/pgtable.h> | ||
28 | #include <asm/sizes.h> | 27 | #include <asm/sizes.h> |
29 | 28 | ||
30 | #include <asm/mach/map.h> | 29 | #include <asm/mach/map.h> |
diff --git a/arch/arm/mach-clps711x/fortunet.c b/arch/arm/mach-clps711x/fortunet.c index 7122b3d21043..7430e4049d87 100644 --- a/arch/arm/mach-clps711x/fortunet.c +++ b/arch/arm/mach-clps711x/fortunet.c | |||
@@ -24,7 +24,6 @@ | |||
24 | #include <linux/initrd.h> | 24 | #include <linux/initrd.h> |
25 | 25 | ||
26 | #include <mach/hardware.h> | 26 | #include <mach/hardware.h> |
27 | #include <asm/irq.h> | ||
28 | #include <asm/setup.h> | 27 | #include <asm/setup.h> |
29 | #include <asm/mach-types.h> | 28 | #include <asm/mach-types.h> |
30 | 29 | ||
diff --git a/arch/arm/mach-davinci/devices.c b/arch/arm/mach-davinci/devices.c index 3d4b1de8f898..808633f9f03c 100644 --- a/arch/arm/mach-davinci/devices.c +++ b/arch/arm/mach-davinci/devices.c | |||
@@ -20,6 +20,7 @@ | |||
20 | 20 | ||
21 | #include <mach/hardware.h> | 21 | #include <mach/hardware.h> |
22 | #include <mach/i2c.h> | 22 | #include <mach/i2c.h> |
23 | #include <mach/irqs.h> | ||
23 | 24 | ||
24 | static struct resource i2c_resources[] = { | 25 | static struct resource i2c_resources[] = { |
25 | { | 26 | { |
diff --git a/arch/arm/mach-davinci/include/mach/gpio.h b/arch/arm/mach-davinci/include/mach/gpio.h index b3a2961f0f46..b456f079f43f 100644 --- a/arch/arm/mach-davinci/include/mach/gpio.h +++ b/arch/arm/mach-davinci/include/mach/gpio.h | |||
@@ -16,6 +16,7 @@ | |||
16 | #include <linux/io.h> | 16 | #include <linux/io.h> |
17 | #include <asm-generic/gpio.h> | 17 | #include <asm-generic/gpio.h> |
18 | #include <mach/hardware.h> | 18 | #include <mach/hardware.h> |
19 | #include <mach/irqs.h> | ||
19 | 20 | ||
20 | /* | 21 | /* |
21 | * basic gpio routines | 22 | * basic gpio routines |
diff --git a/arch/arm/mach-footbridge/common.c b/arch/arm/mach-footbridge/common.c index 36ff06d4df15..b97f529e58e8 100644 --- a/arch/arm/mach-footbridge/common.c +++ b/arch/arm/mach-footbridge/common.c | |||
@@ -28,12 +28,17 @@ | |||
28 | 28 | ||
29 | #include "common.h" | 29 | #include "common.h" |
30 | 30 | ||
31 | extern void __init isa_init_irq(unsigned int irq); | ||
32 | |||
33 | unsigned int mem_fclk_21285 = 50000000; | 31 | unsigned int mem_fclk_21285 = 50000000; |
34 | 32 | ||
35 | EXPORT_SYMBOL(mem_fclk_21285); | 33 | EXPORT_SYMBOL(mem_fclk_21285); |
36 | 34 | ||
35 | static void __init early_fclk(char **arg) | ||
36 | { | ||
37 | mem_fclk_21285 = simple_strtoul(*arg, arg, 0); | ||
38 | } | ||
39 | |||
40 | __early_param("mem_fclk_21285=", early_fclk); | ||
41 | |||
37 | static int __init parse_tag_memclk(const struct tag *tag) | 42 | static int __init parse_tag_memclk(const struct tag *tag) |
38 | { | 43 | { |
39 | mem_fclk_21285 = tag->u.memclk.fmemclk; | 44 | mem_fclk_21285 = tag->u.memclk.fmemclk; |
diff --git a/arch/arm/mach-footbridge/common.h b/arch/arm/mach-footbridge/common.h index 580e31bbc711..b05e662d21ad 100644 --- a/arch/arm/mach-footbridge/common.h +++ b/arch/arm/mach-footbridge/common.h | |||
@@ -7,3 +7,4 @@ extern void isa_rtc_init(void); | |||
7 | extern void footbridge_map_io(void); | 7 | extern void footbridge_map_io(void); |
8 | extern void footbridge_init_irq(void); | 8 | extern void footbridge_init_irq(void); |
9 | 9 | ||
10 | extern void isa_init_irq(unsigned int irq); | ||
diff --git a/arch/arm/mach-footbridge/dc21285.c b/arch/arm/mach-footbridge/dc21285.c index 133086019e3e..3ffa54841ec5 100644 --- a/arch/arm/mach-footbridge/dc21285.c +++ b/arch/arm/mach-footbridge/dc21285.c | |||
@@ -287,6 +287,9 @@ struct pci_bus * __init dc21285_scan_bus(int nr, struct pci_sys_data *sys) | |||
287 | return pci_scan_bus(0, &dc21285_ops, sys); | 287 | return pci_scan_bus(0, &dc21285_ops, sys); |
288 | } | 288 | } |
289 | 289 | ||
290 | #define dc21285_request_irq(_a, _b, _c, _d, _e) \ | ||
291 | WARN_ON(request_irq(_a, _b, _c, _d, _e) < 0) | ||
292 | |||
290 | void __init dc21285_preinit(void) | 293 | void __init dc21285_preinit(void) |
291 | { | 294 | { |
292 | unsigned int mem_size, mem_mask; | 295 | unsigned int mem_size, mem_mask; |
@@ -335,16 +338,16 @@ void __init dc21285_preinit(void) | |||
335 | /* | 338 | /* |
336 | * We don't care if these fail. | 339 | * We don't care if these fail. |
337 | */ | 340 | */ |
338 | request_irq(IRQ_PCI_SERR, dc21285_serr_irq, IRQF_DISABLED, | 341 | dc21285_request_irq(IRQ_PCI_SERR, dc21285_serr_irq, IRQF_DISABLED, |
339 | "PCI system error", &serr_timer); | 342 | "PCI system error", &serr_timer); |
340 | request_irq(IRQ_PCI_PERR, dc21285_parity_irq, IRQF_DISABLED, | 343 | dc21285_request_irq(IRQ_PCI_PERR, dc21285_parity_irq, IRQF_DISABLED, |
341 | "PCI parity error", &perr_timer); | 344 | "PCI parity error", &perr_timer); |
342 | request_irq(IRQ_PCI_ABORT, dc21285_abort_irq, IRQF_DISABLED, | 345 | dc21285_request_irq(IRQ_PCI_ABORT, dc21285_abort_irq, IRQF_DISABLED, |
343 | "PCI abort", NULL); | 346 | "PCI abort", NULL); |
344 | request_irq(IRQ_DISCARD_TIMER, dc21285_discard_irq, IRQF_DISABLED, | 347 | dc21285_request_irq(IRQ_DISCARD_TIMER, dc21285_discard_irq, IRQF_DISABLED, |
345 | "Discard timer", NULL); | 348 | "Discard timer", NULL); |
346 | request_irq(IRQ_PCI_DPERR, dc21285_dparity_irq, IRQF_DISABLED, | 349 | dc21285_request_irq(IRQ_PCI_DPERR, dc21285_dparity_irq, IRQF_DISABLED, |
347 | "PCI data parity", NULL); | 350 | "PCI data parity", NULL); |
348 | 351 | ||
349 | if (cfn_mode) { | 352 | if (cfn_mode) { |
350 | static struct resource csrio; | 353 | static struct resource csrio; |
diff --git a/arch/arm/mach-footbridge/isa-irq.c b/arch/arm/mach-footbridge/isa-irq.c index 9ee80a211d3c..8bfd06aeb64d 100644 --- a/arch/arm/mach-footbridge/isa-irq.c +++ b/arch/arm/mach-footbridge/isa-irq.c | |||
@@ -28,6 +28,8 @@ | |||
28 | #include <asm/irq.h> | 28 | #include <asm/irq.h> |
29 | #include <asm/mach-types.h> | 29 | #include <asm/mach-types.h> |
30 | 30 | ||
31 | #include "common.h" | ||
32 | |||
31 | static void isa_mask_pic_lo_irq(unsigned int irq) | 33 | static void isa_mask_pic_lo_irq(unsigned int irq) |
32 | { | 34 | { |
33 | unsigned int mask = 1 << (irq & 7); | 35 | unsigned int mask = 1 << (irq & 7); |
diff --git a/arch/arm/mach-h720x/h7202-eval.c b/arch/arm/mach-h720x/h7202-eval.c index 56161d55cf47..8c0ba99d683f 100644 --- a/arch/arm/mach-h720x/h7202-eval.c +++ b/arch/arm/mach-h720x/h7202-eval.c | |||
@@ -25,6 +25,7 @@ | |||
25 | #include <asm/page.h> | 25 | #include <asm/page.h> |
26 | #include <asm/pgtable.h> | 26 | #include <asm/pgtable.h> |
27 | #include <asm/mach/arch.h> | 27 | #include <asm/mach/arch.h> |
28 | #include <mach/irqs.h> | ||
28 | #include <mach/hardware.h> | 29 | #include <mach/hardware.h> |
29 | #include "common.h" | 30 | #include "common.h" |
30 | 31 | ||
diff --git a/arch/arm/mach-kirkwood/common.c b/arch/arm/mach-kirkwood/common.c index 7b8ef97fb501..b3404b7775b3 100644 --- a/arch/arm/mach-kirkwood/common.c +++ b/arch/arm/mach-kirkwood/common.c | |||
@@ -698,6 +698,7 @@ void __init kirkwood_init(void) | |||
698 | printk(KERN_INFO "Kirkwood: %s, TCLK=%d.\n", | 698 | printk(KERN_INFO "Kirkwood: %s, TCLK=%d.\n", |
699 | kirkwood_id(), kirkwood_tclk); | 699 | kirkwood_id(), kirkwood_tclk); |
700 | kirkwood_ge00_shared_data.t_clk = kirkwood_tclk; | 700 | kirkwood_ge00_shared_data.t_clk = kirkwood_tclk; |
701 | kirkwood_ge01_shared_data.t_clk = kirkwood_tclk; | ||
701 | kirkwood_spi_plat_data.tclk = kirkwood_tclk; | 702 | kirkwood_spi_plat_data.tclk = kirkwood_tclk; |
702 | kirkwood_uart0_data[0].uartclk = kirkwood_tclk; | 703 | kirkwood_uart0_data[0].uartclk = kirkwood_tclk; |
703 | kirkwood_uart1_data[0].uartclk = kirkwood_tclk; | 704 | kirkwood_uart1_data[0].uartclk = kirkwood_tclk; |
diff --git a/arch/arm/mach-kirkwood/pcie.c b/arch/arm/mach-kirkwood/pcie.c index f6b08f207c89..73fccacd1a73 100644 --- a/arch/arm/mach-kirkwood/pcie.c +++ b/arch/arm/mach-kirkwood/pcie.c | |||
@@ -11,6 +11,7 @@ | |||
11 | #include <linux/kernel.h> | 11 | #include <linux/kernel.h> |
12 | #include <linux/pci.h> | 12 | #include <linux/pci.h> |
13 | #include <linux/mbus.h> | 13 | #include <linux/mbus.h> |
14 | #include <asm/irq.h> | ||
14 | #include <asm/mach/pci.h> | 15 | #include <asm/mach/pci.h> |
15 | #include <plat/pcie.h> | 16 | #include <plat/pcie.h> |
16 | #include "common.h" | 17 | #include "common.h" |
diff --git a/arch/arm/mach-ks8695/devices.c b/arch/arm/mach-ks8695/devices.c index 36ab0fd3d9b6..b89fb6d46ccc 100644 --- a/arch/arm/mach-ks8695/devices.c +++ b/arch/arm/mach-ks8695/devices.c | |||
@@ -22,6 +22,7 @@ | |||
22 | 22 | ||
23 | #include <linux/platform_device.h> | 23 | #include <linux/platform_device.h> |
24 | 24 | ||
25 | #include <mach/irqs.h> | ||
25 | #include <mach/regs-wan.h> | 26 | #include <mach/regs-wan.h> |
26 | #include <mach/regs-lan.h> | 27 | #include <mach/regs-lan.h> |
27 | #include <mach/regs-hpna.h> | 28 | #include <mach/regs-hpna.h> |
diff --git a/arch/arm/mach-msm/devices.c b/arch/arm/mach-msm/devices.c index f2a74b92a97f..31b6b30e98bf 100644 --- a/arch/arm/mach-msm/devices.c +++ b/arch/arm/mach-msm/devices.c | |||
@@ -16,6 +16,7 @@ | |||
16 | #include <linux/kernel.h> | 16 | #include <linux/kernel.h> |
17 | #include <linux/platform_device.h> | 17 | #include <linux/platform_device.h> |
18 | 18 | ||
19 | #include <mach/irqs.h> | ||
19 | #include <mach/msm_iomap.h> | 20 | #include <mach/msm_iomap.h> |
20 | #include "devices.h" | 21 | #include "devices.h" |
21 | 22 | ||
diff --git a/arch/arm/mach-mv78xx0/pcie.c b/arch/arm/mach-mv78xx0/pcie.c index 430ea84d587d..aad3a7a2f830 100644 --- a/arch/arm/mach-mv78xx0/pcie.c +++ b/arch/arm/mach-mv78xx0/pcie.c | |||
@@ -11,6 +11,7 @@ | |||
11 | #include <linux/kernel.h> | 11 | #include <linux/kernel.h> |
12 | #include <linux/pci.h> | 12 | #include <linux/pci.h> |
13 | #include <linux/mbus.h> | 13 | #include <linux/mbus.h> |
14 | #include <asm/irq.h> | ||
14 | #include <asm/mach/pci.h> | 15 | #include <asm/mach/pci.h> |
15 | #include <plat/pcie.h> | 16 | #include <plat/pcie.h> |
16 | #include "common.h" | 17 | #include "common.h" |
diff --git a/arch/arm/mach-mx2/devices.c b/arch/arm/mach-mx2/devices.c index af121f5ab710..2f9240be1c76 100644 --- a/arch/arm/mach-mx2/devices.c +++ b/arch/arm/mach-mx2/devices.c | |||
@@ -32,6 +32,7 @@ | |||
32 | #include <linux/platform_device.h> | 32 | #include <linux/platform_device.h> |
33 | #include <linux/gpio.h> | 33 | #include <linux/gpio.h> |
34 | 34 | ||
35 | #include <mach/irqs.h> | ||
35 | #include <mach/hardware.h> | 36 | #include <mach/hardware.h> |
36 | 37 | ||
37 | /* | 38 | /* |
diff --git a/arch/arm/mach-mx3/devices.c b/arch/arm/mach-mx3/devices.c index 1d46cb4adf96..f8428800f286 100644 --- a/arch/arm/mach-mx3/devices.c +++ b/arch/arm/mach-mx3/devices.c | |||
@@ -22,6 +22,7 @@ | |||
22 | #include <linux/serial.h> | 22 | #include <linux/serial.h> |
23 | #include <linux/gpio.h> | 23 | #include <linux/gpio.h> |
24 | #include <mach/hardware.h> | 24 | #include <mach/hardware.h> |
25 | #include <mach/irqs.h> | ||
25 | #include <mach/imx-uart.h> | 26 | #include <mach/imx-uart.h> |
26 | 27 | ||
27 | static struct resource uart0[] = { | 28 | static struct resource uart0[] = { |
diff --git a/arch/arm/mach-netx/fb.c b/arch/arm/mach-netx/fb.c index ea8fa8898fe8..1d844e228ea9 100644 --- a/arch/arm/mach-netx/fb.c +++ b/arch/arm/mach-netx/fb.c | |||
@@ -24,6 +24,8 @@ | |||
24 | #include <linux/amba/clcd.h> | 24 | #include <linux/amba/clcd.h> |
25 | #include <linux/err.h> | 25 | #include <linux/err.h> |
26 | 26 | ||
27 | #include <asm/irq.h> | ||
28 | |||
27 | #include <mach/netx-regs.h> | 29 | #include <mach/netx-regs.h> |
28 | #include <mach/hardware.h> | 30 | #include <mach/hardware.h> |
29 | 31 | ||
diff --git a/arch/arm/mach-netx/time.c b/arch/arm/mach-netx/time.c index d51d627ce7cf..f201fddb594f 100644 --- a/arch/arm/mach-netx/time.c +++ b/arch/arm/mach-netx/time.c | |||
@@ -163,7 +163,7 @@ static void __init netx_timer_init(void) | |||
163 | * Adding some safety ... */ | 163 | * Adding some safety ... */ |
164 | netx_clockevent.min_delta_ns = | 164 | netx_clockevent.min_delta_ns = |
165 | clockevent_delta2ns(0xa00, &netx_clockevent); | 165 | clockevent_delta2ns(0xa00, &netx_clockevent); |
166 | netx_clockevent.cpumask = cpumask_of_cpu(0); | 166 | netx_clockevent.cpumask = cpumask_of(0); |
167 | clockevents_register_device(&netx_clockevent); | 167 | clockevents_register_device(&netx_clockevent); |
168 | } | 168 | } |
169 | 169 | ||
diff --git a/arch/arm/mach-netx/xc.c b/arch/arm/mach-netx/xc.c index 8fc6205dc3a5..181a78ba8165 100644 --- a/arch/arm/mach-netx/xc.c +++ b/arch/arm/mach-netx/xc.c | |||
@@ -24,6 +24,7 @@ | |||
24 | #include <linux/io.h> | 24 | #include <linux/io.h> |
25 | 25 | ||
26 | #include <mach/hardware.h> | 26 | #include <mach/hardware.h> |
27 | #include <mach/irqs.h> | ||
27 | #include <mach/netx-regs.h> | 28 | #include <mach/netx-regs.h> |
28 | 29 | ||
29 | #include <mach/xc.h> | 30 | #include <mach/xc.h> |
diff --git a/arch/arm/mach-omap1/mcbsp.c b/arch/arm/mach-omap1/mcbsp.c index 7de7c6915584..4474da7bc88a 100644 --- a/arch/arm/mach-omap1/mcbsp.c +++ b/arch/arm/mach-omap1/mcbsp.c | |||
@@ -18,6 +18,7 @@ | |||
18 | #include <linux/platform_device.h> | 18 | #include <linux/platform_device.h> |
19 | 19 | ||
20 | #include <mach/dma.h> | 20 | #include <mach/dma.h> |
21 | #include <mach/irqs.h> | ||
21 | #include <mach/mux.h> | 22 | #include <mach/mux.h> |
22 | #include <mach/cpu.h> | 23 | #include <mach/cpu.h> |
23 | #include <mach/mcbsp.h> | 24 | #include <mach/mcbsp.h> |
diff --git a/arch/arm/mach-omap2/mcbsp.c b/arch/arm/mach-omap2/mcbsp.c index cae3ebe249b3..acdc709901cd 100644 --- a/arch/arm/mach-omap2/mcbsp.c +++ b/arch/arm/mach-omap2/mcbsp.c | |||
@@ -18,6 +18,7 @@ | |||
18 | #include <linux/platform_device.h> | 18 | #include <linux/platform_device.h> |
19 | 19 | ||
20 | #include <mach/dma.h> | 20 | #include <mach/dma.h> |
21 | #include <mach/irqs.h> | ||
21 | #include <mach/mux.h> | 22 | #include <mach/mux.h> |
22 | #include <mach/cpu.h> | 23 | #include <mach/cpu.h> |
23 | #include <mach/mcbsp.h> | 24 | #include <mach/mcbsp.h> |
diff --git a/arch/arm/mach-orion5x/pci.c b/arch/arm/mach-orion5x/pci.c index a7b7d77b1b09..d0a785a3b880 100644 --- a/arch/arm/mach-orion5x/pci.c +++ b/arch/arm/mach-orion5x/pci.c | |||
@@ -13,6 +13,7 @@ | |||
13 | #include <linux/kernel.h> | 13 | #include <linux/kernel.h> |
14 | #include <linux/pci.h> | 14 | #include <linux/pci.h> |
15 | #include <linux/mbus.h> | 15 | #include <linux/mbus.h> |
16 | #include <asm/irq.h> | ||
16 | #include <asm/mach/pci.h> | 17 | #include <asm/mach/pci.h> |
17 | #include <plat/pcie.h> | 18 | #include <plat/pcie.h> |
18 | #include "common.h" | 19 | #include "common.h" |
diff --git a/arch/arm/mach-pnx4008/gpio.c b/arch/arm/mach-pnx4008/gpio.c index 015cc21d5f55..f219914f5b29 100644 --- a/arch/arm/mach-pnx4008/gpio.c +++ b/arch/arm/mach-pnx4008/gpio.c | |||
@@ -18,6 +18,7 @@ | |||
18 | #include <linux/kernel.h> | 18 | #include <linux/kernel.h> |
19 | #include <linux/module.h> | 19 | #include <linux/module.h> |
20 | #include <linux/io.h> | 20 | #include <linux/io.h> |
21 | #include <mach/hardware.h> | ||
21 | #include <mach/platform.h> | 22 | #include <mach/platform.h> |
22 | #include <mach/gpio.h> | 23 | #include <mach/gpio.h> |
23 | 24 | ||
diff --git a/arch/arm/mach-pnx4008/i2c.c b/arch/arm/mach-pnx4008/i2c.c index 87c093286ff9..f3fea29c00d3 100644 --- a/arch/arm/mach-pnx4008/i2c.c +++ b/arch/arm/mach-pnx4008/i2c.c | |||
@@ -15,6 +15,7 @@ | |||
15 | #include <linux/platform_device.h> | 15 | #include <linux/platform_device.h> |
16 | #include <linux/err.h> | 16 | #include <linux/err.h> |
17 | #include <mach/platform.h> | 17 | #include <mach/platform.h> |
18 | #include <mach/irqs.h> | ||
18 | #include <mach/i2c.h> | 19 | #include <mach/i2c.h> |
19 | 20 | ||
20 | static int set_clock_run(struct platform_device *pdev) | 21 | static int set_clock_run(struct platform_device *pdev) |
diff --git a/arch/arm/mach-pxa/e350.c b/arch/arm/mach-pxa/e350.c index 251129391d7d..edcd9d5ce545 100644 --- a/arch/arm/mach-pxa/e350.c +++ b/arch/arm/mach-pxa/e350.c | |||
@@ -20,6 +20,7 @@ | |||
20 | #include <asm/mach/arch.h> | 20 | #include <asm/mach/arch.h> |
21 | #include <asm/mach-types.h> | 21 | #include <asm/mach-types.h> |
22 | 22 | ||
23 | #include <mach/irqs.h> | ||
23 | #include <mach/mfp-pxa25x.h> | 24 | #include <mach/mfp-pxa25x.h> |
24 | #include <mach/pxa-regs.h> | 25 | #include <mach/pxa-regs.h> |
25 | #include <mach/hardware.h> | 26 | #include <mach/hardware.h> |
diff --git a/arch/arm/mach-pxa/e400.c b/arch/arm/mach-pxa/e400.c index bed0336aca3d..77bb8e2c48c0 100644 --- a/arch/arm/mach-pxa/e400.c +++ b/arch/arm/mach-pxa/e400.c | |||
@@ -28,6 +28,7 @@ | |||
28 | #include <mach/eseries-gpio.h> | 28 | #include <mach/eseries-gpio.h> |
29 | #include <mach/pxafb.h> | 29 | #include <mach/pxafb.h> |
30 | #include <mach/udc.h> | 30 | #include <mach/udc.h> |
31 | #include <mach/irqs.h> | ||
31 | 32 | ||
32 | #include "generic.h" | 33 | #include "generic.h" |
33 | #include "eseries.h" | 34 | #include "eseries.h" |
diff --git a/arch/arm/mach-pxa/e740.c b/arch/arm/mach-pxa/e740.c index b00d670b2ea6..6d48e00f4f0b 100644 --- a/arch/arm/mach-pxa/e740.c +++ b/arch/arm/mach-pxa/e740.c | |||
@@ -30,6 +30,7 @@ | |||
30 | #include <mach/eseries-gpio.h> | 30 | #include <mach/eseries-gpio.h> |
31 | #include <mach/udc.h> | 31 | #include <mach/udc.h> |
32 | #include <mach/irda.h> | 32 | #include <mach/irda.h> |
33 | #include <mach/irqs.h> | ||
33 | 34 | ||
34 | #include "generic.h" | 35 | #include "generic.h" |
35 | #include "eseries.h" | 36 | #include "eseries.h" |
diff --git a/arch/arm/mach-pxa/e750.c b/arch/arm/mach-pxa/e750.c index 84d7c1aac58d..be1ab8edb973 100644 --- a/arch/arm/mach-pxa/e750.c +++ b/arch/arm/mach-pxa/e750.c | |||
@@ -29,6 +29,7 @@ | |||
29 | #include <mach/eseries-gpio.h> | 29 | #include <mach/eseries-gpio.h> |
30 | #include <mach/udc.h> | 30 | #include <mach/udc.h> |
31 | #include <mach/irda.h> | 31 | #include <mach/irda.h> |
32 | #include <mach/irqs.h> | ||
32 | 33 | ||
33 | #include "generic.h" | 34 | #include "generic.h" |
34 | #include "eseries.h" | 35 | #include "eseries.h" |
@@ -105,6 +106,57 @@ static struct platform_device e750_fb_device = { | |||
105 | .resource = e750_fb_resources, | 106 | .resource = e750_fb_resources, |
106 | }; | 107 | }; |
107 | 108 | ||
109 | /* -------------------- e750 MFP parameters -------------------- */ | ||
110 | |||
111 | static unsigned long e750_pin_config[] __initdata = { | ||
112 | /* Chip selects */ | ||
113 | GPIO15_nCS_1, /* CS1 - Flash */ | ||
114 | GPIO79_nCS_3, /* CS3 - IMAGEON */ | ||
115 | GPIO80_nCS_4, /* CS4 - TMIO */ | ||
116 | |||
117 | /* Clocks */ | ||
118 | GPIO11_3_6MHz, | ||
119 | |||
120 | /* BTUART */ | ||
121 | GPIO42_BTUART_RXD, | ||
122 | GPIO43_BTUART_TXD, | ||
123 | GPIO44_BTUART_CTS, | ||
124 | |||
125 | /* TMIO controller */ | ||
126 | GPIO19_GPIO, /* t7l66xb #PCLR */ | ||
127 | GPIO45_GPIO, /* t7l66xb #SUSPEND (NOT BTUART!) */ | ||
128 | |||
129 | /* UDC */ | ||
130 | GPIO13_GPIO, | ||
131 | GPIO3_GPIO, | ||
132 | |||
133 | /* IrDA */ | ||
134 | GPIO38_GPIO | MFP_LPM_DRIVE_HIGH, | ||
135 | |||
136 | /* PC Card */ | ||
137 | GPIO8_GPIO, /* CD0 */ | ||
138 | GPIO44_GPIO, /* CD1 */ | ||
139 | GPIO11_GPIO, /* IRQ0 */ | ||
140 | GPIO6_GPIO, /* IRQ1 */ | ||
141 | GPIO27_GPIO, /* RST0 */ | ||
142 | GPIO24_GPIO, /* RST1 */ | ||
143 | GPIO20_GPIO, /* PWR0 */ | ||
144 | GPIO23_GPIO, /* PWR1 */ | ||
145 | GPIO48_nPOE, | ||
146 | GPIO49_nPWE, | ||
147 | GPIO50_nPIOR, | ||
148 | GPIO51_nPIOW, | ||
149 | GPIO52_nPCE_1, | ||
150 | GPIO53_nPCE_2, | ||
151 | GPIO54_nPSKTSEL, | ||
152 | GPIO55_nPREG, | ||
153 | GPIO56_nPWAIT, | ||
154 | GPIO57_nIOIS16, | ||
155 | |||
156 | /* wakeup */ | ||
157 | GPIO0_GPIO | WAKEUP_ON_EDGE_RISE, | ||
158 | }; | ||
159 | |||
108 | /* ----------------- e750 tc6393xb parameters ------------------ */ | 160 | /* ----------------- e750 tc6393xb parameters ------------------ */ |
109 | 161 | ||
110 | static struct tc6393xb_platform_data e750_tc6393xb_info = { | 162 | static struct tc6393xb_platform_data e750_tc6393xb_info = { |
@@ -137,6 +189,7 @@ static struct platform_device *devices[] __initdata = { | |||
137 | 189 | ||
138 | static void __init e750_init(void) | 190 | static void __init e750_init(void) |
139 | { | 191 | { |
192 | pxa2xx_mfp_config(ARRAY_AND_SIZE(e750_pin_config)); | ||
140 | clk_add_alias("CLK_CK3P6MI", &e750_tc6393xb_device.dev, | 193 | clk_add_alias("CLK_CK3P6MI", &e750_tc6393xb_device.dev, |
141 | "GPIO11_CLK", NULL), | 194 | "GPIO11_CLK", NULL), |
142 | eseries_get_tmio_gpios(); | 195 | eseries_get_tmio_gpios(); |
diff --git a/arch/arm/mach-pxa/e800.c b/arch/arm/mach-pxa/e800.c index 9a86a426f924..cc9b1293e866 100644 --- a/arch/arm/mach-pxa/e800.c +++ b/arch/arm/mach-pxa/e800.c | |||
@@ -28,6 +28,7 @@ | |||
28 | #include <mach/hardware.h> | 28 | #include <mach/hardware.h> |
29 | #include <mach/eseries-gpio.h> | 29 | #include <mach/eseries-gpio.h> |
30 | #include <mach/udc.h> | 30 | #include <mach/udc.h> |
31 | #include <mach/irqs.h> | ||
31 | 32 | ||
32 | #include "generic.h" | 33 | #include "generic.h" |
33 | #include "eseries.h" | 34 | #include "eseries.h" |
diff --git a/arch/arm/mach-pxa/include/mach/pxa3xx-regs.h b/arch/arm/mach-pxa/include/mach/pxa3xx-regs.h index b1fcd10ab6c6..bcf3fb2c4b3a 100644 --- a/arch/arm/mach-pxa/include/mach/pxa3xx-regs.h +++ b/arch/arm/mach-pxa/include/mach/pxa3xx-regs.h | |||
@@ -193,10 +193,8 @@ | |||
193 | #define CKEN_MINI_IM 48 /* < Mini-IM */ | 193 | #define CKEN_MINI_IM 48 /* < Mini-IM */ |
194 | #define CKEN_MINI_LCD 49 /* < Mini LCD */ | 194 | #define CKEN_MINI_LCD 49 /* < Mini LCD */ |
195 | 195 | ||
196 | #if defined(CONFIG_CPU_PXA310) | ||
197 | #define CKEN_MMC3 5 /* < MMC3 Clock Enable */ | 196 | #define CKEN_MMC3 5 /* < MMC3 Clock Enable */ |
198 | #define CKEN_MVED 43 /* < MVED clock enable */ | 197 | #define CKEN_MVED 43 /* < MVED clock enable */ |
199 | #endif | ||
200 | 198 | ||
201 | /* Note: GCU clock enable bit differs on PXA300/PXA310 and PXA320 */ | 199 | /* Note: GCU clock enable bit differs on PXA300/PXA310 and PXA320 */ |
202 | #define PXA300_CKEN_GRAPHICS 42 /* Graphics controller clock enable */ | 200 | #define PXA300_CKEN_GRAPHICS 42 /* Graphics controller clock enable */ |
diff --git a/arch/arm/mach-realview/platsmp.c b/arch/arm/mach-realview/platsmp.c index 8fce85f33033..ea3c75595fa9 100644 --- a/arch/arm/mach-realview/platsmp.c +++ b/arch/arm/mach-realview/platsmp.c | |||
@@ -12,6 +12,7 @@ | |||
12 | #include <linux/errno.h> | 12 | #include <linux/errno.h> |
13 | #include <linux/delay.h> | 13 | #include <linux/delay.h> |
14 | #include <linux/device.h> | 14 | #include <linux/device.h> |
15 | #include <linux/jiffies.h> | ||
15 | #include <linux/smp.h> | 16 | #include <linux/smp.h> |
16 | #include <linux/io.h> | 17 | #include <linux/io.h> |
17 | 18 | ||
diff --git a/arch/arm/mach-s3c2410/include/mach/gpio.h b/arch/arm/mach-s3c2410/include/mach/gpio.h index e0349af8a483..00476a573bbe 100644 --- a/arch/arm/mach-s3c2410/include/mach/gpio.h +++ b/arch/arm/mach-s3c2410/include/mach/gpio.h | |||
@@ -14,6 +14,7 @@ | |||
14 | #define gpio_get_value __gpio_get_value | 14 | #define gpio_get_value __gpio_get_value |
15 | #define gpio_set_value __gpio_set_value | 15 | #define gpio_set_value __gpio_set_value |
16 | #define gpio_cansleep __gpio_cansleep | 16 | #define gpio_cansleep __gpio_cansleep |
17 | #define gpio_to_irq __gpio_to_irq | ||
17 | 18 | ||
18 | /* some boards require extra gpio capacity to support external | 19 | /* some boards require extra gpio capacity to support external |
19 | * devices that need GPIO. | 20 | * devices that need GPIO. |
diff --git a/arch/arm/mach-s3c2410/include/mach/irqs.h b/arch/arm/mach-s3c2410/include/mach/irqs.h index 9565903d490b..49efce8cd4a7 100644 --- a/arch/arm/mach-s3c2410/include/mach/irqs.h +++ b/arch/arm/mach-s3c2410/include/mach/irqs.h | |||
@@ -12,10 +12,6 @@ | |||
12 | #ifndef __ASM_ARCH_IRQS_H | 12 | #ifndef __ASM_ARCH_IRQS_H |
13 | #define __ASM_ARCH_IRQS_H __FILE__ | 13 | #define __ASM_ARCH_IRQS_H __FILE__ |
14 | 14 | ||
15 | #ifndef __ASM_ARM_IRQ_H | ||
16 | #error "Do not include this directly, instead #include <asm/irq.h>" | ||
17 | #endif | ||
18 | |||
19 | /* we keep the first set of CPU IRQs out of the range of | 15 | /* we keep the first set of CPU IRQs out of the range of |
20 | * the ISA space, so that the PC104 has them to itself | 16 | * the ISA space, so that the PC104 has them to itself |
21 | * and we don't end up having to do horrible things to the | 17 | * and we don't end up having to do horrible things to the |
diff --git a/arch/arm/mach-s3c2440/mach-at2440evb.c b/arch/arm/mach-s3c2440/mach-at2440evb.c index 0a6d0a5d961b..315c42e31278 100644 --- a/arch/arm/mach-s3c2440/mach-at2440evb.c +++ b/arch/arm/mach-s3c2440/mach-at2440evb.c | |||
@@ -47,7 +47,7 @@ | |||
47 | #include <plat/clock.h> | 47 | #include <plat/clock.h> |
48 | #include <plat/devs.h> | 48 | #include <plat/devs.h> |
49 | #include <plat/cpu.h> | 49 | #include <plat/cpu.h> |
50 | #include <asm/plat-s3c24xx/mci.h> | 50 | #include <plat/mci.h> |
51 | 51 | ||
52 | static struct map_desc at2440evb_iodesc[] __initdata = { | 52 | static struct map_desc at2440evb_iodesc[] __initdata = { |
53 | /* Nothing here */ | 53 | /* Nothing here */ |
diff --git a/arch/arm/mach-s3c6400/include/mach/irqs.h b/arch/arm/mach-s3c6400/include/mach/irqs.h index b38c47cffc28..4c97f9a4370b 100644 --- a/arch/arm/mach-s3c6400/include/mach/irqs.h +++ b/arch/arm/mach-s3c6400/include/mach/irqs.h | |||
@@ -11,10 +11,6 @@ | |||
11 | #ifndef __ASM_ARCH_IRQS_H | 11 | #ifndef __ASM_ARCH_IRQS_H |
12 | #define __ASM_ARCH_IRQS_H __FILE__ | 12 | #define __ASM_ARCH_IRQS_H __FILE__ |
13 | 13 | ||
14 | #ifndef __ASM_ARM_IRQ_H | ||
15 | #error "Do not include this directly, instead #include <asm/irq.h>" | ||
16 | #endif | ||
17 | |||
18 | #include <plat/irqs.h> | 14 | #include <plat/irqs.h> |
19 | 15 | ||
20 | #endif /* __ASM_ARCH_IRQ_H */ | 16 | #endif /* __ASM_ARCH_IRQ_H */ |
diff --git a/arch/arm/plat-omap/i2c.c b/arch/arm/plat-omap/i2c.c index 89a6ab0b7db8..467531edefd3 100644 --- a/arch/arm/plat-omap/i2c.c +++ b/arch/arm/plat-omap/i2c.c | |||
@@ -26,6 +26,7 @@ | |||
26 | #include <linux/kernel.h> | 26 | #include <linux/kernel.h> |
27 | #include <linux/platform_device.h> | 27 | #include <linux/platform_device.h> |
28 | #include <linux/i2c.h> | 28 | #include <linux/i2c.h> |
29 | #include <mach/irqs.h> | ||
29 | #include <mach/mux.h> | 30 | #include <mach/mux.h> |
30 | 31 | ||
31 | #define OMAP_I2C_SIZE 0x3f | 32 | #define OMAP_I2C_SIZE 0x3f |
diff --git a/arch/arm/plat-s3c/dev-fb.c b/arch/arm/plat-s3c/dev-fb.c index 0454b8ec02e2..a90198fc4b0f 100644 --- a/arch/arm/plat-s3c/dev-fb.c +++ b/arch/arm/plat-s3c/dev-fb.c | |||
@@ -16,6 +16,7 @@ | |||
16 | #include <linux/platform_device.h> | 16 | #include <linux/platform_device.h> |
17 | #include <linux/fb.h> | 17 | #include <linux/fb.h> |
18 | 18 | ||
19 | #include <mach/irqs.h> | ||
19 | #include <mach/map.h> | 20 | #include <mach/map.h> |
20 | #include <mach/regs-fb.h> | 21 | #include <mach/regs-fb.h> |
21 | 22 | ||
diff --git a/arch/arm/plat-s3c/dev-i2c0.c b/arch/arm/plat-s3c/dev-i2c0.c index 2c0128c77c6e..fe327074037e 100644 --- a/arch/arm/plat-s3c/dev-i2c0.c +++ b/arch/arm/plat-s3c/dev-i2c0.c | |||
@@ -15,6 +15,7 @@ | |||
15 | #include <linux/string.h> | 15 | #include <linux/string.h> |
16 | #include <linux/platform_device.h> | 16 | #include <linux/platform_device.h> |
17 | 17 | ||
18 | #include <mach/irqs.h> | ||
18 | #include <mach/map.h> | 19 | #include <mach/map.h> |
19 | 20 | ||
20 | #include <plat/regs-iic.h> | 21 | #include <plat/regs-iic.h> |
diff --git a/arch/arm/plat-s3c/dev-i2c1.c b/arch/arm/plat-s3c/dev-i2c1.c index 9658fb0aec95..2387fbf57af6 100644 --- a/arch/arm/plat-s3c/dev-i2c1.c +++ b/arch/arm/plat-s3c/dev-i2c1.c | |||
@@ -15,6 +15,7 @@ | |||
15 | #include <linux/string.h> | 15 | #include <linux/string.h> |
16 | #include <linux/platform_device.h> | 16 | #include <linux/platform_device.h> |
17 | 17 | ||
18 | #include <mach/irqs.h> | ||
18 | #include <mach/map.h> | 19 | #include <mach/map.h> |
19 | 20 | ||
20 | #include <plat/regs-iic.h> | 21 | #include <plat/regs-iic.h> |
diff --git a/arch/arm/plat-s3c24xx/gpiolib.c b/arch/arm/plat-s3c24xx/gpiolib.c index f95c6c9d9f1a..94a341aaa4e4 100644 --- a/arch/arm/plat-s3c24xx/gpiolib.c +++ b/arch/arm/plat-s3c24xx/gpiolib.c | |||
@@ -59,6 +59,22 @@ static int s3c24xx_gpiolib_banka_output(struct gpio_chip *chip, | |||
59 | return 0; | 59 | return 0; |
60 | } | 60 | } |
61 | 61 | ||
62 | static int s3c24xx_gpiolib_bankf_toirq(struct gpio_chip *chip, unsigned offset) | ||
63 | { | ||
64 | if (offset < 4) | ||
65 | return IRQ_EINT0 + offset; | ||
66 | |||
67 | if (offset < 8) | ||
68 | return IRQ_EINT4 + offset - 4; | ||
69 | |||
70 | return -EINVAL; | ||
71 | } | ||
72 | |||
73 | static int s3c24xx_gpiolib_bankg_toirq(struct gpio_chip *chip, unsigned offset) | ||
74 | { | ||
75 | return IRQ_EINT8 + offset; | ||
76 | } | ||
77 | |||
62 | struct s3c_gpio_chip s3c24xx_gpios[] = { | 78 | struct s3c_gpio_chip s3c24xx_gpios[] = { |
63 | [0] = { | 79 | [0] = { |
64 | .base = S3C24XX_GPIO_BASE(S3C2410_GPA0), | 80 | .base = S3C24XX_GPIO_BASE(S3C2410_GPA0), |
@@ -114,6 +130,7 @@ struct s3c_gpio_chip s3c24xx_gpios[] = { | |||
114 | .owner = THIS_MODULE, | 130 | .owner = THIS_MODULE, |
115 | .label = "GPIOF", | 131 | .label = "GPIOF", |
116 | .ngpio = 8, | 132 | .ngpio = 8, |
133 | .to_irq = s3c24xx_gpiolib_bankf_toirq, | ||
117 | }, | 134 | }, |
118 | }, | 135 | }, |
119 | [6] = { | 136 | [6] = { |
@@ -123,6 +140,7 @@ struct s3c_gpio_chip s3c24xx_gpios[] = { | |||
123 | .owner = THIS_MODULE, | 140 | .owner = THIS_MODULE, |
124 | .label = "GPIOG", | 141 | .label = "GPIOG", |
125 | .ngpio = 10, | 142 | .ngpio = 10, |
143 | .to_irq = s3c24xx_gpiolib_bankg_toirq, | ||
126 | }, | 144 | }, |
127 | }, | 145 | }, |
128 | }; | 146 | }; |
diff --git a/arch/arm/plat-s3c24xx/pwm.c b/arch/arm/plat-s3c24xx/pwm.c index ec56b88866c4..0120b760315b 100644 --- a/arch/arm/plat-s3c24xx/pwm.c +++ b/arch/arm/plat-s3c24xx/pwm.c | |||
@@ -19,6 +19,8 @@ | |||
19 | #include <linux/io.h> | 19 | #include <linux/io.h> |
20 | #include <linux/pwm.h> | 20 | #include <linux/pwm.h> |
21 | 21 | ||
22 | #include <mach/irqs.h> | ||
23 | |||
22 | #include <plat/devs.h> | 24 | #include <plat/devs.h> |
23 | #include <plat/regs-timer.h> | 25 | #include <plat/regs-timer.h> |
24 | 26 | ||
diff --git a/arch/arm/plat-s3c64xx/include/plat/irqs.h b/arch/arm/plat-s3c64xx/include/plat/irqs.h index 02e8dd4c97d5..2846f550b727 100644 --- a/arch/arm/plat-s3c64xx/include/plat/irqs.h +++ b/arch/arm/plat-s3c64xx/include/plat/irqs.h | |||
@@ -191,7 +191,7 @@ | |||
191 | #define IRQ_EINT_GROUP8_BASE (IRQ_EINT_GROUP7_BASE + IRQ_EINT_GROUP7_NR) | 191 | #define IRQ_EINT_GROUP8_BASE (IRQ_EINT_GROUP7_BASE + IRQ_EINT_GROUP7_NR) |
192 | #define IRQ_EINT_GROUP9_BASE (IRQ_EINT_GROUP8_BASE + IRQ_EINT_GROUP8_NR) | 192 | #define IRQ_EINT_GROUP9_BASE (IRQ_EINT_GROUP8_BASE + IRQ_EINT_GROUP8_NR) |
193 | 193 | ||
194 | #define IRQ_EINT_GROUP(group, no) (IRQ_EINT_GROUP##group##__BASE + (x)) | 194 | #define IRQ_EINT_GROUP(group, no) (IRQ_EINT_GROUP##group##_BASE + (no)) |
195 | 195 | ||
196 | /* Set the default NR_IRQS */ | 196 | /* Set the default NR_IRQS */ |
197 | 197 | ||
diff --git a/arch/powerpc/Kconfig b/arch/powerpc/Kconfig index 79f25cef32df..84b861316ce7 100644 --- a/arch/powerpc/Kconfig +++ b/arch/powerpc/Kconfig | |||
@@ -108,6 +108,8 @@ config ARCH_NO_VIRT_TO_BUS | |||
108 | config PPC | 108 | config PPC |
109 | bool | 109 | bool |
110 | default y | 110 | default y |
111 | select HAVE_FTRACE_MCOUNT_RECORD | ||
112 | select HAVE_DYNAMIC_FTRACE | ||
111 | select HAVE_FUNCTION_TRACER | 113 | select HAVE_FUNCTION_TRACER |
112 | select ARCH_WANT_OPTIONAL_GPIOLIB | 114 | select ARCH_WANT_OPTIONAL_GPIOLIB |
113 | select HAVE_IDE | 115 | select HAVE_IDE |
@@ -326,7 +328,8 @@ config KEXEC | |||
326 | 328 | ||
327 | config CRASH_DUMP | 329 | config CRASH_DUMP |
328 | bool "Build a kdump crash kernel" | 330 | bool "Build a kdump crash kernel" |
329 | depends on (PPC64 && RELOCATABLE) || 6xx | 331 | depends on PPC64 || 6xx |
332 | select RELOCATABLE if PPC64 | ||
330 | help | 333 | help |
331 | Build a kernel suitable for use as a kdump capture kernel. | 334 | Build a kernel suitable for use as a kdump capture kernel. |
332 | The same kernel binary can be used as production kernel and dump | 335 | The same kernel binary can be used as production kernel and dump |
diff --git a/arch/powerpc/boot/Makefile b/arch/powerpc/boot/Makefile index ab6dda372438..e84df338ea29 100644 --- a/arch/powerpc/boot/Makefile +++ b/arch/powerpc/boot/Makefile | |||
@@ -356,7 +356,7 @@ $(obj)/zImage.initrd: $(addprefix $(obj)/, $(initrd-y)) | |||
356 | @rm -f $@; ln $< $@ | 356 | @rm -f $@; ln $< $@ |
357 | 357 | ||
358 | install: $(CONFIGURE) $(addprefix $(obj)/, $(image-y)) | 358 | install: $(CONFIGURE) $(addprefix $(obj)/, $(image-y)) |
359 | sh -x $(srctree)/$(src)/install.sh "$(KERNELRELEASE)" vmlinux System.map "$(INSTALL_PATH)" $< | 359 | sh -x $(srctree)/$(src)/install.sh "$(KERNELRELEASE)" vmlinux System.map "$(INSTALL_PATH)" $^ |
360 | 360 | ||
361 | # anything not in $(targets) | 361 | # anything not in $(targets) |
362 | clean-files += $(image-) $(initrd-) cuImage.* dtbImage.* treeImage.* \ | 362 | clean-files += $(image-) $(initrd-) cuImage.* dtbImage.* treeImage.* \ |
diff --git a/arch/powerpc/boot/dts/mpc836x_mds.dts b/arch/powerpc/boot/dts/mpc836x_mds.dts index 14534d04e4db..6e34f170fa62 100644 --- a/arch/powerpc/boot/dts/mpc836x_mds.dts +++ b/arch/powerpc/boot/dts/mpc836x_mds.dts | |||
@@ -69,8 +69,18 @@ | |||
69 | }; | 69 | }; |
70 | 70 | ||
71 | bcsr@1,0 { | 71 | bcsr@1,0 { |
72 | #address-cells = <1>; | ||
73 | #size-cells = <1>; | ||
72 | compatible = "fsl,mpc8360mds-bcsr"; | 74 | compatible = "fsl,mpc8360mds-bcsr"; |
73 | reg = <1 0 0x8000>; | 75 | reg = <1 0 0x8000>; |
76 | ranges = <0 1 0 0x8000>; | ||
77 | |||
78 | bcsr13: gpio-controller@d { | ||
79 | #gpio-cells = <2>; | ||
80 | compatible = "fsl,mpc8360mds-bcsr-gpio"; | ||
81 | reg = <0xd 1>; | ||
82 | gpio-controller; | ||
83 | }; | ||
74 | }; | 84 | }; |
75 | }; | 85 | }; |
76 | 86 | ||
@@ -195,10 +205,21 @@ | |||
195 | }; | 205 | }; |
196 | 206 | ||
197 | par_io@1400 { | 207 | par_io@1400 { |
208 | #address-cells = <1>; | ||
209 | #size-cells = <1>; | ||
198 | reg = <0x1400 0x100>; | 210 | reg = <0x1400 0x100>; |
211 | ranges = <0 0x1400 0x100>; | ||
199 | device_type = "par_io"; | 212 | device_type = "par_io"; |
200 | num-ports = <7>; | 213 | num-ports = <7>; |
201 | 214 | ||
215 | qe_pio_b: gpio-controller@18 { | ||
216 | #gpio-cells = <2>; | ||
217 | compatible = "fsl,mpc8360-qe-pario-bank", | ||
218 | "fsl,mpc8323-qe-pario-bank"; | ||
219 | reg = <0x18 0x18>; | ||
220 | gpio-controller; | ||
221 | }; | ||
222 | |||
202 | pio1: ucc_pin@01 { | 223 | pio1: ucc_pin@01 { |
203 | pio-map = < | 224 | pio-map = < |
204 | /* port pin dir open_drain assignment has_irq */ | 225 | /* port pin dir open_drain assignment has_irq */ |
@@ -282,6 +303,15 @@ | |||
282 | }; | 303 | }; |
283 | }; | 304 | }; |
284 | 305 | ||
306 | timer@440 { | ||
307 | compatible = "fsl,mpc8360-qe-gtm", | ||
308 | "fsl,qe-gtm", "fsl,gtm"; | ||
309 | reg = <0x440 0x40>; | ||
310 | clock-frequency = <132000000>; | ||
311 | interrupts = <12 13 14 15>; | ||
312 | interrupt-parent = <&qeic>; | ||
313 | }; | ||
314 | |||
285 | spi@4c0 { | 315 | spi@4c0 { |
286 | cell-index = <0>; | 316 | cell-index = <0>; |
287 | compatible = "fsl,spi"; | 317 | compatible = "fsl,spi"; |
@@ -301,11 +331,20 @@ | |||
301 | }; | 331 | }; |
302 | 332 | ||
303 | usb@6c0 { | 333 | usb@6c0 { |
304 | compatible = "qe_udc"; | 334 | compatible = "fsl,mpc8360-qe-usb", |
335 | "fsl,mpc8323-qe-usb"; | ||
305 | reg = <0x6c0 0x40 0x8b00 0x100>; | 336 | reg = <0x6c0 0x40 0x8b00 0x100>; |
306 | interrupts = <11>; | 337 | interrupts = <11>; |
307 | interrupt-parent = <&qeic>; | 338 | interrupt-parent = <&qeic>; |
308 | mode = "slave"; | 339 | fsl,fullspeed-clock = "clk21"; |
340 | fsl,lowspeed-clock = "brg9"; | ||
341 | gpios = <&qe_pio_b 2 0 /* USBOE */ | ||
342 | &qe_pio_b 3 0 /* USBTP */ | ||
343 | &qe_pio_b 8 0 /* USBTN */ | ||
344 | &qe_pio_b 9 0 /* USBRP */ | ||
345 | &qe_pio_b 11 0 /* USBRN */ | ||
346 | &bcsr13 5 0 /* SPEED */ | ||
347 | &bcsr13 4 1>; /* POWER */ | ||
309 | }; | 348 | }; |
310 | 349 | ||
311 | enet0: ucc@2000 { | 350 | enet0: ucc@2000 { |
diff --git a/arch/powerpc/boot/dts/mpc836x_rdk.dts b/arch/powerpc/boot/dts/mpc836x_rdk.dts index decadf3d9e98..37b789510d68 100644 --- a/arch/powerpc/boot/dts/mpc836x_rdk.dts +++ b/arch/powerpc/boot/dts/mpc836x_rdk.dts | |||
@@ -218,8 +218,23 @@ | |||
218 | reg = <0x440 0x40>; | 218 | reg = <0x440 0x40>; |
219 | interrupts = <12 13 14 15>; | 219 | interrupts = <12 13 14 15>; |
220 | interrupt-parent = <&qeic>; | 220 | interrupt-parent = <&qeic>; |
221 | /* filled by u-boot */ | 221 | clock-frequency = <166666666>; |
222 | clock-frequency = <0>; | 222 | }; |
223 | |||
224 | usb@6c0 { | ||
225 | compatible = "fsl,mpc8360-qe-usb", | ||
226 | "fsl,mpc8323-qe-usb"; | ||
227 | reg = <0x6c0 0x40 0x8b00 0x100>; | ||
228 | interrupts = <11>; | ||
229 | interrupt-parent = <&qeic>; | ||
230 | fsl,fullspeed-clock = "clk21"; | ||
231 | gpios = <&qe_pio_b 2 0 /* USBOE */ | ||
232 | &qe_pio_b 3 0 /* USBTP */ | ||
233 | &qe_pio_b 8 0 /* USBTN */ | ||
234 | &qe_pio_b 9 0 /* USBRP */ | ||
235 | &qe_pio_b 11 0 /* USBRN */ | ||
236 | &qe_pio_e 20 0 /* SPEED */ | ||
237 | &qe_pio_e 21 1 /* POWER */>; | ||
223 | }; | 238 | }; |
224 | 239 | ||
225 | spi@4c0 { | 240 | spi@4c0 { |
diff --git a/arch/powerpc/boot/dts/mpc8641_hpcn.dts b/arch/powerpc/boot/dts/mpc8641_hpcn.dts index 35d5e248ccd7..4481532cbe77 100644 --- a/arch/powerpc/boot/dts/mpc8641_hpcn.dts +++ b/arch/powerpc/boot/dts/mpc8641_hpcn.dts | |||
@@ -26,7 +26,13 @@ | |||
26 | serial1 = &serial1; | 26 | serial1 = &serial1; |
27 | pci0 = &pci0; | 27 | pci0 = &pci0; |
28 | pci1 = &pci1; | 28 | pci1 = &pci1; |
29 | rapidio0 = &rapidio0; | 29 | /* |
30 | * Only one of Rapid IO or PCI can be present due to HW limitations and | ||
31 | * due to the fact that the 2 now share address space in the new memory | ||
32 | * map. The most likely case is that we have PCI, so comment out the | ||
33 | * rapidio node. Leave it here for reference. | ||
34 | */ | ||
35 | /* rapidio0 = &rapidio0; */ | ||
30 | }; | 36 | }; |
31 | 37 | ||
32 | cpus { | 38 | cpus { |
@@ -62,18 +68,17 @@ | |||
62 | reg = <0x00000000 0x40000000>; // 1G at 0x0 | 68 | reg = <0x00000000 0x40000000>; // 1G at 0x0 |
63 | }; | 69 | }; |
64 | 70 | ||
65 | localbus@f8005000 { | 71 | localbus@ffe05000 { |
66 | #address-cells = <2>; | 72 | #address-cells = <2>; |
67 | #size-cells = <1>; | 73 | #size-cells = <1>; |
68 | compatible = "fsl,mpc8641-localbus", "simple-bus"; | 74 | compatible = "fsl,mpc8641-localbus", "simple-bus"; |
69 | reg = <0xf8005000 0x1000>; | 75 | reg = <0xffe05000 0x1000>; |
70 | interrupts = <19 2>; | 76 | interrupts = <19 2>; |
71 | interrupt-parent = <&mpic>; | 77 | interrupt-parent = <&mpic>; |
72 | 78 | ||
73 | ranges = <0 0 0xff800000 0x00800000 | 79 | ranges = <0 0 0xef800000 0x00800000 |
74 | 1 0 0xfe000000 0x01000000 | 80 | 2 0 0xffdf8000 0x00008000 |
75 | 2 0 0xf8200000 0x00100000 | 81 | 3 0 0xffdf0000 0x00008000>; |
76 | 3 0 0xf8100000 0x00100000>; | ||
77 | 82 | ||
78 | flash@0,0 { | 83 | flash@0,0 { |
79 | compatible = "cfi-flash"; | 84 | compatible = "cfi-flash"; |
@@ -103,13 +108,13 @@ | |||
103 | }; | 108 | }; |
104 | }; | 109 | }; |
105 | 110 | ||
106 | soc8641@f8000000 { | 111 | soc8641@ffe00000 { |
107 | #address-cells = <1>; | 112 | #address-cells = <1>; |
108 | #size-cells = <1>; | 113 | #size-cells = <1>; |
109 | device_type = "soc"; | 114 | device_type = "soc"; |
110 | compatible = "simple-bus"; | 115 | compatible = "simple-bus"; |
111 | ranges = <0x00000000 0xf8000000 0x00100000>; | 116 | ranges = <0x00000000 0xffe00000 0x00100000>; |
112 | reg = <0xf8000000 0x00001000>; // CCSRBAR | 117 | reg = <0xffe00000 0x00001000>; // CCSRBAR |
113 | bus-frequency = <0>; | 118 | bus-frequency = <0>; |
114 | 119 | ||
115 | i2c@3000 { | 120 | i2c@3000 { |
@@ -340,17 +345,17 @@ | |||
340 | }; | 345 | }; |
341 | }; | 346 | }; |
342 | 347 | ||
343 | pci0: pcie@f8008000 { | 348 | pci0: pcie@ffe08000 { |
344 | cell-index = <0>; | 349 | cell-index = <0>; |
345 | compatible = "fsl,mpc8641-pcie"; | 350 | compatible = "fsl,mpc8641-pcie"; |
346 | device_type = "pci"; | 351 | device_type = "pci"; |
347 | #interrupt-cells = <1>; | 352 | #interrupt-cells = <1>; |
348 | #size-cells = <2>; | 353 | #size-cells = <2>; |
349 | #address-cells = <3>; | 354 | #address-cells = <3>; |
350 | reg = <0xf8008000 0x1000>; | 355 | reg = <0xffe08000 0x1000>; |
351 | bus-range = <0x0 0xff>; | 356 | bus-range = <0x0 0xff>; |
352 | ranges = <0x02000000 0x0 0x80000000 0x80000000 0x0 0x20000000 | 357 | ranges = <0x02000000 0x0 0x80000000 0x80000000 0x0 0x20000000 |
353 | 0x01000000 0x0 0x00000000 0xe2000000 0x0 0x00100000>; | 358 | 0x01000000 0x0 0x00000000 0xffc00000 0x0 0x00010000>; |
354 | clock-frequency = <33333333>; | 359 | clock-frequency = <33333333>; |
355 | interrupt-parent = <&mpic>; | 360 | interrupt-parent = <&mpic>; |
356 | interrupts = <24 2>; | 361 | interrupts = <24 2>; |
@@ -481,7 +486,7 @@ | |||
481 | 486 | ||
482 | 0x01000000 0x0 0x00000000 | 487 | 0x01000000 0x0 0x00000000 |
483 | 0x01000000 0x0 0x00000000 | 488 | 0x01000000 0x0 0x00000000 |
484 | 0x0 0x00100000>; | 489 | 0x0 0x00010000>; |
485 | uli1575@0 { | 490 | uli1575@0 { |
486 | reg = <0 0 0 0 0>; | 491 | reg = <0 0 0 0 0>; |
487 | #size-cells = <2>; | 492 | #size-cells = <2>; |
@@ -491,7 +496,7 @@ | |||
491 | 0x0 0x20000000 | 496 | 0x0 0x20000000 |
492 | 0x01000000 0x0 0x00000000 | 497 | 0x01000000 0x0 0x00000000 |
493 | 0x01000000 0x0 0x00000000 | 498 | 0x01000000 0x0 0x00000000 |
494 | 0x0 0x00100000>; | 499 | 0x0 0x00010000>; |
495 | isa@1e { | 500 | isa@1e { |
496 | device_type = "isa"; | 501 | device_type = "isa"; |
497 | #interrupt-cells = <2>; | 502 | #interrupt-cells = <2>; |
@@ -549,17 +554,17 @@ | |||
549 | 554 | ||
550 | }; | 555 | }; |
551 | 556 | ||
552 | pci1: pcie@f8009000 { | 557 | pci1: pcie@ffe09000 { |
553 | cell-index = <1>; | 558 | cell-index = <1>; |
554 | compatible = "fsl,mpc8641-pcie"; | 559 | compatible = "fsl,mpc8641-pcie"; |
555 | device_type = "pci"; | 560 | device_type = "pci"; |
556 | #interrupt-cells = <1>; | 561 | #interrupt-cells = <1>; |
557 | #size-cells = <2>; | 562 | #size-cells = <2>; |
558 | #address-cells = <3>; | 563 | #address-cells = <3>; |
559 | reg = <0xf8009000 0x1000>; | 564 | reg = <0xffe09000 0x1000>; |
560 | bus-range = <0 0xff>; | 565 | bus-range = <0 0xff>; |
561 | ranges = <0x02000000 0x0 0xa0000000 0xa0000000 0x0 0x20000000 | 566 | ranges = <0x02000000 0x0 0xa0000000 0xa0000000 0x0 0x20000000 |
562 | 0x01000000 0x0 0x00000000 0xe3000000 0x0 0x00100000>; | 567 | 0x01000000 0x0 0x00000000 0xffc10000 0x0 0x00010000>; |
563 | clock-frequency = <33333333>; | 568 | clock-frequency = <33333333>; |
564 | interrupt-parent = <&mpic>; | 569 | interrupt-parent = <&mpic>; |
565 | interrupts = <25 2>; | 570 | interrupts = <25 2>; |
@@ -582,18 +587,21 @@ | |||
582 | 587 | ||
583 | 0x01000000 0x0 0x00000000 | 588 | 0x01000000 0x0 0x00000000 |
584 | 0x01000000 0x0 0x00000000 | 589 | 0x01000000 0x0 0x00000000 |
585 | 0x0 0x00100000>; | 590 | 0x0 0x00010000>; |
586 | }; | 591 | }; |
587 | }; | 592 | }; |
588 | rapidio0: rapidio@f80c0000 { | 593 | /* |
594 | rapidio0: rapidio@ffec0000 { | ||
589 | #address-cells = <2>; | 595 | #address-cells = <2>; |
590 | #size-cells = <2>; | 596 | #size-cells = <2>; |
591 | compatible = "fsl,rapidio-delta"; | 597 | compatible = "fsl,rapidio-delta"; |
592 | reg = <0xf80c0000 0x20000>; | 598 | reg = <0xffec0000 0x20000>; |
593 | ranges = <0 0 0xc0000000 0 0x20000000>; | 599 | ranges = <0 0 0x80000000 0 0x20000000>; |
594 | interrupt-parent = <&mpic>; | 600 | interrupt-parent = <&mpic>; |
595 | /* err_irq bell_outb_irq bell_inb_irq | 601 | // err_irq bell_outb_irq bell_inb_irq |
596 | msg1_tx_irq msg1_rx_irq msg2_tx_irq msg2_rx_irq */ | 602 | // msg1_tx_irq msg1_rx_irq msg2_tx_irq msg2_rx_irq |
597 | interrupts = <48 2 49 2 50 2 53 2 54 2 55 2 56 2>; | 603 | interrupts = <48 2 49 2 50 2 53 2 54 2 55 2 56 2>; |
598 | }; | 604 | }; |
605 | */ | ||
606 | |||
599 | }; | 607 | }; |
diff --git a/arch/powerpc/boot/install.sh b/arch/powerpc/boot/install.sh index b002bfd56786..51b2387bdba0 100644 --- a/arch/powerpc/boot/install.sh +++ b/arch/powerpc/boot/install.sh | |||
@@ -15,7 +15,7 @@ | |||
15 | # $2 - kernel image file | 15 | # $2 - kernel image file |
16 | # $3 - kernel map file | 16 | # $3 - kernel map file |
17 | # $4 - default install path (blank if root directory) | 17 | # $4 - default install path (blank if root directory) |
18 | # $5 - kernel boot file, the zImage | 18 | # $5 and more - kernel boot files; zImage*, uImage, cuImage.*, etc. |
19 | # | 19 | # |
20 | 20 | ||
21 | # User may have a custom install script | 21 | # User may have a custom install script |
@@ -38,3 +38,15 @@ fi | |||
38 | 38 | ||
39 | cat $2 > $4/$image_name | 39 | cat $2 > $4/$image_name |
40 | cp $3 $4/System.map | 40 | cp $3 $4/System.map |
41 | |||
42 | # Copy all the bootable image files | ||
43 | path=$4 | ||
44 | shift 4 | ||
45 | while [ $# -ne 0 ]; do | ||
46 | image_name=`basename $1` | ||
47 | if [ -f $path/$image_name ]; then | ||
48 | mv $path/$image_name $path/$image_name.old | ||
49 | fi | ||
50 | cat $1 > $path/$image_name | ||
51 | shift | ||
52 | done; | ||
diff --git a/arch/powerpc/configs/85xx/mpc8572_ds_defconfig b/arch/powerpc/configs/85xx/mpc8572_ds_defconfig index 635588319e0d..32aeb79216f7 100644 --- a/arch/powerpc/configs/85xx/mpc8572_ds_defconfig +++ b/arch/powerpc/configs/85xx/mpc8572_ds_defconfig | |||
@@ -1,7 +1,7 @@ | |||
1 | # | 1 | # |
2 | # Automatically generated make config: don't edit | 2 | # Automatically generated make config: don't edit |
3 | # Linux kernel version: 2.6.28-rc3 | 3 | # Linux kernel version: 2.6.28-rc8 |
4 | # Sat Nov 8 12:40:13 2008 | 4 | # Tue Dec 30 11:17:46 2008 |
5 | # | 5 | # |
6 | # CONFIG_PPC64 is not set | 6 | # CONFIG_PPC64 is not set |
7 | 7 | ||
@@ -21,7 +21,10 @@ CONFIG_FSL_BOOKE=y | |||
21 | CONFIG_FSL_EMB_PERFMON=y | 21 | CONFIG_FSL_EMB_PERFMON=y |
22 | # CONFIG_PHYS_64BIT is not set | 22 | # CONFIG_PHYS_64BIT is not set |
23 | CONFIG_SPE=y | 23 | CONFIG_SPE=y |
24 | CONFIG_PPC_MMU_NOHASH=y | ||
24 | # CONFIG_PPC_MM_SLICES is not set | 25 | # CONFIG_PPC_MM_SLICES is not set |
26 | CONFIG_SMP=y | ||
27 | CONFIG_NR_CPUS=2 | ||
25 | CONFIG_PPC32=y | 28 | CONFIG_PPC32=y |
26 | CONFIG_WORD_SIZE=32 | 29 | CONFIG_WORD_SIZE=32 |
27 | # CONFIG_ARCH_PHYS_ADDR_T_64BIT is not set | 30 | # CONFIG_ARCH_PHYS_ADDR_T_64BIT is not set |
@@ -50,7 +53,7 @@ CONFIG_ARCH_MAY_HAVE_PC_FDC=y | |||
50 | CONFIG_PPC_OF=y | 53 | CONFIG_PPC_OF=y |
51 | CONFIG_OF=y | 54 | CONFIG_OF=y |
52 | CONFIG_PPC_UDBG_16550=y | 55 | CONFIG_PPC_UDBG_16550=y |
53 | # CONFIG_GENERIC_TBSYNC is not set | 56 | CONFIG_GENERIC_TBSYNC=y |
54 | CONFIG_AUDIT_ARCH=y | 57 | CONFIG_AUDIT_ARCH=y |
55 | CONFIG_GENERIC_BUG=y | 58 | CONFIG_GENERIC_BUG=y |
56 | CONFIG_DEFAULT_UIMAGE=y | 59 | CONFIG_DEFAULT_UIMAGE=y |
@@ -62,7 +65,7 @@ CONFIG_DEFCONFIG_LIST="/lib/modules/$UNAME_RELEASE/.config" | |||
62 | # General setup | 65 | # General setup |
63 | # | 66 | # |
64 | CONFIG_EXPERIMENTAL=y | 67 | CONFIG_EXPERIMENTAL=y |
65 | CONFIG_BROKEN_ON_SMP=y | 68 | CONFIG_LOCK_KERNEL=y |
66 | CONFIG_INIT_ENV_ARG_LIMIT=32 | 69 | CONFIG_INIT_ENV_ARG_LIMIT=32 |
67 | CONFIG_LOCALVERSION="" | 70 | CONFIG_LOCALVERSION="" |
68 | CONFIG_LOCALVERSION_AUTO=y | 71 | CONFIG_LOCALVERSION_AUTO=y |
@@ -126,6 +129,7 @@ CONFIG_HAVE_IOREMAP_PROT=y | |||
126 | CONFIG_HAVE_KPROBES=y | 129 | CONFIG_HAVE_KPROBES=y |
127 | CONFIG_HAVE_KRETPROBES=y | 130 | CONFIG_HAVE_KRETPROBES=y |
128 | CONFIG_HAVE_ARCH_TRACEHOOK=y | 131 | CONFIG_HAVE_ARCH_TRACEHOOK=y |
132 | CONFIG_USE_GENERIC_SMP_HELPERS=y | ||
129 | # CONFIG_HAVE_GENERIC_DMA_COHERENT is not set | 133 | # CONFIG_HAVE_GENERIC_DMA_COHERENT is not set |
130 | CONFIG_SLABINFO=y | 134 | CONFIG_SLABINFO=y |
131 | CONFIG_RT_MUTEXES=y | 135 | CONFIG_RT_MUTEXES=y |
@@ -138,6 +142,7 @@ CONFIG_MODULE_FORCE_UNLOAD=y | |||
138 | CONFIG_MODVERSIONS=y | 142 | CONFIG_MODVERSIONS=y |
139 | # CONFIG_MODULE_SRCVERSION_ALL is not set | 143 | # CONFIG_MODULE_SRCVERSION_ALL is not set |
140 | CONFIG_KMOD=y | 144 | CONFIG_KMOD=y |
145 | CONFIG_STOP_MACHINE=y | ||
141 | CONFIG_BLOCK=y | 146 | CONFIG_BLOCK=y |
142 | CONFIG_LBD=y | 147 | CONFIG_LBD=y |
143 | # CONFIG_BLK_DEV_IO_TRACE is not set | 148 | # CONFIG_BLK_DEV_IO_TRACE is not set |
@@ -197,6 +202,7 @@ CONFIG_PPC_I8259=y | |||
197 | # CONFIG_CPM2 is not set | 202 | # CONFIG_CPM2 is not set |
198 | CONFIG_FSL_ULI1575=y | 203 | CONFIG_FSL_ULI1575=y |
199 | # CONFIG_MPC8xxx_GPIO is not set | 204 | # CONFIG_MPC8xxx_GPIO is not set |
205 | # CONFIG_SIMPLE_GPIO is not set | ||
200 | 206 | ||
201 | # | 207 | # |
202 | # Kernel options | 208 | # Kernel options |
@@ -224,6 +230,7 @@ CONFIG_MATH_EMULATION=y | |||
224 | CONFIG_ARCH_ENABLE_MEMORY_HOTPLUG=y | 230 | CONFIG_ARCH_ENABLE_MEMORY_HOTPLUG=y |
225 | CONFIG_ARCH_HAS_WALK_MEMORY=y | 231 | CONFIG_ARCH_HAS_WALK_MEMORY=y |
226 | CONFIG_ARCH_ENABLE_MEMORY_HOTREMOVE=y | 232 | CONFIG_ARCH_ENABLE_MEMORY_HOTREMOVE=y |
233 | # CONFIG_IRQ_ALL_CPUS is not set | ||
227 | CONFIG_ARCH_FLATMEM_ENABLE=y | 234 | CONFIG_ARCH_FLATMEM_ENABLE=y |
228 | CONFIG_ARCH_POPULATES_NODE_MAP=y | 235 | CONFIG_ARCH_POPULATES_NODE_MAP=y |
229 | CONFIG_SELECT_MEMORY_MODEL=y | 236 | CONFIG_SELECT_MEMORY_MODEL=y |
@@ -241,6 +248,9 @@ CONFIG_ZONE_DMA_FLAG=1 | |||
241 | CONFIG_BOUNCE=y | 248 | CONFIG_BOUNCE=y |
242 | CONFIG_VIRT_TO_BUS=y | 249 | CONFIG_VIRT_TO_BUS=y |
243 | CONFIG_UNEVICTABLE_LRU=y | 250 | CONFIG_UNEVICTABLE_LRU=y |
251 | CONFIG_PPC_4K_PAGES=y | ||
252 | # CONFIG_PPC_16K_PAGES is not set | ||
253 | # CONFIG_PPC_64K_PAGES is not set | ||
244 | CONFIG_FORCE_MAX_ZONEORDER=11 | 254 | CONFIG_FORCE_MAX_ZONEORDER=11 |
245 | CONFIG_PROC_DEVICETREE=y | 255 | CONFIG_PROC_DEVICETREE=y |
246 | # CONFIG_CMDLINE_BOOL is not set | 256 | # CONFIG_CMDLINE_BOOL is not set |
@@ -443,8 +453,10 @@ CONFIG_MISC_DEVICES=y | |||
443 | # CONFIG_EEPROM_93CX6 is not set | 453 | # CONFIG_EEPROM_93CX6 is not set |
444 | # CONFIG_SGI_IOC4 is not set | 454 | # CONFIG_SGI_IOC4 is not set |
445 | # CONFIG_TIFM_CORE is not set | 455 | # CONFIG_TIFM_CORE is not set |
456 | # CONFIG_ICS932S401 is not set | ||
446 | # CONFIG_ENCLOSURE_SERVICES is not set | 457 | # CONFIG_ENCLOSURE_SERVICES is not set |
447 | # CONFIG_HP_ILO is not set | 458 | # CONFIG_HP_ILO is not set |
459 | # CONFIG_C2PORT is not set | ||
448 | CONFIG_HAVE_IDE=y | 460 | CONFIG_HAVE_IDE=y |
449 | # CONFIG_IDE is not set | 461 | # CONFIG_IDE is not set |
450 | 462 | ||
@@ -784,6 +796,7 @@ CONFIG_SERIAL_CORE_CONSOLE=y | |||
784 | CONFIG_UNIX98_PTYS=y | 796 | CONFIG_UNIX98_PTYS=y |
785 | CONFIG_LEGACY_PTYS=y | 797 | CONFIG_LEGACY_PTYS=y |
786 | CONFIG_LEGACY_PTY_COUNT=256 | 798 | CONFIG_LEGACY_PTY_COUNT=256 |
799 | # CONFIG_HVC_UDBG is not set | ||
787 | # CONFIG_IPMI_HANDLER is not set | 800 | # CONFIG_IPMI_HANDLER is not set |
788 | CONFIG_HW_RANDOM=y | 801 | CONFIG_HW_RANDOM=y |
789 | CONFIG_NVRAM=y | 802 | CONFIG_NVRAM=y |
@@ -869,11 +882,11 @@ CONFIG_ARCH_WANT_OPTIONAL_GPIOLIB=y | |||
869 | # CONFIG_THERMAL is not set | 882 | # CONFIG_THERMAL is not set |
870 | # CONFIG_THERMAL_HWMON is not set | 883 | # CONFIG_THERMAL_HWMON is not set |
871 | # CONFIG_WATCHDOG is not set | 884 | # CONFIG_WATCHDOG is not set |
885 | CONFIG_SSB_POSSIBLE=y | ||
872 | 886 | ||
873 | # | 887 | # |
874 | # Sonics Silicon Backplane | 888 | # Sonics Silicon Backplane |
875 | # | 889 | # |
876 | CONFIG_SSB_POSSIBLE=y | ||
877 | # CONFIG_SSB is not set | 890 | # CONFIG_SSB is not set |
878 | 891 | ||
879 | # | 892 | # |
@@ -886,14 +899,7 @@ CONFIG_SSB_POSSIBLE=y | |||
886 | # CONFIG_PMIC_DA903X is not set | 899 | # CONFIG_PMIC_DA903X is not set |
887 | # CONFIG_MFD_WM8400 is not set | 900 | # CONFIG_MFD_WM8400 is not set |
888 | # CONFIG_MFD_WM8350_I2C is not set | 901 | # CONFIG_MFD_WM8350_I2C is not set |
889 | |||
890 | # | ||
891 | # Voltage and Current regulators | ||
892 | # | ||
893 | # CONFIG_REGULATOR is not set | 902 | # CONFIG_REGULATOR is not set |
894 | # CONFIG_REGULATOR_FIXED_VOLTAGE is not set | ||
895 | # CONFIG_REGULATOR_VIRTUAL_CONSUMER is not set | ||
896 | # CONFIG_REGULATOR_BQ24022 is not set | ||
897 | 903 | ||
898 | # | 904 | # |
899 | # Multimedia devices | 905 | # Multimedia devices |
@@ -1252,11 +1258,11 @@ CONFIG_USB_OHCI_LITTLE_ENDIAN=y | |||
1252 | # CONFIG_USB_TMC is not set | 1258 | # CONFIG_USB_TMC is not set |
1253 | 1259 | ||
1254 | # | 1260 | # |
1255 | # NOTE: USB_STORAGE enables SCSI, and 'SCSI disk support' | 1261 | # NOTE: USB_STORAGE depends on SCSI but BLK_DEV_SD may also be needed; |
1256 | # | 1262 | # |
1257 | 1263 | ||
1258 | # | 1264 | # |
1259 | # may also be needed; see USB_STORAGE Help for more information | 1265 | # see USB_STORAGE Help for more information |
1260 | # | 1266 | # |
1261 | CONFIG_USB_STORAGE=y | 1267 | CONFIG_USB_STORAGE=y |
1262 | # CONFIG_USB_STORAGE_DEBUG is not set | 1268 | # CONFIG_USB_STORAGE_DEBUG is not set |
@@ -1348,6 +1354,7 @@ CONFIG_RTC_INTF_DEV=y | |||
1348 | # CONFIG_RTC_DRV_M41T80 is not set | 1354 | # CONFIG_RTC_DRV_M41T80 is not set |
1349 | # CONFIG_RTC_DRV_S35390A is not set | 1355 | # CONFIG_RTC_DRV_S35390A is not set |
1350 | # CONFIG_RTC_DRV_FM3130 is not set | 1356 | # CONFIG_RTC_DRV_FM3130 is not set |
1357 | # CONFIG_RTC_DRV_RX8581 is not set | ||
1351 | 1358 | ||
1352 | # | 1359 | # |
1353 | # SPI RTC drivers | 1360 | # SPI RTC drivers |
@@ -1624,6 +1631,7 @@ CONFIG_HAVE_FUNCTION_TRACER=y | |||
1624 | # CONFIG_SAMPLES is not set | 1631 | # CONFIG_SAMPLES is not set |
1625 | CONFIG_HAVE_ARCH_KGDB=y | 1632 | CONFIG_HAVE_ARCH_KGDB=y |
1626 | # CONFIG_KGDB is not set | 1633 | # CONFIG_KGDB is not set |
1634 | CONFIG_PRINT_STACK_DEPTH=64 | ||
1627 | # CONFIG_DEBUG_STACKOVERFLOW is not set | 1635 | # CONFIG_DEBUG_STACKOVERFLOW is not set |
1628 | # CONFIG_DEBUG_STACK_USAGE is not set | 1636 | # CONFIG_DEBUG_STACK_USAGE is not set |
1629 | # CONFIG_DEBUG_PAGEALLOC is not set | 1637 | # CONFIG_DEBUG_PAGEALLOC is not set |
@@ -1649,11 +1657,16 @@ CONFIG_CRYPTO=y | |||
1649 | # | 1657 | # |
1650 | # CONFIG_CRYPTO_FIPS is not set | 1658 | # CONFIG_CRYPTO_FIPS is not set |
1651 | CONFIG_CRYPTO_ALGAPI=y | 1659 | CONFIG_CRYPTO_ALGAPI=y |
1660 | CONFIG_CRYPTO_ALGAPI2=y | ||
1652 | CONFIG_CRYPTO_AEAD=y | 1661 | CONFIG_CRYPTO_AEAD=y |
1662 | CONFIG_CRYPTO_AEAD2=y | ||
1653 | CONFIG_CRYPTO_BLKCIPHER=y | 1663 | CONFIG_CRYPTO_BLKCIPHER=y |
1664 | CONFIG_CRYPTO_BLKCIPHER2=y | ||
1654 | CONFIG_CRYPTO_HASH=y | 1665 | CONFIG_CRYPTO_HASH=y |
1655 | CONFIG_CRYPTO_RNG=y | 1666 | CONFIG_CRYPTO_HASH2=y |
1667 | CONFIG_CRYPTO_RNG2=y | ||
1656 | CONFIG_CRYPTO_MANAGER=y | 1668 | CONFIG_CRYPTO_MANAGER=y |
1669 | CONFIG_CRYPTO_MANAGER2=y | ||
1657 | # CONFIG_CRYPTO_GF128MUL is not set | 1670 | # CONFIG_CRYPTO_GF128MUL is not set |
1658 | # CONFIG_CRYPTO_NULL is not set | 1671 | # CONFIG_CRYPTO_NULL is not set |
1659 | # CONFIG_CRYPTO_CRYPTD is not set | 1672 | # CONFIG_CRYPTO_CRYPTD is not set |
diff --git a/arch/powerpc/include/asm/ioctls.h b/arch/powerpc/include/asm/ioctls.h index 279a6229584b..1842186d872c 100644 --- a/arch/powerpc/include/asm/ioctls.h +++ b/arch/powerpc/include/asm/ioctls.h | |||
@@ -89,6 +89,8 @@ | |||
89 | #define TIOCSBRK 0x5427 /* BSD compatibility */ | 89 | #define TIOCSBRK 0x5427 /* BSD compatibility */ |
90 | #define TIOCCBRK 0x5428 /* BSD compatibility */ | 90 | #define TIOCCBRK 0x5428 /* BSD compatibility */ |
91 | #define TIOCGSID 0x5429 /* Return the session ID of FD */ | 91 | #define TIOCGSID 0x5429 /* Return the session ID of FD */ |
92 | #define TIOCGRS485 0x542e | ||
93 | #define TIOCSRS485 0x542f | ||
92 | #define TIOCGPTN _IOR('T',0x30, unsigned int) /* Get Pty Number (of pty-mux device) */ | 94 | #define TIOCGPTN _IOR('T',0x30, unsigned int) /* Get Pty Number (of pty-mux device) */ |
93 | #define TIOCSPTLCK _IOW('T',0x31, int) /* Lock/unlock Pty */ | 95 | #define TIOCSPTLCK _IOW('T',0x31, int) /* Lock/unlock Pty */ |
94 | 96 | ||
diff --git a/arch/powerpc/include/asm/kexec.h b/arch/powerpc/include/asm/kexec.h index 6dbffc981702..7e06b43720d3 100644 --- a/arch/powerpc/include/asm/kexec.h +++ b/arch/powerpc/include/asm/kexec.h | |||
@@ -48,63 +48,8 @@ static inline void crash_setup_regs(struct pt_regs *newregs, | |||
48 | { | 48 | { |
49 | if (oldregs) | 49 | if (oldregs) |
50 | memcpy(newregs, oldregs, sizeof(*newregs)); | 50 | memcpy(newregs, oldregs, sizeof(*newregs)); |
51 | #ifdef __powerpc64__ | ||
52 | else { | ||
53 | /* FIXME Merge this with xmon_save_regs ?? */ | ||
54 | unsigned long tmp1, tmp2; | ||
55 | __asm__ __volatile__ ( | ||
56 | "std 0,0(%2)\n" | ||
57 | "std 1,8(%2)\n" | ||
58 | "std 2,16(%2)\n" | ||
59 | "std 3,24(%2)\n" | ||
60 | "std 4,32(%2)\n" | ||
61 | "std 5,40(%2)\n" | ||
62 | "std 6,48(%2)\n" | ||
63 | "std 7,56(%2)\n" | ||
64 | "std 8,64(%2)\n" | ||
65 | "std 9,72(%2)\n" | ||
66 | "std 10,80(%2)\n" | ||
67 | "std 11,88(%2)\n" | ||
68 | "std 12,96(%2)\n" | ||
69 | "std 13,104(%2)\n" | ||
70 | "std 14,112(%2)\n" | ||
71 | "std 15,120(%2)\n" | ||
72 | "std 16,128(%2)\n" | ||
73 | "std 17,136(%2)\n" | ||
74 | "std 18,144(%2)\n" | ||
75 | "std 19,152(%2)\n" | ||
76 | "std 20,160(%2)\n" | ||
77 | "std 21,168(%2)\n" | ||
78 | "std 22,176(%2)\n" | ||
79 | "std 23,184(%2)\n" | ||
80 | "std 24,192(%2)\n" | ||
81 | "std 25,200(%2)\n" | ||
82 | "std 26,208(%2)\n" | ||
83 | "std 27,216(%2)\n" | ||
84 | "std 28,224(%2)\n" | ||
85 | "std 29,232(%2)\n" | ||
86 | "std 30,240(%2)\n" | ||
87 | "std 31,248(%2)\n" | ||
88 | "mfmsr %0\n" | ||
89 | "std %0, 264(%2)\n" | ||
90 | "mfctr %0\n" | ||
91 | "std %0, 280(%2)\n" | ||
92 | "mflr %0\n" | ||
93 | "std %0, 288(%2)\n" | ||
94 | "bl 1f\n" | ||
95 | "1: mflr %1\n" | ||
96 | "std %1, 256(%2)\n" | ||
97 | "mtlr %0\n" | ||
98 | "mfxer %0\n" | ||
99 | "std %0, 296(%2)\n" | ||
100 | : "=&r" (tmp1), "=&r" (tmp2) | ||
101 | : "b" (newregs) | ||
102 | : "memory"); | ||
103 | } | ||
104 | #else | ||
105 | else | 51 | else |
106 | ppc_save_regs(newregs); | 52 | ppc_save_regs(newregs); |
107 | #endif /* __powerpc64__ */ | ||
108 | } | 53 | } |
109 | 54 | ||
110 | extern void kexec_smp_wait(void); /* get and clear naca physid, wait for | 55 | extern void kexec_smp_wait(void); /* get and clear naca physid, wait for |
diff --git a/arch/powerpc/include/asm/ps3.h b/arch/powerpc/include/asm/ps3.h index cff30c0ef1ff..eead5c67197a 100644 --- a/arch/powerpc/include/asm/ps3.h +++ b/arch/powerpc/include/asm/ps3.h | |||
@@ -320,6 +320,7 @@ enum ps3_match_id { | |||
320 | 320 | ||
321 | enum ps3_match_sub_id { | 321 | enum ps3_match_sub_id { |
322 | PS3_MATCH_SUB_ID_GPU_FB = 1, | 322 | PS3_MATCH_SUB_ID_GPU_FB = 1, |
323 | PS3_MATCH_SUB_ID_GPU_RAMDISK = 2, | ||
323 | }; | 324 | }; |
324 | 325 | ||
325 | #define PS3_MODULE_ALIAS_EHCI "ps3:1:0" | 326 | #define PS3_MODULE_ALIAS_EHCI "ps3:1:0" |
@@ -332,6 +333,7 @@ enum ps3_match_sub_id { | |||
332 | #define PS3_MODULE_ALIAS_STOR_FLASH "ps3:8:0" | 333 | #define PS3_MODULE_ALIAS_STOR_FLASH "ps3:8:0" |
333 | #define PS3_MODULE_ALIAS_SOUND "ps3:9:0" | 334 | #define PS3_MODULE_ALIAS_SOUND "ps3:9:0" |
334 | #define PS3_MODULE_ALIAS_GPU_FB "ps3:10:1" | 335 | #define PS3_MODULE_ALIAS_GPU_FB "ps3:10:1" |
336 | #define PS3_MODULE_ALIAS_GPU_RAMDISK "ps3:10:2" | ||
335 | #define PS3_MODULE_ALIAS_LPM "ps3:11:0" | 337 | #define PS3_MODULE_ALIAS_LPM "ps3:11:0" |
336 | 338 | ||
337 | enum ps3_system_bus_device_type { | 339 | enum ps3_system_bus_device_type { |
diff --git a/arch/powerpc/include/asm/qe.h b/arch/powerpc/include/asm/qe.h index edee15d269ea..a0a15311d0d8 100644 --- a/arch/powerpc/include/asm/qe.h +++ b/arch/powerpc/include/asm/qe.h | |||
@@ -17,6 +17,8 @@ | |||
17 | #ifdef __KERNEL__ | 17 | #ifdef __KERNEL__ |
18 | 18 | ||
19 | #include <linux/spinlock.h> | 19 | #include <linux/spinlock.h> |
20 | #include <linux/errno.h> | ||
21 | #include <linux/err.h> | ||
20 | #include <asm/cpm.h> | 22 | #include <asm/cpm.h> |
21 | #include <asm/immap_qe.h> | 23 | #include <asm/immap_qe.h> |
22 | 24 | ||
@@ -84,7 +86,11 @@ static inline bool qe_clock_is_brg(enum qe_clock clk) | |||
84 | extern spinlock_t cmxgcr_lock; | 86 | extern spinlock_t cmxgcr_lock; |
85 | 87 | ||
86 | /* Export QE common operations */ | 88 | /* Export QE common operations */ |
89 | #ifdef CONFIG_QUICC_ENGINE | ||
87 | extern void __init qe_reset(void); | 90 | extern void __init qe_reset(void); |
91 | #else | ||
92 | static inline void qe_reset(void) {} | ||
93 | #endif | ||
88 | 94 | ||
89 | /* QE PIO */ | 95 | /* QE PIO */ |
90 | #define QE_PIO_PINS 32 | 96 | #define QE_PIO_PINS 32 |
@@ -101,16 +107,43 @@ struct qe_pio_regs { | |||
101 | #endif | 107 | #endif |
102 | }; | 108 | }; |
103 | 109 | ||
104 | extern int par_io_init(struct device_node *np); | ||
105 | extern int par_io_of_config(struct device_node *np); | ||
106 | #define QE_PIO_DIR_IN 2 | 110 | #define QE_PIO_DIR_IN 2 |
107 | #define QE_PIO_DIR_OUT 1 | 111 | #define QE_PIO_DIR_OUT 1 |
108 | extern void __par_io_config_pin(struct qe_pio_regs __iomem *par_io, u8 pin, | 112 | extern void __par_io_config_pin(struct qe_pio_regs __iomem *par_io, u8 pin, |
109 | int dir, int open_drain, int assignment, | 113 | int dir, int open_drain, int assignment, |
110 | int has_irq); | 114 | int has_irq); |
115 | #ifdef CONFIG_QUICC_ENGINE | ||
116 | extern int par_io_init(struct device_node *np); | ||
117 | extern int par_io_of_config(struct device_node *np); | ||
111 | extern int par_io_config_pin(u8 port, u8 pin, int dir, int open_drain, | 118 | extern int par_io_config_pin(u8 port, u8 pin, int dir, int open_drain, |
112 | int assignment, int has_irq); | 119 | int assignment, int has_irq); |
113 | extern int par_io_data_set(u8 port, u8 pin, u8 val); | 120 | extern int par_io_data_set(u8 port, u8 pin, u8 val); |
121 | #else | ||
122 | static inline int par_io_init(struct device_node *np) { return -ENOSYS; } | ||
123 | static inline int par_io_of_config(struct device_node *np) { return -ENOSYS; } | ||
124 | static inline int par_io_config_pin(u8 port, u8 pin, int dir, int open_drain, | ||
125 | int assignment, int has_irq) { return -ENOSYS; } | ||
126 | static inline int par_io_data_set(u8 port, u8 pin, u8 val) { return -ENOSYS; } | ||
127 | #endif /* CONFIG_QUICC_ENGINE */ | ||
128 | |||
129 | /* | ||
130 | * Pin multiplexing functions. | ||
131 | */ | ||
132 | struct qe_pin; | ||
133 | #ifdef CONFIG_QE_GPIO | ||
134 | extern struct qe_pin *qe_pin_request(struct device_node *np, int index); | ||
135 | extern void qe_pin_free(struct qe_pin *qe_pin); | ||
136 | extern void qe_pin_set_gpio(struct qe_pin *qe_pin); | ||
137 | extern void qe_pin_set_dedicated(struct qe_pin *pin); | ||
138 | #else | ||
139 | static inline struct qe_pin *qe_pin_request(struct device_node *np, int index) | ||
140 | { | ||
141 | return ERR_PTR(-ENOSYS); | ||
142 | } | ||
143 | static inline void qe_pin_free(struct qe_pin *qe_pin) {} | ||
144 | static inline void qe_pin_set_gpio(struct qe_pin *qe_pin) {} | ||
145 | static inline void qe_pin_set_dedicated(struct qe_pin *pin) {} | ||
146 | #endif /* CONFIG_QE_GPIO */ | ||
114 | 147 | ||
115 | /* QE internal API */ | 148 | /* QE internal API */ |
116 | int qe_issue_cmd(u32 cmd, u32 device, u8 mcn_protocol, u32 cmd_input); | 149 | int qe_issue_cmd(u32 cmd, u32 device, u8 mcn_protocol, u32 cmd_input); |
diff --git a/arch/powerpc/include/asm/qe_ic.h b/arch/powerpc/include/asm/qe_ic.h index 56a7745ca343..cf519663a791 100644 --- a/arch/powerpc/include/asm/qe_ic.h +++ b/arch/powerpc/include/asm/qe_ic.h | |||
@@ -17,6 +17,9 @@ | |||
17 | 17 | ||
18 | #include <linux/irq.h> | 18 | #include <linux/irq.h> |
19 | 19 | ||
20 | struct device_node; | ||
21 | struct qe_ic; | ||
22 | |||
20 | #define NUM_OF_QE_IC_GROUPS 6 | 23 | #define NUM_OF_QE_IC_GROUPS 6 |
21 | 24 | ||
22 | /* Flags when we init the QE IC */ | 25 | /* Flags when we init the QE IC */ |
@@ -54,17 +57,27 @@ enum qe_ic_grp_id { | |||
54 | QE_IC_GRP_RISCB /* QE interrupt controller RISC group B */ | 57 | QE_IC_GRP_RISCB /* QE interrupt controller RISC group B */ |
55 | }; | 58 | }; |
56 | 59 | ||
60 | #ifdef CONFIG_QUICC_ENGINE | ||
57 | void qe_ic_init(struct device_node *node, unsigned int flags, | 61 | void qe_ic_init(struct device_node *node, unsigned int flags, |
58 | void (*low_handler)(unsigned int irq, struct irq_desc *desc), | 62 | void (*low_handler)(unsigned int irq, struct irq_desc *desc), |
59 | void (*high_handler)(unsigned int irq, struct irq_desc *desc)); | 63 | void (*high_handler)(unsigned int irq, struct irq_desc *desc)); |
64 | unsigned int qe_ic_get_low_irq(struct qe_ic *qe_ic); | ||
65 | unsigned int qe_ic_get_high_irq(struct qe_ic *qe_ic); | ||
66 | #else | ||
67 | static inline void qe_ic_init(struct device_node *node, unsigned int flags, | ||
68 | void (*low_handler)(unsigned int irq, struct irq_desc *desc), | ||
69 | void (*high_handler)(unsigned int irq, struct irq_desc *desc)) | ||
70 | {} | ||
71 | static inline unsigned int qe_ic_get_low_irq(struct qe_ic *qe_ic) | ||
72 | { return 0; } | ||
73 | static inline unsigned int qe_ic_get_high_irq(struct qe_ic *qe_ic) | ||
74 | { return 0; } | ||
75 | #endif /* CONFIG_QUICC_ENGINE */ | ||
76 | |||
60 | void qe_ic_set_highest_priority(unsigned int virq, int high); | 77 | void qe_ic_set_highest_priority(unsigned int virq, int high); |
61 | int qe_ic_set_priority(unsigned int virq, unsigned int priority); | 78 | int qe_ic_set_priority(unsigned int virq, unsigned int priority); |
62 | int qe_ic_set_high_priority(unsigned int virq, unsigned int priority, int high); | 79 | int qe_ic_set_high_priority(unsigned int virq, unsigned int priority, int high); |
63 | 80 | ||
64 | struct qe_ic; | ||
65 | unsigned int qe_ic_get_low_irq(struct qe_ic *qe_ic); | ||
66 | unsigned int qe_ic_get_high_irq(struct qe_ic *qe_ic); | ||
67 | |||
68 | static inline void qe_ic_cascade_low_ipic(unsigned int irq, | 81 | static inline void qe_ic_cascade_low_ipic(unsigned int irq, |
69 | struct irq_desc *desc) | 82 | struct irq_desc *desc) |
70 | { | 83 | { |
diff --git a/arch/powerpc/include/asm/spu.h b/arch/powerpc/include/asm/spu.h index 8b2eb044270a..0ab8d869e3d6 100644 --- a/arch/powerpc/include/asm/spu.h +++ b/arch/powerpc/include/asm/spu.h | |||
@@ -128,7 +128,7 @@ struct spu { | |||
128 | int number; | 128 | int number; |
129 | unsigned int irqs[3]; | 129 | unsigned int irqs[3]; |
130 | u32 node; | 130 | u32 node; |
131 | u64 flags; | 131 | unsigned long flags; |
132 | u64 class_0_pending; | 132 | u64 class_0_pending; |
133 | u64 class_0_dar; | 133 | u64 class_0_dar; |
134 | u64 class_1_dar; | 134 | u64 class_1_dar; |
diff --git a/arch/powerpc/kernel/Makefile b/arch/powerpc/kernel/Makefile index 1308a86e9070..8d1a419df35d 100644 --- a/arch/powerpc/kernel/Makefile +++ b/arch/powerpc/kernel/Makefile | |||
@@ -29,7 +29,7 @@ endif | |||
29 | obj-y := cputable.o ptrace.o syscalls.o \ | 29 | obj-y := cputable.o ptrace.o syscalls.o \ |
30 | irq.o align.o signal_32.o pmc.o vdso.o \ | 30 | irq.o align.o signal_32.o pmc.o vdso.o \ |
31 | init_task.o process.o systbl.o idle.o \ | 31 | init_task.o process.o systbl.o idle.o \ |
32 | signal.o sysfs.o | 32 | signal.o sysfs.o cacheinfo.o |
33 | obj-y += vdso32/ | 33 | obj-y += vdso32/ |
34 | obj-$(CONFIG_PPC64) += setup_64.o sys_ppc32.o \ | 34 | obj-$(CONFIG_PPC64) += setup_64.o sys_ppc32.o \ |
35 | signal_64.o ptrace32.o \ | 35 | signal_64.o ptrace32.o \ |
diff --git a/arch/powerpc/kernel/cacheinfo.c b/arch/powerpc/kernel/cacheinfo.c new file mode 100644 index 000000000000..b33f0417a4bf --- /dev/null +++ b/arch/powerpc/kernel/cacheinfo.c | |||
@@ -0,0 +1,837 @@ | |||
1 | /* | ||
2 | * Processor cache information made available to userspace via sysfs; | ||
3 | * intended to be compatible with x86 intel_cacheinfo implementation. | ||
4 | * | ||
5 | * Copyright 2008 IBM Corporation | ||
6 | * Author: Nathan Lynch | ||
7 | * | ||
8 | * This program is free software; you can redistribute it and/or | ||
9 | * modify it under the terms of the GNU General Public License version | ||
10 | * 2 as published by the Free Software Foundation. | ||
11 | */ | ||
12 | |||
13 | #include <linux/cpu.h> | ||
14 | #include <linux/cpumask.h> | ||
15 | #include <linux/init.h> | ||
16 | #include <linux/kernel.h> | ||
17 | #include <linux/kobject.h> | ||
18 | #include <linux/list.h> | ||
19 | #include <linux/notifier.h> | ||
20 | #include <linux/of.h> | ||
21 | #include <linux/percpu.h> | ||
22 | #include <asm/prom.h> | ||
23 | |||
24 | #include "cacheinfo.h" | ||
25 | |||
26 | /* per-cpu object for tracking: | ||
27 | * - a "cache" kobject for the top-level directory | ||
28 | * - a list of "index" objects representing the cpu's local cache hierarchy | ||
29 | */ | ||
30 | struct cache_dir { | ||
31 | struct kobject *kobj; /* bare (not embedded) kobject for cache | ||
32 | * directory */ | ||
33 | struct cache_index_dir *index; /* list of index objects */ | ||
34 | }; | ||
35 | |||
36 | /* "index" object: each cpu's cache directory has an index | ||
37 | * subdirectory corresponding to a cache object associated with the | ||
38 | * cpu. This object's lifetime is managed via the embedded kobject. | ||
39 | */ | ||
40 | struct cache_index_dir { | ||
41 | struct kobject kobj; | ||
42 | struct cache_index_dir *next; /* next index in parent directory */ | ||
43 | struct cache *cache; | ||
44 | }; | ||
45 | |||
46 | /* Template for determining which OF properties to query for a given | ||
47 | * cache type */ | ||
48 | struct cache_type_info { | ||
49 | const char *name; | ||
50 | const char *size_prop; | ||
51 | |||
52 | /* Allow for both [di]-cache-line-size and | ||
53 | * [di]-cache-block-size properties. According to the PowerPC | ||
54 | * Processor binding, -line-size should be provided if it | ||
55 | * differs from the cache block size (that which is operated | ||
56 | * on by cache instructions), so we look for -line-size first. | ||
57 | * See cache_get_line_size(). */ | ||
58 | |||
59 | const char *line_size_props[2]; | ||
60 | const char *nr_sets_prop; | ||
61 | }; | ||
62 | |||
63 | /* These are used to index the cache_type_info array. */ | ||
64 | #define CACHE_TYPE_UNIFIED 0 | ||
65 | #define CACHE_TYPE_INSTRUCTION 1 | ||
66 | #define CACHE_TYPE_DATA 2 | ||
67 | |||
68 | static const struct cache_type_info cache_type_info[] = { | ||
69 | { | ||
70 | /* PowerPC Processor binding says the [di]-cache-* | ||
71 | * must be equal on unified caches, so just use | ||
72 | * d-cache properties. */ | ||
73 | .name = "Unified", | ||
74 | .size_prop = "d-cache-size", | ||
75 | .line_size_props = { "d-cache-line-size", | ||
76 | "d-cache-block-size", }, | ||
77 | .nr_sets_prop = "d-cache-sets", | ||
78 | }, | ||
79 | { | ||
80 | .name = "Instruction", | ||
81 | .size_prop = "i-cache-size", | ||
82 | .line_size_props = { "i-cache-line-size", | ||
83 | "i-cache-block-size", }, | ||
84 | .nr_sets_prop = "i-cache-sets", | ||
85 | }, | ||
86 | { | ||
87 | .name = "Data", | ||
88 | .size_prop = "d-cache-size", | ||
89 | .line_size_props = { "d-cache-line-size", | ||
90 | "d-cache-block-size", }, | ||
91 | .nr_sets_prop = "d-cache-sets", | ||
92 | }, | ||
93 | }; | ||
94 | |||
95 | /* Cache object: each instance of this corresponds to a distinct cache | ||
96 | * in the system. There are separate objects for Harvard caches: one | ||
97 | * each for instruction and data, and each refers to the same OF node. | ||
98 | * The refcount of the OF node is elevated for the lifetime of the | ||
99 | * cache object. A cache object is released when its shared_cpu_map | ||
100 | * is cleared (see cache_cpu_clear). | ||
101 | * | ||
102 | * A cache object is on two lists: an unsorted global list | ||
103 | * (cache_list) of cache objects; and a singly-linked list | ||
104 | * representing the local cache hierarchy, which is ordered by level | ||
105 | * (e.g. L1d -> L1i -> L2 -> L3). | ||
106 | */ | ||
107 | struct cache { | ||
108 | struct device_node *ofnode; /* OF node for this cache, may be cpu */ | ||
109 | struct cpumask shared_cpu_map; /* online CPUs using this cache */ | ||
110 | int type; /* split cache disambiguation */ | ||
111 | int level; /* level not explicit in device tree */ | ||
112 | struct list_head list; /* global list of cache objects */ | ||
113 | struct cache *next_local; /* next cache of >= level */ | ||
114 | }; | ||
115 | |||
116 | static DEFINE_PER_CPU(struct cache_dir *, cache_dir); | ||
117 | |||
118 | /* traversal/modification of this list occurs only at cpu hotplug time; | ||
119 | * access is serialized by cpu hotplug locking | ||
120 | */ | ||
121 | static LIST_HEAD(cache_list); | ||
122 | |||
123 | static struct cache_index_dir *kobj_to_cache_index_dir(struct kobject *k) | ||
124 | { | ||
125 | return container_of(k, struct cache_index_dir, kobj); | ||
126 | } | ||
127 | |||
128 | static const char *cache_type_string(const struct cache *cache) | ||
129 | { | ||
130 | return cache_type_info[cache->type].name; | ||
131 | } | ||
132 | |||
133 | static void __cpuinit cache_init(struct cache *cache, int type, int level, struct device_node *ofnode) | ||
134 | { | ||
135 | cache->type = type; | ||
136 | cache->level = level; | ||
137 | cache->ofnode = of_node_get(ofnode); | ||
138 | INIT_LIST_HEAD(&cache->list); | ||
139 | list_add(&cache->list, &cache_list); | ||
140 | } | ||
141 | |||
142 | static struct cache *__cpuinit new_cache(int type, int level, struct device_node *ofnode) | ||
143 | { | ||
144 | struct cache *cache; | ||
145 | |||
146 | cache = kzalloc(sizeof(*cache), GFP_KERNEL); | ||
147 | if (cache) | ||
148 | cache_init(cache, type, level, ofnode); | ||
149 | |||
150 | return cache; | ||
151 | } | ||
152 | |||
153 | static void release_cache_debugcheck(struct cache *cache) | ||
154 | { | ||
155 | struct cache *iter; | ||
156 | |||
157 | list_for_each_entry(iter, &cache_list, list) | ||
158 | WARN_ONCE(iter->next_local == cache, | ||
159 | "cache for %s(%s) refers to cache for %s(%s)\n", | ||
160 | iter->ofnode->full_name, | ||
161 | cache_type_string(iter), | ||
162 | cache->ofnode->full_name, | ||
163 | cache_type_string(cache)); | ||
164 | } | ||
165 | |||
166 | static void release_cache(struct cache *cache) | ||
167 | { | ||
168 | if (!cache) | ||
169 | return; | ||
170 | |||
171 | pr_debug("freeing L%d %s cache for %s\n", cache->level, | ||
172 | cache_type_string(cache), cache->ofnode->full_name); | ||
173 | |||
174 | release_cache_debugcheck(cache); | ||
175 | list_del(&cache->list); | ||
176 | of_node_put(cache->ofnode); | ||
177 | kfree(cache); | ||
178 | } | ||
179 | |||
180 | static void cache_cpu_set(struct cache *cache, int cpu) | ||
181 | { | ||
182 | struct cache *next = cache; | ||
183 | |||
184 | while (next) { | ||
185 | WARN_ONCE(cpumask_test_cpu(cpu, &next->shared_cpu_map), | ||
186 | "CPU %i already accounted in %s(%s)\n", | ||
187 | cpu, next->ofnode->full_name, | ||
188 | cache_type_string(next)); | ||
189 | cpumask_set_cpu(cpu, &next->shared_cpu_map); | ||
190 | next = next->next_local; | ||
191 | } | ||
192 | } | ||
193 | |||
194 | static int cache_size(const struct cache *cache, unsigned int *ret) | ||
195 | { | ||
196 | const char *propname; | ||
197 | const u32 *cache_size; | ||
198 | |||
199 | propname = cache_type_info[cache->type].size_prop; | ||
200 | |||
201 | cache_size = of_get_property(cache->ofnode, propname, NULL); | ||
202 | if (!cache_size) | ||
203 | return -ENODEV; | ||
204 | |||
205 | *ret = *cache_size; | ||
206 | return 0; | ||
207 | } | ||
208 | |||
209 | static int cache_size_kb(const struct cache *cache, unsigned int *ret) | ||
210 | { | ||
211 | unsigned int size; | ||
212 | |||
213 | if (cache_size(cache, &size)) | ||
214 | return -ENODEV; | ||
215 | |||
216 | *ret = size / 1024; | ||
217 | return 0; | ||
218 | } | ||
219 | |||
220 | /* not cache_line_size() because that's a macro in include/linux/cache.h */ | ||
221 | static int cache_get_line_size(const struct cache *cache, unsigned int *ret) | ||
222 | { | ||
223 | const u32 *line_size; | ||
224 | int i, lim; | ||
225 | |||
226 | lim = ARRAY_SIZE(cache_type_info[cache->type].line_size_props); | ||
227 | |||
228 | for (i = 0; i < lim; i++) { | ||
229 | const char *propname; | ||
230 | |||
231 | propname = cache_type_info[cache->type].line_size_props[i]; | ||
232 | line_size = of_get_property(cache->ofnode, propname, NULL); | ||
233 | if (line_size) | ||
234 | break; | ||
235 | } | ||
236 | |||
237 | if (!line_size) | ||
238 | return -ENODEV; | ||
239 | |||
240 | *ret = *line_size; | ||
241 | return 0; | ||
242 | } | ||
243 | |||
244 | static int cache_nr_sets(const struct cache *cache, unsigned int *ret) | ||
245 | { | ||
246 | const char *propname; | ||
247 | const u32 *nr_sets; | ||
248 | |||
249 | propname = cache_type_info[cache->type].nr_sets_prop; | ||
250 | |||
251 | nr_sets = of_get_property(cache->ofnode, propname, NULL); | ||
252 | if (!nr_sets) | ||
253 | return -ENODEV; | ||
254 | |||
255 | *ret = *nr_sets; | ||
256 | return 0; | ||
257 | } | ||
258 | |||
259 | static int cache_associativity(const struct cache *cache, unsigned int *ret) | ||
260 | { | ||
261 | unsigned int line_size; | ||
262 | unsigned int nr_sets; | ||
263 | unsigned int size; | ||
264 | |||
265 | if (cache_nr_sets(cache, &nr_sets)) | ||
266 | goto err; | ||
267 | |||
268 | /* If the cache is fully associative, there is no need to | ||
269 | * check the other properties. | ||
270 | */ | ||
271 | if (nr_sets == 1) { | ||
272 | *ret = 0; | ||
273 | return 0; | ||
274 | } | ||
275 | |||
276 | if (cache_get_line_size(cache, &line_size)) | ||
277 | goto err; | ||
278 | if (cache_size(cache, &size)) | ||
279 | goto err; | ||
280 | |||
281 | if (!(nr_sets > 0 && size > 0 && line_size > 0)) | ||
282 | goto err; | ||
283 | |||
284 | *ret = (size / nr_sets) / line_size; | ||
285 | return 0; | ||
286 | err: | ||
287 | return -ENODEV; | ||
288 | } | ||
289 | |||
290 | /* helper for dealing with split caches */ | ||
291 | static struct cache *cache_find_first_sibling(struct cache *cache) | ||
292 | { | ||
293 | struct cache *iter; | ||
294 | |||
295 | if (cache->type == CACHE_TYPE_UNIFIED) | ||
296 | return cache; | ||
297 | |||
298 | list_for_each_entry(iter, &cache_list, list) | ||
299 | if (iter->ofnode == cache->ofnode && iter->next_local == cache) | ||
300 | return iter; | ||
301 | |||
302 | return cache; | ||
303 | } | ||
304 | |||
305 | /* return the first cache on a local list matching node */ | ||
306 | static struct cache *cache_lookup_by_node(const struct device_node *node) | ||
307 | { | ||
308 | struct cache *cache = NULL; | ||
309 | struct cache *iter; | ||
310 | |||
311 | list_for_each_entry(iter, &cache_list, list) { | ||
312 | if (iter->ofnode != node) | ||
313 | continue; | ||
314 | cache = cache_find_first_sibling(iter); | ||
315 | break; | ||
316 | } | ||
317 | |||
318 | return cache; | ||
319 | } | ||
320 | |||
321 | static bool cache_node_is_unified(const struct device_node *np) | ||
322 | { | ||
323 | return of_get_property(np, "cache-unified", NULL); | ||
324 | } | ||
325 | |||
326 | static struct cache *__cpuinit cache_do_one_devnode_unified(struct device_node *node, int level) | ||
327 | { | ||
328 | struct cache *cache; | ||
329 | |||
330 | pr_debug("creating L%d ucache for %s\n", level, node->full_name); | ||
331 | |||
332 | cache = new_cache(CACHE_TYPE_UNIFIED, level, node); | ||
333 | |||
334 | return cache; | ||
335 | } | ||
336 | |||
337 | static struct cache *__cpuinit cache_do_one_devnode_split(struct device_node *node, int level) | ||
338 | { | ||
339 | struct cache *dcache, *icache; | ||
340 | |||
341 | pr_debug("creating L%d dcache and icache for %s\n", level, | ||
342 | node->full_name); | ||
343 | |||
344 | dcache = new_cache(CACHE_TYPE_DATA, level, node); | ||
345 | icache = new_cache(CACHE_TYPE_INSTRUCTION, level, node); | ||
346 | |||
347 | if (!dcache || !icache) | ||
348 | goto err; | ||
349 | |||
350 | dcache->next_local = icache; | ||
351 | |||
352 | return dcache; | ||
353 | err: | ||
354 | release_cache(dcache); | ||
355 | release_cache(icache); | ||
356 | return NULL; | ||
357 | } | ||
358 | |||
359 | static struct cache *__cpuinit cache_do_one_devnode(struct device_node *node, int level) | ||
360 | { | ||
361 | struct cache *cache; | ||
362 | |||
363 | if (cache_node_is_unified(node)) | ||
364 | cache = cache_do_one_devnode_unified(node, level); | ||
365 | else | ||
366 | cache = cache_do_one_devnode_split(node, level); | ||
367 | |||
368 | return cache; | ||
369 | } | ||
370 | |||
371 | static struct cache *__cpuinit cache_lookup_or_instantiate(struct device_node *node, int level) | ||
372 | { | ||
373 | struct cache *cache; | ||
374 | |||
375 | cache = cache_lookup_by_node(node); | ||
376 | |||
377 | WARN_ONCE(cache && cache->level != level, | ||
378 | "cache level mismatch on lookup (got %d, expected %d)\n", | ||
379 | cache->level, level); | ||
380 | |||
381 | if (!cache) | ||
382 | cache = cache_do_one_devnode(node, level); | ||
383 | |||
384 | return cache; | ||
385 | } | ||
386 | |||
387 | static void __cpuinit link_cache_lists(struct cache *smaller, struct cache *bigger) | ||
388 | { | ||
389 | while (smaller->next_local) { | ||
390 | if (smaller->next_local == bigger) | ||
391 | return; /* already linked */ | ||
392 | smaller = smaller->next_local; | ||
393 | } | ||
394 | |||
395 | smaller->next_local = bigger; | ||
396 | } | ||
397 | |||
398 | static void __cpuinit do_subsidiary_caches_debugcheck(struct cache *cache) | ||
399 | { | ||
400 | WARN_ON_ONCE(cache->level != 1); | ||
401 | WARN_ON_ONCE(strcmp(cache->ofnode->type, "cpu")); | ||
402 | } | ||
403 | |||
404 | static void __cpuinit do_subsidiary_caches(struct cache *cache) | ||
405 | { | ||
406 | struct device_node *subcache_node; | ||
407 | int level = cache->level; | ||
408 | |||
409 | do_subsidiary_caches_debugcheck(cache); | ||
410 | |||
411 | while ((subcache_node = of_find_next_cache_node(cache->ofnode))) { | ||
412 | struct cache *subcache; | ||
413 | |||
414 | level++; | ||
415 | subcache = cache_lookup_or_instantiate(subcache_node, level); | ||
416 | of_node_put(subcache_node); | ||
417 | if (!subcache) | ||
418 | break; | ||
419 | |||
420 | link_cache_lists(cache, subcache); | ||
421 | cache = subcache; | ||
422 | } | ||
423 | } | ||
424 | |||
425 | static struct cache *__cpuinit cache_chain_instantiate(unsigned int cpu_id) | ||
426 | { | ||
427 | struct device_node *cpu_node; | ||
428 | struct cache *cpu_cache = NULL; | ||
429 | |||
430 | pr_debug("creating cache object(s) for CPU %i\n", cpu_id); | ||
431 | |||
432 | cpu_node = of_get_cpu_node(cpu_id, NULL); | ||
433 | WARN_ONCE(!cpu_node, "no OF node found for CPU %i\n", cpu_id); | ||
434 | if (!cpu_node) | ||
435 | goto out; | ||
436 | |||
437 | cpu_cache = cache_lookup_or_instantiate(cpu_node, 1); | ||
438 | if (!cpu_cache) | ||
439 | goto out; | ||
440 | |||
441 | do_subsidiary_caches(cpu_cache); | ||
442 | |||
443 | cache_cpu_set(cpu_cache, cpu_id); | ||
444 | out: | ||
445 | of_node_put(cpu_node); | ||
446 | |||
447 | return cpu_cache; | ||
448 | } | ||
449 | |||
450 | static struct cache_dir *__cpuinit cacheinfo_create_cache_dir(unsigned int cpu_id) | ||
451 | { | ||
452 | struct cache_dir *cache_dir; | ||
453 | struct sys_device *sysdev; | ||
454 | struct kobject *kobj = NULL; | ||
455 | |||
456 | sysdev = get_cpu_sysdev(cpu_id); | ||
457 | WARN_ONCE(!sysdev, "no sysdev for CPU %i\n", cpu_id); | ||
458 | if (!sysdev) | ||
459 | goto err; | ||
460 | |||
461 | kobj = kobject_create_and_add("cache", &sysdev->kobj); | ||
462 | if (!kobj) | ||
463 | goto err; | ||
464 | |||
465 | cache_dir = kzalloc(sizeof(*cache_dir), GFP_KERNEL); | ||
466 | if (!cache_dir) | ||
467 | goto err; | ||
468 | |||
469 | cache_dir->kobj = kobj; | ||
470 | |||
471 | WARN_ON_ONCE(per_cpu(cache_dir, cpu_id) != NULL); | ||
472 | |||
473 | per_cpu(cache_dir, cpu_id) = cache_dir; | ||
474 | |||
475 | return cache_dir; | ||
476 | err: | ||
477 | kobject_put(kobj); | ||
478 | return NULL; | ||
479 | } | ||
480 | |||
481 | static void cache_index_release(struct kobject *kobj) | ||
482 | { | ||
483 | struct cache_index_dir *index; | ||
484 | |||
485 | index = kobj_to_cache_index_dir(kobj); | ||
486 | |||
487 | pr_debug("freeing index directory for L%d %s cache\n", | ||
488 | index->cache->level, cache_type_string(index->cache)); | ||
489 | |||
490 | kfree(index); | ||
491 | } | ||
492 | |||
493 | static ssize_t cache_index_show(struct kobject *k, struct attribute *attr, char *buf) | ||
494 | { | ||
495 | struct kobj_attribute *kobj_attr; | ||
496 | |||
497 | kobj_attr = container_of(attr, struct kobj_attribute, attr); | ||
498 | |||
499 | return kobj_attr->show(k, kobj_attr, buf); | ||
500 | } | ||
501 | |||
502 | static struct cache *index_kobj_to_cache(struct kobject *k) | ||
503 | { | ||
504 | struct cache_index_dir *index; | ||
505 | |||
506 | index = kobj_to_cache_index_dir(k); | ||
507 | |||
508 | return index->cache; | ||
509 | } | ||
510 | |||
511 | static ssize_t size_show(struct kobject *k, struct kobj_attribute *attr, char *buf) | ||
512 | { | ||
513 | unsigned int size_kb; | ||
514 | struct cache *cache; | ||
515 | |||
516 | cache = index_kobj_to_cache(k); | ||
517 | |||
518 | if (cache_size_kb(cache, &size_kb)) | ||
519 | return -ENODEV; | ||
520 | |||
521 | return sprintf(buf, "%uK\n", size_kb); | ||
522 | } | ||
523 | |||
524 | static struct kobj_attribute cache_size_attr = | ||
525 | __ATTR(size, 0444, size_show, NULL); | ||
526 | |||
527 | |||
528 | static ssize_t line_size_show(struct kobject *k, struct kobj_attribute *attr, char *buf) | ||
529 | { | ||
530 | unsigned int line_size; | ||
531 | struct cache *cache; | ||
532 | |||
533 | cache = index_kobj_to_cache(k); | ||
534 | |||
535 | if (cache_get_line_size(cache, &line_size)) | ||
536 | return -ENODEV; | ||
537 | |||
538 | return sprintf(buf, "%u\n", line_size); | ||
539 | } | ||
540 | |||
541 | static struct kobj_attribute cache_line_size_attr = | ||
542 | __ATTR(coherency_line_size, 0444, line_size_show, NULL); | ||
543 | |||
544 | static ssize_t nr_sets_show(struct kobject *k, struct kobj_attribute *attr, char *buf) | ||
545 | { | ||
546 | unsigned int nr_sets; | ||
547 | struct cache *cache; | ||
548 | |||
549 | cache = index_kobj_to_cache(k); | ||
550 | |||
551 | if (cache_nr_sets(cache, &nr_sets)) | ||
552 | return -ENODEV; | ||
553 | |||
554 | return sprintf(buf, "%u\n", nr_sets); | ||
555 | } | ||
556 | |||
557 | static struct kobj_attribute cache_nr_sets_attr = | ||
558 | __ATTR(number_of_sets, 0444, nr_sets_show, NULL); | ||
559 | |||
560 | static ssize_t associativity_show(struct kobject *k, struct kobj_attribute *attr, char *buf) | ||
561 | { | ||
562 | unsigned int associativity; | ||
563 | struct cache *cache; | ||
564 | |||
565 | cache = index_kobj_to_cache(k); | ||
566 | |||
567 | if (cache_associativity(cache, &associativity)) | ||
568 | return -ENODEV; | ||
569 | |||
570 | return sprintf(buf, "%u\n", associativity); | ||
571 | } | ||
572 | |||
573 | static struct kobj_attribute cache_assoc_attr = | ||
574 | __ATTR(ways_of_associativity, 0444, associativity_show, NULL); | ||
575 | |||
576 | static ssize_t type_show(struct kobject *k, struct kobj_attribute *attr, char *buf) | ||
577 | { | ||
578 | struct cache *cache; | ||
579 | |||
580 | cache = index_kobj_to_cache(k); | ||
581 | |||
582 | return sprintf(buf, "%s\n", cache_type_string(cache)); | ||
583 | } | ||
584 | |||
585 | static struct kobj_attribute cache_type_attr = | ||
586 | __ATTR(type, 0444, type_show, NULL); | ||
587 | |||
588 | static ssize_t level_show(struct kobject *k, struct kobj_attribute *attr, char *buf) | ||
589 | { | ||
590 | struct cache_index_dir *index; | ||
591 | struct cache *cache; | ||
592 | |||
593 | index = kobj_to_cache_index_dir(k); | ||
594 | cache = index->cache; | ||
595 | |||
596 | return sprintf(buf, "%d\n", cache->level); | ||
597 | } | ||
598 | |||
599 | static struct kobj_attribute cache_level_attr = | ||
600 | __ATTR(level, 0444, level_show, NULL); | ||
601 | |||
602 | static ssize_t shared_cpu_map_show(struct kobject *k, struct kobj_attribute *attr, char *buf) | ||
603 | { | ||
604 | struct cache_index_dir *index; | ||
605 | struct cache *cache; | ||
606 | int len; | ||
607 | int n = 0; | ||
608 | |||
609 | index = kobj_to_cache_index_dir(k); | ||
610 | cache = index->cache; | ||
611 | len = PAGE_SIZE - 2; | ||
612 | |||
613 | if (len > 1) { | ||
614 | n = cpumask_scnprintf(buf, len, &cache->shared_cpu_map); | ||
615 | buf[n++] = '\n'; | ||
616 | buf[n] = '\0'; | ||
617 | } | ||
618 | return n; | ||
619 | } | ||
620 | |||
621 | static struct kobj_attribute cache_shared_cpu_map_attr = | ||
622 | __ATTR(shared_cpu_map, 0444, shared_cpu_map_show, NULL); | ||
623 | |||
624 | /* Attributes which should always be created -- the kobject/sysfs core | ||
625 | * does this automatically via kobj_type->default_attrs. This is the | ||
626 | * minimum data required to uniquely identify a cache. | ||
627 | */ | ||
628 | static struct attribute *cache_index_default_attrs[] = { | ||
629 | &cache_type_attr.attr, | ||
630 | &cache_level_attr.attr, | ||
631 | &cache_shared_cpu_map_attr.attr, | ||
632 | NULL, | ||
633 | }; | ||
634 | |||
635 | /* Attributes which should be created if the cache device node has the | ||
636 | * right properties -- see cacheinfo_create_index_opt_attrs | ||
637 | */ | ||
638 | static struct kobj_attribute *cache_index_opt_attrs[] = { | ||
639 | &cache_size_attr, | ||
640 | &cache_line_size_attr, | ||
641 | &cache_nr_sets_attr, | ||
642 | &cache_assoc_attr, | ||
643 | }; | ||
644 | |||
645 | static struct sysfs_ops cache_index_ops = { | ||
646 | .show = cache_index_show, | ||
647 | }; | ||
648 | |||
649 | static struct kobj_type cache_index_type = { | ||
650 | .release = cache_index_release, | ||
651 | .sysfs_ops = &cache_index_ops, | ||
652 | .default_attrs = cache_index_default_attrs, | ||
653 | }; | ||
654 | |||
655 | static void __cpuinit cacheinfo_create_index_opt_attrs(struct cache_index_dir *dir) | ||
656 | { | ||
657 | const char *cache_name; | ||
658 | const char *cache_type; | ||
659 | struct cache *cache; | ||
660 | char *buf; | ||
661 | int i; | ||
662 | |||
663 | buf = kmalloc(PAGE_SIZE, GFP_KERNEL); | ||
664 | if (!buf) | ||
665 | return; | ||
666 | |||
667 | cache = dir->cache; | ||
668 | cache_name = cache->ofnode->full_name; | ||
669 | cache_type = cache_type_string(cache); | ||
670 | |||
671 | /* We don't want to create an attribute that can't provide a | ||
672 | * meaningful value. Check the return value of each optional | ||
673 | * attribute's ->show method before registering the | ||
674 | * attribute. | ||
675 | */ | ||
676 | for (i = 0; i < ARRAY_SIZE(cache_index_opt_attrs); i++) { | ||
677 | struct kobj_attribute *attr; | ||
678 | ssize_t rc; | ||
679 | |||
680 | attr = cache_index_opt_attrs[i]; | ||
681 | |||
682 | rc = attr->show(&dir->kobj, attr, buf); | ||
683 | if (rc <= 0) { | ||
684 | pr_debug("not creating %s attribute for " | ||
685 | "%s(%s) (rc = %zd)\n", | ||
686 | attr->attr.name, cache_name, | ||
687 | cache_type, rc); | ||
688 | continue; | ||
689 | } | ||
690 | if (sysfs_create_file(&dir->kobj, &attr->attr)) | ||
691 | pr_debug("could not create %s attribute for %s(%s)\n", | ||
692 | attr->attr.name, cache_name, cache_type); | ||
693 | } | ||
694 | |||
695 | kfree(buf); | ||
696 | } | ||
697 | |||
698 | static void __cpuinit cacheinfo_create_index_dir(struct cache *cache, int index, struct cache_dir *cache_dir) | ||
699 | { | ||
700 | struct cache_index_dir *index_dir; | ||
701 | int rc; | ||
702 | |||
703 | index_dir = kzalloc(sizeof(*index_dir), GFP_KERNEL); | ||
704 | if (!index_dir) | ||
705 | goto err; | ||
706 | |||
707 | index_dir->cache = cache; | ||
708 | |||
709 | rc = kobject_init_and_add(&index_dir->kobj, &cache_index_type, | ||
710 | cache_dir->kobj, "index%d", index); | ||
711 | if (rc) | ||
712 | goto err; | ||
713 | |||
714 | index_dir->next = cache_dir->index; | ||
715 | cache_dir->index = index_dir; | ||
716 | |||
717 | cacheinfo_create_index_opt_attrs(index_dir); | ||
718 | |||
719 | return; | ||
720 | err: | ||
721 | kfree(index_dir); | ||
722 | } | ||
723 | |||
724 | static void __cpuinit cacheinfo_sysfs_populate(unsigned int cpu_id, struct cache *cache_list) | ||
725 | { | ||
726 | struct cache_dir *cache_dir; | ||
727 | struct cache *cache; | ||
728 | int index = 0; | ||
729 | |||
730 | cache_dir = cacheinfo_create_cache_dir(cpu_id); | ||
731 | if (!cache_dir) | ||
732 | return; | ||
733 | |||
734 | cache = cache_list; | ||
735 | while (cache) { | ||
736 | cacheinfo_create_index_dir(cache, index, cache_dir); | ||
737 | index++; | ||
738 | cache = cache->next_local; | ||
739 | } | ||
740 | } | ||
741 | |||
742 | void __cpuinit cacheinfo_cpu_online(unsigned int cpu_id) | ||
743 | { | ||
744 | struct cache *cache; | ||
745 | |||
746 | cache = cache_chain_instantiate(cpu_id); | ||
747 | if (!cache) | ||
748 | return; | ||
749 | |||
750 | cacheinfo_sysfs_populate(cpu_id, cache); | ||
751 | } | ||
752 | |||
753 | #ifdef CONFIG_HOTPLUG_CPU /* functions needed for cpu offline */ | ||
754 | |||
755 | static struct cache *cache_lookup_by_cpu(unsigned int cpu_id) | ||
756 | { | ||
757 | struct device_node *cpu_node; | ||
758 | struct cache *cache; | ||
759 | |||
760 | cpu_node = of_get_cpu_node(cpu_id, NULL); | ||
761 | WARN_ONCE(!cpu_node, "no OF node found for CPU %i\n", cpu_id); | ||
762 | if (!cpu_node) | ||
763 | return NULL; | ||
764 | |||
765 | cache = cache_lookup_by_node(cpu_node); | ||
766 | of_node_put(cpu_node); | ||
767 | |||
768 | return cache; | ||
769 | } | ||
770 | |||
771 | static void remove_index_dirs(struct cache_dir *cache_dir) | ||
772 | { | ||
773 | struct cache_index_dir *index; | ||
774 | |||
775 | index = cache_dir->index; | ||
776 | |||
777 | while (index) { | ||
778 | struct cache_index_dir *next; | ||
779 | |||
780 | next = index->next; | ||
781 | kobject_put(&index->kobj); | ||
782 | index = next; | ||
783 | } | ||
784 | } | ||
785 | |||
786 | static void remove_cache_dir(struct cache_dir *cache_dir) | ||
787 | { | ||
788 | remove_index_dirs(cache_dir); | ||
789 | |||
790 | kobject_put(cache_dir->kobj); | ||
791 | |||
792 | kfree(cache_dir); | ||
793 | } | ||
794 | |||
795 | static void cache_cpu_clear(struct cache *cache, int cpu) | ||
796 | { | ||
797 | while (cache) { | ||
798 | struct cache *next = cache->next_local; | ||
799 | |||
800 | WARN_ONCE(!cpumask_test_cpu(cpu, &cache->shared_cpu_map), | ||
801 | "CPU %i not accounted in %s(%s)\n", | ||
802 | cpu, cache->ofnode->full_name, | ||
803 | cache_type_string(cache)); | ||
804 | |||
805 | cpumask_clear_cpu(cpu, &cache->shared_cpu_map); | ||
806 | |||
807 | /* Release the cache object if all the cpus using it | ||
808 | * are offline */ | ||
809 | if (cpumask_empty(&cache->shared_cpu_map)) | ||
810 | release_cache(cache); | ||
811 | |||
812 | cache = next; | ||
813 | } | ||
814 | } | ||
815 | |||
816 | void cacheinfo_cpu_offline(unsigned int cpu_id) | ||
817 | { | ||
818 | struct cache_dir *cache_dir; | ||
819 | struct cache *cache; | ||
820 | |||
821 | /* Prevent userspace from seeing inconsistent state - remove | ||
822 | * the sysfs hierarchy first */ | ||
823 | cache_dir = per_cpu(cache_dir, cpu_id); | ||
824 | |||
825 | /* careful, sysfs population may have failed */ | ||
826 | if (cache_dir) | ||
827 | remove_cache_dir(cache_dir); | ||
828 | |||
829 | per_cpu(cache_dir, cpu_id) = NULL; | ||
830 | |||
831 | /* clear the CPU's bit in its cache chain, possibly freeing | ||
832 | * cache objects */ | ||
833 | cache = cache_lookup_by_cpu(cpu_id); | ||
834 | if (cache) | ||
835 | cache_cpu_clear(cache, cpu_id); | ||
836 | } | ||
837 | #endif /* CONFIG_HOTPLUG_CPU */ | ||
diff --git a/arch/powerpc/kernel/cacheinfo.h b/arch/powerpc/kernel/cacheinfo.h new file mode 100644 index 000000000000..a7b74d36acd7 --- /dev/null +++ b/arch/powerpc/kernel/cacheinfo.h | |||
@@ -0,0 +1,8 @@ | |||
1 | #ifndef _PPC_CACHEINFO_H | ||
2 | #define _PPC_CACHEINFO_H | ||
3 | |||
4 | /* These are just hooks for sysfs.c to use. */ | ||
5 | extern void cacheinfo_cpu_online(unsigned int cpu_id); | ||
6 | extern void cacheinfo_cpu_offline(unsigned int cpu_id); | ||
7 | |||
8 | #endif /* _PPC_CACHEINFO_H */ | ||
diff --git a/arch/powerpc/kernel/pci-common.c b/arch/powerpc/kernel/pci-common.c index 2538030954d8..da5a3855a0c4 100644 --- a/arch/powerpc/kernel/pci-common.c +++ b/arch/powerpc/kernel/pci-common.c | |||
@@ -16,7 +16,7 @@ | |||
16 | * 2 of the License, or (at your option) any later version. | 16 | * 2 of the License, or (at your option) any later version. |
17 | */ | 17 | */ |
18 | 18 | ||
19 | #undef DEBUG | 19 | #define DEBUG |
20 | 20 | ||
21 | #include <linux/kernel.h> | 21 | #include <linux/kernel.h> |
22 | #include <linux/pci.h> | 22 | #include <linux/pci.h> |
@@ -1356,6 +1356,63 @@ static void __init pcibios_allocate_resources(int pass) | |||
1356 | } | 1356 | } |
1357 | } | 1357 | } |
1358 | 1358 | ||
1359 | static void __init pcibios_reserve_legacy_regions(struct pci_bus *bus) | ||
1360 | { | ||
1361 | struct pci_controller *hose = pci_bus_to_host(bus); | ||
1362 | resource_size_t offset; | ||
1363 | struct resource *res, *pres; | ||
1364 | int i; | ||
1365 | |||
1366 | pr_debug("Reserving legacy ranges for domain %04x\n", pci_domain_nr(bus)); | ||
1367 | |||
1368 | /* Check for IO */ | ||
1369 | if (!(hose->io_resource.flags & IORESOURCE_IO)) | ||
1370 | goto no_io; | ||
1371 | offset = (unsigned long)hose->io_base_virt - _IO_BASE; | ||
1372 | res = kzalloc(sizeof(struct resource), GFP_KERNEL); | ||
1373 | BUG_ON(res == NULL); | ||
1374 | res->name = "Legacy IO"; | ||
1375 | res->flags = IORESOURCE_IO; | ||
1376 | res->start = offset; | ||
1377 | res->end = (offset + 0xfff) & 0xfffffffful; | ||
1378 | pr_debug("Candidate legacy IO: %pR\n", res); | ||
1379 | if (request_resource(&hose->io_resource, res)) { | ||
1380 | printk(KERN_DEBUG | ||
1381 | "PCI %04x:%02x Cannot reserve Legacy IO %pR\n", | ||
1382 | pci_domain_nr(bus), bus->number, res); | ||
1383 | kfree(res); | ||
1384 | } | ||
1385 | |||
1386 | no_io: | ||
1387 | /* Check for memory */ | ||
1388 | offset = hose->pci_mem_offset; | ||
1389 | pr_debug("hose mem offset: %016llx\n", (unsigned long long)offset); | ||
1390 | for (i = 0; i < 3; i++) { | ||
1391 | pres = &hose->mem_resources[i]; | ||
1392 | if (!(pres->flags & IORESOURCE_MEM)) | ||
1393 | continue; | ||
1394 | pr_debug("hose mem res: %pR\n", pres); | ||
1395 | if ((pres->start - offset) <= 0xa0000 && | ||
1396 | (pres->end - offset) >= 0xbffff) | ||
1397 | break; | ||
1398 | } | ||
1399 | if (i >= 3) | ||
1400 | return; | ||
1401 | res = kzalloc(sizeof(struct resource), GFP_KERNEL); | ||
1402 | BUG_ON(res == NULL); | ||
1403 | res->name = "Legacy VGA memory"; | ||
1404 | res->flags = IORESOURCE_MEM; | ||
1405 | res->start = 0xa0000 + offset; | ||
1406 | res->end = 0xbffff + offset; | ||
1407 | pr_debug("Candidate VGA memory: %pR\n", res); | ||
1408 | if (request_resource(pres, res)) { | ||
1409 | printk(KERN_DEBUG | ||
1410 | "PCI %04x:%02x Cannot reserve VGA memory %pR\n", | ||
1411 | pci_domain_nr(bus), bus->number, res); | ||
1412 | kfree(res); | ||
1413 | } | ||
1414 | } | ||
1415 | |||
1359 | void __init pcibios_resource_survey(void) | 1416 | void __init pcibios_resource_survey(void) |
1360 | { | 1417 | { |
1361 | struct pci_bus *b; | 1418 | struct pci_bus *b; |
@@ -1371,6 +1428,18 @@ void __init pcibios_resource_survey(void) | |||
1371 | pcibios_allocate_resources(1); | 1428 | pcibios_allocate_resources(1); |
1372 | } | 1429 | } |
1373 | 1430 | ||
1431 | /* Before we start assigning unassigned resource, we try to reserve | ||
1432 | * the low IO area and the VGA memory area if they intersect the | ||
1433 | * bus available resources to avoid allocating things on top of them | ||
1434 | */ | ||
1435 | if (!(ppc_pci_flags & PPC_PCI_PROBE_ONLY)) { | ||
1436 | list_for_each_entry(b, &pci_root_buses, node) | ||
1437 | pcibios_reserve_legacy_regions(b); | ||
1438 | } | ||
1439 | |||
1440 | /* Now, if the platform didn't decide to blindly trust the firmware, | ||
1441 | * we proceed to assigning things that were left unassigned | ||
1442 | */ | ||
1374 | if (!(ppc_pci_flags & PPC_PCI_PROBE_ONLY)) { | 1443 | if (!(ppc_pci_flags & PPC_PCI_PROBE_ONLY)) { |
1375 | pr_debug("PCI: Assigning unassigned resouces...\n"); | 1444 | pr_debug("PCI: Assigning unassigned resouces...\n"); |
1376 | pci_assign_unassigned_resources(); | 1445 | pci_assign_unassigned_resources(); |
diff --git a/arch/powerpc/kernel/pci_64.c b/arch/powerpc/kernel/pci_64.c index 39fadc6e1492..586962f65c2a 100644 --- a/arch/powerpc/kernel/pci_64.c +++ b/arch/powerpc/kernel/pci_64.c | |||
@@ -560,9 +560,14 @@ long sys_pciconfig_iobase(long which, unsigned long in_bus, | |||
560 | * G5 machines... So when something asks for bus 0 io base | 560 | * G5 machines... So when something asks for bus 0 io base |
561 | * (bus 0 is HT root), we return the AGP one instead. | 561 | * (bus 0 is HT root), we return the AGP one instead. |
562 | */ | 562 | */ |
563 | if (machine_is_compatible("MacRISC4")) | 563 | if (in_bus == 0 && machine_is_compatible("MacRISC4")) { |
564 | if (in_bus == 0) | 564 | struct device_node *agp; |
565 | |||
566 | agp = of_find_compatible_node(NULL, NULL, "u3-agp"); | ||
567 | if (agp) | ||
565 | in_bus = 0xf0; | 568 | in_bus = 0xf0; |
569 | of_node_put(agp); | ||
570 | } | ||
566 | 571 | ||
567 | /* That syscall isn't quite compatible with PCI domains, but it's | 572 | /* That syscall isn't quite compatible with PCI domains, but it's |
568 | * used on pre-domains setup. We return the first match | 573 | * used on pre-domains setup. We return the first match |
diff --git a/arch/powerpc/kernel/ppc_ksyms.c b/arch/powerpc/kernel/ppc_ksyms.c index dcec1325d340..c8b27bb4dbde 100644 --- a/arch/powerpc/kernel/ppc_ksyms.c +++ b/arch/powerpc/kernel/ppc_ksyms.c | |||
@@ -165,6 +165,7 @@ EXPORT_SYMBOL(timer_interrupt); | |||
165 | EXPORT_SYMBOL(irq_desc); | 165 | EXPORT_SYMBOL(irq_desc); |
166 | EXPORT_SYMBOL(tb_ticks_per_jiffy); | 166 | EXPORT_SYMBOL(tb_ticks_per_jiffy); |
167 | EXPORT_SYMBOL(cacheable_memcpy); | 167 | EXPORT_SYMBOL(cacheable_memcpy); |
168 | EXPORT_SYMBOL(cacheable_memzero); | ||
168 | #endif | 169 | #endif |
169 | 170 | ||
170 | #ifdef CONFIG_PPC32 | 171 | #ifdef CONFIG_PPC32 |
diff --git a/arch/powerpc/kernel/prom.c b/arch/powerpc/kernel/prom.c index 6f73c739f1e2..c09cffafb6ee 100644 --- a/arch/powerpc/kernel/prom.c +++ b/arch/powerpc/kernel/prom.c | |||
@@ -824,11 +824,11 @@ static int __init early_init_dt_scan_chosen(unsigned long node, | |||
824 | #endif | 824 | #endif |
825 | 825 | ||
826 | #ifdef CONFIG_KEXEC | 826 | #ifdef CONFIG_KEXEC |
827 | lprop = (u64*)of_get_flat_dt_prop(node, "linux,crashkernel-base", NULL); | 827 | lprop = of_get_flat_dt_prop(node, "linux,crashkernel-base", NULL); |
828 | if (lprop) | 828 | if (lprop) |
829 | crashk_res.start = *lprop; | 829 | crashk_res.start = *lprop; |
830 | 830 | ||
831 | lprop = (u64*)of_get_flat_dt_prop(node, "linux,crashkernel-size", NULL); | 831 | lprop = of_get_flat_dt_prop(node, "linux,crashkernel-size", NULL); |
832 | if (lprop) | 832 | if (lprop) |
833 | crashk_res.end = crashk_res.start + *lprop - 1; | 833 | crashk_res.end = crashk_res.start + *lprop - 1; |
834 | #endif | 834 | #endif |
@@ -893,12 +893,12 @@ static int __init early_init_dt_scan_drconf_memory(unsigned long node) | |||
893 | u64 base, size, lmb_size; | 893 | u64 base, size, lmb_size; |
894 | unsigned int is_kexec_kdump = 0, rngs; | 894 | unsigned int is_kexec_kdump = 0, rngs; |
895 | 895 | ||
896 | ls = (cell_t *)of_get_flat_dt_prop(node, "ibm,lmb-size", &l); | 896 | ls = of_get_flat_dt_prop(node, "ibm,lmb-size", &l); |
897 | if (ls == NULL || l < dt_root_size_cells * sizeof(cell_t)) | 897 | if (ls == NULL || l < dt_root_size_cells * sizeof(cell_t)) |
898 | return 0; | 898 | return 0; |
899 | lmb_size = dt_mem_next_cell(dt_root_size_cells, &ls); | 899 | lmb_size = dt_mem_next_cell(dt_root_size_cells, &ls); |
900 | 900 | ||
901 | dm = (cell_t *)of_get_flat_dt_prop(node, "ibm,dynamic-memory", &l); | 901 | dm = of_get_flat_dt_prop(node, "ibm,dynamic-memory", &l); |
902 | if (dm == NULL || l < sizeof(cell_t)) | 902 | if (dm == NULL || l < sizeof(cell_t)) |
903 | return 0; | 903 | return 0; |
904 | 904 | ||
@@ -907,7 +907,7 @@ static int __init early_init_dt_scan_drconf_memory(unsigned long node) | |||
907 | return 0; | 907 | return 0; |
908 | 908 | ||
909 | /* check if this is a kexec/kdump kernel. */ | 909 | /* check if this is a kexec/kdump kernel. */ |
910 | usm = (cell_t *)of_get_flat_dt_prop(node, "linux,drconf-usable-memory", | 910 | usm = of_get_flat_dt_prop(node, "linux,drconf-usable-memory", |
911 | &l); | 911 | &l); |
912 | if (usm != NULL) | 912 | if (usm != NULL) |
913 | is_kexec_kdump = 1; | 913 | is_kexec_kdump = 1; |
@@ -981,9 +981,9 @@ static int __init early_init_dt_scan_memory(unsigned long node, | |||
981 | } else if (strcmp(type, "memory") != 0) | 981 | } else if (strcmp(type, "memory") != 0) |
982 | return 0; | 982 | return 0; |
983 | 983 | ||
984 | reg = (cell_t *)of_get_flat_dt_prop(node, "linux,usable-memory", &l); | 984 | reg = of_get_flat_dt_prop(node, "linux,usable-memory", &l); |
985 | if (reg == NULL) | 985 | if (reg == NULL) |
986 | reg = (cell_t *)of_get_flat_dt_prop(node, "reg", &l); | 986 | reg = of_get_flat_dt_prop(node, "reg", &l); |
987 | if (reg == NULL) | 987 | if (reg == NULL) |
988 | return 0; | 988 | return 0; |
989 | 989 | ||
diff --git a/arch/powerpc/kernel/prom_init.c b/arch/powerpc/kernel/prom_init.c index 2445945d3761..7f1b33d5e30d 100644 --- a/arch/powerpc/kernel/prom_init.c +++ b/arch/powerpc/kernel/prom_init.c | |||
@@ -1210,7 +1210,7 @@ static void __init prom_initialize_tce_table(void) | |||
1210 | /* Initialize the table to have a one-to-one mapping | 1210 | /* Initialize the table to have a one-to-one mapping |
1211 | * over the allocated size. | 1211 | * over the allocated size. |
1212 | */ | 1212 | */ |
1213 | tce_entryp = (unsigned long *)base; | 1213 | tce_entryp = (u64 *)base; |
1214 | for (i = 0; i < (minsize >> 3) ;tce_entryp++, i++) { | 1214 | for (i = 0; i < (minsize >> 3) ;tce_entryp++, i++) { |
1215 | tce_entry = (i << PAGE_SHIFT); | 1215 | tce_entry = (i << PAGE_SHIFT); |
1216 | tce_entry |= 0x3; | 1216 | tce_entry |= 0x3; |
diff --git a/arch/powerpc/kernel/sysfs.c b/arch/powerpc/kernel/sysfs.c index 0c64f10087b9..4a2ee08af6a7 100644 --- a/arch/powerpc/kernel/sysfs.c +++ b/arch/powerpc/kernel/sysfs.c | |||
@@ -18,6 +18,8 @@ | |||
18 | #include <asm/machdep.h> | 18 | #include <asm/machdep.h> |
19 | #include <asm/smp.h> | 19 | #include <asm/smp.h> |
20 | 20 | ||
21 | #include "cacheinfo.h" | ||
22 | |||
21 | #ifdef CONFIG_PPC64 | 23 | #ifdef CONFIG_PPC64 |
22 | #include <asm/paca.h> | 24 | #include <asm/paca.h> |
23 | #include <asm/lppaca.h> | 25 | #include <asm/lppaca.h> |
@@ -25,8 +27,6 @@ | |||
25 | 27 | ||
26 | static DEFINE_PER_CPU(struct cpu, cpu_devices); | 28 | static DEFINE_PER_CPU(struct cpu, cpu_devices); |
27 | 29 | ||
28 | static DEFINE_PER_CPU(struct kobject *, cache_toplevel); | ||
29 | |||
30 | /* | 30 | /* |
31 | * SMT snooze delay stuff, 64-bit only for now | 31 | * SMT snooze delay stuff, 64-bit only for now |
32 | */ | 32 | */ |
@@ -343,283 +343,6 @@ static struct sysdev_attribute pa6t_attrs[] = { | |||
343 | #endif /* HAS_PPC_PMC_PA6T */ | 343 | #endif /* HAS_PPC_PMC_PA6T */ |
344 | #endif /* HAS_PPC_PMC_CLASSIC */ | 344 | #endif /* HAS_PPC_PMC_CLASSIC */ |
345 | 345 | ||
346 | struct cache_desc { | ||
347 | struct kobject kobj; | ||
348 | struct cache_desc *next; | ||
349 | const char *type; /* Instruction, Data, or Unified */ | ||
350 | u32 size; /* total cache size in KB */ | ||
351 | u32 line_size; /* in bytes */ | ||
352 | u32 nr_sets; /* number of sets */ | ||
353 | u32 level; /* e.g. 1, 2, 3... */ | ||
354 | u32 associativity; /* e.g. 8-way... 0 is fully associative */ | ||
355 | }; | ||
356 | |||
357 | DEFINE_PER_CPU(struct cache_desc *, cache_desc); | ||
358 | |||
359 | static struct cache_desc *kobj_to_cache_desc(struct kobject *k) | ||
360 | { | ||
361 | return container_of(k, struct cache_desc, kobj); | ||
362 | } | ||
363 | |||
364 | static void cache_desc_release(struct kobject *k) | ||
365 | { | ||
366 | struct cache_desc *desc = kobj_to_cache_desc(k); | ||
367 | |||
368 | pr_debug("%s: releasing %s\n", __func__, kobject_name(k)); | ||
369 | |||
370 | if (desc->next) | ||
371 | kobject_put(&desc->next->kobj); | ||
372 | |||
373 | kfree(kobj_to_cache_desc(k)); | ||
374 | } | ||
375 | |||
376 | static ssize_t cache_desc_show(struct kobject *k, struct attribute *attr, char *buf) | ||
377 | { | ||
378 | struct kobj_attribute *kobj_attr; | ||
379 | |||
380 | kobj_attr = container_of(attr, struct kobj_attribute, attr); | ||
381 | |||
382 | return kobj_attr->show(k, kobj_attr, buf); | ||
383 | } | ||
384 | |||
385 | static struct sysfs_ops cache_desc_sysfs_ops = { | ||
386 | .show = cache_desc_show, | ||
387 | }; | ||
388 | |||
389 | static struct kobj_type cache_desc_type = { | ||
390 | .release = cache_desc_release, | ||
391 | .sysfs_ops = &cache_desc_sysfs_ops, | ||
392 | }; | ||
393 | |||
394 | static ssize_t cache_size_show(struct kobject *k, struct kobj_attribute *attr, char *buf) | ||
395 | { | ||
396 | struct cache_desc *cache = kobj_to_cache_desc(k); | ||
397 | |||
398 | return sprintf(buf, "%uK\n", cache->size); | ||
399 | } | ||
400 | |||
401 | static struct kobj_attribute cache_size_attr = | ||
402 | __ATTR(size, 0444, cache_size_show, NULL); | ||
403 | |||
404 | static ssize_t cache_line_size_show(struct kobject *k, struct kobj_attribute *attr, char *buf) | ||
405 | { | ||
406 | struct cache_desc *cache = kobj_to_cache_desc(k); | ||
407 | |||
408 | return sprintf(buf, "%u\n", cache->line_size); | ||
409 | } | ||
410 | |||
411 | static struct kobj_attribute cache_line_size_attr = | ||
412 | __ATTR(coherency_line_size, 0444, cache_line_size_show, NULL); | ||
413 | |||
414 | static ssize_t cache_nr_sets_show(struct kobject *k, struct kobj_attribute *attr, char *buf) | ||
415 | { | ||
416 | struct cache_desc *cache = kobj_to_cache_desc(k); | ||
417 | |||
418 | return sprintf(buf, "%u\n", cache->nr_sets); | ||
419 | } | ||
420 | |||
421 | static struct kobj_attribute cache_nr_sets_attr = | ||
422 | __ATTR(number_of_sets, 0444, cache_nr_sets_show, NULL); | ||
423 | |||
424 | static ssize_t cache_type_show(struct kobject *k, struct kobj_attribute *attr, char *buf) | ||
425 | { | ||
426 | struct cache_desc *cache = kobj_to_cache_desc(k); | ||
427 | |||
428 | return sprintf(buf, "%s\n", cache->type); | ||
429 | } | ||
430 | |||
431 | static struct kobj_attribute cache_type_attr = | ||
432 | __ATTR(type, 0444, cache_type_show, NULL); | ||
433 | |||
434 | static ssize_t cache_level_show(struct kobject *k, struct kobj_attribute *attr, char *buf) | ||
435 | { | ||
436 | struct cache_desc *cache = kobj_to_cache_desc(k); | ||
437 | |||
438 | return sprintf(buf, "%u\n", cache->level); | ||
439 | } | ||
440 | |||
441 | static struct kobj_attribute cache_level_attr = | ||
442 | __ATTR(level, 0444, cache_level_show, NULL); | ||
443 | |||
444 | static ssize_t cache_assoc_show(struct kobject *k, struct kobj_attribute *attr, char *buf) | ||
445 | { | ||
446 | struct cache_desc *cache = kobj_to_cache_desc(k); | ||
447 | |||
448 | return sprintf(buf, "%u\n", cache->associativity); | ||
449 | } | ||
450 | |||
451 | static struct kobj_attribute cache_assoc_attr = | ||
452 | __ATTR(ways_of_associativity, 0444, cache_assoc_show, NULL); | ||
453 | |||
454 | struct cache_desc_info { | ||
455 | const char *type; | ||
456 | const char *size_prop; | ||
457 | const char *line_size_prop; | ||
458 | const char *nr_sets_prop; | ||
459 | }; | ||
460 | |||
461 | /* PowerPC Processor binding says the [di]-cache-* must be equal on | ||
462 | * unified caches, so just use d-cache properties. */ | ||
463 | static struct cache_desc_info ucache_info = { | ||
464 | .type = "Unified", | ||
465 | .size_prop = "d-cache-size", | ||
466 | .line_size_prop = "d-cache-line-size", | ||
467 | .nr_sets_prop = "d-cache-sets", | ||
468 | }; | ||
469 | |||
470 | static struct cache_desc_info dcache_info = { | ||
471 | .type = "Data", | ||
472 | .size_prop = "d-cache-size", | ||
473 | .line_size_prop = "d-cache-line-size", | ||
474 | .nr_sets_prop = "d-cache-sets", | ||
475 | }; | ||
476 | |||
477 | static struct cache_desc_info icache_info = { | ||
478 | .type = "Instruction", | ||
479 | .size_prop = "i-cache-size", | ||
480 | .line_size_prop = "i-cache-line-size", | ||
481 | .nr_sets_prop = "i-cache-sets", | ||
482 | }; | ||
483 | |||
484 | static struct cache_desc * __cpuinit create_cache_desc(struct device_node *np, struct kobject *parent, int index, int level, struct cache_desc_info *info) | ||
485 | { | ||
486 | const u32 *cache_line_size; | ||
487 | struct cache_desc *new; | ||
488 | const u32 *cache_size; | ||
489 | const u32 *nr_sets; | ||
490 | int rc; | ||
491 | |||
492 | new = kzalloc(sizeof(*new), GFP_KERNEL); | ||
493 | if (!new) | ||
494 | return NULL; | ||
495 | |||
496 | rc = kobject_init_and_add(&new->kobj, &cache_desc_type, parent, | ||
497 | "index%d", index); | ||
498 | if (rc) | ||
499 | goto err; | ||
500 | |||
501 | /* type */ | ||
502 | new->type = info->type; | ||
503 | rc = sysfs_create_file(&new->kobj, &cache_type_attr.attr); | ||
504 | WARN_ON(rc); | ||
505 | |||
506 | /* level */ | ||
507 | new->level = level; | ||
508 | rc = sysfs_create_file(&new->kobj, &cache_level_attr.attr); | ||
509 | WARN_ON(rc); | ||
510 | |||
511 | /* size */ | ||
512 | cache_size = of_get_property(np, info->size_prop, NULL); | ||
513 | if (cache_size) { | ||
514 | new->size = *cache_size / 1024; | ||
515 | rc = sysfs_create_file(&new->kobj, | ||
516 | &cache_size_attr.attr); | ||
517 | WARN_ON(rc); | ||
518 | } | ||
519 | |||
520 | /* coherency_line_size */ | ||
521 | cache_line_size = of_get_property(np, info->line_size_prop, NULL); | ||
522 | if (cache_line_size) { | ||
523 | new->line_size = *cache_line_size; | ||
524 | rc = sysfs_create_file(&new->kobj, | ||
525 | &cache_line_size_attr.attr); | ||
526 | WARN_ON(rc); | ||
527 | } | ||
528 | |||
529 | /* number_of_sets */ | ||
530 | nr_sets = of_get_property(np, info->nr_sets_prop, NULL); | ||
531 | if (nr_sets) { | ||
532 | new->nr_sets = *nr_sets; | ||
533 | rc = sysfs_create_file(&new->kobj, | ||
534 | &cache_nr_sets_attr.attr); | ||
535 | WARN_ON(rc); | ||
536 | } | ||
537 | |||
538 | /* ways_of_associativity */ | ||
539 | if (new->nr_sets == 1) { | ||
540 | /* fully associative */ | ||
541 | new->associativity = 0; | ||
542 | goto create_assoc; | ||
543 | } | ||
544 | |||
545 | if (new->nr_sets && new->size && new->line_size) { | ||
546 | /* If we have values for all of these we can derive | ||
547 | * the associativity. */ | ||
548 | new->associativity = | ||
549 | ((new->size * 1024) / new->nr_sets) / new->line_size; | ||
550 | create_assoc: | ||
551 | rc = sysfs_create_file(&new->kobj, | ||
552 | &cache_assoc_attr.attr); | ||
553 | WARN_ON(rc); | ||
554 | } | ||
555 | |||
556 | return new; | ||
557 | err: | ||
558 | kfree(new); | ||
559 | return NULL; | ||
560 | } | ||
561 | |||
562 | static bool cache_is_unified(struct device_node *np) | ||
563 | { | ||
564 | return of_get_property(np, "cache-unified", NULL); | ||
565 | } | ||
566 | |||
567 | static struct cache_desc * __cpuinit create_cache_index_info(struct device_node *np, struct kobject *parent, int index, int level) | ||
568 | { | ||
569 | struct device_node *next_cache; | ||
570 | struct cache_desc *new, **end; | ||
571 | |||
572 | pr_debug("%s(node = %s, index = %d)\n", __func__, np->full_name, index); | ||
573 | |||
574 | if (cache_is_unified(np)) { | ||
575 | new = create_cache_desc(np, parent, index, level, | ||
576 | &ucache_info); | ||
577 | } else { | ||
578 | new = create_cache_desc(np, parent, index, level, | ||
579 | &dcache_info); | ||
580 | if (new) { | ||
581 | index++; | ||
582 | new->next = create_cache_desc(np, parent, index, level, | ||
583 | &icache_info); | ||
584 | } | ||
585 | } | ||
586 | if (!new) | ||
587 | return NULL; | ||
588 | |||
589 | end = &new->next; | ||
590 | while (*end) | ||
591 | end = &(*end)->next; | ||
592 | |||
593 | next_cache = of_find_next_cache_node(np); | ||
594 | if (!next_cache) | ||
595 | goto out; | ||
596 | |||
597 | *end = create_cache_index_info(next_cache, parent, ++index, ++level); | ||
598 | |||
599 | of_node_put(next_cache); | ||
600 | out: | ||
601 | return new; | ||
602 | } | ||
603 | |||
604 | static void __cpuinit create_cache_info(struct sys_device *sysdev) | ||
605 | { | ||
606 | struct kobject *cache_toplevel; | ||
607 | struct device_node *np = NULL; | ||
608 | int cpu = sysdev->id; | ||
609 | |||
610 | cache_toplevel = kobject_create_and_add("cache", &sysdev->kobj); | ||
611 | if (!cache_toplevel) | ||
612 | return; | ||
613 | per_cpu(cache_toplevel, cpu) = cache_toplevel; | ||
614 | np = of_get_cpu_node(cpu, NULL); | ||
615 | if (np != NULL) { | ||
616 | per_cpu(cache_desc, cpu) = | ||
617 | create_cache_index_info(np, cache_toplevel, 0, 1); | ||
618 | of_node_put(np); | ||
619 | } | ||
620 | return; | ||
621 | } | ||
622 | |||
623 | static void __cpuinit register_cpu_online(unsigned int cpu) | 346 | static void __cpuinit register_cpu_online(unsigned int cpu) |
624 | { | 347 | { |
625 | struct cpu *c = &per_cpu(cpu_devices, cpu); | 348 | struct cpu *c = &per_cpu(cpu_devices, cpu); |
@@ -684,25 +407,10 @@ static void __cpuinit register_cpu_online(unsigned int cpu) | |||
684 | sysdev_create_file(s, &attr_dscr); | 407 | sysdev_create_file(s, &attr_dscr); |
685 | #endif /* CONFIG_PPC64 */ | 408 | #endif /* CONFIG_PPC64 */ |
686 | 409 | ||
687 | create_cache_info(s); | 410 | cacheinfo_cpu_online(cpu); |
688 | } | 411 | } |
689 | 412 | ||
690 | #ifdef CONFIG_HOTPLUG_CPU | 413 | #ifdef CONFIG_HOTPLUG_CPU |
691 | static void remove_cache_info(struct sys_device *sysdev) | ||
692 | { | ||
693 | struct kobject *cache_toplevel; | ||
694 | struct cache_desc *cache_desc; | ||
695 | int cpu = sysdev->id; | ||
696 | |||
697 | cache_desc = per_cpu(cache_desc, cpu); | ||
698 | if (cache_desc != NULL) | ||
699 | kobject_put(&cache_desc->kobj); | ||
700 | |||
701 | cache_toplevel = per_cpu(cache_toplevel, cpu); | ||
702 | if (cache_toplevel != NULL) | ||
703 | kobject_put(cache_toplevel); | ||
704 | } | ||
705 | |||
706 | static void unregister_cpu_online(unsigned int cpu) | 414 | static void unregister_cpu_online(unsigned int cpu) |
707 | { | 415 | { |
708 | struct cpu *c = &per_cpu(cpu_devices, cpu); | 416 | struct cpu *c = &per_cpu(cpu_devices, cpu); |
@@ -769,7 +477,7 @@ static void unregister_cpu_online(unsigned int cpu) | |||
769 | sysdev_remove_file(s, &attr_dscr); | 477 | sysdev_remove_file(s, &attr_dscr); |
770 | #endif /* CONFIG_PPC64 */ | 478 | #endif /* CONFIG_PPC64 */ |
771 | 479 | ||
772 | remove_cache_info(s); | 480 | cacheinfo_cpu_offline(cpu); |
773 | } | 481 | } |
774 | #endif /* CONFIG_HOTPLUG_CPU */ | 482 | #endif /* CONFIG_HOTPLUG_CPU */ |
775 | 483 | ||
diff --git a/arch/powerpc/mm/mmu_decl.h b/arch/powerpc/mm/mmu_decl.h index 4314b39b6faf..ad123bced404 100644 --- a/arch/powerpc/mm/mmu_decl.h +++ b/arch/powerpc/mm/mmu_decl.h | |||
@@ -30,11 +30,11 @@ | |||
30 | #if defined(CONFIG_40x) || defined(CONFIG_8xx) | 30 | #if defined(CONFIG_40x) || defined(CONFIG_8xx) |
31 | static inline void _tlbil_all(void) | 31 | static inline void _tlbil_all(void) |
32 | { | 32 | { |
33 | asm volatile ("sync; tlbia; isync" : : : "memory") | 33 | asm volatile ("sync; tlbia; isync" : : : "memory"); |
34 | } | 34 | } |
35 | static inline void _tlbil_pid(unsigned int pid) | 35 | static inline void _tlbil_pid(unsigned int pid) |
36 | { | 36 | { |
37 | asm volatile ("sync; tlbia; isync" : : : "memory") | 37 | asm volatile ("sync; tlbia; isync" : : : "memory"); |
38 | } | 38 | } |
39 | #else /* CONFIG_40x || CONFIG_8xx */ | 39 | #else /* CONFIG_40x || CONFIG_8xx */ |
40 | extern void _tlbil_all(void); | 40 | extern void _tlbil_all(void); |
@@ -47,7 +47,7 @@ extern void _tlbil_pid(unsigned int pid); | |||
47 | #ifdef CONFIG_8xx | 47 | #ifdef CONFIG_8xx |
48 | static inline void _tlbil_va(unsigned long address, unsigned int pid) | 48 | static inline void _tlbil_va(unsigned long address, unsigned int pid) |
49 | { | 49 | { |
50 | asm volatile ("tlbie %0; sync" : : "r" (address) : "memory") | 50 | asm volatile ("tlbie %0; sync" : : "r" (address) : "memory"); |
51 | } | 51 | } |
52 | #else /* CONFIG_8xx */ | 52 | #else /* CONFIG_8xx */ |
53 | extern void _tlbil_va(unsigned long address, unsigned int pid); | 53 | extern void _tlbil_va(unsigned long address, unsigned int pid); |
diff --git a/arch/powerpc/mm/numa.c b/arch/powerpc/mm/numa.c index cf81049e1e51..7393bd76d698 100644 --- a/arch/powerpc/mm/numa.c +++ b/arch/powerpc/mm/numa.c | |||
@@ -822,42 +822,50 @@ static void __init dump_numa_memory_topology(void) | |||
822 | * required. nid is the preferred node and end is the physical address of | 822 | * required. nid is the preferred node and end is the physical address of |
823 | * the highest address in the node. | 823 | * the highest address in the node. |
824 | * | 824 | * |
825 | * Returns the physical address of the memory. | 825 | * Returns the virtual address of the memory. |
826 | */ | 826 | */ |
827 | static void __init *careful_allocation(int nid, unsigned long size, | 827 | static void __init *careful_zallocation(int nid, unsigned long size, |
828 | unsigned long align, | 828 | unsigned long align, |
829 | unsigned long end_pfn) | 829 | unsigned long end_pfn) |
830 | { | 830 | { |
831 | void *ret; | ||
831 | int new_nid; | 832 | int new_nid; |
832 | unsigned long ret = __lmb_alloc_base(size, align, end_pfn << PAGE_SHIFT); | 833 | unsigned long ret_paddr; |
834 | |||
835 | ret_paddr = __lmb_alloc_base(size, align, end_pfn << PAGE_SHIFT); | ||
833 | 836 | ||
834 | /* retry over all memory */ | 837 | /* retry over all memory */ |
835 | if (!ret) | 838 | if (!ret_paddr) |
836 | ret = __lmb_alloc_base(size, align, lmb_end_of_DRAM()); | 839 | ret_paddr = __lmb_alloc_base(size, align, lmb_end_of_DRAM()); |
837 | 840 | ||
838 | if (!ret) | 841 | if (!ret_paddr) |
839 | panic("numa.c: cannot allocate %lu bytes on node %d", | 842 | panic("numa.c: cannot allocate %lu bytes for node %d", |
840 | size, nid); | 843 | size, nid); |
841 | 844 | ||
845 | ret = __va(ret_paddr); | ||
846 | |||
842 | /* | 847 | /* |
843 | * If the memory came from a previously allocated node, we must | 848 | * We initialize the nodes in numeric order: 0, 1, 2... |
844 | * retry with the bootmem allocator. | 849 | * and hand over control from the LMB allocator to the |
850 | * bootmem allocator. If this function is called for | ||
851 | * node 5, then we know that all nodes <5 are using the | ||
852 | * bootmem allocator instead of the LMB allocator. | ||
853 | * | ||
854 | * So, check the nid from which this allocation came | ||
855 | * and double check to see if we need to use bootmem | ||
856 | * instead of the LMB. We don't free the LMB memory | ||
857 | * since it would be useless. | ||
845 | */ | 858 | */ |
846 | new_nid = early_pfn_to_nid(ret >> PAGE_SHIFT); | 859 | new_nid = early_pfn_to_nid(ret_paddr >> PAGE_SHIFT); |
847 | if (new_nid < nid) { | 860 | if (new_nid < nid) { |
848 | ret = (unsigned long)__alloc_bootmem_node(NODE_DATA(new_nid), | 861 | ret = __alloc_bootmem_node(NODE_DATA(new_nid), |
849 | size, align, 0); | 862 | size, align, 0); |
850 | 863 | ||
851 | if (!ret) | 864 | dbg("alloc_bootmem %p %lx\n", ret, size); |
852 | panic("numa.c: cannot allocate %lu bytes on node %d", | ||
853 | size, new_nid); | ||
854 | |||
855 | ret = __pa(ret); | ||
856 | |||
857 | dbg("alloc_bootmem %lx %lx\n", ret, size); | ||
858 | } | 865 | } |
859 | 866 | ||
860 | return (void *)ret; | 867 | memset(ret, 0, size); |
868 | return ret; | ||
861 | } | 869 | } |
862 | 870 | ||
863 | static struct notifier_block __cpuinitdata ppc64_numa_nb = { | 871 | static struct notifier_block __cpuinitdata ppc64_numa_nb = { |
@@ -952,7 +960,7 @@ void __init do_init_bootmem(void) | |||
952 | 960 | ||
953 | for_each_online_node(nid) { | 961 | for_each_online_node(nid) { |
954 | unsigned long start_pfn, end_pfn; | 962 | unsigned long start_pfn, end_pfn; |
955 | unsigned long bootmem_paddr; | 963 | void *bootmem_vaddr; |
956 | unsigned long bootmap_pages; | 964 | unsigned long bootmap_pages; |
957 | 965 | ||
958 | get_pfn_range_for_nid(nid, &start_pfn, &end_pfn); | 966 | get_pfn_range_for_nid(nid, &start_pfn, &end_pfn); |
@@ -964,11 +972,9 @@ void __init do_init_bootmem(void) | |||
964 | * previous nodes' bootmem to be initialized and have | 972 | * previous nodes' bootmem to be initialized and have |
965 | * all reserved areas marked. | 973 | * all reserved areas marked. |
966 | */ | 974 | */ |
967 | NODE_DATA(nid) = careful_allocation(nid, | 975 | NODE_DATA(nid) = careful_zallocation(nid, |
968 | sizeof(struct pglist_data), | 976 | sizeof(struct pglist_data), |
969 | SMP_CACHE_BYTES, end_pfn); | 977 | SMP_CACHE_BYTES, end_pfn); |
970 | NODE_DATA(nid) = __va(NODE_DATA(nid)); | ||
971 | memset(NODE_DATA(nid), 0, sizeof(struct pglist_data)); | ||
972 | 978 | ||
973 | dbg("node %d\n", nid); | 979 | dbg("node %d\n", nid); |
974 | dbg("NODE_DATA() = %p\n", NODE_DATA(nid)); | 980 | dbg("NODE_DATA() = %p\n", NODE_DATA(nid)); |
@@ -984,20 +990,20 @@ void __init do_init_bootmem(void) | |||
984 | dbg("end_paddr = %lx\n", end_pfn << PAGE_SHIFT); | 990 | dbg("end_paddr = %lx\n", end_pfn << PAGE_SHIFT); |
985 | 991 | ||
986 | bootmap_pages = bootmem_bootmap_pages(end_pfn - start_pfn); | 992 | bootmap_pages = bootmem_bootmap_pages(end_pfn - start_pfn); |
987 | bootmem_paddr = (unsigned long)careful_allocation(nid, | 993 | bootmem_vaddr = careful_zallocation(nid, |
988 | bootmap_pages << PAGE_SHIFT, | 994 | bootmap_pages << PAGE_SHIFT, |
989 | PAGE_SIZE, end_pfn); | 995 | PAGE_SIZE, end_pfn); |
990 | memset(__va(bootmem_paddr), 0, bootmap_pages << PAGE_SHIFT); | ||
991 | 996 | ||
992 | dbg("bootmap_paddr = %lx\n", bootmem_paddr); | 997 | dbg("bootmap_vaddr = %p\n", bootmem_vaddr); |
993 | 998 | ||
994 | init_bootmem_node(NODE_DATA(nid), bootmem_paddr >> PAGE_SHIFT, | 999 | init_bootmem_node(NODE_DATA(nid), |
1000 | __pa(bootmem_vaddr) >> PAGE_SHIFT, | ||
995 | start_pfn, end_pfn); | 1001 | start_pfn, end_pfn); |
996 | 1002 | ||
997 | free_bootmem_with_active_regions(nid, end_pfn); | 1003 | free_bootmem_with_active_regions(nid, end_pfn); |
998 | /* | 1004 | /* |
999 | * Be very careful about moving this around. Future | 1005 | * Be very careful about moving this around. Future |
1000 | * calls to careful_allocation() depend on this getting | 1006 | * calls to careful_zallocation() depend on this getting |
1001 | * done correctly. | 1007 | * done correctly. |
1002 | */ | 1008 | */ |
1003 | mark_reserved_regions_for_nid(nid); | 1009 | mark_reserved_regions_for_nid(nid); |
diff --git a/arch/powerpc/mm/pgtable_32.c b/arch/powerpc/mm/pgtable_32.c index 38ff35f2142a..22972cd83cc9 100644 --- a/arch/powerpc/mm/pgtable_32.c +++ b/arch/powerpc/mm/pgtable_32.c | |||
@@ -266,7 +266,8 @@ int map_page(unsigned long va, phys_addr_t pa, int flags) | |||
266 | /* The PTE should never be already set nor present in the | 266 | /* The PTE should never be already set nor present in the |
267 | * hash table | 267 | * hash table |
268 | */ | 268 | */ |
269 | BUG_ON(pte_val(*pg) & (_PAGE_PRESENT | _PAGE_HASHPTE)); | 269 | BUG_ON((pte_val(*pg) & (_PAGE_PRESENT | _PAGE_HASHPTE)) && |
270 | flags); | ||
270 | set_pte_at(&init_mm, va, pg, pfn_pte(pa >> PAGE_SHIFT, | 271 | set_pte_at(&init_mm, va, pg, pfn_pte(pa >> PAGE_SHIFT, |
271 | __pgprot(flags))); | 272 | __pgprot(flags))); |
272 | } | 273 | } |
diff --git a/arch/powerpc/mm/tlb_nohash.c b/arch/powerpc/mm/tlb_nohash.c index 803a64c02b06..39ac22b13c73 100644 --- a/arch/powerpc/mm/tlb_nohash.c +++ b/arch/powerpc/mm/tlb_nohash.c | |||
@@ -189,8 +189,9 @@ void flush_tlb_kernel_range(unsigned long start, unsigned long end) | |||
189 | smp_call_function(do_flush_tlb_mm_ipi, NULL, 1); | 189 | smp_call_function(do_flush_tlb_mm_ipi, NULL, 1); |
190 | _tlbil_pid(0); | 190 | _tlbil_pid(0); |
191 | preempt_enable(); | 191 | preempt_enable(); |
192 | #endif | 192 | #else |
193 | _tlbil_pid(0); | 193 | _tlbil_pid(0); |
194 | #endif | ||
194 | } | 195 | } |
195 | EXPORT_SYMBOL(flush_tlb_kernel_range); | 196 | EXPORT_SYMBOL(flush_tlb_kernel_range); |
196 | 197 | ||
diff --git a/arch/powerpc/oprofile/cell/pr_util.h b/arch/powerpc/oprofile/cell/pr_util.h index 628009c01958..dfdbffa06818 100644 --- a/arch/powerpc/oprofile/cell/pr_util.h +++ b/arch/powerpc/oprofile/cell/pr_util.h | |||
@@ -79,7 +79,7 @@ struct spu_buffer { | |||
79 | * the vma-to-fileoffset map. | 79 | * the vma-to-fileoffset map. |
80 | */ | 80 | */ |
81 | struct vma_to_fileoffset_map *create_vma_map(const struct spu *spu, | 81 | struct vma_to_fileoffset_map *create_vma_map(const struct spu *spu, |
82 | u64 objectid); | 82 | unsigned long objectid); |
83 | unsigned int vma_map_lookup(struct vma_to_fileoffset_map *map, | 83 | unsigned int vma_map_lookup(struct vma_to_fileoffset_map *map, |
84 | unsigned int vma, const struct spu *aSpu, | 84 | unsigned int vma, const struct spu *aSpu, |
85 | int *grd_val); | 85 | int *grd_val); |
diff --git a/arch/powerpc/platforms/52xx/mpc52xx_common.c b/arch/powerpc/platforms/52xx/mpc52xx_common.c index ae7c34f37e1c..98367a0255f3 100644 --- a/arch/powerpc/platforms/52xx/mpc52xx_common.c +++ b/arch/powerpc/platforms/52xx/mpc52xx_common.c | |||
@@ -42,7 +42,7 @@ static struct of_device_id mpc52xx_bus_ids[] __initdata = { | |||
42 | * from interrupt context while node mapping (which calls ioremap()) | 42 | * from interrupt context while node mapping (which calls ioremap()) |
43 | * cannot be used at such point. | 43 | * cannot be used at such point. |
44 | */ | 44 | */ |
45 | static spinlock_t mpc52xx_lock = SPIN_LOCK_UNLOCKED; | 45 | static DEFINE_SPINLOCK(mpc52xx_lock); |
46 | static struct mpc52xx_gpt __iomem *mpc52xx_wdt; | 46 | static struct mpc52xx_gpt __iomem *mpc52xx_wdt; |
47 | static struct mpc52xx_cdm __iomem *mpc52xx_cdm; | 47 | static struct mpc52xx_cdm __iomem *mpc52xx_cdm; |
48 | 48 | ||
diff --git a/arch/powerpc/platforms/83xx/mpc831x_rdb.c b/arch/powerpc/platforms/83xx/mpc831x_rdb.c index a428f8d1ac80..5177bdd2c62a 100644 --- a/arch/powerpc/platforms/83xx/mpc831x_rdb.c +++ b/arch/powerpc/platforms/83xx/mpc831x_rdb.c | |||
@@ -42,7 +42,7 @@ static void __init mpc831x_rdb_setup_arch(void) | |||
42 | mpc831x_usb_cfg(); | 42 | mpc831x_usb_cfg(); |
43 | } | 43 | } |
44 | 44 | ||
45 | void __init mpc831x_rdb_init_IRQ(void) | 45 | static void __init mpc831x_rdb_init_IRQ(void) |
46 | { | 46 | { |
47 | struct device_node *np; | 47 | struct device_node *np; |
48 | 48 | ||
diff --git a/arch/powerpc/platforms/83xx/mpc832x_mds.c b/arch/powerpc/platforms/83xx/mpc832x_mds.c index ec43477caa63..ec0b401bc9cf 100644 --- a/arch/powerpc/platforms/83xx/mpc832x_mds.c +++ b/arch/powerpc/platforms/83xx/mpc832x_mds.c | |||
@@ -49,8 +49,6 @@ | |||
49 | #define DBG(fmt...) | 49 | #define DBG(fmt...) |
50 | #endif | 50 | #endif |
51 | 51 | ||
52 | static u8 *bcsr_regs = NULL; | ||
53 | |||
54 | /* ************************************************************************ | 52 | /* ************************************************************************ |
55 | * | 53 | * |
56 | * Setup the architecture | 54 | * Setup the architecture |
@@ -59,13 +57,14 @@ static u8 *bcsr_regs = NULL; | |||
59 | static void __init mpc832x_sys_setup_arch(void) | 57 | static void __init mpc832x_sys_setup_arch(void) |
60 | { | 58 | { |
61 | struct device_node *np; | 59 | struct device_node *np; |
60 | u8 __iomem *bcsr_regs = NULL; | ||
62 | 61 | ||
63 | if (ppc_md.progress) | 62 | if (ppc_md.progress) |
64 | ppc_md.progress("mpc832x_sys_setup_arch()", 0); | 63 | ppc_md.progress("mpc832x_sys_setup_arch()", 0); |
65 | 64 | ||
66 | /* Map BCSR area */ | 65 | /* Map BCSR area */ |
67 | np = of_find_node_by_name(NULL, "bcsr"); | 66 | np = of_find_node_by_name(NULL, "bcsr"); |
68 | if (np != 0) { | 67 | if (np) { |
69 | struct resource res; | 68 | struct resource res; |
70 | 69 | ||
71 | of_address_to_resource(np, 0, &res); | 70 | of_address_to_resource(np, 0, &res); |
@@ -93,9 +92,9 @@ static void __init mpc832x_sys_setup_arch(void) | |||
93 | != NULL){ | 92 | != NULL){ |
94 | /* Reset the Ethernet PHYs */ | 93 | /* Reset the Ethernet PHYs */ |
95 | #define BCSR8_FETH_RST 0x50 | 94 | #define BCSR8_FETH_RST 0x50 |
96 | bcsr_regs[8] &= ~BCSR8_FETH_RST; | 95 | clrbits8(&bcsr_regs[8], BCSR8_FETH_RST); |
97 | udelay(1000); | 96 | udelay(1000); |
98 | bcsr_regs[8] |= BCSR8_FETH_RST; | 97 | setbits8(&bcsr_regs[8], BCSR8_FETH_RST); |
99 | iounmap(bcsr_regs); | 98 | iounmap(bcsr_regs); |
100 | of_node_put(np); | 99 | of_node_put(np); |
101 | } | 100 | } |
diff --git a/arch/powerpc/platforms/83xx/mpc832x_rdb.c b/arch/powerpc/platforms/83xx/mpc832x_rdb.c index 0300268ce5b8..2a1295f19832 100644 --- a/arch/powerpc/platforms/83xx/mpc832x_rdb.c +++ b/arch/powerpc/platforms/83xx/mpc832x_rdb.c | |||
@@ -38,6 +38,7 @@ | |||
38 | #define DBG(fmt...) | 38 | #define DBG(fmt...) |
39 | #endif | 39 | #endif |
40 | 40 | ||
41 | #ifdef CONFIG_QUICC_ENGINE | ||
41 | static void mpc83xx_spi_activate_cs(u8 cs, u8 polarity) | 42 | static void mpc83xx_spi_activate_cs(u8 cs, u8 polarity) |
42 | { | 43 | { |
43 | pr_debug("%s %d %d\n", __func__, cs, polarity); | 44 | pr_debug("%s %d %d\n", __func__, cs, polarity); |
@@ -77,8 +78,8 @@ static int __init mpc832x_spi_init(void) | |||
77 | mpc83xx_spi_activate_cs, | 78 | mpc83xx_spi_activate_cs, |
78 | mpc83xx_spi_deactivate_cs); | 79 | mpc83xx_spi_deactivate_cs); |
79 | } | 80 | } |
80 | |||
81 | machine_device_initcall(mpc832x_rdb, mpc832x_spi_init); | 81 | machine_device_initcall(mpc832x_rdb, mpc832x_spi_init); |
82 | #endif /* CONFIG_QUICC_ENGINE */ | ||
82 | 83 | ||
83 | /* ************************************************************************ | 84 | /* ************************************************************************ |
84 | * | 85 | * |
@@ -130,7 +131,7 @@ static int __init mpc832x_declare_of_platform_devices(void) | |||
130 | } | 131 | } |
131 | machine_device_initcall(mpc832x_rdb, mpc832x_declare_of_platform_devices); | 132 | machine_device_initcall(mpc832x_rdb, mpc832x_declare_of_platform_devices); |
132 | 133 | ||
133 | void __init mpc832x_rdb_init_IRQ(void) | 134 | static void __init mpc832x_rdb_init_IRQ(void) |
134 | { | 135 | { |
135 | 136 | ||
136 | struct device_node *np; | 137 | struct device_node *np; |
diff --git a/arch/powerpc/platforms/83xx/mpc836x_mds.c b/arch/powerpc/platforms/83xx/mpc836x_mds.c index 9d46e5bdd101..09e9d6fb7411 100644 --- a/arch/powerpc/platforms/83xx/mpc836x_mds.c +++ b/arch/powerpc/platforms/83xx/mpc836x_mds.c | |||
@@ -18,6 +18,7 @@ | |||
18 | 18 | ||
19 | #include <linux/stddef.h> | 19 | #include <linux/stddef.h> |
20 | #include <linux/kernel.h> | 20 | #include <linux/kernel.h> |
21 | #include <linux/compiler.h> | ||
21 | #include <linux/init.h> | 22 | #include <linux/init.h> |
22 | #include <linux/errno.h> | 23 | #include <linux/errno.h> |
23 | #include <linux/reboot.h> | 24 | #include <linux/reboot.h> |
@@ -43,6 +44,7 @@ | |||
43 | #include <asm/udbg.h> | 44 | #include <asm/udbg.h> |
44 | #include <sysdev/fsl_soc.h> | 45 | #include <sysdev/fsl_soc.h> |
45 | #include <sysdev/fsl_pci.h> | 46 | #include <sysdev/fsl_pci.h> |
47 | #include <sysdev/simple_gpio.h> | ||
46 | #include <asm/qe.h> | 48 | #include <asm/qe.h> |
47 | #include <asm/qe_ic.h> | 49 | #include <asm/qe_ic.h> |
48 | 50 | ||
@@ -55,8 +57,6 @@ | |||
55 | #define DBG(fmt...) | 57 | #define DBG(fmt...) |
56 | #endif | 58 | #endif |
57 | 59 | ||
58 | static u8 *bcsr_regs = NULL; | ||
59 | |||
60 | /* ************************************************************************ | 60 | /* ************************************************************************ |
61 | * | 61 | * |
62 | * Setup the architecture | 62 | * Setup the architecture |
@@ -65,13 +65,14 @@ static u8 *bcsr_regs = NULL; | |||
65 | static void __init mpc836x_mds_setup_arch(void) | 65 | static void __init mpc836x_mds_setup_arch(void) |
66 | { | 66 | { |
67 | struct device_node *np; | 67 | struct device_node *np; |
68 | u8 __iomem *bcsr_regs = NULL; | ||
68 | 69 | ||
69 | if (ppc_md.progress) | 70 | if (ppc_md.progress) |
70 | ppc_md.progress("mpc836x_mds_setup_arch()", 0); | 71 | ppc_md.progress("mpc836x_mds_setup_arch()", 0); |
71 | 72 | ||
72 | /* Map BCSR area */ | 73 | /* Map BCSR area */ |
73 | np = of_find_node_by_name(NULL, "bcsr"); | 74 | np = of_find_node_by_name(NULL, "bcsr"); |
74 | if (np != 0) { | 75 | if (np) { |
75 | struct resource res; | 76 | struct resource res; |
76 | 77 | ||
77 | of_address_to_resource(np, 0, &res); | 78 | of_address_to_resource(np, 0, &res); |
@@ -93,6 +94,16 @@ static void __init mpc836x_mds_setup_arch(void) | |||
93 | 94 | ||
94 | for (np = NULL; (np = of_find_node_by_name(np, "ucc")) != NULL;) | 95 | for (np = NULL; (np = of_find_node_by_name(np, "ucc")) != NULL;) |
95 | par_io_of_config(np); | 96 | par_io_of_config(np); |
97 | #ifdef CONFIG_QE_USB | ||
98 | /* Must fixup Par IO before QE GPIO chips are registered. */ | ||
99 | par_io_config_pin(1, 2, 1, 0, 3, 0); /* USBOE */ | ||
100 | par_io_config_pin(1, 3, 1, 0, 3, 0); /* USBTP */ | ||
101 | par_io_config_pin(1, 8, 1, 0, 1, 0); /* USBTN */ | ||
102 | par_io_config_pin(1, 10, 2, 0, 3, 0); /* USBRXD */ | ||
103 | par_io_config_pin(1, 9, 2, 1, 3, 0); /* USBRP */ | ||
104 | par_io_config_pin(1, 11, 2, 1, 3, 0); /* USBRN */ | ||
105 | par_io_config_pin(2, 20, 2, 0, 1, 0); /* CLK21 */ | ||
106 | #endif /* CONFIG_QE_USB */ | ||
96 | } | 107 | } |
97 | 108 | ||
98 | if ((np = of_find_compatible_node(NULL, "network", "ucc_geth")) | 109 | if ((np = of_find_compatible_node(NULL, "network", "ucc_geth")) |
@@ -151,6 +162,70 @@ static int __init mpc836x_declare_of_platform_devices(void) | |||
151 | } | 162 | } |
152 | machine_device_initcall(mpc836x_mds, mpc836x_declare_of_platform_devices); | 163 | machine_device_initcall(mpc836x_mds, mpc836x_declare_of_platform_devices); |
153 | 164 | ||
165 | #ifdef CONFIG_QE_USB | ||
166 | static int __init mpc836x_usb_cfg(void) | ||
167 | { | ||
168 | u8 __iomem *bcsr; | ||
169 | struct device_node *np; | ||
170 | const char *mode; | ||
171 | int ret = 0; | ||
172 | |||
173 | np = of_find_compatible_node(NULL, NULL, "fsl,mpc8360mds-bcsr"); | ||
174 | if (!np) | ||
175 | return -ENODEV; | ||
176 | |||
177 | bcsr = of_iomap(np, 0); | ||
178 | of_node_put(np); | ||
179 | if (!bcsr) | ||
180 | return -ENOMEM; | ||
181 | |||
182 | np = of_find_compatible_node(NULL, NULL, "fsl,mpc8323-qe-usb"); | ||
183 | if (!np) { | ||
184 | ret = -ENODEV; | ||
185 | goto err; | ||
186 | } | ||
187 | |||
188 | #define BCSR8_TSEC1M_MASK (0x3 << 6) | ||
189 | #define BCSR8_TSEC1M_RGMII (0x0 << 6) | ||
190 | #define BCSR8_TSEC2M_MASK (0x3 << 4) | ||
191 | #define BCSR8_TSEC2M_RGMII (0x0 << 4) | ||
192 | /* | ||
193 | * Default is GMII (2), but we should set it to RGMII (0) if we use | ||
194 | * USB (Eth PHY is in RGMII mode anyway). | ||
195 | */ | ||
196 | clrsetbits_8(&bcsr[8], BCSR8_TSEC1M_MASK | BCSR8_TSEC2M_MASK, | ||
197 | BCSR8_TSEC1M_RGMII | BCSR8_TSEC2M_RGMII); | ||
198 | |||
199 | #define BCSR13_USBMASK 0x0f | ||
200 | #define BCSR13_nUSBEN 0x08 /* 1 - Disable, 0 - Enable */ | ||
201 | #define BCSR13_USBSPEED 0x04 /* 1 - Full, 0 - Low */ | ||
202 | #define BCSR13_USBMODE 0x02 /* 1 - Host, 0 - Function */ | ||
203 | #define BCSR13_nUSBVCC 0x01 /* 1 - gets VBUS, 0 - supplies VBUS */ | ||
204 | |||
205 | clrsetbits_8(&bcsr[13], BCSR13_USBMASK, BCSR13_USBSPEED); | ||
206 | |||
207 | mode = of_get_property(np, "mode", NULL); | ||
208 | if (mode && !strcmp(mode, "peripheral")) { | ||
209 | setbits8(&bcsr[13], BCSR13_nUSBVCC); | ||
210 | qe_usb_clock_set(QE_CLK21, 48000000); | ||
211 | } else { | ||
212 | setbits8(&bcsr[13], BCSR13_USBMODE); | ||
213 | /* | ||
214 | * The BCSR GPIOs are used to control power and | ||
215 | * speed of the USB transceiver. This is needed for | ||
216 | * the USB Host only. | ||
217 | */ | ||
218 | simple_gpiochip_init("fsl,mpc8360mds-bcsr-gpio"); | ||
219 | } | ||
220 | |||
221 | of_node_put(np); | ||
222 | err: | ||
223 | iounmap(bcsr); | ||
224 | return ret; | ||
225 | } | ||
226 | machine_arch_initcall(mpc836x_mds, mpc836x_usb_cfg); | ||
227 | #endif /* CONFIG_QE_USB */ | ||
228 | |||
154 | static void __init mpc836x_mds_init_IRQ(void) | 229 | static void __init mpc836x_mds_init_IRQ(void) |
155 | { | 230 | { |
156 | struct device_node *np; | 231 | struct device_node *np; |
diff --git a/arch/powerpc/platforms/83xx/mpc836x_rdk.c b/arch/powerpc/platforms/83xx/mpc836x_rdk.c index a5273bb28e1b..b0090aac9642 100644 --- a/arch/powerpc/platforms/83xx/mpc836x_rdk.c +++ b/arch/powerpc/platforms/83xx/mpc836x_rdk.c | |||
@@ -51,8 +51,9 @@ static void __init mpc836x_rdk_setup_arch(void) | |||
51 | for_each_compatible_node(np, "pci", "fsl,mpc8349-pci") | 51 | for_each_compatible_node(np, "pci", "fsl,mpc8349-pci") |
52 | mpc83xx_add_bridge(np); | 52 | mpc83xx_add_bridge(np); |
53 | #endif | 53 | #endif |
54 | 54 | #ifdef CONFIG_QUICC_ENGINE | |
55 | qe_reset(); | 55 | qe_reset(); |
56 | #endif | ||
56 | } | 57 | } |
57 | 58 | ||
58 | static void __init mpc836x_rdk_init_IRQ(void) | 59 | static void __init mpc836x_rdk_init_IRQ(void) |
@@ -71,13 +72,14 @@ static void __init mpc836x_rdk_init_IRQ(void) | |||
71 | */ | 72 | */ |
72 | ipic_set_default_priority(); | 73 | ipic_set_default_priority(); |
73 | of_node_put(np); | 74 | of_node_put(np); |
74 | 75 | #ifdef CONFIG_QUICC_ENGINE | |
75 | np = of_find_compatible_node(NULL, NULL, "fsl,qe-ic"); | 76 | np = of_find_compatible_node(NULL, NULL, "fsl,qe-ic"); |
76 | if (!np) | 77 | if (!np) |
77 | return; | 78 | return; |
78 | 79 | ||
79 | qe_ic_init(np, 0, qe_ic_cascade_low_ipic, qe_ic_cascade_high_ipic); | 80 | qe_ic_init(np, 0, qe_ic_cascade_low_ipic, qe_ic_cascade_high_ipic); |
80 | of_node_put(np); | 81 | of_node_put(np); |
82 | #endif | ||
81 | } | 83 | } |
82 | 84 | ||
83 | /* | 85 | /* |
diff --git a/arch/powerpc/platforms/83xx/mpc837x_mds.c b/arch/powerpc/platforms/83xx/mpc837x_mds.c index 8bb13c807142..530ef990ca7c 100644 --- a/arch/powerpc/platforms/83xx/mpc837x_mds.c +++ b/arch/powerpc/platforms/83xx/mpc837x_mds.c | |||
@@ -26,7 +26,6 @@ | |||
26 | #define BCSR12_USB_SER_MASK 0x8a | 26 | #define BCSR12_USB_SER_MASK 0x8a |
27 | #define BCSR12_USB_SER_PIN 0x80 | 27 | #define BCSR12_USB_SER_PIN 0x80 |
28 | #define BCSR12_USB_SER_DEVICE 0x02 | 28 | #define BCSR12_USB_SER_DEVICE 0x02 |
29 | extern int mpc837x_usb_cfg(void); | ||
30 | 29 | ||
31 | static int mpc837xmds_usb_cfg(void) | 30 | static int mpc837xmds_usb_cfg(void) |
32 | { | 31 | { |
diff --git a/arch/powerpc/platforms/83xx/mpc837x_rdb.c b/arch/powerpc/platforms/83xx/mpc837x_rdb.c index da030afa2e2c..1d096545322b 100644 --- a/arch/powerpc/platforms/83xx/mpc837x_rdb.c +++ b/arch/powerpc/platforms/83xx/mpc837x_rdb.c | |||
@@ -21,8 +21,6 @@ | |||
21 | 21 | ||
22 | #include "mpc83xx.h" | 22 | #include "mpc83xx.h" |
23 | 23 | ||
24 | extern int mpc837x_usb_cfg(void); | ||
25 | |||
26 | /* ************************************************************************ | 24 | /* ************************************************************************ |
27 | * | 25 | * |
28 | * Setup the architecture | 26 | * Setup the architecture |
diff --git a/arch/powerpc/platforms/83xx/mpc83xx.h b/arch/powerpc/platforms/83xx/mpc83xx.h index 2a7cbabb410a..83cfe51526ec 100644 --- a/arch/powerpc/platforms/83xx/mpc83xx.h +++ b/arch/powerpc/platforms/83xx/mpc83xx.h | |||
@@ -61,6 +61,7 @@ | |||
61 | 61 | ||
62 | extern void mpc83xx_restart(char *cmd); | 62 | extern void mpc83xx_restart(char *cmd); |
63 | extern long mpc83xx_time_init(void); | 63 | extern long mpc83xx_time_init(void); |
64 | extern int mpc837x_usb_cfg(void); | ||
64 | extern int mpc834x_usb_cfg(void); | 65 | extern int mpc834x_usb_cfg(void); |
65 | extern int mpc831x_usb_cfg(void); | 66 | extern int mpc831x_usb_cfg(void); |
66 | 67 | ||
diff --git a/arch/powerpc/platforms/85xx/mpc85xx_ds.c b/arch/powerpc/platforms/85xx/mpc85xx_ds.c index a8301c8ad537..7326d904202c 100644 --- a/arch/powerpc/platforms/85xx/mpc85xx_ds.c +++ b/arch/powerpc/platforms/85xx/mpc85xx_ds.c | |||
@@ -148,6 +148,9 @@ static int mpc85xx_exclude_device(struct pci_controller *hose, | |||
148 | /* | 148 | /* |
149 | * Setup the architecture | 149 | * Setup the architecture |
150 | */ | 150 | */ |
151 | #ifdef CONFIG_SMP | ||
152 | extern void __init mpc85xx_smp_init(void); | ||
153 | #endif | ||
151 | static void __init mpc85xx_ds_setup_arch(void) | 154 | static void __init mpc85xx_ds_setup_arch(void) |
152 | { | 155 | { |
153 | #ifdef CONFIG_PCI | 156 | #ifdef CONFIG_PCI |
@@ -173,6 +176,10 @@ static void __init mpc85xx_ds_setup_arch(void) | |||
173 | ppc_md.pci_exclude_device = mpc85xx_exclude_device; | 176 | ppc_md.pci_exclude_device = mpc85xx_exclude_device; |
174 | #endif | 177 | #endif |
175 | 178 | ||
179 | #ifdef CONFIG_SMP | ||
180 | mpc85xx_smp_init(); | ||
181 | #endif | ||
182 | |||
176 | printk("MPC85xx DS board from Freescale Semiconductor\n"); | 183 | printk("MPC85xx DS board from Freescale Semiconductor\n"); |
177 | } | 184 | } |
178 | 185 | ||
diff --git a/arch/powerpc/platforms/85xx/smp.c b/arch/powerpc/platforms/85xx/smp.c index d652c713f496..79a0df17078b 100644 --- a/arch/powerpc/platforms/85xx/smp.c +++ b/arch/powerpc/platforms/85xx/smp.c | |||
@@ -58,6 +58,7 @@ smp_85xx_kick_cpu(int nr) | |||
58 | 58 | ||
59 | if (cpu_rel_addr == NULL) { | 59 | if (cpu_rel_addr == NULL) { |
60 | printk(KERN_ERR "No cpu-release-addr for cpu %d\n", nr); | 60 | printk(KERN_ERR "No cpu-release-addr for cpu %d\n", nr); |
61 | local_irq_restore(flags); | ||
61 | return; | 62 | return; |
62 | } | 63 | } |
63 | 64 | ||
diff --git a/arch/powerpc/platforms/Kconfig b/arch/powerpc/platforms/Kconfig index 47e956c871fe..47fe2bea9865 100644 --- a/arch/powerpc/platforms/Kconfig +++ b/arch/powerpc/platforms/Kconfig | |||
@@ -312,4 +312,15 @@ config MPC8xxx_GPIO | |||
312 | Say Y here if you're going to use hardware that connects to the | 312 | Say Y here if you're going to use hardware that connects to the |
313 | MPC831x/834x/837x/8572/8610 GPIOs. | 313 | MPC831x/834x/837x/8572/8610 GPIOs. |
314 | 314 | ||
315 | config SIMPLE_GPIO | ||
316 | bool "Support for simple, memory-mapped GPIO controllers" | ||
317 | depends on PPC | ||
318 | select GENERIC_GPIO | ||
319 | select ARCH_REQUIRE_GPIOLIB | ||
320 | help | ||
321 | Say Y here to support simple, memory-mapped GPIO controllers. | ||
322 | These are usually BCSRs used to control board's switches, LEDs, | ||
323 | chip-selects, Ethernet/USB PHY's power and various other small | ||
324 | on-board peripherals. | ||
325 | |||
315 | endmenu | 326 | endmenu |
diff --git a/arch/powerpc/platforms/Kconfig.cputype b/arch/powerpc/platforms/Kconfig.cputype index 3d0c776f888d..e868b5c50723 100644 --- a/arch/powerpc/platforms/Kconfig.cputype +++ b/arch/powerpc/platforms/Kconfig.cputype | |||
@@ -231,7 +231,7 @@ config VIRT_CPU_ACCOUNTING | |||
231 | If in doubt, say Y here. | 231 | If in doubt, say Y here. |
232 | 232 | ||
233 | config SMP | 233 | config SMP |
234 | depends on PPC_STD_MMU | 234 | depends on PPC_STD_MMU || FSL_BOOKE |
235 | bool "Symmetric multi-processing support" | 235 | bool "Symmetric multi-processing support" |
236 | ---help--- | 236 | ---help--- |
237 | This enables support for systems with more than one CPU. If you have | 237 | This enables support for systems with more than one CPU. If you have |
diff --git a/arch/powerpc/platforms/cell/beat_htab.c b/arch/powerpc/platforms/cell/beat_htab.c index 2e67bd840e01..35b1ec492715 100644 --- a/arch/powerpc/platforms/cell/beat_htab.c +++ b/arch/powerpc/platforms/cell/beat_htab.c | |||
@@ -44,8 +44,8 @@ static DEFINE_SPINLOCK(beat_htab_lock); | |||
44 | 44 | ||
45 | static inline unsigned int beat_read_mask(unsigned hpte_group) | 45 | static inline unsigned int beat_read_mask(unsigned hpte_group) |
46 | { | 46 | { |
47 | unsigned long hpte_v[5]; | ||
48 | unsigned long rmask = 0; | 47 | unsigned long rmask = 0; |
48 | u64 hpte_v[5]; | ||
49 | 49 | ||
50 | beat_read_htab_entries(0, hpte_group + 0, hpte_v); | 50 | beat_read_htab_entries(0, hpte_group + 0, hpte_v); |
51 | if (!(hpte_v[0] & HPTE_V_BOLTED)) | 51 | if (!(hpte_v[0] & HPTE_V_BOLTED)) |
@@ -93,8 +93,7 @@ static long beat_lpar_hpte_insert(unsigned long hpte_group, | |||
93 | int psize, int ssize) | 93 | int psize, int ssize) |
94 | { | 94 | { |
95 | unsigned long lpar_rc; | 95 | unsigned long lpar_rc; |
96 | unsigned long slot; | 96 | u64 hpte_v, hpte_r, slot; |
97 | unsigned long hpte_v, hpte_r; | ||
98 | 97 | ||
99 | /* same as iseries */ | 98 | /* same as iseries */ |
100 | if (vflags & HPTE_V_SECONDARY) | 99 | if (vflags & HPTE_V_SECONDARY) |
@@ -153,8 +152,9 @@ static long beat_lpar_hpte_remove(unsigned long hpte_group) | |||
153 | 152 | ||
154 | static unsigned long beat_lpar_hpte_getword0(unsigned long slot) | 153 | static unsigned long beat_lpar_hpte_getword0(unsigned long slot) |
155 | { | 154 | { |
156 | unsigned long dword0, dword[5]; | 155 | unsigned long dword0; |
157 | unsigned long lpar_rc; | 156 | unsigned long lpar_rc; |
157 | u64 dword[5]; | ||
158 | 158 | ||
159 | lpar_rc = beat_read_htab_entries(0, slot & ~3UL, dword); | 159 | lpar_rc = beat_read_htab_entries(0, slot & ~3UL, dword); |
160 | 160 | ||
@@ -170,7 +170,7 @@ static void beat_lpar_hptab_clear(void) | |||
170 | unsigned long size_bytes = 1UL << ppc64_pft_size; | 170 | unsigned long size_bytes = 1UL << ppc64_pft_size; |
171 | unsigned long hpte_count = size_bytes >> 4; | 171 | unsigned long hpte_count = size_bytes >> 4; |
172 | int i; | 172 | int i; |
173 | unsigned long dummy0, dummy1; | 173 | u64 dummy0, dummy1; |
174 | 174 | ||
175 | /* TODO: Use bulk call */ | 175 | /* TODO: Use bulk call */ |
176 | for (i = 0; i < hpte_count; i++) | 176 | for (i = 0; i < hpte_count; i++) |
@@ -189,7 +189,8 @@ static long beat_lpar_hpte_updatepp(unsigned long slot, | |||
189 | int psize, int ssize, int local) | 189 | int psize, int ssize, int local) |
190 | { | 190 | { |
191 | unsigned long lpar_rc; | 191 | unsigned long lpar_rc; |
192 | unsigned long dummy0, dummy1, want_v; | 192 | u64 dummy0, dummy1; |
193 | unsigned long want_v; | ||
193 | 194 | ||
194 | want_v = hpte_encode_v(va, psize, MMU_SEGSIZE_256M); | 195 | want_v = hpte_encode_v(va, psize, MMU_SEGSIZE_256M); |
195 | 196 | ||
@@ -255,7 +256,8 @@ static void beat_lpar_hpte_updateboltedpp(unsigned long newpp, | |||
255 | unsigned long ea, | 256 | unsigned long ea, |
256 | int psize, int ssize) | 257 | int psize, int ssize) |
257 | { | 258 | { |
258 | unsigned long lpar_rc, slot, vsid, va, dummy0, dummy1; | 259 | unsigned long lpar_rc, slot, vsid, va; |
260 | u64 dummy0, dummy1; | ||
259 | 261 | ||
260 | vsid = get_kernel_vsid(ea, MMU_SEGSIZE_256M); | 262 | vsid = get_kernel_vsid(ea, MMU_SEGSIZE_256M); |
261 | va = (vsid << 28) | (ea & 0x0fffffff); | 263 | va = (vsid << 28) | (ea & 0x0fffffff); |
@@ -276,7 +278,7 @@ static void beat_lpar_hpte_invalidate(unsigned long slot, unsigned long va, | |||
276 | { | 278 | { |
277 | unsigned long want_v; | 279 | unsigned long want_v; |
278 | unsigned long lpar_rc; | 280 | unsigned long lpar_rc; |
279 | unsigned long dummy1, dummy2; | 281 | u64 dummy1, dummy2; |
280 | unsigned long flags; | 282 | unsigned long flags; |
281 | 283 | ||
282 | DBG_LOW(" inval : slot=%lx, va=%016lx, psize: %d, local: %d\n", | 284 | DBG_LOW(" inval : slot=%lx, va=%016lx, psize: %d, local: %d\n", |
@@ -315,8 +317,7 @@ static long beat_lpar_hpte_insert_v3(unsigned long hpte_group, | |||
315 | int psize, int ssize) | 317 | int psize, int ssize) |
316 | { | 318 | { |
317 | unsigned long lpar_rc; | 319 | unsigned long lpar_rc; |
318 | unsigned long slot; | 320 | u64 hpte_v, hpte_r, slot; |
319 | unsigned long hpte_v, hpte_r; | ||
320 | 321 | ||
321 | /* same as iseries */ | 322 | /* same as iseries */ |
322 | if (vflags & HPTE_V_SECONDARY) | 323 | if (vflags & HPTE_V_SECONDARY) |
diff --git a/arch/powerpc/platforms/cell/beat_udbg.c b/arch/powerpc/platforms/cell/beat_udbg.c index 6b418f6b6175..350735bc8888 100644 --- a/arch/powerpc/platforms/cell/beat_udbg.c +++ b/arch/powerpc/platforms/cell/beat_udbg.c | |||
@@ -40,8 +40,8 @@ static void udbg_putc_beat(char c) | |||
40 | } | 40 | } |
41 | 41 | ||
42 | /* Buffered chars getc */ | 42 | /* Buffered chars getc */ |
43 | static long inbuflen; | 43 | static u64 inbuflen; |
44 | static long inbuf[2]; /* must be 2 longs */ | 44 | static u64 inbuf[2]; /* must be 2 u64s */ |
45 | 45 | ||
46 | static int udbg_getc_poll_beat(void) | 46 | static int udbg_getc_poll_beat(void) |
47 | { | 47 | { |
diff --git a/arch/powerpc/platforms/cell/cbe_cpufreq_pervasive.c b/arch/powerpc/platforms/cell/cbe_cpufreq_pervasive.c index 70fa7aef5edd..20472e487b6f 100644 --- a/arch/powerpc/platforms/cell/cbe_cpufreq_pervasive.c +++ b/arch/powerpc/platforms/cell/cbe_cpufreq_pervasive.c | |||
@@ -54,7 +54,7 @@ int cbe_cpufreq_set_pmode(int cpu, unsigned int pmode) | |||
54 | { | 54 | { |
55 | struct cbe_pmd_regs __iomem *pmd_regs; | 55 | struct cbe_pmd_regs __iomem *pmd_regs; |
56 | struct cbe_mic_tm_regs __iomem *mic_tm_regs; | 56 | struct cbe_mic_tm_regs __iomem *mic_tm_regs; |
57 | u64 flags; | 57 | unsigned long flags; |
58 | u64 value; | 58 | u64 value; |
59 | #ifdef DEBUG | 59 | #ifdef DEBUG |
60 | long time; | 60 | long time; |
diff --git a/arch/powerpc/platforms/cell/interrupt.c b/arch/powerpc/platforms/cell/interrupt.c index 2d5bb22d6c09..28c04dab2633 100644 --- a/arch/powerpc/platforms/cell/interrupt.c +++ b/arch/powerpc/platforms/cell/interrupt.c | |||
@@ -148,7 +148,7 @@ static unsigned int iic_get_irq(void) | |||
148 | 148 | ||
149 | iic = &__get_cpu_var(iic); | 149 | iic = &__get_cpu_var(iic); |
150 | *(unsigned long *) &pending = | 150 | *(unsigned long *) &pending = |
151 | in_be64((unsigned long __iomem *) &iic->regs->pending_destr); | 151 | in_be64((u64 __iomem *) &iic->regs->pending_destr); |
152 | if (!(pending.flags & CBE_IIC_IRQ_VALID)) | 152 | if (!(pending.flags & CBE_IIC_IRQ_VALID)) |
153 | return NO_IRQ; | 153 | return NO_IRQ; |
154 | virq = irq_linear_revmap(iic_host, iic_pending_to_hwnum(pending)); | 154 | virq = irq_linear_revmap(iic_host, iic_pending_to_hwnum(pending)); |
diff --git a/arch/powerpc/platforms/cell/io-workarounds.c b/arch/powerpc/platforms/cell/io-workarounds.c index b5f84e8f0899..059cad6c3f69 100644 --- a/arch/powerpc/platforms/cell/io-workarounds.c +++ b/arch/powerpc/platforms/cell/io-workarounds.c | |||
@@ -130,14 +130,14 @@ static const struct ppc_pci_io __devinitconst iowa_pci_io = { | |||
130 | 130 | ||
131 | }; | 131 | }; |
132 | 132 | ||
133 | static void __iomem *iowa_ioremap(unsigned long addr, unsigned long size, | 133 | static void __iomem *iowa_ioremap(phys_addr_t addr, unsigned long size, |
134 | unsigned long flags) | 134 | unsigned long flags) |
135 | { | 135 | { |
136 | struct iowa_bus *bus; | 136 | struct iowa_bus *bus; |
137 | void __iomem *res = __ioremap(addr, size, flags); | 137 | void __iomem *res = __ioremap(addr, size, flags); |
138 | int busno; | 138 | int busno; |
139 | 139 | ||
140 | bus = iowa_pci_find(0, addr); | 140 | bus = iowa_pci_find(0, (unsigned long)addr); |
141 | if (bus != NULL) { | 141 | if (bus != NULL) { |
142 | busno = bus - iowa_busses; | 142 | busno = bus - iowa_busses; |
143 | PCI_SET_ADDR_TOKEN(res, busno + 1); | 143 | PCI_SET_ADDR_TOKEN(res, busno + 1); |
diff --git a/arch/powerpc/platforms/cell/iommu.c b/arch/powerpc/platforms/cell/iommu.c index 86db4dd170a0..88d94b59a7cb 100644 --- a/arch/powerpc/platforms/cell/iommu.c +++ b/arch/powerpc/platforms/cell/iommu.c | |||
@@ -150,8 +150,8 @@ static int cbe_nr_iommus; | |||
150 | static void invalidate_tce_cache(struct cbe_iommu *iommu, unsigned long *pte, | 150 | static void invalidate_tce_cache(struct cbe_iommu *iommu, unsigned long *pte, |
151 | long n_ptes) | 151 | long n_ptes) |
152 | { | 152 | { |
153 | unsigned long __iomem *reg; | 153 | u64 __iomem *reg; |
154 | unsigned long val; | 154 | u64 val; |
155 | long n; | 155 | long n; |
156 | 156 | ||
157 | reg = iommu->xlate_regs + IOC_IOPT_CacheInvd; | 157 | reg = iommu->xlate_regs + IOC_IOPT_CacheInvd; |
diff --git a/arch/powerpc/platforms/cell/spufs/spufs.h b/arch/powerpc/platforms/cell/spufs/spufs.h index 15c62d3ca129..3bf908e2873a 100644 --- a/arch/powerpc/platforms/cell/spufs/spufs.h +++ b/arch/powerpc/platforms/cell/spufs/spufs.h | |||
@@ -314,7 +314,7 @@ extern char *isolated_loader; | |||
314 | * we need to call spu_release(ctx) before sleeping, and | 314 | * we need to call spu_release(ctx) before sleeping, and |
315 | * then spu_acquire(ctx) when awoken. | 315 | * then spu_acquire(ctx) when awoken. |
316 | * | 316 | * |
317 | * Returns with state_mutex re-acquired when successfull or | 317 | * Returns with state_mutex re-acquired when successful or |
318 | * with -ERESTARTSYS and the state_mutex dropped when interrupted. | 318 | * with -ERESTARTSYS and the state_mutex dropped when interrupted. |
319 | */ | 319 | */ |
320 | 320 | ||
diff --git a/arch/powerpc/platforms/iseries/Kconfig b/arch/powerpc/platforms/iseries/Kconfig index ed3753d8c109..7ddd0a2c8027 100644 --- a/arch/powerpc/platforms/iseries/Kconfig +++ b/arch/powerpc/platforms/iseries/Kconfig | |||
@@ -10,18 +10,21 @@ menu "iSeries device drivers" | |||
10 | config VIODASD | 10 | config VIODASD |
11 | tristate "iSeries Virtual I/O disk support" | 11 | tristate "iSeries Virtual I/O disk support" |
12 | depends on BLOCK | 12 | depends on BLOCK |
13 | select VIOPATH | ||
13 | help | 14 | help |
14 | If you are running on an iSeries system and you want to use | 15 | If you are running on an iSeries system and you want to use |
15 | virtual disks created and managed by OS/400, say Y. | 16 | virtual disks created and managed by OS/400, say Y. |
16 | 17 | ||
17 | config VIOCD | 18 | config VIOCD |
18 | tristate "iSeries Virtual I/O CD support" | 19 | tristate "iSeries Virtual I/O CD support" |
20 | select VIOPATH | ||
19 | help | 21 | help |
20 | If you are running Linux on an IBM iSeries system and you want to | 22 | If you are running Linux on an IBM iSeries system and you want to |
21 | read a CD drive owned by OS/400, say Y here. | 23 | read a CD drive owned by OS/400, say Y here. |
22 | 24 | ||
23 | config VIOTAPE | 25 | config VIOTAPE |
24 | tristate "iSeries Virtual Tape Support" | 26 | tristate "iSeries Virtual Tape Support" |
27 | select VIOPATH | ||
25 | help | 28 | help |
26 | If you are running Linux on an iSeries system and you want Linux | 29 | If you are running Linux on an iSeries system and you want Linux |
27 | to read and/or write a tape drive owned by OS/400, say Y here. | 30 | to read and/or write a tape drive owned by OS/400, say Y here. |
@@ -30,5 +33,3 @@ endmenu | |||
30 | 33 | ||
31 | config VIOPATH | 34 | config VIOPATH |
32 | bool | 35 | bool |
33 | depends on VIODASD || VIOCD || VIOTAPE || ISERIES_VETH | ||
34 | default y | ||
diff --git a/arch/powerpc/platforms/iseries/setup.c b/arch/powerpc/platforms/iseries/setup.c index 70b688c1aefb..24519b96d6ad 100644 --- a/arch/powerpc/platforms/iseries/setup.c +++ b/arch/powerpc/platforms/iseries/setup.c | |||
@@ -23,6 +23,7 @@ | |||
23 | #include <linux/string.h> | 23 | #include <linux/string.h> |
24 | #include <linux/seq_file.h> | 24 | #include <linux/seq_file.h> |
25 | #include <linux/kdev_t.h> | 25 | #include <linux/kdev_t.h> |
26 | #include <linux/kexec.h> | ||
26 | #include <linux/major.h> | 27 | #include <linux/major.h> |
27 | #include <linux/root_dev.h> | 28 | #include <linux/root_dev.h> |
28 | #include <linux/kernel.h> | 29 | #include <linux/kernel.h> |
@@ -638,6 +639,13 @@ static int __init iseries_probe(void) | |||
638 | return 1; | 639 | return 1; |
639 | } | 640 | } |
640 | 641 | ||
642 | #ifdef CONFIG_KEXEC | ||
643 | static int iseries_kexec_prepare(struct kimage *image) | ||
644 | { | ||
645 | return -ENOSYS; | ||
646 | } | ||
647 | #endif | ||
648 | |||
641 | define_machine(iseries) { | 649 | define_machine(iseries) { |
642 | .name = "iSeries", | 650 | .name = "iSeries", |
643 | .setup_arch = iSeries_setup_arch, | 651 | .setup_arch = iSeries_setup_arch, |
@@ -658,6 +666,9 @@ define_machine(iseries) { | |||
658 | .probe = iseries_probe, | 666 | .probe = iseries_probe, |
659 | .ioremap = iseries_ioremap, | 667 | .ioremap = iseries_ioremap, |
660 | .iounmap = iseries_iounmap, | 668 | .iounmap = iseries_iounmap, |
669 | #ifdef CONFIG_KEXEC | ||
670 | .machine_kexec_prepare = iseries_kexec_prepare, | ||
671 | #endif | ||
661 | /* XXX Implement enable_pmcs for iSeries */ | 672 | /* XXX Implement enable_pmcs for iSeries */ |
662 | }; | 673 | }; |
663 | 674 | ||
diff --git a/arch/powerpc/platforms/pasemi/cpufreq.c b/arch/powerpc/platforms/pasemi/cpufreq.c index 58556b028a4c..86db47c1b665 100644 --- a/arch/powerpc/platforms/pasemi/cpufreq.c +++ b/arch/powerpc/platforms/pasemi/cpufreq.c | |||
@@ -112,7 +112,7 @@ static int get_gizmo_latency(void) | |||
112 | 112 | ||
113 | static void set_astate(int cpu, unsigned int astate) | 113 | static void set_astate(int cpu, unsigned int astate) |
114 | { | 114 | { |
115 | u64 flags; | 115 | unsigned long flags; |
116 | 116 | ||
117 | /* Return if called before init has run */ | 117 | /* Return if called before init has run */ |
118 | if (unlikely(!sdcasr_mapbase)) | 118 | if (unlikely(!sdcasr_mapbase)) |
diff --git a/arch/powerpc/platforms/pasemi/dma_lib.c b/arch/powerpc/platforms/pasemi/dma_lib.c index 217af321b0ca..a6152d922243 100644 --- a/arch/powerpc/platforms/pasemi/dma_lib.c +++ b/arch/powerpc/platforms/pasemi/dma_lib.c | |||
@@ -509,7 +509,7 @@ fallback: | |||
509 | */ | 509 | */ |
510 | int pasemi_dma_init(void) | 510 | int pasemi_dma_init(void) |
511 | { | 511 | { |
512 | static spinlock_t init_lock = SPIN_LOCK_UNLOCKED; | 512 | static DEFINE_SPINLOCK(init_lock); |
513 | struct pci_dev *iob_pdev; | 513 | struct pci_dev *iob_pdev; |
514 | struct pci_dev *pdev; | 514 | struct pci_dev *pdev; |
515 | struct resource res; | 515 | struct resource res; |
diff --git a/arch/powerpc/platforms/powermac/pci.c b/arch/powerpc/platforms/powermac/pci.c index 54b7b76ed4f0..04cdd32624d4 100644 --- a/arch/powerpc/platforms/powermac/pci.c +++ b/arch/powerpc/platforms/powermac/pci.c | |||
@@ -661,6 +661,7 @@ static void __init init_second_ohare(void) | |||
661 | pci_find_hose_for_OF_device(np); | 661 | pci_find_hose_for_OF_device(np); |
662 | if (!hose) { | 662 | if (!hose) { |
663 | printk(KERN_ERR "Can't find PCI hose for OHare2 !\n"); | 663 | printk(KERN_ERR "Can't find PCI hose for OHare2 !\n"); |
664 | of_node_put(np); | ||
664 | return; | 665 | return; |
665 | } | 666 | } |
666 | early_read_config_word(hose, bus, devfn, PCI_COMMAND, &cmd); | 667 | early_read_config_word(hose, bus, devfn, PCI_COMMAND, &cmd); |
@@ -669,6 +670,7 @@ static void __init init_second_ohare(void) | |||
669 | early_write_config_word(hose, bus, devfn, PCI_COMMAND, cmd); | 670 | early_write_config_word(hose, bus, devfn, PCI_COMMAND, cmd); |
670 | } | 671 | } |
671 | has_second_ohare = 1; | 672 | has_second_ohare = 1; |
673 | of_node_put(np); | ||
672 | } | 674 | } |
673 | 675 | ||
674 | /* | 676 | /* |
diff --git a/arch/powerpc/platforms/powermac/time.c b/arch/powerpc/platforms/powermac/time.c index 59eb840d8ce2..1810e4226e56 100644 --- a/arch/powerpc/platforms/powermac/time.c +++ b/arch/powerpc/platforms/powermac/time.c | |||
@@ -265,12 +265,15 @@ int __init via_calibrate_decr(void) | |||
265 | struct resource rsrc; | 265 | struct resource rsrc; |
266 | 266 | ||
267 | vias = of_find_node_by_name(NULL, "via-cuda"); | 267 | vias = of_find_node_by_name(NULL, "via-cuda"); |
268 | if (vias == 0) | 268 | if (vias == NULL) |
269 | vias = of_find_node_by_name(NULL, "via-pmu"); | 269 | vias = of_find_node_by_name(NULL, "via-pmu"); |
270 | if (vias == 0) | 270 | if (vias == NULL) |
271 | vias = of_find_node_by_name(NULL, "via"); | 271 | vias = of_find_node_by_name(NULL, "via"); |
272 | if (vias == 0 || of_address_to_resource(vias, 0, &rsrc)) | 272 | if (vias == NULL || of_address_to_resource(vias, 0, &rsrc)) { |
273 | of_node_put(vias); | ||
273 | return 0; | 274 | return 0; |
275 | } | ||
276 | of_node_put(vias); | ||
274 | via = ioremap(rsrc.start, rsrc.end - rsrc.start + 1); | 277 | via = ioremap(rsrc.start, rsrc.end - rsrc.start + 1); |
275 | if (via == NULL) { | 278 | if (via == NULL) { |
276 | printk(KERN_ERR "Failed to map VIA for timer calibration !\n"); | 279 | printk(KERN_ERR "Failed to map VIA for timer calibration !\n"); |
@@ -297,7 +300,7 @@ int __init via_calibrate_decr(void) | |||
297 | ppc_tb_freq = (dstart - dend) * 100 / 6; | 300 | ppc_tb_freq = (dstart - dend) * 100 / 6; |
298 | 301 | ||
299 | iounmap(via); | 302 | iounmap(via); |
300 | 303 | ||
301 | return 1; | 304 | return 1; |
302 | } | 305 | } |
303 | #endif | 306 | #endif |
diff --git a/arch/powerpc/platforms/ps3/device-init.c b/arch/powerpc/platforms/ps3/device-init.c index dbc124e05646..ca71a12b764c 100644 --- a/arch/powerpc/platforms/ps3/device-init.c +++ b/arch/powerpc/platforms/ps3/device-init.c | |||
@@ -518,6 +518,41 @@ fail_device_register: | |||
518 | return result; | 518 | return result; |
519 | } | 519 | } |
520 | 520 | ||
521 | static int __init ps3_register_ramdisk_device(void) | ||
522 | { | ||
523 | int result; | ||
524 | struct layout { | ||
525 | struct ps3_system_bus_device dev; | ||
526 | } *p; | ||
527 | |||
528 | pr_debug(" -> %s:%d\n", __func__, __LINE__); | ||
529 | |||
530 | p = kzalloc(sizeof(struct layout), GFP_KERNEL); | ||
531 | |||
532 | if (!p) | ||
533 | return -ENOMEM; | ||
534 | |||
535 | p->dev.match_id = PS3_MATCH_ID_GPU; | ||
536 | p->dev.match_sub_id = PS3_MATCH_SUB_ID_GPU_RAMDISK; | ||
537 | p->dev.dev_type = PS3_DEVICE_TYPE_IOC0; | ||
538 | |||
539 | result = ps3_system_bus_device_register(&p->dev); | ||
540 | |||
541 | if (result) { | ||
542 | pr_debug("%s:%d ps3_system_bus_device_register failed\n", | ||
543 | __func__, __LINE__); | ||
544 | goto fail_device_register; | ||
545 | } | ||
546 | |||
547 | pr_debug(" <- %s:%d\n", __func__, __LINE__); | ||
548 | return 0; | ||
549 | |||
550 | fail_device_register: | ||
551 | kfree(p); | ||
552 | pr_debug(" <- %s:%d failed\n", __func__, __LINE__); | ||
553 | return result; | ||
554 | } | ||
555 | |||
521 | /** | 556 | /** |
522 | * ps3_setup_dynamic_device - Setup a dynamic device from the repository | 557 | * ps3_setup_dynamic_device - Setup a dynamic device from the repository |
523 | */ | 558 | */ |
@@ -946,6 +981,8 @@ static int __init ps3_register_devices(void) | |||
946 | 981 | ||
947 | ps3_register_lpm_devices(); | 982 | ps3_register_lpm_devices(); |
948 | 983 | ||
984 | ps3_register_ramdisk_device(); | ||
985 | |||
949 | pr_debug(" <- %s:%d\n", __func__, __LINE__); | 986 | pr_debug(" <- %s:%d\n", __func__, __LINE__); |
950 | return 0; | 987 | return 0; |
951 | } | 988 | } |
diff --git a/arch/powerpc/sysdev/Makefile b/arch/powerpc/sysdev/Makefile index 5afce115ab1f..b33b28a6fe12 100644 --- a/arch/powerpc/sysdev/Makefile +++ b/arch/powerpc/sysdev/Makefile | |||
@@ -17,6 +17,7 @@ obj-$(CONFIG_FSL_PCI) += fsl_pci.o $(fsl-msi-obj-y) | |||
17 | obj-$(CONFIG_FSL_LBC) += fsl_lbc.o | 17 | obj-$(CONFIG_FSL_LBC) += fsl_lbc.o |
18 | obj-$(CONFIG_FSL_GTM) += fsl_gtm.o | 18 | obj-$(CONFIG_FSL_GTM) += fsl_gtm.o |
19 | obj-$(CONFIG_MPC8xxx_GPIO) += mpc8xxx_gpio.o | 19 | obj-$(CONFIG_MPC8xxx_GPIO) += mpc8xxx_gpio.o |
20 | obj-$(CONFIG_SIMPLE_GPIO) += simple_gpio.o | ||
20 | obj-$(CONFIG_RAPIDIO) += fsl_rio.o | 21 | obj-$(CONFIG_RAPIDIO) += fsl_rio.o |
21 | obj-$(CONFIG_TSI108_BRIDGE) += tsi108_pci.o tsi108_dev.o | 22 | obj-$(CONFIG_TSI108_BRIDGE) += tsi108_pci.o tsi108_dev.o |
22 | obj-$(CONFIG_QUICC_ENGINE) += qe_lib/ | 23 | obj-$(CONFIG_QUICC_ENGINE) += qe_lib/ |
diff --git a/arch/powerpc/sysdev/fsl_pci.c b/arch/powerpc/sysdev/fsl_pci.c index d5f9ae0f1b75..f611d0369cc8 100644 --- a/arch/powerpc/sysdev/fsl_pci.c +++ b/arch/powerpc/sysdev/fsl_pci.c | |||
@@ -29,7 +29,8 @@ | |||
29 | 29 | ||
30 | #if defined(CONFIG_PPC_85xx) || defined(CONFIG_PPC_86xx) | 30 | #if defined(CONFIG_PPC_85xx) || defined(CONFIG_PPC_86xx) |
31 | /* atmu setup for fsl pci/pcie controller */ | 31 | /* atmu setup for fsl pci/pcie controller */ |
32 | void __init setup_pci_atmu(struct pci_controller *hose, struct resource *rsrc) | 32 | static void __init setup_pci_atmu(struct pci_controller *hose, |
33 | struct resource *rsrc) | ||
33 | { | 34 | { |
34 | struct ccsr_pci __iomem *pci; | 35 | struct ccsr_pci __iomem *pci; |
35 | int i; | 36 | int i; |
@@ -86,7 +87,7 @@ void __init setup_pci_atmu(struct pci_controller *hose, struct resource *rsrc) | |||
86 | out_be32(&pci->piw[2].piwar, PIWAR_2G); | 87 | out_be32(&pci->piw[2].piwar, PIWAR_2G); |
87 | } | 88 | } |
88 | 89 | ||
89 | void __init setup_pci_cmd(struct pci_controller *hose) | 90 | static void __init setup_pci_cmd(struct pci_controller *hose) |
90 | { | 91 | { |
91 | u16 cmd; | 92 | u16 cmd; |
92 | int cap_x; | 93 | int cap_x; |
@@ -130,7 +131,7 @@ static void __init quirk_fsl_pcie_header(struct pci_dev *dev) | |||
130 | return ; | 131 | return ; |
131 | } | 132 | } |
132 | 133 | ||
133 | int __init fsl_pcie_check_link(struct pci_controller *hose) | 134 | static int __init fsl_pcie_check_link(struct pci_controller *hose) |
134 | { | 135 | { |
135 | u32 val; | 136 | u32 val; |
136 | early_read_config_dword(hose, 0, 0, PCIE_LTSSM, &val); | 137 | early_read_config_dword(hose, 0, 0, PCIE_LTSSM, &val); |
diff --git a/arch/powerpc/sysdev/fsl_soc.h b/arch/powerpc/sysdev/fsl_soc.h index 60f7f227327c..9c744e4285a0 100644 --- a/arch/powerpc/sysdev/fsl_soc.h +++ b/arch/powerpc/sysdev/fsl_soc.h | |||
@@ -5,8 +5,13 @@ | |||
5 | #include <asm/mmu.h> | 5 | #include <asm/mmu.h> |
6 | 6 | ||
7 | extern phys_addr_t get_immrbase(void); | 7 | extern phys_addr_t get_immrbase(void); |
8 | #if defined(CONFIG_CPM2) || defined(CONFIG_QUICC_ENGINE) || defined(CONFIG_8xx) | ||
8 | extern u32 get_brgfreq(void); | 9 | extern u32 get_brgfreq(void); |
9 | extern u32 get_baudrate(void); | 10 | extern u32 get_baudrate(void); |
11 | #else | ||
12 | static inline u32 get_brgfreq(void) { return -1; } | ||
13 | static inline u32 get_baudrate(void) { return -1; } | ||
14 | #endif | ||
10 | extern u32 fsl_get_sys_freq(void); | 15 | extern u32 fsl_get_sys_freq(void); |
11 | 16 | ||
12 | struct spi_board_info; | 17 | struct spi_board_info; |
diff --git a/arch/powerpc/sysdev/qe_lib/Kconfig b/arch/powerpc/sysdev/qe_lib/Kconfig index 76ffbc48d4b9..41ac3dfac98e 100644 --- a/arch/powerpc/sysdev/qe_lib/Kconfig +++ b/arch/powerpc/sysdev/qe_lib/Kconfig | |||
@@ -22,5 +22,6 @@ config UCC | |||
22 | 22 | ||
23 | config QE_USB | 23 | config QE_USB |
24 | bool | 24 | bool |
25 | default y if USB_GADGET_FSL_QE | ||
25 | help | 26 | help |
26 | QE USB Host Controller support | 27 | QE USB Controller support |
diff --git a/arch/powerpc/sysdev/qe_lib/gpio.c b/arch/powerpc/sysdev/qe_lib/gpio.c index 8e5a0bc36d0b..3485288dce31 100644 --- a/arch/powerpc/sysdev/qe_lib/gpio.c +++ b/arch/powerpc/sysdev/qe_lib/gpio.c | |||
@@ -14,6 +14,7 @@ | |||
14 | #include <linux/kernel.h> | 14 | #include <linux/kernel.h> |
15 | #include <linux/init.h> | 15 | #include <linux/init.h> |
16 | #include <linux/spinlock.h> | 16 | #include <linux/spinlock.h> |
17 | #include <linux/err.h> | ||
17 | #include <linux/io.h> | 18 | #include <linux/io.h> |
18 | #include <linux/of.h> | 19 | #include <linux/of.h> |
19 | #include <linux/of_gpio.h> | 20 | #include <linux/of_gpio.h> |
@@ -24,8 +25,14 @@ struct qe_gpio_chip { | |||
24 | struct of_mm_gpio_chip mm_gc; | 25 | struct of_mm_gpio_chip mm_gc; |
25 | spinlock_t lock; | 26 | spinlock_t lock; |
26 | 27 | ||
28 | unsigned long pin_flags[QE_PIO_PINS]; | ||
29 | #define QE_PIN_REQUESTED 0 | ||
30 | |||
27 | /* shadowed data register to clear/set bits safely */ | 31 | /* shadowed data register to clear/set bits safely */ |
28 | u32 cpdata; | 32 | u32 cpdata; |
33 | |||
34 | /* saved_regs used to restore dedicated functions */ | ||
35 | struct qe_pio_regs saved_regs; | ||
29 | }; | 36 | }; |
30 | 37 | ||
31 | static inline struct qe_gpio_chip * | 38 | static inline struct qe_gpio_chip * |
@@ -40,6 +47,12 @@ static void qe_gpio_save_regs(struct of_mm_gpio_chip *mm_gc) | |||
40 | struct qe_pio_regs __iomem *regs = mm_gc->regs; | 47 | struct qe_pio_regs __iomem *regs = mm_gc->regs; |
41 | 48 | ||
42 | qe_gc->cpdata = in_be32(®s->cpdata); | 49 | qe_gc->cpdata = in_be32(®s->cpdata); |
50 | qe_gc->saved_regs.cpdata = qe_gc->cpdata; | ||
51 | qe_gc->saved_regs.cpdir1 = in_be32(®s->cpdir1); | ||
52 | qe_gc->saved_regs.cpdir2 = in_be32(®s->cpdir2); | ||
53 | qe_gc->saved_regs.cppar1 = in_be32(®s->cppar1); | ||
54 | qe_gc->saved_regs.cppar2 = in_be32(®s->cppar2); | ||
55 | qe_gc->saved_regs.cpodr = in_be32(®s->cpodr); | ||
43 | } | 56 | } |
44 | 57 | ||
45 | static int qe_gpio_get(struct gpio_chip *gc, unsigned int gpio) | 58 | static int qe_gpio_get(struct gpio_chip *gc, unsigned int gpio) |
@@ -103,6 +116,188 @@ static int qe_gpio_dir_out(struct gpio_chip *gc, unsigned int gpio, int val) | |||
103 | return 0; | 116 | return 0; |
104 | } | 117 | } |
105 | 118 | ||
119 | struct qe_pin { | ||
120 | /* | ||
121 | * The qe_gpio_chip name is unfortunate, we should change that to | ||
122 | * something like qe_pio_controller. Someday. | ||
123 | */ | ||
124 | struct qe_gpio_chip *controller; | ||
125 | int num; | ||
126 | }; | ||
127 | |||
128 | /** | ||
129 | * qe_pin_request - Request a QE pin | ||
130 | * @np: device node to get a pin from | ||
131 | * @index: index of a pin in the device tree | ||
132 | * Context: non-atomic | ||
133 | * | ||
134 | * This function return qe_pin so that you could use it with the rest of | ||
135 | * the QE Pin Multiplexing API. | ||
136 | */ | ||
137 | struct qe_pin *qe_pin_request(struct device_node *np, int index) | ||
138 | { | ||
139 | struct qe_pin *qe_pin; | ||
140 | struct device_node *gc; | ||
141 | struct of_gpio_chip *of_gc = NULL; | ||
142 | struct of_mm_gpio_chip *mm_gc; | ||
143 | struct qe_gpio_chip *qe_gc; | ||
144 | int err; | ||
145 | int size; | ||
146 | const void *gpio_spec; | ||
147 | const u32 *gpio_cells; | ||
148 | unsigned long flags; | ||
149 | |||
150 | qe_pin = kzalloc(sizeof(*qe_pin), GFP_KERNEL); | ||
151 | if (!qe_pin) { | ||
152 | pr_debug("%s: can't allocate memory\n", __func__); | ||
153 | return ERR_PTR(-ENOMEM); | ||
154 | } | ||
155 | |||
156 | err = of_parse_phandles_with_args(np, "gpios", "#gpio-cells", index, | ||
157 | &gc, &gpio_spec); | ||
158 | if (err) { | ||
159 | pr_debug("%s: can't parse gpios property\n", __func__); | ||
160 | goto err0; | ||
161 | } | ||
162 | |||
163 | if (!of_device_is_compatible(gc, "fsl,mpc8323-qe-pario-bank")) { | ||
164 | pr_debug("%s: tried to get a non-qe pin\n", __func__); | ||
165 | err = -EINVAL; | ||
166 | goto err1; | ||
167 | } | ||
168 | |||
169 | of_gc = gc->data; | ||
170 | if (!of_gc) { | ||
171 | pr_debug("%s: gpio controller %s isn't registered\n", | ||
172 | np->full_name, gc->full_name); | ||
173 | err = -ENODEV; | ||
174 | goto err1; | ||
175 | } | ||
176 | |||
177 | gpio_cells = of_get_property(gc, "#gpio-cells", &size); | ||
178 | if (!gpio_cells || size != sizeof(*gpio_cells) || | ||
179 | *gpio_cells != of_gc->gpio_cells) { | ||
180 | pr_debug("%s: wrong #gpio-cells for %s\n", | ||
181 | np->full_name, gc->full_name); | ||
182 | err = -EINVAL; | ||
183 | goto err1; | ||
184 | } | ||
185 | |||
186 | err = of_gc->xlate(of_gc, np, gpio_spec, NULL); | ||
187 | if (err < 0) | ||
188 | goto err1; | ||
189 | |||
190 | mm_gc = to_of_mm_gpio_chip(&of_gc->gc); | ||
191 | qe_gc = to_qe_gpio_chip(mm_gc); | ||
192 | |||
193 | spin_lock_irqsave(&qe_gc->lock, flags); | ||
194 | |||
195 | if (test_and_set_bit(QE_PIN_REQUESTED, &qe_gc->pin_flags[err]) == 0) { | ||
196 | qe_pin->controller = qe_gc; | ||
197 | qe_pin->num = err; | ||
198 | err = 0; | ||
199 | } else { | ||
200 | err = -EBUSY; | ||
201 | } | ||
202 | |||
203 | spin_unlock_irqrestore(&qe_gc->lock, flags); | ||
204 | |||
205 | if (!err) | ||
206 | return qe_pin; | ||
207 | err1: | ||
208 | of_node_put(gc); | ||
209 | err0: | ||
210 | kfree(qe_pin); | ||
211 | pr_debug("%s failed with status %d\n", __func__, err); | ||
212 | return ERR_PTR(err); | ||
213 | } | ||
214 | EXPORT_SYMBOL(qe_pin_request); | ||
215 | |||
216 | /** | ||
217 | * qe_pin_free - Free a pin | ||
218 | * @qe_pin: pointer to the qe_pin structure | ||
219 | * Context: any | ||
220 | * | ||
221 | * This function frees the qe_pin structure and makes a pin available | ||
222 | * for further qe_pin_request() calls. | ||
223 | */ | ||
224 | void qe_pin_free(struct qe_pin *qe_pin) | ||
225 | { | ||
226 | struct qe_gpio_chip *qe_gc = qe_pin->controller; | ||
227 | unsigned long flags; | ||
228 | const int pin = qe_pin->num; | ||
229 | |||
230 | spin_lock_irqsave(&qe_gc->lock, flags); | ||
231 | test_and_clear_bit(QE_PIN_REQUESTED, &qe_gc->pin_flags[pin]); | ||
232 | spin_unlock_irqrestore(&qe_gc->lock, flags); | ||
233 | |||
234 | kfree(qe_pin); | ||
235 | } | ||
236 | EXPORT_SYMBOL(qe_pin_free); | ||
237 | |||
238 | /** | ||
239 | * qe_pin_set_dedicated - Revert a pin to a dedicated peripheral function mode | ||
240 | * @qe_pin: pointer to the qe_pin structure | ||
241 | * Context: any | ||
242 | * | ||
243 | * This function resets a pin to a dedicated peripheral function that | ||
244 | * has been set up by the firmware. | ||
245 | */ | ||
246 | void qe_pin_set_dedicated(struct qe_pin *qe_pin) | ||
247 | { | ||
248 | struct qe_gpio_chip *qe_gc = qe_pin->controller; | ||
249 | struct qe_pio_regs __iomem *regs = qe_gc->mm_gc.regs; | ||
250 | struct qe_pio_regs *sregs = &qe_gc->saved_regs; | ||
251 | int pin = qe_pin->num; | ||
252 | u32 mask1 = 1 << (QE_PIO_PINS - (pin + 1)); | ||
253 | u32 mask2 = 0x3 << (QE_PIO_PINS - (pin % (QE_PIO_PINS / 2) + 1) * 2); | ||
254 | bool second_reg = pin > (QE_PIO_PINS / 2) - 1; | ||
255 | unsigned long flags; | ||
256 | |||
257 | spin_lock_irqsave(&qe_gc->lock, flags); | ||
258 | |||
259 | if (second_reg) { | ||
260 | clrsetbits_be32(®s->cpdir2, mask2, sregs->cpdir2 & mask2); | ||
261 | clrsetbits_be32(®s->cppar2, mask2, sregs->cppar2 & mask2); | ||
262 | } else { | ||
263 | clrsetbits_be32(®s->cpdir1, mask2, sregs->cpdir1 & mask2); | ||
264 | clrsetbits_be32(®s->cppar1, mask2, sregs->cppar1 & mask2); | ||
265 | } | ||
266 | |||
267 | if (sregs->cpdata & mask1) | ||
268 | qe_gc->cpdata |= mask1; | ||
269 | else | ||
270 | qe_gc->cpdata &= ~mask1; | ||
271 | |||
272 | out_be32(®s->cpdata, qe_gc->cpdata); | ||
273 | clrsetbits_be32(®s->cpodr, mask1, sregs->cpodr & mask1); | ||
274 | |||
275 | spin_unlock_irqrestore(&qe_gc->lock, flags); | ||
276 | } | ||
277 | EXPORT_SYMBOL(qe_pin_set_dedicated); | ||
278 | |||
279 | /** | ||
280 | * qe_pin_set_gpio - Set a pin to the GPIO mode | ||
281 | * @qe_pin: pointer to the qe_pin structure | ||
282 | * Context: any | ||
283 | * | ||
284 | * This function sets a pin to the GPIO mode. | ||
285 | */ | ||
286 | void qe_pin_set_gpio(struct qe_pin *qe_pin) | ||
287 | { | ||
288 | struct qe_gpio_chip *qe_gc = qe_pin->controller; | ||
289 | struct qe_pio_regs __iomem *regs = qe_gc->mm_gc.regs; | ||
290 | unsigned long flags; | ||
291 | |||
292 | spin_lock_irqsave(&qe_gc->lock, flags); | ||
293 | |||
294 | /* Let's make it input by default, GPIO API is able to change that. */ | ||
295 | __par_io_config_pin(regs, qe_pin->num, QE_PIO_DIR_IN, 0, 0, 0); | ||
296 | |||
297 | spin_unlock_irqrestore(&qe_gc->lock, flags); | ||
298 | } | ||
299 | EXPORT_SYMBOL(qe_pin_set_gpio); | ||
300 | |||
106 | static int __init qe_add_gpiochips(void) | 301 | static int __init qe_add_gpiochips(void) |
107 | { | 302 | { |
108 | struct device_node *np; | 303 | struct device_node *np; |
diff --git a/arch/powerpc/sysdev/simple_gpio.c b/arch/powerpc/sysdev/simple_gpio.c new file mode 100644 index 000000000000..43c4569e24b7 --- /dev/null +++ b/arch/powerpc/sysdev/simple_gpio.c | |||
@@ -0,0 +1,155 @@ | |||
1 | /* | ||
2 | * Simple Memory-Mapped GPIOs | ||
3 | * | ||
4 | * Copyright (c) MontaVista Software, Inc. 2008. | ||
5 | * | ||
6 | * Author: Anton Vorontsov <avorontsov@ru.mvista.com> | ||
7 | * | ||
8 | * This program is free software; you can redistribute it and/or modify it | ||
9 | * under the terms of the GNU General Public License as published by the | ||
10 | * Free Software Foundation; either version 2 of the License, or (at your | ||
11 | * option) any later version. | ||
12 | */ | ||
13 | |||
14 | #include <linux/init.h> | ||
15 | #include <linux/kernel.h> | ||
16 | #include <linux/module.h> | ||
17 | #include <linux/spinlock.h> | ||
18 | #include <linux/types.h> | ||
19 | #include <linux/ioport.h> | ||
20 | #include <linux/io.h> | ||
21 | #include <linux/of.h> | ||
22 | #include <linux/of_gpio.h> | ||
23 | #include <linux/gpio.h> | ||
24 | #include <asm/prom.h> | ||
25 | #include "simple_gpio.h" | ||
26 | |||
27 | struct u8_gpio_chip { | ||
28 | struct of_mm_gpio_chip mm_gc; | ||
29 | spinlock_t lock; | ||
30 | |||
31 | /* shadowed data register to clear/set bits safely */ | ||
32 | u8 data; | ||
33 | }; | ||
34 | |||
35 | static struct u8_gpio_chip *to_u8_gpio_chip(struct of_mm_gpio_chip *mm_gc) | ||
36 | { | ||
37 | return container_of(mm_gc, struct u8_gpio_chip, mm_gc); | ||
38 | } | ||
39 | |||
40 | static u8 u8_pin2mask(unsigned int pin) | ||
41 | { | ||
42 | return 1 << (8 - 1 - pin); | ||
43 | } | ||
44 | |||
45 | static int u8_gpio_get(struct gpio_chip *gc, unsigned int gpio) | ||
46 | { | ||
47 | struct of_mm_gpio_chip *mm_gc = to_of_mm_gpio_chip(gc); | ||
48 | |||
49 | return in_8(mm_gc->regs) & u8_pin2mask(gpio); | ||
50 | } | ||
51 | |||
52 | static void u8_gpio_set(struct gpio_chip *gc, unsigned int gpio, int val) | ||
53 | { | ||
54 | struct of_mm_gpio_chip *mm_gc = to_of_mm_gpio_chip(gc); | ||
55 | struct u8_gpio_chip *u8_gc = to_u8_gpio_chip(mm_gc); | ||
56 | unsigned long flags; | ||
57 | |||
58 | spin_lock_irqsave(&u8_gc->lock, flags); | ||
59 | |||
60 | if (val) | ||
61 | u8_gc->data |= u8_pin2mask(gpio); | ||
62 | else | ||
63 | u8_gc->data &= ~u8_pin2mask(gpio); | ||
64 | |||
65 | out_8(mm_gc->regs, u8_gc->data); | ||
66 | |||
67 | spin_unlock_irqrestore(&u8_gc->lock, flags); | ||
68 | } | ||
69 | |||
70 | static int u8_gpio_dir_in(struct gpio_chip *gc, unsigned int gpio) | ||
71 | { | ||
72 | return 0; | ||
73 | } | ||
74 | |||
75 | static int u8_gpio_dir_out(struct gpio_chip *gc, unsigned int gpio, int val) | ||
76 | { | ||
77 | u8_gpio_set(gc, gpio, val); | ||
78 | return 0; | ||
79 | } | ||
80 | |||
81 | static void u8_gpio_save_regs(struct of_mm_gpio_chip *mm_gc) | ||
82 | { | ||
83 | struct u8_gpio_chip *u8_gc = to_u8_gpio_chip(mm_gc); | ||
84 | |||
85 | u8_gc->data = in_8(mm_gc->regs); | ||
86 | } | ||
87 | |||
88 | static int __init u8_simple_gpiochip_add(struct device_node *np) | ||
89 | { | ||
90 | int ret; | ||
91 | struct u8_gpio_chip *u8_gc; | ||
92 | struct of_mm_gpio_chip *mm_gc; | ||
93 | struct of_gpio_chip *of_gc; | ||
94 | struct gpio_chip *gc; | ||
95 | |||
96 | u8_gc = kzalloc(sizeof(*u8_gc), GFP_KERNEL); | ||
97 | if (!u8_gc) | ||
98 | return -ENOMEM; | ||
99 | |||
100 | spin_lock_init(&u8_gc->lock); | ||
101 | |||
102 | mm_gc = &u8_gc->mm_gc; | ||
103 | of_gc = &mm_gc->of_gc; | ||
104 | gc = &of_gc->gc; | ||
105 | |||
106 | mm_gc->save_regs = u8_gpio_save_regs; | ||
107 | of_gc->gpio_cells = 2; | ||
108 | gc->ngpio = 8; | ||
109 | gc->direction_input = u8_gpio_dir_in; | ||
110 | gc->direction_output = u8_gpio_dir_out; | ||
111 | gc->get = u8_gpio_get; | ||
112 | gc->set = u8_gpio_set; | ||
113 | |||
114 | ret = of_mm_gpiochip_add(np, mm_gc); | ||
115 | if (ret) | ||
116 | goto err; | ||
117 | return 0; | ||
118 | err: | ||
119 | kfree(u8_gc); | ||
120 | return ret; | ||
121 | } | ||
122 | |||
123 | void __init simple_gpiochip_init(const char *compatible) | ||
124 | { | ||
125 | struct device_node *np; | ||
126 | |||
127 | for_each_compatible_node(np, NULL, compatible) { | ||
128 | int ret; | ||
129 | struct resource r; | ||
130 | |||
131 | ret = of_address_to_resource(np, 0, &r); | ||
132 | if (ret) | ||
133 | goto err; | ||
134 | |||
135 | switch (resource_size(&r)) { | ||
136 | case 1: | ||
137 | ret = u8_simple_gpiochip_add(np); | ||
138 | if (ret) | ||
139 | goto err; | ||
140 | break; | ||
141 | default: | ||
142 | /* | ||
143 | * Whenever you need support for GPIO bank width > 1, | ||
144 | * please just turn u8_ code into huge macros, and | ||
145 | * construct needed uX_ code with it. | ||
146 | */ | ||
147 | ret = -ENOSYS; | ||
148 | goto err; | ||
149 | } | ||
150 | continue; | ||
151 | err: | ||
152 | pr_err("%s: registration failed, status %d\n", | ||
153 | np->full_name, ret); | ||
154 | } | ||
155 | } | ||
diff --git a/arch/powerpc/sysdev/simple_gpio.h b/arch/powerpc/sysdev/simple_gpio.h new file mode 100644 index 000000000000..3a7b0c513c76 --- /dev/null +++ b/arch/powerpc/sysdev/simple_gpio.h | |||
@@ -0,0 +1,12 @@ | |||
1 | #ifndef __SYSDEV_SIMPLE_GPIO_H | ||
2 | #define __SYSDEV_SIMPLE_GPIO_H | ||
3 | |||
4 | #include <linux/errno.h> | ||
5 | |||
6 | #ifdef CONFIG_SIMPLE_GPIO | ||
7 | extern void simple_gpiochip_init(const char *compatible); | ||
8 | #else | ||
9 | static inline void simple_gpiochip_init(const char *compatible) {} | ||
10 | #endif /* CONFIG_SIMPLE_GPIO */ | ||
11 | |||
12 | #endif /* __SYSDEV_SIMPLE_GPIO_H */ | ||
diff --git a/arch/sparc/kernel/sun4m_smp.c b/arch/sparc/kernel/sun4m_smp.c index 4f8d60586b07..8040376c4890 100644 --- a/arch/sparc/kernel/sun4m_smp.c +++ b/arch/sparc/kernel/sun4m_smp.c | |||
@@ -54,7 +54,8 @@ extern int __smp4m_processor_id(void); | |||
54 | #define SMP_PRINTK(x) | 54 | #define SMP_PRINTK(x) |
55 | #endif | 55 | #endif |
56 | 56 | ||
57 | static inline unsigned long swap(volatile unsigned long *ptr, unsigned long val) | 57 | static inline unsigned long |
58 | swap_ulong(volatile unsigned long *ptr, unsigned long val) | ||
58 | { | 59 | { |
59 | __asm__ __volatile__("swap [%1], %0\n\t" : | 60 | __asm__ __volatile__("swap [%1], %0\n\t" : |
60 | "=&r" (val), "=&r" (ptr) : | 61 | "=&r" (val), "=&r" (ptr) : |
@@ -90,7 +91,7 @@ void __cpuinit smp4m_callin(void) | |||
90 | * to call the scheduler code. | 91 | * to call the scheduler code. |
91 | */ | 92 | */ |
92 | /* Allow master to continue. */ | 93 | /* Allow master to continue. */ |
93 | swap(&cpu_callin_map[cpuid], 1); | 94 | swap_ulong(&cpu_callin_map[cpuid], 1); |
94 | 95 | ||
95 | /* XXX: What's up with all the flushes? */ | 96 | /* XXX: What's up with all the flushes? */ |
96 | local_flush_cache_all(); | 97 | local_flush_cache_all(); |
diff --git a/arch/x86/kernel/cpu/cpufreq/longhaul.c b/arch/x86/kernel/cpu/cpufreq/longhaul.c index b0461856acfb..a4cff5d6e380 100644 --- a/arch/x86/kernel/cpu/cpufreq/longhaul.c +++ b/arch/x86/kernel/cpu/cpufreq/longhaul.c | |||
@@ -982,7 +982,7 @@ static int __init longhaul_init(void) | |||
982 | case 10: | 982 | case 10: |
983 | printk(KERN_ERR PFX "Use acpi-cpufreq driver for VIA C7\n"); | 983 | printk(KERN_ERR PFX "Use acpi-cpufreq driver for VIA C7\n"); |
984 | default: | 984 | default: |
985 | ;; | 985 | ; |
986 | } | 986 | } |
987 | 987 | ||
988 | return -ENODEV; | 988 | return -ENODEV; |
diff --git a/drivers/amba/bus.c b/drivers/amba/bus.c index 6b94fb7be5f2..00c46e0b40e4 100644 --- a/drivers/amba/bus.c +++ b/drivers/amba/bus.c | |||
@@ -12,9 +12,10 @@ | |||
12 | #include <linux/device.h> | 12 | #include <linux/device.h> |
13 | #include <linux/string.h> | 13 | #include <linux/string.h> |
14 | #include <linux/slab.h> | 14 | #include <linux/slab.h> |
15 | #include <linux/io.h> | ||
15 | #include <linux/amba/bus.h> | 16 | #include <linux/amba/bus.h> |
16 | 17 | ||
17 | #include <asm/io.h> | 18 | #include <asm/irq.h> |
18 | #include <asm/sizes.h> | 19 | #include <asm/sizes.h> |
19 | 20 | ||
20 | #define to_amba_device(d) container_of(d, struct amba_device, dev) | 21 | #define to_amba_device(d) container_of(d, struct amba_device, dev) |
diff --git a/drivers/ata/ahci.c b/drivers/ata/ahci.c index 656448c7fef9..7f701cbe14ab 100644 --- a/drivers/ata/ahci.c +++ b/drivers/ata/ahci.c | |||
@@ -105,7 +105,7 @@ enum { | |||
105 | board_ahci_ign_iferr = 2, | 105 | board_ahci_ign_iferr = 2, |
106 | board_ahci_sb600 = 3, | 106 | board_ahci_sb600 = 3, |
107 | board_ahci_mv = 4, | 107 | board_ahci_mv = 4, |
108 | board_ahci_sb700 = 5, | 108 | board_ahci_sb700 = 5, /* for SB700 and SB800 */ |
109 | board_ahci_mcp65 = 6, | 109 | board_ahci_mcp65 = 6, |
110 | board_ahci_nopmp = 7, | 110 | board_ahci_nopmp = 7, |
111 | 111 | ||
@@ -439,7 +439,7 @@ static const struct ata_port_info ahci_port_info[] = { | |||
439 | .udma_mask = ATA_UDMA6, | 439 | .udma_mask = ATA_UDMA6, |
440 | .port_ops = &ahci_ops, | 440 | .port_ops = &ahci_ops, |
441 | }, | 441 | }, |
442 | /* board_ahci_sb700 */ | 442 | /* board_ahci_sb700, for SB700 and SB800 */ |
443 | { | 443 | { |
444 | AHCI_HFLAGS (AHCI_HFLAG_IGN_SERR_INTERNAL), | 444 | AHCI_HFLAGS (AHCI_HFLAG_IGN_SERR_INTERNAL), |
445 | .flags = AHCI_FLAG_COMMON, | 445 | .flags = AHCI_FLAG_COMMON, |
@@ -2446,6 +2446,8 @@ static void ahci_print_info(struct ata_host *host) | |||
2446 | speed_s = "1.5"; | 2446 | speed_s = "1.5"; |
2447 | else if (speed == 2) | 2447 | else if (speed == 2) |
2448 | speed_s = "3"; | 2448 | speed_s = "3"; |
2449 | else if (speed == 3) | ||
2450 | speed_s = "6"; | ||
2449 | else | 2451 | else |
2450 | speed_s = "?"; | 2452 | speed_s = "?"; |
2451 | 2453 | ||
@@ -2610,6 +2612,10 @@ static int ahci_init_one(struct pci_dev *pdev, const struct pci_device_id *ent) | |||
2610 | (pdev->revision == 0xa1 || pdev->revision == 0xa2)) | 2612 | (pdev->revision == 0xa1 || pdev->revision == 0xa2)) |
2611 | hpriv->flags |= AHCI_HFLAG_NO_MSI; | 2613 | hpriv->flags |= AHCI_HFLAG_NO_MSI; |
2612 | 2614 | ||
2615 | /* SB800 does NOT need the workaround to ignore SERR_INTERNAL */ | ||
2616 | if (board_id == board_ahci_sb700 && pdev->revision >= 0x40) | ||
2617 | hpriv->flags &= ~AHCI_HFLAG_IGN_SERR_INTERNAL; | ||
2618 | |||
2613 | if ((hpriv->flags & AHCI_HFLAG_NO_MSI) || pci_enable_msi(pdev)) | 2619 | if ((hpriv->flags & AHCI_HFLAG_NO_MSI) || pci_enable_msi(pdev)) |
2614 | pci_intx(pdev, 1); | 2620 | pci_intx(pdev, 1); |
2615 | 2621 | ||
diff --git a/drivers/ata/ata_piix.c b/drivers/ata/ata_piix.c index 5fdf1678d0cc..887d8f46a287 100644 --- a/drivers/ata/ata_piix.c +++ b/drivers/ata/ata_piix.c | |||
@@ -154,11 +154,13 @@ struct piix_map_db { | |||
154 | 154 | ||
155 | struct piix_host_priv { | 155 | struct piix_host_priv { |
156 | const int *map; | 156 | const int *map; |
157 | u32 saved_iocfg; | ||
157 | void __iomem *sidpr; | 158 | void __iomem *sidpr; |
158 | }; | 159 | }; |
159 | 160 | ||
160 | static int piix_init_one(struct pci_dev *pdev, | 161 | static int piix_init_one(struct pci_dev *pdev, |
161 | const struct pci_device_id *ent); | 162 | const struct pci_device_id *ent); |
163 | static void piix_remove_one(struct pci_dev *pdev); | ||
162 | static int piix_pata_prereset(struct ata_link *link, unsigned long deadline); | 164 | static int piix_pata_prereset(struct ata_link *link, unsigned long deadline); |
163 | static void piix_set_piomode(struct ata_port *ap, struct ata_device *adev); | 165 | static void piix_set_piomode(struct ata_port *ap, struct ata_device *adev); |
164 | static void piix_set_dmamode(struct ata_port *ap, struct ata_device *adev); | 166 | static void piix_set_dmamode(struct ata_port *ap, struct ata_device *adev); |
@@ -296,7 +298,7 @@ static struct pci_driver piix_pci_driver = { | |||
296 | .name = DRV_NAME, | 298 | .name = DRV_NAME, |
297 | .id_table = piix_pci_tbl, | 299 | .id_table = piix_pci_tbl, |
298 | .probe = piix_init_one, | 300 | .probe = piix_init_one, |
299 | .remove = ata_pci_remove_one, | 301 | .remove = piix_remove_one, |
300 | #ifdef CONFIG_PM | 302 | #ifdef CONFIG_PM |
301 | .suspend = piix_pci_device_suspend, | 303 | .suspend = piix_pci_device_suspend, |
302 | .resume = piix_pci_device_resume, | 304 | .resume = piix_pci_device_resume, |
@@ -308,7 +310,7 @@ static struct scsi_host_template piix_sht = { | |||
308 | }; | 310 | }; |
309 | 311 | ||
310 | static struct ata_port_operations piix_pata_ops = { | 312 | static struct ata_port_operations piix_pata_ops = { |
311 | .inherits = &ata_bmdma_port_ops, | 313 | .inherits = &ata_bmdma32_port_ops, |
312 | .cable_detect = ata_cable_40wire, | 314 | .cable_detect = ata_cable_40wire, |
313 | .set_piomode = piix_set_piomode, | 315 | .set_piomode = piix_set_piomode, |
314 | .set_dmamode = piix_set_dmamode, | 316 | .set_dmamode = piix_set_dmamode, |
@@ -610,8 +612,9 @@ static const struct ich_laptop ich_laptop[] = { | |||
610 | static int ich_pata_cable_detect(struct ata_port *ap) | 612 | static int ich_pata_cable_detect(struct ata_port *ap) |
611 | { | 613 | { |
612 | struct pci_dev *pdev = to_pci_dev(ap->host->dev); | 614 | struct pci_dev *pdev = to_pci_dev(ap->host->dev); |
615 | struct piix_host_priv *hpriv = ap->host->private_data; | ||
613 | const struct ich_laptop *lap = &ich_laptop[0]; | 616 | const struct ich_laptop *lap = &ich_laptop[0]; |
614 | u8 tmp, mask; | 617 | u8 mask; |
615 | 618 | ||
616 | /* Check for specials - Acer Aspire 5602WLMi */ | 619 | /* Check for specials - Acer Aspire 5602WLMi */ |
617 | while (lap->device) { | 620 | while (lap->device) { |
@@ -625,8 +628,7 @@ static int ich_pata_cable_detect(struct ata_port *ap) | |||
625 | 628 | ||
626 | /* check BIOS cable detect results */ | 629 | /* check BIOS cable detect results */ |
627 | mask = ap->port_no == 0 ? PIIX_80C_PRI : PIIX_80C_SEC; | 630 | mask = ap->port_no == 0 ? PIIX_80C_PRI : PIIX_80C_SEC; |
628 | pci_read_config_byte(pdev, PIIX_IOCFG, &tmp); | 631 | if ((hpriv->saved_iocfg & mask) == 0) |
629 | if ((tmp & mask) == 0) | ||
630 | return ATA_CBL_PATA40; | 632 | return ATA_CBL_PATA40; |
631 | return ATA_CBL_PATA80; | 633 | return ATA_CBL_PATA80; |
632 | } | 634 | } |
@@ -1350,7 +1352,7 @@ static int __devinit piix_init_sidpr(struct ata_host *host) | |||
1350 | return 0; | 1352 | return 0; |
1351 | } | 1353 | } |
1352 | 1354 | ||
1353 | static void piix_iocfg_bit18_quirk(struct pci_dev *pdev) | 1355 | static void piix_iocfg_bit18_quirk(struct ata_host *host) |
1354 | { | 1356 | { |
1355 | static const struct dmi_system_id sysids[] = { | 1357 | static const struct dmi_system_id sysids[] = { |
1356 | { | 1358 | { |
@@ -1367,7 +1369,8 @@ static void piix_iocfg_bit18_quirk(struct pci_dev *pdev) | |||
1367 | 1369 | ||
1368 | { } /* terminate list */ | 1370 | { } /* terminate list */ |
1369 | }; | 1371 | }; |
1370 | u32 iocfg; | 1372 | struct pci_dev *pdev = to_pci_dev(host->dev); |
1373 | struct piix_host_priv *hpriv = host->private_data; | ||
1371 | 1374 | ||
1372 | if (!dmi_check_system(sysids)) | 1375 | if (!dmi_check_system(sysids)) |
1373 | return; | 1376 | return; |
@@ -1376,12 +1379,11 @@ static void piix_iocfg_bit18_quirk(struct pci_dev *pdev) | |||
1376 | * seem to use it to disable a channel. Clear the bit on the | 1379 | * seem to use it to disable a channel. Clear the bit on the |
1377 | * affected systems. | 1380 | * affected systems. |
1378 | */ | 1381 | */ |
1379 | pci_read_config_dword(pdev, PIIX_IOCFG, &iocfg); | 1382 | if (hpriv->saved_iocfg & (1 << 18)) { |
1380 | if (iocfg & (1 << 18)) { | ||
1381 | dev_printk(KERN_INFO, &pdev->dev, | 1383 | dev_printk(KERN_INFO, &pdev->dev, |
1382 | "applying IOCFG bit18 quirk\n"); | 1384 | "applying IOCFG bit18 quirk\n"); |
1383 | iocfg &= ~(1 << 18); | 1385 | pci_write_config_dword(pdev, PIIX_IOCFG, |
1384 | pci_write_config_dword(pdev, PIIX_IOCFG, iocfg); | 1386 | hpriv->saved_iocfg & ~(1 << 18)); |
1385 | } | 1387 | } |
1386 | } | 1388 | } |
1387 | 1389 | ||
@@ -1430,6 +1432,17 @@ static int __devinit piix_init_one(struct pci_dev *pdev, | |||
1430 | if (rc) | 1432 | if (rc) |
1431 | return rc; | 1433 | return rc; |
1432 | 1434 | ||
1435 | hpriv = devm_kzalloc(dev, sizeof(*hpriv), GFP_KERNEL); | ||
1436 | if (!hpriv) | ||
1437 | return -ENOMEM; | ||
1438 | |||
1439 | /* Save IOCFG, this will be used for cable detection, quirk | ||
1440 | * detection and restoration on detach. This is necessary | ||
1441 | * because some ACPI implementations mess up cable related | ||
1442 | * bits on _STM. Reported on kernel bz#11879. | ||
1443 | */ | ||
1444 | pci_read_config_dword(pdev, PIIX_IOCFG, &hpriv->saved_iocfg); | ||
1445 | |||
1433 | /* ICH6R may be driven by either ata_piix or ahci driver | 1446 | /* ICH6R may be driven by either ata_piix or ahci driver |
1434 | * regardless of BIOS configuration. Make sure AHCI mode is | 1447 | * regardless of BIOS configuration. Make sure AHCI mode is |
1435 | * off. | 1448 | * off. |
@@ -1441,10 +1454,6 @@ static int __devinit piix_init_one(struct pci_dev *pdev, | |||
1441 | } | 1454 | } |
1442 | 1455 | ||
1443 | /* SATA map init can change port_info, do it before prepping host */ | 1456 | /* SATA map init can change port_info, do it before prepping host */ |
1444 | hpriv = devm_kzalloc(dev, sizeof(*hpriv), GFP_KERNEL); | ||
1445 | if (!hpriv) | ||
1446 | return -ENOMEM; | ||
1447 | |||
1448 | if (port_flags & ATA_FLAG_SATA) | 1457 | if (port_flags & ATA_FLAG_SATA) |
1449 | hpriv->map = piix_init_sata_map(pdev, port_info, | 1458 | hpriv->map = piix_init_sata_map(pdev, port_info, |
1450 | piix_map_db_table[ent->driver_data]); | 1459 | piix_map_db_table[ent->driver_data]); |
@@ -1463,7 +1472,7 @@ static int __devinit piix_init_one(struct pci_dev *pdev, | |||
1463 | } | 1472 | } |
1464 | 1473 | ||
1465 | /* apply IOCFG bit18 quirk */ | 1474 | /* apply IOCFG bit18 quirk */ |
1466 | piix_iocfg_bit18_quirk(pdev); | 1475 | piix_iocfg_bit18_quirk(host); |
1467 | 1476 | ||
1468 | /* On ICH5, some BIOSen disable the interrupt using the | 1477 | /* On ICH5, some BIOSen disable the interrupt using the |
1469 | * PCI_COMMAND_INTX_DISABLE bit added in PCI 2.3. | 1478 | * PCI_COMMAND_INTX_DISABLE bit added in PCI 2.3. |
@@ -1488,6 +1497,16 @@ static int __devinit piix_init_one(struct pci_dev *pdev, | |||
1488 | return ata_pci_sff_activate_host(host, ata_sff_interrupt, &piix_sht); | 1497 | return ata_pci_sff_activate_host(host, ata_sff_interrupt, &piix_sht); |
1489 | } | 1498 | } |
1490 | 1499 | ||
1500 | static void piix_remove_one(struct pci_dev *pdev) | ||
1501 | { | ||
1502 | struct ata_host *host = dev_get_drvdata(&pdev->dev); | ||
1503 | struct piix_host_priv *hpriv = host->private_data; | ||
1504 | |||
1505 | pci_write_config_dword(pdev, PIIX_IOCFG, hpriv->saved_iocfg); | ||
1506 | |||
1507 | ata_pci_remove_one(pdev); | ||
1508 | } | ||
1509 | |||
1491 | static int __init piix_init(void) | 1510 | static int __init piix_init(void) |
1492 | { | 1511 | { |
1493 | int rc; | 1512 | int rc; |
diff --git a/drivers/ata/libata-core.c b/drivers/ata/libata-core.c index f178a450ec08..175df54eb664 100644 --- a/drivers/ata/libata-core.c +++ b/drivers/ata/libata-core.c | |||
@@ -1007,6 +1007,7 @@ static const char *sata_spd_string(unsigned int spd) | |||
1007 | static const char * const spd_str[] = { | 1007 | static const char * const spd_str[] = { |
1008 | "1.5 Gbps", | 1008 | "1.5 Gbps", |
1009 | "3.0 Gbps", | 1009 | "3.0 Gbps", |
1010 | "6.0 Gbps", | ||
1010 | }; | 1011 | }; |
1011 | 1012 | ||
1012 | if (spd == 0 || (spd - 1) >= ARRAY_SIZE(spd_str)) | 1013 | if (spd == 0 || (spd - 1) >= ARRAY_SIZE(spd_str)) |
@@ -2000,6 +2001,10 @@ unsigned int ata_pio_need_iordy(const struct ata_device *adev) | |||
2000 | as the caller should know this */ | 2001 | as the caller should know this */ |
2001 | if (adev->link->ap->flags & ATA_FLAG_NO_IORDY) | 2002 | if (adev->link->ap->flags & ATA_FLAG_NO_IORDY) |
2002 | return 0; | 2003 | return 0; |
2004 | /* CF spec. r4.1 Table 22 says no iordy on PIO5 and PIO6. */ | ||
2005 | if (ata_id_is_cfa(adev->id) | ||
2006 | && (adev->pio_mode == XFER_PIO_5 || adev->pio_mode == XFER_PIO_6)) | ||
2007 | return 0; | ||
2003 | /* PIO3 and higher it is mandatory */ | 2008 | /* PIO3 and higher it is mandatory */ |
2004 | if (adev->pio_mode > XFER_PIO_2) | 2009 | if (adev->pio_mode > XFER_PIO_2) |
2005 | return 1; | 2010 | return 1; |
diff --git a/drivers/ata/libata-sff.c b/drivers/ata/libata-sff.c index 9033d164c4ec..c59ad76c84b1 100644 --- a/drivers/ata/libata-sff.c +++ b/drivers/ata/libata-sff.c | |||
@@ -66,6 +66,7 @@ const struct ata_port_operations ata_sff_port_ops = { | |||
66 | 66 | ||
67 | .port_start = ata_sff_port_start, | 67 | .port_start = ata_sff_port_start, |
68 | }; | 68 | }; |
69 | EXPORT_SYMBOL_GPL(ata_sff_port_ops); | ||
69 | 70 | ||
70 | const struct ata_port_operations ata_bmdma_port_ops = { | 71 | const struct ata_port_operations ata_bmdma_port_ops = { |
71 | .inherits = &ata_sff_port_ops, | 72 | .inherits = &ata_sff_port_ops, |
@@ -77,6 +78,14 @@ const struct ata_port_operations ata_bmdma_port_ops = { | |||
77 | .bmdma_stop = ata_bmdma_stop, | 78 | .bmdma_stop = ata_bmdma_stop, |
78 | .bmdma_status = ata_bmdma_status, | 79 | .bmdma_status = ata_bmdma_status, |
79 | }; | 80 | }; |
81 | EXPORT_SYMBOL_GPL(ata_bmdma_port_ops); | ||
82 | |||
83 | const struct ata_port_operations ata_bmdma32_port_ops = { | ||
84 | .inherits = &ata_bmdma_port_ops, | ||
85 | |||
86 | .sff_data_xfer = ata_sff_data_xfer32, | ||
87 | }; | ||
88 | EXPORT_SYMBOL_GPL(ata_bmdma32_port_ops); | ||
80 | 89 | ||
81 | /** | 90 | /** |
82 | * ata_fill_sg - Fill PCI IDE PRD table | 91 | * ata_fill_sg - Fill PCI IDE PRD table |
@@ -166,8 +175,9 @@ static void ata_fill_sg_dumb(struct ata_queued_cmd *qc) | |||
166 | blen = len & 0xffff; | 175 | blen = len & 0xffff; |
167 | ap->prd[pi].addr = cpu_to_le32(addr); | 176 | ap->prd[pi].addr = cpu_to_le32(addr); |
168 | if (blen == 0) { | 177 | if (blen == 0) { |
169 | /* Some PATA chipsets like the CS5530 can't | 178 | /* Some PATA chipsets like the CS5530 can't |
170 | cope with 0x0000 meaning 64K as the spec says */ | 179 | cope with 0x0000 meaning 64K as the spec |
180 | says */ | ||
171 | ap->prd[pi].flags_len = cpu_to_le32(0x8000); | 181 | ap->prd[pi].flags_len = cpu_to_le32(0x8000); |
172 | blen = 0x8000; | 182 | blen = 0x8000; |
173 | ap->prd[++pi].addr = cpu_to_le32(addr + 0x8000); | 183 | ap->prd[++pi].addr = cpu_to_le32(addr + 0x8000); |
@@ -200,6 +210,7 @@ void ata_sff_qc_prep(struct ata_queued_cmd *qc) | |||
200 | 210 | ||
201 | ata_fill_sg(qc); | 211 | ata_fill_sg(qc); |
202 | } | 212 | } |
213 | EXPORT_SYMBOL_GPL(ata_sff_qc_prep); | ||
203 | 214 | ||
204 | /** | 215 | /** |
205 | * ata_sff_dumb_qc_prep - Prepare taskfile for submission | 216 | * ata_sff_dumb_qc_prep - Prepare taskfile for submission |
@@ -217,6 +228,7 @@ void ata_sff_dumb_qc_prep(struct ata_queued_cmd *qc) | |||
217 | 228 | ||
218 | ata_fill_sg_dumb(qc); | 229 | ata_fill_sg_dumb(qc); |
219 | } | 230 | } |
231 | EXPORT_SYMBOL_GPL(ata_sff_dumb_qc_prep); | ||
220 | 232 | ||
221 | /** | 233 | /** |
222 | * ata_sff_check_status - Read device status reg & clear interrupt | 234 | * ata_sff_check_status - Read device status reg & clear interrupt |
@@ -233,6 +245,7 @@ u8 ata_sff_check_status(struct ata_port *ap) | |||
233 | { | 245 | { |
234 | return ioread8(ap->ioaddr.status_addr); | 246 | return ioread8(ap->ioaddr.status_addr); |
235 | } | 247 | } |
248 | EXPORT_SYMBOL_GPL(ata_sff_check_status); | ||
236 | 249 | ||
237 | /** | 250 | /** |
238 | * ata_sff_altstatus - Read device alternate status reg | 251 | * ata_sff_altstatus - Read device alternate status reg |
@@ -275,7 +288,7 @@ static u8 ata_sff_irq_status(struct ata_port *ap) | |||
275 | status = ata_sff_altstatus(ap); | 288 | status = ata_sff_altstatus(ap); |
276 | /* Not us: We are busy */ | 289 | /* Not us: We are busy */ |
277 | if (status & ATA_BUSY) | 290 | if (status & ATA_BUSY) |
278 | return status; | 291 | return status; |
279 | } | 292 | } |
280 | /* Clear INTRQ latch */ | 293 | /* Clear INTRQ latch */ |
281 | status = ap->ops->sff_check_status(ap); | 294 | status = ap->ops->sff_check_status(ap); |
@@ -319,6 +332,7 @@ void ata_sff_pause(struct ata_port *ap) | |||
319 | ata_sff_sync(ap); | 332 | ata_sff_sync(ap); |
320 | ndelay(400); | 333 | ndelay(400); |
321 | } | 334 | } |
335 | EXPORT_SYMBOL_GPL(ata_sff_pause); | ||
322 | 336 | ||
323 | /** | 337 | /** |
324 | * ata_sff_dma_pause - Pause before commencing DMA | 338 | * ata_sff_dma_pause - Pause before commencing DMA |
@@ -327,7 +341,7 @@ void ata_sff_pause(struct ata_port *ap) | |||
327 | * Perform I/O fencing and ensure sufficient cycle delays occur | 341 | * Perform I/O fencing and ensure sufficient cycle delays occur |
328 | * for the HDMA1:0 transition | 342 | * for the HDMA1:0 transition |
329 | */ | 343 | */ |
330 | 344 | ||
331 | void ata_sff_dma_pause(struct ata_port *ap) | 345 | void ata_sff_dma_pause(struct ata_port *ap) |
332 | { | 346 | { |
333 | if (ap->ops->sff_check_altstatus || ap->ioaddr.altstatus_addr) { | 347 | if (ap->ops->sff_check_altstatus || ap->ioaddr.altstatus_addr) { |
@@ -341,6 +355,7 @@ void ata_sff_dma_pause(struct ata_port *ap) | |||
341 | corruption. */ | 355 | corruption. */ |
342 | BUG(); | 356 | BUG(); |
343 | } | 357 | } |
358 | EXPORT_SYMBOL_GPL(ata_sff_dma_pause); | ||
344 | 359 | ||
345 | /** | 360 | /** |
346 | * ata_sff_busy_sleep - sleep until BSY clears, or timeout | 361 | * ata_sff_busy_sleep - sleep until BSY clears, or timeout |
@@ -396,6 +411,7 @@ int ata_sff_busy_sleep(struct ata_port *ap, | |||
396 | 411 | ||
397 | return 0; | 412 | return 0; |
398 | } | 413 | } |
414 | EXPORT_SYMBOL_GPL(ata_sff_busy_sleep); | ||
399 | 415 | ||
400 | static int ata_sff_check_ready(struct ata_link *link) | 416 | static int ata_sff_check_ready(struct ata_link *link) |
401 | { | 417 | { |
@@ -422,6 +438,7 @@ int ata_sff_wait_ready(struct ata_link *link, unsigned long deadline) | |||
422 | { | 438 | { |
423 | return ata_wait_ready(link, deadline, ata_sff_check_ready); | 439 | return ata_wait_ready(link, deadline, ata_sff_check_ready); |
424 | } | 440 | } |
441 | EXPORT_SYMBOL_GPL(ata_sff_wait_ready); | ||
425 | 442 | ||
426 | /** | 443 | /** |
427 | * ata_sff_dev_select - Select device 0/1 on ATA bus | 444 | * ata_sff_dev_select - Select device 0/1 on ATA bus |
@@ -449,6 +466,7 @@ void ata_sff_dev_select(struct ata_port *ap, unsigned int device) | |||
449 | iowrite8(tmp, ap->ioaddr.device_addr); | 466 | iowrite8(tmp, ap->ioaddr.device_addr); |
450 | ata_sff_pause(ap); /* needed; also flushes, for mmio */ | 467 | ata_sff_pause(ap); /* needed; also flushes, for mmio */ |
451 | } | 468 | } |
469 | EXPORT_SYMBOL_GPL(ata_sff_dev_select); | ||
452 | 470 | ||
453 | /** | 471 | /** |
454 | * ata_dev_select - Select device 0/1 on ATA bus | 472 | * ata_dev_select - Select device 0/1 on ATA bus |
@@ -513,6 +531,7 @@ u8 ata_sff_irq_on(struct ata_port *ap) | |||
513 | 531 | ||
514 | return tmp; | 532 | return tmp; |
515 | } | 533 | } |
534 | EXPORT_SYMBOL_GPL(ata_sff_irq_on); | ||
516 | 535 | ||
517 | /** | 536 | /** |
518 | * ata_sff_irq_clear - Clear PCI IDE BMDMA interrupt. | 537 | * ata_sff_irq_clear - Clear PCI IDE BMDMA interrupt. |
@@ -534,6 +553,7 @@ void ata_sff_irq_clear(struct ata_port *ap) | |||
534 | 553 | ||
535 | iowrite8(ioread8(mmio + ATA_DMA_STATUS), mmio + ATA_DMA_STATUS); | 554 | iowrite8(ioread8(mmio + ATA_DMA_STATUS), mmio + ATA_DMA_STATUS); |
536 | } | 555 | } |
556 | EXPORT_SYMBOL_GPL(ata_sff_irq_clear); | ||
537 | 557 | ||
538 | /** | 558 | /** |
539 | * ata_sff_tf_load - send taskfile registers to host controller | 559 | * ata_sff_tf_load - send taskfile registers to host controller |
@@ -593,6 +613,7 @@ void ata_sff_tf_load(struct ata_port *ap, const struct ata_taskfile *tf) | |||
593 | 613 | ||
594 | ata_wait_idle(ap); | 614 | ata_wait_idle(ap); |
595 | } | 615 | } |
616 | EXPORT_SYMBOL_GPL(ata_sff_tf_load); | ||
596 | 617 | ||
597 | /** | 618 | /** |
598 | * ata_sff_tf_read - input device's ATA taskfile shadow registers | 619 | * ata_sff_tf_read - input device's ATA taskfile shadow registers |
@@ -633,6 +654,7 @@ void ata_sff_tf_read(struct ata_port *ap, struct ata_taskfile *tf) | |||
633 | WARN_ON(1); | 654 | WARN_ON(1); |
634 | } | 655 | } |
635 | } | 656 | } |
657 | EXPORT_SYMBOL_GPL(ata_sff_tf_read); | ||
636 | 658 | ||
637 | /** | 659 | /** |
638 | * ata_sff_exec_command - issue ATA command to host controller | 660 | * ata_sff_exec_command - issue ATA command to host controller |
@@ -652,6 +674,7 @@ void ata_sff_exec_command(struct ata_port *ap, const struct ata_taskfile *tf) | |||
652 | iowrite8(tf->command, ap->ioaddr.command_addr); | 674 | iowrite8(tf->command, ap->ioaddr.command_addr); |
653 | ata_sff_pause(ap); | 675 | ata_sff_pause(ap); |
654 | } | 676 | } |
677 | EXPORT_SYMBOL_GPL(ata_sff_exec_command); | ||
655 | 678 | ||
656 | /** | 679 | /** |
657 | * ata_tf_to_host - issue ATA taskfile to host controller | 680 | * ata_tf_to_host - issue ATA taskfile to host controller |
@@ -717,6 +740,53 @@ unsigned int ata_sff_data_xfer(struct ata_device *dev, unsigned char *buf, | |||
717 | 740 | ||
718 | return words << 1; | 741 | return words << 1; |
719 | } | 742 | } |
743 | EXPORT_SYMBOL_GPL(ata_sff_data_xfer); | ||
744 | |||
745 | /** | ||
746 | * ata_sff_data_xfer32 - Transfer data by PIO | ||
747 | * @dev: device to target | ||
748 | * @buf: data buffer | ||
749 | * @buflen: buffer length | ||
750 | * @rw: read/write | ||
751 | * | ||
752 | * Transfer data from/to the device data register by PIO using 32bit | ||
753 | * I/O operations. | ||
754 | * | ||
755 | * LOCKING: | ||
756 | * Inherited from caller. | ||
757 | * | ||
758 | * RETURNS: | ||
759 | * Bytes consumed. | ||
760 | */ | ||
761 | |||
762 | unsigned int ata_sff_data_xfer32(struct ata_device *dev, unsigned char *buf, | ||
763 | unsigned int buflen, int rw) | ||
764 | { | ||
765 | struct ata_port *ap = dev->link->ap; | ||
766 | void __iomem *data_addr = ap->ioaddr.data_addr; | ||
767 | unsigned int words = buflen >> 2; | ||
768 | int slop = buflen & 3; | ||
769 | |||
770 | /* Transfer multiple of 4 bytes */ | ||
771 | if (rw == READ) | ||
772 | ioread32_rep(data_addr, buf, words); | ||
773 | else | ||
774 | iowrite32_rep(data_addr, buf, words); | ||
775 | |||
776 | if (unlikely(slop)) { | ||
777 | __le32 pad; | ||
778 | if (rw == READ) { | ||
779 | pad = cpu_to_le32(ioread32(ap->ioaddr.data_addr)); | ||
780 | memcpy(buf + buflen - slop, &pad, slop); | ||
781 | } else { | ||
782 | memcpy(&pad, buf + buflen - slop, slop); | ||
783 | iowrite32(le32_to_cpu(pad), ap->ioaddr.data_addr); | ||
784 | } | ||
785 | words++; | ||
786 | } | ||
787 | return words << 2; | ||
788 | } | ||
789 | EXPORT_SYMBOL_GPL(ata_sff_data_xfer32); | ||
720 | 790 | ||
721 | /** | 791 | /** |
722 | * ata_sff_data_xfer_noirq - Transfer data by PIO | 792 | * ata_sff_data_xfer_noirq - Transfer data by PIO |
@@ -746,6 +816,7 @@ unsigned int ata_sff_data_xfer_noirq(struct ata_device *dev, unsigned char *buf, | |||
746 | 816 | ||
747 | return consumed; | 817 | return consumed; |
748 | } | 818 | } |
819 | EXPORT_SYMBOL_GPL(ata_sff_data_xfer_noirq); | ||
749 | 820 | ||
750 | /** | 821 | /** |
751 | * ata_pio_sector - Transfer a sector of data. | 822 | * ata_pio_sector - Transfer a sector of data. |
@@ -922,13 +993,15 @@ next_sg: | |||
922 | buf = kmap_atomic(page, KM_IRQ0); | 993 | buf = kmap_atomic(page, KM_IRQ0); |
923 | 994 | ||
924 | /* do the actual data transfer */ | 995 | /* do the actual data transfer */ |
925 | consumed = ap->ops->sff_data_xfer(dev, buf + offset, count, rw); | 996 | consumed = ap->ops->sff_data_xfer(dev, buf + offset, |
997 | count, rw); | ||
926 | 998 | ||
927 | kunmap_atomic(buf, KM_IRQ0); | 999 | kunmap_atomic(buf, KM_IRQ0); |
928 | local_irq_restore(flags); | 1000 | local_irq_restore(flags); |
929 | } else { | 1001 | } else { |
930 | buf = page_address(page); | 1002 | buf = page_address(page); |
931 | consumed = ap->ops->sff_data_xfer(dev, buf + offset, count, rw); | 1003 | consumed = ap->ops->sff_data_xfer(dev, buf + offset, |
1004 | count, rw); | ||
932 | } | 1005 | } |
933 | 1006 | ||
934 | bytes -= min(bytes, consumed); | 1007 | bytes -= min(bytes, consumed); |
@@ -1013,18 +1086,19 @@ static void atapi_pio_bytes(struct ata_queued_cmd *qc) | |||
1013 | * RETURNS: | 1086 | * RETURNS: |
1014 | * 1 if ok in workqueue, 0 otherwise. | 1087 | * 1 if ok in workqueue, 0 otherwise. |
1015 | */ | 1088 | */ |
1016 | static inline int ata_hsm_ok_in_wq(struct ata_port *ap, struct ata_queued_cmd *qc) | 1089 | static inline int ata_hsm_ok_in_wq(struct ata_port *ap, |
1090 | struct ata_queued_cmd *qc) | ||
1017 | { | 1091 | { |
1018 | if (qc->tf.flags & ATA_TFLAG_POLLING) | 1092 | if (qc->tf.flags & ATA_TFLAG_POLLING) |
1019 | return 1; | 1093 | return 1; |
1020 | 1094 | ||
1021 | if (ap->hsm_task_state == HSM_ST_FIRST) { | 1095 | if (ap->hsm_task_state == HSM_ST_FIRST) { |
1022 | if (qc->tf.protocol == ATA_PROT_PIO && | 1096 | if (qc->tf.protocol == ATA_PROT_PIO && |
1023 | (qc->tf.flags & ATA_TFLAG_WRITE)) | 1097 | (qc->tf.flags & ATA_TFLAG_WRITE)) |
1024 | return 1; | 1098 | return 1; |
1025 | 1099 | ||
1026 | if (ata_is_atapi(qc->tf.protocol) && | 1100 | if (ata_is_atapi(qc->tf.protocol) && |
1027 | !(qc->dev->flags & ATA_DFLAG_CDB_INTR)) | 1101 | !(qc->dev->flags & ATA_DFLAG_CDB_INTR)) |
1028 | return 1; | 1102 | return 1; |
1029 | } | 1103 | } |
1030 | 1104 | ||
@@ -1338,6 +1412,7 @@ fsm_start: | |||
1338 | 1412 | ||
1339 | return poll_next; | 1413 | return poll_next; |
1340 | } | 1414 | } |
1415 | EXPORT_SYMBOL_GPL(ata_sff_hsm_move); | ||
1341 | 1416 | ||
1342 | void ata_pio_task(struct work_struct *work) | 1417 | void ata_pio_task(struct work_struct *work) |
1343 | { | 1418 | { |
@@ -1507,6 +1582,7 @@ unsigned int ata_sff_qc_issue(struct ata_queued_cmd *qc) | |||
1507 | 1582 | ||
1508 | return 0; | 1583 | return 0; |
1509 | } | 1584 | } |
1585 | EXPORT_SYMBOL_GPL(ata_sff_qc_issue); | ||
1510 | 1586 | ||
1511 | /** | 1587 | /** |
1512 | * ata_sff_qc_fill_rtf - fill result TF using ->sff_tf_read | 1588 | * ata_sff_qc_fill_rtf - fill result TF using ->sff_tf_read |
@@ -1526,6 +1602,7 @@ bool ata_sff_qc_fill_rtf(struct ata_queued_cmd *qc) | |||
1526 | qc->ap->ops->sff_tf_read(qc->ap, &qc->result_tf); | 1602 | qc->ap->ops->sff_tf_read(qc->ap, &qc->result_tf); |
1527 | return true; | 1603 | return true; |
1528 | } | 1604 | } |
1605 | EXPORT_SYMBOL_GPL(ata_sff_qc_fill_rtf); | ||
1529 | 1606 | ||
1530 | /** | 1607 | /** |
1531 | * ata_sff_host_intr - Handle host interrupt for given (port, task) | 1608 | * ata_sff_host_intr - Handle host interrupt for given (port, task) |
@@ -1623,6 +1700,7 @@ idle_irq: | |||
1623 | #endif | 1700 | #endif |
1624 | return 0; /* irq not handled */ | 1701 | return 0; /* irq not handled */ |
1625 | } | 1702 | } |
1703 | EXPORT_SYMBOL_GPL(ata_sff_host_intr); | ||
1626 | 1704 | ||
1627 | /** | 1705 | /** |
1628 | * ata_sff_interrupt - Default ATA host interrupt handler | 1706 | * ata_sff_interrupt - Default ATA host interrupt handler |
@@ -1667,6 +1745,7 @@ irqreturn_t ata_sff_interrupt(int irq, void *dev_instance) | |||
1667 | 1745 | ||
1668 | return IRQ_RETVAL(handled); | 1746 | return IRQ_RETVAL(handled); |
1669 | } | 1747 | } |
1748 | EXPORT_SYMBOL_GPL(ata_sff_interrupt); | ||
1670 | 1749 | ||
1671 | /** | 1750 | /** |
1672 | * ata_sff_freeze - Freeze SFF controller port | 1751 | * ata_sff_freeze - Freeze SFF controller port |
@@ -1695,6 +1774,7 @@ void ata_sff_freeze(struct ata_port *ap) | |||
1695 | 1774 | ||
1696 | ap->ops->sff_irq_clear(ap); | 1775 | ap->ops->sff_irq_clear(ap); |
1697 | } | 1776 | } |
1777 | EXPORT_SYMBOL_GPL(ata_sff_freeze); | ||
1698 | 1778 | ||
1699 | /** | 1779 | /** |
1700 | * ata_sff_thaw - Thaw SFF controller port | 1780 | * ata_sff_thaw - Thaw SFF controller port |
@@ -1712,6 +1792,7 @@ void ata_sff_thaw(struct ata_port *ap) | |||
1712 | ap->ops->sff_irq_clear(ap); | 1792 | ap->ops->sff_irq_clear(ap); |
1713 | ap->ops->sff_irq_on(ap); | 1793 | ap->ops->sff_irq_on(ap); |
1714 | } | 1794 | } |
1795 | EXPORT_SYMBOL_GPL(ata_sff_thaw); | ||
1715 | 1796 | ||
1716 | /** | 1797 | /** |
1717 | * ata_sff_prereset - prepare SFF link for reset | 1798 | * ata_sff_prereset - prepare SFF link for reset |
@@ -1753,6 +1834,7 @@ int ata_sff_prereset(struct ata_link *link, unsigned long deadline) | |||
1753 | 1834 | ||
1754 | return 0; | 1835 | return 0; |
1755 | } | 1836 | } |
1837 | EXPORT_SYMBOL_GPL(ata_sff_prereset); | ||
1756 | 1838 | ||
1757 | /** | 1839 | /** |
1758 | * ata_devchk - PATA device presence detection | 1840 | * ata_devchk - PATA device presence detection |
@@ -1865,6 +1947,7 @@ unsigned int ata_sff_dev_classify(struct ata_device *dev, int present, | |||
1865 | 1947 | ||
1866 | return class; | 1948 | return class; |
1867 | } | 1949 | } |
1950 | EXPORT_SYMBOL_GPL(ata_sff_dev_classify); | ||
1868 | 1951 | ||
1869 | /** | 1952 | /** |
1870 | * ata_sff_wait_after_reset - wait for devices to become ready after reset | 1953 | * ata_sff_wait_after_reset - wait for devices to become ready after reset |
@@ -1941,6 +2024,7 @@ int ata_sff_wait_after_reset(struct ata_link *link, unsigned int devmask, | |||
1941 | 2024 | ||
1942 | return ret; | 2025 | return ret; |
1943 | } | 2026 | } |
2027 | EXPORT_SYMBOL_GPL(ata_sff_wait_after_reset); | ||
1944 | 2028 | ||
1945 | static int ata_bus_softreset(struct ata_port *ap, unsigned int devmask, | 2029 | static int ata_bus_softreset(struct ata_port *ap, unsigned int devmask, |
1946 | unsigned long deadline) | 2030 | unsigned long deadline) |
@@ -2013,6 +2097,7 @@ int ata_sff_softreset(struct ata_link *link, unsigned int *classes, | |||
2013 | DPRINTK("EXIT, classes[0]=%u [1]=%u\n", classes[0], classes[1]); | 2097 | DPRINTK("EXIT, classes[0]=%u [1]=%u\n", classes[0], classes[1]); |
2014 | return 0; | 2098 | return 0; |
2015 | } | 2099 | } |
2100 | EXPORT_SYMBOL_GPL(ata_sff_softreset); | ||
2016 | 2101 | ||
2017 | /** | 2102 | /** |
2018 | * sata_sff_hardreset - reset host port via SATA phy reset | 2103 | * sata_sff_hardreset - reset host port via SATA phy reset |
@@ -2045,6 +2130,7 @@ int sata_sff_hardreset(struct ata_link *link, unsigned int *class, | |||
2045 | DPRINTK("EXIT, class=%u\n", *class); | 2130 | DPRINTK("EXIT, class=%u\n", *class); |
2046 | return rc; | 2131 | return rc; |
2047 | } | 2132 | } |
2133 | EXPORT_SYMBOL_GPL(sata_sff_hardreset); | ||
2048 | 2134 | ||
2049 | /** | 2135 | /** |
2050 | * ata_sff_postreset - SFF postreset callback | 2136 | * ata_sff_postreset - SFF postreset callback |
@@ -2080,6 +2166,7 @@ void ata_sff_postreset(struct ata_link *link, unsigned int *classes) | |||
2080 | if (ap->ioaddr.ctl_addr) | 2166 | if (ap->ioaddr.ctl_addr) |
2081 | iowrite8(ap->ctl, ap->ioaddr.ctl_addr); | 2167 | iowrite8(ap->ctl, ap->ioaddr.ctl_addr); |
2082 | } | 2168 | } |
2169 | EXPORT_SYMBOL_GPL(ata_sff_postreset); | ||
2083 | 2170 | ||
2084 | /** | 2171 | /** |
2085 | * ata_sff_error_handler - Stock error handler for BMDMA controller | 2172 | * ata_sff_error_handler - Stock error handler for BMDMA controller |
@@ -2152,6 +2239,7 @@ void ata_sff_error_handler(struct ata_port *ap) | |||
2152 | ata_do_eh(ap, ap->ops->prereset, softreset, hardreset, | 2239 | ata_do_eh(ap, ap->ops->prereset, softreset, hardreset, |
2153 | ap->ops->postreset); | 2240 | ap->ops->postreset); |
2154 | } | 2241 | } |
2242 | EXPORT_SYMBOL_GPL(ata_sff_error_handler); | ||
2155 | 2243 | ||
2156 | /** | 2244 | /** |
2157 | * ata_sff_post_internal_cmd - Stock post_internal_cmd for SFF controller | 2245 | * ata_sff_post_internal_cmd - Stock post_internal_cmd for SFF controller |
@@ -2174,6 +2262,7 @@ void ata_sff_post_internal_cmd(struct ata_queued_cmd *qc) | |||
2174 | 2262 | ||
2175 | spin_unlock_irqrestore(ap->lock, flags); | 2263 | spin_unlock_irqrestore(ap->lock, flags); |
2176 | } | 2264 | } |
2265 | EXPORT_SYMBOL_GPL(ata_sff_post_internal_cmd); | ||
2177 | 2266 | ||
2178 | /** | 2267 | /** |
2179 | * ata_sff_port_start - Set port up for dma. | 2268 | * ata_sff_port_start - Set port up for dma. |
@@ -2194,6 +2283,7 @@ int ata_sff_port_start(struct ata_port *ap) | |||
2194 | return ata_port_start(ap); | 2283 | return ata_port_start(ap); |
2195 | return 0; | 2284 | return 0; |
2196 | } | 2285 | } |
2286 | EXPORT_SYMBOL_GPL(ata_sff_port_start); | ||
2197 | 2287 | ||
2198 | /** | 2288 | /** |
2199 | * ata_sff_std_ports - initialize ioaddr with standard port offsets. | 2289 | * ata_sff_std_ports - initialize ioaddr with standard port offsets. |
@@ -2219,6 +2309,7 @@ void ata_sff_std_ports(struct ata_ioports *ioaddr) | |||
2219 | ioaddr->status_addr = ioaddr->cmd_addr + ATA_REG_STATUS; | 2309 | ioaddr->status_addr = ioaddr->cmd_addr + ATA_REG_STATUS; |
2220 | ioaddr->command_addr = ioaddr->cmd_addr + ATA_REG_CMD; | 2310 | ioaddr->command_addr = ioaddr->cmd_addr + ATA_REG_CMD; |
2221 | } | 2311 | } |
2312 | EXPORT_SYMBOL_GPL(ata_sff_std_ports); | ||
2222 | 2313 | ||
2223 | unsigned long ata_bmdma_mode_filter(struct ata_device *adev, | 2314 | unsigned long ata_bmdma_mode_filter(struct ata_device *adev, |
2224 | unsigned long xfer_mask) | 2315 | unsigned long xfer_mask) |
@@ -2230,6 +2321,7 @@ unsigned long ata_bmdma_mode_filter(struct ata_device *adev, | |||
2230 | xfer_mask &= ~(ATA_MASK_MWDMA | ATA_MASK_UDMA); | 2321 | xfer_mask &= ~(ATA_MASK_MWDMA | ATA_MASK_UDMA); |
2231 | return xfer_mask; | 2322 | return xfer_mask; |
2232 | } | 2323 | } |
2324 | EXPORT_SYMBOL_GPL(ata_bmdma_mode_filter); | ||
2233 | 2325 | ||
2234 | /** | 2326 | /** |
2235 | * ata_bmdma_setup - Set up PCI IDE BMDMA transaction | 2327 | * ata_bmdma_setup - Set up PCI IDE BMDMA transaction |
@@ -2258,6 +2350,7 @@ void ata_bmdma_setup(struct ata_queued_cmd *qc) | |||
2258 | /* issue r/w command */ | 2350 | /* issue r/w command */ |
2259 | ap->ops->sff_exec_command(ap, &qc->tf); | 2351 | ap->ops->sff_exec_command(ap, &qc->tf); |
2260 | } | 2352 | } |
2353 | EXPORT_SYMBOL_GPL(ata_bmdma_setup); | ||
2261 | 2354 | ||
2262 | /** | 2355 | /** |
2263 | * ata_bmdma_start - Start a PCI IDE BMDMA transaction | 2356 | * ata_bmdma_start - Start a PCI IDE BMDMA transaction |
@@ -2290,6 +2383,7 @@ void ata_bmdma_start(struct ata_queued_cmd *qc) | |||
2290 | * unneccessarily delayed for MMIO | 2383 | * unneccessarily delayed for MMIO |
2291 | */ | 2384 | */ |
2292 | } | 2385 | } |
2386 | EXPORT_SYMBOL_GPL(ata_bmdma_start); | ||
2293 | 2387 | ||
2294 | /** | 2388 | /** |
2295 | * ata_bmdma_stop - Stop PCI IDE BMDMA transfer | 2389 | * ata_bmdma_stop - Stop PCI IDE BMDMA transfer |
@@ -2314,6 +2408,7 @@ void ata_bmdma_stop(struct ata_queued_cmd *qc) | |||
2314 | /* one-PIO-cycle guaranteed wait, per spec, for HDMA1:0 transition */ | 2408 | /* one-PIO-cycle guaranteed wait, per spec, for HDMA1:0 transition */ |
2315 | ata_sff_dma_pause(ap); | 2409 | ata_sff_dma_pause(ap); |
2316 | } | 2410 | } |
2411 | EXPORT_SYMBOL_GPL(ata_bmdma_stop); | ||
2317 | 2412 | ||
2318 | /** | 2413 | /** |
2319 | * ata_bmdma_status - Read PCI IDE BMDMA status | 2414 | * ata_bmdma_status - Read PCI IDE BMDMA status |
@@ -2330,6 +2425,7 @@ u8 ata_bmdma_status(struct ata_port *ap) | |||
2330 | { | 2425 | { |
2331 | return ioread8(ap->ioaddr.bmdma_addr + ATA_DMA_STATUS); | 2426 | return ioread8(ap->ioaddr.bmdma_addr + ATA_DMA_STATUS); |
2332 | } | 2427 | } |
2428 | EXPORT_SYMBOL_GPL(ata_bmdma_status); | ||
2333 | 2429 | ||
2334 | /** | 2430 | /** |
2335 | * ata_bus_reset - reset host port and associated ATA channel | 2431 | * ata_bus_reset - reset host port and associated ATA channel |
@@ -2422,6 +2518,7 @@ err_out: | |||
2422 | 2518 | ||
2423 | DPRINTK("EXIT\n"); | 2519 | DPRINTK("EXIT\n"); |
2424 | } | 2520 | } |
2521 | EXPORT_SYMBOL_GPL(ata_bus_reset); | ||
2425 | 2522 | ||
2426 | #ifdef CONFIG_PCI | 2523 | #ifdef CONFIG_PCI |
2427 | 2524 | ||
@@ -2449,6 +2546,7 @@ int ata_pci_bmdma_clear_simplex(struct pci_dev *pdev) | |||
2449 | return -EOPNOTSUPP; | 2546 | return -EOPNOTSUPP; |
2450 | return 0; | 2547 | return 0; |
2451 | } | 2548 | } |
2549 | EXPORT_SYMBOL_GPL(ata_pci_bmdma_clear_simplex); | ||
2452 | 2550 | ||
2453 | /** | 2551 | /** |
2454 | * ata_pci_bmdma_init - acquire PCI BMDMA resources and init ATA host | 2552 | * ata_pci_bmdma_init - acquire PCI BMDMA resources and init ATA host |
@@ -2501,11 +2599,12 @@ int ata_pci_bmdma_init(struct ata_host *host) | |||
2501 | host->flags |= ATA_HOST_SIMPLEX; | 2599 | host->flags |= ATA_HOST_SIMPLEX; |
2502 | 2600 | ||
2503 | ata_port_desc(ap, "bmdma 0x%llx", | 2601 | ata_port_desc(ap, "bmdma 0x%llx", |
2504 | (unsigned long long)pci_resource_start(pdev, 4) + 8 * i); | 2602 | (unsigned long long)pci_resource_start(pdev, 4) + 8 * i); |
2505 | } | 2603 | } |
2506 | 2604 | ||
2507 | return 0; | 2605 | return 0; |
2508 | } | 2606 | } |
2607 | EXPORT_SYMBOL_GPL(ata_pci_bmdma_init); | ||
2509 | 2608 | ||
2510 | static int ata_resources_present(struct pci_dev *pdev, int port) | 2609 | static int ata_resources_present(struct pci_dev *pdev, int port) |
2511 | { | 2610 | { |
@@ -2513,7 +2612,7 @@ static int ata_resources_present(struct pci_dev *pdev, int port) | |||
2513 | 2612 | ||
2514 | /* Check the PCI resources for this channel are enabled */ | 2613 | /* Check the PCI resources for this channel are enabled */ |
2515 | port = port * 2; | 2614 | port = port * 2; |
2516 | for (i = 0; i < 2; i ++) { | 2615 | for (i = 0; i < 2; i++) { |
2517 | if (pci_resource_start(pdev, port + i) == 0 || | 2616 | if (pci_resource_start(pdev, port + i) == 0 || |
2518 | pci_resource_len(pdev, port + i) == 0) | 2617 | pci_resource_len(pdev, port + i) == 0) |
2519 | return 0; | 2618 | return 0; |
@@ -2598,6 +2697,7 @@ int ata_pci_sff_init_host(struct ata_host *host) | |||
2598 | 2697 | ||
2599 | return 0; | 2698 | return 0; |
2600 | } | 2699 | } |
2700 | EXPORT_SYMBOL_GPL(ata_pci_sff_init_host); | ||
2601 | 2701 | ||
2602 | /** | 2702 | /** |
2603 | * ata_pci_sff_prepare_host - helper to prepare native PCI ATA host | 2703 | * ata_pci_sff_prepare_host - helper to prepare native PCI ATA host |
@@ -2615,7 +2715,7 @@ int ata_pci_sff_init_host(struct ata_host *host) | |||
2615 | * 0 on success, -errno otherwise. | 2715 | * 0 on success, -errno otherwise. |
2616 | */ | 2716 | */ |
2617 | int ata_pci_sff_prepare_host(struct pci_dev *pdev, | 2717 | int ata_pci_sff_prepare_host(struct pci_dev *pdev, |
2618 | const struct ata_port_info * const * ppi, | 2718 | const struct ata_port_info * const *ppi, |
2619 | struct ata_host **r_host) | 2719 | struct ata_host **r_host) |
2620 | { | 2720 | { |
2621 | struct ata_host *host; | 2721 | struct ata_host *host; |
@@ -2645,17 +2745,18 @@ int ata_pci_sff_prepare_host(struct pci_dev *pdev, | |||
2645 | *r_host = host; | 2745 | *r_host = host; |
2646 | return 0; | 2746 | return 0; |
2647 | 2747 | ||
2648 | err_bmdma: | 2748 | err_bmdma: |
2649 | /* This is necessary because PCI and iomap resources are | 2749 | /* This is necessary because PCI and iomap resources are |
2650 | * merged and releasing the top group won't release the | 2750 | * merged and releasing the top group won't release the |
2651 | * acquired resources if some of those have been acquired | 2751 | * acquired resources if some of those have been acquired |
2652 | * before entering this function. | 2752 | * before entering this function. |
2653 | */ | 2753 | */ |
2654 | pcim_iounmap_regions(pdev, 0xf); | 2754 | pcim_iounmap_regions(pdev, 0xf); |
2655 | err_out: | 2755 | err_out: |
2656 | devres_release_group(&pdev->dev, NULL); | 2756 | devres_release_group(&pdev->dev, NULL); |
2657 | return rc; | 2757 | return rc; |
2658 | } | 2758 | } |
2759 | EXPORT_SYMBOL_GPL(ata_pci_sff_prepare_host); | ||
2659 | 2760 | ||
2660 | /** | 2761 | /** |
2661 | * ata_pci_sff_activate_host - start SFF host, request IRQ and register it | 2762 | * ata_pci_sff_activate_host - start SFF host, request IRQ and register it |
@@ -2741,7 +2842,7 @@ int ata_pci_sff_activate_host(struct ata_host *host, | |||
2741 | } | 2842 | } |
2742 | 2843 | ||
2743 | rc = ata_host_register(host, sht); | 2844 | rc = ata_host_register(host, sht); |
2744 | out: | 2845 | out: |
2745 | if (rc == 0) | 2846 | if (rc == 0) |
2746 | devres_remove_group(dev, NULL); | 2847 | devres_remove_group(dev, NULL); |
2747 | else | 2848 | else |
@@ -2749,6 +2850,7 @@ int ata_pci_sff_activate_host(struct ata_host *host, | |||
2749 | 2850 | ||
2750 | return rc; | 2851 | return rc; |
2751 | } | 2852 | } |
2853 | EXPORT_SYMBOL_GPL(ata_pci_sff_activate_host); | ||
2752 | 2854 | ||
2753 | /** | 2855 | /** |
2754 | * ata_pci_sff_init_one - Initialize/register PCI IDE host controller | 2856 | * ata_pci_sff_init_one - Initialize/register PCI IDE host controller |
@@ -2776,7 +2878,7 @@ int ata_pci_sff_activate_host(struct ata_host *host, | |||
2776 | * Zero on success, negative on errno-based value on error. | 2878 | * Zero on success, negative on errno-based value on error. |
2777 | */ | 2879 | */ |
2778 | int ata_pci_sff_init_one(struct pci_dev *pdev, | 2880 | int ata_pci_sff_init_one(struct pci_dev *pdev, |
2779 | const struct ata_port_info * const * ppi, | 2881 | const struct ata_port_info * const *ppi, |
2780 | struct scsi_host_template *sht, void *host_priv) | 2882 | struct scsi_host_template *sht, void *host_priv) |
2781 | { | 2883 | { |
2782 | struct device *dev = &pdev->dev; | 2884 | struct device *dev = &pdev->dev; |
@@ -2815,7 +2917,7 @@ int ata_pci_sff_init_one(struct pci_dev *pdev, | |||
2815 | 2917 | ||
2816 | pci_set_master(pdev); | 2918 | pci_set_master(pdev); |
2817 | rc = ata_pci_sff_activate_host(host, ata_sff_interrupt, sht); | 2919 | rc = ata_pci_sff_activate_host(host, ata_sff_interrupt, sht); |
2818 | out: | 2920 | out: |
2819 | if (rc == 0) | 2921 | if (rc == 0) |
2820 | devres_remove_group(&pdev->dev, NULL); | 2922 | devres_remove_group(&pdev->dev, NULL); |
2821 | else | 2923 | else |
@@ -2823,54 +2925,7 @@ int ata_pci_sff_init_one(struct pci_dev *pdev, | |||
2823 | 2925 | ||
2824 | return rc; | 2926 | return rc; |
2825 | } | 2927 | } |
2928 | EXPORT_SYMBOL_GPL(ata_pci_sff_init_one); | ||
2826 | 2929 | ||
2827 | #endif /* CONFIG_PCI */ | 2930 | #endif /* CONFIG_PCI */ |
2828 | 2931 | ||
2829 | EXPORT_SYMBOL_GPL(ata_sff_port_ops); | ||
2830 | EXPORT_SYMBOL_GPL(ata_bmdma_port_ops); | ||
2831 | EXPORT_SYMBOL_GPL(ata_sff_qc_prep); | ||
2832 | EXPORT_SYMBOL_GPL(ata_sff_dumb_qc_prep); | ||
2833 | EXPORT_SYMBOL_GPL(ata_sff_dev_select); | ||
2834 | EXPORT_SYMBOL_GPL(ata_sff_check_status); | ||
2835 | EXPORT_SYMBOL_GPL(ata_sff_dma_pause); | ||
2836 | EXPORT_SYMBOL_GPL(ata_sff_pause); | ||
2837 | EXPORT_SYMBOL_GPL(ata_sff_busy_sleep); | ||
2838 | EXPORT_SYMBOL_GPL(ata_sff_wait_ready); | ||
2839 | EXPORT_SYMBOL_GPL(ata_sff_tf_load); | ||
2840 | EXPORT_SYMBOL_GPL(ata_sff_tf_read); | ||
2841 | EXPORT_SYMBOL_GPL(ata_sff_exec_command); | ||
2842 | EXPORT_SYMBOL_GPL(ata_sff_data_xfer); | ||
2843 | EXPORT_SYMBOL_GPL(ata_sff_data_xfer_noirq); | ||
2844 | EXPORT_SYMBOL_GPL(ata_sff_irq_on); | ||
2845 | EXPORT_SYMBOL_GPL(ata_sff_irq_clear); | ||
2846 | EXPORT_SYMBOL_GPL(ata_sff_hsm_move); | ||
2847 | EXPORT_SYMBOL_GPL(ata_sff_qc_issue); | ||
2848 | EXPORT_SYMBOL_GPL(ata_sff_qc_fill_rtf); | ||
2849 | EXPORT_SYMBOL_GPL(ata_sff_host_intr); | ||
2850 | EXPORT_SYMBOL_GPL(ata_sff_interrupt); | ||
2851 | EXPORT_SYMBOL_GPL(ata_sff_freeze); | ||
2852 | EXPORT_SYMBOL_GPL(ata_sff_thaw); | ||
2853 | EXPORT_SYMBOL_GPL(ata_sff_prereset); | ||
2854 | EXPORT_SYMBOL_GPL(ata_sff_dev_classify); | ||
2855 | EXPORT_SYMBOL_GPL(ata_sff_wait_after_reset); | ||
2856 | EXPORT_SYMBOL_GPL(ata_sff_softreset); | ||
2857 | EXPORT_SYMBOL_GPL(sata_sff_hardreset); | ||
2858 | EXPORT_SYMBOL_GPL(ata_sff_postreset); | ||
2859 | EXPORT_SYMBOL_GPL(ata_sff_error_handler); | ||
2860 | EXPORT_SYMBOL_GPL(ata_sff_post_internal_cmd); | ||
2861 | EXPORT_SYMBOL_GPL(ata_sff_port_start); | ||
2862 | EXPORT_SYMBOL_GPL(ata_sff_std_ports); | ||
2863 | EXPORT_SYMBOL_GPL(ata_bmdma_mode_filter); | ||
2864 | EXPORT_SYMBOL_GPL(ata_bmdma_setup); | ||
2865 | EXPORT_SYMBOL_GPL(ata_bmdma_start); | ||
2866 | EXPORT_SYMBOL_GPL(ata_bmdma_stop); | ||
2867 | EXPORT_SYMBOL_GPL(ata_bmdma_status); | ||
2868 | EXPORT_SYMBOL_GPL(ata_bus_reset); | ||
2869 | #ifdef CONFIG_PCI | ||
2870 | EXPORT_SYMBOL_GPL(ata_pci_bmdma_clear_simplex); | ||
2871 | EXPORT_SYMBOL_GPL(ata_pci_bmdma_init); | ||
2872 | EXPORT_SYMBOL_GPL(ata_pci_sff_init_host); | ||
2873 | EXPORT_SYMBOL_GPL(ata_pci_sff_prepare_host); | ||
2874 | EXPORT_SYMBOL_GPL(ata_pci_sff_activate_host); | ||
2875 | EXPORT_SYMBOL_GPL(ata_pci_sff_init_one); | ||
2876 | #endif /* CONFIG_PCI */ | ||
diff --git a/drivers/ata/pata_ali.c b/drivers/ata/pata_ali.c index 73c466e452ca..a7999c19f0c9 100644 --- a/drivers/ata/pata_ali.c +++ b/drivers/ata/pata_ali.c | |||
@@ -19,7 +19,9 @@ | |||
19 | * | 19 | * |
20 | * TODO/CHECK | 20 | * TODO/CHECK |
21 | * Cannot have ATAPI on both master & slave for rev < c2 (???) but | 21 | * Cannot have ATAPI on both master & slave for rev < c2 (???) but |
22 | * otherwise should do atapi DMA. | 22 | * otherwise should do atapi DMA (For now for old we do PIO only for |
23 | * ATAPI) | ||
24 | * Review Sunblade workaround. | ||
23 | */ | 25 | */ |
24 | 26 | ||
25 | #include <linux/kernel.h> | 27 | #include <linux/kernel.h> |
@@ -33,12 +35,14 @@ | |||
33 | #include <linux/dmi.h> | 35 | #include <linux/dmi.h> |
34 | 36 | ||
35 | #define DRV_NAME "pata_ali" | 37 | #define DRV_NAME "pata_ali" |
36 | #define DRV_VERSION "0.7.5" | 38 | #define DRV_VERSION "0.7.8" |
37 | 39 | ||
38 | static int ali_atapi_dma = 0; | 40 | static int ali_atapi_dma = 0; |
39 | module_param_named(atapi_dma, ali_atapi_dma, int, 0644); | 41 | module_param_named(atapi_dma, ali_atapi_dma, int, 0644); |
40 | MODULE_PARM_DESC(atapi_dma, "Enable ATAPI DMA (0=disable, 1=enable)"); | 42 | MODULE_PARM_DESC(atapi_dma, "Enable ATAPI DMA (0=disable, 1=enable)"); |
41 | 43 | ||
44 | static struct pci_dev *isa_bridge; | ||
45 | |||
42 | /* | 46 | /* |
43 | * Cable special cases | 47 | * Cable special cases |
44 | */ | 48 | */ |
@@ -147,8 +151,7 @@ static void ali_fifo_control(struct ata_port *ap, struct ata_device *adev, int o | |||
147 | 151 | ||
148 | pci_read_config_byte(pdev, pio_fifo, &fifo); | 152 | pci_read_config_byte(pdev, pio_fifo, &fifo); |
149 | fifo &= ~(0x0F << shift); | 153 | fifo &= ~(0x0F << shift); |
150 | if (on) | 154 | fifo |= (on << shift); |
151 | fifo |= (on << shift); | ||
152 | pci_write_config_byte(pdev, pio_fifo, fifo); | 155 | pci_write_config_byte(pdev, pio_fifo, fifo); |
153 | } | 156 | } |
154 | 157 | ||
@@ -337,6 +340,23 @@ static int ali_check_atapi_dma(struct ata_queued_cmd *qc) | |||
337 | return 0; | 340 | return 0; |
338 | } | 341 | } |
339 | 342 | ||
343 | static void ali_c2_c3_postreset(struct ata_link *link, unsigned int *classes) | ||
344 | { | ||
345 | u8 r; | ||
346 | int port_bit = 4 << link->ap->port_no; | ||
347 | |||
348 | /* If our bridge is an ALI 1533 then do the extra work */ | ||
349 | if (isa_bridge) { | ||
350 | /* Tristate and re-enable the bus signals */ | ||
351 | pci_read_config_byte(isa_bridge, 0x58, &r); | ||
352 | r &= ~port_bit; | ||
353 | pci_write_config_byte(isa_bridge, 0x58, r); | ||
354 | r |= port_bit; | ||
355 | pci_write_config_byte(isa_bridge, 0x58, r); | ||
356 | } | ||
357 | ata_sff_postreset(link, classes); | ||
358 | } | ||
359 | |||
340 | static struct scsi_host_template ali_sht = { | 360 | static struct scsi_host_template ali_sht = { |
341 | ATA_BMDMA_SHT(DRV_NAME), | 361 | ATA_BMDMA_SHT(DRV_NAME), |
342 | }; | 362 | }; |
@@ -349,10 +369,11 @@ static struct ata_port_operations ali_early_port_ops = { | |||
349 | .inherits = &ata_sff_port_ops, | 369 | .inherits = &ata_sff_port_ops, |
350 | .cable_detect = ata_cable_40wire, | 370 | .cable_detect = ata_cable_40wire, |
351 | .set_piomode = ali_set_piomode, | 371 | .set_piomode = ali_set_piomode, |
372 | .sff_data_xfer = ata_sff_data_xfer32, | ||
352 | }; | 373 | }; |
353 | 374 | ||
354 | static const struct ata_port_operations ali_dma_base_ops = { | 375 | static const struct ata_port_operations ali_dma_base_ops = { |
355 | .inherits = &ata_bmdma_port_ops, | 376 | .inherits = &ata_bmdma32_port_ops, |
356 | .set_piomode = ali_set_piomode, | 377 | .set_piomode = ali_set_piomode, |
357 | .set_dmamode = ali_set_dmamode, | 378 | .set_dmamode = ali_set_dmamode, |
358 | }; | 379 | }; |
@@ -377,6 +398,17 @@ static struct ata_port_operations ali_c2_port_ops = { | |||
377 | .check_atapi_dma = ali_check_atapi_dma, | 398 | .check_atapi_dma = ali_check_atapi_dma, |
378 | .cable_detect = ali_c2_cable_detect, | 399 | .cable_detect = ali_c2_cable_detect, |
379 | .dev_config = ali_lock_sectors, | 400 | .dev_config = ali_lock_sectors, |
401 | .postreset = ali_c2_c3_postreset, | ||
402 | }; | ||
403 | |||
404 | /* | ||
405 | * Port operations for DMA capable ALi with cable detect | ||
406 | */ | ||
407 | static struct ata_port_operations ali_c4_port_ops = { | ||
408 | .inherits = &ali_dma_base_ops, | ||
409 | .check_atapi_dma = ali_check_atapi_dma, | ||
410 | .cable_detect = ali_c2_cable_detect, | ||
411 | .dev_config = ali_lock_sectors, | ||
380 | }; | 412 | }; |
381 | 413 | ||
382 | /* | 414 | /* |
@@ -401,52 +433,49 @@ static struct ata_port_operations ali_c5_port_ops = { | |||
401 | static void ali_init_chipset(struct pci_dev *pdev) | 433 | static void ali_init_chipset(struct pci_dev *pdev) |
402 | { | 434 | { |
403 | u8 tmp; | 435 | u8 tmp; |
404 | struct pci_dev *north, *isa_bridge; | 436 | struct pci_dev *north; |
405 | 437 | ||
406 | /* | 438 | /* |
407 | * The chipset revision selects the driver operations and | 439 | * The chipset revision selects the driver operations and |
408 | * mode data. | 440 | * mode data. |
409 | */ | 441 | */ |
410 | 442 | ||
411 | if (pdev->revision >= 0x20 && pdev->revision < 0xC2) { | 443 | if (pdev->revision <= 0x20) { |
412 | /* 1543-E/F, 1543C-C, 1543C-D, 1543C-E */ | 444 | pci_read_config_byte(pdev, 0x53, &tmp); |
413 | pci_read_config_byte(pdev, 0x4B, &tmp); | 445 | tmp |= 0x03; |
414 | /* Clear CD-ROM DMA write bit */ | 446 | pci_write_config_byte(pdev, 0x53, tmp); |
415 | tmp &= 0x7F; | 447 | } else { |
416 | pci_write_config_byte(pdev, 0x4B, tmp); | 448 | pci_read_config_byte(pdev, 0x4a, &tmp); |
417 | } else if (pdev->revision >= 0xC2) { | 449 | pci_write_config_byte(pdev, 0x4a, tmp | 0x20); |
418 | /* Enable cable detection logic */ | ||
419 | pci_read_config_byte(pdev, 0x4B, &tmp); | 450 | pci_read_config_byte(pdev, 0x4B, &tmp); |
420 | pci_write_config_byte(pdev, 0x4B, tmp | 0x08); | 451 | if (pdev->revision < 0xC2) |
421 | } | 452 | /* 1543-E/F, 1543C-C, 1543C-D, 1543C-E */ |
422 | north = pci_get_bus_and_slot(0, PCI_DEVFN(0,0)); | 453 | /* Clear CD-ROM DMA write bit */ |
423 | isa_bridge = pci_get_device(PCI_VENDOR_ID_AL, PCI_DEVICE_ID_AL_M1533, NULL); | 454 | tmp &= 0x7F; |
424 | 455 | /* Cable and UDMA */ | |
425 | if (north && north->vendor == PCI_VENDOR_ID_AL && isa_bridge) { | 456 | pci_write_config_byte(pdev, 0x4B, tmp | 0x09); |
426 | /* Configure the ALi bridge logic. For non ALi rely on BIOS. | ||
427 | Set the south bridge enable bit */ | ||
428 | pci_read_config_byte(isa_bridge, 0x79, &tmp); | ||
429 | if (pdev->revision == 0xC2) | ||
430 | pci_write_config_byte(isa_bridge, 0x79, tmp | 0x04); | ||
431 | else if (pdev->revision > 0xC2 && pdev->revision < 0xC5) | ||
432 | pci_write_config_byte(isa_bridge, 0x79, tmp | 0x02); | ||
433 | } | ||
434 | if (pdev->revision >= 0x20) { | ||
435 | /* | 457 | /* |
436 | * CD_ROM DMA on (0x53 bit 0). Enable this even if we want | 458 | * CD_ROM DMA on (0x53 bit 0). Enable this even if we want |
437 | * to use PIO. 0x53 bit 1 (rev 20 only) - enable FIFO control | 459 | * to use PIO. 0x53 bit 1 (rev 20 only) - enable FIFO control |
438 | * via 0x54/55. | 460 | * via 0x54/55. |
439 | */ | 461 | */ |
440 | pci_read_config_byte(pdev, 0x53, &tmp); | 462 | pci_read_config_byte(pdev, 0x53, &tmp); |
441 | if (pdev->revision <= 0x20) | ||
442 | tmp &= ~0x02; | ||
443 | if (pdev->revision >= 0xc7) | 463 | if (pdev->revision >= 0xc7) |
444 | tmp |= 0x03; | 464 | tmp |= 0x03; |
445 | else | 465 | else |
446 | tmp |= 0x01; /* CD_ROM enable for DMA */ | 466 | tmp |= 0x01; /* CD_ROM enable for DMA */ |
447 | pci_write_config_byte(pdev, 0x53, tmp); | 467 | pci_write_config_byte(pdev, 0x53, tmp); |
448 | } | 468 | } |
449 | pci_dev_put(isa_bridge); | 469 | north = pci_get_bus_and_slot(0, PCI_DEVFN(0,0)); |
470 | if (north && north->vendor == PCI_VENDOR_ID_AL && isa_bridge) { | ||
471 | /* Configure the ALi bridge logic. For non ALi rely on BIOS. | ||
472 | Set the south bridge enable bit */ | ||
473 | pci_read_config_byte(isa_bridge, 0x79, &tmp); | ||
474 | if (pdev->revision == 0xC2) | ||
475 | pci_write_config_byte(isa_bridge, 0x79, tmp | 0x04); | ||
476 | else if (pdev->revision > 0xC2 && pdev->revision < 0xC5) | ||
477 | pci_write_config_byte(isa_bridge, 0x79, tmp | 0x02); | ||
478 | } | ||
450 | pci_dev_put(north); | 479 | pci_dev_put(north); |
451 | ata_pci_bmdma_clear_simplex(pdev); | 480 | ata_pci_bmdma_clear_simplex(pdev); |
452 | } | 481 | } |
@@ -503,7 +532,7 @@ static int ali_init_one(struct pci_dev *pdev, const struct pci_device_id *id) | |||
503 | .pio_mask = 0x1f, | 532 | .pio_mask = 0x1f, |
504 | .mwdma_mask = 0x07, | 533 | .mwdma_mask = 0x07, |
505 | .udma_mask = ATA_UDMA5, | 534 | .udma_mask = ATA_UDMA5, |
506 | .port_ops = &ali_c2_port_ops | 535 | .port_ops = &ali_c4_port_ops |
507 | }; | 536 | }; |
508 | /* Revision 0xC5 is UDMA133 with LBA48 DMA */ | 537 | /* Revision 0xC5 is UDMA133 with LBA48 DMA */ |
509 | static const struct ata_port_info info_c5 = { | 538 | static const struct ata_port_info info_c5 = { |
@@ -516,7 +545,6 @@ static int ali_init_one(struct pci_dev *pdev, const struct pci_device_id *id) | |||
516 | 545 | ||
517 | const struct ata_port_info *ppi[] = { NULL, NULL }; | 546 | const struct ata_port_info *ppi[] = { NULL, NULL }; |
518 | u8 tmp; | 547 | u8 tmp; |
519 | struct pci_dev *isa_bridge; | ||
520 | int rc; | 548 | int rc; |
521 | 549 | ||
522 | rc = pcim_enable_device(pdev); | 550 | rc = pcim_enable_device(pdev); |
@@ -543,14 +571,12 @@ static int ali_init_one(struct pci_dev *pdev, const struct pci_device_id *id) | |||
543 | 571 | ||
544 | ali_init_chipset(pdev); | 572 | ali_init_chipset(pdev); |
545 | 573 | ||
546 | isa_bridge = pci_get_device(PCI_VENDOR_ID_AL, PCI_DEVICE_ID_AL_M1533, NULL); | ||
547 | if (isa_bridge && pdev->revision >= 0x20 && pdev->revision < 0xC2) { | 574 | if (isa_bridge && pdev->revision >= 0x20 && pdev->revision < 0xC2) { |
548 | /* Are we paired with a UDMA capable chip */ | 575 | /* Are we paired with a UDMA capable chip */ |
549 | pci_read_config_byte(isa_bridge, 0x5E, &tmp); | 576 | pci_read_config_byte(isa_bridge, 0x5E, &tmp); |
550 | if ((tmp & 0x1E) == 0x12) | 577 | if ((tmp & 0x1E) == 0x12) |
551 | ppi[0] = &info_20_udma; | 578 | ppi[0] = &info_20_udma; |
552 | } | 579 | } |
553 | pci_dev_put(isa_bridge); | ||
554 | 580 | ||
555 | return ata_pci_sff_init_one(pdev, ppi, &ali_sht, NULL); | 581 | return ata_pci_sff_init_one(pdev, ppi, &ali_sht, NULL); |
556 | } | 582 | } |
@@ -590,13 +616,20 @@ static struct pci_driver ali_pci_driver = { | |||
590 | 616 | ||
591 | static int __init ali_init(void) | 617 | static int __init ali_init(void) |
592 | { | 618 | { |
593 | return pci_register_driver(&ali_pci_driver); | 619 | int ret; |
620 | isa_bridge = pci_get_device(PCI_VENDOR_ID_AL, PCI_DEVICE_ID_AL_M1533, NULL); | ||
621 | |||
622 | ret = pci_register_driver(&ali_pci_driver); | ||
623 | if (ret < 0) | ||
624 | pci_dev_put(isa_bridge); | ||
625 | return ret; | ||
594 | } | 626 | } |
595 | 627 | ||
596 | 628 | ||
597 | static void __exit ali_exit(void) | 629 | static void __exit ali_exit(void) |
598 | { | 630 | { |
599 | pci_unregister_driver(&ali_pci_driver); | 631 | pci_unregister_driver(&ali_pci_driver); |
632 | pci_dev_put(isa_bridge); | ||
600 | } | 633 | } |
601 | 634 | ||
602 | 635 | ||
diff --git a/drivers/ata/pata_amd.c b/drivers/ata/pata_amd.c index 0ec9c7d9fe9d..63719ab9ea44 100644 --- a/drivers/ata/pata_amd.c +++ b/drivers/ata/pata_amd.c | |||
@@ -24,7 +24,7 @@ | |||
24 | #include <linux/libata.h> | 24 | #include <linux/libata.h> |
25 | 25 | ||
26 | #define DRV_NAME "pata_amd" | 26 | #define DRV_NAME "pata_amd" |
27 | #define DRV_VERSION "0.3.10" | 27 | #define DRV_VERSION "0.3.11" |
28 | 28 | ||
29 | /** | 29 | /** |
30 | * timing_setup - shared timing computation and load | 30 | * timing_setup - shared timing computation and load |
@@ -345,7 +345,7 @@ static struct scsi_host_template amd_sht = { | |||
345 | }; | 345 | }; |
346 | 346 | ||
347 | static const struct ata_port_operations amd_base_port_ops = { | 347 | static const struct ata_port_operations amd_base_port_ops = { |
348 | .inherits = &ata_bmdma_port_ops, | 348 | .inherits = &ata_bmdma32_port_ops, |
349 | .prereset = amd_pre_reset, | 349 | .prereset = amd_pre_reset, |
350 | }; | 350 | }; |
351 | 351 | ||
diff --git a/drivers/ata/pata_hpt366.c b/drivers/ata/pata_hpt366.c index e0c4f05d7d57..65c28e5a6cd7 100644 --- a/drivers/ata/pata_hpt366.c +++ b/drivers/ata/pata_hpt366.c | |||
@@ -30,7 +30,7 @@ | |||
30 | #define DRV_VERSION "0.6.2" | 30 | #define DRV_VERSION "0.6.2" |
31 | 31 | ||
32 | struct hpt_clock { | 32 | struct hpt_clock { |
33 | u8 xfer_speed; | 33 | u8 xfer_mode; |
34 | u32 timing; | 34 | u32 timing; |
35 | }; | 35 | }; |
36 | 36 | ||
@@ -189,28 +189,6 @@ static unsigned long hpt366_filter(struct ata_device *adev, unsigned long mask) | |||
189 | return ata_bmdma_mode_filter(adev, mask); | 189 | return ata_bmdma_mode_filter(adev, mask); |
190 | } | 190 | } |
191 | 191 | ||
192 | /** | ||
193 | * hpt36x_find_mode - reset the hpt36x bus | ||
194 | * @ap: ATA port | ||
195 | * @speed: transfer mode | ||
196 | * | ||
197 | * Return the 32bit register programming information for this channel | ||
198 | * that matches the speed provided. | ||
199 | */ | ||
200 | |||
201 | static u32 hpt36x_find_mode(struct ata_port *ap, int speed) | ||
202 | { | ||
203 | struct hpt_clock *clocks = ap->host->private_data; | ||
204 | |||
205 | while(clocks->xfer_speed) { | ||
206 | if (clocks->xfer_speed == speed) | ||
207 | return clocks->timing; | ||
208 | clocks++; | ||
209 | } | ||
210 | BUG(); | ||
211 | return 0xffffffffU; /* silence compiler warning */ | ||
212 | } | ||
213 | |||
214 | static int hpt36x_cable_detect(struct ata_port *ap) | 192 | static int hpt36x_cable_detect(struct ata_port *ap) |
215 | { | 193 | { |
216 | struct pci_dev *pdev = to_pci_dev(ap->host->dev); | 194 | struct pci_dev *pdev = to_pci_dev(ap->host->dev); |
@@ -226,25 +204,16 @@ static int hpt36x_cable_detect(struct ata_port *ap) | |||
226 | return ATA_CBL_PATA80; | 204 | return ATA_CBL_PATA80; |
227 | } | 205 | } |
228 | 206 | ||
229 | /** | 207 | static void hpt366_set_mode(struct ata_port *ap, struct ata_device *adev, |
230 | * hpt366_set_piomode - PIO setup | 208 | u8 mode) |
231 | * @ap: ATA interface | ||
232 | * @adev: device on the interface | ||
233 | * | ||
234 | * Perform PIO mode setup. | ||
235 | */ | ||
236 | |||
237 | static void hpt366_set_piomode(struct ata_port *ap, struct ata_device *adev) | ||
238 | { | 209 | { |
210 | struct hpt_clock *clocks = ap->host->private_data; | ||
239 | struct pci_dev *pdev = to_pci_dev(ap->host->dev); | 211 | struct pci_dev *pdev = to_pci_dev(ap->host->dev); |
240 | u32 addr1, addr2; | 212 | u32 addr1 = 0x40 + 4 * (adev->devno + 2 * ap->port_no); |
241 | u32 reg; | 213 | u32 addr2 = 0x51 + 4 * ap->port_no; |
242 | u32 mode; | 214 | u32 mask, reg; |
243 | u8 fast; | 215 | u8 fast; |
244 | 216 | ||
245 | addr1 = 0x40 + 4 * (adev->devno + 2 * ap->port_no); | ||
246 | addr2 = 0x51 + 4 * ap->port_no; | ||
247 | |||
248 | /* Fast interrupt prediction disable, hold off interrupt disable */ | 217 | /* Fast interrupt prediction disable, hold off interrupt disable */ |
249 | pci_read_config_byte(pdev, addr2, &fast); | 218 | pci_read_config_byte(pdev, addr2, &fast); |
250 | if (fast & 0x80) { | 219 | if (fast & 0x80) { |
@@ -252,12 +221,43 @@ static void hpt366_set_piomode(struct ata_port *ap, struct ata_device *adev) | |||
252 | pci_write_config_byte(pdev, addr2, fast); | 221 | pci_write_config_byte(pdev, addr2, fast); |
253 | } | 222 | } |
254 | 223 | ||
224 | /* determine timing mask and find matching clock entry */ | ||
225 | if (mode < XFER_MW_DMA_0) | ||
226 | mask = 0xc1f8ffff; | ||
227 | else if (mode < XFER_UDMA_0) | ||
228 | mask = 0x303800ff; | ||
229 | else | ||
230 | mask = 0x30070000; | ||
231 | |||
232 | while (clocks->xfer_mode) { | ||
233 | if (clocks->xfer_mode == mode) | ||
234 | break; | ||
235 | clocks++; | ||
236 | } | ||
237 | if (!clocks->xfer_mode) | ||
238 | BUG(); | ||
239 | |||
240 | /* | ||
241 | * Combine new mode bits with old config bits and disable | ||
242 | * on-chip PIO FIFO/buffer (and PIO MST mode as well) to avoid | ||
243 | * problems handling I/O errors later. | ||
244 | */ | ||
255 | pci_read_config_dword(pdev, addr1, ®); | 245 | pci_read_config_dword(pdev, addr1, ®); |
256 | mode = hpt36x_find_mode(ap, adev->pio_mode); | 246 | reg = ((reg & ~mask) | (clocks->timing & mask)) & ~0xc0000000; |
257 | mode &= ~0x8000000; /* No FIFO in PIO */ | 247 | pci_write_config_dword(pdev, addr1, reg); |
258 | mode &= ~0x30070000; /* Leave config bits alone */ | 248 | } |
259 | reg &= 0x30070000; /* Strip timing bits */ | 249 | |
260 | pci_write_config_dword(pdev, addr1, reg | mode); | 250 | /** |
251 | * hpt366_set_piomode - PIO setup | ||
252 | * @ap: ATA interface | ||
253 | * @adev: device on the interface | ||
254 | * | ||
255 | * Perform PIO mode setup. | ||
256 | */ | ||
257 | |||
258 | static void hpt366_set_piomode(struct ata_port *ap, struct ata_device *adev) | ||
259 | { | ||
260 | hpt366_set_mode(ap, adev, adev->pio_mode); | ||
261 | } | 261 | } |
262 | 262 | ||
263 | /** | 263 | /** |
@@ -271,28 +271,7 @@ static void hpt366_set_piomode(struct ata_port *ap, struct ata_device *adev) | |||
271 | 271 | ||
272 | static void hpt366_set_dmamode(struct ata_port *ap, struct ata_device *adev) | 272 | static void hpt366_set_dmamode(struct ata_port *ap, struct ata_device *adev) |
273 | { | 273 | { |
274 | struct pci_dev *pdev = to_pci_dev(ap->host->dev); | 274 | hpt366_set_mode(ap, adev, adev->dma_mode); |
275 | u32 addr1, addr2; | ||
276 | u32 reg; | ||
277 | u32 mode; | ||
278 | u8 fast; | ||
279 | |||
280 | addr1 = 0x40 + 4 * (adev->devno + 2 * ap->port_no); | ||
281 | addr2 = 0x51 + 4 * ap->port_no; | ||
282 | |||
283 | /* Fast interrupt prediction disable, hold off interrupt disable */ | ||
284 | pci_read_config_byte(pdev, addr2, &fast); | ||
285 | if (fast & 0x80) { | ||
286 | fast &= ~0x80; | ||
287 | pci_write_config_byte(pdev, addr2, fast); | ||
288 | } | ||
289 | |||
290 | pci_read_config_dword(pdev, addr1, ®); | ||
291 | mode = hpt36x_find_mode(ap, adev->dma_mode); | ||
292 | mode |= 0x8000000; /* FIFO in MWDMA or UDMA */ | ||
293 | mode &= ~0xC0000000; /* Leave config bits alone */ | ||
294 | reg &= 0xC0000000; /* Strip timing bits */ | ||
295 | pci_write_config_dword(pdev, addr1, reg | mode); | ||
296 | } | 275 | } |
297 | 276 | ||
298 | static struct scsi_host_template hpt36x_sht = { | 277 | static struct scsi_host_template hpt36x_sht = { |
diff --git a/drivers/ata/pata_hpt3x3.c b/drivers/ata/pata_hpt3x3.c index f11a320337c0..f19cc645881a 100644 --- a/drivers/ata/pata_hpt3x3.c +++ b/drivers/ata/pata_hpt3x3.c | |||
@@ -23,7 +23,7 @@ | |||
23 | #include <linux/libata.h> | 23 | #include <linux/libata.h> |
24 | 24 | ||
25 | #define DRV_NAME "pata_hpt3x3" | 25 | #define DRV_NAME "pata_hpt3x3" |
26 | #define DRV_VERSION "0.5.3" | 26 | #define DRV_VERSION "0.6.1" |
27 | 27 | ||
28 | /** | 28 | /** |
29 | * hpt3x3_set_piomode - PIO setup | 29 | * hpt3x3_set_piomode - PIO setup |
@@ -80,14 +80,48 @@ static void hpt3x3_set_dmamode(struct ata_port *ap, struct ata_device *adev) | |||
80 | r2 &= ~(0x11 << dn); /* Clear MWDMA and UDMA bits */ | 80 | r2 &= ~(0x11 << dn); /* Clear MWDMA and UDMA bits */ |
81 | 81 | ||
82 | if (adev->dma_mode >= XFER_UDMA_0) | 82 | if (adev->dma_mode >= XFER_UDMA_0) |
83 | r2 |= (0x10 << dn); /* Ultra mode */ | 83 | r2 |= (0x01 << dn); /* Ultra mode */ |
84 | else | 84 | else |
85 | r2 |= (0x01 << dn); /* MWDMA */ | 85 | r2 |= (0x10 << dn); /* MWDMA */ |
86 | 86 | ||
87 | pci_write_config_dword(pdev, 0x44, r1); | 87 | pci_write_config_dword(pdev, 0x44, r1); |
88 | pci_write_config_dword(pdev, 0x48, r2); | 88 | pci_write_config_dword(pdev, 0x48, r2); |
89 | } | 89 | } |
90 | #endif /* CONFIG_PATA_HPT3X3_DMA */ | 90 | |
91 | /** | ||
92 | * hpt3x3_freeze - DMA workaround | ||
93 | * @ap: port to freeze | ||
94 | * | ||
95 | * When freezing an HPT3x3 we must stop any pending DMA before | ||
96 | * writing to the control register or the chip will hang | ||
97 | */ | ||
98 | |||
99 | static void hpt3x3_freeze(struct ata_port *ap) | ||
100 | { | ||
101 | void __iomem *mmio = ap->ioaddr.bmdma_addr; | ||
102 | |||
103 | iowrite8(ioread8(mmio + ATA_DMA_CMD) & ~ ATA_DMA_START, | ||
104 | mmio + ATA_DMA_CMD); | ||
105 | ata_sff_dma_pause(ap); | ||
106 | ata_sff_freeze(ap); | ||
107 | } | ||
108 | |||
109 | /** | ||
110 | * hpt3x3_bmdma_setup - DMA workaround | ||
111 | * @qc: Queued command | ||
112 | * | ||
113 | * When issuing BMDMA we must clean up the error/active bits in | ||
114 | * software on this device | ||
115 | */ | ||
116 | |||
117 | static void hpt3x3_bmdma_setup(struct ata_queued_cmd *qc) | ||
118 | { | ||
119 | struct ata_port *ap = qc->ap; | ||
120 | u8 r = ioread8(ap->ioaddr.bmdma_addr + ATA_DMA_STATUS); | ||
121 | r |= ATA_DMA_INTR | ATA_DMA_ERR; | ||
122 | iowrite8(r, ap->ioaddr.bmdma_addr + ATA_DMA_STATUS); | ||
123 | return ata_bmdma_setup(qc); | ||
124 | } | ||
91 | 125 | ||
92 | /** | 126 | /** |
93 | * hpt3x3_atapi_dma - ATAPI DMA check | 127 | * hpt3x3_atapi_dma - ATAPI DMA check |
@@ -101,18 +135,23 @@ static int hpt3x3_atapi_dma(struct ata_queued_cmd *qc) | |||
101 | return 1; | 135 | return 1; |
102 | } | 136 | } |
103 | 137 | ||
138 | #endif /* CONFIG_PATA_HPT3X3_DMA */ | ||
139 | |||
104 | static struct scsi_host_template hpt3x3_sht = { | 140 | static struct scsi_host_template hpt3x3_sht = { |
105 | ATA_BMDMA_SHT(DRV_NAME), | 141 | ATA_BMDMA_SHT(DRV_NAME), |
106 | }; | 142 | }; |
107 | 143 | ||
108 | static struct ata_port_operations hpt3x3_port_ops = { | 144 | static struct ata_port_operations hpt3x3_port_ops = { |
109 | .inherits = &ata_bmdma_port_ops, | 145 | .inherits = &ata_bmdma_port_ops, |
110 | .check_atapi_dma= hpt3x3_atapi_dma, | ||
111 | .cable_detect = ata_cable_40wire, | 146 | .cable_detect = ata_cable_40wire, |
112 | .set_piomode = hpt3x3_set_piomode, | 147 | .set_piomode = hpt3x3_set_piomode, |
113 | #if defined(CONFIG_PATA_HPT3X3_DMA) | 148 | #if defined(CONFIG_PATA_HPT3X3_DMA) |
114 | .set_dmamode = hpt3x3_set_dmamode, | 149 | .set_dmamode = hpt3x3_set_dmamode, |
150 | .bmdma_setup = hpt3x3_bmdma_setup, | ||
151 | .check_atapi_dma= hpt3x3_atapi_dma, | ||
152 | .freeze = hpt3x3_freeze, | ||
115 | #endif | 153 | #endif |
154 | |||
116 | }; | 155 | }; |
117 | 156 | ||
118 | /** | 157 | /** |
diff --git a/drivers/ata/pata_mpiix.c b/drivers/ata/pata_mpiix.c index 7c8faa48b5f3..aa576cac4d17 100644 --- a/drivers/ata/pata_mpiix.c +++ b/drivers/ata/pata_mpiix.c | |||
@@ -35,7 +35,7 @@ | |||
35 | #include <linux/libata.h> | 35 | #include <linux/libata.h> |
36 | 36 | ||
37 | #define DRV_NAME "pata_mpiix" | 37 | #define DRV_NAME "pata_mpiix" |
38 | #define DRV_VERSION "0.7.6" | 38 | #define DRV_VERSION "0.7.7" |
39 | 39 | ||
40 | enum { | 40 | enum { |
41 | IDETIM = 0x6C, /* IDE control register */ | 41 | IDETIM = 0x6C, /* IDE control register */ |
@@ -146,6 +146,7 @@ static struct ata_port_operations mpiix_port_ops = { | |||
146 | .cable_detect = ata_cable_40wire, | 146 | .cable_detect = ata_cable_40wire, |
147 | .set_piomode = mpiix_set_piomode, | 147 | .set_piomode = mpiix_set_piomode, |
148 | .prereset = mpiix_pre_reset, | 148 | .prereset = mpiix_pre_reset, |
149 | .sff_data_xfer = ata_sff_data_xfer32, | ||
149 | }; | 150 | }; |
150 | 151 | ||
151 | static int mpiix_init_one(struct pci_dev *dev, const struct pci_device_id *id) | 152 | static int mpiix_init_one(struct pci_dev *dev, const struct pci_device_id *id) |
diff --git a/drivers/ata/pata_platform.c b/drivers/ata/pata_platform.c index 6afa07a37648..d8d743af3225 100644 --- a/drivers/ata/pata_platform.c +++ b/drivers/ata/pata_platform.c | |||
@@ -186,7 +186,7 @@ EXPORT_SYMBOL_GPL(__pata_platform_probe); | |||
186 | * A platform bus ATA device has been unplugged. Perform the needed | 186 | * A platform bus ATA device has been unplugged. Perform the needed |
187 | * cleanup. Also called on module unload for any active devices. | 187 | * cleanup. Also called on module unload for any active devices. |
188 | */ | 188 | */ |
189 | int __devexit __pata_platform_remove(struct device *dev) | 189 | int __pata_platform_remove(struct device *dev) |
190 | { | 190 | { |
191 | struct ata_host *host = dev_get_drvdata(dev); | 191 | struct ata_host *host = dev_get_drvdata(dev); |
192 | 192 | ||
diff --git a/drivers/ata/pata_sil680.c b/drivers/ata/pata_sil680.c index 83580a59db58..9e764e5747e6 100644 --- a/drivers/ata/pata_sil680.c +++ b/drivers/ata/pata_sil680.c | |||
@@ -32,7 +32,7 @@ | |||
32 | #include <linux/libata.h> | 32 | #include <linux/libata.h> |
33 | 33 | ||
34 | #define DRV_NAME "pata_sil680" | 34 | #define DRV_NAME "pata_sil680" |
35 | #define DRV_VERSION "0.4.8" | 35 | #define DRV_VERSION "0.4.9" |
36 | 36 | ||
37 | #define SIL680_MMIO_BAR 5 | 37 | #define SIL680_MMIO_BAR 5 |
38 | 38 | ||
@@ -195,7 +195,7 @@ static struct scsi_host_template sil680_sht = { | |||
195 | }; | 195 | }; |
196 | 196 | ||
197 | static struct ata_port_operations sil680_port_ops = { | 197 | static struct ata_port_operations sil680_port_ops = { |
198 | .inherits = &ata_bmdma_port_ops, | 198 | .inherits = &ata_bmdma32_port_ops, |
199 | .cable_detect = sil680_cable_detect, | 199 | .cable_detect = sil680_cable_detect, |
200 | .set_piomode = sil680_set_piomode, | 200 | .set_piomode = sil680_set_piomode, |
201 | .set_dmamode = sil680_set_dmamode, | 201 | .set_dmamode = sil680_set_dmamode, |
diff --git a/drivers/ata/sata_sil24.c b/drivers/ata/sata_sil24.c index ccee930f1e12..2590c2279fa7 100644 --- a/drivers/ata/sata_sil24.c +++ b/drivers/ata/sata_sil24.c | |||
@@ -51,13 +51,6 @@ struct sil24_sge { | |||
51 | __le32 flags; | 51 | __le32 flags; |
52 | }; | 52 | }; |
53 | 53 | ||
54 | /* | ||
55 | * Port multiplier | ||
56 | */ | ||
57 | struct sil24_port_multiplier { | ||
58 | __le32 diag; | ||
59 | __le32 sactive; | ||
60 | }; | ||
61 | 54 | ||
62 | enum { | 55 | enum { |
63 | SIL24_HOST_BAR = 0, | 56 | SIL24_HOST_BAR = 0, |
diff --git a/drivers/atm/iphase.c b/drivers/atm/iphase.c index 088885ed51b9..e1c7611e9144 100644 --- a/drivers/atm/iphase.c +++ b/drivers/atm/iphase.c | |||
@@ -64,7 +64,7 @@ | |||
64 | #include <linux/jiffies.h> | 64 | #include <linux/jiffies.h> |
65 | #include "iphase.h" | 65 | #include "iphase.h" |
66 | #include "suni.h" | 66 | #include "suni.h" |
67 | #define swap(x) (((x & 0xff) << 8) | ((x & 0xff00) >> 8)) | 67 | #define swap_byte_order(x) (((x & 0xff) << 8) | ((x & 0xff00) >> 8)) |
68 | 68 | ||
69 | #define PRIV(dev) ((struct suni_priv *) dev->phy_data) | 69 | #define PRIV(dev) ((struct suni_priv *) dev->phy_data) |
70 | 70 | ||
@@ -1306,7 +1306,7 @@ static void rx_dle_intr(struct atm_dev *dev) | |||
1306 | // get real pkt length pwang_test | 1306 | // get real pkt length pwang_test |
1307 | trailer = (struct cpcs_trailer*)((u_char *)skb->data + | 1307 | trailer = (struct cpcs_trailer*)((u_char *)skb->data + |
1308 | skb->len - sizeof(*trailer)); | 1308 | skb->len - sizeof(*trailer)); |
1309 | length = swap(trailer->length); | 1309 | length = swap_byte_order(trailer->length); |
1310 | if ((length > iadev->rx_buf_sz) || (length > | 1310 | if ((length > iadev->rx_buf_sz) || (length > |
1311 | (skb->len - sizeof(struct cpcs_trailer)))) | 1311 | (skb->len - sizeof(struct cpcs_trailer)))) |
1312 | { | 1312 | { |
@@ -2995,7 +2995,7 @@ static int ia_pkt_tx (struct atm_vcc *vcc, struct sk_buff *skb) { | |||
2995 | skb->len, PCI_DMA_TODEVICE); | 2995 | skb->len, PCI_DMA_TODEVICE); |
2996 | wr_ptr->local_pkt_addr = (buf_desc_ptr->buf_start_hi << 16) | | 2996 | wr_ptr->local_pkt_addr = (buf_desc_ptr->buf_start_hi << 16) | |
2997 | buf_desc_ptr->buf_start_lo; | 2997 | buf_desc_ptr->buf_start_lo; |
2998 | /* wr_ptr->bytes = swap(total_len); didn't seem to affect ?? */ | 2998 | /* wr_ptr->bytes = swap_byte_order(total_len); didn't seem to affect?? */ |
2999 | wr_ptr->bytes = skb->len; | 2999 | wr_ptr->bytes = skb->len; |
3000 | 3000 | ||
3001 | /* hw bug - DLEs of 0x2d, 0x2e, 0x2f cause DMA lockup */ | 3001 | /* hw bug - DLEs of 0x2d, 0x2e, 0x2f cause DMA lockup */ |
diff --git a/drivers/char/Kconfig b/drivers/char/Kconfig index 35914b6e1d2a..f5be8081cd81 100644 --- a/drivers/char/Kconfig +++ b/drivers/char/Kconfig | |||
@@ -616,6 +616,7 @@ config HVC_ISERIES | |||
616 | default y | 616 | default y |
617 | select HVC_DRIVER | 617 | select HVC_DRIVER |
618 | select HVC_IRQ | 618 | select HVC_IRQ |
619 | select VIOPATH | ||
619 | help | 620 | help |
620 | iSeries machines support a hypervisor virtual console. | 621 | iSeries machines support a hypervisor virtual console. |
621 | 622 | ||
diff --git a/drivers/char/hvc_beat.c b/drivers/char/hvc_beat.c index 91cdb35a9204..0afc8b82212e 100644 --- a/drivers/char/hvc_beat.c +++ b/drivers/char/hvc_beat.c | |||
@@ -44,7 +44,7 @@ static int hvc_beat_get_chars(uint32_t vtermno, char *buf, int cnt) | |||
44 | static unsigned char q[sizeof(unsigned long) * 2] | 44 | static unsigned char q[sizeof(unsigned long) * 2] |
45 | __attribute__((aligned(sizeof(unsigned long)))); | 45 | __attribute__((aligned(sizeof(unsigned long)))); |
46 | static int qlen = 0; | 46 | static int qlen = 0; |
47 | unsigned long got; | 47 | u64 got; |
48 | 48 | ||
49 | again: | 49 | again: |
50 | if (qlen) { | 50 | if (qlen) { |
@@ -63,7 +63,7 @@ again: | |||
63 | } | 63 | } |
64 | } | 64 | } |
65 | if (beat_get_term_char(vtermno, &got, | 65 | if (beat_get_term_char(vtermno, &got, |
66 | ((unsigned long *)q), ((unsigned long *)q) + 1) == 0) { | 66 | ((u64 *)q), ((u64 *)q) + 1) == 0) { |
67 | qlen = got; | 67 | qlen = got; |
68 | goto again; | 68 | goto again; |
69 | } | 69 | } |
diff --git a/drivers/char/pty.c b/drivers/char/pty.c index 112a6ba9a96f..146c97613da0 100644 --- a/drivers/char/pty.c +++ b/drivers/char/pty.c | |||
@@ -32,7 +32,7 @@ | |||
32 | 32 | ||
33 | /* These are global because they are accessed in tty_io.c */ | 33 | /* These are global because they are accessed in tty_io.c */ |
34 | #ifdef CONFIG_UNIX98_PTYS | 34 | #ifdef CONFIG_UNIX98_PTYS |
35 | struct tty_driver *ptm_driver; | 35 | static struct tty_driver *ptm_driver; |
36 | static struct tty_driver *pts_driver; | 36 | static struct tty_driver *pts_driver; |
37 | #endif | 37 | #endif |
38 | 38 | ||
diff --git a/drivers/char/tpm/tpm_nsc.c b/drivers/char/tpm/tpm_nsc.c index ab18c1e7b115..70efba2ee053 100644 --- a/drivers/char/tpm/tpm_nsc.c +++ b/drivers/char/tpm/tpm_nsc.c | |||
@@ -273,12 +273,23 @@ static void tpm_nsc_remove(struct device *dev) | |||
273 | } | 273 | } |
274 | } | 274 | } |
275 | 275 | ||
276 | static struct device_driver nsc_drv = { | 276 | static int tpm_nsc_suspend(struct platform_device *dev, pm_message_t msg) |
277 | .name = "tpm_nsc", | 277 | { |
278 | .bus = &platform_bus_type, | 278 | return tpm_pm_suspend(&dev->dev, msg); |
279 | .owner = THIS_MODULE, | 279 | } |
280 | .suspend = tpm_pm_suspend, | 280 | |
281 | .resume = tpm_pm_resume, | 281 | static int tpm_nsc_resume(struct platform_device *dev) |
282 | { | ||
283 | return tpm_pm_resume(&dev->dev); | ||
284 | } | ||
285 | |||
286 | static struct platform_driver nsc_drv = { | ||
287 | .suspend = tpm_nsc_suspend, | ||
288 | .resume = tpm_nsc_resume, | ||
289 | .driver = { | ||
290 | .name = "tpm_nsc", | ||
291 | .owner = THIS_MODULE, | ||
292 | }, | ||
282 | }; | 293 | }; |
283 | 294 | ||
284 | static int __init init_nsc(void) | 295 | static int __init init_nsc(void) |
@@ -297,7 +308,7 @@ static int __init init_nsc(void) | |||
297 | return -ENODEV; | 308 | return -ENODEV; |
298 | } | 309 | } |
299 | 310 | ||
300 | err = driver_register(&nsc_drv); | 311 | err = platform_driver_register(&nsc_drv); |
301 | if (err) | 312 | if (err) |
302 | return err; | 313 | return err; |
303 | 314 | ||
@@ -308,17 +319,15 @@ static int __init init_nsc(void) | |||
308 | /* enable the DPM module */ | 319 | /* enable the DPM module */ |
309 | tpm_write_index(nscAddrBase, NSC_LDC_INDEX, 0x01); | 320 | tpm_write_index(nscAddrBase, NSC_LDC_INDEX, 0x01); |
310 | 321 | ||
311 | pdev = kzalloc(sizeof(struct platform_device), GFP_KERNEL); | 322 | pdev = platform_device_alloc("tpm_nscl0", -1); |
312 | if (!pdev) { | 323 | if (!pdev) { |
313 | rc = -ENOMEM; | 324 | rc = -ENOMEM; |
314 | goto err_unreg_drv; | 325 | goto err_unreg_drv; |
315 | } | 326 | } |
316 | 327 | ||
317 | pdev->name = "tpm_nscl0"; | ||
318 | pdev->id = -1; | ||
319 | pdev->num_resources = 0; | 328 | pdev->num_resources = 0; |
329 | pdev->dev.driver = &nsc_drv.driver; | ||
320 | pdev->dev.release = tpm_nsc_remove; | 330 | pdev->dev.release = tpm_nsc_remove; |
321 | pdev->dev.driver = &nsc_drv; | ||
322 | 331 | ||
323 | if ((rc = platform_device_register(pdev)) < 0) | 332 | if ((rc = platform_device_register(pdev)) < 0) |
324 | goto err_free_dev; | 333 | goto err_free_dev; |
@@ -377,7 +386,7 @@ err_unreg_dev: | |||
377 | err_free_dev: | 386 | err_free_dev: |
378 | kfree(pdev); | 387 | kfree(pdev); |
379 | err_unreg_drv: | 388 | err_unreg_drv: |
380 | driver_unregister(&nsc_drv); | 389 | platform_driver_unregister(&nsc_drv); |
381 | return rc; | 390 | return rc; |
382 | } | 391 | } |
383 | 392 | ||
@@ -390,7 +399,7 @@ static void __exit cleanup_nsc(void) | |||
390 | pdev = NULL; | 399 | pdev = NULL; |
391 | } | 400 | } |
392 | 401 | ||
393 | driver_unregister(&nsc_drv); | 402 | platform_driver_unregister(&nsc_drv); |
394 | } | 403 | } |
395 | 404 | ||
396 | module_init(init_nsc); | 405 | module_init(init_nsc); |
diff --git a/drivers/char/vt.c b/drivers/char/vt.c index 80014213fb53..7900bd63b36d 100644 --- a/drivers/char/vt.c +++ b/drivers/char/vt.c | |||
@@ -969,8 +969,7 @@ int vc_resize(struct vc_data *vc, unsigned int cols, unsigned int rows) | |||
969 | * Takes the console sem and the called methods then take the tty | 969 | * Takes the console sem and the called methods then take the tty |
970 | * termios_mutex and the tty ctrl_lock in that order. | 970 | * termios_mutex and the tty ctrl_lock in that order. |
971 | */ | 971 | */ |
972 | 972 | static int vt_resize(struct tty_struct *tty, struct winsize *ws) | |
973 | int vt_resize(struct tty_struct *tty, struct winsize *ws) | ||
974 | { | 973 | { |
975 | struct vc_data *vc = tty->driver_data; | 974 | struct vc_data *vc = tty->driver_data; |
976 | int ret; | 975 | int ret; |
diff --git a/drivers/firmware/dcdbas.c b/drivers/firmware/dcdbas.c index 50a071f1c945..777fba48d2d3 100644 --- a/drivers/firmware/dcdbas.c +++ b/drivers/firmware/dcdbas.c | |||
@@ -238,11 +238,11 @@ static ssize_t host_control_on_shutdown_store(struct device *dev, | |||
238 | } | 238 | } |
239 | 239 | ||
240 | /** | 240 | /** |
241 | * smi_request: generate SMI request | 241 | * dcdbas_smi_request: generate SMI request |
242 | * | 242 | * |
243 | * Called with smi_data_lock. | 243 | * Called with smi_data_lock. |
244 | */ | 244 | */ |
245 | static int smi_request(struct smi_cmd *smi_cmd) | 245 | int dcdbas_smi_request(struct smi_cmd *smi_cmd) |
246 | { | 246 | { |
247 | cpumask_t old_mask; | 247 | cpumask_t old_mask; |
248 | int ret = 0; | 248 | int ret = 0; |
@@ -309,14 +309,14 @@ static ssize_t smi_request_store(struct device *dev, | |||
309 | switch (val) { | 309 | switch (val) { |
310 | case 2: | 310 | case 2: |
311 | /* Raw SMI */ | 311 | /* Raw SMI */ |
312 | ret = smi_request(smi_cmd); | 312 | ret = dcdbas_smi_request(smi_cmd); |
313 | if (!ret) | 313 | if (!ret) |
314 | ret = count; | 314 | ret = count; |
315 | break; | 315 | break; |
316 | case 1: | 316 | case 1: |
317 | /* Calling Interface SMI */ | 317 | /* Calling Interface SMI */ |
318 | smi_cmd->ebx = (u32) virt_to_phys(smi_cmd->command_buffer); | 318 | smi_cmd->ebx = (u32) virt_to_phys(smi_cmd->command_buffer); |
319 | ret = smi_request(smi_cmd); | 319 | ret = dcdbas_smi_request(smi_cmd); |
320 | if (!ret) | 320 | if (!ret) |
321 | ret = count; | 321 | ret = count; |
322 | break; | 322 | break; |
@@ -333,6 +333,7 @@ out: | |||
333 | mutex_unlock(&smi_data_lock); | 333 | mutex_unlock(&smi_data_lock); |
334 | return ret; | 334 | return ret; |
335 | } | 335 | } |
336 | EXPORT_SYMBOL(dcdbas_smi_request); | ||
336 | 337 | ||
337 | /** | 338 | /** |
338 | * host_control_smi: generate host control SMI | 339 | * host_control_smi: generate host control SMI |
diff --git a/drivers/firmware/dcdbas.h b/drivers/firmware/dcdbas.h index 87bc3417de27..ca3cb0a54ab6 100644 --- a/drivers/firmware/dcdbas.h +++ b/drivers/firmware/dcdbas.h | |||
@@ -101,5 +101,7 @@ struct apm_cmd { | |||
101 | } __attribute__ ((packed)) parameters; | 101 | } __attribute__ ((packed)) parameters; |
102 | } __attribute__ ((packed)); | 102 | } __attribute__ ((packed)); |
103 | 103 | ||
104 | int dcdbas_smi_request(struct smi_cmd *smi_cmd); | ||
105 | |||
104 | #endif /* _DCDBAS_H_ */ | 106 | #endif /* _DCDBAS_H_ */ |
105 | 107 | ||
diff --git a/drivers/firmware/memmap.c b/drivers/firmware/memmap.c index 3bf8ee120d42..261b9aa3f248 100644 --- a/drivers/firmware/memmap.c +++ b/drivers/firmware/memmap.c | |||
@@ -56,9 +56,9 @@ struct memmap_attribute { | |||
56 | ssize_t (*show)(struct firmware_map_entry *entry, char *buf); | 56 | ssize_t (*show)(struct firmware_map_entry *entry, char *buf); |
57 | }; | 57 | }; |
58 | 58 | ||
59 | struct memmap_attribute memmap_start_attr = __ATTR_RO(start); | 59 | static struct memmap_attribute memmap_start_attr = __ATTR_RO(start); |
60 | struct memmap_attribute memmap_end_attr = __ATTR_RO(end); | 60 | static struct memmap_attribute memmap_end_attr = __ATTR_RO(end); |
61 | struct memmap_attribute memmap_type_attr = __ATTR_RO(type); | 61 | static struct memmap_attribute memmap_type_attr = __ATTR_RO(type); |
62 | 62 | ||
63 | /* | 63 | /* |
64 | * These are default attributes that are added for every memmap entry. | 64 | * These are default attributes that are added for every memmap entry. |
diff --git a/drivers/infiniband/hw/nes/nes_cm.c b/drivers/infiniband/hw/nes/nes_cm.c index a812db243477..6ba57e91d7ab 100644 --- a/drivers/infiniband/hw/nes/nes_cm.c +++ b/drivers/infiniband/hw/nes/nes_cm.c | |||
@@ -2705,7 +2705,7 @@ int nes_accept(struct iw_cm_id *cm_id, struct iw_cm_conn_param *conn_param) | |||
2705 | sizeof(struct ietf_mpa_frame)); | 2705 | sizeof(struct ietf_mpa_frame)); |
2706 | 2706 | ||
2707 | 2707 | ||
2708 | /* notify OF layer that accept event was successfull */ | 2708 | /* notify OF layer that accept event was successful */ |
2709 | cm_id->add_ref(cm_id); | 2709 | cm_id->add_ref(cm_id); |
2710 | 2710 | ||
2711 | cm_event.event = IW_CM_EVENT_ESTABLISHED; | 2711 | cm_event.event = IW_CM_EVENT_ESTABLISHED; |
diff --git a/drivers/input/mouse/pxa930_trkball.c b/drivers/input/mouse/pxa930_trkball.c index a0f45c4fc198..d297accf9a7f 100644 --- a/drivers/input/mouse/pxa930_trkball.c +++ b/drivers/input/mouse/pxa930_trkball.c | |||
@@ -186,7 +186,7 @@ static int __devinit pxa930_trkball_probe(struct platform_device *pdev) | |||
186 | error = request_irq(irq, pxa930_trkball_interrupt, IRQF_DISABLED, | 186 | error = request_irq(irq, pxa930_trkball_interrupt, IRQF_DISABLED, |
187 | pdev->name, trkball); | 187 | pdev->name, trkball); |
188 | if (error) { | 188 | if (error) { |
189 | dev_err(&pdev->dev, "failed to request irq: %d\n", ret); | 189 | dev_err(&pdev->dev, "failed to request irq: %d\n", error); |
190 | goto failed_free_io; | 190 | goto failed_free_io; |
191 | } | 191 | } |
192 | 192 | ||
@@ -227,7 +227,7 @@ failed_free_io: | |||
227 | iounmap(trkball->mmio_base); | 227 | iounmap(trkball->mmio_base); |
228 | failed: | 228 | failed: |
229 | kfree(trkball); | 229 | kfree(trkball); |
230 | return ret; | 230 | return error; |
231 | } | 231 | } |
232 | 232 | ||
233 | static int __devexit pxa930_trkball_remove(struct platform_device *pdev) | 233 | static int __devexit pxa930_trkball_remove(struct platform_device *pdev) |
diff --git a/drivers/isdn/hardware/eicon/debuglib.h b/drivers/isdn/hardware/eicon/debuglib.h index 016410cf2273..8ea587783e14 100644 --- a/drivers/isdn/hardware/eicon/debuglib.h +++ b/drivers/isdn/hardware/eicon/debuglib.h | |||
@@ -235,7 +235,7 @@ typedef void ( * DbgOld) (unsigned short, char *, va_list) ; | |||
235 | typedef void ( * DbgEv) (unsigned short, unsigned long, va_list) ; | 235 | typedef void ( * DbgEv) (unsigned short, unsigned long, va_list) ; |
236 | typedef void ( * DbgIrq) (unsigned short, int, char *, va_list) ; | 236 | typedef void ( * DbgIrq) (unsigned short, int, char *, va_list) ; |
237 | typedef struct _DbgHandle_ | 237 | typedef struct _DbgHandle_ |
238 | { char Registered ; /* driver successfull registered */ | 238 | { char Registered ; /* driver successfully registered */ |
239 | #define DBG_HANDLE_REG_NEW 0x01 /* this (new) structure */ | 239 | #define DBG_HANDLE_REG_NEW 0x01 /* this (new) structure */ |
240 | #define DBG_HANDLE_REG_OLD 0x7f /* old structure (see below) */ | 240 | #define DBG_HANDLE_REG_OLD 0x7f /* old structure (see below) */ |
241 | char Version; /* version of this structure */ | 241 | char Version; /* version of this structure */ |
diff --git a/drivers/isdn/hardware/eicon/os_4bri.c b/drivers/isdn/hardware/eicon/os_4bri.c index 7b4ec3f60dbf..c964b8d91ada 100644 --- a/drivers/isdn/hardware/eicon/os_4bri.c +++ b/drivers/isdn/hardware/eicon/os_4bri.c | |||
@@ -997,7 +997,7 @@ diva_4bri_start_adapter(PISDN_ADAPTER IoAdapter, | |||
997 | diva_xdi_display_adapter_features(IoAdapter->ANum); | 997 | diva_xdi_display_adapter_features(IoAdapter->ANum); |
998 | 998 | ||
999 | for (i = 0; i < IoAdapter->tasks; i++) { | 999 | for (i = 0; i < IoAdapter->tasks; i++) { |
1000 | DBG_LOG(("A(%d) %s adapter successfull started", | 1000 | DBG_LOG(("A(%d) %s adapter successfully started", |
1001 | IoAdapter->QuadroList->QuadroAdapter[i]->ANum, | 1001 | IoAdapter->QuadroList->QuadroAdapter[i]->ANum, |
1002 | (IoAdapter->tasks == 1) ? "BRI 2.0" : "4BRI")) | 1002 | (IoAdapter->tasks == 1) ? "BRI 2.0" : "4BRI")) |
1003 | diva_xdi_didd_register_adapter(IoAdapter->QuadroList->QuadroAdapter[i]->ANum); | 1003 | diva_xdi_didd_register_adapter(IoAdapter->QuadroList->QuadroAdapter[i]->ANum); |
diff --git a/drivers/isdn/hardware/eicon/os_bri.c b/drivers/isdn/hardware/eicon/os_bri.c index f31bba5b16ff..08f01993f46b 100644 --- a/drivers/isdn/hardware/eicon/os_bri.c +++ b/drivers/isdn/hardware/eicon/os_bri.c | |||
@@ -736,7 +736,7 @@ diva_bri_start_adapter(PISDN_ADAPTER IoAdapter, | |||
736 | 736 | ||
737 | IoAdapter->Properties.Features = (word) features; | 737 | IoAdapter->Properties.Features = (word) features; |
738 | diva_xdi_display_adapter_features(IoAdapter->ANum); | 738 | diva_xdi_display_adapter_features(IoAdapter->ANum); |
739 | DBG_LOG(("A(%d) BRI adapter successfull started", IoAdapter->ANum)) | 739 | DBG_LOG(("A(%d) BRI adapter successfully started", IoAdapter->ANum)) |
740 | /* | 740 | /* |
741 | Register with DIDD | 741 | Register with DIDD |
742 | */ | 742 | */ |
diff --git a/drivers/isdn/hardware/eicon/os_pri.c b/drivers/isdn/hardware/eicon/os_pri.c index 903356547b79..5d65405c75f4 100644 --- a/drivers/isdn/hardware/eicon/os_pri.c +++ b/drivers/isdn/hardware/eicon/os_pri.c | |||
@@ -513,7 +513,7 @@ diva_pri_start_adapter(PISDN_ADAPTER IoAdapter, | |||
513 | 513 | ||
514 | diva_xdi_display_adapter_features(IoAdapter->ANum); | 514 | diva_xdi_display_adapter_features(IoAdapter->ANum); |
515 | 515 | ||
516 | DBG_LOG(("A(%d) PRI adapter successfull started", IoAdapter->ANum)) | 516 | DBG_LOG(("A(%d) PRI adapter successfully started", IoAdapter->ANum)) |
517 | /* | 517 | /* |
518 | Register with DIDD | 518 | Register with DIDD |
519 | */ | 519 | */ |
diff --git a/drivers/md/bitmap.c b/drivers/md/bitmap.c index ab7c8e4a61f9..719943763391 100644 --- a/drivers/md/bitmap.c +++ b/drivers/md/bitmap.c | |||
@@ -215,7 +215,6 @@ static struct page *read_sb_page(mddev_t *mddev, long offset, | |||
215 | /* choose a good rdev and read the page from there */ | 215 | /* choose a good rdev and read the page from there */ |
216 | 216 | ||
217 | mdk_rdev_t *rdev; | 217 | mdk_rdev_t *rdev; |
218 | struct list_head *tmp; | ||
219 | sector_t target; | 218 | sector_t target; |
220 | 219 | ||
221 | if (!page) | 220 | if (!page) |
@@ -223,7 +222,7 @@ static struct page *read_sb_page(mddev_t *mddev, long offset, | |||
223 | if (!page) | 222 | if (!page) |
224 | return ERR_PTR(-ENOMEM); | 223 | return ERR_PTR(-ENOMEM); |
225 | 224 | ||
226 | rdev_for_each(rdev, tmp, mddev) { | 225 | list_for_each_entry(rdev, &mddev->disks, same_set) { |
227 | if (! test_bit(In_sync, &rdev->flags) | 226 | if (! test_bit(In_sync, &rdev->flags) |
228 | || test_bit(Faulty, &rdev->flags)) | 227 | || test_bit(Faulty, &rdev->flags)) |
229 | continue; | 228 | continue; |
@@ -964,9 +963,11 @@ static int bitmap_init_from_disk(struct bitmap *bitmap, sector_t start) | |||
964 | */ | 963 | */ |
965 | page = bitmap->sb_page; | 964 | page = bitmap->sb_page; |
966 | offset = sizeof(bitmap_super_t); | 965 | offset = sizeof(bitmap_super_t); |
967 | read_sb_page(bitmap->mddev, bitmap->offset, | 966 | if (!file) |
968 | page, | 967 | read_sb_page(bitmap->mddev, |
969 | index, count); | 968 | bitmap->offset, |
969 | page, | ||
970 | index, count); | ||
970 | } else if (file) { | 971 | } else if (file) { |
971 | page = read_page(file, index, bitmap, count); | 972 | page = read_page(file, index, bitmap, count); |
972 | offset = 0; | 973 | offset = 0; |
diff --git a/drivers/md/faulty.c b/drivers/md/faulty.c index f26c1f9a475b..86d9adf90e79 100644 --- a/drivers/md/faulty.c +++ b/drivers/md/faulty.c | |||
@@ -283,7 +283,6 @@ static int reconfig(mddev_t *mddev, int layout, int chunk_size) | |||
283 | static int run(mddev_t *mddev) | 283 | static int run(mddev_t *mddev) |
284 | { | 284 | { |
285 | mdk_rdev_t *rdev; | 285 | mdk_rdev_t *rdev; |
286 | struct list_head *tmp; | ||
287 | int i; | 286 | int i; |
288 | 287 | ||
289 | conf_t *conf = kmalloc(sizeof(*conf), GFP_KERNEL); | 288 | conf_t *conf = kmalloc(sizeof(*conf), GFP_KERNEL); |
@@ -296,7 +295,7 @@ static int run(mddev_t *mddev) | |||
296 | } | 295 | } |
297 | conf->nfaults = 0; | 296 | conf->nfaults = 0; |
298 | 297 | ||
299 | rdev_for_each(rdev, tmp, mddev) | 298 | list_for_each_entry(rdev, &mddev->disks, same_set) |
300 | conf->rdev = rdev; | 299 | conf->rdev = rdev; |
301 | 300 | ||
302 | mddev->array_sectors = mddev->size * 2; | 301 | mddev->array_sectors = mddev->size * 2; |
diff --git a/drivers/md/linear.c b/drivers/md/linear.c index 3b90c5c924ec..1e3aea9eecf1 100644 --- a/drivers/md/linear.c +++ b/drivers/md/linear.c | |||
@@ -105,7 +105,6 @@ static linear_conf_t *linear_conf(mddev_t *mddev, int raid_disks) | |||
105 | int i, nb_zone, cnt; | 105 | int i, nb_zone, cnt; |
106 | sector_t min_sectors; | 106 | sector_t min_sectors; |
107 | sector_t curr_sector; | 107 | sector_t curr_sector; |
108 | struct list_head *tmp; | ||
109 | 108 | ||
110 | conf = kzalloc (sizeof (*conf) + raid_disks*sizeof(dev_info_t), | 109 | conf = kzalloc (sizeof (*conf) + raid_disks*sizeof(dev_info_t), |
111 | GFP_KERNEL); | 110 | GFP_KERNEL); |
@@ -115,7 +114,7 @@ static linear_conf_t *linear_conf(mddev_t *mddev, int raid_disks) | |||
115 | cnt = 0; | 114 | cnt = 0; |
116 | conf->array_sectors = 0; | 115 | conf->array_sectors = 0; |
117 | 116 | ||
118 | rdev_for_each(rdev, tmp, mddev) { | 117 | list_for_each_entry(rdev, &mddev->disks, same_set) { |
119 | int j = rdev->raid_disk; | 118 | int j = rdev->raid_disk; |
120 | dev_info_t *disk = conf->disks + j; | 119 | dev_info_t *disk = conf->disks + j; |
121 | 120 | ||
diff --git a/drivers/md/md.c b/drivers/md/md.c index 1b1d32694f6f..41e2509bf896 100644 --- a/drivers/md/md.c +++ b/drivers/md/md.c | |||
@@ -214,20 +214,33 @@ static inline mddev_t *mddev_get(mddev_t *mddev) | |||
214 | return mddev; | 214 | return mddev; |
215 | } | 215 | } |
216 | 216 | ||
217 | static void mddev_delayed_delete(struct work_struct *ws) | ||
218 | { | ||
219 | mddev_t *mddev = container_of(ws, mddev_t, del_work); | ||
220 | kobject_del(&mddev->kobj); | ||
221 | kobject_put(&mddev->kobj); | ||
222 | } | ||
223 | |||
217 | static void mddev_put(mddev_t *mddev) | 224 | static void mddev_put(mddev_t *mddev) |
218 | { | 225 | { |
219 | if (!atomic_dec_and_lock(&mddev->active, &all_mddevs_lock)) | 226 | if (!atomic_dec_and_lock(&mddev->active, &all_mddevs_lock)) |
220 | return; | 227 | return; |
221 | if (!mddev->raid_disks && list_empty(&mddev->disks)) { | 228 | if (!mddev->raid_disks && list_empty(&mddev->disks) && |
229 | !mddev->hold_active) { | ||
222 | list_del(&mddev->all_mddevs); | 230 | list_del(&mddev->all_mddevs); |
223 | spin_unlock(&all_mddevs_lock); | 231 | if (mddev->gendisk) { |
224 | blk_cleanup_queue(mddev->queue); | 232 | /* we did a probe so need to clean up. |
225 | if (mddev->sysfs_state) | 233 | * Call schedule_work inside the spinlock |
226 | sysfs_put(mddev->sysfs_state); | 234 | * so that flush_scheduled_work() after |
227 | mddev->sysfs_state = NULL; | 235 | * mddev_find will succeed in waiting for the |
228 | kobject_put(&mddev->kobj); | 236 | * work to be done. |
229 | } else | 237 | */ |
230 | spin_unlock(&all_mddevs_lock); | 238 | INIT_WORK(&mddev->del_work, mddev_delayed_delete); |
239 | schedule_work(&mddev->del_work); | ||
240 | } else | ||
241 | kfree(mddev); | ||
242 | } | ||
243 | spin_unlock(&all_mddevs_lock); | ||
231 | } | 244 | } |
232 | 245 | ||
233 | static mddev_t * mddev_find(dev_t unit) | 246 | static mddev_t * mddev_find(dev_t unit) |
@@ -236,15 +249,50 @@ static mddev_t * mddev_find(dev_t unit) | |||
236 | 249 | ||
237 | retry: | 250 | retry: |
238 | spin_lock(&all_mddevs_lock); | 251 | spin_lock(&all_mddevs_lock); |
239 | list_for_each_entry(mddev, &all_mddevs, all_mddevs) | 252 | |
240 | if (mddev->unit == unit) { | 253 | if (unit) { |
241 | mddev_get(mddev); | 254 | list_for_each_entry(mddev, &all_mddevs, all_mddevs) |
255 | if (mddev->unit == unit) { | ||
256 | mddev_get(mddev); | ||
257 | spin_unlock(&all_mddevs_lock); | ||
258 | kfree(new); | ||
259 | return mddev; | ||
260 | } | ||
261 | |||
262 | if (new) { | ||
263 | list_add(&new->all_mddevs, &all_mddevs); | ||
242 | spin_unlock(&all_mddevs_lock); | 264 | spin_unlock(&all_mddevs_lock); |
243 | kfree(new); | 265 | new->hold_active = UNTIL_IOCTL; |
244 | return mddev; | 266 | return new; |
245 | } | 267 | } |
246 | 268 | } else if (new) { | |
247 | if (new) { | 269 | /* find an unused unit number */ |
270 | static int next_minor = 512; | ||
271 | int start = next_minor; | ||
272 | int is_free = 0; | ||
273 | int dev = 0; | ||
274 | while (!is_free) { | ||
275 | dev = MKDEV(MD_MAJOR, next_minor); | ||
276 | next_minor++; | ||
277 | if (next_minor > MINORMASK) | ||
278 | next_minor = 0; | ||
279 | if (next_minor == start) { | ||
280 | /* Oh dear, all in use. */ | ||
281 | spin_unlock(&all_mddevs_lock); | ||
282 | kfree(new); | ||
283 | return NULL; | ||
284 | } | ||
285 | |||
286 | is_free = 1; | ||
287 | list_for_each_entry(mddev, &all_mddevs, all_mddevs) | ||
288 | if (mddev->unit == dev) { | ||
289 | is_free = 0; | ||
290 | break; | ||
291 | } | ||
292 | } | ||
293 | new->unit = dev; | ||
294 | new->md_minor = MINOR(dev); | ||
295 | new->hold_active = UNTIL_STOP; | ||
248 | list_add(&new->all_mddevs, &all_mddevs); | 296 | list_add(&new->all_mddevs, &all_mddevs); |
249 | spin_unlock(&all_mddevs_lock); | 297 | spin_unlock(&all_mddevs_lock); |
250 | return new; | 298 | return new; |
@@ -275,16 +323,6 @@ static mddev_t * mddev_find(dev_t unit) | |||
275 | new->resync_max = MaxSector; | 323 | new->resync_max = MaxSector; |
276 | new->level = LEVEL_NONE; | 324 | new->level = LEVEL_NONE; |
277 | 325 | ||
278 | new->queue = blk_alloc_queue(GFP_KERNEL); | ||
279 | if (!new->queue) { | ||
280 | kfree(new); | ||
281 | return NULL; | ||
282 | } | ||
283 | /* Can be unlocked because the queue is new: no concurrency */ | ||
284 | queue_flag_set_unlocked(QUEUE_FLAG_CLUSTER, new->queue); | ||
285 | |||
286 | blk_queue_make_request(new->queue, md_fail_request); | ||
287 | |||
288 | goto retry; | 326 | goto retry; |
289 | } | 327 | } |
290 | 328 | ||
@@ -307,25 +345,23 @@ static inline void mddev_unlock(mddev_t * mddev) | |||
307 | 345 | ||
308 | static mdk_rdev_t * find_rdev_nr(mddev_t *mddev, int nr) | 346 | static mdk_rdev_t * find_rdev_nr(mddev_t *mddev, int nr) |
309 | { | 347 | { |
310 | mdk_rdev_t * rdev; | 348 | mdk_rdev_t *rdev; |
311 | struct list_head *tmp; | ||
312 | 349 | ||
313 | rdev_for_each(rdev, tmp, mddev) { | 350 | list_for_each_entry(rdev, &mddev->disks, same_set) |
314 | if (rdev->desc_nr == nr) | 351 | if (rdev->desc_nr == nr) |
315 | return rdev; | 352 | return rdev; |
316 | } | 353 | |
317 | return NULL; | 354 | return NULL; |
318 | } | 355 | } |
319 | 356 | ||
320 | static mdk_rdev_t * find_rdev(mddev_t * mddev, dev_t dev) | 357 | static mdk_rdev_t * find_rdev(mddev_t * mddev, dev_t dev) |
321 | { | 358 | { |
322 | struct list_head *tmp; | ||
323 | mdk_rdev_t *rdev; | 359 | mdk_rdev_t *rdev; |
324 | 360 | ||
325 | rdev_for_each(rdev, tmp, mddev) { | 361 | list_for_each_entry(rdev, &mddev->disks, same_set) |
326 | if (rdev->bdev->bd_dev == dev) | 362 | if (rdev->bdev->bd_dev == dev) |
327 | return rdev; | 363 | return rdev; |
328 | } | 364 | |
329 | return NULL; | 365 | return NULL; |
330 | } | 366 | } |
331 | 367 | ||
@@ -861,7 +897,6 @@ static int super_90_validate(mddev_t *mddev, mdk_rdev_t *rdev) | |||
861 | static void super_90_sync(mddev_t *mddev, mdk_rdev_t *rdev) | 897 | static void super_90_sync(mddev_t *mddev, mdk_rdev_t *rdev) |
862 | { | 898 | { |
863 | mdp_super_t *sb; | 899 | mdp_super_t *sb; |
864 | struct list_head *tmp; | ||
865 | mdk_rdev_t *rdev2; | 900 | mdk_rdev_t *rdev2; |
866 | int next_spare = mddev->raid_disks; | 901 | int next_spare = mddev->raid_disks; |
867 | 902 | ||
@@ -933,7 +968,7 @@ static void super_90_sync(mddev_t *mddev, mdk_rdev_t *rdev) | |||
933 | sb->state |= (1<<MD_SB_BITMAP_PRESENT); | 968 | sb->state |= (1<<MD_SB_BITMAP_PRESENT); |
934 | 969 | ||
935 | sb->disks[0].state = (1<<MD_DISK_REMOVED); | 970 | sb->disks[0].state = (1<<MD_DISK_REMOVED); |
936 | rdev_for_each(rdev2, tmp, mddev) { | 971 | list_for_each_entry(rdev2, &mddev->disks, same_set) { |
937 | mdp_disk_t *d; | 972 | mdp_disk_t *d; |
938 | int desc_nr; | 973 | int desc_nr; |
939 | if (rdev2->raid_disk >= 0 && test_bit(In_sync, &rdev2->flags) | 974 | if (rdev2->raid_disk >= 0 && test_bit(In_sync, &rdev2->flags) |
@@ -1259,7 +1294,6 @@ static int super_1_validate(mddev_t *mddev, mdk_rdev_t *rdev) | |||
1259 | static void super_1_sync(mddev_t *mddev, mdk_rdev_t *rdev) | 1294 | static void super_1_sync(mddev_t *mddev, mdk_rdev_t *rdev) |
1260 | { | 1295 | { |
1261 | struct mdp_superblock_1 *sb; | 1296 | struct mdp_superblock_1 *sb; |
1262 | struct list_head *tmp; | ||
1263 | mdk_rdev_t *rdev2; | 1297 | mdk_rdev_t *rdev2; |
1264 | int max_dev, i; | 1298 | int max_dev, i; |
1265 | /* make rdev->sb match mddev and rdev data. */ | 1299 | /* make rdev->sb match mddev and rdev data. */ |
@@ -1307,7 +1341,7 @@ static void super_1_sync(mddev_t *mddev, mdk_rdev_t *rdev) | |||
1307 | } | 1341 | } |
1308 | 1342 | ||
1309 | max_dev = 0; | 1343 | max_dev = 0; |
1310 | rdev_for_each(rdev2, tmp, mddev) | 1344 | list_for_each_entry(rdev2, &mddev->disks, same_set) |
1311 | if (rdev2->desc_nr+1 > max_dev) | 1345 | if (rdev2->desc_nr+1 > max_dev) |
1312 | max_dev = rdev2->desc_nr+1; | 1346 | max_dev = rdev2->desc_nr+1; |
1313 | 1347 | ||
@@ -1316,7 +1350,7 @@ static void super_1_sync(mddev_t *mddev, mdk_rdev_t *rdev) | |||
1316 | for (i=0; i<max_dev;i++) | 1350 | for (i=0; i<max_dev;i++) |
1317 | sb->dev_roles[i] = cpu_to_le16(0xfffe); | 1351 | sb->dev_roles[i] = cpu_to_le16(0xfffe); |
1318 | 1352 | ||
1319 | rdev_for_each(rdev2, tmp, mddev) { | 1353 | list_for_each_entry(rdev2, &mddev->disks, same_set) { |
1320 | i = rdev2->desc_nr; | 1354 | i = rdev2->desc_nr; |
1321 | if (test_bit(Faulty, &rdev2->flags)) | 1355 | if (test_bit(Faulty, &rdev2->flags)) |
1322 | sb->dev_roles[i] = cpu_to_le16(0xfffe); | 1356 | sb->dev_roles[i] = cpu_to_le16(0xfffe); |
@@ -1466,6 +1500,9 @@ static int bind_rdev_to_array(mdk_rdev_t * rdev, mddev_t * mddev) | |||
1466 | 1500 | ||
1467 | list_add_rcu(&rdev->same_set, &mddev->disks); | 1501 | list_add_rcu(&rdev->same_set, &mddev->disks); |
1468 | bd_claim_by_disk(rdev->bdev, rdev->bdev->bd_holder, mddev->gendisk); | 1502 | bd_claim_by_disk(rdev->bdev, rdev->bdev->bd_holder, mddev->gendisk); |
1503 | |||
1504 | /* May as well allow recovery to be retried once */ | ||
1505 | mddev->recovery_disabled = 0; | ||
1469 | return 0; | 1506 | return 0; |
1470 | 1507 | ||
1471 | fail: | 1508 | fail: |
@@ -1571,8 +1608,7 @@ static void kick_rdev_from_array(mdk_rdev_t * rdev) | |||
1571 | 1608 | ||
1572 | static void export_array(mddev_t *mddev) | 1609 | static void export_array(mddev_t *mddev) |
1573 | { | 1610 | { |
1574 | struct list_head *tmp; | 1611 | mdk_rdev_t *rdev, *tmp; |
1575 | mdk_rdev_t *rdev; | ||
1576 | 1612 | ||
1577 | rdev_for_each(rdev, tmp, mddev) { | 1613 | rdev_for_each(rdev, tmp, mddev) { |
1578 | if (!rdev->mddev) { | 1614 | if (!rdev->mddev) { |
@@ -1593,7 +1629,7 @@ static void print_desc(mdp_disk_t *desc) | |||
1593 | desc->major,desc->minor,desc->raid_disk,desc->state); | 1629 | desc->major,desc->minor,desc->raid_disk,desc->state); |
1594 | } | 1630 | } |
1595 | 1631 | ||
1596 | static void print_sb(mdp_super_t *sb) | 1632 | static void print_sb_90(mdp_super_t *sb) |
1597 | { | 1633 | { |
1598 | int i; | 1634 | int i; |
1599 | 1635 | ||
@@ -1624,10 +1660,57 @@ static void print_sb(mdp_super_t *sb) | |||
1624 | } | 1660 | } |
1625 | printk(KERN_INFO "md: THIS: "); | 1661 | printk(KERN_INFO "md: THIS: "); |
1626 | print_desc(&sb->this_disk); | 1662 | print_desc(&sb->this_disk); |
1627 | |||
1628 | } | 1663 | } |
1629 | 1664 | ||
1630 | static void print_rdev(mdk_rdev_t *rdev) | 1665 | static void print_sb_1(struct mdp_superblock_1 *sb) |
1666 | { | ||
1667 | __u8 *uuid; | ||
1668 | |||
1669 | uuid = sb->set_uuid; | ||
1670 | printk(KERN_INFO "md: SB: (V:%u) (F:0x%08x) Array-ID:<%02x%02x%02x%02x" | ||
1671 | ":%02x%02x:%02x%02x:%02x%02x:%02x%02x%02x%02x%02x%02x>\n" | ||
1672 | KERN_INFO "md: Name: \"%s\" CT:%llu\n", | ||
1673 | le32_to_cpu(sb->major_version), | ||
1674 | le32_to_cpu(sb->feature_map), | ||
1675 | uuid[0], uuid[1], uuid[2], uuid[3], | ||
1676 | uuid[4], uuid[5], uuid[6], uuid[7], | ||
1677 | uuid[8], uuid[9], uuid[10], uuid[11], | ||
1678 | uuid[12], uuid[13], uuid[14], uuid[15], | ||
1679 | sb->set_name, | ||
1680 | (unsigned long long)le64_to_cpu(sb->ctime) | ||
1681 | & MD_SUPERBLOCK_1_TIME_SEC_MASK); | ||
1682 | |||
1683 | uuid = sb->device_uuid; | ||
1684 | printk(KERN_INFO "md: L%u SZ%llu RD:%u LO:%u CS:%u DO:%llu DS:%llu SO:%llu" | ||
1685 | " RO:%llu\n" | ||
1686 | KERN_INFO "md: Dev:%08x UUID: %02x%02x%02x%02x:%02x%02x:%02x%02x:%02x%02x" | ||
1687 | ":%02x%02x%02x%02x%02x%02x\n" | ||
1688 | KERN_INFO "md: (F:0x%08x) UT:%llu Events:%llu ResyncOffset:%llu CSUM:0x%08x\n" | ||
1689 | KERN_INFO "md: (MaxDev:%u) \n", | ||
1690 | le32_to_cpu(sb->level), | ||
1691 | (unsigned long long)le64_to_cpu(sb->size), | ||
1692 | le32_to_cpu(sb->raid_disks), | ||
1693 | le32_to_cpu(sb->layout), | ||
1694 | le32_to_cpu(sb->chunksize), | ||
1695 | (unsigned long long)le64_to_cpu(sb->data_offset), | ||
1696 | (unsigned long long)le64_to_cpu(sb->data_size), | ||
1697 | (unsigned long long)le64_to_cpu(sb->super_offset), | ||
1698 | (unsigned long long)le64_to_cpu(sb->recovery_offset), | ||
1699 | le32_to_cpu(sb->dev_number), | ||
1700 | uuid[0], uuid[1], uuid[2], uuid[3], | ||
1701 | uuid[4], uuid[5], uuid[6], uuid[7], | ||
1702 | uuid[8], uuid[9], uuid[10], uuid[11], | ||
1703 | uuid[12], uuid[13], uuid[14], uuid[15], | ||
1704 | sb->devflags, | ||
1705 | (unsigned long long)le64_to_cpu(sb->utime) & MD_SUPERBLOCK_1_TIME_SEC_MASK, | ||
1706 | (unsigned long long)le64_to_cpu(sb->events), | ||
1707 | (unsigned long long)le64_to_cpu(sb->resync_offset), | ||
1708 | le32_to_cpu(sb->sb_csum), | ||
1709 | le32_to_cpu(sb->max_dev) | ||
1710 | ); | ||
1711 | } | ||
1712 | |||
1713 | static void print_rdev(mdk_rdev_t *rdev, int major_version) | ||
1631 | { | 1714 | { |
1632 | char b[BDEVNAME_SIZE]; | 1715 | char b[BDEVNAME_SIZE]; |
1633 | printk(KERN_INFO "md: rdev %s, SZ:%08llu F:%d S:%d DN:%u\n", | 1716 | printk(KERN_INFO "md: rdev %s, SZ:%08llu F:%d S:%d DN:%u\n", |
@@ -1635,15 +1718,22 @@ static void print_rdev(mdk_rdev_t *rdev) | |||
1635 | test_bit(Faulty, &rdev->flags), test_bit(In_sync, &rdev->flags), | 1718 | test_bit(Faulty, &rdev->flags), test_bit(In_sync, &rdev->flags), |
1636 | rdev->desc_nr); | 1719 | rdev->desc_nr); |
1637 | if (rdev->sb_loaded) { | 1720 | if (rdev->sb_loaded) { |
1638 | printk(KERN_INFO "md: rdev superblock:\n"); | 1721 | printk(KERN_INFO "md: rdev superblock (MJ:%d):\n", major_version); |
1639 | print_sb((mdp_super_t*)page_address(rdev->sb_page)); | 1722 | switch (major_version) { |
1723 | case 0: | ||
1724 | print_sb_90((mdp_super_t*)page_address(rdev->sb_page)); | ||
1725 | break; | ||
1726 | case 1: | ||
1727 | print_sb_1((struct mdp_superblock_1 *)page_address(rdev->sb_page)); | ||
1728 | break; | ||
1729 | } | ||
1640 | } else | 1730 | } else |
1641 | printk(KERN_INFO "md: no rdev superblock!\n"); | 1731 | printk(KERN_INFO "md: no rdev superblock!\n"); |
1642 | } | 1732 | } |
1643 | 1733 | ||
1644 | static void md_print_devices(void) | 1734 | static void md_print_devices(void) |
1645 | { | 1735 | { |
1646 | struct list_head *tmp, *tmp2; | 1736 | struct list_head *tmp; |
1647 | mdk_rdev_t *rdev; | 1737 | mdk_rdev_t *rdev; |
1648 | mddev_t *mddev; | 1738 | mddev_t *mddev; |
1649 | char b[BDEVNAME_SIZE]; | 1739 | char b[BDEVNAME_SIZE]; |
@@ -1658,12 +1748,12 @@ static void md_print_devices(void) | |||
1658 | bitmap_print_sb(mddev->bitmap); | 1748 | bitmap_print_sb(mddev->bitmap); |
1659 | else | 1749 | else |
1660 | printk("%s: ", mdname(mddev)); | 1750 | printk("%s: ", mdname(mddev)); |
1661 | rdev_for_each(rdev, tmp2, mddev) | 1751 | list_for_each_entry(rdev, &mddev->disks, same_set) |
1662 | printk("<%s>", bdevname(rdev->bdev,b)); | 1752 | printk("<%s>", bdevname(rdev->bdev,b)); |
1663 | printk("\n"); | 1753 | printk("\n"); |
1664 | 1754 | ||
1665 | rdev_for_each(rdev, tmp2, mddev) | 1755 | list_for_each_entry(rdev, &mddev->disks, same_set) |
1666 | print_rdev(rdev); | 1756 | print_rdev(rdev, mddev->major_version); |
1667 | } | 1757 | } |
1668 | printk("md: **********************************\n"); | 1758 | printk("md: **********************************\n"); |
1669 | printk("\n"); | 1759 | printk("\n"); |
@@ -1679,9 +1769,8 @@ static void sync_sbs(mddev_t * mddev, int nospares) | |||
1679 | * with the rest of the array) | 1769 | * with the rest of the array) |
1680 | */ | 1770 | */ |
1681 | mdk_rdev_t *rdev; | 1771 | mdk_rdev_t *rdev; |
1682 | struct list_head *tmp; | ||
1683 | 1772 | ||
1684 | rdev_for_each(rdev, tmp, mddev) { | 1773 | list_for_each_entry(rdev, &mddev->disks, same_set) { |
1685 | if (rdev->sb_events == mddev->events || | 1774 | if (rdev->sb_events == mddev->events || |
1686 | (nospares && | 1775 | (nospares && |
1687 | rdev->raid_disk < 0 && | 1776 | rdev->raid_disk < 0 && |
@@ -1699,7 +1788,6 @@ static void sync_sbs(mddev_t * mddev, int nospares) | |||
1699 | 1788 | ||
1700 | static void md_update_sb(mddev_t * mddev, int force_change) | 1789 | static void md_update_sb(mddev_t * mddev, int force_change) |
1701 | { | 1790 | { |
1702 | struct list_head *tmp; | ||
1703 | mdk_rdev_t *rdev; | 1791 | mdk_rdev_t *rdev; |
1704 | int sync_req; | 1792 | int sync_req; |
1705 | int nospares = 0; | 1793 | int nospares = 0; |
@@ -1790,7 +1878,7 @@ repeat: | |||
1790 | mdname(mddev),mddev->in_sync); | 1878 | mdname(mddev),mddev->in_sync); |
1791 | 1879 | ||
1792 | bitmap_update_sb(mddev->bitmap); | 1880 | bitmap_update_sb(mddev->bitmap); |
1793 | rdev_for_each(rdev, tmp, mddev) { | 1881 | list_for_each_entry(rdev, &mddev->disks, same_set) { |
1794 | char b[BDEVNAME_SIZE]; | 1882 | char b[BDEVNAME_SIZE]; |
1795 | dprintk(KERN_INFO "md: "); | 1883 | dprintk(KERN_INFO "md: "); |
1796 | if (rdev->sb_loaded != 1) | 1884 | if (rdev->sb_loaded != 1) |
@@ -1999,7 +2087,6 @@ slot_store(mdk_rdev_t *rdev, const char *buf, size_t len) | |||
1999 | md_wakeup_thread(rdev->mddev->thread); | 2087 | md_wakeup_thread(rdev->mddev->thread); |
2000 | } else if (rdev->mddev->pers) { | 2088 | } else if (rdev->mddev->pers) { |
2001 | mdk_rdev_t *rdev2; | 2089 | mdk_rdev_t *rdev2; |
2002 | struct list_head *tmp; | ||
2003 | /* Activating a spare .. or possibly reactivating | 2090 | /* Activating a spare .. or possibly reactivating |
2004 | * if we every get bitmaps working here. | 2091 | * if we every get bitmaps working here. |
2005 | */ | 2092 | */ |
@@ -2010,7 +2097,7 @@ slot_store(mdk_rdev_t *rdev, const char *buf, size_t len) | |||
2010 | if (rdev->mddev->pers->hot_add_disk == NULL) | 2097 | if (rdev->mddev->pers->hot_add_disk == NULL) |
2011 | return -EINVAL; | 2098 | return -EINVAL; |
2012 | 2099 | ||
2013 | rdev_for_each(rdev2, tmp, rdev->mddev) | 2100 | list_for_each_entry(rdev2, &rdev->mddev->disks, same_set) |
2014 | if (rdev2->raid_disk == slot) | 2101 | if (rdev2->raid_disk == slot) |
2015 | return -EEXIST; | 2102 | return -EEXIST; |
2016 | 2103 | ||
@@ -2125,14 +2212,14 @@ rdev_size_store(mdk_rdev_t *rdev, const char *buf, size_t len) | |||
2125 | */ | 2212 | */ |
2126 | mddev_t *mddev; | 2213 | mddev_t *mddev; |
2127 | int overlap = 0; | 2214 | int overlap = 0; |
2128 | struct list_head *tmp, *tmp2; | 2215 | struct list_head *tmp; |
2129 | 2216 | ||
2130 | mddev_unlock(my_mddev); | 2217 | mddev_unlock(my_mddev); |
2131 | for_each_mddev(mddev, tmp) { | 2218 | for_each_mddev(mddev, tmp) { |
2132 | mdk_rdev_t *rdev2; | 2219 | mdk_rdev_t *rdev2; |
2133 | 2220 | ||
2134 | mddev_lock(mddev); | 2221 | mddev_lock(mddev); |
2135 | rdev_for_each(rdev2, tmp2, mddev) | 2222 | list_for_each_entry(rdev2, &mddev->disks, same_set) |
2136 | if (test_bit(AllReserved, &rdev2->flags) || | 2223 | if (test_bit(AllReserved, &rdev2->flags) || |
2137 | (rdev->bdev == rdev2->bdev && | 2224 | (rdev->bdev == rdev2->bdev && |
2138 | rdev != rdev2 && | 2225 | rdev != rdev2 && |
@@ -2328,8 +2415,7 @@ abort_free: | |||
2328 | static void analyze_sbs(mddev_t * mddev) | 2415 | static void analyze_sbs(mddev_t * mddev) |
2329 | { | 2416 | { |
2330 | int i; | 2417 | int i; |
2331 | struct list_head *tmp; | 2418 | mdk_rdev_t *rdev, *freshest, *tmp; |
2332 | mdk_rdev_t *rdev, *freshest; | ||
2333 | char b[BDEVNAME_SIZE]; | 2419 | char b[BDEVNAME_SIZE]; |
2334 | 2420 | ||
2335 | freshest = NULL; | 2421 | freshest = NULL; |
@@ -3046,7 +3132,7 @@ action_store(mddev_t *mddev, const char *page, size_t len) | |||
3046 | } | 3132 | } |
3047 | set_bit(MD_RECOVERY_NEEDED, &mddev->recovery); | 3133 | set_bit(MD_RECOVERY_NEEDED, &mddev->recovery); |
3048 | md_wakeup_thread(mddev->thread); | 3134 | md_wakeup_thread(mddev->thread); |
3049 | sysfs_notify(&mddev->kobj, NULL, "sync_action"); | 3135 | sysfs_notify_dirent(mddev->sysfs_action); |
3050 | return len; | 3136 | return len; |
3051 | } | 3137 | } |
3052 | 3138 | ||
@@ -3404,6 +3490,8 @@ md_attr_store(struct kobject *kobj, struct attribute *attr, | |||
3404 | if (!capable(CAP_SYS_ADMIN)) | 3490 | if (!capable(CAP_SYS_ADMIN)) |
3405 | return -EACCES; | 3491 | return -EACCES; |
3406 | rv = mddev_lock(mddev); | 3492 | rv = mddev_lock(mddev); |
3493 | if (mddev->hold_active == UNTIL_IOCTL) | ||
3494 | mddev->hold_active = 0; | ||
3407 | if (!rv) { | 3495 | if (!rv) { |
3408 | rv = entry->store(mddev, page, length); | 3496 | rv = entry->store(mddev, page, length); |
3409 | mddev_unlock(mddev); | 3497 | mddev_unlock(mddev); |
@@ -3414,6 +3502,17 @@ md_attr_store(struct kobject *kobj, struct attribute *attr, | |||
3414 | static void md_free(struct kobject *ko) | 3502 | static void md_free(struct kobject *ko) |
3415 | { | 3503 | { |
3416 | mddev_t *mddev = container_of(ko, mddev_t, kobj); | 3504 | mddev_t *mddev = container_of(ko, mddev_t, kobj); |
3505 | |||
3506 | if (mddev->sysfs_state) | ||
3507 | sysfs_put(mddev->sysfs_state); | ||
3508 | |||
3509 | if (mddev->gendisk) { | ||
3510 | del_gendisk(mddev->gendisk); | ||
3511 | put_disk(mddev->gendisk); | ||
3512 | } | ||
3513 | if (mddev->queue) | ||
3514 | blk_cleanup_queue(mddev->queue); | ||
3515 | |||
3417 | kfree(mddev); | 3516 | kfree(mddev); |
3418 | } | 3517 | } |
3419 | 3518 | ||
@@ -3429,34 +3528,74 @@ static struct kobj_type md_ktype = { | |||
3429 | 3528 | ||
3430 | int mdp_major = 0; | 3529 | int mdp_major = 0; |
3431 | 3530 | ||
3432 | static struct kobject *md_probe(dev_t dev, int *part, void *data) | 3531 | static int md_alloc(dev_t dev, char *name) |
3433 | { | 3532 | { |
3434 | static DEFINE_MUTEX(disks_mutex); | 3533 | static DEFINE_MUTEX(disks_mutex); |
3435 | mddev_t *mddev = mddev_find(dev); | 3534 | mddev_t *mddev = mddev_find(dev); |
3436 | struct gendisk *disk; | 3535 | struct gendisk *disk; |
3437 | int partitioned = (MAJOR(dev) != MD_MAJOR); | 3536 | int partitioned; |
3438 | int shift = partitioned ? MdpMinorShift : 0; | 3537 | int shift; |
3439 | int unit = MINOR(dev) >> shift; | 3538 | int unit; |
3440 | int error; | 3539 | int error; |
3441 | 3540 | ||
3442 | if (!mddev) | 3541 | if (!mddev) |
3443 | return NULL; | 3542 | return -ENODEV; |
3543 | |||
3544 | partitioned = (MAJOR(mddev->unit) != MD_MAJOR); | ||
3545 | shift = partitioned ? MdpMinorShift : 0; | ||
3546 | unit = MINOR(mddev->unit) >> shift; | ||
3547 | |||
3548 | /* wait for any previous instance if this device | ||
3549 | * to be completed removed (mddev_delayed_delete). | ||
3550 | */ | ||
3551 | flush_scheduled_work(); | ||
3444 | 3552 | ||
3445 | mutex_lock(&disks_mutex); | 3553 | mutex_lock(&disks_mutex); |
3446 | if (mddev->gendisk) { | 3554 | if (mddev->gendisk) { |
3447 | mutex_unlock(&disks_mutex); | 3555 | mutex_unlock(&disks_mutex); |
3448 | mddev_put(mddev); | 3556 | mddev_put(mddev); |
3449 | return NULL; | 3557 | return -EEXIST; |
3558 | } | ||
3559 | |||
3560 | if (name) { | ||
3561 | /* Need to ensure that 'name' is not a duplicate. | ||
3562 | */ | ||
3563 | mddev_t *mddev2; | ||
3564 | spin_lock(&all_mddevs_lock); | ||
3565 | |||
3566 | list_for_each_entry(mddev2, &all_mddevs, all_mddevs) | ||
3567 | if (mddev2->gendisk && | ||
3568 | strcmp(mddev2->gendisk->disk_name, name) == 0) { | ||
3569 | spin_unlock(&all_mddevs_lock); | ||
3570 | return -EEXIST; | ||
3571 | } | ||
3572 | spin_unlock(&all_mddevs_lock); | ||
3573 | } | ||
3574 | |||
3575 | mddev->queue = blk_alloc_queue(GFP_KERNEL); | ||
3576 | if (!mddev->queue) { | ||
3577 | mutex_unlock(&disks_mutex); | ||
3578 | mddev_put(mddev); | ||
3579 | return -ENOMEM; | ||
3450 | } | 3580 | } |
3581 | /* Can be unlocked because the queue is new: no concurrency */ | ||
3582 | queue_flag_set_unlocked(QUEUE_FLAG_CLUSTER, mddev->queue); | ||
3583 | |||
3584 | blk_queue_make_request(mddev->queue, md_fail_request); | ||
3585 | |||
3451 | disk = alloc_disk(1 << shift); | 3586 | disk = alloc_disk(1 << shift); |
3452 | if (!disk) { | 3587 | if (!disk) { |
3453 | mutex_unlock(&disks_mutex); | 3588 | mutex_unlock(&disks_mutex); |
3589 | blk_cleanup_queue(mddev->queue); | ||
3590 | mddev->queue = NULL; | ||
3454 | mddev_put(mddev); | 3591 | mddev_put(mddev); |
3455 | return NULL; | 3592 | return -ENOMEM; |
3456 | } | 3593 | } |
3457 | disk->major = MAJOR(dev); | 3594 | disk->major = MAJOR(mddev->unit); |
3458 | disk->first_minor = unit << shift; | 3595 | disk->first_minor = unit << shift; |
3459 | if (partitioned) | 3596 | if (name) |
3597 | strcpy(disk->disk_name, name); | ||
3598 | else if (partitioned) | ||
3460 | sprintf(disk->disk_name, "md_d%d", unit); | 3599 | sprintf(disk->disk_name, "md_d%d", unit); |
3461 | else | 3600 | else |
3462 | sprintf(disk->disk_name, "md%d", unit); | 3601 | sprintf(disk->disk_name, "md%d", unit); |
@@ -3464,7 +3603,7 @@ static struct kobject *md_probe(dev_t dev, int *part, void *data) | |||
3464 | disk->private_data = mddev; | 3603 | disk->private_data = mddev; |
3465 | disk->queue = mddev->queue; | 3604 | disk->queue = mddev->queue; |
3466 | /* Allow extended partitions. This makes the | 3605 | /* Allow extended partitions. This makes the |
3467 | * 'mdp' device redundant, but we can really | 3606 | * 'mdp' device redundant, but we can't really |
3468 | * remove it now. | 3607 | * remove it now. |
3469 | */ | 3608 | */ |
3470 | disk->flags |= GENHD_FL_EXT_DEVT; | 3609 | disk->flags |= GENHD_FL_EXT_DEVT; |
@@ -3480,9 +3619,35 @@ static struct kobject *md_probe(dev_t dev, int *part, void *data) | |||
3480 | kobject_uevent(&mddev->kobj, KOBJ_ADD); | 3619 | kobject_uevent(&mddev->kobj, KOBJ_ADD); |
3481 | mddev->sysfs_state = sysfs_get_dirent(mddev->kobj.sd, "array_state"); | 3620 | mddev->sysfs_state = sysfs_get_dirent(mddev->kobj.sd, "array_state"); |
3482 | } | 3621 | } |
3622 | mddev_put(mddev); | ||
3623 | return 0; | ||
3624 | } | ||
3625 | |||
3626 | static struct kobject *md_probe(dev_t dev, int *part, void *data) | ||
3627 | { | ||
3628 | md_alloc(dev, NULL); | ||
3483 | return NULL; | 3629 | return NULL; |
3484 | } | 3630 | } |
3485 | 3631 | ||
3632 | static int add_named_array(const char *val, struct kernel_param *kp) | ||
3633 | { | ||
3634 | /* val must be "md_*" where * is not all digits. | ||
3635 | * We allocate an array with a large free minor number, and | ||
3636 | * set the name to val. val must not already be an active name. | ||
3637 | */ | ||
3638 | int len = strlen(val); | ||
3639 | char buf[DISK_NAME_LEN]; | ||
3640 | |||
3641 | while (len && val[len-1] == '\n') | ||
3642 | len--; | ||
3643 | if (len >= DISK_NAME_LEN) | ||
3644 | return -E2BIG; | ||
3645 | strlcpy(buf, val, len+1); | ||
3646 | if (strncmp(buf, "md_", 3) != 0) | ||
3647 | return -EINVAL; | ||
3648 | return md_alloc(0, buf); | ||
3649 | } | ||
3650 | |||
3486 | static void md_safemode_timeout(unsigned long data) | 3651 | static void md_safemode_timeout(unsigned long data) |
3487 | { | 3652 | { |
3488 | mddev_t *mddev = (mddev_t *) data; | 3653 | mddev_t *mddev = (mddev_t *) data; |
@@ -3501,7 +3666,6 @@ static int do_md_run(mddev_t * mddev) | |||
3501 | { | 3666 | { |
3502 | int err; | 3667 | int err; |
3503 | int chunk_size; | 3668 | int chunk_size; |
3504 | struct list_head *tmp; | ||
3505 | mdk_rdev_t *rdev; | 3669 | mdk_rdev_t *rdev; |
3506 | struct gendisk *disk; | 3670 | struct gendisk *disk; |
3507 | struct mdk_personality *pers; | 3671 | struct mdk_personality *pers; |
@@ -3540,7 +3704,7 @@ static int do_md_run(mddev_t * mddev) | |||
3540 | } | 3704 | } |
3541 | 3705 | ||
3542 | /* devices must have minimum size of one chunk */ | 3706 | /* devices must have minimum size of one chunk */ |
3543 | rdev_for_each(rdev, tmp, mddev) { | 3707 | list_for_each_entry(rdev, &mddev->disks, same_set) { |
3544 | if (test_bit(Faulty, &rdev->flags)) | 3708 | if (test_bit(Faulty, &rdev->flags)) |
3545 | continue; | 3709 | continue; |
3546 | if (rdev->size < chunk_size / 1024) { | 3710 | if (rdev->size < chunk_size / 1024) { |
@@ -3565,7 +3729,7 @@ static int do_md_run(mddev_t * mddev) | |||
3565 | * the only valid external interface is through the md | 3729 | * the only valid external interface is through the md |
3566 | * device. | 3730 | * device. |
3567 | */ | 3731 | */ |
3568 | rdev_for_each(rdev, tmp, mddev) { | 3732 | list_for_each_entry(rdev, &mddev->disks, same_set) { |
3569 | if (test_bit(Faulty, &rdev->flags)) | 3733 | if (test_bit(Faulty, &rdev->flags)) |
3570 | continue; | 3734 | continue; |
3571 | sync_blockdev(rdev->bdev); | 3735 | sync_blockdev(rdev->bdev); |
@@ -3630,10 +3794,10 @@ static int do_md_run(mddev_t * mddev) | |||
3630 | */ | 3794 | */ |
3631 | char b[BDEVNAME_SIZE], b2[BDEVNAME_SIZE]; | 3795 | char b[BDEVNAME_SIZE], b2[BDEVNAME_SIZE]; |
3632 | mdk_rdev_t *rdev2; | 3796 | mdk_rdev_t *rdev2; |
3633 | struct list_head *tmp2; | ||
3634 | int warned = 0; | 3797 | int warned = 0; |
3635 | rdev_for_each(rdev, tmp, mddev) { | 3798 | |
3636 | rdev_for_each(rdev2, tmp2, mddev) { | 3799 | list_for_each_entry(rdev, &mddev->disks, same_set) |
3800 | list_for_each_entry(rdev2, &mddev->disks, same_set) { | ||
3637 | if (rdev < rdev2 && | 3801 | if (rdev < rdev2 && |
3638 | rdev->bdev->bd_contains == | 3802 | rdev->bdev->bd_contains == |
3639 | rdev2->bdev->bd_contains) { | 3803 | rdev2->bdev->bd_contains) { |
@@ -3647,7 +3811,7 @@ static int do_md_run(mddev_t * mddev) | |||
3647 | warned = 1; | 3811 | warned = 1; |
3648 | } | 3812 | } |
3649 | } | 3813 | } |
3650 | } | 3814 | |
3651 | if (warned) | 3815 | if (warned) |
3652 | printk(KERN_WARNING | 3816 | printk(KERN_WARNING |
3653 | "True protection against single-disk" | 3817 | "True protection against single-disk" |
@@ -3684,6 +3848,7 @@ static int do_md_run(mddev_t * mddev) | |||
3684 | printk(KERN_WARNING | 3848 | printk(KERN_WARNING |
3685 | "md: cannot register extra attributes for %s\n", | 3849 | "md: cannot register extra attributes for %s\n", |
3686 | mdname(mddev)); | 3850 | mdname(mddev)); |
3851 | mddev->sysfs_action = sysfs_get_dirent(mddev->kobj.sd, "sync_action"); | ||
3687 | } else if (mddev->ro == 2) /* auto-readonly not meaningful */ | 3852 | } else if (mddev->ro == 2) /* auto-readonly not meaningful */ |
3688 | mddev->ro = 0; | 3853 | mddev->ro = 0; |
3689 | 3854 | ||
@@ -3694,7 +3859,7 @@ static int do_md_run(mddev_t * mddev) | |||
3694 | mddev->safemode_delay = (200 * HZ)/1000 +1; /* 200 msec delay */ | 3859 | mddev->safemode_delay = (200 * HZ)/1000 +1; /* 200 msec delay */ |
3695 | mddev->in_sync = 1; | 3860 | mddev->in_sync = 1; |
3696 | 3861 | ||
3697 | rdev_for_each(rdev, tmp, mddev) | 3862 | list_for_each_entry(rdev, &mddev->disks, same_set) |
3698 | if (rdev->raid_disk >= 0) { | 3863 | if (rdev->raid_disk >= 0) { |
3699 | char nm[20]; | 3864 | char nm[20]; |
3700 | sprintf(nm, "rd%d", rdev->raid_disk); | 3865 | sprintf(nm, "rd%d", rdev->raid_disk); |
@@ -3725,9 +3890,8 @@ static int do_md_run(mddev_t * mddev) | |||
3725 | * it will remove the drives and not do the right thing | 3890 | * it will remove the drives and not do the right thing |
3726 | */ | 3891 | */ |
3727 | if (mddev->degraded && !mddev->sync_thread) { | 3892 | if (mddev->degraded && !mddev->sync_thread) { |
3728 | struct list_head *rtmp; | ||
3729 | int spares = 0; | 3893 | int spares = 0; |
3730 | rdev_for_each(rdev, rtmp, mddev) | 3894 | list_for_each_entry(rdev, &mddev->disks, same_set) |
3731 | if (rdev->raid_disk >= 0 && | 3895 | if (rdev->raid_disk >= 0 && |
3732 | !test_bit(In_sync, &rdev->flags) && | 3896 | !test_bit(In_sync, &rdev->flags) && |
3733 | !test_bit(Faulty, &rdev->flags)) | 3897 | !test_bit(Faulty, &rdev->flags)) |
@@ -3754,7 +3918,8 @@ static int do_md_run(mddev_t * mddev) | |||
3754 | mddev->changed = 1; | 3918 | mddev->changed = 1; |
3755 | md_new_event(mddev); | 3919 | md_new_event(mddev); |
3756 | sysfs_notify_dirent(mddev->sysfs_state); | 3920 | sysfs_notify_dirent(mddev->sysfs_state); |
3757 | sysfs_notify(&mddev->kobj, NULL, "sync_action"); | 3921 | if (mddev->sysfs_action) |
3922 | sysfs_notify_dirent(mddev->sysfs_action); | ||
3758 | sysfs_notify(&mddev->kobj, NULL, "degraded"); | 3923 | sysfs_notify(&mddev->kobj, NULL, "degraded"); |
3759 | kobject_uevent(&disk_to_dev(mddev->gendisk)->kobj, KOBJ_CHANGE); | 3924 | kobject_uevent(&disk_to_dev(mddev->gendisk)->kobj, KOBJ_CHANGE); |
3760 | return 0; | 3925 | return 0; |
@@ -3854,9 +4019,12 @@ static int do_md_stop(mddev_t * mddev, int mode, int is_open) | |||
3854 | mddev->queue->merge_bvec_fn = NULL; | 4019 | mddev->queue->merge_bvec_fn = NULL; |
3855 | mddev->queue->unplug_fn = NULL; | 4020 | mddev->queue->unplug_fn = NULL; |
3856 | mddev->queue->backing_dev_info.congested_fn = NULL; | 4021 | mddev->queue->backing_dev_info.congested_fn = NULL; |
3857 | if (mddev->pers->sync_request) | 4022 | if (mddev->pers->sync_request) { |
3858 | sysfs_remove_group(&mddev->kobj, &md_redundancy_group); | 4023 | sysfs_remove_group(&mddev->kobj, &md_redundancy_group); |
3859 | 4024 | if (mddev->sysfs_action) | |
4025 | sysfs_put(mddev->sysfs_action); | ||
4026 | mddev->sysfs_action = NULL; | ||
4027 | } | ||
3860 | module_put(mddev->pers->owner); | 4028 | module_put(mddev->pers->owner); |
3861 | mddev->pers = NULL; | 4029 | mddev->pers = NULL; |
3862 | /* tell userspace to handle 'inactive' */ | 4030 | /* tell userspace to handle 'inactive' */ |
@@ -3883,7 +4051,6 @@ static int do_md_stop(mddev_t * mddev, int mode, int is_open) | |||
3883 | */ | 4051 | */ |
3884 | if (mode == 0) { | 4052 | if (mode == 0) { |
3885 | mdk_rdev_t *rdev; | 4053 | mdk_rdev_t *rdev; |
3886 | struct list_head *tmp; | ||
3887 | 4054 | ||
3888 | printk(KERN_INFO "md: %s stopped.\n", mdname(mddev)); | 4055 | printk(KERN_INFO "md: %s stopped.\n", mdname(mddev)); |
3889 | 4056 | ||
@@ -3895,7 +4062,7 @@ static int do_md_stop(mddev_t * mddev, int mode, int is_open) | |||
3895 | } | 4062 | } |
3896 | mddev->bitmap_offset = 0; | 4063 | mddev->bitmap_offset = 0; |
3897 | 4064 | ||
3898 | rdev_for_each(rdev, tmp, mddev) | 4065 | list_for_each_entry(rdev, &mddev->disks, same_set) |
3899 | if (rdev->raid_disk >= 0) { | 4066 | if (rdev->raid_disk >= 0) { |
3900 | char nm[20]; | 4067 | char nm[20]; |
3901 | sprintf(nm, "rd%d", rdev->raid_disk); | 4068 | sprintf(nm, "rd%d", rdev->raid_disk); |
@@ -3941,6 +4108,8 @@ static int do_md_stop(mddev_t * mddev, int mode, int is_open) | |||
3941 | mddev->barriers_work = 0; | 4108 | mddev->barriers_work = 0; |
3942 | mddev->safemode = 0; | 4109 | mddev->safemode = 0; |
3943 | kobject_uevent(&disk_to_dev(mddev->gendisk)->kobj, KOBJ_CHANGE); | 4110 | kobject_uevent(&disk_to_dev(mddev->gendisk)->kobj, KOBJ_CHANGE); |
4111 | if (mddev->hold_active == UNTIL_STOP) | ||
4112 | mddev->hold_active = 0; | ||
3944 | 4113 | ||
3945 | } else if (mddev->pers) | 4114 | } else if (mddev->pers) |
3946 | printk(KERN_INFO "md: %s switched to read-only mode.\n", | 4115 | printk(KERN_INFO "md: %s switched to read-only mode.\n", |
@@ -3956,7 +4125,6 @@ out: | |||
3956 | static void autorun_array(mddev_t *mddev) | 4125 | static void autorun_array(mddev_t *mddev) |
3957 | { | 4126 | { |
3958 | mdk_rdev_t *rdev; | 4127 | mdk_rdev_t *rdev; |
3959 | struct list_head *tmp; | ||
3960 | int err; | 4128 | int err; |
3961 | 4129 | ||
3962 | if (list_empty(&mddev->disks)) | 4130 | if (list_empty(&mddev->disks)) |
@@ -3964,7 +4132,7 @@ static void autorun_array(mddev_t *mddev) | |||
3964 | 4132 | ||
3965 | printk(KERN_INFO "md: running: "); | 4133 | printk(KERN_INFO "md: running: "); |
3966 | 4134 | ||
3967 | rdev_for_each(rdev, tmp, mddev) { | 4135 | list_for_each_entry(rdev, &mddev->disks, same_set) { |
3968 | char b[BDEVNAME_SIZE]; | 4136 | char b[BDEVNAME_SIZE]; |
3969 | printk("<%s>", bdevname(rdev->bdev,b)); | 4137 | printk("<%s>", bdevname(rdev->bdev,b)); |
3970 | } | 4138 | } |
@@ -3991,8 +4159,7 @@ static void autorun_array(mddev_t *mddev) | |||
3991 | */ | 4159 | */ |
3992 | static void autorun_devices(int part) | 4160 | static void autorun_devices(int part) |
3993 | { | 4161 | { |
3994 | struct list_head *tmp; | 4162 | mdk_rdev_t *rdev0, *rdev, *tmp; |
3995 | mdk_rdev_t *rdev0, *rdev; | ||
3996 | mddev_t *mddev; | 4163 | mddev_t *mddev; |
3997 | char b[BDEVNAME_SIZE]; | 4164 | char b[BDEVNAME_SIZE]; |
3998 | 4165 | ||
@@ -4007,7 +4174,7 @@ static void autorun_devices(int part) | |||
4007 | printk(KERN_INFO "md: considering %s ...\n", | 4174 | printk(KERN_INFO "md: considering %s ...\n", |
4008 | bdevname(rdev0->bdev,b)); | 4175 | bdevname(rdev0->bdev,b)); |
4009 | INIT_LIST_HEAD(&candidates); | 4176 | INIT_LIST_HEAD(&candidates); |
4010 | rdev_for_each_list(rdev, tmp, pending_raid_disks) | 4177 | rdev_for_each_list(rdev, tmp, &pending_raid_disks) |
4011 | if (super_90_load(rdev, rdev0, 0) >= 0) { | 4178 | if (super_90_load(rdev, rdev0, 0) >= 0) { |
4012 | printk(KERN_INFO "md: adding %s ...\n", | 4179 | printk(KERN_INFO "md: adding %s ...\n", |
4013 | bdevname(rdev->bdev,b)); | 4180 | bdevname(rdev->bdev,b)); |
@@ -4053,7 +4220,7 @@ static void autorun_devices(int part) | |||
4053 | } else { | 4220 | } else { |
4054 | printk(KERN_INFO "md: created %s\n", mdname(mddev)); | 4221 | printk(KERN_INFO "md: created %s\n", mdname(mddev)); |
4055 | mddev->persistent = 1; | 4222 | mddev->persistent = 1; |
4056 | rdev_for_each_list(rdev, tmp, candidates) { | 4223 | rdev_for_each_list(rdev, tmp, &candidates) { |
4057 | list_del_init(&rdev->same_set); | 4224 | list_del_init(&rdev->same_set); |
4058 | if (bind_rdev_to_array(rdev, mddev)) | 4225 | if (bind_rdev_to_array(rdev, mddev)) |
4059 | export_rdev(rdev); | 4226 | export_rdev(rdev); |
@@ -4064,7 +4231,7 @@ static void autorun_devices(int part) | |||
4064 | /* on success, candidates will be empty, on error | 4231 | /* on success, candidates will be empty, on error |
4065 | * it won't... | 4232 | * it won't... |
4066 | */ | 4233 | */ |
4067 | rdev_for_each_list(rdev, tmp, candidates) { | 4234 | rdev_for_each_list(rdev, tmp, &candidates) { |
4068 | list_del_init(&rdev->same_set); | 4235 | list_del_init(&rdev->same_set); |
4069 | export_rdev(rdev); | 4236 | export_rdev(rdev); |
4070 | } | 4237 | } |
@@ -4093,10 +4260,9 @@ static int get_array_info(mddev_t * mddev, void __user * arg) | |||
4093 | mdu_array_info_t info; | 4260 | mdu_array_info_t info; |
4094 | int nr,working,active,failed,spare; | 4261 | int nr,working,active,failed,spare; |
4095 | mdk_rdev_t *rdev; | 4262 | mdk_rdev_t *rdev; |
4096 | struct list_head *tmp; | ||
4097 | 4263 | ||
4098 | nr=working=active=failed=spare=0; | 4264 | nr=working=active=failed=spare=0; |
4099 | rdev_for_each(rdev, tmp, mddev) { | 4265 | list_for_each_entry(rdev, &mddev->disks, same_set) { |
4100 | nr++; | 4266 | nr++; |
4101 | if (test_bit(Faulty, &rdev->flags)) | 4267 | if (test_bit(Faulty, &rdev->flags)) |
4102 | failed++; | 4268 | failed++; |
@@ -4614,9 +4780,8 @@ static int set_array_info(mddev_t * mddev, mdu_array_info_t *info) | |||
4614 | 4780 | ||
4615 | static int update_size(mddev_t *mddev, sector_t num_sectors) | 4781 | static int update_size(mddev_t *mddev, sector_t num_sectors) |
4616 | { | 4782 | { |
4617 | mdk_rdev_t * rdev; | 4783 | mdk_rdev_t *rdev; |
4618 | int rv; | 4784 | int rv; |
4619 | struct list_head *tmp; | ||
4620 | int fit = (num_sectors == 0); | 4785 | int fit = (num_sectors == 0); |
4621 | 4786 | ||
4622 | if (mddev->pers->resize == NULL) | 4787 | if (mddev->pers->resize == NULL) |
@@ -4638,7 +4803,7 @@ static int update_size(mddev_t *mddev, sector_t num_sectors) | |||
4638 | * grow, and re-add. | 4803 | * grow, and re-add. |
4639 | */ | 4804 | */ |
4640 | return -EBUSY; | 4805 | return -EBUSY; |
4641 | rdev_for_each(rdev, tmp, mddev) { | 4806 | list_for_each_entry(rdev, &mddev->disks, same_set) { |
4642 | sector_t avail; | 4807 | sector_t avail; |
4643 | avail = rdev->size * 2; | 4808 | avail = rdev->size * 2; |
4644 | 4809 | ||
@@ -5000,6 +5165,9 @@ static int md_ioctl(struct block_device *bdev, fmode_t mode, | |||
5000 | 5165 | ||
5001 | done_unlock: | 5166 | done_unlock: |
5002 | abort_unlock: | 5167 | abort_unlock: |
5168 | if (mddev->hold_active == UNTIL_IOCTL && | ||
5169 | err != -EINVAL) | ||
5170 | mddev->hold_active = 0; | ||
5003 | mddev_unlock(mddev); | 5171 | mddev_unlock(mddev); |
5004 | 5172 | ||
5005 | return err; | 5173 | return err; |
@@ -5016,14 +5184,25 @@ static int md_open(struct block_device *bdev, fmode_t mode) | |||
5016 | * Succeed if we can lock the mddev, which confirms that | 5184 | * Succeed if we can lock the mddev, which confirms that |
5017 | * it isn't being stopped right now. | 5185 | * it isn't being stopped right now. |
5018 | */ | 5186 | */ |
5019 | mddev_t *mddev = bdev->bd_disk->private_data; | 5187 | mddev_t *mddev = mddev_find(bdev->bd_dev); |
5020 | int err; | 5188 | int err; |
5021 | 5189 | ||
5190 | if (mddev->gendisk != bdev->bd_disk) { | ||
5191 | /* we are racing with mddev_put which is discarding this | ||
5192 | * bd_disk. | ||
5193 | */ | ||
5194 | mddev_put(mddev); | ||
5195 | /* Wait until bdev->bd_disk is definitely gone */ | ||
5196 | flush_scheduled_work(); | ||
5197 | /* Then retry the open from the top */ | ||
5198 | return -ERESTARTSYS; | ||
5199 | } | ||
5200 | BUG_ON(mddev != bdev->bd_disk->private_data); | ||
5201 | |||
5022 | if ((err = mutex_lock_interruptible_nested(&mddev->reconfig_mutex, 1))) | 5202 | if ((err = mutex_lock_interruptible_nested(&mddev->reconfig_mutex, 1))) |
5023 | goto out; | 5203 | goto out; |
5024 | 5204 | ||
5025 | err = 0; | 5205 | err = 0; |
5026 | mddev_get(mddev); | ||
5027 | atomic_inc(&mddev->openers); | 5206 | atomic_inc(&mddev->openers); |
5028 | mddev_unlock(mddev); | 5207 | mddev_unlock(mddev); |
5029 | 5208 | ||
@@ -5187,11 +5366,10 @@ static void status_unused(struct seq_file *seq) | |||
5187 | { | 5366 | { |
5188 | int i = 0; | 5367 | int i = 0; |
5189 | mdk_rdev_t *rdev; | 5368 | mdk_rdev_t *rdev; |
5190 | struct list_head *tmp; | ||
5191 | 5369 | ||
5192 | seq_printf(seq, "unused devices: "); | 5370 | seq_printf(seq, "unused devices: "); |
5193 | 5371 | ||
5194 | rdev_for_each_list(rdev, tmp, pending_raid_disks) { | 5372 | list_for_each_entry(rdev, &pending_raid_disks, same_set) { |
5195 | char b[BDEVNAME_SIZE]; | 5373 | char b[BDEVNAME_SIZE]; |
5196 | i++; | 5374 | i++; |
5197 | seq_printf(seq, "%s ", | 5375 | seq_printf(seq, "%s ", |
@@ -5350,7 +5528,6 @@ static int md_seq_show(struct seq_file *seq, void *v) | |||
5350 | { | 5528 | { |
5351 | mddev_t *mddev = v; | 5529 | mddev_t *mddev = v; |
5352 | sector_t size; | 5530 | sector_t size; |
5353 | struct list_head *tmp2; | ||
5354 | mdk_rdev_t *rdev; | 5531 | mdk_rdev_t *rdev; |
5355 | struct mdstat_info *mi = seq->private; | 5532 | struct mdstat_info *mi = seq->private; |
5356 | struct bitmap *bitmap; | 5533 | struct bitmap *bitmap; |
@@ -5387,7 +5564,7 @@ static int md_seq_show(struct seq_file *seq, void *v) | |||
5387 | } | 5564 | } |
5388 | 5565 | ||
5389 | size = 0; | 5566 | size = 0; |
5390 | rdev_for_each(rdev, tmp2, mddev) { | 5567 | list_for_each_entry(rdev, &mddev->disks, same_set) { |
5391 | char b[BDEVNAME_SIZE]; | 5568 | char b[BDEVNAME_SIZE]; |
5392 | seq_printf(seq, " %s[%d]", | 5569 | seq_printf(seq, " %s[%d]", |
5393 | bdevname(rdev->bdev,b), rdev->desc_nr); | 5570 | bdevname(rdev->bdev,b), rdev->desc_nr); |
@@ -5694,7 +5871,6 @@ void md_do_sync(mddev_t *mddev) | |||
5694 | struct list_head *tmp; | 5871 | struct list_head *tmp; |
5695 | sector_t last_check; | 5872 | sector_t last_check; |
5696 | int skipped = 0; | 5873 | int skipped = 0; |
5697 | struct list_head *rtmp; | ||
5698 | mdk_rdev_t *rdev; | 5874 | mdk_rdev_t *rdev; |
5699 | char *desc; | 5875 | char *desc; |
5700 | 5876 | ||
@@ -5799,7 +5975,7 @@ void md_do_sync(mddev_t *mddev) | |||
5799 | /* recovery follows the physical size of devices */ | 5975 | /* recovery follows the physical size of devices */ |
5800 | max_sectors = mddev->size << 1; | 5976 | max_sectors = mddev->size << 1; |
5801 | j = MaxSector; | 5977 | j = MaxSector; |
5802 | rdev_for_each(rdev, rtmp, mddev) | 5978 | list_for_each_entry(rdev, &mddev->disks, same_set) |
5803 | if (rdev->raid_disk >= 0 && | 5979 | if (rdev->raid_disk >= 0 && |
5804 | !test_bit(Faulty, &rdev->flags) && | 5980 | !test_bit(Faulty, &rdev->flags) && |
5805 | !test_bit(In_sync, &rdev->flags) && | 5981 | !test_bit(In_sync, &rdev->flags) && |
@@ -5949,7 +6125,7 @@ void md_do_sync(mddev_t *mddev) | |||
5949 | } else { | 6125 | } else { |
5950 | if (!test_bit(MD_RECOVERY_INTR, &mddev->recovery)) | 6126 | if (!test_bit(MD_RECOVERY_INTR, &mddev->recovery)) |
5951 | mddev->curr_resync = MaxSector; | 6127 | mddev->curr_resync = MaxSector; |
5952 | rdev_for_each(rdev, rtmp, mddev) | 6128 | list_for_each_entry(rdev, &mddev->disks, same_set) |
5953 | if (rdev->raid_disk >= 0 && | 6129 | if (rdev->raid_disk >= 0 && |
5954 | !test_bit(Faulty, &rdev->flags) && | 6130 | !test_bit(Faulty, &rdev->flags) && |
5955 | !test_bit(In_sync, &rdev->flags) && | 6131 | !test_bit(In_sync, &rdev->flags) && |
@@ -5985,10 +6161,9 @@ EXPORT_SYMBOL_GPL(md_do_sync); | |||
5985 | static int remove_and_add_spares(mddev_t *mddev) | 6161 | static int remove_and_add_spares(mddev_t *mddev) |
5986 | { | 6162 | { |
5987 | mdk_rdev_t *rdev; | 6163 | mdk_rdev_t *rdev; |
5988 | struct list_head *rtmp; | ||
5989 | int spares = 0; | 6164 | int spares = 0; |
5990 | 6165 | ||
5991 | rdev_for_each(rdev, rtmp, mddev) | 6166 | list_for_each_entry(rdev, &mddev->disks, same_set) |
5992 | if (rdev->raid_disk >= 0 && | 6167 | if (rdev->raid_disk >= 0 && |
5993 | !test_bit(Blocked, &rdev->flags) && | 6168 | !test_bit(Blocked, &rdev->flags) && |
5994 | (test_bit(Faulty, &rdev->flags) || | 6169 | (test_bit(Faulty, &rdev->flags) || |
@@ -6003,8 +6178,8 @@ static int remove_and_add_spares(mddev_t *mddev) | |||
6003 | } | 6178 | } |
6004 | } | 6179 | } |
6005 | 6180 | ||
6006 | if (mddev->degraded && ! mddev->ro) { | 6181 | if (mddev->degraded && ! mddev->ro && !mddev->recovery_disabled) { |
6007 | rdev_for_each(rdev, rtmp, mddev) { | 6182 | list_for_each_entry(rdev, &mddev->disks, same_set) { |
6008 | if (rdev->raid_disk >= 0 && | 6183 | if (rdev->raid_disk >= 0 && |
6009 | !test_bit(In_sync, &rdev->flags) && | 6184 | !test_bit(In_sync, &rdev->flags) && |
6010 | !test_bit(Blocked, &rdev->flags)) | 6185 | !test_bit(Blocked, &rdev->flags)) |
@@ -6056,7 +6231,6 @@ static int remove_and_add_spares(mddev_t *mddev) | |||
6056 | void md_check_recovery(mddev_t *mddev) | 6231 | void md_check_recovery(mddev_t *mddev) |
6057 | { | 6232 | { |
6058 | mdk_rdev_t *rdev; | 6233 | mdk_rdev_t *rdev; |
6059 | struct list_head *rtmp; | ||
6060 | 6234 | ||
6061 | 6235 | ||
6062 | if (mddev->bitmap) | 6236 | if (mddev->bitmap) |
@@ -6120,7 +6294,7 @@ void md_check_recovery(mddev_t *mddev) | |||
6120 | if (mddev->flags) | 6294 | if (mddev->flags) |
6121 | md_update_sb(mddev, 0); | 6295 | md_update_sb(mddev, 0); |
6122 | 6296 | ||
6123 | rdev_for_each(rdev, rtmp, mddev) | 6297 | list_for_each_entry(rdev, &mddev->disks, same_set) |
6124 | if (test_and_clear_bit(StateChanged, &rdev->flags)) | 6298 | if (test_and_clear_bit(StateChanged, &rdev->flags)) |
6125 | sysfs_notify_dirent(rdev->sysfs_state); | 6299 | sysfs_notify_dirent(rdev->sysfs_state); |
6126 | 6300 | ||
@@ -6149,13 +6323,13 @@ void md_check_recovery(mddev_t *mddev) | |||
6149 | * information must be scrapped | 6323 | * information must be scrapped |
6150 | */ | 6324 | */ |
6151 | if (!mddev->degraded) | 6325 | if (!mddev->degraded) |
6152 | rdev_for_each(rdev, rtmp, mddev) | 6326 | list_for_each_entry(rdev, &mddev->disks, same_set) |
6153 | rdev->saved_raid_disk = -1; | 6327 | rdev->saved_raid_disk = -1; |
6154 | 6328 | ||
6155 | mddev->recovery = 0; | 6329 | mddev->recovery = 0; |
6156 | /* flag recovery needed just to double check */ | 6330 | /* flag recovery needed just to double check */ |
6157 | set_bit(MD_RECOVERY_NEEDED, &mddev->recovery); | 6331 | set_bit(MD_RECOVERY_NEEDED, &mddev->recovery); |
6158 | sysfs_notify(&mddev->kobj, NULL, "sync_action"); | 6332 | sysfs_notify_dirent(mddev->sysfs_action); |
6159 | md_new_event(mddev); | 6333 | md_new_event(mddev); |
6160 | goto unlock; | 6334 | goto unlock; |
6161 | } | 6335 | } |
@@ -6216,7 +6390,7 @@ void md_check_recovery(mddev_t *mddev) | |||
6216 | mddev->recovery = 0; | 6390 | mddev->recovery = 0; |
6217 | } else | 6391 | } else |
6218 | md_wakeup_thread(mddev->sync_thread); | 6392 | md_wakeup_thread(mddev->sync_thread); |
6219 | sysfs_notify(&mddev->kobj, NULL, "sync_action"); | 6393 | sysfs_notify_dirent(mddev->sysfs_action); |
6220 | md_new_event(mddev); | 6394 | md_new_event(mddev); |
6221 | } | 6395 | } |
6222 | unlock: | 6396 | unlock: |
@@ -6224,7 +6398,8 @@ void md_check_recovery(mddev_t *mddev) | |||
6224 | clear_bit(MD_RECOVERY_RUNNING, &mddev->recovery); | 6398 | clear_bit(MD_RECOVERY_RUNNING, &mddev->recovery); |
6225 | if (test_and_clear_bit(MD_RECOVERY_RECOVER, | 6399 | if (test_and_clear_bit(MD_RECOVERY_RECOVER, |
6226 | &mddev->recovery)) | 6400 | &mddev->recovery)) |
6227 | sysfs_notify(&mddev->kobj, NULL, "sync_action"); | 6401 | if (mddev->sysfs_action) |
6402 | sysfs_notify_dirent(mddev->sysfs_action); | ||
6228 | } | 6403 | } |
6229 | mddev_unlock(mddev); | 6404 | mddev_unlock(mddev); |
6230 | } | 6405 | } |
@@ -6386,14 +6561,8 @@ static __exit void md_exit(void) | |||
6386 | unregister_sysctl_table(raid_table_header); | 6561 | unregister_sysctl_table(raid_table_header); |
6387 | remove_proc_entry("mdstat", NULL); | 6562 | remove_proc_entry("mdstat", NULL); |
6388 | for_each_mddev(mddev, tmp) { | 6563 | for_each_mddev(mddev, tmp) { |
6389 | struct gendisk *disk = mddev->gendisk; | ||
6390 | if (!disk) | ||
6391 | continue; | ||
6392 | export_array(mddev); | 6564 | export_array(mddev); |
6393 | del_gendisk(disk); | 6565 | mddev->hold_active = 0; |
6394 | put_disk(disk); | ||
6395 | mddev->gendisk = NULL; | ||
6396 | mddev_put(mddev); | ||
6397 | } | 6566 | } |
6398 | } | 6567 | } |
6399 | 6568 | ||
@@ -6418,6 +6587,7 @@ static int set_ro(const char *val, struct kernel_param *kp) | |||
6418 | module_param_call(start_ro, set_ro, get_ro, NULL, S_IRUSR|S_IWUSR); | 6587 | module_param_call(start_ro, set_ro, get_ro, NULL, S_IRUSR|S_IWUSR); |
6419 | module_param(start_dirty_degraded, int, S_IRUGO|S_IWUSR); | 6588 | module_param(start_dirty_degraded, int, S_IRUGO|S_IWUSR); |
6420 | 6589 | ||
6590 | module_param_call(new_array, add_named_array, NULL, NULL, S_IWUSR); | ||
6421 | 6591 | ||
6422 | EXPORT_SYMBOL(register_md_personality); | 6592 | EXPORT_SYMBOL(register_md_personality); |
6423 | EXPORT_SYMBOL(unregister_md_personality); | 6593 | EXPORT_SYMBOL(unregister_md_personality); |
diff --git a/drivers/md/multipath.c b/drivers/md/multipath.c index d4ac47d11279..f6d08f241671 100644 --- a/drivers/md/multipath.c +++ b/drivers/md/multipath.c | |||
@@ -408,7 +408,6 @@ static int multipath_run (mddev_t *mddev) | |||
408 | int disk_idx; | 408 | int disk_idx; |
409 | struct multipath_info *disk; | 409 | struct multipath_info *disk; |
410 | mdk_rdev_t *rdev; | 410 | mdk_rdev_t *rdev; |
411 | struct list_head *tmp; | ||
412 | 411 | ||
413 | if (mddev->level != LEVEL_MULTIPATH) { | 412 | if (mddev->level != LEVEL_MULTIPATH) { |
414 | printk("multipath: %s: raid level not set to multipath IO (%d)\n", | 413 | printk("multipath: %s: raid level not set to multipath IO (%d)\n", |
@@ -441,7 +440,7 @@ static int multipath_run (mddev_t *mddev) | |||
441 | } | 440 | } |
442 | 441 | ||
443 | conf->working_disks = 0; | 442 | conf->working_disks = 0; |
444 | rdev_for_each(rdev, tmp, mddev) { | 443 | list_for_each_entry(rdev, &mddev->disks, same_set) { |
445 | disk_idx = rdev->raid_disk; | 444 | disk_idx = rdev->raid_disk; |
446 | if (disk_idx < 0 || | 445 | if (disk_idx < 0 || |
447 | disk_idx >= mddev->raid_disks) | 446 | disk_idx >= mddev->raid_disks) |
diff --git a/drivers/md/raid0.c b/drivers/md/raid0.c index 8ac6488ad0dc..c605ba805586 100644 --- a/drivers/md/raid0.c +++ b/drivers/md/raid0.c | |||
@@ -53,11 +53,10 @@ static int raid0_congested(void *data, int bits) | |||
53 | static int create_strip_zones (mddev_t *mddev) | 53 | static int create_strip_zones (mddev_t *mddev) |
54 | { | 54 | { |
55 | int i, c, j; | 55 | int i, c, j; |
56 | sector_t current_offset, curr_zone_offset; | 56 | sector_t current_start, curr_zone_start; |
57 | sector_t min_spacing; | 57 | sector_t min_spacing; |
58 | raid0_conf_t *conf = mddev_to_conf(mddev); | 58 | raid0_conf_t *conf = mddev_to_conf(mddev); |
59 | mdk_rdev_t *smallest, *rdev1, *rdev2, *rdev; | 59 | mdk_rdev_t *smallest, *rdev1, *rdev2, *rdev; |
60 | struct list_head *tmp1, *tmp2; | ||
61 | struct strip_zone *zone; | 60 | struct strip_zone *zone; |
62 | int cnt; | 61 | int cnt; |
63 | char b[BDEVNAME_SIZE]; | 62 | char b[BDEVNAME_SIZE]; |
@@ -67,19 +66,19 @@ static int create_strip_zones (mddev_t *mddev) | |||
67 | */ | 66 | */ |
68 | conf->nr_strip_zones = 0; | 67 | conf->nr_strip_zones = 0; |
69 | 68 | ||
70 | rdev_for_each(rdev1, tmp1, mddev) { | 69 | list_for_each_entry(rdev1, &mddev->disks, same_set) { |
71 | printk("raid0: looking at %s\n", | 70 | printk(KERN_INFO "raid0: looking at %s\n", |
72 | bdevname(rdev1->bdev,b)); | 71 | bdevname(rdev1->bdev,b)); |
73 | c = 0; | 72 | c = 0; |
74 | rdev_for_each(rdev2, tmp2, mddev) { | 73 | list_for_each_entry(rdev2, &mddev->disks, same_set) { |
75 | printk("raid0: comparing %s(%llu)", | 74 | printk(KERN_INFO "raid0: comparing %s(%llu)", |
76 | bdevname(rdev1->bdev,b), | 75 | bdevname(rdev1->bdev,b), |
77 | (unsigned long long)rdev1->size); | 76 | (unsigned long long)rdev1->size); |
78 | printk(" with %s(%llu)\n", | 77 | printk(KERN_INFO " with %s(%llu)\n", |
79 | bdevname(rdev2->bdev,b), | 78 | bdevname(rdev2->bdev,b), |
80 | (unsigned long long)rdev2->size); | 79 | (unsigned long long)rdev2->size); |
81 | if (rdev2 == rdev1) { | 80 | if (rdev2 == rdev1) { |
82 | printk("raid0: END\n"); | 81 | printk(KERN_INFO "raid0: END\n"); |
83 | break; | 82 | break; |
84 | } | 83 | } |
85 | if (rdev2->size == rdev1->size) | 84 | if (rdev2->size == rdev1->size) |
@@ -88,19 +87,20 @@ static int create_strip_zones (mddev_t *mddev) | |||
88 | * Not unique, don't count it as a new | 87 | * Not unique, don't count it as a new |
89 | * group | 88 | * group |
90 | */ | 89 | */ |
91 | printk("raid0: EQUAL\n"); | 90 | printk(KERN_INFO "raid0: EQUAL\n"); |
92 | c = 1; | 91 | c = 1; |
93 | break; | 92 | break; |
94 | } | 93 | } |
95 | printk("raid0: NOT EQUAL\n"); | 94 | printk(KERN_INFO "raid0: NOT EQUAL\n"); |
96 | } | 95 | } |
97 | if (!c) { | 96 | if (!c) { |
98 | printk("raid0: ==> UNIQUE\n"); | 97 | printk(KERN_INFO "raid0: ==> UNIQUE\n"); |
99 | conf->nr_strip_zones++; | 98 | conf->nr_strip_zones++; |
100 | printk("raid0: %d zones\n", conf->nr_strip_zones); | 99 | printk(KERN_INFO "raid0: %d zones\n", |
100 | conf->nr_strip_zones); | ||
101 | } | 101 | } |
102 | } | 102 | } |
103 | printk("raid0: FINAL %d zones\n", conf->nr_strip_zones); | 103 | printk(KERN_INFO "raid0: FINAL %d zones\n", conf->nr_strip_zones); |
104 | 104 | ||
105 | conf->strip_zone = kzalloc(sizeof(struct strip_zone)* | 105 | conf->strip_zone = kzalloc(sizeof(struct strip_zone)* |
106 | conf->nr_strip_zones, GFP_KERNEL); | 106 | conf->nr_strip_zones, GFP_KERNEL); |
@@ -119,16 +119,17 @@ static int create_strip_zones (mddev_t *mddev) | |||
119 | cnt = 0; | 119 | cnt = 0; |
120 | smallest = NULL; | 120 | smallest = NULL; |
121 | zone->dev = conf->devlist; | 121 | zone->dev = conf->devlist; |
122 | rdev_for_each(rdev1, tmp1, mddev) { | 122 | list_for_each_entry(rdev1, &mddev->disks, same_set) { |
123 | int j = rdev1->raid_disk; | 123 | int j = rdev1->raid_disk; |
124 | 124 | ||
125 | if (j < 0 || j >= mddev->raid_disks) { | 125 | if (j < 0 || j >= mddev->raid_disks) { |
126 | printk("raid0: bad disk number %d - aborting!\n", j); | 126 | printk(KERN_ERR "raid0: bad disk number %d - " |
127 | "aborting!\n", j); | ||
127 | goto abort; | 128 | goto abort; |
128 | } | 129 | } |
129 | if (zone->dev[j]) { | 130 | if (zone->dev[j]) { |
130 | printk("raid0: multiple devices for %d - aborting!\n", | 131 | printk(KERN_ERR "raid0: multiple devices for %d - " |
131 | j); | 132 | "aborting!\n", j); |
132 | goto abort; | 133 | goto abort; |
133 | } | 134 | } |
134 | zone->dev[j] = rdev1; | 135 | zone->dev[j] = rdev1; |
@@ -149,16 +150,16 @@ static int create_strip_zones (mddev_t *mddev) | |||
149 | cnt++; | 150 | cnt++; |
150 | } | 151 | } |
151 | if (cnt != mddev->raid_disks) { | 152 | if (cnt != mddev->raid_disks) { |
152 | printk("raid0: too few disks (%d of %d) - aborting!\n", | 153 | printk(KERN_ERR "raid0: too few disks (%d of %d) - " |
153 | cnt, mddev->raid_disks); | 154 | "aborting!\n", cnt, mddev->raid_disks); |
154 | goto abort; | 155 | goto abort; |
155 | } | 156 | } |
156 | zone->nb_dev = cnt; | 157 | zone->nb_dev = cnt; |
157 | zone->size = smallest->size * cnt; | 158 | zone->sectors = smallest->size * cnt * 2; |
158 | zone->zone_offset = 0; | 159 | zone->zone_start = 0; |
159 | 160 | ||
160 | current_offset = smallest->size; | 161 | current_start = smallest->size * 2; |
161 | curr_zone_offset = zone->size; | 162 | curr_zone_start = zone->sectors; |
162 | 163 | ||
163 | /* now do the other zones */ | 164 | /* now do the other zones */ |
164 | for (i = 1; i < conf->nr_strip_zones; i++) | 165 | for (i = 1; i < conf->nr_strip_zones; i++) |
@@ -166,40 +167,41 @@ static int create_strip_zones (mddev_t *mddev) | |||
166 | zone = conf->strip_zone + i; | 167 | zone = conf->strip_zone + i; |
167 | zone->dev = conf->strip_zone[i-1].dev + mddev->raid_disks; | 168 | zone->dev = conf->strip_zone[i-1].dev + mddev->raid_disks; |
168 | 169 | ||
169 | printk("raid0: zone %d\n", i); | 170 | printk(KERN_INFO "raid0: zone %d\n", i); |
170 | zone->dev_offset = current_offset; | 171 | zone->dev_start = current_start; |
171 | smallest = NULL; | 172 | smallest = NULL; |
172 | c = 0; | 173 | c = 0; |
173 | 174 | ||
174 | for (j=0; j<cnt; j++) { | 175 | for (j=0; j<cnt; j++) { |
175 | char b[BDEVNAME_SIZE]; | 176 | char b[BDEVNAME_SIZE]; |
176 | rdev = conf->strip_zone[0].dev[j]; | 177 | rdev = conf->strip_zone[0].dev[j]; |
177 | printk("raid0: checking %s ...", bdevname(rdev->bdev,b)); | 178 | printk(KERN_INFO "raid0: checking %s ...", |
178 | if (rdev->size > current_offset) | 179 | bdevname(rdev->bdev, b)); |
179 | { | 180 | if (rdev->size > current_start / 2) { |
180 | printk(" contained as device %d\n", c); | 181 | printk(KERN_INFO " contained as device %d\n", |
182 | c); | ||
181 | zone->dev[c] = rdev; | 183 | zone->dev[c] = rdev; |
182 | c++; | 184 | c++; |
183 | if (!smallest || (rdev->size <smallest->size)) { | 185 | if (!smallest || (rdev->size <smallest->size)) { |
184 | smallest = rdev; | 186 | smallest = rdev; |
185 | printk(" (%llu) is smallest!.\n", | 187 | printk(KERN_INFO " (%llu) is smallest!.\n", |
186 | (unsigned long long)rdev->size); | 188 | (unsigned long long)rdev->size); |
187 | } | 189 | } |
188 | } else | 190 | } else |
189 | printk(" nope.\n"); | 191 | printk(KERN_INFO " nope.\n"); |
190 | } | 192 | } |
191 | 193 | ||
192 | zone->nb_dev = c; | 194 | zone->nb_dev = c; |
193 | zone->size = (smallest->size - current_offset) * c; | 195 | zone->sectors = (smallest->size * 2 - current_start) * c; |
194 | printk("raid0: zone->nb_dev: %d, size: %llu\n", | 196 | printk(KERN_INFO "raid0: zone->nb_dev: %d, sectors: %llu\n", |
195 | zone->nb_dev, (unsigned long long)zone->size); | 197 | zone->nb_dev, (unsigned long long)zone->sectors); |
196 | 198 | ||
197 | zone->zone_offset = curr_zone_offset; | 199 | zone->zone_start = curr_zone_start; |
198 | curr_zone_offset += zone->size; | 200 | curr_zone_start += zone->sectors; |
199 | 201 | ||
200 | current_offset = smallest->size; | 202 | current_start = smallest->size * 2; |
201 | printk("raid0: current zone offset: %llu\n", | 203 | printk(KERN_INFO "raid0: current zone start: %llu\n", |
202 | (unsigned long long)current_offset); | 204 | (unsigned long long)current_start); |
203 | } | 205 | } |
204 | 206 | ||
205 | /* Now find appropriate hash spacing. | 207 | /* Now find appropriate hash spacing. |
@@ -210,16 +212,16 @@ static int create_strip_zones (mddev_t *mddev) | |||
210 | * strip though as it's size has no bearing on the efficacy of the hash | 212 | * strip though as it's size has no bearing on the efficacy of the hash |
211 | * table. | 213 | * table. |
212 | */ | 214 | */ |
213 | conf->hash_spacing = curr_zone_offset; | 215 | conf->spacing = curr_zone_start; |
214 | min_spacing = curr_zone_offset; | 216 | min_spacing = curr_zone_start; |
215 | sector_div(min_spacing, PAGE_SIZE/sizeof(struct strip_zone*)); | 217 | sector_div(min_spacing, PAGE_SIZE/sizeof(struct strip_zone*)); |
216 | for (i=0; i < conf->nr_strip_zones-1; i++) { | 218 | for (i=0; i < conf->nr_strip_zones-1; i++) { |
217 | sector_t sz = 0; | 219 | sector_t s = 0; |
218 | for (j=i; j<conf->nr_strip_zones-1 && | 220 | for (j = i; j < conf->nr_strip_zones - 1 && |
219 | sz < min_spacing ; j++) | 221 | s < min_spacing; j++) |
220 | sz += conf->strip_zone[j].size; | 222 | s += conf->strip_zone[j].sectors; |
221 | if (sz >= min_spacing && sz < conf->hash_spacing) | 223 | if (s >= min_spacing && s < conf->spacing) |
222 | conf->hash_spacing = sz; | 224 | conf->spacing = s; |
223 | } | 225 | } |
224 | 226 | ||
225 | mddev->queue->unplug_fn = raid0_unplug; | 227 | mddev->queue->unplug_fn = raid0_unplug; |
@@ -227,7 +229,7 @@ static int create_strip_zones (mddev_t *mddev) | |||
227 | mddev->queue->backing_dev_info.congested_fn = raid0_congested; | 229 | mddev->queue->backing_dev_info.congested_fn = raid0_congested; |
228 | mddev->queue->backing_dev_info.congested_data = mddev; | 230 | mddev->queue->backing_dev_info.congested_data = mddev; |
229 | 231 | ||
230 | printk("raid0: done.\n"); | 232 | printk(KERN_INFO "raid0: done.\n"); |
231 | return 0; | 233 | return 0; |
232 | abort: | 234 | abort: |
233 | return 1; | 235 | return 1; |
@@ -262,10 +264,9 @@ static int raid0_mergeable_bvec(struct request_queue *q, | |||
262 | static int raid0_run (mddev_t *mddev) | 264 | static int raid0_run (mddev_t *mddev) |
263 | { | 265 | { |
264 | unsigned cur=0, i=0, nb_zone; | 266 | unsigned cur=0, i=0, nb_zone; |
265 | s64 size; | 267 | s64 sectors; |
266 | raid0_conf_t *conf; | 268 | raid0_conf_t *conf; |
267 | mdk_rdev_t *rdev; | 269 | mdk_rdev_t *rdev; |
268 | struct list_head *tmp; | ||
269 | 270 | ||
270 | if (mddev->chunk_size == 0) { | 271 | if (mddev->chunk_size == 0) { |
271 | printk(KERN_ERR "md/raid0: non-zero chunk size required.\n"); | 272 | printk(KERN_ERR "md/raid0: non-zero chunk size required.\n"); |
@@ -291,54 +292,54 @@ static int raid0_run (mddev_t *mddev) | |||
291 | 292 | ||
292 | /* calculate array device size */ | 293 | /* calculate array device size */ |
293 | mddev->array_sectors = 0; | 294 | mddev->array_sectors = 0; |
294 | rdev_for_each(rdev, tmp, mddev) | 295 | list_for_each_entry(rdev, &mddev->disks, same_set) |
295 | mddev->array_sectors += rdev->size * 2; | 296 | mddev->array_sectors += rdev->size * 2; |
296 | 297 | ||
297 | printk("raid0 : md_size is %llu blocks.\n", | 298 | printk(KERN_INFO "raid0 : md_size is %llu sectors.\n", |
298 | (unsigned long long)mddev->array_sectors / 2); | 299 | (unsigned long long)mddev->array_sectors); |
299 | printk("raid0 : conf->hash_spacing is %llu blocks.\n", | 300 | printk(KERN_INFO "raid0 : conf->spacing is %llu sectors.\n", |
300 | (unsigned long long)conf->hash_spacing); | 301 | (unsigned long long)conf->spacing); |
301 | { | 302 | { |
302 | sector_t s = mddev->array_sectors / 2; | 303 | sector_t s = mddev->array_sectors; |
303 | sector_t space = conf->hash_spacing; | 304 | sector_t space = conf->spacing; |
304 | int round; | 305 | int round; |
305 | conf->preshift = 0; | 306 | conf->sector_shift = 0; |
306 | if (sizeof(sector_t) > sizeof(u32)) { | 307 | if (sizeof(sector_t) > sizeof(u32)) { |
307 | /*shift down space and s so that sector_div will work */ | 308 | /*shift down space and s so that sector_div will work */ |
308 | while (space > (sector_t) (~(u32)0)) { | 309 | while (space > (sector_t) (~(u32)0)) { |
309 | s >>= 1; | 310 | s >>= 1; |
310 | space >>= 1; | 311 | space >>= 1; |
311 | s += 1; /* force round-up */ | 312 | s += 1; /* force round-up */ |
312 | conf->preshift++; | 313 | conf->sector_shift++; |
313 | } | 314 | } |
314 | } | 315 | } |
315 | round = sector_div(s, (u32)space) ? 1 : 0; | 316 | round = sector_div(s, (u32)space) ? 1 : 0; |
316 | nb_zone = s + round; | 317 | nb_zone = s + round; |
317 | } | 318 | } |
318 | printk("raid0 : nb_zone is %d.\n", nb_zone); | 319 | printk(KERN_INFO "raid0 : nb_zone is %d.\n", nb_zone); |
319 | 320 | ||
320 | printk("raid0 : Allocating %Zd bytes for hash.\n", | 321 | printk(KERN_INFO "raid0 : Allocating %zu bytes for hash.\n", |
321 | nb_zone*sizeof(struct strip_zone*)); | 322 | nb_zone*sizeof(struct strip_zone*)); |
322 | conf->hash_table = kmalloc (sizeof (struct strip_zone *)*nb_zone, GFP_KERNEL); | 323 | conf->hash_table = kmalloc (sizeof (struct strip_zone *)*nb_zone, GFP_KERNEL); |
323 | if (!conf->hash_table) | 324 | if (!conf->hash_table) |
324 | goto out_free_conf; | 325 | goto out_free_conf; |
325 | size = conf->strip_zone[cur].size; | 326 | sectors = conf->strip_zone[cur].sectors; |
326 | 327 | ||
327 | conf->hash_table[0] = conf->strip_zone + cur; | 328 | conf->hash_table[0] = conf->strip_zone + cur; |
328 | for (i=1; i< nb_zone; i++) { | 329 | for (i=1; i< nb_zone; i++) { |
329 | while (size <= conf->hash_spacing) { | 330 | while (sectors <= conf->spacing) { |
330 | cur++; | 331 | cur++; |
331 | size += conf->strip_zone[cur].size; | 332 | sectors += conf->strip_zone[cur].sectors; |
332 | } | 333 | } |
333 | size -= conf->hash_spacing; | 334 | sectors -= conf->spacing; |
334 | conf->hash_table[i] = conf->strip_zone + cur; | 335 | conf->hash_table[i] = conf->strip_zone + cur; |
335 | } | 336 | } |
336 | if (conf->preshift) { | 337 | if (conf->sector_shift) { |
337 | conf->hash_spacing >>= conf->preshift; | 338 | conf->spacing >>= conf->sector_shift; |
338 | /* round hash_spacing up so when we divide by it, we | 339 | /* round spacing up so when we divide by it, we |
339 | * err on the side of too-low, which is safest | 340 | * err on the side of too-low, which is safest |
340 | */ | 341 | */ |
341 | conf->hash_spacing++; | 342 | conf->spacing++; |
342 | } | 343 | } |
343 | 344 | ||
344 | /* calculate the max read-ahead size. | 345 | /* calculate the max read-ahead size. |
@@ -387,12 +388,12 @@ static int raid0_stop (mddev_t *mddev) | |||
387 | static int raid0_make_request (struct request_queue *q, struct bio *bio) | 388 | static int raid0_make_request (struct request_queue *q, struct bio *bio) |
388 | { | 389 | { |
389 | mddev_t *mddev = q->queuedata; | 390 | mddev_t *mddev = q->queuedata; |
390 | unsigned int sect_in_chunk, chunksize_bits, chunk_size, chunk_sects; | 391 | unsigned int sect_in_chunk, chunksect_bits, chunk_sects; |
391 | raid0_conf_t *conf = mddev_to_conf(mddev); | 392 | raid0_conf_t *conf = mddev_to_conf(mddev); |
392 | struct strip_zone *zone; | 393 | struct strip_zone *zone; |
393 | mdk_rdev_t *tmp_dev; | 394 | mdk_rdev_t *tmp_dev; |
394 | sector_t chunk; | 395 | sector_t chunk; |
395 | sector_t block, rsect; | 396 | sector_t sector, rsect; |
396 | const int rw = bio_data_dir(bio); | 397 | const int rw = bio_data_dir(bio); |
397 | int cpu; | 398 | int cpu; |
398 | 399 | ||
@@ -407,11 +408,9 @@ static int raid0_make_request (struct request_queue *q, struct bio *bio) | |||
407 | bio_sectors(bio)); | 408 | bio_sectors(bio)); |
408 | part_stat_unlock(); | 409 | part_stat_unlock(); |
409 | 410 | ||
410 | chunk_size = mddev->chunk_size >> 10; | ||
411 | chunk_sects = mddev->chunk_size >> 9; | 411 | chunk_sects = mddev->chunk_size >> 9; |
412 | chunksize_bits = ffz(~chunk_size); | 412 | chunksect_bits = ffz(~chunk_sects); |
413 | block = bio->bi_sector >> 1; | 413 | sector = bio->bi_sector; |
414 | |||
415 | 414 | ||
416 | if (unlikely(chunk_sects < (bio->bi_sector & (chunk_sects - 1)) + (bio->bi_size >> 9))) { | 415 | if (unlikely(chunk_sects < (bio->bi_sector & (chunk_sects - 1)) + (bio->bi_size >> 9))) { |
417 | struct bio_pair *bp; | 416 | struct bio_pair *bp; |
@@ -434,28 +433,27 @@ static int raid0_make_request (struct request_queue *q, struct bio *bio) | |||
434 | 433 | ||
435 | 434 | ||
436 | { | 435 | { |
437 | sector_t x = block >> conf->preshift; | 436 | sector_t x = sector >> conf->sector_shift; |
438 | sector_div(x, (u32)conf->hash_spacing); | 437 | sector_div(x, (u32)conf->spacing); |
439 | zone = conf->hash_table[x]; | 438 | zone = conf->hash_table[x]; |
440 | } | 439 | } |
441 | 440 | ||
442 | while (block >= (zone->zone_offset + zone->size)) | 441 | while (sector >= zone->zone_start + zone->sectors) |
443 | zone++; | 442 | zone++; |
444 | 443 | ||
445 | sect_in_chunk = bio->bi_sector & ((chunk_size<<1) -1); | 444 | sect_in_chunk = bio->bi_sector & (chunk_sects - 1); |
446 | 445 | ||
447 | 446 | ||
448 | { | 447 | { |
449 | sector_t x = (block - zone->zone_offset) >> chunksize_bits; | 448 | sector_t x = (sector - zone->zone_start) >> chunksect_bits; |
450 | 449 | ||
451 | sector_div(x, zone->nb_dev); | 450 | sector_div(x, zone->nb_dev); |
452 | chunk = x; | 451 | chunk = x; |
453 | 452 | ||
454 | x = block >> chunksize_bits; | 453 | x = sector >> chunksect_bits; |
455 | tmp_dev = zone->dev[sector_div(x, zone->nb_dev)]; | 454 | tmp_dev = zone->dev[sector_div(x, zone->nb_dev)]; |
456 | } | 455 | } |
457 | rsect = (((chunk << chunksize_bits) + zone->dev_offset)<<1) | 456 | rsect = (chunk << chunksect_bits) + zone->dev_start + sect_in_chunk; |
458 | + sect_in_chunk; | ||
459 | 457 | ||
460 | bio->bi_bdev = tmp_dev->bdev; | 458 | bio->bi_bdev = tmp_dev->bdev; |
461 | bio->bi_sector = rsect + tmp_dev->data_offset; | 459 | bio->bi_sector = rsect + tmp_dev->data_offset; |
@@ -467,7 +465,7 @@ static int raid0_make_request (struct request_queue *q, struct bio *bio) | |||
467 | 465 | ||
468 | bad_map: | 466 | bad_map: |
469 | printk("raid0_make_request bug: can't convert block across chunks" | 467 | printk("raid0_make_request bug: can't convert block across chunks" |
470 | " or bigger than %dk %llu %d\n", chunk_size, | 468 | " or bigger than %dk %llu %d\n", chunk_sects / 2, |
471 | (unsigned long long)bio->bi_sector, bio->bi_size >> 10); | 469 | (unsigned long long)bio->bi_sector, bio->bi_size >> 10); |
472 | 470 | ||
473 | bio_io_error(bio); | 471 | bio_io_error(bio); |
@@ -492,10 +490,10 @@ static void raid0_status (struct seq_file *seq, mddev_t *mddev) | |||
492 | seq_printf(seq, "%s/", bdevname( | 490 | seq_printf(seq, "%s/", bdevname( |
493 | conf->strip_zone[j].dev[k]->bdev,b)); | 491 | conf->strip_zone[j].dev[k]->bdev,b)); |
494 | 492 | ||
495 | seq_printf(seq, "] zo=%d do=%d s=%d\n", | 493 | seq_printf(seq, "] zs=%d ds=%d s=%d\n", |
496 | conf->strip_zone[j].zone_offset, | 494 | conf->strip_zone[j].zone_start, |
497 | conf->strip_zone[j].dev_offset, | 495 | conf->strip_zone[j].dev_start, |
498 | conf->strip_zone[j].size); | 496 | conf->strip_zone[j].sectors); |
499 | } | 497 | } |
500 | #endif | 498 | #endif |
501 | seq_printf(seq, " %dk chunks", mddev->chunk_size/1024); | 499 | seq_printf(seq, " %dk chunks", mddev->chunk_size/1024); |
diff --git a/drivers/md/raid1.c b/drivers/md/raid1.c index 9c788e2489b1..7b4f5f7155d8 100644 --- a/drivers/md/raid1.c +++ b/drivers/md/raid1.c | |||
@@ -1016,12 +1016,16 @@ static void error(mddev_t *mddev, mdk_rdev_t *rdev) | |||
1016 | * else mark the drive as failed | 1016 | * else mark the drive as failed |
1017 | */ | 1017 | */ |
1018 | if (test_bit(In_sync, &rdev->flags) | 1018 | if (test_bit(In_sync, &rdev->flags) |
1019 | && (conf->raid_disks - mddev->degraded) == 1) | 1019 | && (conf->raid_disks - mddev->degraded) == 1) { |
1020 | /* | 1020 | /* |
1021 | * Don't fail the drive, act as though we were just a | 1021 | * Don't fail the drive, act as though we were just a |
1022 | * normal single drive | 1022 | * normal single drive. |
1023 | * However don't try a recovery from this drive as | ||
1024 | * it is very likely to fail. | ||
1023 | */ | 1025 | */ |
1026 | mddev->recovery_disabled = 1; | ||
1024 | return; | 1027 | return; |
1028 | } | ||
1025 | if (test_and_clear_bit(In_sync, &rdev->flags)) { | 1029 | if (test_and_clear_bit(In_sync, &rdev->flags)) { |
1026 | unsigned long flags; | 1030 | unsigned long flags; |
1027 | spin_lock_irqsave(&conf->device_lock, flags); | 1031 | spin_lock_irqsave(&conf->device_lock, flags); |
@@ -1919,7 +1923,6 @@ static int run(mddev_t *mddev) | |||
1919 | int i, j, disk_idx; | 1923 | int i, j, disk_idx; |
1920 | mirror_info_t *disk; | 1924 | mirror_info_t *disk; |
1921 | mdk_rdev_t *rdev; | 1925 | mdk_rdev_t *rdev; |
1922 | struct list_head *tmp; | ||
1923 | 1926 | ||
1924 | if (mddev->level != 1) { | 1927 | if (mddev->level != 1) { |
1925 | printk("raid1: %s: raid level not set to mirroring (%d)\n", | 1928 | printk("raid1: %s: raid level not set to mirroring (%d)\n", |
@@ -1964,7 +1967,7 @@ static int run(mddev_t *mddev) | |||
1964 | spin_lock_init(&conf->device_lock); | 1967 | spin_lock_init(&conf->device_lock); |
1965 | mddev->queue->queue_lock = &conf->device_lock; | 1968 | mddev->queue->queue_lock = &conf->device_lock; |
1966 | 1969 | ||
1967 | rdev_for_each(rdev, tmp, mddev) { | 1970 | list_for_each_entry(rdev, &mddev->disks, same_set) { |
1968 | disk_idx = rdev->raid_disk; | 1971 | disk_idx = rdev->raid_disk; |
1969 | if (disk_idx >= mddev->raid_disks | 1972 | if (disk_idx >= mddev->raid_disks |
1970 | || disk_idx < 0) | 1973 | || disk_idx < 0) |
diff --git a/drivers/md/raid10.c b/drivers/md/raid10.c index 970a96ef9b18..6736d6dff981 100644 --- a/drivers/md/raid10.c +++ b/drivers/md/raid10.c | |||
@@ -2025,7 +2025,6 @@ static int run(mddev_t *mddev) | |||
2025 | int i, disk_idx; | 2025 | int i, disk_idx; |
2026 | mirror_info_t *disk; | 2026 | mirror_info_t *disk; |
2027 | mdk_rdev_t *rdev; | 2027 | mdk_rdev_t *rdev; |
2028 | struct list_head *tmp; | ||
2029 | int nc, fc, fo; | 2028 | int nc, fc, fo; |
2030 | sector_t stride, size; | 2029 | sector_t stride, size; |
2031 | 2030 | ||
@@ -2108,7 +2107,7 @@ static int run(mddev_t *mddev) | |||
2108 | spin_lock_init(&conf->device_lock); | 2107 | spin_lock_init(&conf->device_lock); |
2109 | mddev->queue->queue_lock = &conf->device_lock; | 2108 | mddev->queue->queue_lock = &conf->device_lock; |
2110 | 2109 | ||
2111 | rdev_for_each(rdev, tmp, mddev) { | 2110 | list_for_each_entry(rdev, &mddev->disks, same_set) { |
2112 | disk_idx = rdev->raid_disk; | 2111 | disk_idx = rdev->raid_disk; |
2113 | if (disk_idx >= mddev->raid_disks | 2112 | if (disk_idx >= mddev->raid_disks |
2114 | || disk_idx < 0) | 2113 | || disk_idx < 0) |
diff --git a/drivers/md/raid5.c b/drivers/md/raid5.c index a36a7435edf5..a5ba080d303b 100644 --- a/drivers/md/raid5.c +++ b/drivers/md/raid5.c | |||
@@ -3998,7 +3998,6 @@ static int run(mddev_t *mddev) | |||
3998 | int raid_disk, memory; | 3998 | int raid_disk, memory; |
3999 | mdk_rdev_t *rdev; | 3999 | mdk_rdev_t *rdev; |
4000 | struct disk_info *disk; | 4000 | struct disk_info *disk; |
4001 | struct list_head *tmp; | ||
4002 | int working_disks = 0; | 4001 | int working_disks = 0; |
4003 | 4002 | ||
4004 | if (mddev->level != 5 && mddev->level != 4 && mddev->level != 6) { | 4003 | if (mddev->level != 5 && mddev->level != 4 && mddev->level != 6) { |
@@ -4108,7 +4107,7 @@ static int run(mddev_t *mddev) | |||
4108 | 4107 | ||
4109 | pr_debug("raid5: run(%s) called.\n", mdname(mddev)); | 4108 | pr_debug("raid5: run(%s) called.\n", mdname(mddev)); |
4110 | 4109 | ||
4111 | rdev_for_each(rdev, tmp, mddev) { | 4110 | list_for_each_entry(rdev, &mddev->disks, same_set) { |
4112 | raid_disk = rdev->raid_disk; | 4111 | raid_disk = rdev->raid_disk; |
4113 | if (raid_disk >= conf->raid_disks | 4112 | if (raid_disk >= conf->raid_disks |
4114 | || raid_disk < 0) | 4113 | || raid_disk < 0) |
@@ -4533,7 +4532,6 @@ static int raid5_start_reshape(mddev_t *mddev) | |||
4533 | { | 4532 | { |
4534 | raid5_conf_t *conf = mddev_to_conf(mddev); | 4533 | raid5_conf_t *conf = mddev_to_conf(mddev); |
4535 | mdk_rdev_t *rdev; | 4534 | mdk_rdev_t *rdev; |
4536 | struct list_head *rtmp; | ||
4537 | int spares = 0; | 4535 | int spares = 0; |
4538 | int added_devices = 0; | 4536 | int added_devices = 0; |
4539 | unsigned long flags; | 4537 | unsigned long flags; |
@@ -4541,7 +4539,7 @@ static int raid5_start_reshape(mddev_t *mddev) | |||
4541 | if (test_bit(MD_RECOVERY_RUNNING, &mddev->recovery)) | 4539 | if (test_bit(MD_RECOVERY_RUNNING, &mddev->recovery)) |
4542 | return -EBUSY; | 4540 | return -EBUSY; |
4543 | 4541 | ||
4544 | rdev_for_each(rdev, rtmp, mddev) | 4542 | list_for_each_entry(rdev, &mddev->disks, same_set) |
4545 | if (rdev->raid_disk < 0 && | 4543 | if (rdev->raid_disk < 0 && |
4546 | !test_bit(Faulty, &rdev->flags)) | 4544 | !test_bit(Faulty, &rdev->flags)) |
4547 | spares++; | 4545 | spares++; |
@@ -4563,7 +4561,7 @@ static int raid5_start_reshape(mddev_t *mddev) | |||
4563 | /* Add some new drives, as many as will fit. | 4561 | /* Add some new drives, as many as will fit. |
4564 | * We know there are enough to make the newly sized array work. | 4562 | * We know there are enough to make the newly sized array work. |
4565 | */ | 4563 | */ |
4566 | rdev_for_each(rdev, rtmp, mddev) | 4564 | list_for_each_entry(rdev, &mddev->disks, same_set) |
4567 | if (rdev->raid_disk < 0 && | 4565 | if (rdev->raid_disk < 0 && |
4568 | !test_bit(Faulty, &rdev->flags)) { | 4566 | !test_bit(Faulty, &rdev->flags)) { |
4569 | if (raid5_add_disk(mddev, rdev) == 0) { | 4567 | if (raid5_add_disk(mddev, rdev) == 0) { |
diff --git a/drivers/misc/Kconfig b/drivers/misc/Kconfig index fee7304102af..3949a1c73451 100644 --- a/drivers/misc/Kconfig +++ b/drivers/misc/Kconfig | |||
@@ -498,6 +498,18 @@ config SGI_GRU_DEBUG | |||
498 | This option enables addition debugging code for the SGI GRU driver. If | 498 | This option enables addition debugging code for the SGI GRU driver. If |
499 | you are unsure, say N. | 499 | you are unsure, say N. |
500 | 500 | ||
501 | config DELL_LAPTOP | ||
502 | tristate "Dell Laptop Extras (EXPERIMENTAL)" | ||
503 | depends on X86 | ||
504 | depends on DCDBAS | ||
505 | depends on EXPERIMENTAL | ||
506 | depends on BACKLIGHT_CLASS_DEVICE | ||
507 | depends on RFKILL | ||
508 | default n | ||
509 | ---help--- | ||
510 | This driver adds support for rfkill and backlight control to Dell | ||
511 | laptops. | ||
512 | |||
501 | source "drivers/misc/c2port/Kconfig" | 513 | source "drivers/misc/c2port/Kconfig" |
502 | 514 | ||
503 | endif # MISC_DEVICES | 515 | endif # MISC_DEVICES |
diff --git a/drivers/misc/Makefile b/drivers/misc/Makefile index 817f7f5ab3bd..5de863a0e395 100644 --- a/drivers/misc/Makefile +++ b/drivers/misc/Makefile | |||
@@ -18,6 +18,7 @@ obj-$(CONFIG_ICS932S401) += ics932s401.o | |||
18 | obj-$(CONFIG_TC1100_WMI) += tc1100-wmi.o | 18 | obj-$(CONFIG_TC1100_WMI) += tc1100-wmi.o |
19 | obj-$(CONFIG_LKDTM) += lkdtm.o | 19 | obj-$(CONFIG_LKDTM) += lkdtm.o |
20 | obj-$(CONFIG_TIFM_CORE) += tifm_core.o | 20 | obj-$(CONFIG_TIFM_CORE) += tifm_core.o |
21 | obj-$(CONFIG_DELL_LAPTOP) += dell-laptop.o | ||
21 | obj-$(CONFIG_TIFM_7XX1) += tifm_7xx1.o | 22 | obj-$(CONFIG_TIFM_7XX1) += tifm_7xx1.o |
22 | obj-$(CONFIG_PHANTOM) += phantom.o | 23 | obj-$(CONFIG_PHANTOM) += phantom.o |
23 | obj-$(CONFIG_SGI_IOC4) += ioc4.o | 24 | obj-$(CONFIG_SGI_IOC4) += ioc4.o |
diff --git a/drivers/misc/dell-laptop.c b/drivers/misc/dell-laptop.c new file mode 100644 index 000000000000..4d33a2068b7a --- /dev/null +++ b/drivers/misc/dell-laptop.c | |||
@@ -0,0 +1,436 @@ | |||
1 | /* | ||
2 | * Driver for Dell laptop extras | ||
3 | * | ||
4 | * Copyright (c) Red Hat <mjg@redhat.com> | ||
5 | * | ||
6 | * Based on documentation in the libsmbios package, Copyright (C) 2005 Dell | ||
7 | * Inc. | ||
8 | * | ||
9 | * This program is free software; you can redistribute it and/or modify | ||
10 | * it under the terms of the GNU General Public License version 2 as | ||
11 | * published by the Free Software Foundation. | ||
12 | */ | ||
13 | |||
14 | #include <linux/module.h> | ||
15 | #include <linux/kernel.h> | ||
16 | #include <linux/init.h> | ||
17 | #include <linux/platform_device.h> | ||
18 | #include <linux/backlight.h> | ||
19 | #include <linux/err.h> | ||
20 | #include <linux/dmi.h> | ||
21 | #include <linux/io.h> | ||
22 | #include <linux/rfkill.h> | ||
23 | #include <linux/power_supply.h> | ||
24 | #include <linux/acpi.h> | ||
25 | #include "../firmware/dcdbas.h" | ||
26 | |||
27 | #define BRIGHTNESS_TOKEN 0x7d | ||
28 | |||
29 | /* This structure will be modified by the firmware when we enter | ||
30 | * system management mode, hence the volatiles */ | ||
31 | |||
32 | struct calling_interface_buffer { | ||
33 | u16 class; | ||
34 | u16 select; | ||
35 | volatile u32 input[4]; | ||
36 | volatile u32 output[4]; | ||
37 | } __packed; | ||
38 | |||
39 | struct calling_interface_token { | ||
40 | u16 tokenID; | ||
41 | u16 location; | ||
42 | union { | ||
43 | u16 value; | ||
44 | u16 stringlength; | ||
45 | }; | ||
46 | }; | ||
47 | |||
48 | struct calling_interface_structure { | ||
49 | struct dmi_header header; | ||
50 | u16 cmdIOAddress; | ||
51 | u8 cmdIOCode; | ||
52 | u32 supportedCmds; | ||
53 | struct calling_interface_token tokens[]; | ||
54 | } __packed; | ||
55 | |||
56 | static int da_command_address; | ||
57 | static int da_command_code; | ||
58 | static int da_num_tokens; | ||
59 | static struct calling_interface_token *da_tokens; | ||
60 | |||
61 | static struct backlight_device *dell_backlight_device; | ||
62 | static struct rfkill *wifi_rfkill; | ||
63 | static struct rfkill *bluetooth_rfkill; | ||
64 | static struct rfkill *wwan_rfkill; | ||
65 | |||
66 | static const struct dmi_system_id __initdata dell_device_table[] = { | ||
67 | { | ||
68 | .ident = "Dell laptop", | ||
69 | .matches = { | ||
70 | DMI_MATCH(DMI_SYS_VENDOR, "Dell Inc."), | ||
71 | DMI_MATCH(DMI_CHASSIS_TYPE, "8"), | ||
72 | }, | ||
73 | }, | ||
74 | { } | ||
75 | }; | ||
76 | |||
77 | static void parse_da_table(const struct dmi_header *dm) | ||
78 | { | ||
79 | /* Final token is a terminator, so we don't want to copy it */ | ||
80 | int tokens = (dm->length-11)/sizeof(struct calling_interface_token)-1; | ||
81 | struct calling_interface_structure *table = | ||
82 | container_of(dm, struct calling_interface_structure, header); | ||
83 | |||
84 | /* 4 bytes of table header, plus 7 bytes of Dell header, plus at least | ||
85 | 6 bytes of entry */ | ||
86 | |||
87 | if (dm->length < 17) | ||
88 | return; | ||
89 | |||
90 | da_command_address = table->cmdIOAddress; | ||
91 | da_command_code = table->cmdIOCode; | ||
92 | |||
93 | da_tokens = krealloc(da_tokens, (da_num_tokens + tokens) * | ||
94 | sizeof(struct calling_interface_token), | ||
95 | GFP_KERNEL); | ||
96 | |||
97 | if (!da_tokens) | ||
98 | return; | ||
99 | |||
100 | memcpy(da_tokens+da_num_tokens, table->tokens, | ||
101 | sizeof(struct calling_interface_token) * tokens); | ||
102 | |||
103 | da_num_tokens += tokens; | ||
104 | } | ||
105 | |||
106 | static void find_tokens(const struct dmi_header *dm) | ||
107 | { | ||
108 | switch (dm->type) { | ||
109 | case 0xd4: /* Indexed IO */ | ||
110 | break; | ||
111 | case 0xd5: /* Protected Area Type 1 */ | ||
112 | break; | ||
113 | case 0xd6: /* Protected Area Type 2 */ | ||
114 | break; | ||
115 | case 0xda: /* Calling interface */ | ||
116 | parse_da_table(dm); | ||
117 | break; | ||
118 | } | ||
119 | } | ||
120 | |||
121 | static int find_token_location(int tokenid) | ||
122 | { | ||
123 | int i; | ||
124 | for (i = 0; i < da_num_tokens; i++) { | ||
125 | if (da_tokens[i].tokenID == tokenid) | ||
126 | return da_tokens[i].location; | ||
127 | } | ||
128 | |||
129 | return -1; | ||
130 | } | ||
131 | |||
132 | static struct calling_interface_buffer * | ||
133 | dell_send_request(struct calling_interface_buffer *buffer, int class, | ||
134 | int select) | ||
135 | { | ||
136 | struct smi_cmd command; | ||
137 | |||
138 | command.magic = SMI_CMD_MAGIC; | ||
139 | command.command_address = da_command_address; | ||
140 | command.command_code = da_command_code; | ||
141 | command.ebx = virt_to_phys(buffer); | ||
142 | command.ecx = 0x42534931; | ||
143 | |||
144 | buffer->class = class; | ||
145 | buffer->select = select; | ||
146 | |||
147 | dcdbas_smi_request(&command); | ||
148 | |||
149 | return buffer; | ||
150 | } | ||
151 | |||
152 | /* Derived from information in DellWirelessCtl.cpp: | ||
153 | Class 17, select 11 is radio control. It returns an array of 32-bit values. | ||
154 | |||
155 | result[0]: return code | ||
156 | result[1]: | ||
157 | Bit 0: Hardware switch supported | ||
158 | Bit 1: Wifi locator supported | ||
159 | Bit 2: Wifi is supported | ||
160 | Bit 3: Bluetooth is supported | ||
161 | Bit 4: WWAN is supported | ||
162 | Bit 5: Wireless keyboard supported | ||
163 | Bits 6-7: Reserved | ||
164 | Bit 8: Wifi is installed | ||
165 | Bit 9: Bluetooth is installed | ||
166 | Bit 10: WWAN is installed | ||
167 | Bits 11-15: Reserved | ||
168 | Bit 16: Hardware switch is on | ||
169 | Bit 17: Wifi is blocked | ||
170 | Bit 18: Bluetooth is blocked | ||
171 | Bit 19: WWAN is blocked | ||
172 | Bits 20-31: Reserved | ||
173 | result[2]: NVRAM size in bytes | ||
174 | result[3]: NVRAM format version number | ||
175 | */ | ||
176 | |||
177 | static int dell_rfkill_set(int radio, enum rfkill_state state) | ||
178 | { | ||
179 | struct calling_interface_buffer buffer; | ||
180 | int disable = (state == RFKILL_STATE_UNBLOCKED) ? 0 : 1; | ||
181 | |||
182 | memset(&buffer, 0, sizeof(struct calling_interface_buffer)); | ||
183 | buffer.input[0] = (1 | (radio<<8) | (disable << 16)); | ||
184 | dell_send_request(&buffer, 17, 11); | ||
185 | |||
186 | return 0; | ||
187 | } | ||
188 | |||
189 | static int dell_wifi_set(void *data, enum rfkill_state state) | ||
190 | { | ||
191 | return dell_rfkill_set(1, state); | ||
192 | } | ||
193 | |||
194 | static int dell_bluetooth_set(void *data, enum rfkill_state state) | ||
195 | { | ||
196 | return dell_rfkill_set(2, state); | ||
197 | } | ||
198 | |||
199 | static int dell_wwan_set(void *data, enum rfkill_state state) | ||
200 | { | ||
201 | return dell_rfkill_set(3, state); | ||
202 | } | ||
203 | |||
204 | static int dell_rfkill_get(int bit, enum rfkill_state *state) | ||
205 | { | ||
206 | struct calling_interface_buffer buffer; | ||
207 | int status; | ||
208 | int new_state = RFKILL_STATE_HARD_BLOCKED; | ||
209 | |||
210 | memset(&buffer, 0, sizeof(struct calling_interface_buffer)); | ||
211 | dell_send_request(&buffer, 17, 11); | ||
212 | status = buffer.output[1]; | ||
213 | |||
214 | if (status & (1<<16)) | ||
215 | new_state = RFKILL_STATE_SOFT_BLOCKED; | ||
216 | |||
217 | if (status & (1<<bit)) | ||
218 | *state = new_state; | ||
219 | else | ||
220 | *state = RFKILL_STATE_UNBLOCKED; | ||
221 | |||
222 | return 0; | ||
223 | } | ||
224 | |||
225 | static int dell_wifi_get(void *data, enum rfkill_state *state) | ||
226 | { | ||
227 | return dell_rfkill_get(17, state); | ||
228 | } | ||
229 | |||
230 | static int dell_bluetooth_get(void *data, enum rfkill_state *state) | ||
231 | { | ||
232 | return dell_rfkill_get(18, state); | ||
233 | } | ||
234 | |||
235 | static int dell_wwan_get(void *data, enum rfkill_state *state) | ||
236 | { | ||
237 | return dell_rfkill_get(19, state); | ||
238 | } | ||
239 | |||
240 | static int dell_setup_rfkill(void) | ||
241 | { | ||
242 | struct calling_interface_buffer buffer; | ||
243 | int status; | ||
244 | int ret; | ||
245 | |||
246 | memset(&buffer, 0, sizeof(struct calling_interface_buffer)); | ||
247 | dell_send_request(&buffer, 17, 11); | ||
248 | status = buffer.output[1]; | ||
249 | |||
250 | if ((status & (1<<2|1<<8)) == (1<<2|1<<8)) { | ||
251 | wifi_rfkill = rfkill_allocate(NULL, RFKILL_TYPE_WLAN); | ||
252 | if (!wifi_rfkill) | ||
253 | goto err_wifi; | ||
254 | wifi_rfkill->name = "dell-wifi"; | ||
255 | wifi_rfkill->toggle_radio = dell_wifi_set; | ||
256 | wifi_rfkill->get_state = dell_wifi_get; | ||
257 | ret = rfkill_register(wifi_rfkill); | ||
258 | if (ret) | ||
259 | goto err_wifi; | ||
260 | } | ||
261 | |||
262 | if ((status & (1<<3|1<<9)) == (1<<3|1<<9)) { | ||
263 | bluetooth_rfkill = rfkill_allocate(NULL, RFKILL_TYPE_BLUETOOTH); | ||
264 | if (!bluetooth_rfkill) | ||
265 | goto err_bluetooth; | ||
266 | bluetooth_rfkill->name = "dell-bluetooth"; | ||
267 | bluetooth_rfkill->toggle_radio = dell_bluetooth_set; | ||
268 | bluetooth_rfkill->get_state = dell_bluetooth_get; | ||
269 | ret = rfkill_register(bluetooth_rfkill); | ||
270 | if (ret) | ||
271 | goto err_bluetooth; | ||
272 | } | ||
273 | |||
274 | if ((status & (1<<4|1<<10)) == (1<<4|1<<10)) { | ||
275 | wwan_rfkill = rfkill_allocate(NULL, RFKILL_TYPE_WWAN); | ||
276 | if (!wwan_rfkill) | ||
277 | goto err_wwan; | ||
278 | wwan_rfkill->name = "dell-wwan"; | ||
279 | wwan_rfkill->toggle_radio = dell_wwan_set; | ||
280 | wwan_rfkill->get_state = dell_wwan_get; | ||
281 | ret = rfkill_register(wwan_rfkill); | ||
282 | if (ret) | ||
283 | goto err_wwan; | ||
284 | } | ||
285 | |||
286 | return 0; | ||
287 | err_wwan: | ||
288 | if (wwan_rfkill) | ||
289 | rfkill_free(wwan_rfkill); | ||
290 | if (bluetooth_rfkill) { | ||
291 | rfkill_unregister(bluetooth_rfkill); | ||
292 | bluetooth_rfkill = NULL; | ||
293 | } | ||
294 | err_bluetooth: | ||
295 | if (bluetooth_rfkill) | ||
296 | rfkill_free(bluetooth_rfkill); | ||
297 | if (wifi_rfkill) { | ||
298 | rfkill_unregister(wifi_rfkill); | ||
299 | wifi_rfkill = NULL; | ||
300 | } | ||
301 | err_wifi: | ||
302 | if (wifi_rfkill) | ||
303 | rfkill_free(wifi_rfkill); | ||
304 | |||
305 | return ret; | ||
306 | } | ||
307 | |||
308 | static int dell_send_intensity(struct backlight_device *bd) | ||
309 | { | ||
310 | struct calling_interface_buffer buffer; | ||
311 | |||
312 | memset(&buffer, 0, sizeof(struct calling_interface_buffer)); | ||
313 | buffer.input[0] = find_token_location(BRIGHTNESS_TOKEN); | ||
314 | buffer.input[1] = bd->props.brightness; | ||
315 | |||
316 | if (buffer.input[0] == -1) | ||
317 | return -ENODEV; | ||
318 | |||
319 | if (power_supply_is_system_supplied() > 0) | ||
320 | dell_send_request(&buffer, 1, 2); | ||
321 | else | ||
322 | dell_send_request(&buffer, 1, 1); | ||
323 | |||
324 | return 0; | ||
325 | } | ||
326 | |||
327 | static int dell_get_intensity(struct backlight_device *bd) | ||
328 | { | ||
329 | struct calling_interface_buffer buffer; | ||
330 | |||
331 | memset(&buffer, 0, sizeof(struct calling_interface_buffer)); | ||
332 | buffer.input[0] = find_token_location(BRIGHTNESS_TOKEN); | ||
333 | |||
334 | if (buffer.input[0] == -1) | ||
335 | return -ENODEV; | ||
336 | |||
337 | if (power_supply_is_system_supplied() > 0) | ||
338 | dell_send_request(&buffer, 0, 2); | ||
339 | else | ||
340 | dell_send_request(&buffer, 0, 1); | ||
341 | |||
342 | return buffer.output[1]; | ||
343 | } | ||
344 | |||
345 | static struct backlight_ops dell_ops = { | ||
346 | .get_brightness = dell_get_intensity, | ||
347 | .update_status = dell_send_intensity, | ||
348 | }; | ||
349 | |||
350 | static int __init dell_init(void) | ||
351 | { | ||
352 | struct calling_interface_buffer buffer; | ||
353 | int max_intensity = 0; | ||
354 | int ret; | ||
355 | |||
356 | if (!dmi_check_system(dell_device_table)) | ||
357 | return -ENODEV; | ||
358 | |||
359 | dmi_walk(find_tokens); | ||
360 | |||
361 | if (!da_tokens) { | ||
362 | printk(KERN_INFO "dell-laptop: Unable to find dmi tokens\n"); | ||
363 | return -ENODEV; | ||
364 | } | ||
365 | |||
366 | ret = dell_setup_rfkill(); | ||
367 | |||
368 | if (ret) { | ||
369 | printk(KERN_WARNING "dell-laptop: Unable to setup rfkill\n"); | ||
370 | goto out; | ||
371 | } | ||
372 | |||
373 | #ifdef CONFIG_ACPI | ||
374 | /* In the event of an ACPI backlight being available, don't | ||
375 | * register the platform controller. | ||
376 | */ | ||
377 | if (acpi_video_backlight_support()) | ||
378 | return 0; | ||
379 | #endif | ||
380 | |||
381 | memset(&buffer, 0, sizeof(struct calling_interface_buffer)); | ||
382 | buffer.input[0] = find_token_location(BRIGHTNESS_TOKEN); | ||
383 | |||
384 | if (buffer.input[0] != -1) { | ||
385 | dell_send_request(&buffer, 0, 2); | ||
386 | max_intensity = buffer.output[3]; | ||
387 | } | ||
388 | |||
389 | if (max_intensity) { | ||
390 | dell_backlight_device = backlight_device_register( | ||
391 | "dell_backlight", | ||
392 | NULL, NULL, | ||
393 | &dell_ops); | ||
394 | |||
395 | if (IS_ERR(dell_backlight_device)) { | ||
396 | ret = PTR_ERR(dell_backlight_device); | ||
397 | dell_backlight_device = NULL; | ||
398 | goto out; | ||
399 | } | ||
400 | |||
401 | dell_backlight_device->props.max_brightness = max_intensity; | ||
402 | dell_backlight_device->props.brightness = | ||
403 | dell_get_intensity(dell_backlight_device); | ||
404 | backlight_update_status(dell_backlight_device); | ||
405 | } | ||
406 | |||
407 | return 0; | ||
408 | out: | ||
409 | if (wifi_rfkill) | ||
410 | rfkill_unregister(wifi_rfkill); | ||
411 | if (bluetooth_rfkill) | ||
412 | rfkill_unregister(bluetooth_rfkill); | ||
413 | if (wwan_rfkill) | ||
414 | rfkill_unregister(wwan_rfkill); | ||
415 | kfree(da_tokens); | ||
416 | return ret; | ||
417 | } | ||
418 | |||
419 | static void __exit dell_exit(void) | ||
420 | { | ||
421 | backlight_device_unregister(dell_backlight_device); | ||
422 | if (wifi_rfkill) | ||
423 | rfkill_unregister(wifi_rfkill); | ||
424 | if (bluetooth_rfkill) | ||
425 | rfkill_unregister(bluetooth_rfkill); | ||
426 | if (wwan_rfkill) | ||
427 | rfkill_unregister(wwan_rfkill); | ||
428 | } | ||
429 | |||
430 | module_init(dell_init); | ||
431 | module_exit(dell_exit); | ||
432 | |||
433 | MODULE_AUTHOR("Matthew Garrett <mjg@redhat.com>"); | ||
434 | MODULE_DESCRIPTION("Dell laptop driver"); | ||
435 | MODULE_LICENSE("GPL"); | ||
436 | MODULE_ALIAS("dmi:*svnDellInc.:*:ct8:*"); | ||
diff --git a/drivers/mtd/devices/Kconfig b/drivers/mtd/devices/Kconfig index 6fde0a2e3567..bc33200535fc 100644 --- a/drivers/mtd/devices/Kconfig +++ b/drivers/mtd/devices/Kconfig | |||
@@ -120,6 +120,13 @@ config MTD_PHRAM | |||
120 | doesn't have access to, memory beyond the mem=xxx limit, nvram, | 120 | doesn't have access to, memory beyond the mem=xxx limit, nvram, |
121 | memory on the video card, etc... | 121 | memory on the video card, etc... |
122 | 122 | ||
123 | config MTD_PS3VRAM | ||
124 | tristate "PS3 video RAM" | ||
125 | depends on FB_PS3 | ||
126 | help | ||
127 | This driver allows you to use excess PS3 video RAM as volatile | ||
128 | storage or system swap. | ||
129 | |||
123 | config MTD_LART | 130 | config MTD_LART |
124 | tristate "28F160xx flash driver for LART" | 131 | tristate "28F160xx flash driver for LART" |
125 | depends on SA1100_LART | 132 | depends on SA1100_LART |
diff --git a/drivers/mtd/devices/Makefile b/drivers/mtd/devices/Makefile index 0993d5cf3923..e51521df4e40 100644 --- a/drivers/mtd/devices/Makefile +++ b/drivers/mtd/devices/Makefile | |||
@@ -16,3 +16,4 @@ obj-$(CONFIG_MTD_LART) += lart.o | |||
16 | obj-$(CONFIG_MTD_BLOCK2MTD) += block2mtd.o | 16 | obj-$(CONFIG_MTD_BLOCK2MTD) += block2mtd.o |
17 | obj-$(CONFIG_MTD_DATAFLASH) += mtd_dataflash.o | 17 | obj-$(CONFIG_MTD_DATAFLASH) += mtd_dataflash.o |
18 | obj-$(CONFIG_MTD_M25P80) += m25p80.o | 18 | obj-$(CONFIG_MTD_M25P80) += m25p80.o |
19 | obj-$(CONFIG_MTD_PS3VRAM) += ps3vram.o | ||
diff --git a/drivers/mtd/devices/ps3vram.c b/drivers/mtd/devices/ps3vram.c new file mode 100644 index 000000000000..d21e9beb7ed2 --- /dev/null +++ b/drivers/mtd/devices/ps3vram.c | |||
@@ -0,0 +1,768 @@ | |||
1 | /** | ||
2 | * ps3vram - Use extra PS3 video ram as MTD block device. | ||
3 | * | ||
4 | * Copyright (c) 2007-2008 Jim Paris <jim@jtan.com> | ||
5 | * Added support RSX DMA Vivien Chappelier <vivien.chappelier@free.fr> | ||
6 | */ | ||
7 | |||
8 | #include <linux/io.h> | ||
9 | #include <linux/mm.h> | ||
10 | #include <linux/init.h> | ||
11 | #include <linux/kernel.h> | ||
12 | #include <linux/list.h> | ||
13 | #include <linux/module.h> | ||
14 | #include <linux/moduleparam.h> | ||
15 | #include <linux/slab.h> | ||
16 | #include <linux/version.h> | ||
17 | #include <linux/gfp.h> | ||
18 | #include <linux/delay.h> | ||
19 | #include <linux/mtd/mtd.h> | ||
20 | |||
21 | #include <asm/lv1call.h> | ||
22 | #include <asm/ps3.h> | ||
23 | |||
24 | #define DEVICE_NAME "ps3vram" | ||
25 | |||
26 | #define XDR_BUF_SIZE (2 * 1024 * 1024) /* XDR buffer (must be 1MiB aligned) */ | ||
27 | #define XDR_IOIF 0x0c000000 | ||
28 | |||
29 | #define FIFO_BASE XDR_IOIF | ||
30 | #define FIFO_SIZE (64 * 1024) | ||
31 | |||
32 | #define DMA_PAGE_SIZE (4 * 1024) | ||
33 | |||
34 | #define CACHE_PAGE_SIZE (256 * 1024) | ||
35 | #define CACHE_PAGE_COUNT ((XDR_BUF_SIZE - FIFO_SIZE) / CACHE_PAGE_SIZE) | ||
36 | |||
37 | #define CACHE_OFFSET CACHE_PAGE_SIZE | ||
38 | #define FIFO_OFFSET 0 | ||
39 | |||
40 | #define CTRL_PUT 0x10 | ||
41 | #define CTRL_GET 0x11 | ||
42 | #define CTRL_TOP 0x15 | ||
43 | |||
44 | #define UPLOAD_SUBCH 1 | ||
45 | #define DOWNLOAD_SUBCH 2 | ||
46 | |||
47 | #define NV_MEMORY_TO_MEMORY_FORMAT_OFFSET_IN 0x0000030c | ||
48 | #define NV_MEMORY_TO_MEMORY_FORMAT_NOTIFY 0x00000104 | ||
49 | |||
50 | #define L1GPU_CONTEXT_ATTRIBUTE_FB_BLIT 0x601 | ||
51 | |||
52 | struct mtd_info ps3vram_mtd; | ||
53 | |||
54 | #define CACHE_PAGE_PRESENT 1 | ||
55 | #define CACHE_PAGE_DIRTY 2 | ||
56 | |||
57 | struct ps3vram_tag { | ||
58 | unsigned int address; | ||
59 | unsigned int flags; | ||
60 | }; | ||
61 | |||
62 | struct ps3vram_cache { | ||
63 | unsigned int page_count; | ||
64 | unsigned int page_size; | ||
65 | struct ps3vram_tag *tags; | ||
66 | }; | ||
67 | |||
68 | struct ps3vram_priv { | ||
69 | u64 memory_handle; | ||
70 | u64 context_handle; | ||
71 | u32 *ctrl; | ||
72 | u32 *reports; | ||
73 | u8 __iomem *ddr_base; | ||
74 | u8 *xdr_buf; | ||
75 | |||
76 | u32 *fifo_base; | ||
77 | u32 *fifo_ptr; | ||
78 | |||
79 | struct device *dev; | ||
80 | struct ps3vram_cache cache; | ||
81 | |||
82 | /* Used to serialize cache/DMA operations */ | ||
83 | struct mutex lock; | ||
84 | }; | ||
85 | |||
86 | #define DMA_NOTIFIER_HANDLE_BASE 0x66604200 /* first DMA notifier handle */ | ||
87 | #define DMA_NOTIFIER_OFFSET_BASE 0x1000 /* first DMA notifier offset */ | ||
88 | #define DMA_NOTIFIER_SIZE 0x40 | ||
89 | #define NOTIFIER 7 /* notifier used for completion report */ | ||
90 | |||
91 | /* A trailing '-' means to subtract off ps3fb_videomemory.size */ | ||
92 | char *size = "256M-"; | ||
93 | module_param(size, charp, 0); | ||
94 | MODULE_PARM_DESC(size, "memory size"); | ||
95 | |||
96 | static u32 *ps3vram_get_notifier(u32 *reports, int notifier) | ||
97 | { | ||
98 | return (void *) reports + | ||
99 | DMA_NOTIFIER_OFFSET_BASE + | ||
100 | DMA_NOTIFIER_SIZE * notifier; | ||
101 | } | ||
102 | |||
103 | static void ps3vram_notifier_reset(struct mtd_info *mtd) | ||
104 | { | ||
105 | int i; | ||
106 | |||
107 | struct ps3vram_priv *priv = mtd->priv; | ||
108 | u32 *notify = ps3vram_get_notifier(priv->reports, NOTIFIER); | ||
109 | for (i = 0; i < 4; i++) | ||
110 | notify[i] = 0xffffffff; | ||
111 | } | ||
112 | |||
113 | static int ps3vram_notifier_wait(struct mtd_info *mtd, unsigned int timeout_ms) | ||
114 | { | ||
115 | struct ps3vram_priv *priv = mtd->priv; | ||
116 | u32 *notify = ps3vram_get_notifier(priv->reports, NOTIFIER); | ||
117 | unsigned long timeout = jiffies + msecs_to_jiffies(timeout_ms); | ||
118 | |||
119 | do { | ||
120 | if (!notify[3]) | ||
121 | return 0; | ||
122 | msleep(1); | ||
123 | } while (time_before(jiffies, timeout)); | ||
124 | |||
125 | return -ETIMEDOUT; | ||
126 | } | ||
127 | |||
128 | static void ps3vram_init_ring(struct mtd_info *mtd) | ||
129 | { | ||
130 | struct ps3vram_priv *priv = mtd->priv; | ||
131 | |||
132 | priv->ctrl[CTRL_PUT] = FIFO_BASE + FIFO_OFFSET; | ||
133 | priv->ctrl[CTRL_GET] = FIFO_BASE + FIFO_OFFSET; | ||
134 | } | ||
135 | |||
136 | static int ps3vram_wait_ring(struct mtd_info *mtd, unsigned int timeout_ms) | ||
137 | { | ||
138 | struct ps3vram_priv *priv = mtd->priv; | ||
139 | unsigned long timeout = jiffies + msecs_to_jiffies(timeout_ms); | ||
140 | |||
141 | do { | ||
142 | if (priv->ctrl[CTRL_PUT] == priv->ctrl[CTRL_GET]) | ||
143 | return 0; | ||
144 | msleep(1); | ||
145 | } while (time_before(jiffies, timeout)); | ||
146 | |||
147 | dev_dbg(priv->dev, "%s:%d: FIFO timeout (%08x/%08x/%08x)\n", __func__, | ||
148 | __LINE__, priv->ctrl[CTRL_PUT], priv->ctrl[CTRL_GET], | ||
149 | priv->ctrl[CTRL_TOP]); | ||
150 | |||
151 | return -ETIMEDOUT; | ||
152 | } | ||
153 | |||
154 | static void ps3vram_out_ring(struct ps3vram_priv *priv, u32 data) | ||
155 | { | ||
156 | *(priv->fifo_ptr)++ = data; | ||
157 | } | ||
158 | |||
159 | static void ps3vram_begin_ring(struct ps3vram_priv *priv, u32 chan, | ||
160 | u32 tag, u32 size) | ||
161 | { | ||
162 | ps3vram_out_ring(priv, (size << 18) | (chan << 13) | tag); | ||
163 | } | ||
164 | |||
165 | static void ps3vram_rewind_ring(struct mtd_info *mtd) | ||
166 | { | ||
167 | struct ps3vram_priv *priv = mtd->priv; | ||
168 | u64 status; | ||
169 | |||
170 | ps3vram_out_ring(priv, 0x20000000 | (FIFO_BASE + FIFO_OFFSET)); | ||
171 | |||
172 | priv->ctrl[CTRL_PUT] = FIFO_BASE + FIFO_OFFSET; | ||
173 | |||
174 | /* asking the HV for a blit will kick the fifo */ | ||
175 | status = lv1_gpu_context_attribute(priv->context_handle, | ||
176 | L1GPU_CONTEXT_ATTRIBUTE_FB_BLIT, | ||
177 | 0, 0, 0, 0); | ||
178 | if (status) | ||
179 | dev_err(priv->dev, "%s:%d: lv1_gpu_context_attribute failed\n", | ||
180 | __func__, __LINE__); | ||
181 | |||
182 | priv->fifo_ptr = priv->fifo_base; | ||
183 | } | ||
184 | |||
185 | static void ps3vram_fire_ring(struct mtd_info *mtd) | ||
186 | { | ||
187 | struct ps3vram_priv *priv = mtd->priv; | ||
188 | u64 status; | ||
189 | |||
190 | mutex_lock(&ps3_gpu_mutex); | ||
191 | |||
192 | priv->ctrl[CTRL_PUT] = FIFO_BASE + FIFO_OFFSET + | ||
193 | (priv->fifo_ptr - priv->fifo_base) * sizeof(u32); | ||
194 | |||
195 | /* asking the HV for a blit will kick the fifo */ | ||
196 | status = lv1_gpu_context_attribute(priv->context_handle, | ||
197 | L1GPU_CONTEXT_ATTRIBUTE_FB_BLIT, | ||
198 | 0, 0, 0, 0); | ||
199 | if (status) | ||
200 | dev_err(priv->dev, "%s:%d: lv1_gpu_context_attribute failed\n", | ||
201 | __func__, __LINE__); | ||
202 | |||
203 | if ((priv->fifo_ptr - priv->fifo_base) * sizeof(u32) > | ||
204 | FIFO_SIZE - 1024) { | ||
205 | dev_dbg(priv->dev, "%s:%d: fifo full, rewinding\n", __func__, | ||
206 | __LINE__); | ||
207 | ps3vram_wait_ring(mtd, 200); | ||
208 | ps3vram_rewind_ring(mtd); | ||
209 | } | ||
210 | |||
211 | mutex_unlock(&ps3_gpu_mutex); | ||
212 | } | ||
213 | |||
214 | static void ps3vram_bind(struct mtd_info *mtd) | ||
215 | { | ||
216 | struct ps3vram_priv *priv = mtd->priv; | ||
217 | |||
218 | ps3vram_begin_ring(priv, UPLOAD_SUBCH, 0, 1); | ||
219 | ps3vram_out_ring(priv, 0x31337303); | ||
220 | ps3vram_begin_ring(priv, UPLOAD_SUBCH, 0x180, 3); | ||
221 | ps3vram_out_ring(priv, DMA_NOTIFIER_HANDLE_BASE + NOTIFIER); | ||
222 | ps3vram_out_ring(priv, 0xfeed0001); /* DMA system RAM instance */ | ||
223 | ps3vram_out_ring(priv, 0xfeed0000); /* DMA video RAM instance */ | ||
224 | |||
225 | ps3vram_begin_ring(priv, DOWNLOAD_SUBCH, 0, 1); | ||
226 | ps3vram_out_ring(priv, 0x3137c0de); | ||
227 | ps3vram_begin_ring(priv, DOWNLOAD_SUBCH, 0x180, 3); | ||
228 | ps3vram_out_ring(priv, DMA_NOTIFIER_HANDLE_BASE + NOTIFIER); | ||
229 | ps3vram_out_ring(priv, 0xfeed0000); /* DMA video RAM instance */ | ||
230 | ps3vram_out_ring(priv, 0xfeed0001); /* DMA system RAM instance */ | ||
231 | |||
232 | ps3vram_fire_ring(mtd); | ||
233 | } | ||
234 | |||
235 | static int ps3vram_upload(struct mtd_info *mtd, unsigned int src_offset, | ||
236 | unsigned int dst_offset, int len, int count) | ||
237 | { | ||
238 | struct ps3vram_priv *priv = mtd->priv; | ||
239 | |||
240 | ps3vram_begin_ring(priv, UPLOAD_SUBCH, | ||
241 | NV_MEMORY_TO_MEMORY_FORMAT_OFFSET_IN, 8); | ||
242 | ps3vram_out_ring(priv, XDR_IOIF + src_offset); | ||
243 | ps3vram_out_ring(priv, dst_offset); | ||
244 | ps3vram_out_ring(priv, len); | ||
245 | ps3vram_out_ring(priv, len); | ||
246 | ps3vram_out_ring(priv, len); | ||
247 | ps3vram_out_ring(priv, count); | ||
248 | ps3vram_out_ring(priv, (1 << 8) | 1); | ||
249 | ps3vram_out_ring(priv, 0); | ||
250 | |||
251 | ps3vram_notifier_reset(mtd); | ||
252 | ps3vram_begin_ring(priv, UPLOAD_SUBCH, | ||
253 | NV_MEMORY_TO_MEMORY_FORMAT_NOTIFY, 1); | ||
254 | ps3vram_out_ring(priv, 0); | ||
255 | ps3vram_begin_ring(priv, UPLOAD_SUBCH, 0x100, 1); | ||
256 | ps3vram_out_ring(priv, 0); | ||
257 | ps3vram_fire_ring(mtd); | ||
258 | if (ps3vram_notifier_wait(mtd, 200) < 0) { | ||
259 | dev_dbg(priv->dev, "%s:%d: notifier timeout\n", __func__, | ||
260 | __LINE__); | ||
261 | return -1; | ||
262 | } | ||
263 | |||
264 | return 0; | ||
265 | } | ||
266 | |||
267 | static int ps3vram_download(struct mtd_info *mtd, unsigned int src_offset, | ||
268 | unsigned int dst_offset, int len, int count) | ||
269 | { | ||
270 | struct ps3vram_priv *priv = mtd->priv; | ||
271 | |||
272 | ps3vram_begin_ring(priv, DOWNLOAD_SUBCH, | ||
273 | NV_MEMORY_TO_MEMORY_FORMAT_OFFSET_IN, 8); | ||
274 | ps3vram_out_ring(priv, src_offset); | ||
275 | ps3vram_out_ring(priv, XDR_IOIF + dst_offset); | ||
276 | ps3vram_out_ring(priv, len); | ||
277 | ps3vram_out_ring(priv, len); | ||
278 | ps3vram_out_ring(priv, len); | ||
279 | ps3vram_out_ring(priv, count); | ||
280 | ps3vram_out_ring(priv, (1 << 8) | 1); | ||
281 | ps3vram_out_ring(priv, 0); | ||
282 | |||
283 | ps3vram_notifier_reset(mtd); | ||
284 | ps3vram_begin_ring(priv, DOWNLOAD_SUBCH, | ||
285 | NV_MEMORY_TO_MEMORY_FORMAT_NOTIFY, 1); | ||
286 | ps3vram_out_ring(priv, 0); | ||
287 | ps3vram_begin_ring(priv, DOWNLOAD_SUBCH, 0x100, 1); | ||
288 | ps3vram_out_ring(priv, 0); | ||
289 | ps3vram_fire_ring(mtd); | ||
290 | if (ps3vram_notifier_wait(mtd, 200) < 0) { | ||
291 | dev_dbg(priv->dev, "%s:%d: notifier timeout\n", __func__, | ||
292 | __LINE__); | ||
293 | return -1; | ||
294 | } | ||
295 | |||
296 | return 0; | ||
297 | } | ||
298 | |||
299 | static void ps3vram_cache_evict(struct mtd_info *mtd, int entry) | ||
300 | { | ||
301 | struct ps3vram_priv *priv = mtd->priv; | ||
302 | struct ps3vram_cache *cache = &priv->cache; | ||
303 | |||
304 | if (cache->tags[entry].flags & CACHE_PAGE_DIRTY) { | ||
305 | dev_dbg(priv->dev, "%s:%d: flushing %d : 0x%08x\n", __func__, | ||
306 | __LINE__, entry, cache->tags[entry].address); | ||
307 | if (ps3vram_upload(mtd, | ||
308 | CACHE_OFFSET + entry * cache->page_size, | ||
309 | cache->tags[entry].address, | ||
310 | DMA_PAGE_SIZE, | ||
311 | cache->page_size / DMA_PAGE_SIZE) < 0) { | ||
312 | dev_dbg(priv->dev, "%s:%d: failed to upload from " | ||
313 | "0x%x to 0x%x size 0x%x\n", __func__, __LINE__, | ||
314 | entry * cache->page_size, | ||
315 | cache->tags[entry].address, cache->page_size); | ||
316 | } | ||
317 | cache->tags[entry].flags &= ~CACHE_PAGE_DIRTY; | ||
318 | } | ||
319 | } | ||
320 | |||
321 | static void ps3vram_cache_load(struct mtd_info *mtd, int entry, | ||
322 | unsigned int address) | ||
323 | { | ||
324 | struct ps3vram_priv *priv = mtd->priv; | ||
325 | struct ps3vram_cache *cache = &priv->cache; | ||
326 | |||
327 | dev_dbg(priv->dev, "%s:%d: fetching %d : 0x%08x\n", __func__, __LINE__, | ||
328 | entry, address); | ||
329 | if (ps3vram_download(mtd, | ||
330 | address, | ||
331 | CACHE_OFFSET + entry * cache->page_size, | ||
332 | DMA_PAGE_SIZE, | ||
333 | cache->page_size / DMA_PAGE_SIZE) < 0) { | ||
334 | dev_err(priv->dev, "%s:%d: failed to download from " | ||
335 | "0x%x to 0x%x size 0x%x\n", __func__, __LINE__, address, | ||
336 | entry * cache->page_size, cache->page_size); | ||
337 | } | ||
338 | |||
339 | cache->tags[entry].address = address; | ||
340 | cache->tags[entry].flags |= CACHE_PAGE_PRESENT; | ||
341 | } | ||
342 | |||
343 | |||
344 | static void ps3vram_cache_flush(struct mtd_info *mtd) | ||
345 | { | ||
346 | struct ps3vram_priv *priv = mtd->priv; | ||
347 | struct ps3vram_cache *cache = &priv->cache; | ||
348 | int i; | ||
349 | |||
350 | dev_dbg(priv->dev, "%s:%d: FLUSH\n", __func__, __LINE__); | ||
351 | for (i = 0; i < cache->page_count; i++) { | ||
352 | ps3vram_cache_evict(mtd, i); | ||
353 | cache->tags[i].flags = 0; | ||
354 | } | ||
355 | } | ||
356 | |||
357 | static unsigned int ps3vram_cache_match(struct mtd_info *mtd, loff_t address) | ||
358 | { | ||
359 | struct ps3vram_priv *priv = mtd->priv; | ||
360 | struct ps3vram_cache *cache = &priv->cache; | ||
361 | unsigned int base; | ||
362 | unsigned int offset; | ||
363 | int i; | ||
364 | static int counter; | ||
365 | |||
366 | offset = (unsigned int) (address & (cache->page_size - 1)); | ||
367 | base = (unsigned int) (address - offset); | ||
368 | |||
369 | /* fully associative check */ | ||
370 | for (i = 0; i < cache->page_count; i++) { | ||
371 | if ((cache->tags[i].flags & CACHE_PAGE_PRESENT) && | ||
372 | cache->tags[i].address == base) { | ||
373 | dev_dbg(priv->dev, "%s:%d: found entry %d : 0x%08x\n", | ||
374 | __func__, __LINE__, i, cache->tags[i].address); | ||
375 | return i; | ||
376 | } | ||
377 | } | ||
378 | |||
379 | /* choose a random entry */ | ||
380 | i = (jiffies + (counter++)) % cache->page_count; | ||
381 | dev_dbg(priv->dev, "%s:%d: using entry %d\n", __func__, __LINE__, i); | ||
382 | |||
383 | ps3vram_cache_evict(mtd, i); | ||
384 | ps3vram_cache_load(mtd, i, base); | ||
385 | |||
386 | return i; | ||
387 | } | ||
388 | |||
389 | static int ps3vram_cache_init(struct mtd_info *mtd) | ||
390 | { | ||
391 | struct ps3vram_priv *priv = mtd->priv; | ||
392 | |||
393 | priv->cache.page_count = CACHE_PAGE_COUNT; | ||
394 | priv->cache.page_size = CACHE_PAGE_SIZE; | ||
395 | priv->cache.tags = kzalloc(sizeof(struct ps3vram_tag) * | ||
396 | CACHE_PAGE_COUNT, GFP_KERNEL); | ||
397 | if (priv->cache.tags == NULL) { | ||
398 | dev_err(priv->dev, "%s:%d: could not allocate cache tags\n", | ||
399 | __func__, __LINE__); | ||
400 | return -ENOMEM; | ||
401 | } | ||
402 | |||
403 | dev_info(priv->dev, "created ram cache: %d entries, %d KiB each\n", | ||
404 | CACHE_PAGE_COUNT, CACHE_PAGE_SIZE / 1024); | ||
405 | |||
406 | return 0; | ||
407 | } | ||
408 | |||
409 | static void ps3vram_cache_cleanup(struct mtd_info *mtd) | ||
410 | { | ||
411 | struct ps3vram_priv *priv = mtd->priv; | ||
412 | |||
413 | ps3vram_cache_flush(mtd); | ||
414 | kfree(priv->cache.tags); | ||
415 | } | ||
416 | |||
417 | static int ps3vram_erase(struct mtd_info *mtd, struct erase_info *instr) | ||
418 | { | ||
419 | struct ps3vram_priv *priv = mtd->priv; | ||
420 | |||
421 | if (instr->addr + instr->len > mtd->size) | ||
422 | return -EINVAL; | ||
423 | |||
424 | mutex_lock(&priv->lock); | ||
425 | |||
426 | ps3vram_cache_flush(mtd); | ||
427 | |||
428 | /* Set bytes to 0xFF */ | ||
429 | memset_io(priv->ddr_base + instr->addr, 0xFF, instr->len); | ||
430 | |||
431 | mutex_unlock(&priv->lock); | ||
432 | |||
433 | instr->state = MTD_ERASE_DONE; | ||
434 | mtd_erase_callback(instr); | ||
435 | |||
436 | return 0; | ||
437 | } | ||
438 | |||
439 | static int ps3vram_read(struct mtd_info *mtd, loff_t from, size_t len, | ||
440 | size_t *retlen, u_char *buf) | ||
441 | { | ||
442 | struct ps3vram_priv *priv = mtd->priv; | ||
443 | unsigned int cached, count; | ||
444 | |||
445 | dev_dbg(priv->dev, "%s:%d: from=0x%08x len=0x%zx\n", __func__, __LINE__, | ||
446 | (unsigned int)from, len); | ||
447 | |||
448 | if (from >= mtd->size) | ||
449 | return -EINVAL; | ||
450 | |||
451 | if (len > mtd->size - from) | ||
452 | len = mtd->size - from; | ||
453 | |||
454 | /* Copy from vram to buf */ | ||
455 | count = len; | ||
456 | while (count) { | ||
457 | unsigned int offset, avail; | ||
458 | unsigned int entry; | ||
459 | |||
460 | offset = (unsigned int) (from & (priv->cache.page_size - 1)); | ||
461 | avail = priv->cache.page_size - offset; | ||
462 | |||
463 | mutex_lock(&priv->lock); | ||
464 | |||
465 | entry = ps3vram_cache_match(mtd, from); | ||
466 | cached = CACHE_OFFSET + entry * priv->cache.page_size + offset; | ||
467 | |||
468 | dev_dbg(priv->dev, "%s:%d: from=%08x cached=%08x offset=%08x " | ||
469 | "avail=%08x count=%08x\n", __func__, __LINE__, | ||
470 | (unsigned int)from, cached, offset, avail, count); | ||
471 | |||
472 | if (avail > count) | ||
473 | avail = count; | ||
474 | memcpy(buf, priv->xdr_buf + cached, avail); | ||
475 | |||
476 | mutex_unlock(&priv->lock); | ||
477 | |||
478 | buf += avail; | ||
479 | count -= avail; | ||
480 | from += avail; | ||
481 | } | ||
482 | |||
483 | *retlen = len; | ||
484 | return 0; | ||
485 | } | ||
486 | |||
487 | static int ps3vram_write(struct mtd_info *mtd, loff_t to, size_t len, | ||
488 | size_t *retlen, const u_char *buf) | ||
489 | { | ||
490 | struct ps3vram_priv *priv = mtd->priv; | ||
491 | unsigned int cached, count; | ||
492 | |||
493 | if (to >= mtd->size) | ||
494 | return -EINVAL; | ||
495 | |||
496 | if (len > mtd->size - to) | ||
497 | len = mtd->size - to; | ||
498 | |||
499 | /* Copy from buf to vram */ | ||
500 | count = len; | ||
501 | while (count) { | ||
502 | unsigned int offset, avail; | ||
503 | unsigned int entry; | ||
504 | |||
505 | offset = (unsigned int) (to & (priv->cache.page_size - 1)); | ||
506 | avail = priv->cache.page_size - offset; | ||
507 | |||
508 | mutex_lock(&priv->lock); | ||
509 | |||
510 | entry = ps3vram_cache_match(mtd, to); | ||
511 | cached = CACHE_OFFSET + entry * priv->cache.page_size + offset; | ||
512 | |||
513 | dev_dbg(priv->dev, "%s:%d: to=%08x cached=%08x offset=%08x " | ||
514 | "avail=%08x count=%08x\n", __func__, __LINE__, | ||
515 | (unsigned int)to, cached, offset, avail, count); | ||
516 | |||
517 | if (avail > count) | ||
518 | avail = count; | ||
519 | memcpy(priv->xdr_buf + cached, buf, avail); | ||
520 | |||
521 | priv->cache.tags[entry].flags |= CACHE_PAGE_DIRTY; | ||
522 | |||
523 | mutex_unlock(&priv->lock); | ||
524 | |||
525 | buf += avail; | ||
526 | count -= avail; | ||
527 | to += avail; | ||
528 | } | ||
529 | |||
530 | *retlen = len; | ||
531 | return 0; | ||
532 | } | ||
533 | |||
534 | static int __devinit ps3vram_probe(struct ps3_system_bus_device *dev) | ||
535 | { | ||
536 | struct ps3vram_priv *priv; | ||
537 | int status; | ||
538 | u64 ddr_lpar; | ||
539 | u64 ctrl_lpar; | ||
540 | u64 info_lpar; | ||
541 | u64 reports_lpar; | ||
542 | u64 ddr_size; | ||
543 | u64 reports_size; | ||
544 | int ret = -ENOMEM; | ||
545 | char *rest; | ||
546 | |||
547 | ret = -EIO; | ||
548 | ps3vram_mtd.priv = kzalloc(sizeof(struct ps3vram_priv), GFP_KERNEL); | ||
549 | if (!ps3vram_mtd.priv) | ||
550 | goto out; | ||
551 | priv = ps3vram_mtd.priv; | ||
552 | |||
553 | mutex_init(&priv->lock); | ||
554 | priv->dev = &dev->core; | ||
555 | |||
556 | /* Allocate XDR buffer (1MiB aligned) */ | ||
557 | priv->xdr_buf = (void *)__get_free_pages(GFP_KERNEL, | ||
558 | get_order(XDR_BUF_SIZE)); | ||
559 | if (priv->xdr_buf == NULL) { | ||
560 | dev_dbg(&dev->core, "%s:%d: could not allocate XDR buffer\n", | ||
561 | __func__, __LINE__); | ||
562 | ret = -ENOMEM; | ||
563 | goto out_free_priv; | ||
564 | } | ||
565 | |||
566 | /* Put FIFO at begginning of XDR buffer */ | ||
567 | priv->fifo_base = (u32 *) (priv->xdr_buf + FIFO_OFFSET); | ||
568 | priv->fifo_ptr = priv->fifo_base; | ||
569 | |||
570 | /* XXX: Need to open GPU, in case ps3fb or snd_ps3 aren't loaded */ | ||
571 | if (ps3_open_hv_device(dev)) { | ||
572 | dev_err(&dev->core, "%s:%d: ps3_open_hv_device failed\n", | ||
573 | __func__, __LINE__); | ||
574 | ret = -EAGAIN; | ||
575 | goto out_close_gpu; | ||
576 | } | ||
577 | |||
578 | /* Request memory */ | ||
579 | status = -1; | ||
580 | ddr_size = memparse(size, &rest); | ||
581 | if (*rest == '-') | ||
582 | ddr_size -= ps3fb_videomemory.size; | ||
583 | ddr_size = ALIGN(ddr_size, 1024*1024); | ||
584 | if (ddr_size <= 0) { | ||
585 | dev_err(&dev->core, "%s:%d: specified size is too small\n", | ||
586 | __func__, __LINE__); | ||
587 | ret = -EINVAL; | ||
588 | goto out_close_gpu; | ||
589 | } | ||
590 | |||
591 | while (ddr_size > 0) { | ||
592 | status = lv1_gpu_memory_allocate(ddr_size, 0, 0, 0, 0, | ||
593 | &priv->memory_handle, | ||
594 | &ddr_lpar); | ||
595 | if (!status) | ||
596 | break; | ||
597 | ddr_size -= 1024*1024; | ||
598 | } | ||
599 | if (status || ddr_size <= 0) { | ||
600 | dev_err(&dev->core, "%s:%d: lv1_gpu_memory_allocate failed\n", | ||
601 | __func__, __LINE__); | ||
602 | ret = -ENOMEM; | ||
603 | goto out_free_xdr_buf; | ||
604 | } | ||
605 | |||
606 | /* Request context */ | ||
607 | status = lv1_gpu_context_allocate(priv->memory_handle, | ||
608 | 0, | ||
609 | &priv->context_handle, | ||
610 | &ctrl_lpar, | ||
611 | &info_lpar, | ||
612 | &reports_lpar, | ||
613 | &reports_size); | ||
614 | if (status) { | ||
615 | dev_err(&dev->core, "%s:%d: lv1_gpu_context_allocate failed\n", | ||
616 | __func__, __LINE__); | ||
617 | ret = -ENOMEM; | ||
618 | goto out_free_memory; | ||
619 | } | ||
620 | |||
621 | /* Map XDR buffer to RSX */ | ||
622 | status = lv1_gpu_context_iomap(priv->context_handle, XDR_IOIF, | ||
623 | ps3_mm_phys_to_lpar(__pa(priv->xdr_buf)), | ||
624 | XDR_BUF_SIZE, 0); | ||
625 | if (status) { | ||
626 | dev_err(&dev->core, "%s:%d: lv1_gpu_context_iomap failed\n", | ||
627 | __func__, __LINE__); | ||
628 | ret = -ENOMEM; | ||
629 | goto out_free_context; | ||
630 | } | ||
631 | |||
632 | priv->ddr_base = ioremap_flags(ddr_lpar, ddr_size, _PAGE_NO_CACHE); | ||
633 | |||
634 | if (!priv->ddr_base) { | ||
635 | dev_err(&dev->core, "%s:%d: ioremap failed\n", __func__, | ||
636 | __LINE__); | ||
637 | ret = -ENOMEM; | ||
638 | goto out_free_context; | ||
639 | } | ||
640 | |||
641 | priv->ctrl = ioremap(ctrl_lpar, 64 * 1024); | ||
642 | if (!priv->ctrl) { | ||
643 | dev_err(&dev->core, "%s:%d: ioremap failed\n", __func__, | ||
644 | __LINE__); | ||
645 | ret = -ENOMEM; | ||
646 | goto out_unmap_vram; | ||
647 | } | ||
648 | |||
649 | priv->reports = ioremap(reports_lpar, reports_size); | ||
650 | if (!priv->reports) { | ||
651 | dev_err(&dev->core, "%s:%d: ioremap failed\n", __func__, | ||
652 | __LINE__); | ||
653 | ret = -ENOMEM; | ||
654 | goto out_unmap_ctrl; | ||
655 | } | ||
656 | |||
657 | mutex_lock(&ps3_gpu_mutex); | ||
658 | ps3vram_init_ring(&ps3vram_mtd); | ||
659 | mutex_unlock(&ps3_gpu_mutex); | ||
660 | |||
661 | ps3vram_mtd.name = "ps3vram"; | ||
662 | ps3vram_mtd.size = ddr_size; | ||
663 | ps3vram_mtd.flags = MTD_CAP_RAM; | ||
664 | ps3vram_mtd.erase = ps3vram_erase; | ||
665 | ps3vram_mtd.point = NULL; | ||
666 | ps3vram_mtd.unpoint = NULL; | ||
667 | ps3vram_mtd.read = ps3vram_read; | ||
668 | ps3vram_mtd.write = ps3vram_write; | ||
669 | ps3vram_mtd.owner = THIS_MODULE; | ||
670 | ps3vram_mtd.type = MTD_RAM; | ||
671 | ps3vram_mtd.erasesize = CACHE_PAGE_SIZE; | ||
672 | ps3vram_mtd.writesize = 1; | ||
673 | |||
674 | ps3vram_bind(&ps3vram_mtd); | ||
675 | |||
676 | mutex_lock(&ps3_gpu_mutex); | ||
677 | ret = ps3vram_wait_ring(&ps3vram_mtd, 100); | ||
678 | mutex_unlock(&ps3_gpu_mutex); | ||
679 | if (ret < 0) { | ||
680 | dev_err(&dev->core, "%s:%d: failed to initialize channels\n", | ||
681 | __func__, __LINE__); | ||
682 | ret = -ETIMEDOUT; | ||
683 | goto out_unmap_reports; | ||
684 | } | ||
685 | |||
686 | ps3vram_cache_init(&ps3vram_mtd); | ||
687 | |||
688 | if (add_mtd_device(&ps3vram_mtd)) { | ||
689 | dev_err(&dev->core, "%s:%d: add_mtd_device failed\n", | ||
690 | __func__, __LINE__); | ||
691 | ret = -EAGAIN; | ||
692 | goto out_cache_cleanup; | ||
693 | } | ||
694 | |||
695 | dev_info(&dev->core, "reserved %u MiB of gpu memory\n", | ||
696 | (unsigned int)(ddr_size / 1024 / 1024)); | ||
697 | |||
698 | return 0; | ||
699 | |||
700 | out_cache_cleanup: | ||
701 | ps3vram_cache_cleanup(&ps3vram_mtd); | ||
702 | out_unmap_reports: | ||
703 | iounmap(priv->reports); | ||
704 | out_unmap_ctrl: | ||
705 | iounmap(priv->ctrl); | ||
706 | out_unmap_vram: | ||
707 | iounmap(priv->ddr_base); | ||
708 | out_free_context: | ||
709 | lv1_gpu_context_free(priv->context_handle); | ||
710 | out_free_memory: | ||
711 | lv1_gpu_memory_free(priv->memory_handle); | ||
712 | out_close_gpu: | ||
713 | ps3_close_hv_device(dev); | ||
714 | out_free_xdr_buf: | ||
715 | free_pages((unsigned long) priv->xdr_buf, get_order(XDR_BUF_SIZE)); | ||
716 | out_free_priv: | ||
717 | kfree(ps3vram_mtd.priv); | ||
718 | ps3vram_mtd.priv = NULL; | ||
719 | out: | ||
720 | return ret; | ||
721 | } | ||
722 | |||
723 | static int ps3vram_shutdown(struct ps3_system_bus_device *dev) | ||
724 | { | ||
725 | struct ps3vram_priv *priv; | ||
726 | |||
727 | priv = ps3vram_mtd.priv; | ||
728 | |||
729 | del_mtd_device(&ps3vram_mtd); | ||
730 | ps3vram_cache_cleanup(&ps3vram_mtd); | ||
731 | iounmap(priv->reports); | ||
732 | iounmap(priv->ctrl); | ||
733 | iounmap(priv->ddr_base); | ||
734 | lv1_gpu_context_free(priv->context_handle); | ||
735 | lv1_gpu_memory_free(priv->memory_handle); | ||
736 | ps3_close_hv_device(dev); | ||
737 | free_pages((unsigned long) priv->xdr_buf, get_order(XDR_BUF_SIZE)); | ||
738 | kfree(priv); | ||
739 | return 0; | ||
740 | } | ||
741 | |||
742 | static struct ps3_system_bus_driver ps3vram_driver = { | ||
743 | .match_id = PS3_MATCH_ID_GPU, | ||
744 | .match_sub_id = PS3_MATCH_SUB_ID_GPU_RAMDISK, | ||
745 | .core.name = DEVICE_NAME, | ||
746 | .core.owner = THIS_MODULE, | ||
747 | .probe = ps3vram_probe, | ||
748 | .remove = ps3vram_shutdown, | ||
749 | .shutdown = ps3vram_shutdown, | ||
750 | }; | ||
751 | |||
752 | static int __init ps3vram_init(void) | ||
753 | { | ||
754 | return ps3_system_bus_driver_register(&ps3vram_driver); | ||
755 | } | ||
756 | |||
757 | static void __exit ps3vram_exit(void) | ||
758 | { | ||
759 | ps3_system_bus_driver_unregister(&ps3vram_driver); | ||
760 | } | ||
761 | |||
762 | module_init(ps3vram_init); | ||
763 | module_exit(ps3vram_exit); | ||
764 | |||
765 | MODULE_LICENSE("GPL"); | ||
766 | MODULE_AUTHOR("Jim Paris <jim@jtan.com>"); | ||
767 | MODULE_DESCRIPTION("MTD driver for PS3 video RAM"); | ||
768 | MODULE_ALIAS(PS3_MODULE_ALIAS_GPU_RAMDISK); | ||
diff --git a/drivers/mtd/ubi/kapi.c b/drivers/mtd/ubi/kapi.c index 5d9bcf109c13..4abbe573fa40 100644 --- a/drivers/mtd/ubi/kapi.c +++ b/drivers/mtd/ubi/kapi.c | |||
@@ -564,7 +564,7 @@ EXPORT_SYMBOL_GPL(ubi_leb_unmap); | |||
564 | * @dtype: expected data type | 564 | * @dtype: expected data type |
565 | * | 565 | * |
566 | * This function maps an un-mapped logical eraseblock @lnum to a physical | 566 | * This function maps an un-mapped logical eraseblock @lnum to a physical |
567 | * eraseblock. This means, that after a successfull invocation of this | 567 | * eraseblock. This means, that after a successful invocation of this |
568 | * function the logical eraseblock @lnum will be empty (contain only %0xFF | 568 | * function the logical eraseblock @lnum will be empty (contain only %0xFF |
569 | * bytes) and be mapped to a physical eraseblock, even if an unclean reboot | 569 | * bytes) and be mapped to a physical eraseblock, even if an unclean reboot |
570 | * happens. | 570 | * happens. |
diff --git a/drivers/net/wireless/ath5k/dma.c b/drivers/net/wireless/ath5k/dma.c index 7e2b1a67e5da..b65b4feb2d28 100644 --- a/drivers/net/wireless/ath5k/dma.c +++ b/drivers/net/wireless/ath5k/dma.c | |||
@@ -594,7 +594,7 @@ int ath5k_hw_get_isr(struct ath5k_hw *ah, enum ath5k_int *interrupt_mask) | |||
594 | * XXX: BMISS interrupts may occur after association. | 594 | * XXX: BMISS interrupts may occur after association. |
595 | * I found this on 5210 code but it needs testing. If this is | 595 | * I found this on 5210 code but it needs testing. If this is |
596 | * true we should disable them before assoc and re-enable them | 596 | * true we should disable them before assoc and re-enable them |
597 | * after a successfull assoc + some jiffies. | 597 | * after a successful assoc + some jiffies. |
598 | interrupt_mask &= ~AR5K_INT_BMISS; | 598 | interrupt_mask &= ~AR5K_INT_BMISS; |
599 | */ | 599 | */ |
600 | } | 600 | } |
diff --git a/drivers/net/wireless/zd1211rw/zd_mac.c b/drivers/net/wireless/zd1211rw/zd_mac.c index 9caa96a13586..a611ad857983 100644 --- a/drivers/net/wireless/zd1211rw/zd_mac.c +++ b/drivers/net/wireless/zd1211rw/zd_mac.c | |||
@@ -287,7 +287,7 @@ static void zd_op_stop(struct ieee80211_hw *hw) | |||
287 | * @skb - a sk-buffer | 287 | * @skb - a sk-buffer |
288 | * @flags: extra flags to set in the TX status info | 288 | * @flags: extra flags to set in the TX status info |
289 | * @ackssi: ACK signal strength | 289 | * @ackssi: ACK signal strength |
290 | * @success - True for successfull transmission of the frame | 290 | * @success - True for successful transmission of the frame |
291 | * | 291 | * |
292 | * This information calls ieee80211_tx_status_irqsafe() if required by the | 292 | * This information calls ieee80211_tx_status_irqsafe() if required by the |
293 | * control information. It copies the control information into the status | 293 | * control information. It copies the control information into the status |
diff --git a/drivers/rtc/rtc-ds1307.c b/drivers/rtc/rtc-ds1307.c index 162330b9d1dc..7e5155e88ac7 100644 --- a/drivers/rtc/rtc-ds1307.c +++ b/drivers/rtc/rtc-ds1307.c | |||
@@ -86,13 +86,11 @@ enum ds_type { | |||
86 | 86 | ||
87 | 87 | ||
88 | struct ds1307 { | 88 | struct ds1307 { |
89 | u8 reg_addr; | ||
90 | u8 regs[11]; | 89 | u8 regs[11]; |
91 | enum ds_type type; | 90 | enum ds_type type; |
92 | unsigned long flags; | 91 | unsigned long flags; |
93 | #define HAS_NVRAM 0 /* bit 0 == sysfs file active */ | 92 | #define HAS_NVRAM 0 /* bit 0 == sysfs file active */ |
94 | #define HAS_ALARM 1 /* bit 1 == irq claimed */ | 93 | #define HAS_ALARM 1 /* bit 1 == irq claimed */ |
95 | struct i2c_msg msg[2]; | ||
96 | struct i2c_client *client; | 94 | struct i2c_client *client; |
97 | struct rtc_device *rtc; | 95 | struct rtc_device *rtc; |
98 | struct work_struct work; | 96 | struct work_struct work; |
@@ -204,13 +202,9 @@ static int ds1307_get_time(struct device *dev, struct rtc_time *t) | |||
204 | int tmp; | 202 | int tmp; |
205 | 203 | ||
206 | /* read the RTC date and time registers all at once */ | 204 | /* read the RTC date and time registers all at once */ |
207 | ds1307->reg_addr = 0; | 205 | tmp = i2c_smbus_read_i2c_block_data(ds1307->client, |
208 | ds1307->msg[1].flags = I2C_M_RD; | 206 | DS1307_REG_SECS, 7, ds1307->regs); |
209 | ds1307->msg[1].len = 7; | 207 | if (tmp != 7) { |
210 | |||
211 | tmp = i2c_transfer(to_i2c_adapter(ds1307->client->dev.parent), | ||
212 | ds1307->msg, 2); | ||
213 | if (tmp != 2) { | ||
214 | dev_err(dev, "%s error %d\n", "read", tmp); | 208 | dev_err(dev, "%s error %d\n", "read", tmp); |
215 | return -EIO; | 209 | return -EIO; |
216 | } | 210 | } |
@@ -257,7 +251,6 @@ static int ds1307_set_time(struct device *dev, struct rtc_time *t) | |||
257 | t->tm_hour, t->tm_mday, | 251 | t->tm_hour, t->tm_mday, |
258 | t->tm_mon, t->tm_year, t->tm_wday); | 252 | t->tm_mon, t->tm_year, t->tm_wday); |
259 | 253 | ||
260 | *buf++ = 0; /* first register addr */ | ||
261 | buf[DS1307_REG_SECS] = bin2bcd(t->tm_sec); | 254 | buf[DS1307_REG_SECS] = bin2bcd(t->tm_sec); |
262 | buf[DS1307_REG_MIN] = bin2bcd(t->tm_min); | 255 | buf[DS1307_REG_MIN] = bin2bcd(t->tm_min); |
263 | buf[DS1307_REG_HOUR] = bin2bcd(t->tm_hour); | 256 | buf[DS1307_REG_HOUR] = bin2bcd(t->tm_hour); |
@@ -282,23 +275,19 @@ static int ds1307_set_time(struct device *dev, struct rtc_time *t) | |||
282 | break; | 275 | break; |
283 | } | 276 | } |
284 | 277 | ||
285 | ds1307->msg[1].flags = 0; | ||
286 | ds1307->msg[1].len = 8; | ||
287 | |||
288 | dev_dbg(dev, "%s: %02x %02x %02x %02x %02x %02x %02x\n", | 278 | dev_dbg(dev, "%s: %02x %02x %02x %02x %02x %02x %02x\n", |
289 | "write", buf[0], buf[1], buf[2], buf[3], | 279 | "write", buf[0], buf[1], buf[2], buf[3], |
290 | buf[4], buf[5], buf[6]); | 280 | buf[4], buf[5], buf[6]); |
291 | 281 | ||
292 | result = i2c_transfer(to_i2c_adapter(ds1307->client->dev.parent), | 282 | result = i2c_smbus_write_i2c_block_data(ds1307->client, 0, 7, buf); |
293 | &ds1307->msg[1], 1); | 283 | if (result < 0) { |
294 | if (result != 1) { | 284 | dev_err(dev, "%s error %d\n", "write", result); |
295 | dev_err(dev, "%s error %d\n", "write", tmp); | 285 | return result; |
296 | return -EIO; | ||
297 | } | 286 | } |
298 | return 0; | 287 | return 0; |
299 | } | 288 | } |
300 | 289 | ||
301 | static int ds1307_read_alarm(struct device *dev, struct rtc_wkalrm *t) | 290 | static int ds1337_read_alarm(struct device *dev, struct rtc_wkalrm *t) |
302 | { | 291 | { |
303 | struct i2c_client *client = to_i2c_client(dev); | 292 | struct i2c_client *client = to_i2c_client(dev); |
304 | struct ds1307 *ds1307 = i2c_get_clientdata(client); | 293 | struct ds1307 *ds1307 = i2c_get_clientdata(client); |
@@ -308,13 +297,9 @@ static int ds1307_read_alarm(struct device *dev, struct rtc_wkalrm *t) | |||
308 | return -EINVAL; | 297 | return -EINVAL; |
309 | 298 | ||
310 | /* read all ALARM1, ALARM2, and status registers at once */ | 299 | /* read all ALARM1, ALARM2, and status registers at once */ |
311 | ds1307->reg_addr = DS1339_REG_ALARM1_SECS; | 300 | ret = i2c_smbus_read_i2c_block_data(client, |
312 | ds1307->msg[1].flags = I2C_M_RD; | 301 | DS1339_REG_ALARM1_SECS, 9, ds1307->regs); |
313 | ds1307->msg[1].len = 9; | 302 | if (ret != 9) { |
314 | |||
315 | ret = i2c_transfer(to_i2c_adapter(client->dev.parent), | ||
316 | ds1307->msg, 2); | ||
317 | if (ret != 2) { | ||
318 | dev_err(dev, "%s error %d\n", "alarm read", ret); | 303 | dev_err(dev, "%s error %d\n", "alarm read", ret); |
319 | return -EIO; | 304 | return -EIO; |
320 | } | 305 | } |
@@ -353,7 +338,7 @@ static int ds1307_read_alarm(struct device *dev, struct rtc_wkalrm *t) | |||
353 | return 0; | 338 | return 0; |
354 | } | 339 | } |
355 | 340 | ||
356 | static int ds1307_set_alarm(struct device *dev, struct rtc_wkalrm *t) | 341 | static int ds1337_set_alarm(struct device *dev, struct rtc_wkalrm *t) |
357 | { | 342 | { |
358 | struct i2c_client *client = to_i2c_client(dev); | 343 | struct i2c_client *client = to_i2c_client(dev); |
359 | struct ds1307 *ds1307 = i2c_get_clientdata(client); | 344 | struct ds1307 *ds1307 = i2c_get_clientdata(client); |
@@ -371,13 +356,9 @@ static int ds1307_set_alarm(struct device *dev, struct rtc_wkalrm *t) | |||
371 | t->enabled, t->pending); | 356 | t->enabled, t->pending); |
372 | 357 | ||
373 | /* read current status of both alarms and the chip */ | 358 | /* read current status of both alarms and the chip */ |
374 | ds1307->reg_addr = DS1339_REG_ALARM1_SECS; | 359 | ret = i2c_smbus_read_i2c_block_data(client, |
375 | ds1307->msg[1].flags = I2C_M_RD; | 360 | DS1339_REG_ALARM1_SECS, 9, buf); |
376 | ds1307->msg[1].len = 9; | 361 | if (ret != 9) { |
377 | |||
378 | ret = i2c_transfer(to_i2c_adapter(client->dev.parent), | ||
379 | ds1307->msg, 2); | ||
380 | if (ret != 2) { | ||
381 | dev_err(dev, "%s error %d\n", "alarm write", ret); | 362 | dev_err(dev, "%s error %d\n", "alarm write", ret); |
382 | return -EIO; | 363 | return -EIO; |
383 | } | 364 | } |
@@ -392,7 +373,6 @@ static int ds1307_set_alarm(struct device *dev, struct rtc_wkalrm *t) | |||
392 | ds1307->regs[6], control, status); | 373 | ds1307->regs[6], control, status); |
393 | 374 | ||
394 | /* set ALARM1, using 24 hour and day-of-month modes */ | 375 | /* set ALARM1, using 24 hour and day-of-month modes */ |
395 | *buf++ = DS1339_REG_ALARM1_SECS; /* first register addr */ | ||
396 | buf[0] = bin2bcd(t->time.tm_sec); | 376 | buf[0] = bin2bcd(t->time.tm_sec); |
397 | buf[1] = bin2bcd(t->time.tm_min); | 377 | buf[1] = bin2bcd(t->time.tm_min); |
398 | buf[2] = bin2bcd(t->time.tm_hour); | 378 | buf[2] = bin2bcd(t->time.tm_hour); |
@@ -411,14 +391,11 @@ static int ds1307_set_alarm(struct device *dev, struct rtc_wkalrm *t) | |||
411 | } | 391 | } |
412 | buf[8] = status & ~(DS1337_BIT_A1I | DS1337_BIT_A2I); | 392 | buf[8] = status & ~(DS1337_BIT_A1I | DS1337_BIT_A2I); |
413 | 393 | ||
414 | ds1307->msg[1].flags = 0; | 394 | ret = i2c_smbus_write_i2c_block_data(client, |
415 | ds1307->msg[1].len = 10; | 395 | DS1339_REG_ALARM1_SECS, 9, buf); |
416 | 396 | if (ret < 0) { | |
417 | ret = i2c_transfer(to_i2c_adapter(client->dev.parent), | ||
418 | &ds1307->msg[1], 1); | ||
419 | if (ret != 1) { | ||
420 | dev_err(dev, "can't set alarm time\n"); | 397 | dev_err(dev, "can't set alarm time\n"); |
421 | return -EIO; | 398 | return ret; |
422 | } | 399 | } |
423 | 400 | ||
424 | return 0; | 401 | return 0; |
@@ -475,8 +452,8 @@ static int ds1307_ioctl(struct device *dev, unsigned int cmd, unsigned long arg) | |||
475 | static const struct rtc_class_ops ds13xx_rtc_ops = { | 452 | static const struct rtc_class_ops ds13xx_rtc_ops = { |
476 | .read_time = ds1307_get_time, | 453 | .read_time = ds1307_get_time, |
477 | .set_time = ds1307_set_time, | 454 | .set_time = ds1307_set_time, |
478 | .read_alarm = ds1307_read_alarm, | 455 | .read_alarm = ds1337_read_alarm, |
479 | .set_alarm = ds1307_set_alarm, | 456 | .set_alarm = ds1337_set_alarm, |
480 | .ioctl = ds1307_ioctl, | 457 | .ioctl = ds1307_ioctl, |
481 | }; | 458 | }; |
482 | 459 | ||
@@ -490,7 +467,6 @@ ds1307_nvram_read(struct kobject *kobj, struct bin_attribute *attr, | |||
490 | { | 467 | { |
491 | struct i2c_client *client; | 468 | struct i2c_client *client; |
492 | struct ds1307 *ds1307; | 469 | struct ds1307 *ds1307; |
493 | struct i2c_msg msg[2]; | ||
494 | int result; | 470 | int result; |
495 | 471 | ||
496 | client = kobj_to_i2c_client(kobj); | 472 | client = kobj_to_i2c_client(kobj); |
@@ -503,24 +479,10 @@ ds1307_nvram_read(struct kobject *kobj, struct bin_attribute *attr, | |||
503 | if (unlikely(!count)) | 479 | if (unlikely(!count)) |
504 | return count; | 480 | return count; |
505 | 481 | ||
506 | msg[0].addr = client->addr; | 482 | result = i2c_smbus_read_i2c_block_data(client, 8 + off, count, buf); |
507 | msg[0].flags = 0; | 483 | if (result < 0) |
508 | msg[0].len = 1; | ||
509 | msg[0].buf = buf; | ||
510 | |||
511 | buf[0] = 8 + off; | ||
512 | |||
513 | msg[1].addr = client->addr; | ||
514 | msg[1].flags = I2C_M_RD; | ||
515 | msg[1].len = count; | ||
516 | msg[1].buf = buf; | ||
517 | |||
518 | result = i2c_transfer(to_i2c_adapter(client->dev.parent), msg, 2); | ||
519 | if (result != 2) { | ||
520 | dev_err(&client->dev, "%s error %d\n", "nvram read", result); | 484 | dev_err(&client->dev, "%s error %d\n", "nvram read", result); |
521 | return -EIO; | 485 | return result; |
522 | } | ||
523 | return count; | ||
524 | } | 486 | } |
525 | 487 | ||
526 | static ssize_t | 488 | static ssize_t |
@@ -528,8 +490,7 @@ ds1307_nvram_write(struct kobject *kobj, struct bin_attribute *attr, | |||
528 | char *buf, loff_t off, size_t count) | 490 | char *buf, loff_t off, size_t count) |
529 | { | 491 | { |
530 | struct i2c_client *client; | 492 | struct i2c_client *client; |
531 | u8 buffer[NVRAM_SIZE + 1]; | 493 | int result; |
532 | int ret; | ||
533 | 494 | ||
534 | client = kobj_to_i2c_client(kobj); | 495 | client = kobj_to_i2c_client(kobj); |
535 | 496 | ||
@@ -540,11 +501,12 @@ ds1307_nvram_write(struct kobject *kobj, struct bin_attribute *attr, | |||
540 | if (unlikely(!count)) | 501 | if (unlikely(!count)) |
541 | return count; | 502 | return count; |
542 | 503 | ||
543 | buffer[0] = 8 + off; | 504 | result = i2c_smbus_write_i2c_block_data(client, 8 + off, count, buf); |
544 | memcpy(buffer + 1, buf, count); | 505 | if (result < 0) { |
545 | 506 | dev_err(&client->dev, "%s error %d\n", "nvram write", result); | |
546 | ret = i2c_master_send(client, buffer, count + 1); | 507 | return result; |
547 | return (ret < 0) ? ret : (ret - 1); | 508 | } |
509 | return count; | ||
548 | } | 510 | } |
549 | 511 | ||
550 | static struct bin_attribute nvram = { | 512 | static struct bin_attribute nvram = { |
@@ -571,9 +533,11 @@ static int __devinit ds1307_probe(struct i2c_client *client, | |||
571 | const struct chip_desc *chip = &chips[id->driver_data]; | 533 | const struct chip_desc *chip = &chips[id->driver_data]; |
572 | struct i2c_adapter *adapter = to_i2c_adapter(client->dev.parent); | 534 | struct i2c_adapter *adapter = to_i2c_adapter(client->dev.parent); |
573 | int want_irq = false; | 535 | int want_irq = false; |
536 | unsigned char *buf; | ||
574 | 537 | ||
575 | if (!i2c_check_functionality(adapter, | 538 | if (!i2c_check_functionality(adapter, |
576 | I2C_FUNC_I2C | I2C_FUNC_SMBUS_WRITE_BYTE_DATA)) | 539 | I2C_FUNC_SMBUS_WRITE_BYTE_DATA | |
540 | I2C_FUNC_SMBUS_I2C_BLOCK)) | ||
577 | return -EIO; | 541 | return -EIO; |
578 | 542 | ||
579 | if (!(ds1307 = kzalloc(sizeof(struct ds1307), GFP_KERNEL))) | 543 | if (!(ds1307 = kzalloc(sizeof(struct ds1307), GFP_KERNEL))) |
@@ -581,18 +545,8 @@ static int __devinit ds1307_probe(struct i2c_client *client, | |||
581 | 545 | ||
582 | ds1307->client = client; | 546 | ds1307->client = client; |
583 | i2c_set_clientdata(client, ds1307); | 547 | i2c_set_clientdata(client, ds1307); |
584 | |||
585 | ds1307->msg[0].addr = client->addr; | ||
586 | ds1307->msg[0].flags = 0; | ||
587 | ds1307->msg[0].len = 1; | ||
588 | ds1307->msg[0].buf = &ds1307->reg_addr; | ||
589 | |||
590 | ds1307->msg[1].addr = client->addr; | ||
591 | ds1307->msg[1].flags = I2C_M_RD; | ||
592 | ds1307->msg[1].len = sizeof(ds1307->regs); | ||
593 | ds1307->msg[1].buf = ds1307->regs; | ||
594 | |||
595 | ds1307->type = id->driver_data; | 548 | ds1307->type = id->driver_data; |
549 | buf = ds1307->regs; | ||
596 | 550 | ||
597 | switch (ds1307->type) { | 551 | switch (ds1307->type) { |
598 | case ds_1337: | 552 | case ds_1337: |
@@ -602,21 +556,15 @@ static int __devinit ds1307_probe(struct i2c_client *client, | |||
602 | INIT_WORK(&ds1307->work, ds1307_work); | 556 | INIT_WORK(&ds1307->work, ds1307_work); |
603 | want_irq = true; | 557 | want_irq = true; |
604 | } | 558 | } |
605 | |||
606 | ds1307->reg_addr = DS1337_REG_CONTROL; | ||
607 | ds1307->msg[1].len = 2; | ||
608 | |||
609 | /* get registers that the "rtc" read below won't read... */ | 559 | /* get registers that the "rtc" read below won't read... */ |
610 | tmp = i2c_transfer(adapter, ds1307->msg, 2); | 560 | tmp = i2c_smbus_read_i2c_block_data(ds1307->client, |
561 | DS1337_REG_CONTROL, 2, buf); | ||
611 | if (tmp != 2) { | 562 | if (tmp != 2) { |
612 | pr_debug("read error %d\n", tmp); | 563 | pr_debug("read error %d\n", tmp); |
613 | err = -EIO; | 564 | err = -EIO; |
614 | goto exit_free; | 565 | goto exit_free; |
615 | } | 566 | } |
616 | 567 | ||
617 | ds1307->reg_addr = 0; | ||
618 | ds1307->msg[1].len = sizeof(ds1307->regs); | ||
619 | |||
620 | /* oscillator off? turn it on, so clock can tick. */ | 568 | /* oscillator off? turn it on, so clock can tick. */ |
621 | if (ds1307->regs[0] & DS1337_BIT_nEOSC) | 569 | if (ds1307->regs[0] & DS1337_BIT_nEOSC) |
622 | ds1307->regs[0] &= ~DS1337_BIT_nEOSC; | 570 | ds1307->regs[0] &= ~DS1337_BIT_nEOSC; |
@@ -647,9 +595,8 @@ static int __devinit ds1307_probe(struct i2c_client *client, | |||
647 | 595 | ||
648 | read_rtc: | 596 | read_rtc: |
649 | /* read RTC registers */ | 597 | /* read RTC registers */ |
650 | 598 | tmp = i2c_smbus_read_i2c_block_data(ds1307->client, 0, 8, buf); | |
651 | tmp = i2c_transfer(adapter, ds1307->msg, 2); | 599 | if (tmp != 8) { |
652 | if (tmp != 2) { | ||
653 | pr_debug("read error %d\n", tmp); | 600 | pr_debug("read error %d\n", tmp); |
654 | err = -EIO; | 601 | err = -EIO; |
655 | goto exit_free; | 602 | goto exit_free; |
@@ -707,22 +654,6 @@ read_rtc: | |||
707 | break; | 654 | break; |
708 | } | 655 | } |
709 | 656 | ||
710 | tmp = ds1307->regs[DS1307_REG_SECS]; | ||
711 | tmp = bcd2bin(tmp & 0x7f); | ||
712 | if (tmp > 60) | ||
713 | goto exit_bad; | ||
714 | tmp = bcd2bin(ds1307->regs[DS1307_REG_MIN] & 0x7f); | ||
715 | if (tmp > 60) | ||
716 | goto exit_bad; | ||
717 | |||
718 | tmp = bcd2bin(ds1307->regs[DS1307_REG_MDAY] & 0x3f); | ||
719 | if (tmp == 0 || tmp > 31) | ||
720 | goto exit_bad; | ||
721 | |||
722 | tmp = bcd2bin(ds1307->regs[DS1307_REG_MONTH] & 0x1f); | ||
723 | if (tmp == 0 || tmp > 12) | ||
724 | goto exit_bad; | ||
725 | |||
726 | tmp = ds1307->regs[DS1307_REG_HOUR]; | 657 | tmp = ds1307->regs[DS1307_REG_HOUR]; |
727 | switch (ds1307->type) { | 658 | switch (ds1307->type) { |
728 | case ds_1340: | 659 | case ds_1340: |
@@ -779,13 +710,6 @@ read_rtc: | |||
779 | 710 | ||
780 | return 0; | 711 | return 0; |
781 | 712 | ||
782 | exit_bad: | ||
783 | dev_dbg(&client->dev, "%s: %02x %02x %02x %02x %02x %02x %02x\n", | ||
784 | "bogus register", | ||
785 | ds1307->regs[0], ds1307->regs[1], | ||
786 | ds1307->regs[2], ds1307->regs[3], | ||
787 | ds1307->regs[4], ds1307->regs[5], | ||
788 | ds1307->regs[6]); | ||
789 | exit_irq: | 713 | exit_irq: |
790 | if (ds1307->rtc) | 714 | if (ds1307->rtc) |
791 | rtc_device_unregister(ds1307->rtc); | 715 | rtc_device_unregister(ds1307->rtc); |
diff --git a/drivers/s390/block/dasd_3990_erp.c b/drivers/s390/block/dasd_3990_erp.c index b8f9c00633f3..d82aad5224f0 100644 --- a/drivers/s390/block/dasd_3990_erp.c +++ b/drivers/s390/block/dasd_3990_erp.c | |||
@@ -2621,7 +2621,7 @@ dasd_3990_erp_action(struct dasd_ccw_req * cqr) | |||
2621 | } | 2621 | } |
2622 | } | 2622 | } |
2623 | 2623 | ||
2624 | /* double-check if current erp/cqr was successfull */ | 2624 | /* double-check if current erp/cqr was successful */ |
2625 | if ((cqr->irb.scsw.cmd.cstat == 0x00) && | 2625 | if ((cqr->irb.scsw.cmd.cstat == 0x00) && |
2626 | (cqr->irb.scsw.cmd.dstat == | 2626 | (cqr->irb.scsw.cmd.dstat == |
2627 | (DEV_STAT_CHN_END | DEV_STAT_DEV_END))) { | 2627 | (DEV_STAT_CHN_END | DEV_STAT_DEV_END))) { |
diff --git a/drivers/s390/block/dasd_int.h b/drivers/s390/block/dasd_int.h index 05a14536c369..4a39084d9c95 100644 --- a/drivers/s390/block/dasd_int.h +++ b/drivers/s390/block/dasd_int.h | |||
@@ -199,7 +199,7 @@ struct dasd_ccw_req { | |||
199 | #define DASD_CQR_ERROR 0x82 /* request is completed with error */ | 199 | #define DASD_CQR_ERROR 0x82 /* request is completed with error */ |
200 | #define DASD_CQR_CLEAR_PENDING 0x83 /* request is clear pending */ | 200 | #define DASD_CQR_CLEAR_PENDING 0x83 /* request is clear pending */ |
201 | #define DASD_CQR_CLEARED 0x84 /* request was cleared */ | 201 | #define DASD_CQR_CLEARED 0x84 /* request was cleared */ |
202 | #define DASD_CQR_SUCCESS 0x85 /* request was successfull */ | 202 | #define DASD_CQR_SUCCESS 0x85 /* request was successful */ |
203 | 203 | ||
204 | 204 | ||
205 | /* per dasd_ccw_req flags */ | 205 | /* per dasd_ccw_req flags */ |
diff --git a/drivers/s390/char/tape_3590.c b/drivers/s390/char/tape_3590.c index 4005c44a404c..71605a179d65 100644 --- a/drivers/s390/char/tape_3590.c +++ b/drivers/s390/char/tape_3590.c | |||
@@ -801,7 +801,7 @@ tape_3590_done(struct tape_device *device, struct tape_request *request) | |||
801 | static inline int | 801 | static inline int |
802 | tape_3590_erp_succeded(struct tape_device *device, struct tape_request *request) | 802 | tape_3590_erp_succeded(struct tape_device *device, struct tape_request *request) |
803 | { | 803 | { |
804 | DBF_EVENT(3, "Error Recovery successfull for %s\n", | 804 | DBF_EVENT(3, "Error Recovery successful for %s\n", |
805 | tape_op_verbose[request->op]); | 805 | tape_op_verbose[request->op]); |
806 | return tape_3590_done(device, request); | 806 | return tape_3590_done(device, request); |
807 | } | 807 | } |
diff --git a/drivers/s390/cio/cio.c b/drivers/s390/cio/cio.c index 06b71823f399..659f8a791656 100644 --- a/drivers/s390/cio/cio.c +++ b/drivers/s390/cio/cio.c | |||
@@ -379,7 +379,7 @@ int cio_commit_config(struct subchannel *sch) | |||
379 | if (ccode < 0) /* -EIO if msch gets a program check. */ | 379 | if (ccode < 0) /* -EIO if msch gets a program check. */ |
380 | return ccode; | 380 | return ccode; |
381 | switch (ccode) { | 381 | switch (ccode) { |
382 | case 0: /* successfull */ | 382 | case 0: /* successful */ |
383 | if (stsch(sch->schid, &schib) || | 383 | if (stsch(sch->schid, &schib) || |
384 | !css_sch_is_valid(&schib)) | 384 | !css_sch_is_valid(&schib)) |
385 | return -ENODEV; | 385 | return -ENODEV; |
diff --git a/drivers/s390/cio/qdio_main.c b/drivers/s390/cio/qdio_main.c index 744f928a59ea..10cb0f8726e5 100644 --- a/drivers/s390/cio/qdio_main.c +++ b/drivers/s390/cio/qdio_main.c | |||
@@ -114,7 +114,7 @@ static inline int qdio_check_ccq(struct qdio_q *q, unsigned int ccq) | |||
114 | * @count: count of buffers to examine | 114 | * @count: count of buffers to examine |
115 | * @auto_ack: automatically acknowledge buffers | 115 | * @auto_ack: automatically acknowledge buffers |
116 | * | 116 | * |
117 | * Returns the number of successfull extracted equal buffer states. | 117 | * Returns the number of successfully extracted equal buffer states. |
118 | * Stops processing if a state is different from the last buffers state. | 118 | * Stops processing if a state is different from the last buffers state. |
119 | */ | 119 | */ |
120 | static int qdio_do_eqbs(struct qdio_q *q, unsigned char *state, | 120 | static int qdio_do_eqbs(struct qdio_q *q, unsigned char *state, |
diff --git a/drivers/scsi/Kconfig b/drivers/scsi/Kconfig index b7322976d2b7..256c7bec7bd7 100644 --- a/drivers/scsi/Kconfig +++ b/drivers/scsi/Kconfig | |||
@@ -884,6 +884,7 @@ config SCSI_IBMVSCSI | |||
884 | tristate "IBM Virtual SCSI support" | 884 | tristate "IBM Virtual SCSI support" |
885 | depends on PPC_PSERIES || PPC_ISERIES | 885 | depends on PPC_PSERIES || PPC_ISERIES |
886 | select SCSI_SRP_ATTRS | 886 | select SCSI_SRP_ATTRS |
887 | select VIOPATH if PPC_ISERIES | ||
887 | help | 888 | help |
888 | This is the IBM POWER Virtual SCSI Client | 889 | This is the IBM POWER Virtual SCSI Client |
889 | 890 | ||
diff --git a/drivers/serial/Kconfig b/drivers/serial/Kconfig index b695ab3142d8..3e525e38a5d9 100644 --- a/drivers/serial/Kconfig +++ b/drivers/serial/Kconfig | |||
@@ -457,7 +457,7 @@ config SERIAL_SAMSUNG | |||
457 | 457 | ||
458 | config SERIAL_SAMSUNG_UARTS | 458 | config SERIAL_SAMSUNG_UARTS |
459 | int | 459 | int |
460 | depends on SERIAL_SAMSUNG | 460 | depends on ARM && PLAT_S3C |
461 | default 2 if ARCH_S3C2400 | 461 | default 2 if ARCH_S3C2400 |
462 | default 4 if ARCH_S3C64XX || CPU_S3C2443 | 462 | default 4 if ARCH_S3C64XX || CPU_S3C2443 |
463 | default 3 | 463 | default 3 |
@@ -1320,13 +1320,30 @@ config SERIAL_NETX_CONSOLE | |||
1320 | config SERIAL_OF_PLATFORM | 1320 | config SERIAL_OF_PLATFORM |
1321 | tristate "Serial port on Open Firmware platform bus" | 1321 | tristate "Serial port on Open Firmware platform bus" |
1322 | depends on PPC_OF | 1322 | depends on PPC_OF |
1323 | depends on SERIAL_8250 | 1323 | depends on SERIAL_8250 || SERIAL_OF_PLATFORM_NWPSERIAL |
1324 | help | 1324 | help |
1325 | If you have a PowerPC based system that has serial ports | 1325 | If you have a PowerPC based system that has serial ports |
1326 | on a platform specific bus, you should enable this option. | 1326 | on a platform specific bus, you should enable this option. |
1327 | Currently, only 8250 compatible ports are supported, but | 1327 | Currently, only 8250 compatible ports are supported, but |
1328 | others can easily be added. | 1328 | others can easily be added. |
1329 | 1329 | ||
1330 | config SERIAL_OF_PLATFORM_NWPSERIAL | ||
1331 | tristate "NWP serial port driver" | ||
1332 | depends on PPC_OF && PPC_DCR | ||
1333 | select SERIAL_OF_PLATFORM | ||
1334 | select SERIAL_CORE_CONSOLE | ||
1335 | select SERIAL_CORE | ||
1336 | help | ||
1337 | This driver supports the cell network processor nwp serial | ||
1338 | device. | ||
1339 | |||
1340 | config SERIAL_OF_PLATFORM_NWPSERIAL_CONSOLE | ||
1341 | bool "Console on NWP serial port" | ||
1342 | depends on SERIAL_OF_PLATFORM_NWPSERIAL=y | ||
1343 | select SERIAL_CORE_CONSOLE | ||
1344 | help | ||
1345 | Support for Console on the NWP serial ports. | ||
1346 | |||
1330 | config SERIAL_QE | 1347 | config SERIAL_QE |
1331 | tristate "Freescale QUICC Engine serial port support" | 1348 | tristate "Freescale QUICC Engine serial port support" |
1332 | depends on QUICC_ENGINE | 1349 | depends on QUICC_ENGINE |
diff --git a/drivers/serial/Makefile b/drivers/serial/Makefile index dfe775ac45b2..8844c0a03929 100644 --- a/drivers/serial/Makefile +++ b/drivers/serial/Makefile | |||
@@ -72,6 +72,7 @@ obj-$(CONFIG_SERIAL_ATMEL) += atmel_serial.o | |||
72 | obj-$(CONFIG_SERIAL_UARTLITE) += uartlite.o | 72 | obj-$(CONFIG_SERIAL_UARTLITE) += uartlite.o |
73 | obj-$(CONFIG_SERIAL_NETX) += netx-serial.o | 73 | obj-$(CONFIG_SERIAL_NETX) += netx-serial.o |
74 | obj-$(CONFIG_SERIAL_OF_PLATFORM) += of_serial.o | 74 | obj-$(CONFIG_SERIAL_OF_PLATFORM) += of_serial.o |
75 | obj-$(CONFIG_SERIAL_OF_PLATFORM_NWPSERIAL) += nwpserial.o | ||
75 | obj-$(CONFIG_SERIAL_KS8695) += serial_ks8695.o | 76 | obj-$(CONFIG_SERIAL_KS8695) += serial_ks8695.o |
76 | obj-$(CONFIG_KGDB_SERIAL_CONSOLE) += kgdboc.o | 77 | obj-$(CONFIG_KGDB_SERIAL_CONSOLE) += kgdboc.o |
77 | obj-$(CONFIG_SERIAL_QE) += ucc_uart.o | 78 | obj-$(CONFIG_SERIAL_QE) += ucc_uart.o |
diff --git a/drivers/serial/nwpserial.c b/drivers/serial/nwpserial.c new file mode 100644 index 000000000000..32f3eaf0d262 --- /dev/null +++ b/drivers/serial/nwpserial.c | |||
@@ -0,0 +1,475 @@ | |||
1 | /* | ||
2 | * Serial Port driver for a NWP uart device | ||
3 | * | ||
4 | * Copyright (C) 2008 IBM Corp., Benjamin Krill <ben@codiert.org> | ||
5 | * | ||
6 | * This program is free software; you can redistribute it and/or | ||
7 | * modify it under the terms of the GNU General Public License | ||
8 | * as published by the Free Software Foundation; either version | ||
9 | * 2 of the License, or (at your option) any later version. | ||
10 | * | ||
11 | */ | ||
12 | #include <linux/init.h> | ||
13 | #include <linux/console.h> | ||
14 | #include <linux/serial.h> | ||
15 | #include <linux/serial_reg.h> | ||
16 | #include <linux/serial_core.h> | ||
17 | #include <linux/tty.h> | ||
18 | #include <linux/irqreturn.h> | ||
19 | #include <linux/mutex.h> | ||
20 | #include <linux/of_platform.h> | ||
21 | #include <linux/of_device.h> | ||
22 | #include <linux/nwpserial.h> | ||
23 | #include <asm/prom.h> | ||
24 | #include <asm/dcr.h> | ||
25 | |||
26 | #define NWPSERIAL_NR 2 | ||
27 | |||
28 | #define NWPSERIAL_STATUS_RXVALID 0x1 | ||
29 | #define NWPSERIAL_STATUS_TXFULL 0x2 | ||
30 | |||
31 | struct nwpserial_port { | ||
32 | struct uart_port port; | ||
33 | dcr_host_t dcr_host; | ||
34 | unsigned int ier; | ||
35 | unsigned int mcr; | ||
36 | }; | ||
37 | |||
38 | static DEFINE_MUTEX(nwpserial_mutex); | ||
39 | static struct nwpserial_port nwpserial_ports[NWPSERIAL_NR]; | ||
40 | |||
41 | static void wait_for_bits(struct nwpserial_port *up, int bits) | ||
42 | { | ||
43 | unsigned int status, tmout = 10000; | ||
44 | |||
45 | /* Wait up to 10ms for the character(s) to be sent. */ | ||
46 | do { | ||
47 | status = dcr_read(up->dcr_host, UART_LSR); | ||
48 | |||
49 | if (--tmout == 0) | ||
50 | break; | ||
51 | udelay(1); | ||
52 | } while ((status & bits) != bits); | ||
53 | } | ||
54 | |||
55 | #ifdef CONFIG_SERIAL_OF_PLATFORM_NWPSERIAL_CONSOLE | ||
56 | static void nwpserial_console_putchar(struct uart_port *port, int c) | ||
57 | { | ||
58 | struct nwpserial_port *up; | ||
59 | up = container_of(port, struct nwpserial_port, port); | ||
60 | /* check if tx buffer is full */ | ||
61 | wait_for_bits(up, UART_LSR_THRE); | ||
62 | dcr_write(up->dcr_host, UART_TX, c); | ||
63 | up->port.icount.tx++; | ||
64 | } | ||
65 | |||
66 | static void | ||
67 | nwpserial_console_write(struct console *co, const char *s, unsigned int count) | ||
68 | { | ||
69 | struct nwpserial_port *up = &nwpserial_ports[co->index]; | ||
70 | unsigned long flags; | ||
71 | int locked = 1; | ||
72 | |||
73 | if (oops_in_progress) | ||
74 | locked = spin_trylock_irqsave(&up->port.lock, flags); | ||
75 | else | ||
76 | spin_lock_irqsave(&up->port.lock, flags); | ||
77 | |||
78 | /* save and disable interrupt */ | ||
79 | up->ier = dcr_read(up->dcr_host, UART_IER); | ||
80 | dcr_write(up->dcr_host, UART_IER, up->ier & ~UART_IER_RDI); | ||
81 | |||
82 | uart_console_write(&up->port, s, count, nwpserial_console_putchar); | ||
83 | |||
84 | /* wait for transmitter to become emtpy */ | ||
85 | while ((dcr_read(up->dcr_host, UART_LSR) & UART_LSR_THRE) == 0) | ||
86 | cpu_relax(); | ||
87 | |||
88 | /* restore interrupt state */ | ||
89 | dcr_write(up->dcr_host, UART_IER, up->ier); | ||
90 | |||
91 | if (locked) | ||
92 | spin_unlock_irqrestore(&up->port.lock, flags); | ||
93 | } | ||
94 | |||
95 | static struct uart_driver nwpserial_reg; | ||
96 | static struct console nwpserial_console = { | ||
97 | .name = "ttySQ", | ||
98 | .write = nwpserial_console_write, | ||
99 | .device = uart_console_device, | ||
100 | .flags = CON_PRINTBUFFER, | ||
101 | .index = -1, | ||
102 | .data = &nwpserial_reg, | ||
103 | }; | ||
104 | #define NWPSERIAL_CONSOLE (&nwpserial_console) | ||
105 | #else | ||
106 | #define NWPSERIAL_CONSOLE NULL | ||
107 | #endif /* CONFIG_SERIAL_OF_PLATFORM_NWPSERIAL_CONSOLE */ | ||
108 | |||
109 | /**************************************************************************/ | ||
110 | |||
111 | static int nwpserial_request_port(struct uart_port *port) | ||
112 | { | ||
113 | return 0; | ||
114 | } | ||
115 | |||
116 | static void nwpserial_release_port(struct uart_port *port) | ||
117 | { | ||
118 | /* N/A */ | ||
119 | } | ||
120 | |||
121 | static void nwpserial_config_port(struct uart_port *port, int flags) | ||
122 | { | ||
123 | port->type = PORT_NWPSERIAL; | ||
124 | } | ||
125 | |||
126 | static irqreturn_t nwpserial_interrupt(int irq, void *dev_id) | ||
127 | { | ||
128 | struct nwpserial_port *up = dev_id; | ||
129 | struct tty_struct *tty = up->port.info->port.tty; | ||
130 | irqreturn_t ret; | ||
131 | unsigned int iir; | ||
132 | unsigned char ch; | ||
133 | |||
134 | spin_lock(&up->port.lock); | ||
135 | |||
136 | /* check if the uart was the interrupt source. */ | ||
137 | iir = dcr_read(up->dcr_host, UART_IIR); | ||
138 | if (!iir) { | ||
139 | ret = IRQ_NONE; | ||
140 | goto out; | ||
141 | } | ||
142 | |||
143 | do { | ||
144 | up->port.icount.rx++; | ||
145 | ch = dcr_read(up->dcr_host, UART_RX); | ||
146 | if (up->port.ignore_status_mask != NWPSERIAL_STATUS_RXVALID) | ||
147 | tty_insert_flip_char(tty, ch, TTY_NORMAL); | ||
148 | } while (dcr_read(up->dcr_host, UART_RX) & UART_LSR_DR); | ||
149 | |||
150 | tty_flip_buffer_push(tty); | ||
151 | ret = IRQ_HANDLED; | ||
152 | |||
153 | out: | ||
154 | spin_unlock(&up->port.lock); | ||
155 | return ret; | ||
156 | } | ||
157 | |||
158 | static int nwpserial_startup(struct uart_port *port) | ||
159 | { | ||
160 | struct nwpserial_port *up; | ||
161 | int err; | ||
162 | |||
163 | up = container_of(port, struct nwpserial_port, port); | ||
164 | |||
165 | /* disable flow control by default */ | ||
166 | up->mcr = dcr_read(up->dcr_host, UART_MCR) & ~UART_MCR_AFE; | ||
167 | dcr_write(up->dcr_host, UART_MCR, up->mcr); | ||
168 | |||
169 | /* register interrupt handler */ | ||
170 | err = request_irq(up->port.irq, nwpserial_interrupt, | ||
171 | IRQF_SHARED, "nwpserial", up); | ||
172 | if (err) | ||
173 | return err; | ||
174 | |||
175 | /* enable interrupts */ | ||
176 | up->ier = UART_IER_RDI; | ||
177 | dcr_write(up->dcr_host, UART_IER, up->ier); | ||
178 | |||
179 | /* enable receiving */ | ||
180 | up->port.ignore_status_mask &= ~NWPSERIAL_STATUS_RXVALID; | ||
181 | |||
182 | return 0; | ||
183 | } | ||
184 | |||
185 | static void nwpserial_shutdown(struct uart_port *port) | ||
186 | { | ||
187 | struct nwpserial_port *up; | ||
188 | up = container_of(port, struct nwpserial_port, port); | ||
189 | |||
190 | /* disable receiving */ | ||
191 | up->port.ignore_status_mask |= NWPSERIAL_STATUS_RXVALID; | ||
192 | |||
193 | /* disable interrupts from this port */ | ||
194 | up->ier = 0; | ||
195 | dcr_write(up->dcr_host, UART_IER, up->ier); | ||
196 | |||
197 | /* free irq */ | ||
198 | free_irq(up->port.irq, port); | ||
199 | } | ||
200 | |||
201 | static int nwpserial_verify_port(struct uart_port *port, | ||
202 | struct serial_struct *ser) | ||
203 | { | ||
204 | return -EINVAL; | ||
205 | } | ||
206 | |||
207 | static const char *nwpserial_type(struct uart_port *port) | ||
208 | { | ||
209 | return port->type == PORT_NWPSERIAL ? "nwpserial" : NULL; | ||
210 | } | ||
211 | |||
212 | static void nwpserial_set_termios(struct uart_port *port, | ||
213 | struct ktermios *termios, struct ktermios *old) | ||
214 | { | ||
215 | struct nwpserial_port *up; | ||
216 | up = container_of(port, struct nwpserial_port, port); | ||
217 | |||
218 | up->port.read_status_mask = NWPSERIAL_STATUS_RXVALID | ||
219 | | NWPSERIAL_STATUS_TXFULL; | ||
220 | |||
221 | up->port.ignore_status_mask = 0; | ||
222 | /* ignore all characters if CREAD is not set */ | ||
223 | if ((termios->c_cflag & CREAD) == 0) | ||
224 | up->port.ignore_status_mask |= NWPSERIAL_STATUS_RXVALID; | ||
225 | |||
226 | /* Copy back the old hardware settings */ | ||
227 | if (old) | ||
228 | tty_termios_copy_hw(termios, old); | ||
229 | } | ||
230 | |||
231 | static void nwpserial_break_ctl(struct uart_port *port, int ctl) | ||
232 | { | ||
233 | /* N/A */ | ||
234 | } | ||
235 | |||
236 | static void nwpserial_enable_ms(struct uart_port *port) | ||
237 | { | ||
238 | /* N/A */ | ||
239 | } | ||
240 | |||
241 | static void nwpserial_stop_rx(struct uart_port *port) | ||
242 | { | ||
243 | struct nwpserial_port *up; | ||
244 | up = container_of(port, struct nwpserial_port, port); | ||
245 | /* don't forward any more data (like !CREAD) */ | ||
246 | up->port.ignore_status_mask = NWPSERIAL_STATUS_RXVALID; | ||
247 | } | ||
248 | |||
249 | static void nwpserial_putchar(struct nwpserial_port *up, unsigned char c) | ||
250 | { | ||
251 | /* check if tx buffer is full */ | ||
252 | wait_for_bits(up, UART_LSR_THRE); | ||
253 | dcr_write(up->dcr_host, UART_TX, c); | ||
254 | up->port.icount.tx++; | ||
255 | } | ||
256 | |||
257 | static void nwpserial_start_tx(struct uart_port *port) | ||
258 | { | ||
259 | struct nwpserial_port *up; | ||
260 | struct circ_buf *xmit; | ||
261 | up = container_of(port, struct nwpserial_port, port); | ||
262 | xmit = &up->port.info->xmit; | ||
263 | |||
264 | if (port->x_char) { | ||
265 | nwpserial_putchar(up, up->port.x_char); | ||
266 | port->x_char = 0; | ||
267 | } | ||
268 | |||
269 | while (!(uart_circ_empty(xmit) || uart_tx_stopped(&up->port))) { | ||
270 | nwpserial_putchar(up, xmit->buf[xmit->tail]); | ||
271 | xmit->tail = (xmit->tail + 1) & (UART_XMIT_SIZE-1); | ||
272 | } | ||
273 | } | ||
274 | |||
275 | static unsigned int nwpserial_get_mctrl(struct uart_port *port) | ||
276 | { | ||
277 | return 0; | ||
278 | } | ||
279 | |||
280 | static void nwpserial_set_mctrl(struct uart_port *port, unsigned int mctrl) | ||
281 | { | ||
282 | /* N/A */ | ||
283 | } | ||
284 | |||
285 | static void nwpserial_stop_tx(struct uart_port *port) | ||
286 | { | ||
287 | /* N/A */ | ||
288 | } | ||
289 | |||
290 | static unsigned int nwpserial_tx_empty(struct uart_port *port) | ||
291 | { | ||
292 | struct nwpserial_port *up; | ||
293 | unsigned long flags; | ||
294 | int ret; | ||
295 | up = container_of(port, struct nwpserial_port, port); | ||
296 | |||
297 | spin_lock_irqsave(&up->port.lock, flags); | ||
298 | ret = dcr_read(up->dcr_host, UART_LSR); | ||
299 | spin_unlock_irqrestore(&up->port.lock, flags); | ||
300 | |||
301 | return ret & UART_LSR_TEMT ? TIOCSER_TEMT : 0; | ||
302 | } | ||
303 | |||
304 | static struct uart_ops nwpserial_pops = { | ||
305 | .tx_empty = nwpserial_tx_empty, | ||
306 | .set_mctrl = nwpserial_set_mctrl, | ||
307 | .get_mctrl = nwpserial_get_mctrl, | ||
308 | .stop_tx = nwpserial_stop_tx, | ||
309 | .start_tx = nwpserial_start_tx, | ||
310 | .stop_rx = nwpserial_stop_rx, | ||
311 | .enable_ms = nwpserial_enable_ms, | ||
312 | .break_ctl = nwpserial_break_ctl, | ||
313 | .startup = nwpserial_startup, | ||
314 | .shutdown = nwpserial_shutdown, | ||
315 | .set_termios = nwpserial_set_termios, | ||
316 | .type = nwpserial_type, | ||
317 | .release_port = nwpserial_release_port, | ||
318 | .request_port = nwpserial_request_port, | ||
319 | .config_port = nwpserial_config_port, | ||
320 | .verify_port = nwpserial_verify_port, | ||
321 | }; | ||
322 | |||
323 | static struct uart_driver nwpserial_reg = { | ||
324 | .owner = THIS_MODULE, | ||
325 | .driver_name = "nwpserial", | ||
326 | .dev_name = "ttySQ", | ||
327 | .major = TTY_MAJOR, | ||
328 | .minor = 68, | ||
329 | .nr = NWPSERIAL_NR, | ||
330 | .cons = NWPSERIAL_CONSOLE, | ||
331 | }; | ||
332 | |||
333 | int nwpserial_register_port(struct uart_port *port) | ||
334 | { | ||
335 | struct nwpserial_port *up = NULL; | ||
336 | int ret = -1; | ||
337 | int i; | ||
338 | static int first = 1; | ||
339 | int dcr_len; | ||
340 | int dcr_base; | ||
341 | struct device_node *dn; | ||
342 | |||
343 | mutex_lock(&nwpserial_mutex); | ||
344 | |||
345 | dn = to_of_device(port->dev)->node; | ||
346 | if (dn == NULL) | ||
347 | goto out; | ||
348 | |||
349 | /* get dcr base. */ | ||
350 | dcr_base = dcr_resource_start(dn, 0); | ||
351 | |||
352 | /* find matching entry */ | ||
353 | for (i = 0; i < NWPSERIAL_NR; i++) | ||
354 | if (nwpserial_ports[i].port.iobase == dcr_base) { | ||
355 | up = &nwpserial_ports[i]; | ||
356 | break; | ||
357 | } | ||
358 | |||
359 | /* we didn't find a mtching entry, search for a free port */ | ||
360 | if (up == NULL) | ||
361 | for (i = 0; i < NWPSERIAL_NR; i++) | ||
362 | if (nwpserial_ports[i].port.type == PORT_UNKNOWN && | ||
363 | nwpserial_ports[i].port.iobase == 0) { | ||
364 | up = &nwpserial_ports[i]; | ||
365 | break; | ||
366 | } | ||
367 | |||
368 | if (up == NULL) { | ||
369 | ret = -EBUSY; | ||
370 | goto out; | ||
371 | } | ||
372 | |||
373 | if (first) | ||
374 | uart_register_driver(&nwpserial_reg); | ||
375 | first = 0; | ||
376 | |||
377 | up->port.membase = port->membase; | ||
378 | up->port.irq = port->irq; | ||
379 | up->port.uartclk = port->uartclk; | ||
380 | up->port.fifosize = port->fifosize; | ||
381 | up->port.regshift = port->regshift; | ||
382 | up->port.iotype = port->iotype; | ||
383 | up->port.flags = port->flags; | ||
384 | up->port.mapbase = port->mapbase; | ||
385 | up->port.private_data = port->private_data; | ||
386 | |||
387 | if (port->dev) | ||
388 | up->port.dev = port->dev; | ||
389 | |||
390 | if (up->port.iobase != dcr_base) { | ||
391 | up->port.ops = &nwpserial_pops; | ||
392 | up->port.fifosize = 16; | ||
393 | |||
394 | spin_lock_init(&up->port.lock); | ||
395 | |||
396 | up->port.iobase = dcr_base; | ||
397 | dcr_len = dcr_resource_len(dn, 0); | ||
398 | |||
399 | up->dcr_host = dcr_map(dn, dcr_base, dcr_len); | ||
400 | if (!DCR_MAP_OK(up->dcr_host)) { | ||
401 | printk(KERN_ERR "Cannot map DCR resources for NWPSERIAL"); | ||
402 | goto out; | ||
403 | } | ||
404 | } | ||
405 | |||
406 | ret = uart_add_one_port(&nwpserial_reg, &up->port); | ||
407 | if (ret == 0) | ||
408 | ret = up->port.line; | ||
409 | |||
410 | out: | ||
411 | mutex_unlock(&nwpserial_mutex); | ||
412 | |||
413 | return ret; | ||
414 | } | ||
415 | EXPORT_SYMBOL(nwpserial_register_port); | ||
416 | |||
417 | void nwpserial_unregister_port(int line) | ||
418 | { | ||
419 | struct nwpserial_port *up = &nwpserial_ports[line]; | ||
420 | mutex_lock(&nwpserial_mutex); | ||
421 | uart_remove_one_port(&nwpserial_reg, &up->port); | ||
422 | |||
423 | up->port.type = PORT_UNKNOWN; | ||
424 | |||
425 | mutex_unlock(&nwpserial_mutex); | ||
426 | } | ||
427 | EXPORT_SYMBOL(nwpserial_unregister_port); | ||
428 | |||
429 | #ifdef CONFIG_SERIAL_OF_PLATFORM_NWPSERIAL_CONSOLE | ||
430 | static int __init nwpserial_console_init(void) | ||
431 | { | ||
432 | struct nwpserial_port *up = NULL; | ||
433 | struct device_node *dn; | ||
434 | const char *name; | ||
435 | int dcr_base; | ||
436 | int dcr_len; | ||
437 | int i; | ||
438 | |||
439 | /* search for a free port */ | ||
440 | for (i = 0; i < NWPSERIAL_NR; i++) | ||
441 | if (nwpserial_ports[i].port.type == PORT_UNKNOWN) { | ||
442 | up = &nwpserial_ports[i]; | ||
443 | break; | ||
444 | } | ||
445 | |||
446 | if (up == NULL) | ||
447 | return -1; | ||
448 | |||
449 | name = of_get_property(of_chosen, "linux,stdout-path", NULL); | ||
450 | if (name == NULL) | ||
451 | return -1; | ||
452 | |||
453 | dn = of_find_node_by_path(name); | ||
454 | if (!dn) | ||
455 | return -1; | ||
456 | |||
457 | spin_lock_init(&up->port.lock); | ||
458 | up->port.ops = &nwpserial_pops; | ||
459 | up->port.type = PORT_NWPSERIAL; | ||
460 | up->port.fifosize = 16; | ||
461 | |||
462 | dcr_base = dcr_resource_start(dn, 0); | ||
463 | dcr_len = dcr_resource_len(dn, 0); | ||
464 | up->port.iobase = dcr_base; | ||
465 | |||
466 | up->dcr_host = dcr_map(dn, dcr_base, dcr_len); | ||
467 | if (!DCR_MAP_OK(up->dcr_host)) { | ||
468 | printk("Cannot map DCR resources for SERIAL"); | ||
469 | return -1; | ||
470 | } | ||
471 | register_console(&nwpserial_console); | ||
472 | return 0; | ||
473 | } | ||
474 | console_initcall(nwpserial_console_init); | ||
475 | #endif /* CONFIG_SERIAL_OF_PLATFORM_NWPSERIAL_CONSOLE */ | ||
diff --git a/drivers/serial/of_serial.c b/drivers/serial/of_serial.c index 8fa0ff561e9f..a821e3a3d664 100644 --- a/drivers/serial/of_serial.c +++ b/drivers/serial/of_serial.c | |||
@@ -14,6 +14,7 @@ | |||
14 | #include <linux/serial_core.h> | 14 | #include <linux/serial_core.h> |
15 | #include <linux/serial_8250.h> | 15 | #include <linux/serial_8250.h> |
16 | #include <linux/of_platform.h> | 16 | #include <linux/of_platform.h> |
17 | #include <linux/nwpserial.h> | ||
17 | 18 | ||
18 | #include <asm/prom.h> | 19 | #include <asm/prom.h> |
19 | 20 | ||
@@ -99,9 +100,16 @@ static int __devinit of_platform_serial_probe(struct of_device *ofdev, | |||
99 | goto out; | 100 | goto out; |
100 | 101 | ||
101 | switch (port_type) { | 102 | switch (port_type) { |
103 | #ifdef CONFIG_SERIAL_8250 | ||
102 | case PORT_8250 ... PORT_MAX_8250: | 104 | case PORT_8250 ... PORT_MAX_8250: |
103 | ret = serial8250_register_port(&port); | 105 | ret = serial8250_register_port(&port); |
104 | break; | 106 | break; |
107 | #endif | ||
108 | #ifdef CONFIG_SERIAL_OF_PLATFORM_NWPSERIAL | ||
109 | case PORT_NWPSERIAL: | ||
110 | ret = nwpserial_register_port(&port); | ||
111 | break; | ||
112 | #endif | ||
105 | default: | 113 | default: |
106 | /* need to add code for these */ | 114 | /* need to add code for these */ |
107 | case PORT_UNKNOWN: | 115 | case PORT_UNKNOWN: |
@@ -129,9 +137,16 @@ static int of_platform_serial_remove(struct of_device *ofdev) | |||
129 | { | 137 | { |
130 | struct of_serial_info *info = ofdev->dev.driver_data; | 138 | struct of_serial_info *info = ofdev->dev.driver_data; |
131 | switch (info->type) { | 139 | switch (info->type) { |
140 | #ifdef CONFIG_SERIAL_8250 | ||
132 | case PORT_8250 ... PORT_MAX_8250: | 141 | case PORT_8250 ... PORT_MAX_8250: |
133 | serial8250_unregister_port(info->line); | 142 | serial8250_unregister_port(info->line); |
134 | break; | 143 | break; |
144 | #endif | ||
145 | #ifdef CONFIG_SERIAL_OF_PLATFORM_NWPSERIAL | ||
146 | case PORT_NWPSERIAL: | ||
147 | nwpserial_unregister_port(info->line); | ||
148 | break; | ||
149 | #endif | ||
135 | default: | 150 | default: |
136 | /* need to add code for these */ | 151 | /* need to add code for these */ |
137 | break; | 152 | break; |
@@ -148,6 +163,10 @@ static struct of_device_id __devinitdata of_platform_serial_table[] = { | |||
148 | { .type = "serial", .compatible = "ns16450", .data = (void *)PORT_16450, }, | 163 | { .type = "serial", .compatible = "ns16450", .data = (void *)PORT_16450, }, |
149 | { .type = "serial", .compatible = "ns16550", .data = (void *)PORT_16550, }, | 164 | { .type = "serial", .compatible = "ns16550", .data = (void *)PORT_16550, }, |
150 | { .type = "serial", .compatible = "ns16750", .data = (void *)PORT_16750, }, | 165 | { .type = "serial", .compatible = "ns16750", .data = (void *)PORT_16750, }, |
166 | #ifdef CONFIG_SERIAL_OF_PLATFORM_NWPSERIAL | ||
167 | { .type = "serial", .compatible = "ibm,qpace-nwp-serial", | ||
168 | .data = (void *)PORT_NWPSERIAL, }, | ||
169 | #endif | ||
151 | { .type = "serial", .data = (void *)PORT_UNKNOWN, }, | 170 | { .type = "serial", .data = (void *)PORT_UNKNOWN, }, |
152 | { /* end of list */ }, | 171 | { /* end of list */ }, |
153 | }; | 172 | }; |
diff --git a/drivers/video/amba-clcd.c b/drivers/video/amba-clcd.c index 2ac52fd8cc11..4e046fed1380 100644 --- a/drivers/video/amba-clcd.c +++ b/drivers/video/amba-clcd.c | |||
@@ -24,6 +24,7 @@ | |||
24 | #include <linux/amba/bus.h> | 24 | #include <linux/amba/bus.h> |
25 | #include <linux/amba/clcd.h> | 25 | #include <linux/amba/clcd.h> |
26 | #include <linux/clk.h> | 26 | #include <linux/clk.h> |
27 | #include <linux/hardirq.h> | ||
27 | 28 | ||
28 | #include <asm/sizes.h> | 29 | #include <asm/sizes.h> |
29 | 30 | ||
diff --git a/drivers/w1/masters/Kconfig b/drivers/w1/masters/Kconfig index 90616822cd20..96d2f8e4c275 100644 --- a/drivers/w1/masters/Kconfig +++ b/drivers/w1/masters/Kconfig | |||
@@ -34,6 +34,12 @@ config W1_MASTER_DS2482 | |||
34 | This driver can also be built as a module. If so, the module | 34 | This driver can also be built as a module. If so, the module |
35 | will be called ds2482. | 35 | will be called ds2482. |
36 | 36 | ||
37 | config W1_MASTER_MXC | ||
38 | tristate "Freescale MXC 1-wire busmaster" | ||
39 | depends on W1 && ARCH_MXC | ||
40 | help | ||
41 | Say Y here to enable MXC 1-wire host | ||
42 | |||
37 | config W1_MASTER_DS1WM | 43 | config W1_MASTER_DS1WM |
38 | tristate "Maxim DS1WM 1-wire busmaster" | 44 | tristate "Maxim DS1WM 1-wire busmaster" |
39 | depends on W1 && ARM && HAVE_CLK | 45 | depends on W1 && ARM && HAVE_CLK |
diff --git a/drivers/w1/masters/Makefile b/drivers/w1/masters/Makefile index bc4714a75f3a..c5a3e96fcbab 100644 --- a/drivers/w1/masters/Makefile +++ b/drivers/w1/masters/Makefile | |||
@@ -5,6 +5,8 @@ | |||
5 | obj-$(CONFIG_W1_MASTER_MATROX) += matrox_w1.o | 5 | obj-$(CONFIG_W1_MASTER_MATROX) += matrox_w1.o |
6 | obj-$(CONFIG_W1_MASTER_DS2490) += ds2490.o | 6 | obj-$(CONFIG_W1_MASTER_DS2490) += ds2490.o |
7 | obj-$(CONFIG_W1_MASTER_DS2482) += ds2482.o | 7 | obj-$(CONFIG_W1_MASTER_DS2482) += ds2482.o |
8 | obj-$(CONFIG_W1_MASTER_MXC) += mxc_w1.o | ||
9 | |||
8 | obj-$(CONFIG_W1_MASTER_DS1WM) += ds1wm.o | 10 | obj-$(CONFIG_W1_MASTER_DS1WM) += ds1wm.o |
9 | obj-$(CONFIG_W1_MASTER_GPIO) += w1-gpio.o | 11 | obj-$(CONFIG_W1_MASTER_GPIO) += w1-gpio.o |
10 | obj-$(CONFIG_HDQ_MASTER_OMAP) += omap_hdq.o | 12 | obj-$(CONFIG_HDQ_MASTER_OMAP) += omap_hdq.o |
diff --git a/drivers/w1/masters/mxc_w1.c b/drivers/w1/masters/mxc_w1.c new file mode 100644 index 000000000000..b9d74d0b353e --- /dev/null +++ b/drivers/w1/masters/mxc_w1.c | |||
@@ -0,0 +1,211 @@ | |||
1 | /* | ||
2 | * Copyright 2005-2008 Freescale Semiconductor, Inc. All Rights Reserved. | ||
3 | * Copyright 2008 Luotao Fu, kernel@pengutronix.de | ||
4 | * | ||
5 | * This program is free software; you can redistribute it and/or | ||
6 | * modify it under the terms of the GNU General Public License | ||
7 | * as published by the Free Software Foundation; either version 2 | ||
8 | * of the License, or (at your option) any later version. | ||
9 | * This program is distributed in the hope that it will be useful, | ||
10 | * but WITHOUT ANY WARRANTY; without even the implied warranty of | ||
11 | * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the | ||
12 | * GNU General Public License for more details. | ||
13 | * | ||
14 | * You should have received a copy of the GNU General Public License | ||
15 | * along with this program; if not, write to the Free Software | ||
16 | * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA. | ||
17 | * | ||
18 | */ | ||
19 | |||
20 | #include <linux/module.h> | ||
21 | #include <linux/interrupt.h> | ||
22 | #include <linux/platform_device.h> | ||
23 | #include <linux/clk.h> | ||
24 | #include <linux/delay.h> | ||
25 | #include <linux/io.h> | ||
26 | |||
27 | #include "../w1.h" | ||
28 | #include "../w1_int.h" | ||
29 | #include "../w1_log.h" | ||
30 | |||
31 | /* According to the mx27 Datasheet the reset procedure should take up to about | ||
32 | * 1350us. We set the timeout to 500*100us = 50ms for sure */ | ||
33 | #define MXC_W1_RESET_TIMEOUT 500 | ||
34 | |||
35 | /* | ||
36 | * MXC W1 Register offsets | ||
37 | */ | ||
38 | #define MXC_W1_CONTROL 0x00 | ||
39 | #define MXC_W1_TIME_DIVIDER 0x02 | ||
40 | #define MXC_W1_RESET 0x04 | ||
41 | #define MXC_W1_COMMAND 0x06 | ||
42 | #define MXC_W1_TXRX 0x08 | ||
43 | #define MXC_W1_INTERRUPT 0x0A | ||
44 | #define MXC_W1_INTERRUPT_EN 0x0C | ||
45 | |||
46 | struct mxc_w1_device { | ||
47 | void __iomem *regs; | ||
48 | unsigned int clkdiv; | ||
49 | struct clk *clk; | ||
50 | struct w1_bus_master bus_master; | ||
51 | }; | ||
52 | |||
53 | /* | ||
54 | * this is the low level routine to | ||
55 | * reset the device on the One Wire interface | ||
56 | * on the hardware | ||
57 | */ | ||
58 | static u8 mxc_w1_ds2_reset_bus(void *data) | ||
59 | { | ||
60 | u8 reg_val; | ||
61 | unsigned int timeout_cnt = 0; | ||
62 | struct mxc_w1_device *dev = data; | ||
63 | |||
64 | __raw_writeb(0x80, (dev->regs + MXC_W1_CONTROL)); | ||
65 | |||
66 | while (1) { | ||
67 | reg_val = __raw_readb(dev->regs + MXC_W1_CONTROL); | ||
68 | |||
69 | if (((reg_val >> 7) & 0x1) == 0 || | ||
70 | timeout_cnt > MXC_W1_RESET_TIMEOUT) | ||
71 | break; | ||
72 | else | ||
73 | timeout_cnt++; | ||
74 | |||
75 | udelay(100); | ||
76 | } | ||
77 | return (reg_val >> 7) & 0x1; | ||
78 | } | ||
79 | |||
80 | /* | ||
81 | * this is the low level routine to read/write a bit on the One Wire | ||
82 | * interface on the hardware. It does write 0 if parameter bit is set | ||
83 | * to 0, otherwise a write 1/read. | ||
84 | */ | ||
85 | static u8 mxc_w1_ds2_touch_bit(void *data, u8 bit) | ||
86 | { | ||
87 | struct mxc_w1_device *mdev = data; | ||
88 | void __iomem *ctrl_addr = mdev->regs + MXC_W1_CONTROL; | ||
89 | unsigned int timeout_cnt = 400; /* Takes max. 120us according to | ||
90 | * datasheet. | ||
91 | */ | ||
92 | |||
93 | __raw_writeb((1 << (5 - bit)), ctrl_addr); | ||
94 | |||
95 | while (timeout_cnt--) { | ||
96 | if (!((__raw_readb(ctrl_addr) >> (5 - bit)) & 0x1)) | ||
97 | break; | ||
98 | |||
99 | udelay(1); | ||
100 | } | ||
101 | |||
102 | return ((__raw_readb(ctrl_addr)) >> 3) & 0x1; | ||
103 | } | ||
104 | |||
105 | static int __init mxc_w1_probe(struct platform_device *pdev) | ||
106 | { | ||
107 | struct mxc_w1_device *mdev; | ||
108 | struct resource *res; | ||
109 | int err = 0; | ||
110 | |||
111 | res = platform_get_resource(pdev, IORESOURCE_MEM, 0); | ||
112 | if (!res) | ||
113 | return -ENODEV; | ||
114 | |||
115 | mdev = kzalloc(sizeof(struct mxc_w1_device), GFP_KERNEL); | ||
116 | if (!mdev) | ||
117 | return -ENOMEM; | ||
118 | |||
119 | mdev->clk = clk_get(&pdev->dev, "owire_clk"); | ||
120 | if (!mdev->clk) { | ||
121 | err = -ENODEV; | ||
122 | goto failed_clk; | ||
123 | } | ||
124 | |||
125 | mdev->clkdiv = (clk_get_rate(mdev->clk) / 1000000) - 1; | ||
126 | |||
127 | res = request_mem_region(res->start, resource_size(res), | ||
128 | "mxc_w1"); | ||
129 | if (!res) { | ||
130 | err = -EBUSY; | ||
131 | goto failed_req; | ||
132 | } | ||
133 | |||
134 | mdev->regs = ioremap(res->start, resource_size(res)); | ||
135 | if (!mdev->regs) { | ||
136 | printk(KERN_ERR "Cannot map frame buffer registers\n"); | ||
137 | goto failed_ioremap; | ||
138 | } | ||
139 | |||
140 | clk_enable(mdev->clk); | ||
141 | __raw_writeb(mdev->clkdiv, mdev->regs + MXC_W1_TIME_DIVIDER); | ||
142 | |||
143 | mdev->bus_master.data = mdev; | ||
144 | mdev->bus_master.reset_bus = mxc_w1_ds2_reset_bus; | ||
145 | mdev->bus_master.touch_bit = mxc_w1_ds2_touch_bit; | ||
146 | |||
147 | err = w1_add_master_device(&mdev->bus_master); | ||
148 | |||
149 | if (err) | ||
150 | goto failed_add; | ||
151 | |||
152 | platform_set_drvdata(pdev, mdev); | ||
153 | return 0; | ||
154 | |||
155 | failed_add: | ||
156 | iounmap(mdev->regs); | ||
157 | failed_ioremap: | ||
158 | release_mem_region(res->start, resource_size(res)); | ||
159 | failed_req: | ||
160 | clk_put(mdev->clk); | ||
161 | failed_clk: | ||
162 | kfree(mdev); | ||
163 | return err; | ||
164 | } | ||
165 | |||
166 | /* | ||
167 | * disassociate the w1 device from the driver | ||
168 | */ | ||
169 | static int mxc_w1_remove(struct platform_device *pdev) | ||
170 | { | ||
171 | struct mxc_w1_device *mdev = platform_get_drvdata(pdev); | ||
172 | struct resource *res; | ||
173 | |||
174 | res = platform_get_resource(pdev, IORESOURCE_MEM, 0); | ||
175 | |||
176 | w1_remove_master_device(&mdev->bus_master); | ||
177 | |||
178 | iounmap(mdev->regs); | ||
179 | release_mem_region(res->start, resource_size(res)); | ||
180 | clk_disable(mdev->clk); | ||
181 | clk_put(mdev->clk); | ||
182 | |||
183 | platform_set_drvdata(pdev, NULL); | ||
184 | |||
185 | return 0; | ||
186 | } | ||
187 | |||
188 | static struct platform_driver mxc_w1_driver = { | ||
189 | .driver = { | ||
190 | .name = "mxc_w1", | ||
191 | }, | ||
192 | .probe = mxc_w1_probe, | ||
193 | .remove = mxc_w1_remove, | ||
194 | }; | ||
195 | |||
196 | static int __init mxc_w1_init(void) | ||
197 | { | ||
198 | return platform_driver_register(&mxc_w1_driver); | ||
199 | } | ||
200 | |||
201 | static void mxc_w1_exit(void) | ||
202 | { | ||
203 | platform_driver_unregister(&mxc_w1_driver); | ||
204 | } | ||
205 | |||
206 | module_init(mxc_w1_init); | ||
207 | module_exit(mxc_w1_exit); | ||
208 | |||
209 | MODULE_LICENSE("GPL"); | ||
210 | MODULE_AUTHOR("Freescale Semiconductors Inc"); | ||
211 | MODULE_DESCRIPTION("Driver for One-Wire on MXC"); | ||
diff --git a/drivers/w1/w1.h b/drivers/w1/w1.h index 97304bd83ec9..d8a9709f3449 100644 --- a/drivers/w1/w1.h +++ b/drivers/w1/w1.h | |||
@@ -210,6 +210,7 @@ u8 w1_read_8(struct w1_master *); | |||
210 | int w1_reset_bus(struct w1_master *); | 210 | int w1_reset_bus(struct w1_master *); |
211 | u8 w1_calc_crc8(u8 *, int); | 211 | u8 w1_calc_crc8(u8 *, int); |
212 | void w1_write_block(struct w1_master *, const u8 *, int); | 212 | void w1_write_block(struct w1_master *, const u8 *, int); |
213 | void w1_touch_block(struct w1_master *, u8 *, int); | ||
213 | u8 w1_read_block(struct w1_master *, u8 *, int); | 214 | u8 w1_read_block(struct w1_master *, u8 *, int); |
214 | int w1_reset_select_slave(struct w1_slave *sl); | 215 | int w1_reset_select_slave(struct w1_slave *sl); |
215 | void w1_next_pullup(struct w1_master *, int); | 216 | void w1_next_pullup(struct w1_master *, int); |
diff --git a/drivers/w1/w1_io.c b/drivers/w1/w1_io.c index 5139c25ca962..442bd8bbd4a5 100644 --- a/drivers/w1/w1_io.c +++ b/drivers/w1/w1_io.c | |||
@@ -238,7 +238,6 @@ EXPORT_SYMBOL_GPL(w1_read_8); | |||
238 | * @param dev the master device | 238 | * @param dev the master device |
239 | * @param buf pointer to the data to write | 239 | * @param buf pointer to the data to write |
240 | * @param len the number of bytes to write | 240 | * @param len the number of bytes to write |
241 | * @return the byte read | ||
242 | */ | 241 | */ |
243 | void w1_write_block(struct w1_master *dev, const u8 *buf, int len) | 242 | void w1_write_block(struct w1_master *dev, const u8 *buf, int len) |
244 | { | 243 | { |
@@ -256,6 +255,31 @@ void w1_write_block(struct w1_master *dev, const u8 *buf, int len) | |||
256 | EXPORT_SYMBOL_GPL(w1_write_block); | 255 | EXPORT_SYMBOL_GPL(w1_write_block); |
257 | 256 | ||
258 | /** | 257 | /** |
258 | * Touches a series of bytes. | ||
259 | * | ||
260 | * @param dev the master device | ||
261 | * @param buf pointer to the data to write | ||
262 | * @param len the number of bytes to write | ||
263 | */ | ||
264 | void w1_touch_block(struct w1_master *dev, u8 *buf, int len) | ||
265 | { | ||
266 | int i, j; | ||
267 | u8 tmp; | ||
268 | |||
269 | for (i = 0; i < len; ++i) { | ||
270 | tmp = 0; | ||
271 | for (j = 0; j < 8; ++j) { | ||
272 | if (j == 7) | ||
273 | w1_pre_write(dev); | ||
274 | tmp |= w1_touch_bit(dev, (buf[i] >> j) & 0x1) << j; | ||
275 | } | ||
276 | |||
277 | buf[i] = tmp; | ||
278 | } | ||
279 | } | ||
280 | EXPORT_SYMBOL_GPL(w1_touch_block); | ||
281 | |||
282 | /** | ||
259 | * Reads a series of bytes. | 283 | * Reads a series of bytes. |
260 | * | 284 | * |
261 | * @param dev the master device | 285 | * @param dev the master device |
diff --git a/drivers/w1/w1_netlink.c b/drivers/w1/w1_netlink.c index 65c5ebd0787e..fdf72851c574 100644 --- a/drivers/w1/w1_netlink.c +++ b/drivers/w1/w1_netlink.c | |||
@@ -47,21 +47,56 @@ void w1_netlink_send(struct w1_master *dev, struct w1_netlink_msg *msg) | |||
47 | cn_netlink_send(m, 0, GFP_KERNEL); | 47 | cn_netlink_send(m, 0, GFP_KERNEL); |
48 | } | 48 | } |
49 | 49 | ||
50 | static int w1_process_command_master(struct w1_master *dev, struct cn_msg *msg, | 50 | static void w1_send_slave(struct w1_master *dev, u64 rn) |
51 | struct w1_netlink_msg *hdr, struct w1_netlink_cmd *cmd) | 51 | { |
52 | struct cn_msg *msg = dev->priv; | ||
53 | struct w1_netlink_msg *hdr = (struct w1_netlink_msg *)(msg + 1); | ||
54 | struct w1_netlink_cmd *cmd = (struct w1_netlink_cmd *)(hdr + 1); | ||
55 | int avail; | ||
56 | |||
57 | avail = dev->priv_size - cmd->len; | ||
58 | |||
59 | if (avail > 8) { | ||
60 | u64 *data = (void *)(cmd + 1) + cmd->len; | ||
61 | |||
62 | *data = rn; | ||
63 | cmd->len += 8; | ||
64 | hdr->len += 8; | ||
65 | msg->len += 8; | ||
66 | return; | ||
67 | } | ||
68 | |||
69 | msg->ack++; | ||
70 | cn_netlink_send(msg, 0, GFP_KERNEL); | ||
71 | |||
72 | msg->len = sizeof(struct w1_netlink_msg) + sizeof(struct w1_netlink_cmd); | ||
73 | hdr->len = sizeof(struct w1_netlink_cmd); | ||
74 | cmd->len = 0; | ||
75 | } | ||
76 | |||
77 | static int w1_process_search_command(struct w1_master *dev, struct cn_msg *msg, | ||
78 | unsigned int avail) | ||
52 | { | 79 | { |
53 | dev_dbg(&dev->dev, "%s: %s: cmd=%02x, len=%u.\n", | 80 | struct w1_netlink_msg *hdr = (struct w1_netlink_msg *)(msg + 1); |
54 | __func__, dev->name, cmd->cmd, cmd->len); | 81 | struct w1_netlink_cmd *cmd = (struct w1_netlink_cmd *)(hdr + 1); |
82 | int search_type = (cmd->cmd == W1_CMD_ALARM_SEARCH)?W1_ALARM_SEARCH:W1_SEARCH; | ||
55 | 83 | ||
56 | if (cmd->cmd != W1_CMD_SEARCH && cmd->cmd != W1_CMD_ALARM_SEARCH) | 84 | dev->priv = msg; |
57 | return -EINVAL; | 85 | dev->priv_size = avail; |
86 | |||
87 | w1_search_devices(dev, search_type, w1_send_slave); | ||
88 | |||
89 | msg->ack = 0; | ||
90 | cn_netlink_send(msg, 0, GFP_KERNEL); | ||
91 | |||
92 | dev->priv = NULL; | ||
93 | dev->priv_size = 0; | ||
58 | 94 | ||
59 | w1_search_process(dev, (cmd->cmd == W1_CMD_ALARM_SEARCH)?W1_ALARM_SEARCH:W1_SEARCH); | ||
60 | return 0; | 95 | return 0; |
61 | } | 96 | } |
62 | 97 | ||
63 | static int w1_send_read_reply(struct w1_slave *sl, struct cn_msg *msg, | 98 | static int w1_send_read_reply(struct cn_msg *msg, struct w1_netlink_msg *hdr, |
64 | struct w1_netlink_msg *hdr, struct w1_netlink_cmd *cmd) | 99 | struct w1_netlink_cmd *cmd) |
65 | { | 100 | { |
66 | void *data; | 101 | void *data; |
67 | struct w1_netlink_msg *h; | 102 | struct w1_netlink_msg *h; |
@@ -85,7 +120,8 @@ static int w1_send_read_reply(struct w1_slave *sl, struct cn_msg *msg, | |||
85 | memcpy(c, cmd, sizeof(struct w1_netlink_cmd)); | 120 | memcpy(c, cmd, sizeof(struct w1_netlink_cmd)); |
86 | 121 | ||
87 | cm->ack = msg->seq+1; | 122 | cm->ack = msg->seq+1; |
88 | cm->len = sizeof(struct w1_netlink_msg) + sizeof(struct w1_netlink_cmd) + cmd->len; | 123 | cm->len = sizeof(struct w1_netlink_msg) + |
124 | sizeof(struct w1_netlink_cmd) + cmd->len; | ||
89 | 125 | ||
90 | h->len = sizeof(struct w1_netlink_cmd) + cmd->len; | 126 | h->len = sizeof(struct w1_netlink_cmd) + cmd->len; |
91 | 127 | ||
@@ -98,36 +134,178 @@ static int w1_send_read_reply(struct w1_slave *sl, struct cn_msg *msg, | |||
98 | return err; | 134 | return err; |
99 | } | 135 | } |
100 | 136 | ||
101 | static int w1_process_command_slave(struct w1_slave *sl, struct cn_msg *msg, | 137 | static int w1_process_command_io(struct w1_master *dev, struct cn_msg *msg, |
102 | struct w1_netlink_msg *hdr, struct w1_netlink_cmd *cmd) | 138 | struct w1_netlink_msg *hdr, struct w1_netlink_cmd *cmd) |
103 | { | 139 | { |
104 | int err = 0; | 140 | int err = 0; |
105 | 141 | ||
106 | dev_dbg(&sl->master->dev, "%s: %02x.%012llx.%02x: cmd=%02x, len=%u.\n", | 142 | switch (cmd->cmd) { |
107 | __func__, sl->reg_num.family, (unsigned long long)sl->reg_num.id, sl->reg_num.crc, | 143 | case W1_CMD_TOUCH: |
108 | cmd->cmd, cmd->len); | 144 | w1_touch_block(dev, cmd->data, cmd->len); |
145 | w1_send_read_reply(msg, hdr, cmd); | ||
146 | break; | ||
147 | case W1_CMD_READ: | ||
148 | w1_read_block(dev, cmd->data, cmd->len); | ||
149 | w1_send_read_reply(msg, hdr, cmd); | ||
150 | break; | ||
151 | case W1_CMD_WRITE: | ||
152 | w1_write_block(dev, cmd->data, cmd->len); | ||
153 | break; | ||
154 | default: | ||
155 | err = -EINVAL; | ||
156 | break; | ||
157 | } | ||
158 | |||
159 | return err; | ||
160 | } | ||
161 | |||
162 | static int w1_process_command_master(struct w1_master *dev, struct cn_msg *req_msg, | ||
163 | struct w1_netlink_msg *req_hdr, struct w1_netlink_cmd *req_cmd) | ||
164 | { | ||
165 | int err = -EINVAL; | ||
166 | struct cn_msg *msg; | ||
167 | struct w1_netlink_msg *hdr; | ||
168 | struct w1_netlink_cmd *cmd; | ||
169 | |||
170 | msg = kzalloc(PAGE_SIZE, GFP_KERNEL); | ||
171 | if (!msg) | ||
172 | return -ENOMEM; | ||
173 | |||
174 | msg->id = req_msg->id; | ||
175 | msg->seq = req_msg->seq; | ||
176 | msg->ack = 0; | ||
177 | msg->len = sizeof(struct w1_netlink_msg) + sizeof(struct w1_netlink_cmd); | ||
178 | |||
179 | hdr = (struct w1_netlink_msg *)(msg + 1); | ||
180 | cmd = (struct w1_netlink_cmd *)(hdr + 1); | ||
181 | |||
182 | hdr->type = W1_MASTER_CMD; | ||
183 | hdr->id = req_hdr->id; | ||
184 | hdr->len = sizeof(struct w1_netlink_cmd); | ||
185 | |||
186 | cmd->cmd = req_cmd->cmd; | ||
187 | cmd->len = 0; | ||
109 | 188 | ||
110 | switch (cmd->cmd) { | 189 | switch (cmd->cmd) { |
111 | case W1_CMD_READ: | 190 | case W1_CMD_SEARCH: |
112 | w1_read_block(sl->master, cmd->data, cmd->len); | 191 | case W1_CMD_ALARM_SEARCH: |
113 | w1_send_read_reply(sl, msg, hdr, cmd); | 192 | err = w1_process_search_command(dev, msg, |
114 | break; | 193 | PAGE_SIZE - msg->len - sizeof(struct cn_msg)); |
115 | case W1_CMD_WRITE: | 194 | break; |
116 | w1_write_block(sl->master, cmd->data, cmd->len); | 195 | case W1_CMD_READ: |
117 | break; | 196 | case W1_CMD_WRITE: |
118 | case W1_CMD_SEARCH: | 197 | case W1_CMD_TOUCH: |
119 | case W1_CMD_ALARM_SEARCH: | 198 | err = w1_process_command_io(dev, req_msg, req_hdr, req_cmd); |
120 | w1_search_process(sl->master, | 199 | break; |
121 | (cmd->cmd == W1_CMD_ALARM_SEARCH)?W1_ALARM_SEARCH:W1_SEARCH); | 200 | case W1_CMD_RESET: |
122 | break; | 201 | err = w1_reset_bus(dev); |
123 | default: | 202 | break; |
124 | err = -1; | 203 | default: |
125 | break; | 204 | err = -EINVAL; |
205 | break; | ||
126 | } | 206 | } |
127 | 207 | ||
208 | kfree(msg); | ||
128 | return err; | 209 | return err; |
129 | } | 210 | } |
130 | 211 | ||
212 | static int w1_process_command_slave(struct w1_slave *sl, struct cn_msg *msg, | ||
213 | struct w1_netlink_msg *hdr, struct w1_netlink_cmd *cmd) | ||
214 | { | ||
215 | dev_dbg(&sl->master->dev, "%s: %02x.%012llx.%02x: cmd=%02x, len=%u.\n", | ||
216 | __func__, sl->reg_num.family, (unsigned long long)sl->reg_num.id, | ||
217 | sl->reg_num.crc, cmd->cmd, cmd->len); | ||
218 | |||
219 | return w1_process_command_io(sl->master, msg, hdr, cmd); | ||
220 | } | ||
221 | |||
222 | static int w1_process_command_root(struct cn_msg *msg, struct w1_netlink_msg *mcmd) | ||
223 | { | ||
224 | struct w1_master *m; | ||
225 | struct cn_msg *cn; | ||
226 | struct w1_netlink_msg *w; | ||
227 | u32 *id; | ||
228 | |||
229 | if (mcmd->type != W1_LIST_MASTERS) { | ||
230 | printk(KERN_NOTICE "%s: msg: %x.%x, wrong type: %u, len: %u.\n", | ||
231 | __func__, msg->id.idx, msg->id.val, mcmd->type, mcmd->len); | ||
232 | return -EPROTO; | ||
233 | } | ||
234 | |||
235 | cn = kmalloc(PAGE_SIZE, GFP_KERNEL); | ||
236 | if (!cn) | ||
237 | return -ENOMEM; | ||
238 | |||
239 | cn->id.idx = CN_W1_IDX; | ||
240 | cn->id.val = CN_W1_VAL; | ||
241 | |||
242 | cn->seq = msg->seq; | ||
243 | cn->ack = 1; | ||
244 | cn->len = sizeof(struct w1_netlink_msg); | ||
245 | w = (struct w1_netlink_msg *)(cn + 1); | ||
246 | |||
247 | w->type = W1_LIST_MASTERS; | ||
248 | w->status = 0; | ||
249 | w->len = 0; | ||
250 | id = (u32 *)(w + 1); | ||
251 | |||
252 | mutex_lock(&w1_mlock); | ||
253 | list_for_each_entry(m, &w1_masters, w1_master_entry) { | ||
254 | if (cn->len + sizeof(*id) > PAGE_SIZE - sizeof(struct cn_msg)) { | ||
255 | cn_netlink_send(cn, 0, GFP_KERNEL); | ||
256 | cn->ack++; | ||
257 | cn->len = sizeof(struct w1_netlink_msg); | ||
258 | w->len = 0; | ||
259 | id = (u32 *)(w + 1); | ||
260 | } | ||
261 | |||
262 | *id = m->id; | ||
263 | w->len += sizeof(*id); | ||
264 | cn->len += sizeof(*id); | ||
265 | id++; | ||
266 | } | ||
267 | cn->ack = 0; | ||
268 | cn_netlink_send(cn, 0, GFP_KERNEL); | ||
269 | mutex_unlock(&w1_mlock); | ||
270 | |||
271 | kfree(cn); | ||
272 | return 0; | ||
273 | } | ||
274 | |||
275 | static int w1_netlink_send_error(struct cn_msg *rcmsg, struct w1_netlink_msg *rmsg, | ||
276 | struct w1_netlink_cmd *rcmd, int error) | ||
277 | { | ||
278 | struct cn_msg *cmsg; | ||
279 | struct w1_netlink_msg *msg; | ||
280 | struct w1_netlink_cmd *cmd; | ||
281 | |||
282 | cmsg = kzalloc(sizeof(*msg) + sizeof(*cmd) + sizeof(*cmsg), GFP_KERNEL); | ||
283 | if (!cmsg) | ||
284 | return -ENOMEM; | ||
285 | |||
286 | msg = (struct w1_netlink_msg *)(cmsg + 1); | ||
287 | cmd = (struct w1_netlink_cmd *)(msg + 1); | ||
288 | |||
289 | memcpy(cmsg, rcmsg, sizeof(*cmsg)); | ||
290 | cmsg->len = sizeof(*msg); | ||
291 | |||
292 | memcpy(msg, rmsg, sizeof(*msg)); | ||
293 | msg->len = 0; | ||
294 | msg->status = (short)-error; | ||
295 | |||
296 | if (rcmd) { | ||
297 | memcpy(cmd, rcmd, sizeof(*cmd)); | ||
298 | cmd->len = 0; | ||
299 | msg->len += sizeof(*cmd); | ||
300 | cmsg->len += sizeof(*cmd); | ||
301 | } | ||
302 | |||
303 | error = cn_netlink_send(cmsg, 0, GFP_KERNEL); | ||
304 | kfree(cmsg); | ||
305 | |||
306 | return error; | ||
307 | } | ||
308 | |||
131 | static void w1_cn_callback(void *data) | 309 | static void w1_cn_callback(void *data) |
132 | { | 310 | { |
133 | struct cn_msg *msg = data; | 311 | struct cn_msg *msg = data; |
@@ -144,6 +322,7 @@ static void w1_cn_callback(void *data) | |||
144 | 322 | ||
145 | dev = NULL; | 323 | dev = NULL; |
146 | sl = NULL; | 324 | sl = NULL; |
325 | cmd = NULL; | ||
147 | 326 | ||
148 | memcpy(&id, m->id.id, sizeof(id)); | 327 | memcpy(&id, m->id.id, sizeof(id)); |
149 | #if 0 | 328 | #if 0 |
@@ -155,15 +334,15 @@ static void w1_cn_callback(void *data) | |||
155 | break; | 334 | break; |
156 | } | 335 | } |
157 | 336 | ||
158 | if (!mlen) | ||
159 | goto out_cont; | ||
160 | |||
161 | if (m->type == W1_MASTER_CMD) { | 337 | if (m->type == W1_MASTER_CMD) { |
162 | dev = w1_search_master_id(m->id.mst.id); | 338 | dev = w1_search_master_id(m->id.mst.id); |
163 | } else if (m->type == W1_SLAVE_CMD) { | 339 | } else if (m->type == W1_SLAVE_CMD) { |
164 | sl = w1_search_slave(&id); | 340 | sl = w1_search_slave(&id); |
165 | if (sl) | 341 | if (sl) |
166 | dev = sl->master; | 342 | dev = sl->master; |
343 | } else { | ||
344 | err = w1_process_command_root(msg, m); | ||
345 | goto out_cont; | ||
167 | } | 346 | } |
168 | 347 | ||
169 | if (!dev) { | 348 | if (!dev) { |
@@ -171,6 +350,10 @@ static void w1_cn_callback(void *data) | |||
171 | goto out_cont; | 350 | goto out_cont; |
172 | } | 351 | } |
173 | 352 | ||
353 | err = 0; | ||
354 | if (!mlen) | ||
355 | goto out_cont; | ||
356 | |||
174 | mutex_lock(&dev->mutex); | 357 | mutex_lock(&dev->mutex); |
175 | 358 | ||
176 | if (sl && w1_reset_select_slave(sl)) { | 359 | if (sl && w1_reset_select_slave(sl)) { |
@@ -187,9 +370,12 @@ static void w1_cn_callback(void *data) | |||
187 | } | 370 | } |
188 | 371 | ||
189 | if (sl) | 372 | if (sl) |
190 | w1_process_command_slave(sl, msg, m, cmd); | 373 | err = w1_process_command_slave(sl, msg, m, cmd); |
191 | else | 374 | else |
192 | w1_process_command_master(dev, msg, m, cmd); | 375 | err = w1_process_command_master(dev, msg, m, cmd); |
376 | |||
377 | w1_netlink_send_error(msg, m, cmd, err); | ||
378 | err = 0; | ||
193 | 379 | ||
194 | cmd_data += cmd->len + sizeof(struct w1_netlink_cmd); | 380 | cmd_data += cmd->len + sizeof(struct w1_netlink_cmd); |
195 | mlen -= cmd->len + sizeof(struct w1_netlink_cmd); | 381 | mlen -= cmd->len + sizeof(struct w1_netlink_cmd); |
@@ -200,6 +386,8 @@ out_up: | |||
200 | atomic_dec(&sl->refcnt); | 386 | atomic_dec(&sl->refcnt); |
201 | mutex_unlock(&dev->mutex); | 387 | mutex_unlock(&dev->mutex); |
202 | out_cont: | 388 | out_cont: |
389 | if (!cmd || err) | ||
390 | w1_netlink_send_error(msg, m, cmd, err); | ||
203 | msg->len -= sizeof(struct w1_netlink_msg) + m->len; | 391 | msg->len -= sizeof(struct w1_netlink_msg) + m->len; |
204 | m = (struct w1_netlink_msg *)(((u8 *)m) + sizeof(struct w1_netlink_msg) + m->len); | 392 | m = (struct w1_netlink_msg *)(((u8 *)m) + sizeof(struct w1_netlink_msg) + m->len); |
205 | 393 | ||
@@ -209,11 +397,6 @@ out_cont: | |||
209 | if (err == -ENODEV) | 397 | if (err == -ENODEV) |
210 | err = 0; | 398 | err = 0; |
211 | } | 399 | } |
212 | #if 0 | ||
213 | if (err) { | ||
214 | printk("%s: malformed message. Dropping.\n", __func__); | ||
215 | } | ||
216 | #endif | ||
217 | } | 400 | } |
218 | 401 | ||
219 | int w1_init_netlink(void) | 402 | int w1_init_netlink(void) |
diff --git a/drivers/w1/w1_netlink.h b/drivers/w1/w1_netlink.h index 56122b9e9294..27e950f935b1 100644 --- a/drivers/w1/w1_netlink.h +++ b/drivers/w1/w1_netlink.h | |||
@@ -34,12 +34,13 @@ enum w1_netlink_message_types { | |||
34 | W1_MASTER_REMOVE, | 34 | W1_MASTER_REMOVE, |
35 | W1_MASTER_CMD, | 35 | W1_MASTER_CMD, |
36 | W1_SLAVE_CMD, | 36 | W1_SLAVE_CMD, |
37 | W1_LIST_MASTERS, | ||
37 | }; | 38 | }; |
38 | 39 | ||
39 | struct w1_netlink_msg | 40 | struct w1_netlink_msg |
40 | { | 41 | { |
41 | __u8 type; | 42 | __u8 type; |
42 | __u8 reserved; | 43 | __u8 status; |
43 | __u16 len; | 44 | __u16 len; |
44 | union { | 45 | union { |
45 | __u8 id[8]; | 46 | __u8 id[8]; |
@@ -51,10 +52,15 @@ struct w1_netlink_msg | |||
51 | __u8 data[0]; | 52 | __u8 data[0]; |
52 | }; | 53 | }; |
53 | 54 | ||
54 | #define W1_CMD_READ 0x0 | 55 | enum w1_commands { |
55 | #define W1_CMD_WRITE 0x1 | 56 | W1_CMD_READ = 0, |
56 | #define W1_CMD_SEARCH 0x2 | 57 | W1_CMD_WRITE, |
57 | #define W1_CMD_ALARM_SEARCH 0x3 | 58 | W1_CMD_SEARCH, |
59 | W1_CMD_ALARM_SEARCH, | ||
60 | W1_CMD_TOUCH, | ||
61 | W1_CMD_RESET, | ||
62 | W1_CMD_MAX, | ||
63 | }; | ||
58 | 64 | ||
59 | struct w1_netlink_cmd | 65 | struct w1_netlink_cmd |
60 | { | 66 | { |
diff --git a/drivers/xen/Kconfig b/drivers/xen/Kconfig index 4b75a16de009..526187c8a12d 100644 --- a/drivers/xen/Kconfig +++ b/drivers/xen/Kconfig | |||
@@ -17,3 +17,27 @@ config XEN_SCRUB_PAGES | |||
17 | is not accidentally visible to other domains. Is it more | 17 | is not accidentally visible to other domains. Is it more |
18 | secure, but slightly less efficient. | 18 | secure, but slightly less efficient. |
19 | If in doubt, say yes. | 19 | If in doubt, say yes. |
20 | |||
21 | config XENFS | ||
22 | tristate "Xen filesystem" | ||
23 | depends on XEN | ||
24 | default y | ||
25 | help | ||
26 | The xen filesystem provides a way for domains to share | ||
27 | information with each other and with the hypervisor. | ||
28 | For example, by reading and writing the "xenbus" file, guests | ||
29 | may pass arbitrary information to the initial domain. | ||
30 | If in doubt, say yes. | ||
31 | |||
32 | config XEN_COMPAT_XENFS | ||
33 | bool "Create compatibility mount point /proc/xen" | ||
34 | depends on XENFS | ||
35 | default y | ||
36 | help | ||
37 | The old xenstore userspace tools expect to find "xenbus" | ||
38 | under /proc/xen, but "xenbus" is now found at the root of the | ||
39 | xenfs filesystem. Selecting this causes the kernel to create | ||
40 | the compatibilty mount point /proc/xen if it is running on | ||
41 | a xen platform. | ||
42 | If in doubt, say yes. | ||
43 | |||
diff --git a/drivers/xen/Makefile b/drivers/xen/Makefile index d2a8fdf0e191..ff8accc9e103 100644 --- a/drivers/xen/Makefile +++ b/drivers/xen/Makefile | |||
@@ -1,5 +1,7 @@ | |||
1 | obj-y += grant-table.o features.o events.o manage.o | 1 | obj-y += grant-table.o features.o events.o manage.o |
2 | obj-y += xenbus/ | 2 | obj-y += xenbus/ |
3 | |||
3 | obj-$(CONFIG_HOTPLUG_CPU) += cpu_hotplug.o | 4 | obj-$(CONFIG_HOTPLUG_CPU) += cpu_hotplug.o |
4 | obj-$(CONFIG_XEN_XENCOMM) += xencomm.o | 5 | obj-$(CONFIG_XEN_XENCOMM) += xencomm.o |
5 | obj-$(CONFIG_XEN_BALLOON) += balloon.o | 6 | obj-$(CONFIG_XEN_BALLOON) += balloon.o |
7 | obj-$(CONFIG_XENFS) += xenfs/ \ No newline at end of file | ||
diff --git a/drivers/xen/xenbus/xenbus_client.c b/drivers/xen/xenbus/xenbus_client.c index 9678b3e98c63..92a1ef80a288 100644 --- a/drivers/xen/xenbus/xenbus_client.c +++ b/drivers/xen/xenbus/xenbus_client.c | |||
@@ -136,7 +136,6 @@ EXPORT_SYMBOL_GPL(xenbus_watch_pathfmt); | |||
136 | /** | 136 | /** |
137 | * xenbus_switch_state | 137 | * xenbus_switch_state |
138 | * @dev: xenbus device | 138 | * @dev: xenbus device |
139 | * @xbt: transaction handle | ||
140 | * @state: new state | 139 | * @state: new state |
141 | * | 140 | * |
142 | * Advertise in the store a change of the given driver to the given new_state. | 141 | * Advertise in the store a change of the given driver to the given new_state. |
@@ -267,7 +266,7 @@ EXPORT_SYMBOL_GPL(xenbus_dev_error); | |||
267 | * @fmt: error message format | 266 | * @fmt: error message format |
268 | * | 267 | * |
269 | * Equivalent to xenbus_dev_error(dev, err, fmt, args), followed by | 268 | * Equivalent to xenbus_dev_error(dev, err, fmt, args), followed by |
270 | * xenbus_switch_state(dev, NULL, XenbusStateClosing) to schedule an orderly | 269 | * xenbus_switch_state(dev, XenbusStateClosing) to schedule an orderly |
271 | * closedown of this driver and its peer. | 270 | * closedown of this driver and its peer. |
272 | */ | 271 | */ |
273 | 272 | ||
diff --git a/drivers/xen/xenbus/xenbus_probe.c b/drivers/xen/xenbus/xenbus_probe.c index b2a03184a246..773d1cf23283 100644 --- a/drivers/xen/xenbus/xenbus_probe.c +++ b/drivers/xen/xenbus/xenbus_probe.c | |||
@@ -40,6 +40,7 @@ | |||
40 | #include <linux/ctype.h> | 40 | #include <linux/ctype.h> |
41 | #include <linux/fcntl.h> | 41 | #include <linux/fcntl.h> |
42 | #include <linux/mm.h> | 42 | #include <linux/mm.h> |
43 | #include <linux/proc_fs.h> | ||
43 | #include <linux/notifier.h> | 44 | #include <linux/notifier.h> |
44 | #include <linux/kthread.h> | 45 | #include <linux/kthread.h> |
45 | #include <linux/mutex.h> | 46 | #include <linux/mutex.h> |
@@ -55,7 +56,10 @@ | |||
55 | #include "xenbus_comms.h" | 56 | #include "xenbus_comms.h" |
56 | #include "xenbus_probe.h" | 57 | #include "xenbus_probe.h" |
57 | 58 | ||
59 | |||
58 | int xen_store_evtchn; | 60 | int xen_store_evtchn; |
61 | EXPORT_SYMBOL(xen_store_evtchn); | ||
62 | |||
59 | struct xenstore_domain_interface *xen_store_interface; | 63 | struct xenstore_domain_interface *xen_store_interface; |
60 | static unsigned long xen_store_mfn; | 64 | static unsigned long xen_store_mfn; |
61 | 65 | ||
@@ -166,6 +170,9 @@ static int read_backend_details(struct xenbus_device *xendev) | |||
166 | return read_otherend_details(xendev, "backend-id", "backend"); | 170 | return read_otherend_details(xendev, "backend-id", "backend"); |
167 | } | 171 | } |
168 | 172 | ||
173 | static struct device_attribute xenbus_dev_attrs[] = { | ||
174 | __ATTR_NULL | ||
175 | }; | ||
169 | 176 | ||
170 | /* Bus type for frontend drivers. */ | 177 | /* Bus type for frontend drivers. */ |
171 | static struct xen_bus_type xenbus_frontend = { | 178 | static struct xen_bus_type xenbus_frontend = { |
@@ -174,12 +181,13 @@ static struct xen_bus_type xenbus_frontend = { | |||
174 | .get_bus_id = frontend_bus_id, | 181 | .get_bus_id = frontend_bus_id, |
175 | .probe = xenbus_probe_frontend, | 182 | .probe = xenbus_probe_frontend, |
176 | .bus = { | 183 | .bus = { |
177 | .name = "xen", | 184 | .name = "xen", |
178 | .match = xenbus_match, | 185 | .match = xenbus_match, |
179 | .uevent = xenbus_uevent, | 186 | .uevent = xenbus_uevent, |
180 | .probe = xenbus_dev_probe, | 187 | .probe = xenbus_dev_probe, |
181 | .remove = xenbus_dev_remove, | 188 | .remove = xenbus_dev_remove, |
182 | .shutdown = xenbus_dev_shutdown, | 189 | .shutdown = xenbus_dev_shutdown, |
190 | .dev_attrs = xenbus_dev_attrs, | ||
183 | }, | 191 | }, |
184 | }; | 192 | }; |
185 | 193 | ||
@@ -852,6 +860,14 @@ static int __init xenbus_probe_init(void) | |||
852 | if (!xen_initial_domain()) | 860 | if (!xen_initial_domain()) |
853 | xenbus_probe(NULL); | 861 | xenbus_probe(NULL); |
854 | 862 | ||
863 | #ifdef CONFIG_XEN_COMPAT_XENFS | ||
864 | /* | ||
865 | * Create xenfs mountpoint in /proc for compatibility with | ||
866 | * utilities that expect to find "xenbus" under "/proc/xen". | ||
867 | */ | ||
868 | proc_mkdir("xen", NULL); | ||
869 | #endif | ||
870 | |||
855 | return 0; | 871 | return 0; |
856 | 872 | ||
857 | out_unreg_back: | 873 | out_unreg_back: |
diff --git a/drivers/xen/xenbus/xenbus_xs.c b/drivers/xen/xenbus/xenbus_xs.c index 7f2f91c0e11d..e325eab4724d 100644 --- a/drivers/xen/xenbus/xenbus_xs.c +++ b/drivers/xen/xenbus/xenbus_xs.c | |||
@@ -184,6 +184,7 @@ void *xenbus_dev_request_and_reply(struct xsd_sockmsg *msg) | |||
184 | 184 | ||
185 | return ret; | 185 | return ret; |
186 | } | 186 | } |
187 | EXPORT_SYMBOL(xenbus_dev_request_and_reply); | ||
187 | 188 | ||
188 | /* Send message to xs, get kmalloc'ed reply. ERR_PTR() on error. */ | 189 | /* Send message to xs, get kmalloc'ed reply. ERR_PTR() on error. */ |
189 | static void *xs_talkv(struct xenbus_transaction t, | 190 | static void *xs_talkv(struct xenbus_transaction t, |
diff --git a/drivers/xen/xenfs/Makefile b/drivers/xen/xenfs/Makefile new file mode 100644 index 000000000000..25275c3bbdff --- /dev/null +++ b/drivers/xen/xenfs/Makefile | |||
@@ -0,0 +1,3 @@ | |||
1 | obj-$(CONFIG_XENFS) += xenfs.o | ||
2 | |||
3 | xenfs-objs = super.o xenbus.o \ No newline at end of file | ||
diff --git a/drivers/xen/xenfs/super.c b/drivers/xen/xenfs/super.c new file mode 100644 index 000000000000..515741a8e6b8 --- /dev/null +++ b/drivers/xen/xenfs/super.c | |||
@@ -0,0 +1,64 @@ | |||
1 | /* | ||
2 | * xenfs.c - a filesystem for passing info between the a domain and | ||
3 | * the hypervisor. | ||
4 | * | ||
5 | * 2008-10-07 Alex Zeffertt Replaced /proc/xen/xenbus with xenfs filesystem | ||
6 | * and /proc/xen compatibility mount point. | ||
7 | * Turned xenfs into a loadable module. | ||
8 | */ | ||
9 | |||
10 | #include <linux/kernel.h> | ||
11 | #include <linux/errno.h> | ||
12 | #include <linux/module.h> | ||
13 | #include <linux/fs.h> | ||
14 | #include <linux/magic.h> | ||
15 | |||
16 | #include "xenfs.h" | ||
17 | |||
18 | #include <asm/xen/hypervisor.h> | ||
19 | |||
20 | MODULE_DESCRIPTION("Xen filesystem"); | ||
21 | MODULE_LICENSE("GPL"); | ||
22 | |||
23 | static int xenfs_fill_super(struct super_block *sb, void *data, int silent) | ||
24 | { | ||
25 | static struct tree_descr xenfs_files[] = { | ||
26 | [2] = {"xenbus", &xenbus_file_ops, S_IRUSR|S_IWUSR}, | ||
27 | {""}, | ||
28 | }; | ||
29 | |||
30 | return simple_fill_super(sb, XENFS_SUPER_MAGIC, xenfs_files); | ||
31 | } | ||
32 | |||
33 | static int xenfs_get_sb(struct file_system_type *fs_type, | ||
34 | int flags, const char *dev_name, | ||
35 | void *data, struct vfsmount *mnt) | ||
36 | { | ||
37 | return get_sb_single(fs_type, flags, data, xenfs_fill_super, mnt); | ||
38 | } | ||
39 | |||
40 | static struct file_system_type xenfs_type = { | ||
41 | .owner = THIS_MODULE, | ||
42 | .name = "xenfs", | ||
43 | .get_sb = xenfs_get_sb, | ||
44 | .kill_sb = kill_litter_super, | ||
45 | }; | ||
46 | |||
47 | static int __init xenfs_init(void) | ||
48 | { | ||
49 | if (xen_pv_domain()) | ||
50 | return register_filesystem(&xenfs_type); | ||
51 | |||
52 | printk(KERN_INFO "XENFS: not registering filesystem on non-xen platform\n"); | ||
53 | return 0; | ||
54 | } | ||
55 | |||
56 | static void __exit xenfs_exit(void) | ||
57 | { | ||
58 | if (xen_pv_domain()) | ||
59 | unregister_filesystem(&xenfs_type); | ||
60 | } | ||
61 | |||
62 | module_init(xenfs_init); | ||
63 | module_exit(xenfs_exit); | ||
64 | |||
diff --git a/drivers/xen/xenfs/xenbus.c b/drivers/xen/xenfs/xenbus.c new file mode 100644 index 000000000000..875a4c59c594 --- /dev/null +++ b/drivers/xen/xenfs/xenbus.c | |||
@@ -0,0 +1,593 @@ | |||
1 | /* | ||
2 | * Driver giving user-space access to the kernel's xenbus connection | ||
3 | * to xenstore. | ||
4 | * | ||
5 | * Copyright (c) 2005, Christian Limpach | ||
6 | * Copyright (c) 2005, Rusty Russell, IBM Corporation | ||
7 | * | ||
8 | * This program is free software; you can redistribute it and/or | ||
9 | * modify it under the terms of the GNU General Public License version 2 | ||
10 | * as published by the Free Software Foundation; or, when distributed | ||
11 | * separately from the Linux kernel or incorporated into other | ||
12 | * software packages, subject to the following license: | ||
13 | * | ||
14 | * Permission is hereby granted, free of charge, to any person obtaining a copy | ||
15 | * of this source file (the "Software"), to deal in the Software without | ||
16 | * restriction, including without limitation the rights to use, copy, modify, | ||
17 | * merge, publish, distribute, sublicense, and/or sell copies of the Software, | ||
18 | * and to permit persons to whom the Software is furnished to do so, subject to | ||
19 | * the following conditions: | ||
20 | * | ||
21 | * The above copyright notice and this permission notice shall be included in | ||
22 | * all copies or substantial portions of the Software. | ||
23 | * | ||
24 | * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR | ||
25 | * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, | ||
26 | * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE | ||
27 | * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER | ||
28 | * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING | ||
29 | * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS | ||
30 | * IN THE SOFTWARE. | ||
31 | * | ||
32 | * Changes: | ||
33 | * 2008-10-07 Alex Zeffertt Replaced /proc/xen/xenbus with xenfs filesystem | ||
34 | * and /proc/xen compatibility mount point. | ||
35 | * Turned xenfs into a loadable module. | ||
36 | */ | ||
37 | |||
38 | #include <linux/kernel.h> | ||
39 | #include <linux/errno.h> | ||
40 | #include <linux/uio.h> | ||
41 | #include <linux/notifier.h> | ||
42 | #include <linux/wait.h> | ||
43 | #include <linux/fs.h> | ||
44 | #include <linux/poll.h> | ||
45 | #include <linux/mutex.h> | ||
46 | #include <linux/spinlock.h> | ||
47 | #include <linux/mount.h> | ||
48 | #include <linux/pagemap.h> | ||
49 | #include <linux/uaccess.h> | ||
50 | #include <linux/init.h> | ||
51 | #include <linux/namei.h> | ||
52 | #include <linux/string.h> | ||
53 | |||
54 | #include "xenfs.h" | ||
55 | #include "../xenbus/xenbus_comms.h" | ||
56 | |||
57 | #include <xen/xenbus.h> | ||
58 | #include <asm/xen/hypervisor.h> | ||
59 | |||
60 | /* | ||
61 | * An element of a list of outstanding transactions, for which we're | ||
62 | * still waiting a reply. | ||
63 | */ | ||
64 | struct xenbus_transaction_holder { | ||
65 | struct list_head list; | ||
66 | struct xenbus_transaction handle; | ||
67 | }; | ||
68 | |||
69 | /* | ||
70 | * A buffer of data on the queue. | ||
71 | */ | ||
72 | struct read_buffer { | ||
73 | struct list_head list; | ||
74 | unsigned int cons; | ||
75 | unsigned int len; | ||
76 | char msg[]; | ||
77 | }; | ||
78 | |||
79 | struct xenbus_file_priv { | ||
80 | /* | ||
81 | * msgbuffer_mutex is held while partial requests are built up | ||
82 | * and complete requests are acted on. It therefore protects | ||
83 | * the "transactions" and "watches" lists, and the partial | ||
84 | * request length and buffer. | ||
85 | * | ||
86 | * reply_mutex protects the reply being built up to return to | ||
87 | * usermode. It nests inside msgbuffer_mutex but may be held | ||
88 | * alone during a watch callback. | ||
89 | */ | ||
90 | struct mutex msgbuffer_mutex; | ||
91 | |||
92 | /* In-progress transactions */ | ||
93 | struct list_head transactions; | ||
94 | |||
95 | /* Active watches. */ | ||
96 | struct list_head watches; | ||
97 | |||
98 | /* Partial request. */ | ||
99 | unsigned int len; | ||
100 | union { | ||
101 | struct xsd_sockmsg msg; | ||
102 | char buffer[PAGE_SIZE]; | ||
103 | } u; | ||
104 | |||
105 | /* Response queue. */ | ||
106 | struct mutex reply_mutex; | ||
107 | struct list_head read_buffers; | ||
108 | wait_queue_head_t read_waitq; | ||
109 | |||
110 | }; | ||
111 | |||
112 | /* Read out any raw xenbus messages queued up. */ | ||
113 | static ssize_t xenbus_file_read(struct file *filp, | ||
114 | char __user *ubuf, | ||
115 | size_t len, loff_t *ppos) | ||
116 | { | ||
117 | struct xenbus_file_priv *u = filp->private_data; | ||
118 | struct read_buffer *rb; | ||
119 | unsigned i; | ||
120 | int ret; | ||
121 | |||
122 | mutex_lock(&u->reply_mutex); | ||
123 | while (list_empty(&u->read_buffers)) { | ||
124 | mutex_unlock(&u->reply_mutex); | ||
125 | ret = wait_event_interruptible(u->read_waitq, | ||
126 | !list_empty(&u->read_buffers)); | ||
127 | if (ret) | ||
128 | return ret; | ||
129 | mutex_lock(&u->reply_mutex); | ||
130 | } | ||
131 | |||
132 | rb = list_entry(u->read_buffers.next, struct read_buffer, list); | ||
133 | i = 0; | ||
134 | while (i < len) { | ||
135 | unsigned sz = min((unsigned)len - i, rb->len - rb->cons); | ||
136 | |||
137 | ret = copy_to_user(ubuf + i, &rb->msg[rb->cons], sz); | ||
138 | |||
139 | i += sz - ret; | ||
140 | rb->cons += sz - ret; | ||
141 | |||
142 | if (ret != sz) { | ||
143 | if (i == 0) | ||
144 | i = -EFAULT; | ||
145 | goto out; | ||
146 | } | ||
147 | |||
148 | /* Clear out buffer if it has been consumed */ | ||
149 | if (rb->cons == rb->len) { | ||
150 | list_del(&rb->list); | ||
151 | kfree(rb); | ||
152 | if (list_empty(&u->read_buffers)) | ||
153 | break; | ||
154 | rb = list_entry(u->read_buffers.next, | ||
155 | struct read_buffer, list); | ||
156 | } | ||
157 | } | ||
158 | |||
159 | out: | ||
160 | mutex_unlock(&u->reply_mutex); | ||
161 | return i; | ||
162 | } | ||
163 | |||
164 | /* | ||
165 | * Add a buffer to the queue. Caller must hold the appropriate lock | ||
166 | * if the queue is not local. (Commonly the caller will build up | ||
167 | * multiple queued buffers on a temporary local list, and then add it | ||
168 | * to the appropriate list under lock once all the buffers have een | ||
169 | * successfully allocated.) | ||
170 | */ | ||
171 | static int queue_reply(struct list_head *queue, const void *data, size_t len) | ||
172 | { | ||
173 | struct read_buffer *rb; | ||
174 | |||
175 | if (len == 0) | ||
176 | return 0; | ||
177 | |||
178 | rb = kmalloc(sizeof(*rb) + len, GFP_KERNEL); | ||
179 | if (rb == NULL) | ||
180 | return -ENOMEM; | ||
181 | |||
182 | rb->cons = 0; | ||
183 | rb->len = len; | ||
184 | |||
185 | memcpy(rb->msg, data, len); | ||
186 | |||
187 | list_add_tail(&rb->list, queue); | ||
188 | return 0; | ||
189 | } | ||
190 | |||
191 | /* | ||
192 | * Free all the read_buffer s on a list. | ||
193 | * Caller must have sole reference to list. | ||
194 | */ | ||
195 | static void queue_cleanup(struct list_head *list) | ||
196 | { | ||
197 | struct read_buffer *rb; | ||
198 | |||
199 | while (!list_empty(list)) { | ||
200 | rb = list_entry(list->next, struct read_buffer, list); | ||
201 | list_del(list->next); | ||
202 | kfree(rb); | ||
203 | } | ||
204 | } | ||
205 | |||
206 | struct watch_adapter { | ||
207 | struct list_head list; | ||
208 | struct xenbus_watch watch; | ||
209 | struct xenbus_file_priv *dev_data; | ||
210 | char *token; | ||
211 | }; | ||
212 | |||
213 | static void free_watch_adapter(struct watch_adapter *watch) | ||
214 | { | ||
215 | kfree(watch->watch.node); | ||
216 | kfree(watch->token); | ||
217 | kfree(watch); | ||
218 | } | ||
219 | |||
220 | static struct watch_adapter *alloc_watch_adapter(const char *path, | ||
221 | const char *token) | ||
222 | { | ||
223 | struct watch_adapter *watch; | ||
224 | |||
225 | watch = kzalloc(sizeof(*watch), GFP_KERNEL); | ||
226 | if (watch == NULL) | ||
227 | goto out_fail; | ||
228 | |||
229 | watch->watch.node = kstrdup(path, GFP_KERNEL); | ||
230 | if (watch->watch.node == NULL) | ||
231 | goto out_free; | ||
232 | |||
233 | watch->token = kstrdup(token, GFP_KERNEL); | ||
234 | if (watch->token == NULL) | ||
235 | goto out_free; | ||
236 | |||
237 | return watch; | ||
238 | |||
239 | out_free: | ||
240 | free_watch_adapter(watch); | ||
241 | |||
242 | out_fail: | ||
243 | return NULL; | ||
244 | } | ||
245 | |||
246 | static void watch_fired(struct xenbus_watch *watch, | ||
247 | const char **vec, | ||
248 | unsigned int len) | ||
249 | { | ||
250 | struct watch_adapter *adap; | ||
251 | struct xsd_sockmsg hdr; | ||
252 | const char *path, *token; | ||
253 | int path_len, tok_len, body_len, data_len = 0; | ||
254 | int ret; | ||
255 | LIST_HEAD(staging_q); | ||
256 | |||
257 | adap = container_of(watch, struct watch_adapter, watch); | ||
258 | |||
259 | path = vec[XS_WATCH_PATH]; | ||
260 | token = adap->token; | ||
261 | |||
262 | path_len = strlen(path) + 1; | ||
263 | tok_len = strlen(token) + 1; | ||
264 | if (len > 2) | ||
265 | data_len = vec[len] - vec[2] + 1; | ||
266 | body_len = path_len + tok_len + data_len; | ||
267 | |||
268 | hdr.type = XS_WATCH_EVENT; | ||
269 | hdr.len = body_len; | ||
270 | |||
271 | mutex_lock(&adap->dev_data->reply_mutex); | ||
272 | |||
273 | ret = queue_reply(&staging_q, &hdr, sizeof(hdr)); | ||
274 | if (!ret) | ||
275 | ret = queue_reply(&staging_q, path, path_len); | ||
276 | if (!ret) | ||
277 | ret = queue_reply(&staging_q, token, tok_len); | ||
278 | if (!ret && len > 2) | ||
279 | ret = queue_reply(&staging_q, vec[2], data_len); | ||
280 | |||
281 | if (!ret) { | ||
282 | /* success: pass reply list onto watcher */ | ||
283 | list_splice_tail(&staging_q, &adap->dev_data->read_buffers); | ||
284 | wake_up(&adap->dev_data->read_waitq); | ||
285 | } else | ||
286 | queue_cleanup(&staging_q); | ||
287 | |||
288 | mutex_unlock(&adap->dev_data->reply_mutex); | ||
289 | } | ||
290 | |||
291 | static int xenbus_write_transaction(unsigned msg_type, | ||
292 | struct xenbus_file_priv *u) | ||
293 | { | ||
294 | int rc, ret; | ||
295 | void *reply; | ||
296 | struct xenbus_transaction_holder *trans = NULL; | ||
297 | LIST_HEAD(staging_q); | ||
298 | |||
299 | if (msg_type == XS_TRANSACTION_START) { | ||
300 | trans = kmalloc(sizeof(*trans), GFP_KERNEL); | ||
301 | if (!trans) { | ||
302 | rc = -ENOMEM; | ||
303 | goto out; | ||
304 | } | ||
305 | } | ||
306 | |||
307 | reply = xenbus_dev_request_and_reply(&u->u.msg); | ||
308 | if (IS_ERR(reply)) { | ||
309 | kfree(trans); | ||
310 | rc = PTR_ERR(reply); | ||
311 | goto out; | ||
312 | } | ||
313 | |||
314 | if (msg_type == XS_TRANSACTION_START) { | ||
315 | trans->handle.id = simple_strtoul(reply, NULL, 0); | ||
316 | |||
317 | list_add(&trans->list, &u->transactions); | ||
318 | } else if (msg_type == XS_TRANSACTION_END) { | ||
319 | list_for_each_entry(trans, &u->transactions, list) | ||
320 | if (trans->handle.id == u->u.msg.tx_id) | ||
321 | break; | ||
322 | BUG_ON(&trans->list == &u->transactions); | ||
323 | list_del(&trans->list); | ||
324 | |||
325 | kfree(trans); | ||
326 | } | ||
327 | |||
328 | mutex_lock(&u->reply_mutex); | ||
329 | ret = queue_reply(&staging_q, &u->u.msg, sizeof(u->u.msg)); | ||
330 | if (!ret) | ||
331 | ret = queue_reply(&staging_q, reply, u->u.msg.len); | ||
332 | if (!ret) { | ||
333 | list_splice_tail(&staging_q, &u->read_buffers); | ||
334 | wake_up(&u->read_waitq); | ||
335 | } else { | ||
336 | queue_cleanup(&staging_q); | ||
337 | rc = ret; | ||
338 | } | ||
339 | mutex_unlock(&u->reply_mutex); | ||
340 | |||
341 | kfree(reply); | ||
342 | |||
343 | out: | ||
344 | return rc; | ||
345 | } | ||
346 | |||
347 | static int xenbus_write_watch(unsigned msg_type, struct xenbus_file_priv *u) | ||
348 | { | ||
349 | struct watch_adapter *watch, *tmp_watch; | ||
350 | char *path, *token; | ||
351 | int err, rc; | ||
352 | LIST_HEAD(staging_q); | ||
353 | |||
354 | path = u->u.buffer + sizeof(u->u.msg); | ||
355 | token = memchr(path, 0, u->u.msg.len); | ||
356 | if (token == NULL) { | ||
357 | rc = -EILSEQ; | ||
358 | goto out; | ||
359 | } | ||
360 | token++; | ||
361 | |||
362 | if (msg_type == XS_WATCH) { | ||
363 | watch = alloc_watch_adapter(path, token); | ||
364 | if (watch == NULL) { | ||
365 | rc = -ENOMEM; | ||
366 | goto out; | ||
367 | } | ||
368 | |||
369 | watch->watch.callback = watch_fired; | ||
370 | watch->dev_data = u; | ||
371 | |||
372 | err = register_xenbus_watch(&watch->watch); | ||
373 | if (err) { | ||
374 | free_watch_adapter(watch); | ||
375 | rc = err; | ||
376 | goto out; | ||
377 | } | ||
378 | list_add(&watch->list, &u->watches); | ||
379 | } else { | ||
380 | list_for_each_entry_safe(watch, tmp_watch, &u->watches, list) { | ||
381 | if (!strcmp(watch->token, token) && | ||
382 | !strcmp(watch->watch.node, path)) { | ||
383 | unregister_xenbus_watch(&watch->watch); | ||
384 | list_del(&watch->list); | ||
385 | free_watch_adapter(watch); | ||
386 | break; | ||
387 | } | ||
388 | } | ||
389 | } | ||
390 | |||
391 | /* Success. Synthesize a reply to say all is OK. */ | ||
392 | { | ||
393 | struct { | ||
394 | struct xsd_sockmsg hdr; | ||
395 | char body[3]; | ||
396 | } __packed reply = { | ||
397 | { | ||
398 | .type = msg_type, | ||
399 | .len = sizeof(reply.body) | ||
400 | }, | ||
401 | "OK" | ||
402 | }; | ||
403 | |||
404 | mutex_lock(&u->reply_mutex); | ||
405 | rc = queue_reply(&u->read_buffers, &reply, sizeof(reply)); | ||
406 | mutex_unlock(&u->reply_mutex); | ||
407 | } | ||
408 | |||
409 | out: | ||
410 | return rc; | ||
411 | } | ||
412 | |||
413 | static ssize_t xenbus_file_write(struct file *filp, | ||
414 | const char __user *ubuf, | ||
415 | size_t len, loff_t *ppos) | ||
416 | { | ||
417 | struct xenbus_file_priv *u = filp->private_data; | ||
418 | uint32_t msg_type; | ||
419 | int rc = len; | ||
420 | int ret; | ||
421 | LIST_HEAD(staging_q); | ||
422 | |||
423 | /* | ||
424 | * We're expecting usermode to be writing properly formed | ||
425 | * xenbus messages. If they write an incomplete message we | ||
426 | * buffer it up. Once it is complete, we act on it. | ||
427 | */ | ||
428 | |||
429 | /* | ||
430 | * Make sure concurrent writers can't stomp all over each | ||
431 | * other's messages and make a mess of our partial message | ||
432 | * buffer. We don't make any attemppt to stop multiple | ||
433 | * writers from making a mess of each other's incomplete | ||
434 | * messages; we're just trying to guarantee our own internal | ||
435 | * consistency and make sure that single writes are handled | ||
436 | * atomically. | ||
437 | */ | ||
438 | mutex_lock(&u->msgbuffer_mutex); | ||
439 | |||
440 | /* Get this out of the way early to avoid confusion */ | ||
441 | if (len == 0) | ||
442 | goto out; | ||
443 | |||
444 | /* Can't write a xenbus message larger we can buffer */ | ||
445 | if ((len + u->len) > sizeof(u->u.buffer)) { | ||
446 | /* On error, dump existing buffer */ | ||
447 | u->len = 0; | ||
448 | rc = -EINVAL; | ||
449 | goto out; | ||
450 | } | ||
451 | |||
452 | ret = copy_from_user(u->u.buffer + u->len, ubuf, len); | ||
453 | |||
454 | if (ret == len) { | ||
455 | rc = -EFAULT; | ||
456 | goto out; | ||
457 | } | ||
458 | |||
459 | /* Deal with a partial copy. */ | ||
460 | len -= ret; | ||
461 | rc = len; | ||
462 | |||
463 | u->len += len; | ||
464 | |||
465 | /* Return if we haven't got a full message yet */ | ||
466 | if (u->len < sizeof(u->u.msg)) | ||
467 | goto out; /* not even the header yet */ | ||
468 | |||
469 | /* If we're expecting a message that's larger than we can | ||
470 | possibly send, dump what we have and return an error. */ | ||
471 | if ((sizeof(u->u.msg) + u->u.msg.len) > sizeof(u->u.buffer)) { | ||
472 | rc = -E2BIG; | ||
473 | u->len = 0; | ||
474 | goto out; | ||
475 | } | ||
476 | |||
477 | if (u->len < (sizeof(u->u.msg) + u->u.msg.len)) | ||
478 | goto out; /* incomplete data portion */ | ||
479 | |||
480 | /* | ||
481 | * OK, now we have a complete message. Do something with it. | ||
482 | */ | ||
483 | |||
484 | msg_type = u->u.msg.type; | ||
485 | |||
486 | switch (msg_type) { | ||
487 | case XS_TRANSACTION_START: | ||
488 | case XS_TRANSACTION_END: | ||
489 | case XS_DIRECTORY: | ||
490 | case XS_READ: | ||
491 | case XS_GET_PERMS: | ||
492 | case XS_RELEASE: | ||
493 | case XS_GET_DOMAIN_PATH: | ||
494 | case XS_WRITE: | ||
495 | case XS_MKDIR: | ||
496 | case XS_RM: | ||
497 | case XS_SET_PERMS: | ||
498 | /* Send out a transaction */ | ||
499 | ret = xenbus_write_transaction(msg_type, u); | ||
500 | break; | ||
501 | |||
502 | case XS_WATCH: | ||
503 | case XS_UNWATCH: | ||
504 | /* (Un)Ask for some path to be watched for changes */ | ||
505 | ret = xenbus_write_watch(msg_type, u); | ||
506 | break; | ||
507 | |||
508 | default: | ||
509 | ret = -EINVAL; | ||
510 | break; | ||
511 | } | ||
512 | if (ret != 0) | ||
513 | rc = ret; | ||
514 | |||
515 | /* Buffered message consumed */ | ||
516 | u->len = 0; | ||
517 | |||
518 | out: | ||
519 | mutex_unlock(&u->msgbuffer_mutex); | ||
520 | return rc; | ||
521 | } | ||
522 | |||
523 | static int xenbus_file_open(struct inode *inode, struct file *filp) | ||
524 | { | ||
525 | struct xenbus_file_priv *u; | ||
526 | |||
527 | if (xen_store_evtchn == 0) | ||
528 | return -ENOENT; | ||
529 | |||
530 | nonseekable_open(inode, filp); | ||
531 | |||
532 | u = kzalloc(sizeof(*u), GFP_KERNEL); | ||
533 | if (u == NULL) | ||
534 | return -ENOMEM; | ||
535 | |||
536 | INIT_LIST_HEAD(&u->transactions); | ||
537 | INIT_LIST_HEAD(&u->watches); | ||
538 | INIT_LIST_HEAD(&u->read_buffers); | ||
539 | init_waitqueue_head(&u->read_waitq); | ||
540 | |||
541 | mutex_init(&u->reply_mutex); | ||
542 | mutex_init(&u->msgbuffer_mutex); | ||
543 | |||
544 | filp->private_data = u; | ||
545 | |||
546 | return 0; | ||
547 | } | ||
548 | |||
549 | static int xenbus_file_release(struct inode *inode, struct file *filp) | ||
550 | { | ||
551 | struct xenbus_file_priv *u = filp->private_data; | ||
552 | struct xenbus_transaction_holder *trans, *tmp; | ||
553 | struct watch_adapter *watch, *tmp_watch; | ||
554 | |||
555 | /* | ||
556 | * No need for locking here because there are no other users, | ||
557 | * by definition. | ||
558 | */ | ||
559 | |||
560 | list_for_each_entry_safe(trans, tmp, &u->transactions, list) { | ||
561 | xenbus_transaction_end(trans->handle, 1); | ||
562 | list_del(&trans->list); | ||
563 | kfree(trans); | ||
564 | } | ||
565 | |||
566 | list_for_each_entry_safe(watch, tmp_watch, &u->watches, list) { | ||
567 | unregister_xenbus_watch(&watch->watch); | ||
568 | list_del(&watch->list); | ||
569 | free_watch_adapter(watch); | ||
570 | } | ||
571 | |||
572 | kfree(u); | ||
573 | |||
574 | return 0; | ||
575 | } | ||
576 | |||
577 | static unsigned int xenbus_file_poll(struct file *file, poll_table *wait) | ||
578 | { | ||
579 | struct xenbus_file_priv *u = file->private_data; | ||
580 | |||
581 | poll_wait(file, &u->read_waitq, wait); | ||
582 | if (!list_empty(&u->read_buffers)) | ||
583 | return POLLIN | POLLRDNORM; | ||
584 | return 0; | ||
585 | } | ||
586 | |||
587 | const struct file_operations xenbus_file_ops = { | ||
588 | .read = xenbus_file_read, | ||
589 | .write = xenbus_file_write, | ||
590 | .open = xenbus_file_open, | ||
591 | .release = xenbus_file_release, | ||
592 | .poll = xenbus_file_poll, | ||
593 | }; | ||
diff --git a/drivers/xen/xenfs/xenfs.h b/drivers/xen/xenfs/xenfs.h new file mode 100644 index 000000000000..51f08b2d0bf1 --- /dev/null +++ b/drivers/xen/xenfs/xenfs.h | |||
@@ -0,0 +1,6 @@ | |||
1 | #ifndef _XENFS_XENBUS_H | ||
2 | #define _XENFS_XENBUS_H | ||
3 | |||
4 | extern const struct file_operations xenbus_file_ops; | ||
5 | |||
6 | #endif /* _XENFS_XENBUS_H */ | ||
diff --git a/fs/binfmt_elf.c b/fs/binfmt_elf.c index c41fa2af7677..e3ff2b9e602f 100644 --- a/fs/binfmt_elf.c +++ b/fs/binfmt_elf.c | |||
@@ -152,8 +152,10 @@ create_elf_tables(struct linux_binprm *bprm, struct elfhdr *exec, | |||
152 | elf_addr_t __user *sp; | 152 | elf_addr_t __user *sp; |
153 | elf_addr_t __user *u_platform; | 153 | elf_addr_t __user *u_platform; |
154 | elf_addr_t __user *u_base_platform; | 154 | elf_addr_t __user *u_base_platform; |
155 | elf_addr_t __user *u_rand_bytes; | ||
155 | const char *k_platform = ELF_PLATFORM; | 156 | const char *k_platform = ELF_PLATFORM; |
156 | const char *k_base_platform = ELF_BASE_PLATFORM; | 157 | const char *k_base_platform = ELF_BASE_PLATFORM; |
158 | unsigned char k_rand_bytes[16]; | ||
157 | int items; | 159 | int items; |
158 | elf_addr_t *elf_info; | 160 | elf_addr_t *elf_info; |
159 | int ei_index = 0; | 161 | int ei_index = 0; |
@@ -196,6 +198,15 @@ create_elf_tables(struct linux_binprm *bprm, struct elfhdr *exec, | |||
196 | return -EFAULT; | 198 | return -EFAULT; |
197 | } | 199 | } |
198 | 200 | ||
201 | /* | ||
202 | * Generate 16 random bytes for userspace PRNG seeding. | ||
203 | */ | ||
204 | get_random_bytes(k_rand_bytes, sizeof(k_rand_bytes)); | ||
205 | u_rand_bytes = (elf_addr_t __user *) | ||
206 | STACK_ALLOC(p, sizeof(k_rand_bytes)); | ||
207 | if (__copy_to_user(u_rand_bytes, k_rand_bytes, sizeof(k_rand_bytes))) | ||
208 | return -EFAULT; | ||
209 | |||
199 | /* Create the ELF interpreter info */ | 210 | /* Create the ELF interpreter info */ |
200 | elf_info = (elf_addr_t *)current->mm->saved_auxv; | 211 | elf_info = (elf_addr_t *)current->mm->saved_auxv; |
201 | /* update AT_VECTOR_SIZE_BASE if the number of NEW_AUX_ENT() changes */ | 212 | /* update AT_VECTOR_SIZE_BASE if the number of NEW_AUX_ENT() changes */ |
@@ -228,6 +239,7 @@ create_elf_tables(struct linux_binprm *bprm, struct elfhdr *exec, | |||
228 | NEW_AUX_ENT(AT_GID, cred->gid); | 239 | NEW_AUX_ENT(AT_GID, cred->gid); |
229 | NEW_AUX_ENT(AT_EGID, cred->egid); | 240 | NEW_AUX_ENT(AT_EGID, cred->egid); |
230 | NEW_AUX_ENT(AT_SECURE, security_bprm_secureexec(bprm)); | 241 | NEW_AUX_ENT(AT_SECURE, security_bprm_secureexec(bprm)); |
242 | NEW_AUX_ENT(AT_RANDOM, (elf_addr_t)(unsigned long)u_rand_bytes); | ||
231 | NEW_AUX_ENT(AT_EXECFN, bprm->exec); | 243 | NEW_AUX_ENT(AT_EXECFN, bprm->exec); |
232 | if (k_platform) { | 244 | if (k_platform) { |
233 | NEW_AUX_ENT(AT_PLATFORM, | 245 | NEW_AUX_ENT(AT_PLATFORM, |
diff --git a/fs/block_dev.c b/fs/block_dev.c index b957717e25ab..8ebbfdf708c2 100644 --- a/fs/block_dev.c +++ b/fs/block_dev.c | |||
@@ -1005,6 +1005,7 @@ static int __blkdev_get(struct block_device *bdev, fmode_t mode, int for_part) | |||
1005 | } | 1005 | } |
1006 | 1006 | ||
1007 | lock_kernel(); | 1007 | lock_kernel(); |
1008 | restart: | ||
1008 | 1009 | ||
1009 | ret = -ENXIO; | 1010 | ret = -ENXIO; |
1010 | disk = get_gendisk(bdev->bd_dev, &partno); | 1011 | disk = get_gendisk(bdev->bd_dev, &partno); |
@@ -1025,6 +1026,19 @@ static int __blkdev_get(struct block_device *bdev, fmode_t mode, int for_part) | |||
1025 | 1026 | ||
1026 | if (disk->fops->open) { | 1027 | if (disk->fops->open) { |
1027 | ret = disk->fops->open(bdev, mode); | 1028 | ret = disk->fops->open(bdev, mode); |
1029 | if (ret == -ERESTARTSYS) { | ||
1030 | /* Lost a race with 'disk' being | ||
1031 | * deleted, try again. | ||
1032 | * See md.c | ||
1033 | */ | ||
1034 | disk_put_part(bdev->bd_part); | ||
1035 | bdev->bd_part = NULL; | ||
1036 | module_put(disk->fops->owner); | ||
1037 | put_disk(disk); | ||
1038 | bdev->bd_disk = NULL; | ||
1039 | mutex_unlock(&bdev->bd_mutex); | ||
1040 | goto restart; | ||
1041 | } | ||
1028 | if (ret) | 1042 | if (ret) |
1029 | goto out_clear; | 1043 | goto out_clear; |
1030 | } | 1044 | } |
diff --git a/fs/coda/sysctl.c b/fs/coda/sysctl.c index 81b7771c6465..43c96ce29614 100644 --- a/fs/coda/sysctl.c +++ b/fs/coda/sysctl.c | |||
@@ -11,7 +11,9 @@ | |||
11 | 11 | ||
12 | #include "coda_int.h" | 12 | #include "coda_int.h" |
13 | 13 | ||
14 | #ifdef CONFIG_SYSCTL | ||
14 | static struct ctl_table_header *fs_table_header; | 15 | static struct ctl_table_header *fs_table_header; |
16 | #endif | ||
15 | 17 | ||
16 | static ctl_table coda_table[] = { | 18 | static ctl_table coda_table[] = { |
17 | { | 19 | { |
@@ -41,6 +43,7 @@ static ctl_table coda_table[] = { | |||
41 | {} | 43 | {} |
42 | }; | 44 | }; |
43 | 45 | ||
46 | #ifdef CONFIG_SYSCTL | ||
44 | static ctl_table fs_table[] = { | 47 | static ctl_table fs_table[] = { |
45 | { | 48 | { |
46 | .ctl_name = CTL_UNNUMBERED, | 49 | .ctl_name = CTL_UNNUMBERED, |
@@ -50,7 +53,7 @@ static ctl_table fs_table[] = { | |||
50 | }, | 53 | }, |
51 | {} | 54 | {} |
52 | }; | 55 | }; |
53 | 56 | #endif | |
54 | 57 | ||
55 | void coda_sysctl_init(void) | 58 | void coda_sysctl_init(void) |
56 | { | 59 | { |
diff --git a/fs/dcache.c b/fs/dcache.c index e88c23b85a32..4547f66884a0 100644 --- a/fs/dcache.c +++ b/fs/dcache.c | |||
@@ -1567,10 +1567,6 @@ void d_rehash(struct dentry * entry) | |||
1567 | spin_unlock(&dcache_lock); | 1567 | spin_unlock(&dcache_lock); |
1568 | } | 1568 | } |
1569 | 1569 | ||
1570 | #define do_switch(x,y) do { \ | ||
1571 | __typeof__ (x) __tmp = x; \ | ||
1572 | x = y; y = __tmp; } while (0) | ||
1573 | |||
1574 | /* | 1570 | /* |
1575 | * When switching names, the actual string doesn't strictly have to | 1571 | * When switching names, the actual string doesn't strictly have to |
1576 | * be preserved in the target - because we're dropping the target | 1572 | * be preserved in the target - because we're dropping the target |
@@ -1589,7 +1585,7 @@ static void switch_names(struct dentry *dentry, struct dentry *target) | |||
1589 | /* | 1585 | /* |
1590 | * Both external: swap the pointers | 1586 | * Both external: swap the pointers |
1591 | */ | 1587 | */ |
1592 | do_switch(target->d_name.name, dentry->d_name.name); | 1588 | swap(target->d_name.name, dentry->d_name.name); |
1593 | } else { | 1589 | } else { |
1594 | /* | 1590 | /* |
1595 | * dentry:internal, target:external. Steal target's | 1591 | * dentry:internal, target:external. Steal target's |
@@ -1620,7 +1616,7 @@ static void switch_names(struct dentry *dentry, struct dentry *target) | |||
1620 | return; | 1616 | return; |
1621 | } | 1617 | } |
1622 | } | 1618 | } |
1623 | do_switch(dentry->d_name.len, target->d_name.len); | 1619 | swap(dentry->d_name.len, target->d_name.len); |
1624 | } | 1620 | } |
1625 | 1621 | ||
1626 | /* | 1622 | /* |
@@ -1680,7 +1676,7 @@ already_unhashed: | |||
1680 | 1676 | ||
1681 | /* Switch the names.. */ | 1677 | /* Switch the names.. */ |
1682 | switch_names(dentry, target); | 1678 | switch_names(dentry, target); |
1683 | do_switch(dentry->d_name.hash, target->d_name.hash); | 1679 | swap(dentry->d_name.hash, target->d_name.hash); |
1684 | 1680 | ||
1685 | /* ... and switch the parents */ | 1681 | /* ... and switch the parents */ |
1686 | if (IS_ROOT(dentry)) { | 1682 | if (IS_ROOT(dentry)) { |
@@ -1688,7 +1684,7 @@ already_unhashed: | |||
1688 | target->d_parent = target; | 1684 | target->d_parent = target; |
1689 | INIT_LIST_HEAD(&target->d_u.d_child); | 1685 | INIT_LIST_HEAD(&target->d_u.d_child); |
1690 | } else { | 1686 | } else { |
1691 | do_switch(dentry->d_parent, target->d_parent); | 1687 | swap(dentry->d_parent, target->d_parent); |
1692 | 1688 | ||
1693 | /* And add them back to the (new) parent lists */ | 1689 | /* And add them back to the (new) parent lists */ |
1694 | list_add(&target->d_u.d_child, &target->d_parent->d_subdirs); | 1690 | list_add(&target->d_u.d_child, &target->d_parent->d_subdirs); |
@@ -1789,7 +1785,7 @@ static void __d_materialise_dentry(struct dentry *dentry, struct dentry *anon) | |||
1789 | struct dentry *dparent, *aparent; | 1785 | struct dentry *dparent, *aparent; |
1790 | 1786 | ||
1791 | switch_names(dentry, anon); | 1787 | switch_names(dentry, anon); |
1792 | do_switch(dentry->d_name.hash, anon->d_name.hash); | 1788 | swap(dentry->d_name.hash, anon->d_name.hash); |
1793 | 1789 | ||
1794 | dparent = dentry->d_parent; | 1790 | dparent = dentry->d_parent; |
1795 | aparent = anon->d_parent; | 1791 | aparent = anon->d_parent; |
diff --git a/fs/dquot.c b/fs/dquot.c index 61bfff64e5af..48c0571f831d 100644 --- a/fs/dquot.c +++ b/fs/dquot.c | |||
@@ -2090,10 +2090,12 @@ static int do_set_dqblk(struct dquot *dquot, struct if_dqblk *di) | |||
2090 | } | 2090 | } |
2091 | if (di->dqb_valid & QIF_BTIME) { | 2091 | if (di->dqb_valid & QIF_BTIME) { |
2092 | dm->dqb_btime = di->dqb_btime; | 2092 | dm->dqb_btime = di->dqb_btime; |
2093 | check_blim = 1; | ||
2093 | __set_bit(DQ_LASTSET_B + QIF_BTIME_B, &dquot->dq_flags); | 2094 | __set_bit(DQ_LASTSET_B + QIF_BTIME_B, &dquot->dq_flags); |
2094 | } | 2095 | } |
2095 | if (di->dqb_valid & QIF_ITIME) { | 2096 | if (di->dqb_valid & QIF_ITIME) { |
2096 | dm->dqb_itime = di->dqb_itime; | 2097 | dm->dqb_itime = di->dqb_itime; |
2098 | check_ilim = 1; | ||
2097 | __set_bit(DQ_LASTSET_B + QIF_ITIME_B, &dquot->dq_flags); | 2099 | __set_bit(DQ_LASTSET_B + QIF_ITIME_B, &dquot->dq_flags); |
2098 | } | 2100 | } |
2099 | 2101 | ||
diff --git a/fs/ext2/ialloc.c b/fs/ext2/ialloc.c index c454d5db28a5..66321a877e74 100644 --- a/fs/ext2/ialloc.c +++ b/fs/ext2/ialloc.c | |||
@@ -565,12 +565,8 @@ got: | |||
565 | inode->i_blocks = 0; | 565 | inode->i_blocks = 0; |
566 | inode->i_mtime = inode->i_atime = inode->i_ctime = CURRENT_TIME_SEC; | 566 | inode->i_mtime = inode->i_atime = inode->i_ctime = CURRENT_TIME_SEC; |
567 | memset(ei->i_data, 0, sizeof(ei->i_data)); | 567 | memset(ei->i_data, 0, sizeof(ei->i_data)); |
568 | ei->i_flags = EXT2_I(dir)->i_flags & ~EXT2_BTREE_FL; | 568 | ei->i_flags = |
569 | if (S_ISLNK(mode)) | 569 | ext2_mask_flags(mode, EXT2_I(dir)->i_flags & EXT2_FL_INHERITED); |
570 | ei->i_flags &= ~(EXT2_IMMUTABLE_FL|EXT2_APPEND_FL); | ||
571 | /* dirsync is only applied to directories */ | ||
572 | if (!S_ISDIR(mode)) | ||
573 | ei->i_flags &= ~EXT2_DIRSYNC_FL; | ||
574 | ei->i_faddr = 0; | 570 | ei->i_faddr = 0; |
575 | ei->i_frag_no = 0; | 571 | ei->i_frag_no = 0; |
576 | ei->i_frag_size = 0; | 572 | ei->i_frag_size = 0; |
diff --git a/fs/ext2/inode.c b/fs/ext2/inode.c index 02b39a5deb74..23fff2f87783 100644 --- a/fs/ext2/inode.c +++ b/fs/ext2/inode.c | |||
@@ -498,8 +498,6 @@ static int ext2_alloc_branch(struct inode *inode, | |||
498 | * ext2_splice_branch - splice the allocated branch onto inode. | 498 | * ext2_splice_branch - splice the allocated branch onto inode. |
499 | * @inode: owner | 499 | * @inode: owner |
500 | * @block: (logical) number of block we are adding | 500 | * @block: (logical) number of block we are adding |
501 | * @chain: chain of indirect blocks (with a missing link - see | ||
502 | * ext2_alloc_branch) | ||
503 | * @where: location of missing link | 501 | * @where: location of missing link |
504 | * @num: number of indirect blocks we are adding | 502 | * @num: number of indirect blocks we are adding |
505 | * @blks: number of direct blocks we are adding | 503 | * @blks: number of direct blocks we are adding |
diff --git a/fs/ext2/ioctl.c b/fs/ext2/ioctl.c index de876fa793e1..7cb4badef927 100644 --- a/fs/ext2/ioctl.c +++ b/fs/ext2/ioctl.c | |||
@@ -50,8 +50,7 @@ long ext2_ioctl(struct file *filp, unsigned int cmd, unsigned long arg) | |||
50 | goto setflags_out; | 50 | goto setflags_out; |
51 | } | 51 | } |
52 | 52 | ||
53 | if (!S_ISDIR(inode->i_mode)) | 53 | flags = ext2_mask_flags(inode->i_mode, flags); |
54 | flags &= ~EXT2_DIRSYNC_FL; | ||
55 | 54 | ||
56 | mutex_lock(&inode->i_mutex); | 55 | mutex_lock(&inode->i_mutex); |
57 | /* Is it quota file? Do not allow user to mess with it */ | 56 | /* Is it quota file? Do not allow user to mess with it */ |
diff --git a/fs/ext2/super.c b/fs/ext2/super.c index 647cd888ac87..da8bdeaa2e6d 100644 --- a/fs/ext2/super.c +++ b/fs/ext2/super.c | |||
@@ -132,6 +132,7 @@ static void ext2_put_super (struct super_block * sb) | |||
132 | percpu_counter_destroy(&sbi->s_dirs_counter); | 132 | percpu_counter_destroy(&sbi->s_dirs_counter); |
133 | brelse (sbi->s_sbh); | 133 | brelse (sbi->s_sbh); |
134 | sb->s_fs_info = NULL; | 134 | sb->s_fs_info = NULL; |
135 | kfree(sbi->s_blockgroup_lock); | ||
135 | kfree(sbi); | 136 | kfree(sbi); |
136 | 137 | ||
137 | return; | 138 | return; |
@@ -756,6 +757,13 @@ static int ext2_fill_super(struct super_block *sb, void *data, int silent) | |||
756 | sbi = kzalloc(sizeof(*sbi), GFP_KERNEL); | 757 | sbi = kzalloc(sizeof(*sbi), GFP_KERNEL); |
757 | if (!sbi) | 758 | if (!sbi) |
758 | return -ENOMEM; | 759 | return -ENOMEM; |
760 | |||
761 | sbi->s_blockgroup_lock = | ||
762 | kzalloc(sizeof(struct blockgroup_lock), GFP_KERNEL); | ||
763 | if (!sbi->s_blockgroup_lock) { | ||
764 | kfree(sbi); | ||
765 | return -ENOMEM; | ||
766 | } | ||
759 | sb->s_fs_info = sbi; | 767 | sb->s_fs_info = sbi; |
760 | sbi->s_sb_block = sb_block; | 768 | sbi->s_sb_block = sb_block; |
761 | 769 | ||
@@ -983,7 +991,7 @@ static int ext2_fill_super(struct super_block *sb, void *data, int silent) | |||
983 | printk ("EXT2-fs: not enough memory\n"); | 991 | printk ("EXT2-fs: not enough memory\n"); |
984 | goto failed_mount; | 992 | goto failed_mount; |
985 | } | 993 | } |
986 | bgl_lock_init(&sbi->s_blockgroup_lock); | 994 | bgl_lock_init(sbi->s_blockgroup_lock); |
987 | sbi->s_debts = kcalloc(sbi->s_groups_count, sizeof(*sbi->s_debts), GFP_KERNEL); | 995 | sbi->s_debts = kcalloc(sbi->s_groups_count, sizeof(*sbi->s_debts), GFP_KERNEL); |
988 | if (!sbi->s_debts) { | 996 | if (!sbi->s_debts) { |
989 | printk ("EXT2-fs: not enough memory\n"); | 997 | printk ("EXT2-fs: not enough memory\n"); |
diff --git a/fs/ext3/ialloc.c b/fs/ext3/ialloc.c index 5655fbcbd11f..8de6c720e510 100644 --- a/fs/ext3/ialloc.c +++ b/fs/ext3/ialloc.c | |||
@@ -559,12 +559,8 @@ got: | |||
559 | ei->i_dir_start_lookup = 0; | 559 | ei->i_dir_start_lookup = 0; |
560 | ei->i_disksize = 0; | 560 | ei->i_disksize = 0; |
561 | 561 | ||
562 | ei->i_flags = EXT3_I(dir)->i_flags & ~EXT3_INDEX_FL; | 562 | ei->i_flags = |
563 | if (S_ISLNK(mode)) | 563 | ext3_mask_flags(mode, EXT3_I(dir)->i_flags & EXT3_FL_INHERITED); |
564 | ei->i_flags &= ~(EXT3_IMMUTABLE_FL|EXT3_APPEND_FL); | ||
565 | /* dirsync only applies to directories */ | ||
566 | if (!S_ISDIR(mode)) | ||
567 | ei->i_flags &= ~EXT3_DIRSYNC_FL; | ||
568 | #ifdef EXT3_FRAGMENTS | 564 | #ifdef EXT3_FRAGMENTS |
569 | ei->i_faddr = 0; | 565 | ei->i_faddr = 0; |
570 | ei->i_frag_no = 0; | 566 | ei->i_frag_no = 0; |
diff --git a/fs/ext3/ioctl.c b/fs/ext3/ioctl.c index b7394d05ee8e..5e86ce9a86e0 100644 --- a/fs/ext3/ioctl.c +++ b/fs/ext3/ioctl.c | |||
@@ -53,8 +53,7 @@ int ext3_ioctl (struct inode * inode, struct file * filp, unsigned int cmd, | |||
53 | goto flags_out; | 53 | goto flags_out; |
54 | } | 54 | } |
55 | 55 | ||
56 | if (!S_ISDIR(inode->i_mode)) | 56 | flags = ext3_mask_flags(inode->i_mode, flags); |
57 | flags &= ~EXT3_DIRSYNC_FL; | ||
58 | 57 | ||
59 | mutex_lock(&inode->i_mutex); | 58 | mutex_lock(&inode->i_mutex); |
60 | /* Is it quota file? Do not allow user to mess with it */ | 59 | /* Is it quota file? Do not allow user to mess with it */ |
diff --git a/fs/ext3/namei.c b/fs/ext3/namei.c index 1dd2abe6313e..8d6f965e502c 100644 --- a/fs/ext3/namei.c +++ b/fs/ext3/namei.c | |||
@@ -74,10 +74,6 @@ static struct buffer_head *ext3_append(handle_t *handle, | |||
74 | #define assert(test) J_ASSERT(test) | 74 | #define assert(test) J_ASSERT(test) |
75 | #endif | 75 | #endif |
76 | 76 | ||
77 | #ifndef swap | ||
78 | #define swap(x, y) do { typeof(x) z = x; x = y; y = z; } while (0) | ||
79 | #endif | ||
80 | |||
81 | #ifdef DX_DEBUG | 77 | #ifdef DX_DEBUG |
82 | #define dxtrace(command) command | 78 | #define dxtrace(command) command |
83 | #else | 79 | #else |
diff --git a/fs/ext3/super.c b/fs/ext3/super.c index c22d01467bd1..01c235bc2054 100644 --- a/fs/ext3/super.c +++ b/fs/ext3/super.c | |||
@@ -439,6 +439,7 @@ static void ext3_put_super (struct super_block * sb) | |||
439 | ext3_blkdev_remove(sbi); | 439 | ext3_blkdev_remove(sbi); |
440 | } | 440 | } |
441 | sb->s_fs_info = NULL; | 441 | sb->s_fs_info = NULL; |
442 | kfree(sbi->s_blockgroup_lock); | ||
442 | kfree(sbi); | 443 | kfree(sbi); |
443 | return; | 444 | return; |
444 | } | 445 | } |
@@ -1546,6 +1547,13 @@ static int ext3_fill_super (struct super_block *sb, void *data, int silent) | |||
1546 | sbi = kzalloc(sizeof(*sbi), GFP_KERNEL); | 1547 | sbi = kzalloc(sizeof(*sbi), GFP_KERNEL); |
1547 | if (!sbi) | 1548 | if (!sbi) |
1548 | return -ENOMEM; | 1549 | return -ENOMEM; |
1550 | |||
1551 | sbi->s_blockgroup_lock = | ||
1552 | kzalloc(sizeof(struct blockgroup_lock), GFP_KERNEL); | ||
1553 | if (!sbi->s_blockgroup_lock) { | ||
1554 | kfree(sbi); | ||
1555 | return -ENOMEM; | ||
1556 | } | ||
1549 | sb->s_fs_info = sbi; | 1557 | sb->s_fs_info = sbi; |
1550 | sbi->s_mount_opt = 0; | 1558 | sbi->s_mount_opt = 0; |
1551 | sbi->s_resuid = EXT3_DEF_RESUID; | 1559 | sbi->s_resuid = EXT3_DEF_RESUID; |
@@ -1786,7 +1794,7 @@ static int ext3_fill_super (struct super_block *sb, void *data, int silent) | |||
1786 | goto failed_mount; | 1794 | goto failed_mount; |
1787 | } | 1795 | } |
1788 | 1796 | ||
1789 | bgl_lock_init(&sbi->s_blockgroup_lock); | 1797 | bgl_lock_init(sbi->s_blockgroup_lock); |
1790 | 1798 | ||
1791 | for (i = 0; i < db_count; i++) { | 1799 | for (i = 0; i < db_count; i++) { |
1792 | block = descriptor_loc(sb, logic_sb_block, i); | 1800 | block = descriptor_loc(sb, logic_sb_block, i); |
diff --git a/fs/ext4/extents.c b/fs/ext4/extents.c index ea2ce3c0ae66..3f54db31cdc2 100644 --- a/fs/ext4/extents.c +++ b/fs/ext4/extents.c | |||
@@ -2536,7 +2536,7 @@ static int ext4_ext_convert_to_initialized(handle_t *handle, | |||
2536 | */ | 2536 | */ |
2537 | newdepth = ext_depth(inode); | 2537 | newdepth = ext_depth(inode); |
2538 | /* | 2538 | /* |
2539 | * update the extent length after successfull insert of the | 2539 | * update the extent length after successful insert of the |
2540 | * split extent | 2540 | * split extent |
2541 | */ | 2541 | */ |
2542 | orig_ex.ee_len = cpu_to_le16(ee_len - | 2542 | orig_ex.ee_len = cpu_to_le16(ee_len - |
diff --git a/fs/ext4/namei.c b/fs/ext4/namei.c index 9fd2a5e1be4d..4b8d431d7dff 100644 --- a/fs/ext4/namei.c +++ b/fs/ext4/namei.c | |||
@@ -74,10 +74,6 @@ static struct buffer_head *ext4_append(handle_t *handle, | |||
74 | #define assert(test) J_ASSERT(test) | 74 | #define assert(test) J_ASSERT(test) |
75 | #endif | 75 | #endif |
76 | 76 | ||
77 | #ifndef swap | ||
78 | #define swap(x, y) do { typeof(x) z = x; x = y; y = z; } while (0) | ||
79 | #endif | ||
80 | |||
81 | #ifdef DX_DEBUG | 77 | #ifdef DX_DEBUG |
82 | #define dxtrace(command) command | 78 | #define dxtrace(command) command |
83 | #else | 79 | #else |
diff --git a/fs/jbd/commit.c b/fs/jbd/commit.c index 25719d902c51..3fbffb1ea714 100644 --- a/fs/jbd/commit.c +++ b/fs/jbd/commit.c | |||
@@ -306,6 +306,8 @@ void journal_commit_transaction(journal_t *journal) | |||
306 | int flags; | 306 | int flags; |
307 | int err; | 307 | int err; |
308 | unsigned long blocknr; | 308 | unsigned long blocknr; |
309 | ktime_t start_time; | ||
310 | u64 commit_time; | ||
309 | char *tagp = NULL; | 311 | char *tagp = NULL; |
310 | journal_header_t *header; | 312 | journal_header_t *header; |
311 | journal_block_tag_t *tag = NULL; | 313 | journal_block_tag_t *tag = NULL; |
@@ -418,6 +420,7 @@ void journal_commit_transaction(journal_t *journal) | |||
418 | commit_transaction->t_state = T_FLUSH; | 420 | commit_transaction->t_state = T_FLUSH; |
419 | journal->j_committing_transaction = commit_transaction; | 421 | journal->j_committing_transaction = commit_transaction; |
420 | journal->j_running_transaction = NULL; | 422 | journal->j_running_transaction = NULL; |
423 | start_time = ktime_get(); | ||
421 | commit_transaction->t_log_start = journal->j_head; | 424 | commit_transaction->t_log_start = journal->j_head; |
422 | wake_up(&journal->j_wait_transaction_locked); | 425 | wake_up(&journal->j_wait_transaction_locked); |
423 | spin_unlock(&journal->j_state_lock); | 426 | spin_unlock(&journal->j_state_lock); |
@@ -913,6 +916,18 @@ restart_loop: | |||
913 | J_ASSERT(commit_transaction == journal->j_committing_transaction); | 916 | J_ASSERT(commit_transaction == journal->j_committing_transaction); |
914 | journal->j_commit_sequence = commit_transaction->t_tid; | 917 | journal->j_commit_sequence = commit_transaction->t_tid; |
915 | journal->j_committing_transaction = NULL; | 918 | journal->j_committing_transaction = NULL; |
919 | commit_time = ktime_to_ns(ktime_sub(ktime_get(), start_time)); | ||
920 | |||
921 | /* | ||
922 | * weight the commit time higher than the average time so we don't | ||
923 | * react too strongly to vast changes in commit time | ||
924 | */ | ||
925 | if (likely(journal->j_average_commit_time)) | ||
926 | journal->j_average_commit_time = (commit_time*3 + | ||
927 | journal->j_average_commit_time) / 4; | ||
928 | else | ||
929 | journal->j_average_commit_time = commit_time; | ||
930 | |||
916 | spin_unlock(&journal->j_state_lock); | 931 | spin_unlock(&journal->j_state_lock); |
917 | 932 | ||
918 | if (commit_transaction->t_checkpoint_list == NULL && | 933 | if (commit_transaction->t_checkpoint_list == NULL && |
diff --git a/fs/jbd/transaction.c b/fs/jbd/transaction.c index 60d4c32c8808..e6a117431277 100644 --- a/fs/jbd/transaction.c +++ b/fs/jbd/transaction.c | |||
@@ -25,6 +25,7 @@ | |||
25 | #include <linux/timer.h> | 25 | #include <linux/timer.h> |
26 | #include <linux/mm.h> | 26 | #include <linux/mm.h> |
27 | #include <linux/highmem.h> | 27 | #include <linux/highmem.h> |
28 | #include <linux/hrtimer.h> | ||
28 | 29 | ||
29 | static void __journal_temp_unlink_buffer(struct journal_head *jh); | 30 | static void __journal_temp_unlink_buffer(struct journal_head *jh); |
30 | 31 | ||
@@ -49,6 +50,7 @@ get_transaction(journal_t *journal, transaction_t *transaction) | |||
49 | { | 50 | { |
50 | transaction->t_journal = journal; | 51 | transaction->t_journal = journal; |
51 | transaction->t_state = T_RUNNING; | 52 | transaction->t_state = T_RUNNING; |
53 | transaction->t_start_time = ktime_get(); | ||
52 | transaction->t_tid = journal->j_transaction_sequence++; | 54 | transaction->t_tid = journal->j_transaction_sequence++; |
53 | transaction->t_expires = jiffies + journal->j_commit_interval; | 55 | transaction->t_expires = jiffies + journal->j_commit_interval; |
54 | spin_lock_init(&transaction->t_handle_lock); | 56 | spin_lock_init(&transaction->t_handle_lock); |
@@ -752,7 +754,6 @@ out: | |||
752 | * int journal_get_write_access() - notify intent to modify a buffer for metadata (not data) update. | 754 | * int journal_get_write_access() - notify intent to modify a buffer for metadata (not data) update. |
753 | * @handle: transaction to add buffer modifications to | 755 | * @handle: transaction to add buffer modifications to |
754 | * @bh: bh to be used for metadata writes | 756 | * @bh: bh to be used for metadata writes |
755 | * @credits: variable that will receive credits for the buffer | ||
756 | * | 757 | * |
757 | * Returns an error code or 0 on success. | 758 | * Returns an error code or 0 on success. |
758 | * | 759 | * |
@@ -1370,7 +1371,7 @@ int journal_stop(handle_t *handle) | |||
1370 | { | 1371 | { |
1371 | transaction_t *transaction = handle->h_transaction; | 1372 | transaction_t *transaction = handle->h_transaction; |
1372 | journal_t *journal = transaction->t_journal; | 1373 | journal_t *journal = transaction->t_journal; |
1373 | int old_handle_count, err; | 1374 | int err; |
1374 | pid_t pid; | 1375 | pid_t pid; |
1375 | 1376 | ||
1376 | J_ASSERT(journal_current_handle() == handle); | 1377 | J_ASSERT(journal_current_handle() == handle); |
@@ -1399,6 +1400,17 @@ int journal_stop(handle_t *handle) | |||
1399 | * on IO anyway. Speeds up many-threaded, many-dir operations | 1400 | * on IO anyway. Speeds up many-threaded, many-dir operations |
1400 | * by 30x or more... | 1401 | * by 30x or more... |
1401 | * | 1402 | * |
1403 | * We try and optimize the sleep time against what the underlying disk | ||
1404 | * can do, instead of having a static sleep time. This is usefull for | ||
1405 | * the case where our storage is so fast that it is more optimal to go | ||
1406 | * ahead and force a flush and wait for the transaction to be committed | ||
1407 | * than it is to wait for an arbitrary amount of time for new writers to | ||
1408 | * join the transaction. We acheive this by measuring how long it takes | ||
1409 | * to commit a transaction, and compare it with how long this | ||
1410 | * transaction has been running, and if run time < commit time then we | ||
1411 | * sleep for the delta and commit. This greatly helps super fast disks | ||
1412 | * that would see slowdowns as more threads started doing fsyncs. | ||
1413 | * | ||
1402 | * But don't do this if this process was the most recent one to | 1414 | * But don't do this if this process was the most recent one to |
1403 | * perform a synchronous write. We do this to detect the case where a | 1415 | * perform a synchronous write. We do this to detect the case where a |
1404 | * single process is doing a stream of sync writes. No point in waiting | 1416 | * single process is doing a stream of sync writes. No point in waiting |
@@ -1406,11 +1418,26 @@ int journal_stop(handle_t *handle) | |||
1406 | */ | 1418 | */ |
1407 | pid = current->pid; | 1419 | pid = current->pid; |
1408 | if (handle->h_sync && journal->j_last_sync_writer != pid) { | 1420 | if (handle->h_sync && journal->j_last_sync_writer != pid) { |
1421 | u64 commit_time, trans_time; | ||
1422 | |||
1409 | journal->j_last_sync_writer = pid; | 1423 | journal->j_last_sync_writer = pid; |
1410 | do { | 1424 | |
1411 | old_handle_count = transaction->t_handle_count; | 1425 | spin_lock(&journal->j_state_lock); |
1412 | schedule_timeout_uninterruptible(1); | 1426 | commit_time = journal->j_average_commit_time; |
1413 | } while (old_handle_count != transaction->t_handle_count); | 1427 | spin_unlock(&journal->j_state_lock); |
1428 | |||
1429 | trans_time = ktime_to_ns(ktime_sub(ktime_get(), | ||
1430 | transaction->t_start_time)); | ||
1431 | |||
1432 | commit_time = min_t(u64, commit_time, | ||
1433 | 1000*jiffies_to_usecs(1)); | ||
1434 | |||
1435 | if (trans_time < commit_time) { | ||
1436 | ktime_t expires = ktime_add_ns(ktime_get(), | ||
1437 | commit_time); | ||
1438 | set_current_state(TASK_UNINTERRUPTIBLE); | ||
1439 | schedule_hrtimeout(&expires, HRTIMER_MODE_ABS); | ||
1440 | } | ||
1414 | } | 1441 | } |
1415 | 1442 | ||
1416 | current->journal_info = NULL; | 1443 | current->journal_info = NULL; |
diff --git a/fs/ocfs2/alloc.c b/fs/ocfs2/alloc.c index 54ff4c77aaa3..d861096c9d81 100644 --- a/fs/ocfs2/alloc.c +++ b/fs/ocfs2/alloc.c | |||
@@ -3868,7 +3868,7 @@ static void ocfs2_split_record(struct inode *inode, | |||
3868 | struct ocfs2_extent_list *left_el = NULL, *right_el, *insert_el, *el; | 3868 | struct ocfs2_extent_list *left_el = NULL, *right_el, *insert_el, *el; |
3869 | struct ocfs2_extent_rec *rec, *tmprec; | 3869 | struct ocfs2_extent_rec *rec, *tmprec; |
3870 | 3870 | ||
3871 | right_el = path_leaf_el(right_path);; | 3871 | right_el = path_leaf_el(right_path); |
3872 | if (left_path) | 3872 | if (left_path) |
3873 | left_el = path_leaf_el(left_path); | 3873 | left_el = path_leaf_el(left_path); |
3874 | 3874 | ||
diff --git a/fs/ocfs2/dlmglue.c b/fs/ocfs2/dlmglue.c index f731ab491795..b0c4cadd4c45 100644 --- a/fs/ocfs2/dlmglue.c +++ b/fs/ocfs2/dlmglue.c | |||
@@ -1324,7 +1324,7 @@ again: | |||
1324 | goto out; | 1324 | goto out; |
1325 | } | 1325 | } |
1326 | 1326 | ||
1327 | mlog(0, "lock %s, successfull return from ocfs2_dlm_lock\n", | 1327 | mlog(0, "lock %s, successful return from ocfs2_dlm_lock\n", |
1328 | lockres->l_name); | 1328 | lockres->l_name); |
1329 | 1329 | ||
1330 | /* At this point we've gone inside the dlm and need to | 1330 | /* At this point we've gone inside the dlm and need to |
@@ -2951,7 +2951,7 @@ static int ocfs2_drop_lock(struct ocfs2_super *osb, | |||
2951 | ocfs2_dlm_dump_lksb(&lockres->l_lksb); | 2951 | ocfs2_dlm_dump_lksb(&lockres->l_lksb); |
2952 | BUG(); | 2952 | BUG(); |
2953 | } | 2953 | } |
2954 | mlog(0, "lock %s, successfull return from ocfs2_dlm_unlock\n", | 2954 | mlog(0, "lock %s, successful return from ocfs2_dlm_unlock\n", |
2955 | lockres->l_name); | 2955 | lockres->l_name); |
2956 | 2956 | ||
2957 | ocfs2_wait_on_busy_lock(lockres); | 2957 | ocfs2_wait_on_busy_lock(lockres); |
diff --git a/fs/ocfs2/file.c b/fs/ocfs2/file.c index e8f795f978aa..a5887df2cd8a 100644 --- a/fs/ocfs2/file.c +++ b/fs/ocfs2/file.c | |||
@@ -1605,7 +1605,7 @@ int ocfs2_change_file_space(struct file *file, unsigned int cmd, | |||
1605 | struct ocfs2_space_resv *sr) | 1605 | struct ocfs2_space_resv *sr) |
1606 | { | 1606 | { |
1607 | struct inode *inode = file->f_path.dentry->d_inode; | 1607 | struct inode *inode = file->f_path.dentry->d_inode; |
1608 | struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);; | 1608 | struct ocfs2_super *osb = OCFS2_SB(inode->i_sb); |
1609 | 1609 | ||
1610 | if ((cmd == OCFS2_IOC_RESVSP || cmd == OCFS2_IOC_RESVSP64) && | 1610 | if ((cmd == OCFS2_IOC_RESVSP || cmd == OCFS2_IOC_RESVSP64) && |
1611 | !ocfs2_writes_unwritten_extents(osb)) | 1611 | !ocfs2_writes_unwritten_extents(osb)) |
diff --git a/fs/proc/vmcore.c b/fs/proc/vmcore.c index 03ec59504906..5edcc3f92ba7 100644 --- a/fs/proc/vmcore.c +++ b/fs/proc/vmcore.c | |||
@@ -47,8 +47,6 @@ static ssize_t read_from_oldmem(char *buf, size_t count, | |||
47 | 47 | ||
48 | offset = (unsigned long)(*ppos % PAGE_SIZE); | 48 | offset = (unsigned long)(*ppos % PAGE_SIZE); |
49 | pfn = (unsigned long)(*ppos / PAGE_SIZE); | 49 | pfn = (unsigned long)(*ppos / PAGE_SIZE); |
50 | if (pfn > saved_max_pfn) | ||
51 | return -EINVAL; | ||
52 | 50 | ||
53 | do { | 51 | do { |
54 | if (count > (PAGE_SIZE - offset)) | 52 | if (count > (PAGE_SIZE - offset)) |
diff --git a/fs/romfs/inode.c b/fs/romfs/inode.c index c97d4c931715..98a232f7196b 100644 --- a/fs/romfs/inode.c +++ b/fs/romfs/inode.c | |||
@@ -490,7 +490,7 @@ static mode_t romfs_modemap[] = | |||
490 | static struct inode * | 490 | static struct inode * |
491 | romfs_iget(struct super_block *sb, unsigned long ino) | 491 | romfs_iget(struct super_block *sb, unsigned long ino) |
492 | { | 492 | { |
493 | int nextfh; | 493 | int nextfh, ret; |
494 | struct romfs_inode ri; | 494 | struct romfs_inode ri; |
495 | struct inode *i; | 495 | struct inode *i; |
496 | 496 | ||
@@ -526,11 +526,11 @@ romfs_iget(struct super_block *sb, unsigned long ino) | |||
526 | i->i_mtime.tv_nsec = i->i_atime.tv_nsec = i->i_ctime.tv_nsec = 0; | 526 | i->i_mtime.tv_nsec = i->i_atime.tv_nsec = i->i_ctime.tv_nsec = 0; |
527 | 527 | ||
528 | /* Precalculate the data offset */ | 528 | /* Precalculate the data offset */ |
529 | ino = romfs_strnlen(i, ino+ROMFH_SIZE, ROMFS_MAXFN); | 529 | ret = romfs_strnlen(i, ino + ROMFH_SIZE, ROMFS_MAXFN); |
530 | if (ino >= 0) | 530 | if (ret >= 0) |
531 | ino = ((ROMFH_SIZE+ino+1+ROMFH_PAD)&ROMFH_MASK); | 531 | ino = (ROMFH_SIZE + ret + 1 + ROMFH_PAD) & ROMFH_MASK; |
532 | else | 532 | else |
533 | ino = 0; | 533 | ino = 0; |
534 | 534 | ||
535 | ROMFS_I(i)->i_metasize = ino; | 535 | ROMFS_I(i)->i_metasize = ino; |
536 | ROMFS_I(i)->i_dataoffset = ino+(i->i_ino&ROMFH_MASK); | 536 | ROMFS_I(i)->i_dataoffset = ino+(i->i_ino&ROMFH_MASK); |
diff --git a/fs/splice.c b/fs/splice.c index 1abab5cee4ba..a54b3e3f10a7 100644 --- a/fs/splice.c +++ b/fs/splice.c | |||
@@ -21,6 +21,7 @@ | |||
21 | #include <linux/file.h> | 21 | #include <linux/file.h> |
22 | #include <linux/pagemap.h> | 22 | #include <linux/pagemap.h> |
23 | #include <linux/splice.h> | 23 | #include <linux/splice.h> |
24 | #include <linux/memcontrol.h> | ||
24 | #include <linux/mm_inline.h> | 25 | #include <linux/mm_inline.h> |
25 | #include <linux/swap.h> | 26 | #include <linux/swap.h> |
26 | #include <linux/writeback.h> | 27 | #include <linux/writeback.h> |
diff --git a/fs/super.c b/fs/super.c index cb20744ec789..7d67387496cb 100644 --- a/fs/super.c +++ b/fs/super.c | |||
@@ -458,7 +458,6 @@ void sync_filesystems(int wait) | |||
458 | if (sb->s_flags & MS_RDONLY) | 458 | if (sb->s_flags & MS_RDONLY) |
459 | continue; | 459 | continue; |
460 | sb->s_need_sync_fs = 1; | 460 | sb->s_need_sync_fs = 1; |
461 | async_synchronize_full_special(&sb->s_async_list); | ||
462 | } | 461 | } |
463 | 462 | ||
464 | restart: | 463 | restart: |
@@ -471,6 +470,7 @@ restart: | |||
471 | sb->s_count++; | 470 | sb->s_count++; |
472 | spin_unlock(&sb_lock); | 471 | spin_unlock(&sb_lock); |
473 | down_read(&sb->s_umount); | 472 | down_read(&sb->s_umount); |
473 | async_synchronize_full_special(&sb->s_async_list); | ||
474 | if (sb->s_root && (wait || sb->s_dirt)) | 474 | if (sb->s_root && (wait || sb->s_dirt)) |
475 | sb->s_op->sync_fs(sb, wait); | 475 | sb->s_op->sync_fs(sb, wait); |
476 | up_read(&sb->s_umount); | 476 | up_read(&sb->s_umount); |
diff --git a/include/linux/auxvec.h b/include/linux/auxvec.h index d7afa9dd6635..f3b5d4e3a2ac 100644 --- a/include/linux/auxvec.h +++ b/include/linux/auxvec.h | |||
@@ -23,16 +23,16 @@ | |||
23 | #define AT_PLATFORM 15 /* string identifying CPU for optimizations */ | 23 | #define AT_PLATFORM 15 /* string identifying CPU for optimizations */ |
24 | #define AT_HWCAP 16 /* arch dependent hints at CPU capabilities */ | 24 | #define AT_HWCAP 16 /* arch dependent hints at CPU capabilities */ |
25 | #define AT_CLKTCK 17 /* frequency at which times() increments */ | 25 | #define AT_CLKTCK 17 /* frequency at which times() increments */ |
26 | 26 | /* AT_* values 18 through 22 are reserved */ | |
27 | #define AT_SECURE 23 /* secure mode boolean */ | 27 | #define AT_SECURE 23 /* secure mode boolean */ |
28 | |||
29 | #define AT_BASE_PLATFORM 24 /* string identifying real platform, may | 28 | #define AT_BASE_PLATFORM 24 /* string identifying real platform, may |
30 | * differ from AT_PLATFORM. */ | 29 | * differ from AT_PLATFORM. */ |
30 | #define AT_RANDOM 25 /* address of 16 random bytes */ | ||
31 | 31 | ||
32 | #define AT_EXECFN 31 /* filename of program */ | 32 | #define AT_EXECFN 31 /* filename of program */ |
33 | 33 | ||
34 | #ifdef __KERNEL__ | 34 | #ifdef __KERNEL__ |
35 | #define AT_VECTOR_SIZE_BASE 18 /* NEW_AUX_ENT entries in auxiliary table */ | 35 | #define AT_VECTOR_SIZE_BASE 19 /* NEW_AUX_ENT entries in auxiliary table */ |
36 | /* number of "#define AT_.*" above, minus {AT_NULL, AT_IGNORE, AT_NOTELF} */ | 36 | /* number of "#define AT_.*" above, minus {AT_NULL, AT_IGNORE, AT_NOTELF} */ |
37 | #endif | 37 | #endif |
38 | 38 | ||
diff --git a/include/linux/cgroup.h b/include/linux/cgroup.h index 08b78c09b09a..e267e62827bb 100644 --- a/include/linux/cgroup.h +++ b/include/linux/cgroup.h | |||
@@ -52,9 +52,9 @@ struct cgroup_subsys_state { | |||
52 | * hierarchy structure */ | 52 | * hierarchy structure */ |
53 | struct cgroup *cgroup; | 53 | struct cgroup *cgroup; |
54 | 54 | ||
55 | /* State maintained by the cgroup system to allow | 55 | /* State maintained by the cgroup system to allow subsystems |
56 | * subsystems to be "busy". Should be accessed via css_get() | 56 | * to be "busy". Should be accessed via css_get(), |
57 | * and css_put() */ | 57 | * css_tryget() and and css_put(). */ |
58 | 58 | ||
59 | atomic_t refcnt; | 59 | atomic_t refcnt; |
60 | 60 | ||
@@ -64,11 +64,14 @@ struct cgroup_subsys_state { | |||
64 | /* bits in struct cgroup_subsys_state flags field */ | 64 | /* bits in struct cgroup_subsys_state flags field */ |
65 | enum { | 65 | enum { |
66 | CSS_ROOT, /* This CSS is the root of the subsystem */ | 66 | CSS_ROOT, /* This CSS is the root of the subsystem */ |
67 | CSS_REMOVED, /* This CSS is dead */ | ||
67 | }; | 68 | }; |
68 | 69 | ||
69 | /* | 70 | /* |
70 | * Call css_get() to hold a reference on the cgroup; | 71 | * Call css_get() to hold a reference on the css; it can be used |
71 | * | 72 | * for a reference obtained via: |
73 | * - an existing ref-counted reference to the css | ||
74 | * - task->cgroups for a locked task | ||
72 | */ | 75 | */ |
73 | 76 | ||
74 | static inline void css_get(struct cgroup_subsys_state *css) | 77 | static inline void css_get(struct cgroup_subsys_state *css) |
@@ -77,9 +80,32 @@ static inline void css_get(struct cgroup_subsys_state *css) | |||
77 | if (!test_bit(CSS_ROOT, &css->flags)) | 80 | if (!test_bit(CSS_ROOT, &css->flags)) |
78 | atomic_inc(&css->refcnt); | 81 | atomic_inc(&css->refcnt); |
79 | } | 82 | } |
83 | |||
84 | static inline bool css_is_removed(struct cgroup_subsys_state *css) | ||
85 | { | ||
86 | return test_bit(CSS_REMOVED, &css->flags); | ||
87 | } | ||
88 | |||
89 | /* | ||
90 | * Call css_tryget() to take a reference on a css if your existing | ||
91 | * (known-valid) reference isn't already ref-counted. Returns false if | ||
92 | * the css has been destroyed. | ||
93 | */ | ||
94 | |||
95 | static inline bool css_tryget(struct cgroup_subsys_state *css) | ||
96 | { | ||
97 | if (test_bit(CSS_ROOT, &css->flags)) | ||
98 | return true; | ||
99 | while (!atomic_inc_not_zero(&css->refcnt)) { | ||
100 | if (test_bit(CSS_REMOVED, &css->flags)) | ||
101 | return false; | ||
102 | } | ||
103 | return true; | ||
104 | } | ||
105 | |||
80 | /* | 106 | /* |
81 | * css_put() should be called to release a reference taken by | 107 | * css_put() should be called to release a reference taken by |
82 | * css_get() | 108 | * css_get() or css_tryget() |
83 | */ | 109 | */ |
84 | 110 | ||
85 | extern void __css_put(struct cgroup_subsys_state *css); | 111 | extern void __css_put(struct cgroup_subsys_state *css); |
@@ -116,7 +142,7 @@ struct cgroup { | |||
116 | struct list_head children; /* my children */ | 142 | struct list_head children; /* my children */ |
117 | 143 | ||
118 | struct cgroup *parent; /* my parent */ | 144 | struct cgroup *parent; /* my parent */ |
119 | struct dentry *dentry; /* cgroup fs entry */ | 145 | struct dentry *dentry; /* cgroup fs entry, RCU protected */ |
120 | 146 | ||
121 | /* Private pointers for each registered subsystem */ | 147 | /* Private pointers for each registered subsystem */ |
122 | struct cgroup_subsys_state *subsys[CGROUP_SUBSYS_COUNT]; | 148 | struct cgroup_subsys_state *subsys[CGROUP_SUBSYS_COUNT]; |
@@ -145,6 +171,9 @@ struct cgroup { | |||
145 | int pids_use_count; | 171 | int pids_use_count; |
146 | /* Length of the current tasks_pids array */ | 172 | /* Length of the current tasks_pids array */ |
147 | int pids_length; | 173 | int pids_length; |
174 | |||
175 | /* For RCU-protected deletion */ | ||
176 | struct rcu_head rcu_head; | ||
148 | }; | 177 | }; |
149 | 178 | ||
150 | /* A css_set is a structure holding pointers to a set of | 179 | /* A css_set is a structure holding pointers to a set of |
@@ -337,9 +366,23 @@ struct cgroup_subsys { | |||
337 | #define MAX_CGROUP_TYPE_NAMELEN 32 | 366 | #define MAX_CGROUP_TYPE_NAMELEN 32 |
338 | const char *name; | 367 | const char *name; |
339 | 368 | ||
340 | /* Protected by RCU */ | 369 | /* |
341 | struct cgroupfs_root *root; | 370 | * Protects sibling/children links of cgroups in this |
371 | * hierarchy, plus protects which hierarchy (or none) the | ||
372 | * subsystem is a part of (i.e. root/sibling). To avoid | ||
373 | * potential deadlocks, the following operations should not be | ||
374 | * undertaken while holding any hierarchy_mutex: | ||
375 | * | ||
376 | * - allocating memory | ||
377 | * - initiating hotplug events | ||
378 | */ | ||
379 | struct mutex hierarchy_mutex; | ||
342 | 380 | ||
381 | /* | ||
382 | * Link to parent, and list entry in parent's children. | ||
383 | * Protected by this->hierarchy_mutex and cgroup_lock() | ||
384 | */ | ||
385 | struct cgroupfs_root *root; | ||
343 | struct list_head sibling; | 386 | struct list_head sibling; |
344 | }; | 387 | }; |
345 | 388 | ||
diff --git a/include/linux/cpuset.h b/include/linux/cpuset.h index 51ea2bdea0f9..90c6074a36ca 100644 --- a/include/linux/cpuset.h +++ b/include/linux/cpuset.h | |||
@@ -20,8 +20,9 @@ extern int number_of_cpusets; /* How many cpusets are defined in system? */ | |||
20 | extern int cpuset_init_early(void); | 20 | extern int cpuset_init_early(void); |
21 | extern int cpuset_init(void); | 21 | extern int cpuset_init(void); |
22 | extern void cpuset_init_smp(void); | 22 | extern void cpuset_init_smp(void); |
23 | extern void cpuset_cpus_allowed(struct task_struct *p, cpumask_t *mask); | 23 | extern void cpuset_cpus_allowed(struct task_struct *p, struct cpumask *mask); |
24 | extern void cpuset_cpus_allowed_locked(struct task_struct *p, cpumask_t *mask); | 24 | extern void cpuset_cpus_allowed_locked(struct task_struct *p, |
25 | struct cpumask *mask); | ||
25 | extern nodemask_t cpuset_mems_allowed(struct task_struct *p); | 26 | extern nodemask_t cpuset_mems_allowed(struct task_struct *p); |
26 | #define cpuset_current_mems_allowed (current->mems_allowed) | 27 | #define cpuset_current_mems_allowed (current->mems_allowed) |
27 | void cpuset_init_current_mems_allowed(void); | 28 | void cpuset_init_current_mems_allowed(void); |
@@ -86,12 +87,13 @@ static inline int cpuset_init_early(void) { return 0; } | |||
86 | static inline int cpuset_init(void) { return 0; } | 87 | static inline int cpuset_init(void) { return 0; } |
87 | static inline void cpuset_init_smp(void) {} | 88 | static inline void cpuset_init_smp(void) {} |
88 | 89 | ||
89 | static inline void cpuset_cpus_allowed(struct task_struct *p, cpumask_t *mask) | 90 | static inline void cpuset_cpus_allowed(struct task_struct *p, |
91 | struct cpumask *mask) | ||
90 | { | 92 | { |
91 | *mask = cpu_possible_map; | 93 | *mask = cpu_possible_map; |
92 | } | 94 | } |
93 | static inline void cpuset_cpus_allowed_locked(struct task_struct *p, | 95 | static inline void cpuset_cpus_allowed_locked(struct task_struct *p, |
94 | cpumask_t *mask) | 96 | struct cpumask *mask) |
95 | { | 97 | { |
96 | *mask = cpu_possible_map; | 98 | *mask = cpu_possible_map; |
97 | } | 99 | } |
diff --git a/include/linux/ext2_fs.h b/include/linux/ext2_fs.h index 78c775a83f7c..121720d74e15 100644 --- a/include/linux/ext2_fs.h +++ b/include/linux/ext2_fs.h | |||
@@ -194,6 +194,30 @@ struct ext2_group_desc | |||
194 | #define EXT2_FL_USER_VISIBLE FS_FL_USER_VISIBLE /* User visible flags */ | 194 | #define EXT2_FL_USER_VISIBLE FS_FL_USER_VISIBLE /* User visible flags */ |
195 | #define EXT2_FL_USER_MODIFIABLE FS_FL_USER_MODIFIABLE /* User modifiable flags */ | 195 | #define EXT2_FL_USER_MODIFIABLE FS_FL_USER_MODIFIABLE /* User modifiable flags */ |
196 | 196 | ||
197 | /* Flags that should be inherited by new inodes from their parent. */ | ||
198 | #define EXT2_FL_INHERITED (EXT2_SECRM_FL | EXT2_UNRM_FL | EXT2_COMPR_FL |\ | ||
199 | EXT2_SYNC_FL | EXT2_IMMUTABLE_FL | EXT2_APPEND_FL |\ | ||
200 | EXT2_NODUMP_FL | EXT2_NOATIME_FL | EXT2_COMPRBLK_FL|\ | ||
201 | EXT2_NOCOMP_FL | EXT2_JOURNAL_DATA_FL |\ | ||
202 | EXT2_NOTAIL_FL | EXT2_DIRSYNC_FL) | ||
203 | |||
204 | /* Flags that are appropriate for regular files (all but dir-specific ones). */ | ||
205 | #define EXT2_REG_FLMASK (~(EXT2_DIRSYNC_FL | EXT2_TOPDIR_FL)) | ||
206 | |||
207 | /* Flags that are appropriate for non-directories/regular files. */ | ||
208 | #define EXT2_OTHER_FLMASK (EXT2_NODUMP_FL | EXT2_NOATIME_FL) | ||
209 | |||
210 | /* Mask out flags that are inappropriate for the given type of inode. */ | ||
211 | static inline __u32 ext2_mask_flags(umode_t mode, __u32 flags) | ||
212 | { | ||
213 | if (S_ISDIR(mode)) | ||
214 | return flags; | ||
215 | else if (S_ISREG(mode)) | ||
216 | return flags & EXT2_REG_FLMASK; | ||
217 | else | ||
218 | return flags & EXT2_OTHER_FLMASK; | ||
219 | } | ||
220 | |||
197 | /* | 221 | /* |
198 | * ioctl commands | 222 | * ioctl commands |
199 | */ | 223 | */ |
diff --git a/include/linux/ext2_fs_sb.h b/include/linux/ext2_fs_sb.h index dc541f3653d1..1cdb66367c98 100644 --- a/include/linux/ext2_fs_sb.h +++ b/include/linux/ext2_fs_sb.h | |||
@@ -101,7 +101,7 @@ struct ext2_sb_info { | |||
101 | struct percpu_counter s_freeblocks_counter; | 101 | struct percpu_counter s_freeblocks_counter; |
102 | struct percpu_counter s_freeinodes_counter; | 102 | struct percpu_counter s_freeinodes_counter; |
103 | struct percpu_counter s_dirs_counter; | 103 | struct percpu_counter s_dirs_counter; |
104 | struct blockgroup_lock s_blockgroup_lock; | 104 | struct blockgroup_lock *s_blockgroup_lock; |
105 | /* root of the per fs reservation window tree */ | 105 | /* root of the per fs reservation window tree */ |
106 | spinlock_t s_rsv_window_lock; | 106 | spinlock_t s_rsv_window_lock; |
107 | struct rb_root s_rsv_window_root; | 107 | struct rb_root s_rsv_window_root; |
@@ -111,7 +111,7 @@ struct ext2_sb_info { | |||
111 | static inline spinlock_t * | 111 | static inline spinlock_t * |
112 | sb_bgl_lock(struct ext2_sb_info *sbi, unsigned int block_group) | 112 | sb_bgl_lock(struct ext2_sb_info *sbi, unsigned int block_group) |
113 | { | 113 | { |
114 | return bgl_lock_ptr(&sbi->s_blockgroup_lock, block_group); | 114 | return bgl_lock_ptr(sbi->s_blockgroup_lock, block_group); |
115 | } | 115 | } |
116 | 116 | ||
117 | #endif /* _LINUX_EXT2_FS_SB */ | 117 | #endif /* _LINUX_EXT2_FS_SB */ |
diff --git a/include/linux/ext3_fs.h b/include/linux/ext3_fs.h index d14f02918483..d76800f6ecf0 100644 --- a/include/linux/ext3_fs.h +++ b/include/linux/ext3_fs.h | |||
@@ -178,6 +178,30 @@ struct ext3_group_desc | |||
178 | #define EXT3_FL_USER_VISIBLE 0x0003DFFF /* User visible flags */ | 178 | #define EXT3_FL_USER_VISIBLE 0x0003DFFF /* User visible flags */ |
179 | #define EXT3_FL_USER_MODIFIABLE 0x000380FF /* User modifiable flags */ | 179 | #define EXT3_FL_USER_MODIFIABLE 0x000380FF /* User modifiable flags */ |
180 | 180 | ||
181 | /* Flags that should be inherited by new inodes from their parent. */ | ||
182 | #define EXT3_FL_INHERITED (EXT3_SECRM_FL | EXT3_UNRM_FL | EXT3_COMPR_FL |\ | ||
183 | EXT3_SYNC_FL | EXT3_IMMUTABLE_FL | EXT3_APPEND_FL |\ | ||
184 | EXT3_NODUMP_FL | EXT3_NOATIME_FL | EXT3_COMPRBLK_FL|\ | ||
185 | EXT3_NOCOMPR_FL | EXT3_JOURNAL_DATA_FL |\ | ||
186 | EXT3_NOTAIL_FL | EXT3_DIRSYNC_FL) | ||
187 | |||
188 | /* Flags that are appropriate for regular files (all but dir-specific ones). */ | ||
189 | #define EXT3_REG_FLMASK (~(EXT3_DIRSYNC_FL | EXT3_TOPDIR_FL)) | ||
190 | |||
191 | /* Flags that are appropriate for non-directories/regular files. */ | ||
192 | #define EXT3_OTHER_FLMASK (EXT3_NODUMP_FL | EXT3_NOATIME_FL) | ||
193 | |||
194 | /* Mask out flags that are inappropriate for the given type of inode. */ | ||
195 | static inline __u32 ext3_mask_flags(umode_t mode, __u32 flags) | ||
196 | { | ||
197 | if (S_ISDIR(mode)) | ||
198 | return flags; | ||
199 | else if (S_ISREG(mode)) | ||
200 | return flags & EXT3_REG_FLMASK; | ||
201 | else | ||
202 | return flags & EXT3_OTHER_FLMASK; | ||
203 | } | ||
204 | |||
181 | /* | 205 | /* |
182 | * Inode dynamic state flags | 206 | * Inode dynamic state flags |
183 | */ | 207 | */ |
diff --git a/include/linux/ext3_fs_sb.h b/include/linux/ext3_fs_sb.h index e024e38248ff..76fdc0f4b028 100644 --- a/include/linux/ext3_fs_sb.h +++ b/include/linux/ext3_fs_sb.h | |||
@@ -60,7 +60,7 @@ struct ext3_sb_info { | |||
60 | struct percpu_counter s_freeblocks_counter; | 60 | struct percpu_counter s_freeblocks_counter; |
61 | struct percpu_counter s_freeinodes_counter; | 61 | struct percpu_counter s_freeinodes_counter; |
62 | struct percpu_counter s_dirs_counter; | 62 | struct percpu_counter s_dirs_counter; |
63 | struct blockgroup_lock s_blockgroup_lock; | 63 | struct blockgroup_lock *s_blockgroup_lock; |
64 | 64 | ||
65 | /* root of the per fs reservation window tree */ | 65 | /* root of the per fs reservation window tree */ |
66 | spinlock_t s_rsv_window_lock; | 66 | spinlock_t s_rsv_window_lock; |
@@ -86,7 +86,7 @@ struct ext3_sb_info { | |||
86 | static inline spinlock_t * | 86 | static inline spinlock_t * |
87 | sb_bgl_lock(struct ext3_sb_info *sbi, unsigned int block_group) | 87 | sb_bgl_lock(struct ext3_sb_info *sbi, unsigned int block_group) |
88 | { | 88 | { |
89 | return bgl_lock_ptr(&sbi->s_blockgroup_lock, block_group); | 89 | return bgl_lock_ptr(sbi->s_blockgroup_lock, block_group); |
90 | } | 90 | } |
91 | 91 | ||
92 | #endif /* _LINUX_EXT3_FS_SB */ | 92 | #endif /* _LINUX_EXT3_FS_SB */ |
diff --git a/include/linux/jbd.h b/include/linux/jbd.h index 346e2b80be7d..6384b19efe64 100644 --- a/include/linux/jbd.h +++ b/include/linux/jbd.h | |||
@@ -543,6 +543,11 @@ struct transaction_s | |||
543 | unsigned long t_expires; | 543 | unsigned long t_expires; |
544 | 544 | ||
545 | /* | 545 | /* |
546 | * When this transaction started, in nanoseconds [no locking] | ||
547 | */ | ||
548 | ktime_t t_start_time; | ||
549 | |||
550 | /* | ||
546 | * How many handles used this transaction? [t_handle_lock] | 551 | * How many handles used this transaction? [t_handle_lock] |
547 | */ | 552 | */ |
548 | int t_handle_count; | 553 | int t_handle_count; |
@@ -798,9 +803,19 @@ struct journal_s | |||
798 | struct buffer_head **j_wbuf; | 803 | struct buffer_head **j_wbuf; |
799 | int j_wbufsize; | 804 | int j_wbufsize; |
800 | 805 | ||
806 | /* | ||
807 | * this is the pid of the last person to run a synchronous operation | ||
808 | * through the journal. | ||
809 | */ | ||
801 | pid_t j_last_sync_writer; | 810 | pid_t j_last_sync_writer; |
802 | 811 | ||
803 | /* | 812 | /* |
813 | * the average amount of time in nanoseconds it takes to commit a | ||
814 | * transaction to the disk. [j_state_lock] | ||
815 | */ | ||
816 | u64 j_average_commit_time; | ||
817 | |||
818 | /* | ||
804 | * An opaque pointer to fs-private information. ext3 puts its | 819 | * An opaque pointer to fs-private information. ext3 puts its |
805 | * superblock pointer here | 820 | * superblock pointer here |
806 | */ | 821 | */ |
diff --git a/include/linux/kernel.h b/include/linux/kernel.h index 6b8e2027165e..343df9ef2412 100644 --- a/include/linux/kernel.h +++ b/include/linux/kernel.h | |||
@@ -476,6 +476,12 @@ static inline char *pack_hex_byte(char *buf, u8 byte) | |||
476 | __val = __val < __min ? __min: __val; \ | 476 | __val = __val < __min ? __min: __val; \ |
477 | __val > __max ? __max: __val; }) | 477 | __val > __max ? __max: __val; }) |
478 | 478 | ||
479 | |||
480 | /* | ||
481 | * swap - swap value of @a and @b | ||
482 | */ | ||
483 | #define swap(a, b) ({ typeof(a) __tmp = (a); (a) = (b); (b) = __tmp; }) | ||
484 | |||
479 | /** | 485 | /** |
480 | * container_of - cast a member of a structure out to the containing structure | 486 | * container_of - cast a member of a structure out to the containing structure |
481 | * @ptr: the pointer to the member. | 487 | * @ptr: the pointer to the member. |
diff --git a/include/linux/libata.h b/include/linux/libata.h index 3449de597eff..4f7c8fb4d3fe 100644 --- a/include/linux/libata.h +++ b/include/linux/libata.h | |||
@@ -1518,6 +1518,7 @@ extern void sata_pmp_error_handler(struct ata_port *ap); | |||
1518 | 1518 | ||
1519 | extern const struct ata_port_operations ata_sff_port_ops; | 1519 | extern const struct ata_port_operations ata_sff_port_ops; |
1520 | extern const struct ata_port_operations ata_bmdma_port_ops; | 1520 | extern const struct ata_port_operations ata_bmdma_port_ops; |
1521 | extern const struct ata_port_operations ata_bmdma32_port_ops; | ||
1521 | 1522 | ||
1522 | /* PIO only, sg_tablesize and dma_boundary limits can be removed */ | 1523 | /* PIO only, sg_tablesize and dma_boundary limits can be removed */ |
1523 | #define ATA_PIO_SHT(drv_name) \ | 1524 | #define ATA_PIO_SHT(drv_name) \ |
@@ -1545,6 +1546,8 @@ extern void ata_sff_exec_command(struct ata_port *ap, | |||
1545 | const struct ata_taskfile *tf); | 1546 | const struct ata_taskfile *tf); |
1546 | extern unsigned int ata_sff_data_xfer(struct ata_device *dev, | 1547 | extern unsigned int ata_sff_data_xfer(struct ata_device *dev, |
1547 | unsigned char *buf, unsigned int buflen, int rw); | 1548 | unsigned char *buf, unsigned int buflen, int rw); |
1549 | extern unsigned int ata_sff_data_xfer32(struct ata_device *dev, | ||
1550 | unsigned char *buf, unsigned int buflen, int rw); | ||
1548 | extern unsigned int ata_sff_data_xfer_noirq(struct ata_device *dev, | 1551 | extern unsigned int ata_sff_data_xfer_noirq(struct ata_device *dev, |
1549 | unsigned char *buf, unsigned int buflen, int rw); | 1552 | unsigned char *buf, unsigned int buflen, int rw); |
1550 | extern u8 ata_sff_irq_on(struct ata_port *ap); | 1553 | extern u8 ata_sff_irq_on(struct ata_port *ap); |
diff --git a/include/linux/magic.h b/include/linux/magic.h index f7f3fdddbef0..439f6f3cb0c4 100644 --- a/include/linux/magic.h +++ b/include/linux/magic.h | |||
@@ -13,6 +13,7 @@ | |||
13 | #define EFS_SUPER_MAGIC 0x414A53 | 13 | #define EFS_SUPER_MAGIC 0x414A53 |
14 | #define EXT2_SUPER_MAGIC 0xEF53 | 14 | #define EXT2_SUPER_MAGIC 0xEF53 |
15 | #define EXT3_SUPER_MAGIC 0xEF53 | 15 | #define EXT3_SUPER_MAGIC 0xEF53 |
16 | #define XENFS_SUPER_MAGIC 0xabba1974 | ||
16 | #define EXT4_SUPER_MAGIC 0xEF53 | 17 | #define EXT4_SUPER_MAGIC 0xEF53 |
17 | #define HPFS_SUPER_MAGIC 0xf995e849 | 18 | #define HPFS_SUPER_MAGIC 0xf995e849 |
18 | #define ISOFS_SUPER_MAGIC 0x9660 | 19 | #define ISOFS_SUPER_MAGIC 0x9660 |
diff --git a/include/linux/memcontrol.h b/include/linux/memcontrol.h index 1fbe14d39521..326f45c86530 100644 --- a/include/linux/memcontrol.h +++ b/include/linux/memcontrol.h | |||
@@ -19,22 +19,45 @@ | |||
19 | 19 | ||
20 | #ifndef _LINUX_MEMCONTROL_H | 20 | #ifndef _LINUX_MEMCONTROL_H |
21 | #define _LINUX_MEMCONTROL_H | 21 | #define _LINUX_MEMCONTROL_H |
22 | 22 | #include <linux/cgroup.h> | |
23 | struct mem_cgroup; | 23 | struct mem_cgroup; |
24 | struct page_cgroup; | 24 | struct page_cgroup; |
25 | struct page; | 25 | struct page; |
26 | struct mm_struct; | 26 | struct mm_struct; |
27 | 27 | ||
28 | #ifdef CONFIG_CGROUP_MEM_RES_CTLR | 28 | #ifdef CONFIG_CGROUP_MEM_RES_CTLR |
29 | /* | ||
30 | * All "charge" functions with gfp_mask should use GFP_KERNEL or | ||
31 | * (gfp_mask & GFP_RECLAIM_MASK). In current implementatin, memcg doesn't | ||
32 | * alloc memory but reclaims memory from all available zones. So, "where I want | ||
33 | * memory from" bits of gfp_mask has no meaning. So any bits of that field is | ||
34 | * available but adding a rule is better. charge functions' gfp_mask should | ||
35 | * be set to GFP_KERNEL or gfp_mask & GFP_RECLAIM_MASK for avoiding ambiguous | ||
36 | * codes. | ||
37 | * (Of course, if memcg does memory allocation in future, GFP_KERNEL is sane.) | ||
38 | */ | ||
29 | 39 | ||
30 | extern int mem_cgroup_charge(struct page *page, struct mm_struct *mm, | 40 | extern int mem_cgroup_newpage_charge(struct page *page, struct mm_struct *mm, |
31 | gfp_t gfp_mask); | 41 | gfp_t gfp_mask); |
42 | /* for swap handling */ | ||
43 | extern int mem_cgroup_try_charge_swapin(struct mm_struct *mm, | ||
44 | struct page *page, gfp_t mask, struct mem_cgroup **ptr); | ||
45 | extern void mem_cgroup_commit_charge_swapin(struct page *page, | ||
46 | struct mem_cgroup *ptr); | ||
47 | extern void mem_cgroup_cancel_charge_swapin(struct mem_cgroup *ptr); | ||
48 | |||
32 | extern int mem_cgroup_cache_charge(struct page *page, struct mm_struct *mm, | 49 | extern int mem_cgroup_cache_charge(struct page *page, struct mm_struct *mm, |
33 | gfp_t gfp_mask); | 50 | gfp_t gfp_mask); |
34 | extern void mem_cgroup_move_lists(struct page *page, enum lru_list lru); | 51 | extern void mem_cgroup_add_lru_list(struct page *page, enum lru_list lru); |
52 | extern void mem_cgroup_del_lru_list(struct page *page, enum lru_list lru); | ||
53 | extern void mem_cgroup_rotate_lru_list(struct page *page, enum lru_list lru); | ||
54 | extern void mem_cgroup_del_lru(struct page *page); | ||
55 | extern void mem_cgroup_move_lists(struct page *page, | ||
56 | enum lru_list from, enum lru_list to); | ||
35 | extern void mem_cgroup_uncharge_page(struct page *page); | 57 | extern void mem_cgroup_uncharge_page(struct page *page); |
36 | extern void mem_cgroup_uncharge_cache_page(struct page *page); | 58 | extern void mem_cgroup_uncharge_cache_page(struct page *page); |
37 | extern int mem_cgroup_shrink_usage(struct mm_struct *mm, gfp_t gfp_mask); | 59 | extern int mem_cgroup_shrink_usage(struct page *page, |
60 | struct mm_struct *mm, gfp_t gfp_mask); | ||
38 | 61 | ||
39 | extern unsigned long mem_cgroup_isolate_pages(unsigned long nr_to_scan, | 62 | extern unsigned long mem_cgroup_isolate_pages(unsigned long nr_to_scan, |
40 | struct list_head *dst, | 63 | struct list_head *dst, |
@@ -47,12 +70,20 @@ int task_in_mem_cgroup(struct task_struct *task, const struct mem_cgroup *mem); | |||
47 | 70 | ||
48 | extern struct mem_cgroup *mem_cgroup_from_task(struct task_struct *p); | 71 | extern struct mem_cgroup *mem_cgroup_from_task(struct task_struct *p); |
49 | 72 | ||
50 | #define mm_match_cgroup(mm, cgroup) \ | 73 | static inline |
51 | ((cgroup) == mem_cgroup_from_task((mm)->owner)) | 74 | int mm_match_cgroup(const struct mm_struct *mm, const struct mem_cgroup *cgroup) |
75 | { | ||
76 | struct mem_cgroup *mem; | ||
77 | rcu_read_lock(); | ||
78 | mem = mem_cgroup_from_task((mm)->owner); | ||
79 | rcu_read_unlock(); | ||
80 | return cgroup == mem; | ||
81 | } | ||
52 | 82 | ||
53 | extern int | 83 | extern int |
54 | mem_cgroup_prepare_migration(struct page *page, struct page *newpage); | 84 | mem_cgroup_prepare_migration(struct page *page, struct mem_cgroup **ptr); |
55 | extern void mem_cgroup_end_migration(struct page *page); | 85 | extern void mem_cgroup_end_migration(struct mem_cgroup *mem, |
86 | struct page *oldpage, struct page *newpage); | ||
56 | 87 | ||
57 | /* | 88 | /* |
58 | * For memory reclaim. | 89 | * For memory reclaim. |
@@ -65,13 +96,32 @@ extern void mem_cgroup_note_reclaim_priority(struct mem_cgroup *mem, | |||
65 | int priority); | 96 | int priority); |
66 | extern void mem_cgroup_record_reclaim_priority(struct mem_cgroup *mem, | 97 | extern void mem_cgroup_record_reclaim_priority(struct mem_cgroup *mem, |
67 | int priority); | 98 | int priority); |
99 | int mem_cgroup_inactive_anon_is_low(struct mem_cgroup *memcg); | ||
100 | unsigned long mem_cgroup_zone_nr_pages(struct mem_cgroup *memcg, | ||
101 | struct zone *zone, | ||
102 | enum lru_list lru); | ||
103 | struct zone_reclaim_stat *mem_cgroup_get_reclaim_stat(struct mem_cgroup *memcg, | ||
104 | struct zone *zone); | ||
105 | struct zone_reclaim_stat* | ||
106 | mem_cgroup_get_reclaim_stat_from_page(struct page *page); | ||
68 | 107 | ||
69 | extern long mem_cgroup_calc_reclaim(struct mem_cgroup *mem, struct zone *zone, | 108 | #ifdef CONFIG_CGROUP_MEM_RES_CTLR_SWAP |
70 | int priority, enum lru_list lru); | 109 | extern int do_swap_account; |
110 | #endif | ||
71 | 111 | ||
112 | static inline bool mem_cgroup_disabled(void) | ||
113 | { | ||
114 | if (mem_cgroup_subsys.disabled) | ||
115 | return true; | ||
116 | return false; | ||
117 | } | ||
118 | |||
119 | extern bool mem_cgroup_oom_called(struct task_struct *task); | ||
72 | 120 | ||
73 | #else /* CONFIG_CGROUP_MEM_RES_CTLR */ | 121 | #else /* CONFIG_CGROUP_MEM_RES_CTLR */ |
74 | static inline int mem_cgroup_charge(struct page *page, | 122 | struct mem_cgroup; |
123 | |||
124 | static inline int mem_cgroup_newpage_charge(struct page *page, | ||
75 | struct mm_struct *mm, gfp_t gfp_mask) | 125 | struct mm_struct *mm, gfp_t gfp_mask) |
76 | { | 126 | { |
77 | return 0; | 127 | return 0; |
@@ -83,6 +133,21 @@ static inline int mem_cgroup_cache_charge(struct page *page, | |||
83 | return 0; | 133 | return 0; |
84 | } | 134 | } |
85 | 135 | ||
136 | static inline int mem_cgroup_try_charge_swapin(struct mm_struct *mm, | ||
137 | struct page *page, gfp_t gfp_mask, struct mem_cgroup **ptr) | ||
138 | { | ||
139 | return 0; | ||
140 | } | ||
141 | |||
142 | static inline void mem_cgroup_commit_charge_swapin(struct page *page, | ||
143 | struct mem_cgroup *ptr) | ||
144 | { | ||
145 | } | ||
146 | |||
147 | static inline void mem_cgroup_cancel_charge_swapin(struct mem_cgroup *ptr) | ||
148 | { | ||
149 | } | ||
150 | |||
86 | static inline void mem_cgroup_uncharge_page(struct page *page) | 151 | static inline void mem_cgroup_uncharge_page(struct page *page) |
87 | { | 152 | { |
88 | } | 153 | } |
@@ -91,12 +156,33 @@ static inline void mem_cgroup_uncharge_cache_page(struct page *page) | |||
91 | { | 156 | { |
92 | } | 157 | } |
93 | 158 | ||
94 | static inline int mem_cgroup_shrink_usage(struct mm_struct *mm, gfp_t gfp_mask) | 159 | static inline int mem_cgroup_shrink_usage(struct page *page, |
160 | struct mm_struct *mm, gfp_t gfp_mask) | ||
95 | { | 161 | { |
96 | return 0; | 162 | return 0; |
97 | } | 163 | } |
98 | 164 | ||
99 | static inline void mem_cgroup_move_lists(struct page *page, bool active) | 165 | static inline void mem_cgroup_add_lru_list(struct page *page, int lru) |
166 | { | ||
167 | } | ||
168 | |||
169 | static inline void mem_cgroup_del_lru_list(struct page *page, int lru) | ||
170 | { | ||
171 | return ; | ||
172 | } | ||
173 | |||
174 | static inline void mem_cgroup_rotate_lru_list(struct page *page, int lru) | ||
175 | { | ||
176 | return ; | ||
177 | } | ||
178 | |||
179 | static inline void mem_cgroup_del_lru(struct page *page) | ||
180 | { | ||
181 | return ; | ||
182 | } | ||
183 | |||
184 | static inline void | ||
185 | mem_cgroup_move_lists(struct page *page, enum lru_list from, enum lru_list to) | ||
100 | { | 186 | { |
101 | } | 187 | } |
102 | 188 | ||
@@ -112,12 +198,14 @@ static inline int task_in_mem_cgroup(struct task_struct *task, | |||
112 | } | 198 | } |
113 | 199 | ||
114 | static inline int | 200 | static inline int |
115 | mem_cgroup_prepare_migration(struct page *page, struct page *newpage) | 201 | mem_cgroup_prepare_migration(struct page *page, struct mem_cgroup **ptr) |
116 | { | 202 | { |
117 | return 0; | 203 | return 0; |
118 | } | 204 | } |
119 | 205 | ||
120 | static inline void mem_cgroup_end_migration(struct page *page) | 206 | static inline void mem_cgroup_end_migration(struct mem_cgroup *mem, |
207 | struct page *oldpage, | ||
208 | struct page *newpage) | ||
121 | { | 209 | { |
122 | } | 210 | } |
123 | 211 | ||
@@ -146,12 +234,42 @@ static inline void mem_cgroup_record_reclaim_priority(struct mem_cgroup *mem, | |||
146 | { | 234 | { |
147 | } | 235 | } |
148 | 236 | ||
149 | static inline long mem_cgroup_calc_reclaim(struct mem_cgroup *mem, | 237 | static inline bool mem_cgroup_disabled(void) |
150 | struct zone *zone, int priority, | 238 | { |
151 | enum lru_list lru) | 239 | return true; |
240 | } | ||
241 | |||
242 | static inline bool mem_cgroup_oom_called(struct task_struct *task) | ||
243 | { | ||
244 | return false; | ||
245 | } | ||
246 | |||
247 | static inline int | ||
248 | mem_cgroup_inactive_anon_is_low(struct mem_cgroup *memcg) | ||
249 | { | ||
250 | return 1; | ||
251 | } | ||
252 | |||
253 | static inline unsigned long | ||
254 | mem_cgroup_zone_nr_pages(struct mem_cgroup *memcg, struct zone *zone, | ||
255 | enum lru_list lru) | ||
152 | { | 256 | { |
153 | return 0; | 257 | return 0; |
154 | } | 258 | } |
259 | |||
260 | |||
261 | static inline struct zone_reclaim_stat* | ||
262 | mem_cgroup_get_reclaim_stat(struct mem_cgroup *memcg, struct zone *zone) | ||
263 | { | ||
264 | return NULL; | ||
265 | } | ||
266 | |||
267 | static inline struct zone_reclaim_stat* | ||
268 | mem_cgroup_get_reclaim_stat_from_page(struct page *page) | ||
269 | { | ||
270 | return NULL; | ||
271 | } | ||
272 | |||
155 | #endif /* CONFIG_CGROUP_MEM_CONT */ | 273 | #endif /* CONFIG_CGROUP_MEM_CONT */ |
156 | 274 | ||
157 | #endif /* _LINUX_MEMCONTROL_H */ | 275 | #endif /* _LINUX_MEMCONTROL_H */ |
diff --git a/include/linux/mm_inline.h b/include/linux/mm_inline.h index c948350c378e..7fbb97267556 100644 --- a/include/linux/mm_inline.h +++ b/include/linux/mm_inline.h | |||
@@ -28,6 +28,7 @@ add_page_to_lru_list(struct zone *zone, struct page *page, enum lru_list l) | |||
28 | { | 28 | { |
29 | list_add(&page->lru, &zone->lru[l].list); | 29 | list_add(&page->lru, &zone->lru[l].list); |
30 | __inc_zone_state(zone, NR_LRU_BASE + l); | 30 | __inc_zone_state(zone, NR_LRU_BASE + l); |
31 | mem_cgroup_add_lru_list(page, l); | ||
31 | } | 32 | } |
32 | 33 | ||
33 | static inline void | 34 | static inline void |
@@ -35,6 +36,7 @@ del_page_from_lru_list(struct zone *zone, struct page *page, enum lru_list l) | |||
35 | { | 36 | { |
36 | list_del(&page->lru); | 37 | list_del(&page->lru); |
37 | __dec_zone_state(zone, NR_LRU_BASE + l); | 38 | __dec_zone_state(zone, NR_LRU_BASE + l); |
39 | mem_cgroup_del_lru_list(page, l); | ||
38 | } | 40 | } |
39 | 41 | ||
40 | static inline void | 42 | static inline void |
@@ -54,6 +56,7 @@ del_page_from_lru(struct zone *zone, struct page *page) | |||
54 | l += page_is_file_cache(page); | 56 | l += page_is_file_cache(page); |
55 | } | 57 | } |
56 | __dec_zone_state(zone, NR_LRU_BASE + l); | 58 | __dec_zone_state(zone, NR_LRU_BASE + l); |
59 | mem_cgroup_del_lru_list(page, l); | ||
57 | } | 60 | } |
58 | 61 | ||
59 | /** | 62 | /** |
@@ -78,23 +81,4 @@ static inline enum lru_list page_lru(struct page *page) | |||
78 | return lru; | 81 | return lru; |
79 | } | 82 | } |
80 | 83 | ||
81 | /** | ||
82 | * inactive_anon_is_low - check if anonymous pages need to be deactivated | ||
83 | * @zone: zone to check | ||
84 | * | ||
85 | * Returns true if the zone does not have enough inactive anon pages, | ||
86 | * meaning some active anon pages need to be deactivated. | ||
87 | */ | ||
88 | static inline int inactive_anon_is_low(struct zone *zone) | ||
89 | { | ||
90 | unsigned long active, inactive; | ||
91 | |||
92 | active = zone_page_state(zone, NR_ACTIVE_ANON); | ||
93 | inactive = zone_page_state(zone, NR_INACTIVE_ANON); | ||
94 | |||
95 | if (inactive * zone->inactive_ratio < active) | ||
96 | return 1; | ||
97 | |||
98 | return 0; | ||
99 | } | ||
100 | #endif | 84 | #endif |
diff --git a/include/linux/mmzone.h b/include/linux/mmzone.h index 35a7b5e19465..09c14e213b63 100644 --- a/include/linux/mmzone.h +++ b/include/linux/mmzone.h | |||
@@ -263,6 +263,19 @@ enum zone_type { | |||
263 | #error ZONES_SHIFT -- too many zones configured adjust calculation | 263 | #error ZONES_SHIFT -- too many zones configured adjust calculation |
264 | #endif | 264 | #endif |
265 | 265 | ||
266 | struct zone_reclaim_stat { | ||
267 | /* | ||
268 | * The pageout code in vmscan.c keeps track of how many of the | ||
269 | * mem/swap backed and file backed pages are refeferenced. | ||
270 | * The higher the rotated/scanned ratio, the more valuable | ||
271 | * that cache is. | ||
272 | * | ||
273 | * The anon LRU stats live in [0], file LRU stats in [1] | ||
274 | */ | ||
275 | unsigned long recent_rotated[2]; | ||
276 | unsigned long recent_scanned[2]; | ||
277 | }; | ||
278 | |||
266 | struct zone { | 279 | struct zone { |
267 | /* Fields commonly accessed by the page allocator */ | 280 | /* Fields commonly accessed by the page allocator */ |
268 | unsigned long pages_min, pages_low, pages_high; | 281 | unsigned long pages_min, pages_low, pages_high; |
@@ -315,16 +328,7 @@ struct zone { | |||
315 | unsigned long nr_scan; | 328 | unsigned long nr_scan; |
316 | } lru[NR_LRU_LISTS]; | 329 | } lru[NR_LRU_LISTS]; |
317 | 330 | ||
318 | /* | 331 | struct zone_reclaim_stat reclaim_stat; |
319 | * The pageout code in vmscan.c keeps track of how many of the | ||
320 | * mem/swap backed and file backed pages are refeferenced. | ||
321 | * The higher the rotated/scanned ratio, the more valuable | ||
322 | * that cache is. | ||
323 | * | ||
324 | * The anon LRU stats live in [0], file LRU stats in [1] | ||
325 | */ | ||
326 | unsigned long recent_rotated[2]; | ||
327 | unsigned long recent_scanned[2]; | ||
328 | 332 | ||
329 | unsigned long pages_scanned; /* since last reclaim */ | 333 | unsigned long pages_scanned; /* since last reclaim */ |
330 | unsigned long flags; /* zone flags, see below */ | 334 | unsigned long flags; /* zone flags, see below */ |
diff --git a/include/linux/nwpserial.h b/include/linux/nwpserial.h new file mode 100644 index 000000000000..9acb21572eaf --- /dev/null +++ b/include/linux/nwpserial.h | |||
@@ -0,0 +1,18 @@ | |||
1 | /* | ||
2 | * Serial Port driver for a NWP uart device | ||
3 | * | ||
4 | * Copyright (C) 2008 IBM Corp., Benjamin Krill <ben@codiert.org> | ||
5 | * | ||
6 | * This program is free software; you can redistribute it and/or | ||
7 | * modify it under the terms of the GNU General Public License | ||
8 | * as published by the Free Software Foundation; either version | ||
9 | * 2 of the License, or (at your option) any later version. | ||
10 | * | ||
11 | */ | ||
12 | #ifndef _NWPSERIAL_H | ||
13 | #define _NWPSERIAL_H | ||
14 | |||
15 | int nwpserial_register_port(struct uart_port *port); | ||
16 | void nwpserial_unregister_port(int line); | ||
17 | |||
18 | #endif /* _NWPSERIAL_H */ | ||
diff --git a/include/linux/page_cgroup.h b/include/linux/page_cgroup.h index 1e6d34bfa094..602cc1fdee90 100644 --- a/include/linux/page_cgroup.h +++ b/include/linux/page_cgroup.h | |||
@@ -26,10 +26,6 @@ enum { | |||
26 | PCG_LOCK, /* page cgroup is locked */ | 26 | PCG_LOCK, /* page cgroup is locked */ |
27 | PCG_CACHE, /* charged as cache */ | 27 | PCG_CACHE, /* charged as cache */ |
28 | PCG_USED, /* this object is in use. */ | 28 | PCG_USED, /* this object is in use. */ |
29 | /* flags for LRU placement */ | ||
30 | PCG_ACTIVE, /* page is active in this cgroup */ | ||
31 | PCG_FILE, /* page is file system backed */ | ||
32 | PCG_UNEVICTABLE, /* page is unevictableable */ | ||
33 | }; | 29 | }; |
34 | 30 | ||
35 | #define TESTPCGFLAG(uname, lname) \ | 31 | #define TESTPCGFLAG(uname, lname) \ |
@@ -50,19 +46,6 @@ TESTPCGFLAG(Cache, CACHE) | |||
50 | TESTPCGFLAG(Used, USED) | 46 | TESTPCGFLAG(Used, USED) |
51 | CLEARPCGFLAG(Used, USED) | 47 | CLEARPCGFLAG(Used, USED) |
52 | 48 | ||
53 | /* LRU management flags (from global-lru definition) */ | ||
54 | TESTPCGFLAG(File, FILE) | ||
55 | SETPCGFLAG(File, FILE) | ||
56 | CLEARPCGFLAG(File, FILE) | ||
57 | |||
58 | TESTPCGFLAG(Active, ACTIVE) | ||
59 | SETPCGFLAG(Active, ACTIVE) | ||
60 | CLEARPCGFLAG(Active, ACTIVE) | ||
61 | |||
62 | TESTPCGFLAG(Unevictable, UNEVICTABLE) | ||
63 | SETPCGFLAG(Unevictable, UNEVICTABLE) | ||
64 | CLEARPCGFLAG(Unevictable, UNEVICTABLE) | ||
65 | |||
66 | static inline int page_cgroup_nid(struct page_cgroup *pc) | 49 | static inline int page_cgroup_nid(struct page_cgroup *pc) |
67 | { | 50 | { |
68 | return page_to_nid(pc->page); | 51 | return page_to_nid(pc->page); |
@@ -105,4 +88,39 @@ static inline void page_cgroup_init(void) | |||
105 | } | 88 | } |
106 | 89 | ||
107 | #endif | 90 | #endif |
91 | |||
92 | #ifdef CONFIG_CGROUP_MEM_RES_CTLR_SWAP | ||
93 | #include <linux/swap.h> | ||
94 | extern struct mem_cgroup * | ||
95 | swap_cgroup_record(swp_entry_t ent, struct mem_cgroup *mem); | ||
96 | extern struct mem_cgroup *lookup_swap_cgroup(swp_entry_t ent); | ||
97 | extern int swap_cgroup_swapon(int type, unsigned long max_pages); | ||
98 | extern void swap_cgroup_swapoff(int type); | ||
99 | #else | ||
100 | #include <linux/swap.h> | ||
101 | |||
102 | static inline | ||
103 | struct mem_cgroup *swap_cgroup_record(swp_entry_t ent, struct mem_cgroup *mem) | ||
104 | { | ||
105 | return NULL; | ||
106 | } | ||
107 | |||
108 | static inline | ||
109 | struct mem_cgroup *lookup_swap_cgroup(swp_entry_t ent) | ||
110 | { | ||
111 | return NULL; | ||
112 | } | ||
113 | |||
114 | static inline int | ||
115 | swap_cgroup_swapon(int type, unsigned long max_pages) | ||
116 | { | ||
117 | return 0; | ||
118 | } | ||
119 | |||
120 | static inline void swap_cgroup_swapoff(int type) | ||
121 | { | ||
122 | return; | ||
123 | } | ||
124 | |||
125 | #endif | ||
108 | #endif | 126 | #endif |
diff --git a/include/linux/pid.h b/include/linux/pid.h index bb206c56d1f0..49f1c2f66e95 100644 --- a/include/linux/pid.h +++ b/include/linux/pid.h | |||
@@ -123,6 +123,24 @@ extern struct pid *alloc_pid(struct pid_namespace *ns); | |||
123 | extern void free_pid(struct pid *pid); | 123 | extern void free_pid(struct pid *pid); |
124 | 124 | ||
125 | /* | 125 | /* |
126 | * ns_of_pid() returns the pid namespace in which the specified pid was | ||
127 | * allocated. | ||
128 | * | ||
129 | * NOTE: | ||
130 | * ns_of_pid() is expected to be called for a process (task) that has | ||
131 | * an attached 'struct pid' (see attach_pid(), detach_pid()) i.e @pid | ||
132 | * is expected to be non-NULL. If @pid is NULL, caller should handle | ||
133 | * the resulting NULL pid-ns. | ||
134 | */ | ||
135 | static inline struct pid_namespace *ns_of_pid(struct pid *pid) | ||
136 | { | ||
137 | struct pid_namespace *ns = NULL; | ||
138 | if (pid) | ||
139 | ns = pid->numbers[pid->level].ns; | ||
140 | return ns; | ||
141 | } | ||
142 | |||
143 | /* | ||
126 | * the helpers to get the pid's id seen from different namespaces | 144 | * the helpers to get the pid's id seen from different namespaces |
127 | * | 145 | * |
128 | * pid_nr() : global id, i.e. the id seen from the init namespace; | 146 | * pid_nr() : global id, i.e. the id seen from the init namespace; |
diff --git a/include/linux/pid_namespace.h b/include/linux/pid_namespace.h index d82fe825d62f..38d10326246a 100644 --- a/include/linux/pid_namespace.h +++ b/include/linux/pid_namespace.h | |||
@@ -79,11 +79,7 @@ static inline void zap_pid_ns_processes(struct pid_namespace *ns) | |||
79 | } | 79 | } |
80 | #endif /* CONFIG_PID_NS */ | 80 | #endif /* CONFIG_PID_NS */ |
81 | 81 | ||
82 | static inline struct pid_namespace *task_active_pid_ns(struct task_struct *tsk) | 82 | extern struct pid_namespace *task_active_pid_ns(struct task_struct *tsk); |
83 | { | ||
84 | return tsk->nsproxy->pid_ns; | ||
85 | } | ||
86 | |||
87 | void pidhash_init(void); | 83 | void pidhash_init(void); |
88 | void pidmap_init(void); | 84 | void pidmap_init(void); |
89 | 85 | ||
diff --git a/include/linux/raid/md_k.h b/include/linux/raid/md_k.h index 8fc909ef6787..9743e4dbc918 100644 --- a/include/linux/raid/md_k.h +++ b/include/linux/raid/md_k.h | |||
@@ -137,6 +137,9 @@ struct mddev_s | |||
137 | struct gendisk *gendisk; | 137 | struct gendisk *gendisk; |
138 | 138 | ||
139 | struct kobject kobj; | 139 | struct kobject kobj; |
140 | int hold_active; | ||
141 | #define UNTIL_IOCTL 1 | ||
142 | #define UNTIL_STOP 2 | ||
140 | 143 | ||
141 | /* Superblock information */ | 144 | /* Superblock information */ |
142 | int major_version, | 145 | int major_version, |
@@ -215,6 +218,9 @@ struct mddev_s | |||
215 | #define MD_RECOVERY_FROZEN 9 | 218 | #define MD_RECOVERY_FROZEN 9 |
216 | 219 | ||
217 | unsigned long recovery; | 220 | unsigned long recovery; |
221 | int recovery_disabled; /* if we detect that recovery | ||
222 | * will always fail, set this | ||
223 | * so we don't loop trying */ | ||
218 | 224 | ||
219 | int in_sync; /* know to not need resync */ | 225 | int in_sync; /* know to not need resync */ |
220 | struct mutex reconfig_mutex; | 226 | struct mutex reconfig_mutex; |
@@ -244,6 +250,9 @@ struct mddev_s | |||
244 | struct sysfs_dirent *sysfs_state; /* handle for 'array_state' | 250 | struct sysfs_dirent *sysfs_state; /* handle for 'array_state' |
245 | * file in sysfs. | 251 | * file in sysfs. |
246 | */ | 252 | */ |
253 | struct sysfs_dirent *sysfs_action; /* handle for 'sync_action' */ | ||
254 | |||
255 | struct work_struct del_work; /* used for delayed sysfs removal */ | ||
247 | 256 | ||
248 | spinlock_t write_lock; | 257 | spinlock_t write_lock; |
249 | wait_queue_head_t sb_wait; /* for waiting on superblock updates */ | 258 | wait_queue_head_t sb_wait; /* for waiting on superblock updates */ |
@@ -334,17 +343,14 @@ static inline char * mdname (mddev_t * mddev) | |||
334 | * iterates through some rdev ringlist. It's safe to remove the | 343 | * iterates through some rdev ringlist. It's safe to remove the |
335 | * current 'rdev'. Dont touch 'tmp' though. | 344 | * current 'rdev'. Dont touch 'tmp' though. |
336 | */ | 345 | */ |
337 | #define rdev_for_each_list(rdev, tmp, list) \ | 346 | #define rdev_for_each_list(rdev, tmp, head) \ |
338 | \ | 347 | list_for_each_entry_safe(rdev, tmp, head, same_set) |
339 | for ((tmp) = (list).next; \ | 348 | |
340 | (rdev) = (list_entry((tmp), mdk_rdev_t, same_set)), \ | ||
341 | (tmp) = (tmp)->next, (tmp)->prev != &(list) \ | ||
342 | ; ) | ||
343 | /* | 349 | /* |
344 | * iterates through the 'same array disks' ringlist | 350 | * iterates through the 'same array disks' ringlist |
345 | */ | 351 | */ |
346 | #define rdev_for_each(rdev, tmp, mddev) \ | 352 | #define rdev_for_each(rdev, tmp, mddev) \ |
347 | rdev_for_each_list(rdev, tmp, (mddev)->disks) | 353 | list_for_each_entry_safe(rdev, tmp, &((mddev)->disks), same_set) |
348 | 354 | ||
349 | #define rdev_for_each_rcu(rdev, mddev) \ | 355 | #define rdev_for_each_rcu(rdev, mddev) \ |
350 | list_for_each_entry_rcu(rdev, &((mddev)->disks), same_set) | 356 | list_for_each_entry_rcu(rdev, &((mddev)->disks), same_set) |
diff --git a/include/linux/raid/md_p.h b/include/linux/raid/md_p.h index 8b4de4a41ff1..9491026afe66 100644 --- a/include/linux/raid/md_p.h +++ b/include/linux/raid/md_p.h | |||
@@ -194,6 +194,8 @@ static inline __u64 md_event(mdp_super_t *sb) { | |||
194 | return (ev<<32)| sb->events_lo; | 194 | return (ev<<32)| sb->events_lo; |
195 | } | 195 | } |
196 | 196 | ||
197 | #define MD_SUPERBLOCK_1_TIME_SEC_MASK ((1ULL<<40) - 1) | ||
198 | |||
197 | /* | 199 | /* |
198 | * The version-1 superblock : | 200 | * The version-1 superblock : |
199 | * All numeric fields are little-endian. | 201 | * All numeric fields are little-endian. |
diff --git a/include/linux/raid/raid0.h b/include/linux/raid/raid0.h index 1b2dda035f8e..fd42aa87c391 100644 --- a/include/linux/raid/raid0.h +++ b/include/linux/raid/raid0.h | |||
@@ -5,9 +5,9 @@ | |||
5 | 5 | ||
6 | struct strip_zone | 6 | struct strip_zone |
7 | { | 7 | { |
8 | sector_t zone_offset; /* Zone offset in md_dev */ | 8 | sector_t zone_start; /* Zone offset in md_dev (in sectors) */ |
9 | sector_t dev_offset; /* Zone offset in real dev */ | 9 | sector_t dev_start; /* Zone offset in real dev (in sectors) */ |
10 | sector_t size; /* Zone size */ | 10 | sector_t sectors; /* Zone size in sectors */ |
11 | int nb_dev; /* # of devices attached to the zone */ | 11 | int nb_dev; /* # of devices attached to the zone */ |
12 | mdk_rdev_t **dev; /* Devices attached to the zone */ | 12 | mdk_rdev_t **dev; /* Devices attached to the zone */ |
13 | }; | 13 | }; |
@@ -19,8 +19,8 @@ struct raid0_private_data | |||
19 | mdk_rdev_t **devlist; /* lists of rdevs, pointed to by strip_zone->dev */ | 19 | mdk_rdev_t **devlist; /* lists of rdevs, pointed to by strip_zone->dev */ |
20 | int nr_strip_zones; | 20 | int nr_strip_zones; |
21 | 21 | ||
22 | sector_t hash_spacing; | 22 | sector_t spacing; |
23 | int preshift; /* shift this before divide by hash_spacing */ | 23 | int sector_shift; /* shift this before divide by spacing */ |
24 | }; | 24 | }; |
25 | 25 | ||
26 | typedef struct raid0_private_data raid0_conf_t; | 26 | typedef struct raid0_private_data raid0_conf_t; |
diff --git a/include/linux/res_counter.h b/include/linux/res_counter.h index 271c1c2c9f6f..dede0a2cfc45 100644 --- a/include/linux/res_counter.h +++ b/include/linux/res_counter.h | |||
@@ -43,6 +43,10 @@ struct res_counter { | |||
43 | * the routines below consider this to be IRQ-safe | 43 | * the routines below consider this to be IRQ-safe |
44 | */ | 44 | */ |
45 | spinlock_t lock; | 45 | spinlock_t lock; |
46 | /* | ||
47 | * Parent counter, used for hierarchial resource accounting | ||
48 | */ | ||
49 | struct res_counter *parent; | ||
46 | }; | 50 | }; |
47 | 51 | ||
48 | /** | 52 | /** |
@@ -87,7 +91,7 @@ enum { | |||
87 | * helpers for accounting | 91 | * helpers for accounting |
88 | */ | 92 | */ |
89 | 93 | ||
90 | void res_counter_init(struct res_counter *counter); | 94 | void res_counter_init(struct res_counter *counter, struct res_counter *parent); |
91 | 95 | ||
92 | /* | 96 | /* |
93 | * charge - try to consume more resource. | 97 | * charge - try to consume more resource. |
@@ -103,7 +107,7 @@ void res_counter_init(struct res_counter *counter); | |||
103 | int __must_check res_counter_charge_locked(struct res_counter *counter, | 107 | int __must_check res_counter_charge_locked(struct res_counter *counter, |
104 | unsigned long val); | 108 | unsigned long val); |
105 | int __must_check res_counter_charge(struct res_counter *counter, | 109 | int __must_check res_counter_charge(struct res_counter *counter, |
106 | unsigned long val); | 110 | unsigned long val, struct res_counter **limit_fail_at); |
107 | 111 | ||
108 | /* | 112 | /* |
109 | * uncharge - tell that some portion of the resource is released | 113 | * uncharge - tell that some portion of the resource is released |
diff --git a/include/linux/serial_core.h b/include/linux/serial_core.h index b4199841f1fc..90bbbf0b1161 100644 --- a/include/linux/serial_core.h +++ b/include/linux/serial_core.h | |||
@@ -161,6 +161,9 @@ | |||
161 | 161 | ||
162 | #define PORT_S3C6400 84 | 162 | #define PORT_S3C6400 84 |
163 | 163 | ||
164 | /* NWPSERIAL */ | ||
165 | #define PORT_NWPSERIAL 85 | ||
166 | |||
164 | #ifdef __KERNEL__ | 167 | #ifdef __KERNEL__ |
165 | 168 | ||
166 | #include <linux/compiler.h> | 169 | #include <linux/compiler.h> |
diff --git a/include/linux/swap.h b/include/linux/swap.h index 91dee50fe260..d30215578877 100644 --- a/include/linux/swap.h +++ b/include/linux/swap.h | |||
@@ -214,7 +214,8 @@ static inline void lru_cache_add_active_file(struct page *page) | |||
214 | extern unsigned long try_to_free_pages(struct zonelist *zonelist, int order, | 214 | extern unsigned long try_to_free_pages(struct zonelist *zonelist, int order, |
215 | gfp_t gfp_mask); | 215 | gfp_t gfp_mask); |
216 | extern unsigned long try_to_free_mem_cgroup_pages(struct mem_cgroup *mem, | 216 | extern unsigned long try_to_free_mem_cgroup_pages(struct mem_cgroup *mem, |
217 | gfp_t gfp_mask); | 217 | gfp_t gfp_mask, bool noswap, |
218 | unsigned int swappiness); | ||
218 | extern int __isolate_lru_page(struct page *page, int mode, int file); | 219 | extern int __isolate_lru_page(struct page *page, int mode, int file); |
219 | extern unsigned long shrink_all_memory(unsigned long nr_pages); | 220 | extern unsigned long shrink_all_memory(unsigned long nr_pages); |
220 | extern int vm_swappiness; | 221 | extern int vm_swappiness; |
@@ -333,6 +334,22 @@ static inline void disable_swap_token(void) | |||
333 | put_swap_token(swap_token_mm); | 334 | put_swap_token(swap_token_mm); |
334 | } | 335 | } |
335 | 336 | ||
337 | #ifdef CONFIG_CGROUP_MEM_RES_CTLR | ||
338 | extern void mem_cgroup_uncharge_swapcache(struct page *page, swp_entry_t ent); | ||
339 | #else | ||
340 | static inline void | ||
341 | mem_cgroup_uncharge_swapcache(struct page *page, swp_entry_t ent) | ||
342 | { | ||
343 | } | ||
344 | #endif | ||
345 | #ifdef CONFIG_CGROUP_MEM_RES_CTLR_SWAP | ||
346 | extern void mem_cgroup_uncharge_swap(swp_entry_t ent); | ||
347 | #else | ||
348 | static inline void mem_cgroup_uncharge_swap(swp_entry_t ent) | ||
349 | { | ||
350 | } | ||
351 | #endif | ||
352 | |||
336 | #else /* CONFIG_SWAP */ | 353 | #else /* CONFIG_SWAP */ |
337 | 354 | ||
338 | #define nr_swap_pages 0L | 355 | #define nr_swap_pages 0L |
@@ -409,6 +426,12 @@ static inline swp_entry_t get_swap_page(void) | |||
409 | #define has_swap_token(x) 0 | 426 | #define has_swap_token(x) 0 |
410 | #define disable_swap_token() do { } while(0) | 427 | #define disable_swap_token() do { } while(0) |
411 | 428 | ||
429 | static inline int mem_cgroup_cache_charge_swapin(struct page *page, | ||
430 | struct mm_struct *mm, gfp_t mask, bool locked) | ||
431 | { | ||
432 | return 0; | ||
433 | } | ||
434 | |||
412 | #endif /* CONFIG_SWAP */ | 435 | #endif /* CONFIG_SWAP */ |
413 | #endif /* __KERNEL__*/ | 436 | #endif /* __KERNEL__*/ |
414 | #endif /* _LINUX_SWAP_H */ | 437 | #endif /* _LINUX_SWAP_H */ |
diff --git a/include/xen/xenbus.h b/include/xen/xenbus.h index 6369d89c25d5..f87f9614844d 100644 --- a/include/xen/xenbus.h +++ b/include/xen/xenbus.h | |||
@@ -136,8 +136,6 @@ struct xenbus_transaction | |||
136 | /* Nil transaction ID. */ | 136 | /* Nil transaction ID. */ |
137 | #define XBT_NIL ((struct xenbus_transaction) { 0 }) | 137 | #define XBT_NIL ((struct xenbus_transaction) { 0 }) |
138 | 138 | ||
139 | int __init xenbus_dev_init(void); | ||
140 | |||
141 | char **xenbus_directory(struct xenbus_transaction t, | 139 | char **xenbus_directory(struct xenbus_transaction t, |
142 | const char *dir, const char *node, unsigned int *num); | 140 | const char *dir, const char *node, unsigned int *num); |
143 | void *xenbus_read(struct xenbus_transaction t, | 141 | void *xenbus_read(struct xenbus_transaction t, |
diff --git a/init/Kconfig b/init/Kconfig index e7893b1d3e42..a724a149bf3f 100644 --- a/init/Kconfig +++ b/init/Kconfig | |||
@@ -271,59 +271,6 @@ config LOG_BUF_SHIFT | |||
271 | 13 => 8 KB | 271 | 13 => 8 KB |
272 | 12 => 4 KB | 272 | 12 => 4 KB |
273 | 273 | ||
274 | config CGROUPS | ||
275 | bool "Control Group support" | ||
276 | help | ||
277 | This option will let you use process cgroup subsystems | ||
278 | such as Cpusets | ||
279 | |||
280 | Say N if unsure. | ||
281 | |||
282 | config CGROUP_DEBUG | ||
283 | bool "Example debug cgroup subsystem" | ||
284 | depends on CGROUPS | ||
285 | default n | ||
286 | help | ||
287 | This option enables a simple cgroup subsystem that | ||
288 | exports useful debugging information about the cgroups | ||
289 | framework | ||
290 | |||
291 | Say N if unsure | ||
292 | |||
293 | config CGROUP_NS | ||
294 | bool "Namespace cgroup subsystem" | ||
295 | depends on CGROUPS | ||
296 | help | ||
297 | Provides a simple namespace cgroup subsystem to | ||
298 | provide hierarchical naming of sets of namespaces, | ||
299 | for instance virtual servers and checkpoint/restart | ||
300 | jobs. | ||
301 | |||
302 | config CGROUP_FREEZER | ||
303 | bool "control group freezer subsystem" | ||
304 | depends on CGROUPS | ||
305 | help | ||
306 | Provides a way to freeze and unfreeze all tasks in a | ||
307 | cgroup. | ||
308 | |||
309 | config CGROUP_DEVICE | ||
310 | bool "Device controller for cgroups" | ||
311 | depends on CGROUPS && EXPERIMENTAL | ||
312 | help | ||
313 | Provides a cgroup implementing whitelists for devices which | ||
314 | a process in the cgroup can mknod or open. | ||
315 | |||
316 | config CPUSETS | ||
317 | bool "Cpuset support" | ||
318 | depends on SMP && CGROUPS | ||
319 | help | ||
320 | This option will let you create and manage CPUSETs which | ||
321 | allow dynamically partitioning a system into sets of CPUs and | ||
322 | Memory Nodes and assigning tasks to run only within those sets. | ||
323 | This is primarily useful on large SMP or NUMA systems. | ||
324 | |||
325 | Say N if unsure. | ||
326 | |||
327 | # | 274 | # |
328 | # Architectures with an unreliable sched_clock() should select this: | 275 | # Architectures with an unreliable sched_clock() should select this: |
329 | # | 276 | # |
@@ -337,6 +284,8 @@ config GROUP_SCHED | |||
337 | help | 284 | help |
338 | This feature lets CPU scheduler recognize task groups and control CPU | 285 | This feature lets CPU scheduler recognize task groups and control CPU |
339 | bandwidth allocation to such task groups. | 286 | bandwidth allocation to such task groups. |
287 | In order to create a group from arbitrary set of processes, use | ||
288 | CONFIG_CGROUPS. (See Control Group support.) | ||
340 | 289 | ||
341 | config FAIR_GROUP_SCHED | 290 | config FAIR_GROUP_SCHED |
342 | bool "Group scheduling for SCHED_OTHER" | 291 | bool "Group scheduling for SCHED_OTHER" |
@@ -379,6 +328,66 @@ config CGROUP_SCHED | |||
379 | 328 | ||
380 | endchoice | 329 | endchoice |
381 | 330 | ||
331 | menu "Control Group support" | ||
332 | config CGROUPS | ||
333 | bool "Control Group support" | ||
334 | help | ||
335 | This option add support for grouping sets of processes together, for | ||
336 | use with process control subsystems such as Cpusets, CFS, memory | ||
337 | controls or device isolation. | ||
338 | See | ||
339 | - Documentation/cpusets.txt (Cpusets) | ||
340 | - Documentation/scheduler/sched-design-CFS.txt (CFS) | ||
341 | - Documentation/cgroups/ (features for grouping, isolation) | ||
342 | - Documentation/controllers/ (features for resource control) | ||
343 | |||
344 | Say N if unsure. | ||
345 | |||
346 | config CGROUP_DEBUG | ||
347 | bool "Example debug cgroup subsystem" | ||
348 | depends on CGROUPS | ||
349 | default n | ||
350 | help | ||
351 | This option enables a simple cgroup subsystem that | ||
352 | exports useful debugging information about the cgroups | ||
353 | framework | ||
354 | |||
355 | Say N if unsure | ||
356 | |||
357 | config CGROUP_NS | ||
358 | bool "Namespace cgroup subsystem" | ||
359 | depends on CGROUPS | ||
360 | help | ||
361 | Provides a simple namespace cgroup subsystem to | ||
362 | provide hierarchical naming of sets of namespaces, | ||
363 | for instance virtual servers and checkpoint/restart | ||
364 | jobs. | ||
365 | |||
366 | config CGROUP_FREEZER | ||
367 | bool "control group freezer subsystem" | ||
368 | depends on CGROUPS | ||
369 | help | ||
370 | Provides a way to freeze and unfreeze all tasks in a | ||
371 | cgroup. | ||
372 | |||
373 | config CGROUP_DEVICE | ||
374 | bool "Device controller for cgroups" | ||
375 | depends on CGROUPS && EXPERIMENTAL | ||
376 | help | ||
377 | Provides a cgroup implementing whitelists for devices which | ||
378 | a process in the cgroup can mknod or open. | ||
379 | |||
380 | config CPUSETS | ||
381 | bool "Cpuset support" | ||
382 | depends on SMP && CGROUPS | ||
383 | help | ||
384 | This option will let you create and manage CPUSETs which | ||
385 | allow dynamically partitioning a system into sets of CPUs and | ||
386 | Memory Nodes and assigning tasks to run only within those sets. | ||
387 | This is primarily useful on large SMP or NUMA systems. | ||
388 | |||
389 | Say N if unsure. | ||
390 | |||
382 | config CGROUP_CPUACCT | 391 | config CGROUP_CPUACCT |
383 | bool "Simple CPU accounting cgroup subsystem" | 392 | bool "Simple CPU accounting cgroup subsystem" |
384 | depends on CGROUPS | 393 | depends on CGROUPS |
@@ -393,9 +402,6 @@ config RESOURCE_COUNTERS | |||
393 | infrastructure that works with cgroups | 402 | infrastructure that works with cgroups |
394 | depends on CGROUPS | 403 | depends on CGROUPS |
395 | 404 | ||
396 | config MM_OWNER | ||
397 | bool | ||
398 | |||
399 | config CGROUP_MEM_RES_CTLR | 405 | config CGROUP_MEM_RES_CTLR |
400 | bool "Memory Resource Controller for Control Groups" | 406 | bool "Memory Resource Controller for Control Groups" |
401 | depends on CGROUPS && RESOURCE_COUNTERS | 407 | depends on CGROUPS && RESOURCE_COUNTERS |
@@ -414,11 +420,33 @@ config CGROUP_MEM_RES_CTLR | |||
414 | sure you need the memory resource controller. Even when you enable | 420 | sure you need the memory resource controller. Even when you enable |
415 | this, you can set "cgroup_disable=memory" at your boot option to | 421 | this, you can set "cgroup_disable=memory" at your boot option to |
416 | disable memory resource controller and you can avoid overheads. | 422 | disable memory resource controller and you can avoid overheads. |
417 | (and lose benefits of memory resource contoller) | 423 | (and lose benefits of memory resource controller) |
418 | 424 | ||
419 | This config option also selects MM_OWNER config option, which | 425 | This config option also selects MM_OWNER config option, which |
420 | could in turn add some fork/exit overhead. | 426 | could in turn add some fork/exit overhead. |
421 | 427 | ||
428 | config MM_OWNER | ||
429 | bool | ||
430 | |||
431 | config CGROUP_MEM_RES_CTLR_SWAP | ||
432 | bool "Memory Resource Controller Swap Extension(EXPERIMENTAL)" | ||
433 | depends on CGROUP_MEM_RES_CTLR && SWAP && EXPERIMENTAL | ||
434 | help | ||
435 | Add swap management feature to memory resource controller. When you | ||
436 | enable this, you can limit mem+swap usage per cgroup. In other words, | ||
437 | when you disable this, memory resource controller has no cares to | ||
438 | usage of swap...a process can exhaust all of the swap. This extension | ||
439 | is useful when you want to avoid exhaustion swap but this itself | ||
440 | adds more overheads and consumes memory for remembering information. | ||
441 | Especially if you use 32bit system or small memory system, please | ||
442 | be careful about enabling this. When memory resource controller | ||
443 | is disabled by boot option, this will be automatically disabled and | ||
444 | there will be no overhead from this. Even when you set this config=y, | ||
445 | if boot option "noswapaccount" is set, swap will not be accounted. | ||
446 | |||
447 | |||
448 | endmenu | ||
449 | |||
422 | config SYSFS_DEPRECATED | 450 | config SYSFS_DEPRECATED |
423 | bool | 451 | bool |
424 | 452 | ||
diff --git a/ipc/mqueue.c b/ipc/mqueue.c index eddb6247a553..23fdb8492b8e 100644 --- a/ipc/mqueue.c +++ b/ipc/mqueue.c | |||
@@ -505,7 +505,8 @@ static void __do_notify(struct mqueue_inode_info *info) | |||
505 | sig_i.si_errno = 0; | 505 | sig_i.si_errno = 0; |
506 | sig_i.si_code = SI_MESGQ; | 506 | sig_i.si_code = SI_MESGQ; |
507 | sig_i.si_value = info->notify.sigev_value; | 507 | sig_i.si_value = info->notify.sigev_value; |
508 | sig_i.si_pid = task_tgid_vnr(current); | 508 | sig_i.si_pid = task_tgid_nr_ns(current, |
509 | ns_of_pid(info->notify_owner)); | ||
509 | sig_i.si_uid = current_uid(); | 510 | sig_i.si_uid = current_uid(); |
510 | 511 | ||
511 | kill_pid_info(info->notify.sigev_signo, | 512 | kill_pid_info(info->notify.sigev_signo, |
diff --git a/kernel/async.c b/kernel/async.c index 97373380c9e7..64cc916299a5 100644 --- a/kernel/async.c +++ b/kernel/async.c | |||
@@ -206,7 +206,9 @@ EXPORT_SYMBOL_GPL(async_schedule_special); | |||
206 | 206 | ||
207 | void async_synchronize_full(void) | 207 | void async_synchronize_full(void) |
208 | { | 208 | { |
209 | async_synchronize_cookie(next_cookie); | 209 | do { |
210 | async_synchronize_cookie(next_cookie); | ||
211 | } while (!list_empty(&async_running) || !list_empty(&async_pending)); | ||
210 | } | 212 | } |
211 | EXPORT_SYMBOL_GPL(async_synchronize_full); | 213 | EXPORT_SYMBOL_GPL(async_synchronize_full); |
212 | 214 | ||
diff --git a/kernel/cgroup.c b/kernel/cgroup.c index f221446aa02d..c29831076e7a 100644 --- a/kernel/cgroup.c +++ b/kernel/cgroup.c | |||
@@ -84,7 +84,7 @@ struct cgroupfs_root { | |||
84 | /* Tracks how many cgroups are currently defined in hierarchy.*/ | 84 | /* Tracks how many cgroups are currently defined in hierarchy.*/ |
85 | int number_of_cgroups; | 85 | int number_of_cgroups; |
86 | 86 | ||
87 | /* A list running through the mounted hierarchies */ | 87 | /* A list running through the active hierarchies */ |
88 | struct list_head root_list; | 88 | struct list_head root_list; |
89 | 89 | ||
90 | /* Hierarchy-specific flags */ | 90 | /* Hierarchy-specific flags */ |
@@ -148,8 +148,8 @@ static int notify_on_release(const struct cgroup *cgrp) | |||
148 | #define for_each_subsys(_root, _ss) \ | 148 | #define for_each_subsys(_root, _ss) \ |
149 | list_for_each_entry(_ss, &_root->subsys_list, sibling) | 149 | list_for_each_entry(_ss, &_root->subsys_list, sibling) |
150 | 150 | ||
151 | /* for_each_root() allows you to iterate across the active hierarchies */ | 151 | /* for_each_active_root() allows you to iterate across the active hierarchies */ |
152 | #define for_each_root(_root) \ | 152 | #define for_each_active_root(_root) \ |
153 | list_for_each_entry(_root, &roots, root_list) | 153 | list_for_each_entry(_root, &roots, root_list) |
154 | 154 | ||
155 | /* the list of cgroups eligible for automatic release. Protected by | 155 | /* the list of cgroups eligible for automatic release. Protected by |
@@ -271,7 +271,7 @@ static void __put_css_set(struct css_set *cg, int taskexit) | |||
271 | 271 | ||
272 | rcu_read_lock(); | 272 | rcu_read_lock(); |
273 | for (i = 0; i < CGROUP_SUBSYS_COUNT; i++) { | 273 | for (i = 0; i < CGROUP_SUBSYS_COUNT; i++) { |
274 | struct cgroup *cgrp = cg->subsys[i]->cgroup; | 274 | struct cgroup *cgrp = rcu_dereference(cg->subsys[i]->cgroup); |
275 | if (atomic_dec_and_test(&cgrp->count) && | 275 | if (atomic_dec_and_test(&cgrp->count) && |
276 | notify_on_release(cgrp)) { | 276 | notify_on_release(cgrp)) { |
277 | if (taskexit) | 277 | if (taskexit) |
@@ -384,6 +384,25 @@ static int allocate_cg_links(int count, struct list_head *tmp) | |||
384 | return 0; | 384 | return 0; |
385 | } | 385 | } |
386 | 386 | ||
387 | /** | ||
388 | * link_css_set - a helper function to link a css_set to a cgroup | ||
389 | * @tmp_cg_links: cg_cgroup_link objects allocated by allocate_cg_links() | ||
390 | * @cg: the css_set to be linked | ||
391 | * @cgrp: the destination cgroup | ||
392 | */ | ||
393 | static void link_css_set(struct list_head *tmp_cg_links, | ||
394 | struct css_set *cg, struct cgroup *cgrp) | ||
395 | { | ||
396 | struct cg_cgroup_link *link; | ||
397 | |||
398 | BUG_ON(list_empty(tmp_cg_links)); | ||
399 | link = list_first_entry(tmp_cg_links, struct cg_cgroup_link, | ||
400 | cgrp_link_list); | ||
401 | link->cg = cg; | ||
402 | list_move(&link->cgrp_link_list, &cgrp->css_sets); | ||
403 | list_add(&link->cg_link_list, &cg->cg_links); | ||
404 | } | ||
405 | |||
387 | /* | 406 | /* |
388 | * find_css_set() takes an existing cgroup group and a | 407 | * find_css_set() takes an existing cgroup group and a |
389 | * cgroup object, and returns a css_set object that's | 408 | * cgroup object, and returns a css_set object that's |
@@ -399,7 +418,6 @@ static struct css_set *find_css_set( | |||
399 | int i; | 418 | int i; |
400 | 419 | ||
401 | struct list_head tmp_cg_links; | 420 | struct list_head tmp_cg_links; |
402 | struct cg_cgroup_link *link; | ||
403 | 421 | ||
404 | struct hlist_head *hhead; | 422 | struct hlist_head *hhead; |
405 | 423 | ||
@@ -444,26 +462,11 @@ static struct css_set *find_css_set( | |||
444 | * only do it for the first subsystem in each | 462 | * only do it for the first subsystem in each |
445 | * hierarchy | 463 | * hierarchy |
446 | */ | 464 | */ |
447 | if (ss->root->subsys_list.next == &ss->sibling) { | 465 | if (ss->root->subsys_list.next == &ss->sibling) |
448 | BUG_ON(list_empty(&tmp_cg_links)); | 466 | link_css_set(&tmp_cg_links, res, cgrp); |
449 | link = list_entry(tmp_cg_links.next, | ||
450 | struct cg_cgroup_link, | ||
451 | cgrp_link_list); | ||
452 | list_del(&link->cgrp_link_list); | ||
453 | list_add(&link->cgrp_link_list, &cgrp->css_sets); | ||
454 | link->cg = res; | ||
455 | list_add(&link->cg_link_list, &res->cg_links); | ||
456 | } | ||
457 | } | ||
458 | if (list_empty(&rootnode.subsys_list)) { | ||
459 | link = list_entry(tmp_cg_links.next, | ||
460 | struct cg_cgroup_link, | ||
461 | cgrp_link_list); | ||
462 | list_del(&link->cgrp_link_list); | ||
463 | list_add(&link->cgrp_link_list, &dummytop->css_sets); | ||
464 | link->cg = res; | ||
465 | list_add(&link->cg_link_list, &res->cg_links); | ||
466 | } | 467 | } |
468 | if (list_empty(&rootnode.subsys_list)) | ||
469 | link_css_set(&tmp_cg_links, res, dummytop); | ||
467 | 470 | ||
468 | BUG_ON(!list_empty(&tmp_cg_links)); | 471 | BUG_ON(!list_empty(&tmp_cg_links)); |
469 | 472 | ||
@@ -586,11 +589,18 @@ static void cgroup_call_pre_destroy(struct cgroup *cgrp) | |||
586 | { | 589 | { |
587 | struct cgroup_subsys *ss; | 590 | struct cgroup_subsys *ss; |
588 | for_each_subsys(cgrp->root, ss) | 591 | for_each_subsys(cgrp->root, ss) |
589 | if (ss->pre_destroy && cgrp->subsys[ss->subsys_id]) | 592 | if (ss->pre_destroy) |
590 | ss->pre_destroy(ss, cgrp); | 593 | ss->pre_destroy(ss, cgrp); |
591 | return; | 594 | return; |
592 | } | 595 | } |
593 | 596 | ||
597 | static void free_cgroup_rcu(struct rcu_head *obj) | ||
598 | { | ||
599 | struct cgroup *cgrp = container_of(obj, struct cgroup, rcu_head); | ||
600 | |||
601 | kfree(cgrp); | ||
602 | } | ||
603 | |||
594 | static void cgroup_diput(struct dentry *dentry, struct inode *inode) | 604 | static void cgroup_diput(struct dentry *dentry, struct inode *inode) |
595 | { | 605 | { |
596 | /* is dentry a directory ? if so, kfree() associated cgroup */ | 606 | /* is dentry a directory ? if so, kfree() associated cgroup */ |
@@ -610,19 +620,19 @@ static void cgroup_diput(struct dentry *dentry, struct inode *inode) | |||
610 | /* | 620 | /* |
611 | * Release the subsystem state objects. | 621 | * Release the subsystem state objects. |
612 | */ | 622 | */ |
613 | for_each_subsys(cgrp->root, ss) { | 623 | for_each_subsys(cgrp->root, ss) |
614 | if (cgrp->subsys[ss->subsys_id]) | 624 | ss->destroy(ss, cgrp); |
615 | ss->destroy(ss, cgrp); | ||
616 | } | ||
617 | 625 | ||
618 | cgrp->root->number_of_cgroups--; | 626 | cgrp->root->number_of_cgroups--; |
619 | mutex_unlock(&cgroup_mutex); | 627 | mutex_unlock(&cgroup_mutex); |
620 | 628 | ||
621 | /* Drop the active superblock reference that we took when we | 629 | /* |
622 | * created the cgroup */ | 630 | * Drop the active superblock reference that we took when we |
631 | * created the cgroup | ||
632 | */ | ||
623 | deactivate_super(cgrp->root->sb); | 633 | deactivate_super(cgrp->root->sb); |
624 | 634 | ||
625 | kfree(cgrp); | 635 | call_rcu(&cgrp->rcu_head, free_cgroup_rcu); |
626 | } | 636 | } |
627 | iput(inode); | 637 | iput(inode); |
628 | } | 638 | } |
@@ -712,23 +722,26 @@ static int rebind_subsystems(struct cgroupfs_root *root, | |||
712 | BUG_ON(cgrp->subsys[i]); | 722 | BUG_ON(cgrp->subsys[i]); |
713 | BUG_ON(!dummytop->subsys[i]); | 723 | BUG_ON(!dummytop->subsys[i]); |
714 | BUG_ON(dummytop->subsys[i]->cgroup != dummytop); | 724 | BUG_ON(dummytop->subsys[i]->cgroup != dummytop); |
725 | mutex_lock(&ss->hierarchy_mutex); | ||
715 | cgrp->subsys[i] = dummytop->subsys[i]; | 726 | cgrp->subsys[i] = dummytop->subsys[i]; |
716 | cgrp->subsys[i]->cgroup = cgrp; | 727 | cgrp->subsys[i]->cgroup = cgrp; |
717 | list_add(&ss->sibling, &root->subsys_list); | 728 | list_move(&ss->sibling, &root->subsys_list); |
718 | rcu_assign_pointer(ss->root, root); | 729 | ss->root = root; |
719 | if (ss->bind) | 730 | if (ss->bind) |
720 | ss->bind(ss, cgrp); | 731 | ss->bind(ss, cgrp); |
721 | 732 | mutex_unlock(&ss->hierarchy_mutex); | |
722 | } else if (bit & removed_bits) { | 733 | } else if (bit & removed_bits) { |
723 | /* We're removing this subsystem */ | 734 | /* We're removing this subsystem */ |
724 | BUG_ON(cgrp->subsys[i] != dummytop->subsys[i]); | 735 | BUG_ON(cgrp->subsys[i] != dummytop->subsys[i]); |
725 | BUG_ON(cgrp->subsys[i]->cgroup != cgrp); | 736 | BUG_ON(cgrp->subsys[i]->cgroup != cgrp); |
737 | mutex_lock(&ss->hierarchy_mutex); | ||
726 | if (ss->bind) | 738 | if (ss->bind) |
727 | ss->bind(ss, dummytop); | 739 | ss->bind(ss, dummytop); |
728 | dummytop->subsys[i]->cgroup = dummytop; | 740 | dummytop->subsys[i]->cgroup = dummytop; |
729 | cgrp->subsys[i] = NULL; | 741 | cgrp->subsys[i] = NULL; |
730 | rcu_assign_pointer(subsys[i]->root, &rootnode); | 742 | subsys[i]->root = &rootnode; |
731 | list_del(&ss->sibling); | 743 | list_move(&ss->sibling, &rootnode.subsys_list); |
744 | mutex_unlock(&ss->hierarchy_mutex); | ||
732 | } else if (bit & final_bits) { | 745 | } else if (bit & final_bits) { |
733 | /* Subsystem state should already exist */ | 746 | /* Subsystem state should already exist */ |
734 | BUG_ON(!cgrp->subsys[i]); | 747 | BUG_ON(!cgrp->subsys[i]); |
@@ -990,7 +1003,7 @@ static int cgroup_get_sb(struct file_system_type *fs_type, | |||
990 | root = NULL; | 1003 | root = NULL; |
991 | } else { | 1004 | } else { |
992 | /* New superblock */ | 1005 | /* New superblock */ |
993 | struct cgroup *cgrp = &root->top_cgroup; | 1006 | struct cgroup *root_cgrp = &root->top_cgroup; |
994 | struct inode *inode; | 1007 | struct inode *inode; |
995 | int i; | 1008 | int i; |
996 | 1009 | ||
@@ -1031,7 +1044,7 @@ static int cgroup_get_sb(struct file_system_type *fs_type, | |||
1031 | list_add(&root->root_list, &roots); | 1044 | list_add(&root->root_list, &roots); |
1032 | root_count++; | 1045 | root_count++; |
1033 | 1046 | ||
1034 | sb->s_root->d_fsdata = &root->top_cgroup; | 1047 | sb->s_root->d_fsdata = root_cgrp; |
1035 | root->top_cgroup.dentry = sb->s_root; | 1048 | root->top_cgroup.dentry = sb->s_root; |
1036 | 1049 | ||
1037 | /* Link the top cgroup in this hierarchy into all | 1050 | /* Link the top cgroup in this hierarchy into all |
@@ -1042,29 +1055,18 @@ static int cgroup_get_sb(struct file_system_type *fs_type, | |||
1042 | struct hlist_node *node; | 1055 | struct hlist_node *node; |
1043 | struct css_set *cg; | 1056 | struct css_set *cg; |
1044 | 1057 | ||
1045 | hlist_for_each_entry(cg, node, hhead, hlist) { | 1058 | hlist_for_each_entry(cg, node, hhead, hlist) |
1046 | struct cg_cgroup_link *link; | 1059 | link_css_set(&tmp_cg_links, cg, root_cgrp); |
1047 | |||
1048 | BUG_ON(list_empty(&tmp_cg_links)); | ||
1049 | link = list_entry(tmp_cg_links.next, | ||
1050 | struct cg_cgroup_link, | ||
1051 | cgrp_link_list); | ||
1052 | list_del(&link->cgrp_link_list); | ||
1053 | link->cg = cg; | ||
1054 | list_add(&link->cgrp_link_list, | ||
1055 | &root->top_cgroup.css_sets); | ||
1056 | list_add(&link->cg_link_list, &cg->cg_links); | ||
1057 | } | ||
1058 | } | 1060 | } |
1059 | write_unlock(&css_set_lock); | 1061 | write_unlock(&css_set_lock); |
1060 | 1062 | ||
1061 | free_cg_links(&tmp_cg_links); | 1063 | free_cg_links(&tmp_cg_links); |
1062 | 1064 | ||
1063 | BUG_ON(!list_empty(&cgrp->sibling)); | 1065 | BUG_ON(!list_empty(&root_cgrp->sibling)); |
1064 | BUG_ON(!list_empty(&cgrp->children)); | 1066 | BUG_ON(!list_empty(&root_cgrp->children)); |
1065 | BUG_ON(root->number_of_cgroups != 1); | 1067 | BUG_ON(root->number_of_cgroups != 1); |
1066 | 1068 | ||
1067 | cgroup_populate_dir(cgrp); | 1069 | cgroup_populate_dir(root_cgrp); |
1068 | mutex_unlock(&inode->i_mutex); | 1070 | mutex_unlock(&inode->i_mutex); |
1069 | mutex_unlock(&cgroup_mutex); | 1071 | mutex_unlock(&cgroup_mutex); |
1070 | } | 1072 | } |
@@ -1113,10 +1115,9 @@ static void cgroup_kill_sb(struct super_block *sb) { | |||
1113 | } | 1115 | } |
1114 | write_unlock(&css_set_lock); | 1116 | write_unlock(&css_set_lock); |
1115 | 1117 | ||
1116 | if (!list_empty(&root->root_list)) { | 1118 | list_del(&root->root_list); |
1117 | list_del(&root->root_list); | 1119 | root_count--; |
1118 | root_count--; | 1120 | |
1119 | } | ||
1120 | mutex_unlock(&cgroup_mutex); | 1121 | mutex_unlock(&cgroup_mutex); |
1121 | 1122 | ||
1122 | kfree(root); | 1123 | kfree(root); |
@@ -1145,14 +1146,16 @@ static inline struct cftype *__d_cft(struct dentry *dentry) | |||
1145 | * @buf: the buffer to write the path into | 1146 | * @buf: the buffer to write the path into |
1146 | * @buflen: the length of the buffer | 1147 | * @buflen: the length of the buffer |
1147 | * | 1148 | * |
1148 | * Called with cgroup_mutex held. Writes path of cgroup into buf. | 1149 | * Called with cgroup_mutex held or else with an RCU-protected cgroup |
1149 | * Returns 0 on success, -errno on error. | 1150 | * reference. Writes path of cgroup into buf. Returns 0 on success, |
1151 | * -errno on error. | ||
1150 | */ | 1152 | */ |
1151 | int cgroup_path(const struct cgroup *cgrp, char *buf, int buflen) | 1153 | int cgroup_path(const struct cgroup *cgrp, char *buf, int buflen) |
1152 | { | 1154 | { |
1153 | char *start; | 1155 | char *start; |
1156 | struct dentry *dentry = rcu_dereference(cgrp->dentry); | ||
1154 | 1157 | ||
1155 | if (cgrp == dummytop) { | 1158 | if (!dentry || cgrp == dummytop) { |
1156 | /* | 1159 | /* |
1157 | * Inactive subsystems have no dentry for their root | 1160 | * Inactive subsystems have no dentry for their root |
1158 | * cgroup | 1161 | * cgroup |
@@ -1165,13 +1168,14 @@ int cgroup_path(const struct cgroup *cgrp, char *buf, int buflen) | |||
1165 | 1168 | ||
1166 | *--start = '\0'; | 1169 | *--start = '\0'; |
1167 | for (;;) { | 1170 | for (;;) { |
1168 | int len = cgrp->dentry->d_name.len; | 1171 | int len = dentry->d_name.len; |
1169 | if ((start -= len) < buf) | 1172 | if ((start -= len) < buf) |
1170 | return -ENAMETOOLONG; | 1173 | return -ENAMETOOLONG; |
1171 | memcpy(start, cgrp->dentry->d_name.name, len); | 1174 | memcpy(start, cgrp->dentry->d_name.name, len); |
1172 | cgrp = cgrp->parent; | 1175 | cgrp = cgrp->parent; |
1173 | if (!cgrp) | 1176 | if (!cgrp) |
1174 | break; | 1177 | break; |
1178 | dentry = rcu_dereference(cgrp->dentry); | ||
1175 | if (!cgrp->parent) | 1179 | if (!cgrp->parent) |
1176 | continue; | 1180 | continue; |
1177 | if (--start < buf) | 1181 | if (--start < buf) |
@@ -1216,7 +1220,7 @@ int cgroup_attach_task(struct cgroup *cgrp, struct task_struct *tsk) | |||
1216 | int retval = 0; | 1220 | int retval = 0; |
1217 | struct cgroup_subsys *ss; | 1221 | struct cgroup_subsys *ss; |
1218 | struct cgroup *oldcgrp; | 1222 | struct cgroup *oldcgrp; |
1219 | struct css_set *cg = tsk->cgroups; | 1223 | struct css_set *cg; |
1220 | struct css_set *newcg; | 1224 | struct css_set *newcg; |
1221 | struct cgroupfs_root *root = cgrp->root; | 1225 | struct cgroupfs_root *root = cgrp->root; |
1222 | int subsys_id; | 1226 | int subsys_id; |
@@ -1236,11 +1240,16 @@ int cgroup_attach_task(struct cgroup *cgrp, struct task_struct *tsk) | |||
1236 | } | 1240 | } |
1237 | } | 1241 | } |
1238 | 1242 | ||
1243 | task_lock(tsk); | ||
1244 | cg = tsk->cgroups; | ||
1245 | get_css_set(cg); | ||
1246 | task_unlock(tsk); | ||
1239 | /* | 1247 | /* |
1240 | * Locate or allocate a new css_set for this task, | 1248 | * Locate or allocate a new css_set for this task, |
1241 | * based on its final set of cgroups | 1249 | * based on its final set of cgroups |
1242 | */ | 1250 | */ |
1243 | newcg = find_css_set(cg, cgrp); | 1251 | newcg = find_css_set(cg, cgrp); |
1252 | put_css_set(cg); | ||
1244 | if (!newcg) | 1253 | if (!newcg) |
1245 | return -ENOMEM; | 1254 | return -ENOMEM; |
1246 | 1255 | ||
@@ -1445,7 +1454,7 @@ static ssize_t cgroup_file_write(struct file *file, const char __user *buf, | |||
1445 | struct cftype *cft = __d_cft(file->f_dentry); | 1454 | struct cftype *cft = __d_cft(file->f_dentry); |
1446 | struct cgroup *cgrp = __d_cgrp(file->f_dentry->d_parent); | 1455 | struct cgroup *cgrp = __d_cgrp(file->f_dentry->d_parent); |
1447 | 1456 | ||
1448 | if (!cft || cgroup_is_removed(cgrp)) | 1457 | if (cgroup_is_removed(cgrp)) |
1449 | return -ENODEV; | 1458 | return -ENODEV; |
1450 | if (cft->write) | 1459 | if (cft->write) |
1451 | return cft->write(cgrp, cft, file, buf, nbytes, ppos); | 1460 | return cft->write(cgrp, cft, file, buf, nbytes, ppos); |
@@ -1490,7 +1499,7 @@ static ssize_t cgroup_file_read(struct file *file, char __user *buf, | |||
1490 | struct cftype *cft = __d_cft(file->f_dentry); | 1499 | struct cftype *cft = __d_cft(file->f_dentry); |
1491 | struct cgroup *cgrp = __d_cgrp(file->f_dentry->d_parent); | 1500 | struct cgroup *cgrp = __d_cgrp(file->f_dentry->d_parent); |
1492 | 1501 | ||
1493 | if (!cft || cgroup_is_removed(cgrp)) | 1502 | if (cgroup_is_removed(cgrp)) |
1494 | return -ENODEV; | 1503 | return -ENODEV; |
1495 | 1504 | ||
1496 | if (cft->read) | 1505 | if (cft->read) |
@@ -1554,10 +1563,8 @@ static int cgroup_file_open(struct inode *inode, struct file *file) | |||
1554 | err = generic_file_open(inode, file); | 1563 | err = generic_file_open(inode, file); |
1555 | if (err) | 1564 | if (err) |
1556 | return err; | 1565 | return err; |
1557 | |||
1558 | cft = __d_cft(file->f_dentry); | 1566 | cft = __d_cft(file->f_dentry); |
1559 | if (!cft) | 1567 | |
1560 | return -ENODEV; | ||
1561 | if (cft->read_map || cft->read_seq_string) { | 1568 | if (cft->read_map || cft->read_seq_string) { |
1562 | struct cgroup_seqfile_state *state = | 1569 | struct cgroup_seqfile_state *state = |
1563 | kzalloc(sizeof(*state), GFP_USER); | 1570 | kzalloc(sizeof(*state), GFP_USER); |
@@ -1671,7 +1678,7 @@ static int cgroup_create_dir(struct cgroup *cgrp, struct dentry *dentry, | |||
1671 | if (!error) { | 1678 | if (!error) { |
1672 | dentry->d_fsdata = cgrp; | 1679 | dentry->d_fsdata = cgrp; |
1673 | inc_nlink(parent->d_inode); | 1680 | inc_nlink(parent->d_inode); |
1674 | cgrp->dentry = dentry; | 1681 | rcu_assign_pointer(cgrp->dentry, dentry); |
1675 | dget(dentry); | 1682 | dget(dentry); |
1676 | } | 1683 | } |
1677 | dput(dentry); | 1684 | dput(dentry); |
@@ -1812,6 +1819,7 @@ struct task_struct *cgroup_iter_next(struct cgroup *cgrp, | |||
1812 | { | 1819 | { |
1813 | struct task_struct *res; | 1820 | struct task_struct *res; |
1814 | struct list_head *l = it->task; | 1821 | struct list_head *l = it->task; |
1822 | struct cg_cgroup_link *link; | ||
1815 | 1823 | ||
1816 | /* If the iterator cg is NULL, we have no tasks */ | 1824 | /* If the iterator cg is NULL, we have no tasks */ |
1817 | if (!it->cg_link) | 1825 | if (!it->cg_link) |
@@ -1819,7 +1827,8 @@ struct task_struct *cgroup_iter_next(struct cgroup *cgrp, | |||
1819 | res = list_entry(l, struct task_struct, cg_list); | 1827 | res = list_entry(l, struct task_struct, cg_list); |
1820 | /* Advance iterator to find next entry */ | 1828 | /* Advance iterator to find next entry */ |
1821 | l = l->next; | 1829 | l = l->next; |
1822 | if (l == &res->cgroups->tasks) { | 1830 | link = list_entry(it->cg_link, struct cg_cgroup_link, cgrp_link_list); |
1831 | if (l == &link->cg->tasks) { | ||
1823 | /* We reached the end of this task list - move on to | 1832 | /* We reached the end of this task list - move on to |
1824 | * the next cg_cgroup_link */ | 1833 | * the next cg_cgroup_link */ |
1825 | cgroup_advance_iter(cgrp, it); | 1834 | cgroup_advance_iter(cgrp, it); |
@@ -2013,14 +2022,16 @@ int cgroup_scan_tasks(struct cgroup_scanner *scan) | |||
2013 | */ | 2022 | */ |
2014 | static int pid_array_load(pid_t *pidarray, int npids, struct cgroup *cgrp) | 2023 | static int pid_array_load(pid_t *pidarray, int npids, struct cgroup *cgrp) |
2015 | { | 2024 | { |
2016 | int n = 0; | 2025 | int n = 0, pid; |
2017 | struct cgroup_iter it; | 2026 | struct cgroup_iter it; |
2018 | struct task_struct *tsk; | 2027 | struct task_struct *tsk; |
2019 | cgroup_iter_start(cgrp, &it); | 2028 | cgroup_iter_start(cgrp, &it); |
2020 | while ((tsk = cgroup_iter_next(cgrp, &it))) { | 2029 | while ((tsk = cgroup_iter_next(cgrp, &it))) { |
2021 | if (unlikely(n == npids)) | 2030 | if (unlikely(n == npids)) |
2022 | break; | 2031 | break; |
2023 | pidarray[n++] = task_pid_vnr(tsk); | 2032 | pid = task_pid_vnr(tsk); |
2033 | if (pid > 0) | ||
2034 | pidarray[n++] = pid; | ||
2024 | } | 2035 | } |
2025 | cgroup_iter_end(cgrp, &it); | 2036 | cgroup_iter_end(cgrp, &it); |
2026 | return n; | 2037 | return n; |
@@ -2052,7 +2063,6 @@ int cgroupstats_build(struct cgroupstats *stats, struct dentry *dentry) | |||
2052 | 2063 | ||
2053 | ret = 0; | 2064 | ret = 0; |
2054 | cgrp = dentry->d_fsdata; | 2065 | cgrp = dentry->d_fsdata; |
2055 | rcu_read_lock(); | ||
2056 | 2066 | ||
2057 | cgroup_iter_start(cgrp, &it); | 2067 | cgroup_iter_start(cgrp, &it); |
2058 | while ((tsk = cgroup_iter_next(cgrp, &it))) { | 2068 | while ((tsk = cgroup_iter_next(cgrp, &it))) { |
@@ -2077,7 +2087,6 @@ int cgroupstats_build(struct cgroupstats *stats, struct dentry *dentry) | |||
2077 | } | 2087 | } |
2078 | cgroup_iter_end(cgrp, &it); | 2088 | cgroup_iter_end(cgrp, &it); |
2079 | 2089 | ||
2080 | rcu_read_unlock(); | ||
2081 | err: | 2090 | err: |
2082 | return ret; | 2091 | return ret; |
2083 | } | 2092 | } |
@@ -2324,7 +2333,7 @@ static void init_cgroup_css(struct cgroup_subsys_state *css, | |||
2324 | struct cgroup *cgrp) | 2333 | struct cgroup *cgrp) |
2325 | { | 2334 | { |
2326 | css->cgroup = cgrp; | 2335 | css->cgroup = cgrp; |
2327 | atomic_set(&css->refcnt, 0); | 2336 | atomic_set(&css->refcnt, 1); |
2328 | css->flags = 0; | 2337 | css->flags = 0; |
2329 | if (cgrp == dummytop) | 2338 | if (cgrp == dummytop) |
2330 | set_bit(CSS_ROOT, &css->flags); | 2339 | set_bit(CSS_ROOT, &css->flags); |
@@ -2332,6 +2341,29 @@ static void init_cgroup_css(struct cgroup_subsys_state *css, | |||
2332 | cgrp->subsys[ss->subsys_id] = css; | 2341 | cgrp->subsys[ss->subsys_id] = css; |
2333 | } | 2342 | } |
2334 | 2343 | ||
2344 | static void cgroup_lock_hierarchy(struct cgroupfs_root *root) | ||
2345 | { | ||
2346 | /* We need to take each hierarchy_mutex in a consistent order */ | ||
2347 | int i; | ||
2348 | |||
2349 | for (i = 0; i < CGROUP_SUBSYS_COUNT; i++) { | ||
2350 | struct cgroup_subsys *ss = subsys[i]; | ||
2351 | if (ss->root == root) | ||
2352 | mutex_lock_nested(&ss->hierarchy_mutex, i); | ||
2353 | } | ||
2354 | } | ||
2355 | |||
2356 | static void cgroup_unlock_hierarchy(struct cgroupfs_root *root) | ||
2357 | { | ||
2358 | int i; | ||
2359 | |||
2360 | for (i = 0; i < CGROUP_SUBSYS_COUNT; i++) { | ||
2361 | struct cgroup_subsys *ss = subsys[i]; | ||
2362 | if (ss->root == root) | ||
2363 | mutex_unlock(&ss->hierarchy_mutex); | ||
2364 | } | ||
2365 | } | ||
2366 | |||
2335 | /* | 2367 | /* |
2336 | * cgroup_create - create a cgroup | 2368 | * cgroup_create - create a cgroup |
2337 | * @parent: cgroup that will be parent of the new cgroup | 2369 | * @parent: cgroup that will be parent of the new cgroup |
@@ -2380,7 +2412,9 @@ static long cgroup_create(struct cgroup *parent, struct dentry *dentry, | |||
2380 | init_cgroup_css(css, ss, cgrp); | 2412 | init_cgroup_css(css, ss, cgrp); |
2381 | } | 2413 | } |
2382 | 2414 | ||
2415 | cgroup_lock_hierarchy(root); | ||
2383 | list_add(&cgrp->sibling, &cgrp->parent->children); | 2416 | list_add(&cgrp->sibling, &cgrp->parent->children); |
2417 | cgroup_unlock_hierarchy(root); | ||
2384 | root->number_of_cgroups++; | 2418 | root->number_of_cgroups++; |
2385 | 2419 | ||
2386 | err = cgroup_create_dir(cgrp, dentry, mode); | 2420 | err = cgroup_create_dir(cgrp, dentry, mode); |
@@ -2431,7 +2465,7 @@ static int cgroup_has_css_refs(struct cgroup *cgrp) | |||
2431 | { | 2465 | { |
2432 | /* Check the reference count on each subsystem. Since we | 2466 | /* Check the reference count on each subsystem. Since we |
2433 | * already established that there are no tasks in the | 2467 | * already established that there are no tasks in the |
2434 | * cgroup, if the css refcount is also 0, then there should | 2468 | * cgroup, if the css refcount is also 1, then there should |
2435 | * be no outstanding references, so the subsystem is safe to | 2469 | * be no outstanding references, so the subsystem is safe to |
2436 | * destroy. We scan across all subsystems rather than using | 2470 | * destroy. We scan across all subsystems rather than using |
2437 | * the per-hierarchy linked list of mounted subsystems since | 2471 | * the per-hierarchy linked list of mounted subsystems since |
@@ -2452,19 +2486,67 @@ static int cgroup_has_css_refs(struct cgroup *cgrp) | |||
2452 | * matter, since it can only happen if the cgroup | 2486 | * matter, since it can only happen if the cgroup |
2453 | * has been deleted and hence no longer needs the | 2487 | * has been deleted and hence no longer needs the |
2454 | * release agent to be called anyway. */ | 2488 | * release agent to be called anyway. */ |
2455 | if (css && atomic_read(&css->refcnt)) | 2489 | if (css && (atomic_read(&css->refcnt) > 1)) |
2456 | return 1; | 2490 | return 1; |
2457 | } | 2491 | } |
2458 | return 0; | 2492 | return 0; |
2459 | } | 2493 | } |
2460 | 2494 | ||
2495 | /* | ||
2496 | * Atomically mark all (or else none) of the cgroup's CSS objects as | ||
2497 | * CSS_REMOVED. Return true on success, or false if the cgroup has | ||
2498 | * busy subsystems. Call with cgroup_mutex held | ||
2499 | */ | ||
2500 | |||
2501 | static int cgroup_clear_css_refs(struct cgroup *cgrp) | ||
2502 | { | ||
2503 | struct cgroup_subsys *ss; | ||
2504 | unsigned long flags; | ||
2505 | bool failed = false; | ||
2506 | local_irq_save(flags); | ||
2507 | for_each_subsys(cgrp->root, ss) { | ||
2508 | struct cgroup_subsys_state *css = cgrp->subsys[ss->subsys_id]; | ||
2509 | int refcnt; | ||
2510 | do { | ||
2511 | /* We can only remove a CSS with a refcnt==1 */ | ||
2512 | refcnt = atomic_read(&css->refcnt); | ||
2513 | if (refcnt > 1) { | ||
2514 | failed = true; | ||
2515 | goto done; | ||
2516 | } | ||
2517 | BUG_ON(!refcnt); | ||
2518 | /* | ||
2519 | * Drop the refcnt to 0 while we check other | ||
2520 | * subsystems. This will cause any racing | ||
2521 | * css_tryget() to spin until we set the | ||
2522 | * CSS_REMOVED bits or abort | ||
2523 | */ | ||
2524 | } while (atomic_cmpxchg(&css->refcnt, refcnt, 0) != refcnt); | ||
2525 | } | ||
2526 | done: | ||
2527 | for_each_subsys(cgrp->root, ss) { | ||
2528 | struct cgroup_subsys_state *css = cgrp->subsys[ss->subsys_id]; | ||
2529 | if (failed) { | ||
2530 | /* | ||
2531 | * Restore old refcnt if we previously managed | ||
2532 | * to clear it from 1 to 0 | ||
2533 | */ | ||
2534 | if (!atomic_read(&css->refcnt)) | ||
2535 | atomic_set(&css->refcnt, 1); | ||
2536 | } else { | ||
2537 | /* Commit the fact that the CSS is removed */ | ||
2538 | set_bit(CSS_REMOVED, &css->flags); | ||
2539 | } | ||
2540 | } | ||
2541 | local_irq_restore(flags); | ||
2542 | return !failed; | ||
2543 | } | ||
2544 | |||
2461 | static int cgroup_rmdir(struct inode *unused_dir, struct dentry *dentry) | 2545 | static int cgroup_rmdir(struct inode *unused_dir, struct dentry *dentry) |
2462 | { | 2546 | { |
2463 | struct cgroup *cgrp = dentry->d_fsdata; | 2547 | struct cgroup *cgrp = dentry->d_fsdata; |
2464 | struct dentry *d; | 2548 | struct dentry *d; |
2465 | struct cgroup *parent; | 2549 | struct cgroup *parent; |
2466 | struct super_block *sb; | ||
2467 | struct cgroupfs_root *root; | ||
2468 | 2550 | ||
2469 | /* the vfs holds both inode->i_mutex already */ | 2551 | /* the vfs holds both inode->i_mutex already */ |
2470 | 2552 | ||
@@ -2487,12 +2569,10 @@ static int cgroup_rmdir(struct inode *unused_dir, struct dentry *dentry) | |||
2487 | 2569 | ||
2488 | mutex_lock(&cgroup_mutex); | 2570 | mutex_lock(&cgroup_mutex); |
2489 | parent = cgrp->parent; | 2571 | parent = cgrp->parent; |
2490 | root = cgrp->root; | ||
2491 | sb = root->sb; | ||
2492 | 2572 | ||
2493 | if (atomic_read(&cgrp->count) | 2573 | if (atomic_read(&cgrp->count) |
2494 | || !list_empty(&cgrp->children) | 2574 | || !list_empty(&cgrp->children) |
2495 | || cgroup_has_css_refs(cgrp)) { | 2575 | || !cgroup_clear_css_refs(cgrp)) { |
2496 | mutex_unlock(&cgroup_mutex); | 2576 | mutex_unlock(&cgroup_mutex); |
2497 | return -EBUSY; | 2577 | return -EBUSY; |
2498 | } | 2578 | } |
@@ -2502,8 +2582,12 @@ static int cgroup_rmdir(struct inode *unused_dir, struct dentry *dentry) | |||
2502 | if (!list_empty(&cgrp->release_list)) | 2582 | if (!list_empty(&cgrp->release_list)) |
2503 | list_del(&cgrp->release_list); | 2583 | list_del(&cgrp->release_list); |
2504 | spin_unlock(&release_list_lock); | 2584 | spin_unlock(&release_list_lock); |
2505 | /* delete my sibling from parent->children */ | 2585 | |
2586 | cgroup_lock_hierarchy(cgrp->root); | ||
2587 | /* delete this cgroup from parent->children */ | ||
2506 | list_del(&cgrp->sibling); | 2588 | list_del(&cgrp->sibling); |
2589 | cgroup_unlock_hierarchy(cgrp->root); | ||
2590 | |||
2507 | spin_lock(&cgrp->dentry->d_lock); | 2591 | spin_lock(&cgrp->dentry->d_lock); |
2508 | d = dget(cgrp->dentry); | 2592 | d = dget(cgrp->dentry); |
2509 | spin_unlock(&d->d_lock); | 2593 | spin_unlock(&d->d_lock); |
@@ -2525,6 +2609,7 @@ static void __init cgroup_init_subsys(struct cgroup_subsys *ss) | |||
2525 | printk(KERN_INFO "Initializing cgroup subsys %s\n", ss->name); | 2609 | printk(KERN_INFO "Initializing cgroup subsys %s\n", ss->name); |
2526 | 2610 | ||
2527 | /* Create the top cgroup state for this subsystem */ | 2611 | /* Create the top cgroup state for this subsystem */ |
2612 | list_add(&ss->sibling, &rootnode.subsys_list); | ||
2528 | ss->root = &rootnode; | 2613 | ss->root = &rootnode; |
2529 | css = ss->create(ss, dummytop); | 2614 | css = ss->create(ss, dummytop); |
2530 | /* We don't handle early failures gracefully */ | 2615 | /* We don't handle early failures gracefully */ |
@@ -2544,6 +2629,7 @@ static void __init cgroup_init_subsys(struct cgroup_subsys *ss) | |||
2544 | * need to invoke fork callbacks here. */ | 2629 | * need to invoke fork callbacks here. */ |
2545 | BUG_ON(!list_empty(&init_task.tasks)); | 2630 | BUG_ON(!list_empty(&init_task.tasks)); |
2546 | 2631 | ||
2632 | mutex_init(&ss->hierarchy_mutex); | ||
2547 | ss->active = 1; | 2633 | ss->active = 1; |
2548 | } | 2634 | } |
2549 | 2635 | ||
@@ -2562,7 +2648,6 @@ int __init cgroup_init_early(void) | |||
2562 | INIT_HLIST_NODE(&init_css_set.hlist); | 2648 | INIT_HLIST_NODE(&init_css_set.hlist); |
2563 | css_set_count = 1; | 2649 | css_set_count = 1; |
2564 | init_cgroup_root(&rootnode); | 2650 | init_cgroup_root(&rootnode); |
2565 | list_add(&rootnode.root_list, &roots); | ||
2566 | root_count = 1; | 2651 | root_count = 1; |
2567 | init_task.cgroups = &init_css_set; | 2652 | init_task.cgroups = &init_css_set; |
2568 | 2653 | ||
@@ -2669,15 +2754,12 @@ static int proc_cgroup_show(struct seq_file *m, void *v) | |||
2669 | 2754 | ||
2670 | mutex_lock(&cgroup_mutex); | 2755 | mutex_lock(&cgroup_mutex); |
2671 | 2756 | ||
2672 | for_each_root(root) { | 2757 | for_each_active_root(root) { |
2673 | struct cgroup_subsys *ss; | 2758 | struct cgroup_subsys *ss; |
2674 | struct cgroup *cgrp; | 2759 | struct cgroup *cgrp; |
2675 | int subsys_id; | 2760 | int subsys_id; |
2676 | int count = 0; | 2761 | int count = 0; |
2677 | 2762 | ||
2678 | /* Skip this hierarchy if it has no active subsystems */ | ||
2679 | if (!root->actual_subsys_bits) | ||
2680 | continue; | ||
2681 | seq_printf(m, "%lu:", root->subsys_bits); | 2763 | seq_printf(m, "%lu:", root->subsys_bits); |
2682 | for_each_subsys(root, ss) | 2764 | for_each_subsys(root, ss) |
2683 | seq_printf(m, "%s%s", count++ ? "," : "", ss->name); | 2765 | seq_printf(m, "%s%s", count++ ? "," : "", ss->name); |
@@ -2800,8 +2882,10 @@ void cgroup_post_fork(struct task_struct *child) | |||
2800 | { | 2882 | { |
2801 | if (use_task_css_set_links) { | 2883 | if (use_task_css_set_links) { |
2802 | write_lock(&css_set_lock); | 2884 | write_lock(&css_set_lock); |
2885 | task_lock(child); | ||
2803 | if (list_empty(&child->cg_list)) | 2886 | if (list_empty(&child->cg_list)) |
2804 | list_add(&child->cg_list, &child->cgroups->tasks); | 2887 | list_add(&child->cg_list, &child->cgroups->tasks); |
2888 | task_unlock(child); | ||
2805 | write_unlock(&css_set_lock); | 2889 | write_unlock(&css_set_lock); |
2806 | } | 2890 | } |
2807 | } | 2891 | } |
@@ -2907,6 +2991,7 @@ int cgroup_clone(struct task_struct *tsk, struct cgroup_subsys *subsys, | |||
2907 | mutex_unlock(&cgroup_mutex); | 2991 | mutex_unlock(&cgroup_mutex); |
2908 | return 0; | 2992 | return 0; |
2909 | } | 2993 | } |
2994 | task_lock(tsk); | ||
2910 | cg = tsk->cgroups; | 2995 | cg = tsk->cgroups; |
2911 | parent = task_cgroup(tsk, subsys->subsys_id); | 2996 | parent = task_cgroup(tsk, subsys->subsys_id); |
2912 | 2997 | ||
@@ -2919,6 +3004,7 @@ int cgroup_clone(struct task_struct *tsk, struct cgroup_subsys *subsys, | |||
2919 | 3004 | ||
2920 | /* Keep the cgroup alive */ | 3005 | /* Keep the cgroup alive */ |
2921 | get_css_set(cg); | 3006 | get_css_set(cg); |
3007 | task_unlock(tsk); | ||
2922 | mutex_unlock(&cgroup_mutex); | 3008 | mutex_unlock(&cgroup_mutex); |
2923 | 3009 | ||
2924 | /* Now do the VFS work to create a cgroup */ | 3010 | /* Now do the VFS work to create a cgroup */ |
@@ -2937,7 +3023,7 @@ int cgroup_clone(struct task_struct *tsk, struct cgroup_subsys *subsys, | |||
2937 | } | 3023 | } |
2938 | 3024 | ||
2939 | /* Create the cgroup directory, which also creates the cgroup */ | 3025 | /* Create the cgroup directory, which also creates the cgroup */ |
2940 | ret = vfs_mkdir(inode, dentry, S_IFDIR | 0755); | 3026 | ret = vfs_mkdir(inode, dentry, 0755); |
2941 | child = __d_cgrp(dentry); | 3027 | child = __d_cgrp(dentry); |
2942 | dput(dentry); | 3028 | dput(dentry); |
2943 | if (ret) { | 3029 | if (ret) { |
@@ -2947,13 +3033,6 @@ int cgroup_clone(struct task_struct *tsk, struct cgroup_subsys *subsys, | |||
2947 | goto out_release; | 3033 | goto out_release; |
2948 | } | 3034 | } |
2949 | 3035 | ||
2950 | if (!child) { | ||
2951 | printk(KERN_INFO | ||
2952 | "Couldn't find new cgroup %s\n", nodename); | ||
2953 | ret = -ENOMEM; | ||
2954 | goto out_release; | ||
2955 | } | ||
2956 | |||
2957 | /* The cgroup now exists. Retake cgroup_mutex and check | 3036 | /* The cgroup now exists. Retake cgroup_mutex and check |
2958 | * that we're still in the same state that we thought we | 3037 | * that we're still in the same state that we thought we |
2959 | * were. */ | 3038 | * were. */ |
@@ -3049,7 +3128,8 @@ void __css_put(struct cgroup_subsys_state *css) | |||
3049 | { | 3128 | { |
3050 | struct cgroup *cgrp = css->cgroup; | 3129 | struct cgroup *cgrp = css->cgroup; |
3051 | rcu_read_lock(); | 3130 | rcu_read_lock(); |
3052 | if (atomic_dec_and_test(&css->refcnt) && notify_on_release(cgrp)) { | 3131 | if ((atomic_dec_return(&css->refcnt) == 1) && |
3132 | notify_on_release(cgrp)) { | ||
3053 | set_bit(CGRP_RELEASABLE, &cgrp->flags); | 3133 | set_bit(CGRP_RELEASABLE, &cgrp->flags); |
3054 | check_for_release(cgrp); | 3134 | check_for_release(cgrp); |
3055 | } | 3135 | } |
diff --git a/kernel/cpuset.c b/kernel/cpuset.c index 345ace5117de..647c77a88fcb 100644 --- a/kernel/cpuset.c +++ b/kernel/cpuset.c | |||
@@ -84,7 +84,7 @@ struct cpuset { | |||
84 | struct cgroup_subsys_state css; | 84 | struct cgroup_subsys_state css; |
85 | 85 | ||
86 | unsigned long flags; /* "unsigned long" so bitops work */ | 86 | unsigned long flags; /* "unsigned long" so bitops work */ |
87 | cpumask_t cpus_allowed; /* CPUs allowed to tasks in cpuset */ | 87 | cpumask_var_t cpus_allowed; /* CPUs allowed to tasks in cpuset */ |
88 | nodemask_t mems_allowed; /* Memory Nodes allowed to tasks */ | 88 | nodemask_t mems_allowed; /* Memory Nodes allowed to tasks */ |
89 | 89 | ||
90 | struct cpuset *parent; /* my parent */ | 90 | struct cpuset *parent; /* my parent */ |
@@ -195,8 +195,6 @@ static int cpuset_mems_generation; | |||
195 | 195 | ||
196 | static struct cpuset top_cpuset = { | 196 | static struct cpuset top_cpuset = { |
197 | .flags = ((1 << CS_CPU_EXCLUSIVE) | (1 << CS_MEM_EXCLUSIVE)), | 197 | .flags = ((1 << CS_CPU_EXCLUSIVE) | (1 << CS_MEM_EXCLUSIVE)), |
198 | .cpus_allowed = CPU_MASK_ALL, | ||
199 | .mems_allowed = NODE_MASK_ALL, | ||
200 | }; | 198 | }; |
201 | 199 | ||
202 | /* | 200 | /* |
@@ -278,7 +276,7 @@ static struct file_system_type cpuset_fs_type = { | |||
278 | }; | 276 | }; |
279 | 277 | ||
280 | /* | 278 | /* |
281 | * Return in *pmask the portion of a cpusets's cpus_allowed that | 279 | * Return in pmask the portion of a cpusets's cpus_allowed that |
282 | * are online. If none are online, walk up the cpuset hierarchy | 280 | * are online. If none are online, walk up the cpuset hierarchy |
283 | * until we find one that does have some online cpus. If we get | 281 | * until we find one that does have some online cpus. If we get |
284 | * all the way to the top and still haven't found any online cpus, | 282 | * all the way to the top and still haven't found any online cpus, |
@@ -291,15 +289,16 @@ static struct file_system_type cpuset_fs_type = { | |||
291 | * Call with callback_mutex held. | 289 | * Call with callback_mutex held. |
292 | */ | 290 | */ |
293 | 291 | ||
294 | static void guarantee_online_cpus(const struct cpuset *cs, cpumask_t *pmask) | 292 | static void guarantee_online_cpus(const struct cpuset *cs, |
293 | struct cpumask *pmask) | ||
295 | { | 294 | { |
296 | while (cs && !cpus_intersects(cs->cpus_allowed, cpu_online_map)) | 295 | while (cs && !cpumask_intersects(cs->cpus_allowed, cpu_online_mask)) |
297 | cs = cs->parent; | 296 | cs = cs->parent; |
298 | if (cs) | 297 | if (cs) |
299 | cpus_and(*pmask, cs->cpus_allowed, cpu_online_map); | 298 | cpumask_and(pmask, cs->cpus_allowed, cpu_online_mask); |
300 | else | 299 | else |
301 | *pmask = cpu_online_map; | 300 | cpumask_copy(pmask, cpu_online_mask); |
302 | BUG_ON(!cpus_intersects(*pmask, cpu_online_map)); | 301 | BUG_ON(!cpumask_intersects(pmask, cpu_online_mask)); |
303 | } | 302 | } |
304 | 303 | ||
305 | /* | 304 | /* |
@@ -375,14 +374,9 @@ void cpuset_update_task_memory_state(void) | |||
375 | struct task_struct *tsk = current; | 374 | struct task_struct *tsk = current; |
376 | struct cpuset *cs; | 375 | struct cpuset *cs; |
377 | 376 | ||
378 | if (task_cs(tsk) == &top_cpuset) { | 377 | rcu_read_lock(); |
379 | /* Don't need rcu for top_cpuset. It's never freed. */ | 378 | my_cpusets_mem_gen = task_cs(tsk)->mems_generation; |
380 | my_cpusets_mem_gen = top_cpuset.mems_generation; | 379 | rcu_read_unlock(); |
381 | } else { | ||
382 | rcu_read_lock(); | ||
383 | my_cpusets_mem_gen = task_cs(tsk)->mems_generation; | ||
384 | rcu_read_unlock(); | ||
385 | } | ||
386 | 380 | ||
387 | if (my_cpusets_mem_gen != tsk->cpuset_mems_generation) { | 381 | if (my_cpusets_mem_gen != tsk->cpuset_mems_generation) { |
388 | mutex_lock(&callback_mutex); | 382 | mutex_lock(&callback_mutex); |
@@ -414,12 +408,43 @@ void cpuset_update_task_memory_state(void) | |||
414 | 408 | ||
415 | static int is_cpuset_subset(const struct cpuset *p, const struct cpuset *q) | 409 | static int is_cpuset_subset(const struct cpuset *p, const struct cpuset *q) |
416 | { | 410 | { |
417 | return cpus_subset(p->cpus_allowed, q->cpus_allowed) && | 411 | return cpumask_subset(p->cpus_allowed, q->cpus_allowed) && |
418 | nodes_subset(p->mems_allowed, q->mems_allowed) && | 412 | nodes_subset(p->mems_allowed, q->mems_allowed) && |
419 | is_cpu_exclusive(p) <= is_cpu_exclusive(q) && | 413 | is_cpu_exclusive(p) <= is_cpu_exclusive(q) && |
420 | is_mem_exclusive(p) <= is_mem_exclusive(q); | 414 | is_mem_exclusive(p) <= is_mem_exclusive(q); |
421 | } | 415 | } |
422 | 416 | ||
417 | /** | ||
418 | * alloc_trial_cpuset - allocate a trial cpuset | ||
419 | * @cs: the cpuset that the trial cpuset duplicates | ||
420 | */ | ||
421 | static struct cpuset *alloc_trial_cpuset(const struct cpuset *cs) | ||
422 | { | ||
423 | struct cpuset *trial; | ||
424 | |||
425 | trial = kmemdup(cs, sizeof(*cs), GFP_KERNEL); | ||
426 | if (!trial) | ||
427 | return NULL; | ||
428 | |||
429 | if (!alloc_cpumask_var(&trial->cpus_allowed, GFP_KERNEL)) { | ||
430 | kfree(trial); | ||
431 | return NULL; | ||
432 | } | ||
433 | cpumask_copy(trial->cpus_allowed, cs->cpus_allowed); | ||
434 | |||
435 | return trial; | ||
436 | } | ||
437 | |||
438 | /** | ||
439 | * free_trial_cpuset - free the trial cpuset | ||
440 | * @trial: the trial cpuset to be freed | ||
441 | */ | ||
442 | static void free_trial_cpuset(struct cpuset *trial) | ||
443 | { | ||
444 | free_cpumask_var(trial->cpus_allowed); | ||
445 | kfree(trial); | ||
446 | } | ||
447 | |||
423 | /* | 448 | /* |
424 | * validate_change() - Used to validate that any proposed cpuset change | 449 | * validate_change() - Used to validate that any proposed cpuset change |
425 | * follows the structural rules for cpusets. | 450 | * follows the structural rules for cpusets. |
@@ -469,7 +494,7 @@ static int validate_change(const struct cpuset *cur, const struct cpuset *trial) | |||
469 | c = cgroup_cs(cont); | 494 | c = cgroup_cs(cont); |
470 | if ((is_cpu_exclusive(trial) || is_cpu_exclusive(c)) && | 495 | if ((is_cpu_exclusive(trial) || is_cpu_exclusive(c)) && |
471 | c != cur && | 496 | c != cur && |
472 | cpus_intersects(trial->cpus_allowed, c->cpus_allowed)) | 497 | cpumask_intersects(trial->cpus_allowed, c->cpus_allowed)) |
473 | return -EINVAL; | 498 | return -EINVAL; |
474 | if ((is_mem_exclusive(trial) || is_mem_exclusive(c)) && | 499 | if ((is_mem_exclusive(trial) || is_mem_exclusive(c)) && |
475 | c != cur && | 500 | c != cur && |
@@ -479,7 +504,7 @@ static int validate_change(const struct cpuset *cur, const struct cpuset *trial) | |||
479 | 504 | ||
480 | /* Cpusets with tasks can't have empty cpus_allowed or mems_allowed */ | 505 | /* Cpusets with tasks can't have empty cpus_allowed or mems_allowed */ |
481 | if (cgroup_task_count(cur->css.cgroup)) { | 506 | if (cgroup_task_count(cur->css.cgroup)) { |
482 | if (cpus_empty(trial->cpus_allowed) || | 507 | if (cpumask_empty(trial->cpus_allowed) || |
483 | nodes_empty(trial->mems_allowed)) { | 508 | nodes_empty(trial->mems_allowed)) { |
484 | return -ENOSPC; | 509 | return -ENOSPC; |
485 | } | 510 | } |
@@ -494,7 +519,7 @@ static int validate_change(const struct cpuset *cur, const struct cpuset *trial) | |||
494 | */ | 519 | */ |
495 | static int cpusets_overlap(struct cpuset *a, struct cpuset *b) | 520 | static int cpusets_overlap(struct cpuset *a, struct cpuset *b) |
496 | { | 521 | { |
497 | return cpus_intersects(a->cpus_allowed, b->cpus_allowed); | 522 | return cpumask_intersects(a->cpus_allowed, b->cpus_allowed); |
498 | } | 523 | } |
499 | 524 | ||
500 | static void | 525 | static void |
@@ -519,7 +544,7 @@ update_domain_attr_tree(struct sched_domain_attr *dattr, struct cpuset *c) | |||
519 | cp = list_first_entry(&q, struct cpuset, stack_list); | 544 | cp = list_first_entry(&q, struct cpuset, stack_list); |
520 | list_del(q.next); | 545 | list_del(q.next); |
521 | 546 | ||
522 | if (cpus_empty(cp->cpus_allowed)) | 547 | if (cpumask_empty(cp->cpus_allowed)) |
523 | continue; | 548 | continue; |
524 | 549 | ||
525 | if (is_sched_load_balance(cp)) | 550 | if (is_sched_load_balance(cp)) |
@@ -586,7 +611,8 @@ update_domain_attr_tree(struct sched_domain_attr *dattr, struct cpuset *c) | |||
586 | * element of the partition (one sched domain) to be passed to | 611 | * element of the partition (one sched domain) to be passed to |
587 | * partition_sched_domains(). | 612 | * partition_sched_domains(). |
588 | */ | 613 | */ |
589 | static int generate_sched_domains(cpumask_t **domains, | 614 | /* FIXME: see the FIXME in partition_sched_domains() */ |
615 | static int generate_sched_domains(struct cpumask **domains, | ||
590 | struct sched_domain_attr **attributes) | 616 | struct sched_domain_attr **attributes) |
591 | { | 617 | { |
592 | LIST_HEAD(q); /* queue of cpusets to be scanned */ | 618 | LIST_HEAD(q); /* queue of cpusets to be scanned */ |
@@ -594,10 +620,10 @@ static int generate_sched_domains(cpumask_t **domains, | |||
594 | struct cpuset **csa; /* array of all cpuset ptrs */ | 620 | struct cpuset **csa; /* array of all cpuset ptrs */ |
595 | int csn; /* how many cpuset ptrs in csa so far */ | 621 | int csn; /* how many cpuset ptrs in csa so far */ |
596 | int i, j, k; /* indices for partition finding loops */ | 622 | int i, j, k; /* indices for partition finding loops */ |
597 | cpumask_t *doms; /* resulting partition; i.e. sched domains */ | 623 | struct cpumask *doms; /* resulting partition; i.e. sched domains */ |
598 | struct sched_domain_attr *dattr; /* attributes for custom domains */ | 624 | struct sched_domain_attr *dattr; /* attributes for custom domains */ |
599 | int ndoms = 0; /* number of sched domains in result */ | 625 | int ndoms = 0; /* number of sched domains in result */ |
600 | int nslot; /* next empty doms[] cpumask_t slot */ | 626 | int nslot; /* next empty doms[] struct cpumask slot */ |
601 | 627 | ||
602 | doms = NULL; | 628 | doms = NULL; |
603 | dattr = NULL; | 629 | dattr = NULL; |
@@ -605,7 +631,7 @@ static int generate_sched_domains(cpumask_t **domains, | |||
605 | 631 | ||
606 | /* Special case for the 99% of systems with one, full, sched domain */ | 632 | /* Special case for the 99% of systems with one, full, sched domain */ |
607 | if (is_sched_load_balance(&top_cpuset)) { | 633 | if (is_sched_load_balance(&top_cpuset)) { |
608 | doms = kmalloc(sizeof(cpumask_t), GFP_KERNEL); | 634 | doms = kmalloc(cpumask_size(), GFP_KERNEL); |
609 | if (!doms) | 635 | if (!doms) |
610 | goto done; | 636 | goto done; |
611 | 637 | ||
@@ -614,7 +640,7 @@ static int generate_sched_domains(cpumask_t **domains, | |||
614 | *dattr = SD_ATTR_INIT; | 640 | *dattr = SD_ATTR_INIT; |
615 | update_domain_attr_tree(dattr, &top_cpuset); | 641 | update_domain_attr_tree(dattr, &top_cpuset); |
616 | } | 642 | } |
617 | *doms = top_cpuset.cpus_allowed; | 643 | cpumask_copy(doms, top_cpuset.cpus_allowed); |
618 | 644 | ||
619 | ndoms = 1; | 645 | ndoms = 1; |
620 | goto done; | 646 | goto done; |
@@ -633,7 +659,7 @@ static int generate_sched_domains(cpumask_t **domains, | |||
633 | cp = list_first_entry(&q, struct cpuset, stack_list); | 659 | cp = list_first_entry(&q, struct cpuset, stack_list); |
634 | list_del(q.next); | 660 | list_del(q.next); |
635 | 661 | ||
636 | if (cpus_empty(cp->cpus_allowed)) | 662 | if (cpumask_empty(cp->cpus_allowed)) |
637 | continue; | 663 | continue; |
638 | 664 | ||
639 | /* | 665 | /* |
@@ -684,7 +710,7 @@ restart: | |||
684 | * Now we know how many domains to create. | 710 | * Now we know how many domains to create. |
685 | * Convert <csn, csa> to <ndoms, doms> and populate cpu masks. | 711 | * Convert <csn, csa> to <ndoms, doms> and populate cpu masks. |
686 | */ | 712 | */ |
687 | doms = kmalloc(ndoms * sizeof(cpumask_t), GFP_KERNEL); | 713 | doms = kmalloc(ndoms * cpumask_size(), GFP_KERNEL); |
688 | if (!doms) | 714 | if (!doms) |
689 | goto done; | 715 | goto done; |
690 | 716 | ||
@@ -696,7 +722,7 @@ restart: | |||
696 | 722 | ||
697 | for (nslot = 0, i = 0; i < csn; i++) { | 723 | for (nslot = 0, i = 0; i < csn; i++) { |
698 | struct cpuset *a = csa[i]; | 724 | struct cpuset *a = csa[i]; |
699 | cpumask_t *dp; | 725 | struct cpumask *dp; |
700 | int apn = a->pn; | 726 | int apn = a->pn; |
701 | 727 | ||
702 | if (apn < 0) { | 728 | if (apn < 0) { |
@@ -719,14 +745,14 @@ restart: | |||
719 | continue; | 745 | continue; |
720 | } | 746 | } |
721 | 747 | ||
722 | cpus_clear(*dp); | 748 | cpumask_clear(dp); |
723 | if (dattr) | 749 | if (dattr) |
724 | *(dattr + nslot) = SD_ATTR_INIT; | 750 | *(dattr + nslot) = SD_ATTR_INIT; |
725 | for (j = i; j < csn; j++) { | 751 | for (j = i; j < csn; j++) { |
726 | struct cpuset *b = csa[j]; | 752 | struct cpuset *b = csa[j]; |
727 | 753 | ||
728 | if (apn == b->pn) { | 754 | if (apn == b->pn) { |
729 | cpus_or(*dp, *dp, b->cpus_allowed); | 755 | cpumask_or(dp, dp, b->cpus_allowed); |
730 | if (dattr) | 756 | if (dattr) |
731 | update_domain_attr_tree(dattr + nslot, b); | 757 | update_domain_attr_tree(dattr + nslot, b); |
732 | 758 | ||
@@ -766,7 +792,7 @@ done: | |||
766 | static void do_rebuild_sched_domains(struct work_struct *unused) | 792 | static void do_rebuild_sched_domains(struct work_struct *unused) |
767 | { | 793 | { |
768 | struct sched_domain_attr *attr; | 794 | struct sched_domain_attr *attr; |
769 | cpumask_t *doms; | 795 | struct cpumask *doms; |
770 | int ndoms; | 796 | int ndoms; |
771 | 797 | ||
772 | get_online_cpus(); | 798 | get_online_cpus(); |
@@ -835,7 +861,7 @@ void rebuild_sched_domains(void) | |||
835 | static int cpuset_test_cpumask(struct task_struct *tsk, | 861 | static int cpuset_test_cpumask(struct task_struct *tsk, |
836 | struct cgroup_scanner *scan) | 862 | struct cgroup_scanner *scan) |
837 | { | 863 | { |
838 | return !cpus_equal(tsk->cpus_allowed, | 864 | return !cpumask_equal(&tsk->cpus_allowed, |
839 | (cgroup_cs(scan->cg))->cpus_allowed); | 865 | (cgroup_cs(scan->cg))->cpus_allowed); |
840 | } | 866 | } |
841 | 867 | ||
@@ -853,7 +879,7 @@ static int cpuset_test_cpumask(struct task_struct *tsk, | |||
853 | static void cpuset_change_cpumask(struct task_struct *tsk, | 879 | static void cpuset_change_cpumask(struct task_struct *tsk, |
854 | struct cgroup_scanner *scan) | 880 | struct cgroup_scanner *scan) |
855 | { | 881 | { |
856 | set_cpus_allowed_ptr(tsk, &((cgroup_cs(scan->cg))->cpus_allowed)); | 882 | set_cpus_allowed_ptr(tsk, ((cgroup_cs(scan->cg))->cpus_allowed)); |
857 | } | 883 | } |
858 | 884 | ||
859 | /** | 885 | /** |
@@ -885,10 +911,10 @@ static void update_tasks_cpumask(struct cpuset *cs, struct ptr_heap *heap) | |||
885 | * @cs: the cpuset to consider | 911 | * @cs: the cpuset to consider |
886 | * @buf: buffer of cpu numbers written to this cpuset | 912 | * @buf: buffer of cpu numbers written to this cpuset |
887 | */ | 913 | */ |
888 | static int update_cpumask(struct cpuset *cs, const char *buf) | 914 | static int update_cpumask(struct cpuset *cs, struct cpuset *trialcs, |
915 | const char *buf) | ||
889 | { | 916 | { |
890 | struct ptr_heap heap; | 917 | struct ptr_heap heap; |
891 | struct cpuset trialcs; | ||
892 | int retval; | 918 | int retval; |
893 | int is_load_balanced; | 919 | int is_load_balanced; |
894 | 920 | ||
@@ -896,8 +922,6 @@ static int update_cpumask(struct cpuset *cs, const char *buf) | |||
896 | if (cs == &top_cpuset) | 922 | if (cs == &top_cpuset) |
897 | return -EACCES; | 923 | return -EACCES; |
898 | 924 | ||
899 | trialcs = *cs; | ||
900 | |||
901 | /* | 925 | /* |
902 | * An empty cpus_allowed is ok only if the cpuset has no tasks. | 926 | * An empty cpus_allowed is ok only if the cpuset has no tasks. |
903 | * Since cpulist_parse() fails on an empty mask, we special case | 927 | * Since cpulist_parse() fails on an empty mask, we special case |
@@ -905,31 +929,31 @@ static int update_cpumask(struct cpuset *cs, const char *buf) | |||
905 | * with tasks have cpus. | 929 | * with tasks have cpus. |
906 | */ | 930 | */ |
907 | if (!*buf) { | 931 | if (!*buf) { |
908 | cpus_clear(trialcs.cpus_allowed); | 932 | cpumask_clear(trialcs->cpus_allowed); |
909 | } else { | 933 | } else { |
910 | retval = cpulist_parse(buf, &trialcs.cpus_allowed); | 934 | retval = cpulist_parse(buf, trialcs->cpus_allowed); |
911 | if (retval < 0) | 935 | if (retval < 0) |
912 | return retval; | 936 | return retval; |
913 | 937 | ||
914 | if (!cpus_subset(trialcs.cpus_allowed, cpu_online_map)) | 938 | if (!cpumask_subset(trialcs->cpus_allowed, cpu_online_mask)) |
915 | return -EINVAL; | 939 | return -EINVAL; |
916 | } | 940 | } |
917 | retval = validate_change(cs, &trialcs); | 941 | retval = validate_change(cs, trialcs); |
918 | if (retval < 0) | 942 | if (retval < 0) |
919 | return retval; | 943 | return retval; |
920 | 944 | ||
921 | /* Nothing to do if the cpus didn't change */ | 945 | /* Nothing to do if the cpus didn't change */ |
922 | if (cpus_equal(cs->cpus_allowed, trialcs.cpus_allowed)) | 946 | if (cpumask_equal(cs->cpus_allowed, trialcs->cpus_allowed)) |
923 | return 0; | 947 | return 0; |
924 | 948 | ||
925 | retval = heap_init(&heap, PAGE_SIZE, GFP_KERNEL, NULL); | 949 | retval = heap_init(&heap, PAGE_SIZE, GFP_KERNEL, NULL); |
926 | if (retval) | 950 | if (retval) |
927 | return retval; | 951 | return retval; |
928 | 952 | ||
929 | is_load_balanced = is_sched_load_balance(&trialcs); | 953 | is_load_balanced = is_sched_load_balance(trialcs); |
930 | 954 | ||
931 | mutex_lock(&callback_mutex); | 955 | mutex_lock(&callback_mutex); |
932 | cs->cpus_allowed = trialcs.cpus_allowed; | 956 | cpumask_copy(cs->cpus_allowed, trialcs->cpus_allowed); |
933 | mutex_unlock(&callback_mutex); | 957 | mutex_unlock(&callback_mutex); |
934 | 958 | ||
935 | /* | 959 | /* |
@@ -1017,7 +1041,7 @@ static int update_tasks_nodemask(struct cpuset *cs, const nodemask_t *oldmem) | |||
1017 | cpuset_being_rebound = cs; /* causes mpol_dup() rebind */ | 1041 | cpuset_being_rebound = cs; /* causes mpol_dup() rebind */ |
1018 | 1042 | ||
1019 | fudge = 10; /* spare mmarray[] slots */ | 1043 | fudge = 10; /* spare mmarray[] slots */ |
1020 | fudge += cpus_weight(cs->cpus_allowed); /* imagine one fork-bomb/cpu */ | 1044 | fudge += cpumask_weight(cs->cpus_allowed);/* imagine 1 fork-bomb/cpu */ |
1021 | retval = -ENOMEM; | 1045 | retval = -ENOMEM; |
1022 | 1046 | ||
1023 | /* | 1047 | /* |
@@ -1104,9 +1128,9 @@ done: | |||
1104 | * lock each such tasks mm->mmap_sem, scan its vma's and rebind | 1128 | * lock each such tasks mm->mmap_sem, scan its vma's and rebind |
1105 | * their mempolicies to the cpusets new mems_allowed. | 1129 | * their mempolicies to the cpusets new mems_allowed. |
1106 | */ | 1130 | */ |
1107 | static int update_nodemask(struct cpuset *cs, const char *buf) | 1131 | static int update_nodemask(struct cpuset *cs, struct cpuset *trialcs, |
1132 | const char *buf) | ||
1108 | { | 1133 | { |
1109 | struct cpuset trialcs; | ||
1110 | nodemask_t oldmem; | 1134 | nodemask_t oldmem; |
1111 | int retval; | 1135 | int retval; |
1112 | 1136 | ||
@@ -1117,8 +1141,6 @@ static int update_nodemask(struct cpuset *cs, const char *buf) | |||
1117 | if (cs == &top_cpuset) | 1141 | if (cs == &top_cpuset) |
1118 | return -EACCES; | 1142 | return -EACCES; |
1119 | 1143 | ||
1120 | trialcs = *cs; | ||
1121 | |||
1122 | /* | 1144 | /* |
1123 | * An empty mems_allowed is ok iff there are no tasks in the cpuset. | 1145 | * An empty mems_allowed is ok iff there are no tasks in the cpuset. |
1124 | * Since nodelist_parse() fails on an empty mask, we special case | 1146 | * Since nodelist_parse() fails on an empty mask, we special case |
@@ -1126,27 +1148,27 @@ static int update_nodemask(struct cpuset *cs, const char *buf) | |||
1126 | * with tasks have memory. | 1148 | * with tasks have memory. |
1127 | */ | 1149 | */ |
1128 | if (!*buf) { | 1150 | if (!*buf) { |
1129 | nodes_clear(trialcs.mems_allowed); | 1151 | nodes_clear(trialcs->mems_allowed); |
1130 | } else { | 1152 | } else { |
1131 | retval = nodelist_parse(buf, trialcs.mems_allowed); | 1153 | retval = nodelist_parse(buf, trialcs->mems_allowed); |
1132 | if (retval < 0) | 1154 | if (retval < 0) |
1133 | goto done; | 1155 | goto done; |
1134 | 1156 | ||
1135 | if (!nodes_subset(trialcs.mems_allowed, | 1157 | if (!nodes_subset(trialcs->mems_allowed, |
1136 | node_states[N_HIGH_MEMORY])) | 1158 | node_states[N_HIGH_MEMORY])) |
1137 | return -EINVAL; | 1159 | return -EINVAL; |
1138 | } | 1160 | } |
1139 | oldmem = cs->mems_allowed; | 1161 | oldmem = cs->mems_allowed; |
1140 | if (nodes_equal(oldmem, trialcs.mems_allowed)) { | 1162 | if (nodes_equal(oldmem, trialcs->mems_allowed)) { |
1141 | retval = 0; /* Too easy - nothing to do */ | 1163 | retval = 0; /* Too easy - nothing to do */ |
1142 | goto done; | 1164 | goto done; |
1143 | } | 1165 | } |
1144 | retval = validate_change(cs, &trialcs); | 1166 | retval = validate_change(cs, trialcs); |
1145 | if (retval < 0) | 1167 | if (retval < 0) |
1146 | goto done; | 1168 | goto done; |
1147 | 1169 | ||
1148 | mutex_lock(&callback_mutex); | 1170 | mutex_lock(&callback_mutex); |
1149 | cs->mems_allowed = trialcs.mems_allowed; | 1171 | cs->mems_allowed = trialcs->mems_allowed; |
1150 | cs->mems_generation = cpuset_mems_generation++; | 1172 | cs->mems_generation = cpuset_mems_generation++; |
1151 | mutex_unlock(&callback_mutex); | 1173 | mutex_unlock(&callback_mutex); |
1152 | 1174 | ||
@@ -1167,7 +1189,8 @@ static int update_relax_domain_level(struct cpuset *cs, s64 val) | |||
1167 | 1189 | ||
1168 | if (val != cs->relax_domain_level) { | 1190 | if (val != cs->relax_domain_level) { |
1169 | cs->relax_domain_level = val; | 1191 | cs->relax_domain_level = val; |
1170 | if (!cpus_empty(cs->cpus_allowed) && is_sched_load_balance(cs)) | 1192 | if (!cpumask_empty(cs->cpus_allowed) && |
1193 | is_sched_load_balance(cs)) | ||
1171 | async_rebuild_sched_domains(); | 1194 | async_rebuild_sched_domains(); |
1172 | } | 1195 | } |
1173 | 1196 | ||
@@ -1186,31 +1209,36 @@ static int update_relax_domain_level(struct cpuset *cs, s64 val) | |||
1186 | static int update_flag(cpuset_flagbits_t bit, struct cpuset *cs, | 1209 | static int update_flag(cpuset_flagbits_t bit, struct cpuset *cs, |
1187 | int turning_on) | 1210 | int turning_on) |
1188 | { | 1211 | { |
1189 | struct cpuset trialcs; | 1212 | struct cpuset *trialcs; |
1190 | int err; | 1213 | int err; |
1191 | int balance_flag_changed; | 1214 | int balance_flag_changed; |
1192 | 1215 | ||
1193 | trialcs = *cs; | 1216 | trialcs = alloc_trial_cpuset(cs); |
1217 | if (!trialcs) | ||
1218 | return -ENOMEM; | ||
1219 | |||
1194 | if (turning_on) | 1220 | if (turning_on) |
1195 | set_bit(bit, &trialcs.flags); | 1221 | set_bit(bit, &trialcs->flags); |
1196 | else | 1222 | else |
1197 | clear_bit(bit, &trialcs.flags); | 1223 | clear_bit(bit, &trialcs->flags); |
1198 | 1224 | ||
1199 | err = validate_change(cs, &trialcs); | 1225 | err = validate_change(cs, trialcs); |
1200 | if (err < 0) | 1226 | if (err < 0) |
1201 | return err; | 1227 | goto out; |
1202 | 1228 | ||
1203 | balance_flag_changed = (is_sched_load_balance(cs) != | 1229 | balance_flag_changed = (is_sched_load_balance(cs) != |
1204 | is_sched_load_balance(&trialcs)); | 1230 | is_sched_load_balance(trialcs)); |
1205 | 1231 | ||
1206 | mutex_lock(&callback_mutex); | 1232 | mutex_lock(&callback_mutex); |
1207 | cs->flags = trialcs.flags; | 1233 | cs->flags = trialcs->flags; |
1208 | mutex_unlock(&callback_mutex); | 1234 | mutex_unlock(&callback_mutex); |
1209 | 1235 | ||
1210 | if (!cpus_empty(trialcs.cpus_allowed) && balance_flag_changed) | 1236 | if (!cpumask_empty(trialcs->cpus_allowed) && balance_flag_changed) |
1211 | async_rebuild_sched_domains(); | 1237 | async_rebuild_sched_domains(); |
1212 | 1238 | ||
1213 | return 0; | 1239 | out: |
1240 | free_trial_cpuset(trialcs); | ||
1241 | return err; | ||
1214 | } | 1242 | } |
1215 | 1243 | ||
1216 | /* | 1244 | /* |
@@ -1311,42 +1339,47 @@ static int fmeter_getrate(struct fmeter *fmp) | |||
1311 | return val; | 1339 | return val; |
1312 | } | 1340 | } |
1313 | 1341 | ||
1342 | /* Protected by cgroup_lock */ | ||
1343 | static cpumask_var_t cpus_attach; | ||
1344 | |||
1314 | /* Called by cgroups to determine if a cpuset is usable; cgroup_mutex held */ | 1345 | /* Called by cgroups to determine if a cpuset is usable; cgroup_mutex held */ |
1315 | static int cpuset_can_attach(struct cgroup_subsys *ss, | 1346 | static int cpuset_can_attach(struct cgroup_subsys *ss, |
1316 | struct cgroup *cont, struct task_struct *tsk) | 1347 | struct cgroup *cont, struct task_struct *tsk) |
1317 | { | 1348 | { |
1318 | struct cpuset *cs = cgroup_cs(cont); | 1349 | struct cpuset *cs = cgroup_cs(cont); |
1350 | int ret = 0; | ||
1319 | 1351 | ||
1320 | if (cpus_empty(cs->cpus_allowed) || nodes_empty(cs->mems_allowed)) | 1352 | if (cpumask_empty(cs->cpus_allowed) || nodes_empty(cs->mems_allowed)) |
1321 | return -ENOSPC; | 1353 | return -ENOSPC; |
1322 | if (tsk->flags & PF_THREAD_BOUND) { | ||
1323 | cpumask_t mask; | ||
1324 | 1354 | ||
1355 | if (tsk->flags & PF_THREAD_BOUND) { | ||
1325 | mutex_lock(&callback_mutex); | 1356 | mutex_lock(&callback_mutex); |
1326 | mask = cs->cpus_allowed; | 1357 | if (!cpumask_equal(&tsk->cpus_allowed, cs->cpus_allowed)) |
1358 | ret = -EINVAL; | ||
1327 | mutex_unlock(&callback_mutex); | 1359 | mutex_unlock(&callback_mutex); |
1328 | if (!cpus_equal(tsk->cpus_allowed, mask)) | ||
1329 | return -EINVAL; | ||
1330 | } | 1360 | } |
1331 | 1361 | ||
1332 | return security_task_setscheduler(tsk, 0, NULL); | 1362 | return ret < 0 ? ret : security_task_setscheduler(tsk, 0, NULL); |
1333 | } | 1363 | } |
1334 | 1364 | ||
1335 | static void cpuset_attach(struct cgroup_subsys *ss, | 1365 | static void cpuset_attach(struct cgroup_subsys *ss, |
1336 | struct cgroup *cont, struct cgroup *oldcont, | 1366 | struct cgroup *cont, struct cgroup *oldcont, |
1337 | struct task_struct *tsk) | 1367 | struct task_struct *tsk) |
1338 | { | 1368 | { |
1339 | cpumask_t cpus; | ||
1340 | nodemask_t from, to; | 1369 | nodemask_t from, to; |
1341 | struct mm_struct *mm; | 1370 | struct mm_struct *mm; |
1342 | struct cpuset *cs = cgroup_cs(cont); | 1371 | struct cpuset *cs = cgroup_cs(cont); |
1343 | struct cpuset *oldcs = cgroup_cs(oldcont); | 1372 | struct cpuset *oldcs = cgroup_cs(oldcont); |
1344 | int err; | 1373 | int err; |
1345 | 1374 | ||
1346 | mutex_lock(&callback_mutex); | 1375 | if (cs == &top_cpuset) { |
1347 | guarantee_online_cpus(cs, &cpus); | 1376 | cpumask_copy(cpus_attach, cpu_possible_mask); |
1348 | err = set_cpus_allowed_ptr(tsk, &cpus); | 1377 | } else { |
1349 | mutex_unlock(&callback_mutex); | 1378 | mutex_lock(&callback_mutex); |
1379 | guarantee_online_cpus(cs, cpus_attach); | ||
1380 | mutex_unlock(&callback_mutex); | ||
1381 | } | ||
1382 | err = set_cpus_allowed_ptr(tsk, cpus_attach); | ||
1350 | if (err) | 1383 | if (err) |
1351 | return; | 1384 | return; |
1352 | 1385 | ||
@@ -1359,7 +1392,6 @@ static void cpuset_attach(struct cgroup_subsys *ss, | |||
1359 | cpuset_migrate_mm(mm, &from, &to); | 1392 | cpuset_migrate_mm(mm, &from, &to); |
1360 | mmput(mm); | 1393 | mmput(mm); |
1361 | } | 1394 | } |
1362 | |||
1363 | } | 1395 | } |
1364 | 1396 | ||
1365 | /* The various types of files and directories in a cpuset file system */ | 1397 | /* The various types of files and directories in a cpuset file system */ |
@@ -1454,21 +1486,29 @@ static int cpuset_write_resmask(struct cgroup *cgrp, struct cftype *cft, | |||
1454 | const char *buf) | 1486 | const char *buf) |
1455 | { | 1487 | { |
1456 | int retval = 0; | 1488 | int retval = 0; |
1489 | struct cpuset *cs = cgroup_cs(cgrp); | ||
1490 | struct cpuset *trialcs; | ||
1457 | 1491 | ||
1458 | if (!cgroup_lock_live_group(cgrp)) | 1492 | if (!cgroup_lock_live_group(cgrp)) |
1459 | return -ENODEV; | 1493 | return -ENODEV; |
1460 | 1494 | ||
1495 | trialcs = alloc_trial_cpuset(cs); | ||
1496 | if (!trialcs) | ||
1497 | return -ENOMEM; | ||
1498 | |||
1461 | switch (cft->private) { | 1499 | switch (cft->private) { |
1462 | case FILE_CPULIST: | 1500 | case FILE_CPULIST: |
1463 | retval = update_cpumask(cgroup_cs(cgrp), buf); | 1501 | retval = update_cpumask(cs, trialcs, buf); |
1464 | break; | 1502 | break; |
1465 | case FILE_MEMLIST: | 1503 | case FILE_MEMLIST: |
1466 | retval = update_nodemask(cgroup_cs(cgrp), buf); | 1504 | retval = update_nodemask(cs, trialcs, buf); |
1467 | break; | 1505 | break; |
1468 | default: | 1506 | default: |
1469 | retval = -EINVAL; | 1507 | retval = -EINVAL; |
1470 | break; | 1508 | break; |
1471 | } | 1509 | } |
1510 | |||
1511 | free_trial_cpuset(trialcs); | ||
1472 | cgroup_unlock(); | 1512 | cgroup_unlock(); |
1473 | return retval; | 1513 | return retval; |
1474 | } | 1514 | } |
@@ -1487,13 +1527,13 @@ static int cpuset_write_resmask(struct cgroup *cgrp, struct cftype *cft, | |||
1487 | 1527 | ||
1488 | static int cpuset_sprintf_cpulist(char *page, struct cpuset *cs) | 1528 | static int cpuset_sprintf_cpulist(char *page, struct cpuset *cs) |
1489 | { | 1529 | { |
1490 | cpumask_t mask; | 1530 | int ret; |
1491 | 1531 | ||
1492 | mutex_lock(&callback_mutex); | 1532 | mutex_lock(&callback_mutex); |
1493 | mask = cs->cpus_allowed; | 1533 | ret = cpulist_scnprintf(page, PAGE_SIZE, cs->cpus_allowed); |
1494 | mutex_unlock(&callback_mutex); | 1534 | mutex_unlock(&callback_mutex); |
1495 | 1535 | ||
1496 | return cpulist_scnprintf(page, PAGE_SIZE, &mask); | 1536 | return ret; |
1497 | } | 1537 | } |
1498 | 1538 | ||
1499 | static int cpuset_sprintf_memlist(char *page, struct cpuset *cs) | 1539 | static int cpuset_sprintf_memlist(char *page, struct cpuset *cs) |
@@ -1729,7 +1769,7 @@ static void cpuset_post_clone(struct cgroup_subsys *ss, | |||
1729 | parent_cs = cgroup_cs(parent); | 1769 | parent_cs = cgroup_cs(parent); |
1730 | 1770 | ||
1731 | cs->mems_allowed = parent_cs->mems_allowed; | 1771 | cs->mems_allowed = parent_cs->mems_allowed; |
1732 | cs->cpus_allowed = parent_cs->cpus_allowed; | 1772 | cpumask_copy(cs->cpus_allowed, parent_cs->cpus_allowed); |
1733 | return; | 1773 | return; |
1734 | } | 1774 | } |
1735 | 1775 | ||
@@ -1755,6 +1795,10 @@ static struct cgroup_subsys_state *cpuset_create( | |||
1755 | cs = kmalloc(sizeof(*cs), GFP_KERNEL); | 1795 | cs = kmalloc(sizeof(*cs), GFP_KERNEL); |
1756 | if (!cs) | 1796 | if (!cs) |
1757 | return ERR_PTR(-ENOMEM); | 1797 | return ERR_PTR(-ENOMEM); |
1798 | if (!alloc_cpumask_var(&cs->cpus_allowed, GFP_KERNEL)) { | ||
1799 | kfree(cs); | ||
1800 | return ERR_PTR(-ENOMEM); | ||
1801 | } | ||
1758 | 1802 | ||
1759 | cpuset_update_task_memory_state(); | 1803 | cpuset_update_task_memory_state(); |
1760 | cs->flags = 0; | 1804 | cs->flags = 0; |
@@ -1763,7 +1807,7 @@ static struct cgroup_subsys_state *cpuset_create( | |||
1763 | if (is_spread_slab(parent)) | 1807 | if (is_spread_slab(parent)) |
1764 | set_bit(CS_SPREAD_SLAB, &cs->flags); | 1808 | set_bit(CS_SPREAD_SLAB, &cs->flags); |
1765 | set_bit(CS_SCHED_LOAD_BALANCE, &cs->flags); | 1809 | set_bit(CS_SCHED_LOAD_BALANCE, &cs->flags); |
1766 | cpus_clear(cs->cpus_allowed); | 1810 | cpumask_clear(cs->cpus_allowed); |
1767 | nodes_clear(cs->mems_allowed); | 1811 | nodes_clear(cs->mems_allowed); |
1768 | cs->mems_generation = cpuset_mems_generation++; | 1812 | cs->mems_generation = cpuset_mems_generation++; |
1769 | fmeter_init(&cs->fmeter); | 1813 | fmeter_init(&cs->fmeter); |
@@ -1790,6 +1834,7 @@ static void cpuset_destroy(struct cgroup_subsys *ss, struct cgroup *cont) | |||
1790 | update_flag(CS_SCHED_LOAD_BALANCE, cs, 0); | 1834 | update_flag(CS_SCHED_LOAD_BALANCE, cs, 0); |
1791 | 1835 | ||
1792 | number_of_cpusets--; | 1836 | number_of_cpusets--; |
1837 | free_cpumask_var(cs->cpus_allowed); | ||
1793 | kfree(cs); | 1838 | kfree(cs); |
1794 | } | 1839 | } |
1795 | 1840 | ||
@@ -1813,6 +1858,8 @@ struct cgroup_subsys cpuset_subsys = { | |||
1813 | 1858 | ||
1814 | int __init cpuset_init_early(void) | 1859 | int __init cpuset_init_early(void) |
1815 | { | 1860 | { |
1861 | alloc_bootmem_cpumask_var(&top_cpuset.cpus_allowed); | ||
1862 | |||
1816 | top_cpuset.mems_generation = cpuset_mems_generation++; | 1863 | top_cpuset.mems_generation = cpuset_mems_generation++; |
1817 | return 0; | 1864 | return 0; |
1818 | } | 1865 | } |
@@ -1828,7 +1875,7 @@ int __init cpuset_init(void) | |||
1828 | { | 1875 | { |
1829 | int err = 0; | 1876 | int err = 0; |
1830 | 1877 | ||
1831 | cpus_setall(top_cpuset.cpus_allowed); | 1878 | cpumask_setall(top_cpuset.cpus_allowed); |
1832 | nodes_setall(top_cpuset.mems_allowed); | 1879 | nodes_setall(top_cpuset.mems_allowed); |
1833 | 1880 | ||
1834 | fmeter_init(&top_cpuset.fmeter); | 1881 | fmeter_init(&top_cpuset.fmeter); |
@@ -1840,6 +1887,9 @@ int __init cpuset_init(void) | |||
1840 | if (err < 0) | 1887 | if (err < 0) |
1841 | return err; | 1888 | return err; |
1842 | 1889 | ||
1890 | if (!alloc_cpumask_var(&cpus_attach, GFP_KERNEL)) | ||
1891 | BUG(); | ||
1892 | |||
1843 | number_of_cpusets = 1; | 1893 | number_of_cpusets = 1; |
1844 | return 0; | 1894 | return 0; |
1845 | } | 1895 | } |
@@ -1914,7 +1964,7 @@ static void remove_tasks_in_empty_cpuset(struct cpuset *cs) | |||
1914 | * has online cpus, so can't be empty). | 1964 | * has online cpus, so can't be empty). |
1915 | */ | 1965 | */ |
1916 | parent = cs->parent; | 1966 | parent = cs->parent; |
1917 | while (cpus_empty(parent->cpus_allowed) || | 1967 | while (cpumask_empty(parent->cpus_allowed) || |
1918 | nodes_empty(parent->mems_allowed)) | 1968 | nodes_empty(parent->mems_allowed)) |
1919 | parent = parent->parent; | 1969 | parent = parent->parent; |
1920 | 1970 | ||
@@ -1955,7 +2005,7 @@ static void scan_for_empty_cpusets(struct cpuset *root) | |||
1955 | } | 2005 | } |
1956 | 2006 | ||
1957 | /* Continue past cpusets with all cpus, mems online */ | 2007 | /* Continue past cpusets with all cpus, mems online */ |
1958 | if (cpus_subset(cp->cpus_allowed, cpu_online_map) && | 2008 | if (cpumask_subset(cp->cpus_allowed, cpu_online_mask) && |
1959 | nodes_subset(cp->mems_allowed, node_states[N_HIGH_MEMORY])) | 2009 | nodes_subset(cp->mems_allowed, node_states[N_HIGH_MEMORY])) |
1960 | continue; | 2010 | continue; |
1961 | 2011 | ||
@@ -1963,13 +2013,14 @@ static void scan_for_empty_cpusets(struct cpuset *root) | |||
1963 | 2013 | ||
1964 | /* Remove offline cpus and mems from this cpuset. */ | 2014 | /* Remove offline cpus and mems from this cpuset. */ |
1965 | mutex_lock(&callback_mutex); | 2015 | mutex_lock(&callback_mutex); |
1966 | cpus_and(cp->cpus_allowed, cp->cpus_allowed, cpu_online_map); | 2016 | cpumask_and(cp->cpus_allowed, cp->cpus_allowed, |
2017 | cpu_online_mask); | ||
1967 | nodes_and(cp->mems_allowed, cp->mems_allowed, | 2018 | nodes_and(cp->mems_allowed, cp->mems_allowed, |
1968 | node_states[N_HIGH_MEMORY]); | 2019 | node_states[N_HIGH_MEMORY]); |
1969 | mutex_unlock(&callback_mutex); | 2020 | mutex_unlock(&callback_mutex); |
1970 | 2021 | ||
1971 | /* Move tasks from the empty cpuset to a parent */ | 2022 | /* Move tasks from the empty cpuset to a parent */ |
1972 | if (cpus_empty(cp->cpus_allowed) || | 2023 | if (cpumask_empty(cp->cpus_allowed) || |
1973 | nodes_empty(cp->mems_allowed)) | 2024 | nodes_empty(cp->mems_allowed)) |
1974 | remove_tasks_in_empty_cpuset(cp); | 2025 | remove_tasks_in_empty_cpuset(cp); |
1975 | else { | 2026 | else { |
@@ -1995,7 +2046,7 @@ static int cpuset_track_online_cpus(struct notifier_block *unused_nb, | |||
1995 | unsigned long phase, void *unused_cpu) | 2046 | unsigned long phase, void *unused_cpu) |
1996 | { | 2047 | { |
1997 | struct sched_domain_attr *attr; | 2048 | struct sched_domain_attr *attr; |
1998 | cpumask_t *doms; | 2049 | struct cpumask *doms; |
1999 | int ndoms; | 2050 | int ndoms; |
2000 | 2051 | ||
2001 | switch (phase) { | 2052 | switch (phase) { |
@@ -2010,7 +2061,7 @@ static int cpuset_track_online_cpus(struct notifier_block *unused_nb, | |||
2010 | } | 2061 | } |
2011 | 2062 | ||
2012 | cgroup_lock(); | 2063 | cgroup_lock(); |
2013 | top_cpuset.cpus_allowed = cpu_online_map; | 2064 | cpumask_copy(top_cpuset.cpus_allowed, cpu_online_mask); |
2014 | scan_for_empty_cpusets(&top_cpuset); | 2065 | scan_for_empty_cpusets(&top_cpuset); |
2015 | ndoms = generate_sched_domains(&doms, &attr); | 2066 | ndoms = generate_sched_domains(&doms, &attr); |
2016 | cgroup_unlock(); | 2067 | cgroup_unlock(); |
@@ -2055,7 +2106,7 @@ static int cpuset_track_online_nodes(struct notifier_block *self, | |||
2055 | 2106 | ||
2056 | void __init cpuset_init_smp(void) | 2107 | void __init cpuset_init_smp(void) |
2057 | { | 2108 | { |
2058 | top_cpuset.cpus_allowed = cpu_online_map; | 2109 | cpumask_copy(top_cpuset.cpus_allowed, cpu_online_mask); |
2059 | top_cpuset.mems_allowed = node_states[N_HIGH_MEMORY]; | 2110 | top_cpuset.mems_allowed = node_states[N_HIGH_MEMORY]; |
2060 | 2111 | ||
2061 | hotcpu_notifier(cpuset_track_online_cpus, 0); | 2112 | hotcpu_notifier(cpuset_track_online_cpus, 0); |
@@ -2065,15 +2116,15 @@ void __init cpuset_init_smp(void) | |||
2065 | /** | 2116 | /** |
2066 | * cpuset_cpus_allowed - return cpus_allowed mask from a tasks cpuset. | 2117 | * cpuset_cpus_allowed - return cpus_allowed mask from a tasks cpuset. |
2067 | * @tsk: pointer to task_struct from which to obtain cpuset->cpus_allowed. | 2118 | * @tsk: pointer to task_struct from which to obtain cpuset->cpus_allowed. |
2068 | * @pmask: pointer to cpumask_t variable to receive cpus_allowed set. | 2119 | * @pmask: pointer to struct cpumask variable to receive cpus_allowed set. |
2069 | * | 2120 | * |
2070 | * Description: Returns the cpumask_t cpus_allowed of the cpuset | 2121 | * Description: Returns the cpumask_var_t cpus_allowed of the cpuset |
2071 | * attached to the specified @tsk. Guaranteed to return some non-empty | 2122 | * attached to the specified @tsk. Guaranteed to return some non-empty |
2072 | * subset of cpu_online_map, even if this means going outside the | 2123 | * subset of cpu_online_map, even if this means going outside the |
2073 | * tasks cpuset. | 2124 | * tasks cpuset. |
2074 | **/ | 2125 | **/ |
2075 | 2126 | ||
2076 | void cpuset_cpus_allowed(struct task_struct *tsk, cpumask_t *pmask) | 2127 | void cpuset_cpus_allowed(struct task_struct *tsk, struct cpumask *pmask) |
2077 | { | 2128 | { |
2078 | mutex_lock(&callback_mutex); | 2129 | mutex_lock(&callback_mutex); |
2079 | cpuset_cpus_allowed_locked(tsk, pmask); | 2130 | cpuset_cpus_allowed_locked(tsk, pmask); |
@@ -2084,7 +2135,7 @@ void cpuset_cpus_allowed(struct task_struct *tsk, cpumask_t *pmask) | |||
2084 | * cpuset_cpus_allowed_locked - return cpus_allowed mask from a tasks cpuset. | 2135 | * cpuset_cpus_allowed_locked - return cpus_allowed mask from a tasks cpuset. |
2085 | * Must be called with callback_mutex held. | 2136 | * Must be called with callback_mutex held. |
2086 | **/ | 2137 | **/ |
2087 | void cpuset_cpus_allowed_locked(struct task_struct *tsk, cpumask_t *pmask) | 2138 | void cpuset_cpus_allowed_locked(struct task_struct *tsk, struct cpumask *pmask) |
2088 | { | 2139 | { |
2089 | task_lock(tsk); | 2140 | task_lock(tsk); |
2090 | guarantee_online_cpus(task_cs(tsk), pmask); | 2141 | guarantee_online_cpus(task_cs(tsk), pmask); |
diff --git a/kernel/fork.c b/kernel/fork.c index 7b8f2a78be3d..4018308048cf 100644 --- a/kernel/fork.c +++ b/kernel/fork.c | |||
@@ -1126,12 +1126,12 @@ static struct task_struct *copy_process(unsigned long clone_flags, | |||
1126 | 1126 | ||
1127 | if (pid != &init_struct_pid) { | 1127 | if (pid != &init_struct_pid) { |
1128 | retval = -ENOMEM; | 1128 | retval = -ENOMEM; |
1129 | pid = alloc_pid(task_active_pid_ns(p)); | 1129 | pid = alloc_pid(p->nsproxy->pid_ns); |
1130 | if (!pid) | 1130 | if (!pid) |
1131 | goto bad_fork_cleanup_io; | 1131 | goto bad_fork_cleanup_io; |
1132 | 1132 | ||
1133 | if (clone_flags & CLONE_NEWPID) { | 1133 | if (clone_flags & CLONE_NEWPID) { |
1134 | retval = pid_ns_prepare_proc(task_active_pid_ns(p)); | 1134 | retval = pid_ns_prepare_proc(p->nsproxy->pid_ns); |
1135 | if (retval < 0) | 1135 | if (retval < 0) |
1136 | goto bad_fork_free_pid; | 1136 | goto bad_fork_free_pid; |
1137 | } | 1137 | } |
diff --git a/kernel/ns_cgroup.c b/kernel/ns_cgroup.c index 43c2111cd54d..78bc3fdac0d2 100644 --- a/kernel/ns_cgroup.c +++ b/kernel/ns_cgroup.c | |||
@@ -13,7 +13,6 @@ | |||
13 | 13 | ||
14 | struct ns_cgroup { | 14 | struct ns_cgroup { |
15 | struct cgroup_subsys_state css; | 15 | struct cgroup_subsys_state css; |
16 | spinlock_t lock; | ||
17 | }; | 16 | }; |
18 | 17 | ||
19 | struct cgroup_subsys ns_subsys; | 18 | struct cgroup_subsys ns_subsys; |
@@ -84,7 +83,6 @@ static struct cgroup_subsys_state *ns_create(struct cgroup_subsys *ss, | |||
84 | ns_cgroup = kzalloc(sizeof(*ns_cgroup), GFP_KERNEL); | 83 | ns_cgroup = kzalloc(sizeof(*ns_cgroup), GFP_KERNEL); |
85 | if (!ns_cgroup) | 84 | if (!ns_cgroup) |
86 | return ERR_PTR(-ENOMEM); | 85 | return ERR_PTR(-ENOMEM); |
87 | spin_lock_init(&ns_cgroup->lock); | ||
88 | return &ns_cgroup->css; | 86 | return &ns_cgroup->css; |
89 | } | 87 | } |
90 | 88 | ||
diff --git a/kernel/pid.c b/kernel/pid.c index af9224cdd6c0..1b3586fe753a 100644 --- a/kernel/pid.c +++ b/kernel/pid.c | |||
@@ -474,6 +474,12 @@ pid_t task_session_nr_ns(struct task_struct *tsk, struct pid_namespace *ns) | |||
474 | } | 474 | } |
475 | EXPORT_SYMBOL(task_session_nr_ns); | 475 | EXPORT_SYMBOL(task_session_nr_ns); |
476 | 476 | ||
477 | struct pid_namespace *task_active_pid_ns(struct task_struct *tsk) | ||
478 | { | ||
479 | return ns_of_pid(task_pid(tsk)); | ||
480 | } | ||
481 | EXPORT_SYMBOL_GPL(task_active_pid_ns); | ||
482 | |||
477 | /* | 483 | /* |
478 | * Used by proc to find the first pid that is greater than or equal to nr. | 484 | * Used by proc to find the first pid that is greater than or equal to nr. |
479 | * | 485 | * |
diff --git a/kernel/res_counter.c b/kernel/res_counter.c index f275c8eca772..bf8e7534c803 100644 --- a/kernel/res_counter.c +++ b/kernel/res_counter.c | |||
@@ -15,10 +15,11 @@ | |||
15 | #include <linux/uaccess.h> | 15 | #include <linux/uaccess.h> |
16 | #include <linux/mm.h> | 16 | #include <linux/mm.h> |
17 | 17 | ||
18 | void res_counter_init(struct res_counter *counter) | 18 | void res_counter_init(struct res_counter *counter, struct res_counter *parent) |
19 | { | 19 | { |
20 | spin_lock_init(&counter->lock); | 20 | spin_lock_init(&counter->lock); |
21 | counter->limit = (unsigned long long)LLONG_MAX; | 21 | counter->limit = (unsigned long long)LLONG_MAX; |
22 | counter->parent = parent; | ||
22 | } | 23 | } |
23 | 24 | ||
24 | int res_counter_charge_locked(struct res_counter *counter, unsigned long val) | 25 | int res_counter_charge_locked(struct res_counter *counter, unsigned long val) |
@@ -34,14 +35,34 @@ int res_counter_charge_locked(struct res_counter *counter, unsigned long val) | |||
34 | return 0; | 35 | return 0; |
35 | } | 36 | } |
36 | 37 | ||
37 | int res_counter_charge(struct res_counter *counter, unsigned long val) | 38 | int res_counter_charge(struct res_counter *counter, unsigned long val, |
39 | struct res_counter **limit_fail_at) | ||
38 | { | 40 | { |
39 | int ret; | 41 | int ret; |
40 | unsigned long flags; | 42 | unsigned long flags; |
41 | 43 | struct res_counter *c, *u; | |
42 | spin_lock_irqsave(&counter->lock, flags); | 44 | |
43 | ret = res_counter_charge_locked(counter, val); | 45 | *limit_fail_at = NULL; |
44 | spin_unlock_irqrestore(&counter->lock, flags); | 46 | local_irq_save(flags); |
47 | for (c = counter; c != NULL; c = c->parent) { | ||
48 | spin_lock(&c->lock); | ||
49 | ret = res_counter_charge_locked(c, val); | ||
50 | spin_unlock(&c->lock); | ||
51 | if (ret < 0) { | ||
52 | *limit_fail_at = c; | ||
53 | goto undo; | ||
54 | } | ||
55 | } | ||
56 | ret = 0; | ||
57 | goto done; | ||
58 | undo: | ||
59 | for (u = counter; u != c; u = u->parent) { | ||
60 | spin_lock(&u->lock); | ||
61 | res_counter_uncharge_locked(u, val); | ||
62 | spin_unlock(&u->lock); | ||
63 | } | ||
64 | done: | ||
65 | local_irq_restore(flags); | ||
45 | return ret; | 66 | return ret; |
46 | } | 67 | } |
47 | 68 | ||
@@ -56,10 +77,15 @@ void res_counter_uncharge_locked(struct res_counter *counter, unsigned long val) | |||
56 | void res_counter_uncharge(struct res_counter *counter, unsigned long val) | 77 | void res_counter_uncharge(struct res_counter *counter, unsigned long val) |
57 | { | 78 | { |
58 | unsigned long flags; | 79 | unsigned long flags; |
80 | struct res_counter *c; | ||
59 | 81 | ||
60 | spin_lock_irqsave(&counter->lock, flags); | 82 | local_irq_save(flags); |
61 | res_counter_uncharge_locked(counter, val); | 83 | for (c = counter; c != NULL; c = c->parent) { |
62 | spin_unlock_irqrestore(&counter->lock, flags); | 84 | spin_lock(&c->lock); |
85 | res_counter_uncharge_locked(c, val); | ||
86 | spin_unlock(&c->lock); | ||
87 | } | ||
88 | local_irq_restore(flags); | ||
63 | } | 89 | } |
64 | 90 | ||
65 | 91 | ||
diff --git a/kernel/sched_fair.c b/kernel/sched_fair.c index e0c0b4bc3f08..8e1352c75557 100644 --- a/kernel/sched_fair.c +++ b/kernel/sched_fair.c | |||
@@ -1617,8 +1617,6 @@ static void task_tick_fair(struct rq *rq, struct task_struct *curr, int queued) | |||
1617 | } | 1617 | } |
1618 | } | 1618 | } |
1619 | 1619 | ||
1620 | #define swap(a, b) do { typeof(a) tmp = (a); (a) = (b); (b) = tmp; } while (0) | ||
1621 | |||
1622 | /* | 1620 | /* |
1623 | * Share the fairness runtime between parent and child, thus the | 1621 | * Share the fairness runtime between parent and child, thus the |
1624 | * total amount of pressure for CPU stays equal - new tasks | 1622 | * total amount of pressure for CPU stays equal - new tasks |
diff --git a/lib/sort.c b/lib/sort.c index 6abbaf3d5858..926d00429ed2 100644 --- a/lib/sort.c +++ b/lib/sort.c | |||
@@ -32,11 +32,11 @@ static void generic_swap(void *a, void *b, int size) | |||
32 | * @base: pointer to data to sort | 32 | * @base: pointer to data to sort |
33 | * @num: number of elements | 33 | * @num: number of elements |
34 | * @size: size of each element | 34 | * @size: size of each element |
35 | * @cmp: pointer to comparison function | 35 | * @cmp_func: pointer to comparison function |
36 | * @swap: pointer to swap function or NULL | 36 | * @swap_func: pointer to swap function or NULL |
37 | * | 37 | * |
38 | * This function does a heapsort on the given array. You may provide a | 38 | * This function does a heapsort on the given array. You may provide a |
39 | * swap function optimized to your element type. | 39 | * swap_func function optimized to your element type. |
40 | * | 40 | * |
41 | * Sorting time is O(n log n) both on average and worst-case. While | 41 | * Sorting time is O(n log n) both on average and worst-case. While |
42 | * qsort is about 20% faster on average, it suffers from exploitable | 42 | * qsort is about 20% faster on average, it suffers from exploitable |
@@ -45,37 +45,39 @@ static void generic_swap(void *a, void *b, int size) | |||
45 | */ | 45 | */ |
46 | 46 | ||
47 | void sort(void *base, size_t num, size_t size, | 47 | void sort(void *base, size_t num, size_t size, |
48 | int (*cmp)(const void *, const void *), | 48 | int (*cmp_func)(const void *, const void *), |
49 | void (*swap)(void *, void *, int size)) | 49 | void (*swap_func)(void *, void *, int size)) |
50 | { | 50 | { |
51 | /* pre-scale counters for performance */ | 51 | /* pre-scale counters for performance */ |
52 | int i = (num/2 - 1) * size, n = num * size, c, r; | 52 | int i = (num/2 - 1) * size, n = num * size, c, r; |
53 | 53 | ||
54 | if (!swap) | 54 | if (!swap_func) |
55 | swap = (size == 4 ? u32_swap : generic_swap); | 55 | swap_func = (size == 4 ? u32_swap : generic_swap); |
56 | 56 | ||
57 | /* heapify */ | 57 | /* heapify */ |
58 | for ( ; i >= 0; i -= size) { | 58 | for ( ; i >= 0; i -= size) { |
59 | for (r = i; r * 2 + size < n; r = c) { | 59 | for (r = i; r * 2 + size < n; r = c) { |
60 | c = r * 2 + size; | 60 | c = r * 2 + size; |
61 | if (c < n - size && cmp(base + c, base + c + size) < 0) | 61 | if (c < n - size && |
62 | cmp_func(base + c, base + c + size) < 0) | ||
62 | c += size; | 63 | c += size; |
63 | if (cmp(base + r, base + c) >= 0) | 64 | if (cmp_func(base + r, base + c) >= 0) |
64 | break; | 65 | break; |
65 | swap(base + r, base + c, size); | 66 | swap_func(base + r, base + c, size); |
66 | } | 67 | } |
67 | } | 68 | } |
68 | 69 | ||
69 | /* sort */ | 70 | /* sort */ |
70 | for (i = n - size; i > 0; i -= size) { | 71 | for (i = n - size; i > 0; i -= size) { |
71 | swap(base, base + i, size); | 72 | swap_func(base, base + i, size); |
72 | for (r = 0; r * 2 + size < i; r = c) { | 73 | for (r = 0; r * 2 + size < i; r = c) { |
73 | c = r * 2 + size; | 74 | c = r * 2 + size; |
74 | if (c < i - size && cmp(base + c, base + c + size) < 0) | 75 | if (c < i - size && |
76 | cmp_func(base + c, base + c + size) < 0) | ||
75 | c += size; | 77 | c += size; |
76 | if (cmp(base + r, base + c) >= 0) | 78 | if (cmp_func(base + r, base + c) >= 0) |
77 | break; | 79 | break; |
78 | swap(base + r, base + c, size); | 80 | swap_func(base + r, base + c, size); |
79 | } | 81 | } |
80 | } | 82 | } |
81 | } | 83 | } |
diff --git a/mm/filemap.c b/mm/filemap.c index 2f55a1e2baf7..ceba0bd03662 100644 --- a/mm/filemap.c +++ b/mm/filemap.c | |||
@@ -460,7 +460,7 @@ int add_to_page_cache_locked(struct page *page, struct address_space *mapping, | |||
460 | VM_BUG_ON(!PageLocked(page)); | 460 | VM_BUG_ON(!PageLocked(page)); |
461 | 461 | ||
462 | error = mem_cgroup_cache_charge(page, current->mm, | 462 | error = mem_cgroup_cache_charge(page, current->mm, |
463 | gfp_mask & ~__GFP_HIGHMEM); | 463 | gfp_mask & GFP_RECLAIM_MASK); |
464 | if (error) | 464 | if (error) |
465 | goto out; | 465 | goto out; |
466 | 466 | ||
diff --git a/mm/memcontrol.c b/mm/memcontrol.c index 51ee96545579..e2996b80601f 100644 --- a/mm/memcontrol.c +++ b/mm/memcontrol.c | |||
@@ -21,11 +21,13 @@ | |||
21 | #include <linux/memcontrol.h> | 21 | #include <linux/memcontrol.h> |
22 | #include <linux/cgroup.h> | 22 | #include <linux/cgroup.h> |
23 | #include <linux/mm.h> | 23 | #include <linux/mm.h> |
24 | #include <linux/pagemap.h> | ||
24 | #include <linux/smp.h> | 25 | #include <linux/smp.h> |
25 | #include <linux/page-flags.h> | 26 | #include <linux/page-flags.h> |
26 | #include <linux/backing-dev.h> | 27 | #include <linux/backing-dev.h> |
27 | #include <linux/bit_spinlock.h> | 28 | #include <linux/bit_spinlock.h> |
28 | #include <linux/rcupdate.h> | 29 | #include <linux/rcupdate.h> |
30 | #include <linux/mutex.h> | ||
29 | #include <linux/slab.h> | 31 | #include <linux/slab.h> |
30 | #include <linux/swap.h> | 32 | #include <linux/swap.h> |
31 | #include <linux/spinlock.h> | 33 | #include <linux/spinlock.h> |
@@ -34,12 +36,23 @@ | |||
34 | #include <linux/vmalloc.h> | 36 | #include <linux/vmalloc.h> |
35 | #include <linux/mm_inline.h> | 37 | #include <linux/mm_inline.h> |
36 | #include <linux/page_cgroup.h> | 38 | #include <linux/page_cgroup.h> |
39 | #include "internal.h" | ||
37 | 40 | ||
38 | #include <asm/uaccess.h> | 41 | #include <asm/uaccess.h> |
39 | 42 | ||
40 | struct cgroup_subsys mem_cgroup_subsys __read_mostly; | 43 | struct cgroup_subsys mem_cgroup_subsys __read_mostly; |
41 | #define MEM_CGROUP_RECLAIM_RETRIES 5 | 44 | #define MEM_CGROUP_RECLAIM_RETRIES 5 |
42 | 45 | ||
46 | #ifdef CONFIG_CGROUP_MEM_RES_CTLR_SWAP | ||
47 | /* Turned on only when memory cgroup is enabled && really_do_swap_account = 0 */ | ||
48 | int do_swap_account __read_mostly; | ||
49 | static int really_do_swap_account __initdata = 1; /* for remember boot option*/ | ||
50 | #else | ||
51 | #define do_swap_account (0) | ||
52 | #endif | ||
53 | |||
54 | static DEFINE_MUTEX(memcg_tasklist); /* can be hold under cgroup_mutex */ | ||
55 | |||
43 | /* | 56 | /* |
44 | * Statistics for memory cgroup. | 57 | * Statistics for memory cgroup. |
45 | */ | 58 | */ |
@@ -60,7 +73,7 @@ struct mem_cgroup_stat_cpu { | |||
60 | } ____cacheline_aligned_in_smp; | 73 | } ____cacheline_aligned_in_smp; |
61 | 74 | ||
62 | struct mem_cgroup_stat { | 75 | struct mem_cgroup_stat { |
63 | struct mem_cgroup_stat_cpu cpustat[NR_CPUS]; | 76 | struct mem_cgroup_stat_cpu cpustat[0]; |
64 | }; | 77 | }; |
65 | 78 | ||
66 | /* | 79 | /* |
@@ -89,9 +102,10 @@ struct mem_cgroup_per_zone { | |||
89 | /* | 102 | /* |
90 | * spin_lock to protect the per cgroup LRU | 103 | * spin_lock to protect the per cgroup LRU |
91 | */ | 104 | */ |
92 | spinlock_t lru_lock; | ||
93 | struct list_head lists[NR_LRU_LISTS]; | 105 | struct list_head lists[NR_LRU_LISTS]; |
94 | unsigned long count[NR_LRU_LISTS]; | 106 | unsigned long count[NR_LRU_LISTS]; |
107 | |||
108 | struct zone_reclaim_stat reclaim_stat; | ||
95 | }; | 109 | }; |
96 | /* Macro for accessing counter */ | 110 | /* Macro for accessing counter */ |
97 | #define MEM_CGROUP_ZSTAT(mz, idx) ((mz)->count[(idx)]) | 111 | #define MEM_CGROUP_ZSTAT(mz, idx) ((mz)->count[(idx)]) |
@@ -122,44 +136,73 @@ struct mem_cgroup { | |||
122 | */ | 136 | */ |
123 | struct res_counter res; | 137 | struct res_counter res; |
124 | /* | 138 | /* |
139 | * the counter to account for mem+swap usage. | ||
140 | */ | ||
141 | struct res_counter memsw; | ||
142 | /* | ||
125 | * Per cgroup active and inactive list, similar to the | 143 | * Per cgroup active and inactive list, similar to the |
126 | * per zone LRU lists. | 144 | * per zone LRU lists. |
127 | */ | 145 | */ |
128 | struct mem_cgroup_lru_info info; | 146 | struct mem_cgroup_lru_info info; |
129 | 147 | ||
148 | /* | ||
149 | protect against reclaim related member. | ||
150 | */ | ||
151 | spinlock_t reclaim_param_lock; | ||
152 | |||
130 | int prev_priority; /* for recording reclaim priority */ | 153 | int prev_priority; /* for recording reclaim priority */ |
154 | |||
155 | /* | ||
156 | * While reclaiming in a hiearchy, we cache the last child we | ||
157 | * reclaimed from. Protected by hierarchy_mutex | ||
158 | */ | ||
159 | struct mem_cgroup *last_scanned_child; | ||
131 | /* | 160 | /* |
132 | * statistics. | 161 | * Should the accounting and control be hierarchical, per subtree? |
162 | */ | ||
163 | bool use_hierarchy; | ||
164 | unsigned long last_oom_jiffies; | ||
165 | atomic_t refcnt; | ||
166 | |||
167 | unsigned int swappiness; | ||
168 | |||
169 | /* | ||
170 | * statistics. This must be placed at the end of memcg. | ||
133 | */ | 171 | */ |
134 | struct mem_cgroup_stat stat; | 172 | struct mem_cgroup_stat stat; |
135 | }; | 173 | }; |
136 | static struct mem_cgroup init_mem_cgroup; | ||
137 | 174 | ||
138 | enum charge_type { | 175 | enum charge_type { |
139 | MEM_CGROUP_CHARGE_TYPE_CACHE = 0, | 176 | MEM_CGROUP_CHARGE_TYPE_CACHE = 0, |
140 | MEM_CGROUP_CHARGE_TYPE_MAPPED, | 177 | MEM_CGROUP_CHARGE_TYPE_MAPPED, |
141 | MEM_CGROUP_CHARGE_TYPE_SHMEM, /* used by page migration of shmem */ | 178 | MEM_CGROUP_CHARGE_TYPE_SHMEM, /* used by page migration of shmem */ |
142 | MEM_CGROUP_CHARGE_TYPE_FORCE, /* used by force_empty */ | 179 | MEM_CGROUP_CHARGE_TYPE_FORCE, /* used by force_empty */ |
180 | MEM_CGROUP_CHARGE_TYPE_SWAPOUT, /* for accounting swapcache */ | ||
143 | NR_CHARGE_TYPE, | 181 | NR_CHARGE_TYPE, |
144 | }; | 182 | }; |
145 | 183 | ||
146 | /* only for here (for easy reading.) */ | 184 | /* only for here (for easy reading.) */ |
147 | #define PCGF_CACHE (1UL << PCG_CACHE) | 185 | #define PCGF_CACHE (1UL << PCG_CACHE) |
148 | #define PCGF_USED (1UL << PCG_USED) | 186 | #define PCGF_USED (1UL << PCG_USED) |
149 | #define PCGF_ACTIVE (1UL << PCG_ACTIVE) | ||
150 | #define PCGF_LOCK (1UL << PCG_LOCK) | 187 | #define PCGF_LOCK (1UL << PCG_LOCK) |
151 | #define PCGF_FILE (1UL << PCG_FILE) | ||
152 | static const unsigned long | 188 | static const unsigned long |
153 | pcg_default_flags[NR_CHARGE_TYPE] = { | 189 | pcg_default_flags[NR_CHARGE_TYPE] = { |
154 | PCGF_CACHE | PCGF_FILE | PCGF_USED | PCGF_LOCK, /* File Cache */ | 190 | PCGF_CACHE | PCGF_USED | PCGF_LOCK, /* File Cache */ |
155 | PCGF_ACTIVE | PCGF_USED | PCGF_LOCK, /* Anon */ | 191 | PCGF_USED | PCGF_LOCK, /* Anon */ |
156 | PCGF_ACTIVE | PCGF_CACHE | PCGF_USED | PCGF_LOCK, /* Shmem */ | 192 | PCGF_CACHE | PCGF_USED | PCGF_LOCK, /* Shmem */ |
157 | 0, /* FORCE */ | 193 | 0, /* FORCE */ |
158 | }; | 194 | }; |
159 | 195 | ||
160 | /* | 196 | /* for encoding cft->private value on file */ |
161 | * Always modified under lru lock. Then, not necessary to preempt_disable() | 197 | #define _MEM (0) |
162 | */ | 198 | #define _MEMSWAP (1) |
199 | #define MEMFILE_PRIVATE(x, val) (((x) << 16) | (val)) | ||
200 | #define MEMFILE_TYPE(val) (((val) >> 16) & 0xffff) | ||
201 | #define MEMFILE_ATTR(val) ((val) & 0xffff) | ||
202 | |||
203 | static void mem_cgroup_get(struct mem_cgroup *mem); | ||
204 | static void mem_cgroup_put(struct mem_cgroup *mem); | ||
205 | |||
163 | static void mem_cgroup_charge_statistics(struct mem_cgroup *mem, | 206 | static void mem_cgroup_charge_statistics(struct mem_cgroup *mem, |
164 | struct page_cgroup *pc, | 207 | struct page_cgroup *pc, |
165 | bool charge) | 208 | bool charge) |
@@ -167,10 +210,9 @@ static void mem_cgroup_charge_statistics(struct mem_cgroup *mem, | |||
167 | int val = (charge)? 1 : -1; | 210 | int val = (charge)? 1 : -1; |
168 | struct mem_cgroup_stat *stat = &mem->stat; | 211 | struct mem_cgroup_stat *stat = &mem->stat; |
169 | struct mem_cgroup_stat_cpu *cpustat; | 212 | struct mem_cgroup_stat_cpu *cpustat; |
213 | int cpu = get_cpu(); | ||
170 | 214 | ||
171 | VM_BUG_ON(!irqs_disabled()); | 215 | cpustat = &stat->cpustat[cpu]; |
172 | |||
173 | cpustat = &stat->cpustat[smp_processor_id()]; | ||
174 | if (PageCgroupCache(pc)) | 216 | if (PageCgroupCache(pc)) |
175 | __mem_cgroup_stat_add_safe(cpustat, MEM_CGROUP_STAT_CACHE, val); | 217 | __mem_cgroup_stat_add_safe(cpustat, MEM_CGROUP_STAT_CACHE, val); |
176 | else | 218 | else |
@@ -182,6 +224,7 @@ static void mem_cgroup_charge_statistics(struct mem_cgroup *mem, | |||
182 | else | 224 | else |
183 | __mem_cgroup_stat_add_safe(cpustat, | 225 | __mem_cgroup_stat_add_safe(cpustat, |
184 | MEM_CGROUP_STAT_PGPGOUT_COUNT, 1); | 226 | MEM_CGROUP_STAT_PGPGOUT_COUNT, 1); |
227 | put_cpu(); | ||
185 | } | 228 | } |
186 | 229 | ||
187 | static struct mem_cgroup_per_zone * | 230 | static struct mem_cgroup_per_zone * |
@@ -197,6 +240,9 @@ page_cgroup_zoneinfo(struct page_cgroup *pc) | |||
197 | int nid = page_cgroup_nid(pc); | 240 | int nid = page_cgroup_nid(pc); |
198 | int zid = page_cgroup_zid(pc); | 241 | int zid = page_cgroup_zid(pc); |
199 | 242 | ||
243 | if (!mem) | ||
244 | return NULL; | ||
245 | |||
200 | return mem_cgroup_zoneinfo(mem, nid, zid); | 246 | return mem_cgroup_zoneinfo(mem, nid, zid); |
201 | } | 247 | } |
202 | 248 | ||
@@ -236,77 +282,152 @@ struct mem_cgroup *mem_cgroup_from_task(struct task_struct *p) | |||
236 | struct mem_cgroup, css); | 282 | struct mem_cgroup, css); |
237 | } | 283 | } |
238 | 284 | ||
239 | static void __mem_cgroup_remove_list(struct mem_cgroup_per_zone *mz, | 285 | static struct mem_cgroup *try_get_mem_cgroup_from_mm(struct mm_struct *mm) |
240 | struct page_cgroup *pc) | ||
241 | { | 286 | { |
242 | int lru = LRU_BASE; | 287 | struct mem_cgroup *mem = NULL; |
288 | /* | ||
289 | * Because we have no locks, mm->owner's may be being moved to other | ||
290 | * cgroup. We use css_tryget() here even if this looks | ||
291 | * pessimistic (rather than adding locks here). | ||
292 | */ | ||
293 | rcu_read_lock(); | ||
294 | do { | ||
295 | mem = mem_cgroup_from_task(rcu_dereference(mm->owner)); | ||
296 | if (unlikely(!mem)) | ||
297 | break; | ||
298 | } while (!css_tryget(&mem->css)); | ||
299 | rcu_read_unlock(); | ||
300 | return mem; | ||
301 | } | ||
243 | 302 | ||
244 | if (PageCgroupUnevictable(pc)) | 303 | static bool mem_cgroup_is_obsolete(struct mem_cgroup *mem) |
245 | lru = LRU_UNEVICTABLE; | 304 | { |
246 | else { | 305 | if (!mem) |
247 | if (PageCgroupActive(pc)) | 306 | return true; |
248 | lru += LRU_ACTIVE; | 307 | return css_is_removed(&mem->css); |
249 | if (PageCgroupFile(pc)) | 308 | } |
250 | lru += LRU_FILE; | ||
251 | } | ||
252 | 309 | ||
253 | MEM_CGROUP_ZSTAT(mz, lru) -= 1; | 310 | /* |
311 | * Following LRU functions are allowed to be used without PCG_LOCK. | ||
312 | * Operations are called by routine of global LRU independently from memcg. | ||
313 | * What we have to take care of here is validness of pc->mem_cgroup. | ||
314 | * | ||
315 | * Changes to pc->mem_cgroup happens when | ||
316 | * 1. charge | ||
317 | * 2. moving account | ||
318 | * In typical case, "charge" is done before add-to-lru. Exception is SwapCache. | ||
319 | * It is added to LRU before charge. | ||
320 | * If PCG_USED bit is not set, page_cgroup is not added to this private LRU. | ||
321 | * When moving account, the page is not on LRU. It's isolated. | ||
322 | */ | ||
254 | 323 | ||
255 | mem_cgroup_charge_statistics(pc->mem_cgroup, pc, false); | 324 | void mem_cgroup_del_lru_list(struct page *page, enum lru_list lru) |
256 | list_del(&pc->lru); | 325 | { |
326 | struct page_cgroup *pc; | ||
327 | struct mem_cgroup *mem; | ||
328 | struct mem_cgroup_per_zone *mz; | ||
329 | |||
330 | if (mem_cgroup_disabled()) | ||
331 | return; | ||
332 | pc = lookup_page_cgroup(page); | ||
333 | /* can happen while we handle swapcache. */ | ||
334 | if (list_empty(&pc->lru) || !pc->mem_cgroup) | ||
335 | return; | ||
336 | /* | ||
337 | * We don't check PCG_USED bit. It's cleared when the "page" is finally | ||
338 | * removed from global LRU. | ||
339 | */ | ||
340 | mz = page_cgroup_zoneinfo(pc); | ||
341 | mem = pc->mem_cgroup; | ||
342 | MEM_CGROUP_ZSTAT(mz, lru) -= 1; | ||
343 | list_del_init(&pc->lru); | ||
344 | return; | ||
257 | } | 345 | } |
258 | 346 | ||
259 | static void __mem_cgroup_add_list(struct mem_cgroup_per_zone *mz, | 347 | void mem_cgroup_del_lru(struct page *page) |
260 | struct page_cgroup *pc) | ||
261 | { | 348 | { |
262 | int lru = LRU_BASE; | 349 | mem_cgroup_del_lru_list(page, page_lru(page)); |
350 | } | ||
263 | 351 | ||
264 | if (PageCgroupUnevictable(pc)) | 352 | void mem_cgroup_rotate_lru_list(struct page *page, enum lru_list lru) |
265 | lru = LRU_UNEVICTABLE; | 353 | { |
266 | else { | 354 | struct mem_cgroup_per_zone *mz; |
267 | if (PageCgroupActive(pc)) | 355 | struct page_cgroup *pc; |
268 | lru += LRU_ACTIVE; | ||
269 | if (PageCgroupFile(pc)) | ||
270 | lru += LRU_FILE; | ||
271 | } | ||
272 | 356 | ||
273 | MEM_CGROUP_ZSTAT(mz, lru) += 1; | 357 | if (mem_cgroup_disabled()) |
274 | list_add(&pc->lru, &mz->lists[lru]); | 358 | return; |
275 | 359 | ||
276 | mem_cgroup_charge_statistics(pc->mem_cgroup, pc, true); | 360 | pc = lookup_page_cgroup(page); |
361 | smp_rmb(); | ||
362 | /* unused page is not rotated. */ | ||
363 | if (!PageCgroupUsed(pc)) | ||
364 | return; | ||
365 | mz = page_cgroup_zoneinfo(pc); | ||
366 | list_move(&pc->lru, &mz->lists[lru]); | ||
277 | } | 367 | } |
278 | 368 | ||
279 | static void __mem_cgroup_move_lists(struct page_cgroup *pc, enum lru_list lru) | 369 | void mem_cgroup_add_lru_list(struct page *page, enum lru_list lru) |
280 | { | 370 | { |
281 | struct mem_cgroup_per_zone *mz = page_cgroup_zoneinfo(pc); | 371 | struct page_cgroup *pc; |
282 | int active = PageCgroupActive(pc); | 372 | struct mem_cgroup_per_zone *mz; |
283 | int file = PageCgroupFile(pc); | ||
284 | int unevictable = PageCgroupUnevictable(pc); | ||
285 | enum lru_list from = unevictable ? LRU_UNEVICTABLE : | ||
286 | (LRU_FILE * !!file + !!active); | ||
287 | 373 | ||
288 | if (lru == from) | 374 | if (mem_cgroup_disabled()) |
375 | return; | ||
376 | pc = lookup_page_cgroup(page); | ||
377 | /* barrier to sync with "charge" */ | ||
378 | smp_rmb(); | ||
379 | if (!PageCgroupUsed(pc)) | ||
289 | return; | 380 | return; |
290 | 381 | ||
291 | MEM_CGROUP_ZSTAT(mz, from) -= 1; | 382 | mz = page_cgroup_zoneinfo(pc); |
383 | MEM_CGROUP_ZSTAT(mz, lru) += 1; | ||
384 | list_add(&pc->lru, &mz->lists[lru]); | ||
385 | } | ||
386 | |||
387 | /* | ||
388 | * At handling SwapCache, pc->mem_cgroup may be changed while it's linked to | ||
389 | * lru because the page may.be reused after it's fully uncharged (because of | ||
390 | * SwapCache behavior).To handle that, unlink page_cgroup from LRU when charge | ||
391 | * it again. This function is only used to charge SwapCache. It's done under | ||
392 | * lock_page and expected that zone->lru_lock is never held. | ||
393 | */ | ||
394 | static void mem_cgroup_lru_del_before_commit_swapcache(struct page *page) | ||
395 | { | ||
396 | unsigned long flags; | ||
397 | struct zone *zone = page_zone(page); | ||
398 | struct page_cgroup *pc = lookup_page_cgroup(page); | ||
399 | |||
400 | spin_lock_irqsave(&zone->lru_lock, flags); | ||
292 | /* | 401 | /* |
293 | * However this is done under mz->lru_lock, another flags, which | 402 | * Forget old LRU when this page_cgroup is *not* used. This Used bit |
294 | * are not related to LRU, will be modified from out-of-lock. | 403 | * is guarded by lock_page() because the page is SwapCache. |
295 | * We have to use atomic set/clear flags. | ||
296 | */ | 404 | */ |
297 | if (is_unevictable_lru(lru)) { | 405 | if (!PageCgroupUsed(pc)) |
298 | ClearPageCgroupActive(pc); | 406 | mem_cgroup_del_lru_list(page, page_lru(page)); |
299 | SetPageCgroupUnevictable(pc); | 407 | spin_unlock_irqrestore(&zone->lru_lock, flags); |
300 | } else { | 408 | } |
301 | if (is_active_lru(lru)) | ||
302 | SetPageCgroupActive(pc); | ||
303 | else | ||
304 | ClearPageCgroupActive(pc); | ||
305 | ClearPageCgroupUnevictable(pc); | ||
306 | } | ||
307 | 409 | ||
308 | MEM_CGROUP_ZSTAT(mz, lru) += 1; | 410 | static void mem_cgroup_lru_add_after_commit_swapcache(struct page *page) |
309 | list_move(&pc->lru, &mz->lists[lru]); | 411 | { |
412 | unsigned long flags; | ||
413 | struct zone *zone = page_zone(page); | ||
414 | struct page_cgroup *pc = lookup_page_cgroup(page); | ||
415 | |||
416 | spin_lock_irqsave(&zone->lru_lock, flags); | ||
417 | /* link when the page is linked to LRU but page_cgroup isn't */ | ||
418 | if (PageLRU(page) && list_empty(&pc->lru)) | ||
419 | mem_cgroup_add_lru_list(page, page_lru(page)); | ||
420 | spin_unlock_irqrestore(&zone->lru_lock, flags); | ||
421 | } | ||
422 | |||
423 | |||
424 | void mem_cgroup_move_lists(struct page *page, | ||
425 | enum lru_list from, enum lru_list to) | ||
426 | { | ||
427 | if (mem_cgroup_disabled()) | ||
428 | return; | ||
429 | mem_cgroup_del_lru_list(page, from); | ||
430 | mem_cgroup_add_lru_list(page, to); | ||
310 | } | 431 | } |
311 | 432 | ||
312 | int task_in_mem_cgroup(struct task_struct *task, const struct mem_cgroup *mem) | 433 | int task_in_mem_cgroup(struct task_struct *task, const struct mem_cgroup *mem) |
@@ -320,37 +441,6 @@ int task_in_mem_cgroup(struct task_struct *task, const struct mem_cgroup *mem) | |||
320 | } | 441 | } |
321 | 442 | ||
322 | /* | 443 | /* |
323 | * This routine assumes that the appropriate zone's lru lock is already held | ||
324 | */ | ||
325 | void mem_cgroup_move_lists(struct page *page, enum lru_list lru) | ||
326 | { | ||
327 | struct page_cgroup *pc; | ||
328 | struct mem_cgroup_per_zone *mz; | ||
329 | unsigned long flags; | ||
330 | |||
331 | if (mem_cgroup_subsys.disabled) | ||
332 | return; | ||
333 | |||
334 | /* | ||
335 | * We cannot lock_page_cgroup while holding zone's lru_lock, | ||
336 | * because other holders of lock_page_cgroup can be interrupted | ||
337 | * with an attempt to rotate_reclaimable_page. But we cannot | ||
338 | * safely get to page_cgroup without it, so just try_lock it: | ||
339 | * mem_cgroup_isolate_pages allows for page left on wrong list. | ||
340 | */ | ||
341 | pc = lookup_page_cgroup(page); | ||
342 | if (!trylock_page_cgroup(pc)) | ||
343 | return; | ||
344 | if (pc && PageCgroupUsed(pc)) { | ||
345 | mz = page_cgroup_zoneinfo(pc); | ||
346 | spin_lock_irqsave(&mz->lru_lock, flags); | ||
347 | __mem_cgroup_move_lists(pc, lru); | ||
348 | spin_unlock_irqrestore(&mz->lru_lock, flags); | ||
349 | } | ||
350 | unlock_page_cgroup(pc); | ||
351 | } | ||
352 | |||
353 | /* | ||
354 | * Calculate mapped_ratio under memory controller. This will be used in | 444 | * Calculate mapped_ratio under memory controller. This will be used in |
355 | * vmscan.c for deteremining we have to reclaim mapped pages. | 445 | * vmscan.c for deteremining we have to reclaim mapped pages. |
356 | */ | 446 | */ |
@@ -372,39 +462,108 @@ int mem_cgroup_calc_mapped_ratio(struct mem_cgroup *mem) | |||
372 | */ | 462 | */ |
373 | int mem_cgroup_get_reclaim_priority(struct mem_cgroup *mem) | 463 | int mem_cgroup_get_reclaim_priority(struct mem_cgroup *mem) |
374 | { | 464 | { |
375 | return mem->prev_priority; | 465 | int prev_priority; |
466 | |||
467 | spin_lock(&mem->reclaim_param_lock); | ||
468 | prev_priority = mem->prev_priority; | ||
469 | spin_unlock(&mem->reclaim_param_lock); | ||
470 | |||
471 | return prev_priority; | ||
376 | } | 472 | } |
377 | 473 | ||
378 | void mem_cgroup_note_reclaim_priority(struct mem_cgroup *mem, int priority) | 474 | void mem_cgroup_note_reclaim_priority(struct mem_cgroup *mem, int priority) |
379 | { | 475 | { |
476 | spin_lock(&mem->reclaim_param_lock); | ||
380 | if (priority < mem->prev_priority) | 477 | if (priority < mem->prev_priority) |
381 | mem->prev_priority = priority; | 478 | mem->prev_priority = priority; |
479 | spin_unlock(&mem->reclaim_param_lock); | ||
382 | } | 480 | } |
383 | 481 | ||
384 | void mem_cgroup_record_reclaim_priority(struct mem_cgroup *mem, int priority) | 482 | void mem_cgroup_record_reclaim_priority(struct mem_cgroup *mem, int priority) |
385 | { | 483 | { |
484 | spin_lock(&mem->reclaim_param_lock); | ||
386 | mem->prev_priority = priority; | 485 | mem->prev_priority = priority; |
486 | spin_unlock(&mem->reclaim_param_lock); | ||
387 | } | 487 | } |
388 | 488 | ||
389 | /* | 489 | static int calc_inactive_ratio(struct mem_cgroup *memcg, unsigned long *present_pages) |
390 | * Calculate # of pages to be scanned in this priority/zone. | 490 | { |
391 | * See also vmscan.c | 491 | unsigned long active; |
392 | * | 492 | unsigned long inactive; |
393 | * priority starts from "DEF_PRIORITY" and decremented in each loop. | 493 | unsigned long gb; |
394 | * (see include/linux/mmzone.h) | 494 | unsigned long inactive_ratio; |
395 | */ | 495 | |
496 | inactive = mem_cgroup_get_all_zonestat(memcg, LRU_INACTIVE_ANON); | ||
497 | active = mem_cgroup_get_all_zonestat(memcg, LRU_ACTIVE_ANON); | ||
498 | |||
499 | gb = (inactive + active) >> (30 - PAGE_SHIFT); | ||
500 | if (gb) | ||
501 | inactive_ratio = int_sqrt(10 * gb); | ||
502 | else | ||
503 | inactive_ratio = 1; | ||
504 | |||
505 | if (present_pages) { | ||
506 | present_pages[0] = inactive; | ||
507 | present_pages[1] = active; | ||
508 | } | ||
509 | |||
510 | return inactive_ratio; | ||
511 | } | ||
512 | |||
513 | int mem_cgroup_inactive_anon_is_low(struct mem_cgroup *memcg) | ||
514 | { | ||
515 | unsigned long active; | ||
516 | unsigned long inactive; | ||
517 | unsigned long present_pages[2]; | ||
518 | unsigned long inactive_ratio; | ||
396 | 519 | ||
397 | long mem_cgroup_calc_reclaim(struct mem_cgroup *mem, struct zone *zone, | 520 | inactive_ratio = calc_inactive_ratio(memcg, present_pages); |
398 | int priority, enum lru_list lru) | 521 | |
522 | inactive = present_pages[0]; | ||
523 | active = present_pages[1]; | ||
524 | |||
525 | if (inactive * inactive_ratio < active) | ||
526 | return 1; | ||
527 | |||
528 | return 0; | ||
529 | } | ||
530 | |||
531 | unsigned long mem_cgroup_zone_nr_pages(struct mem_cgroup *memcg, | ||
532 | struct zone *zone, | ||
533 | enum lru_list lru) | ||
399 | { | 534 | { |
400 | long nr_pages; | ||
401 | int nid = zone->zone_pgdat->node_id; | 535 | int nid = zone->zone_pgdat->node_id; |
402 | int zid = zone_idx(zone); | 536 | int zid = zone_idx(zone); |
403 | struct mem_cgroup_per_zone *mz = mem_cgroup_zoneinfo(mem, nid, zid); | 537 | struct mem_cgroup_per_zone *mz = mem_cgroup_zoneinfo(memcg, nid, zid); |
404 | 538 | ||
405 | nr_pages = MEM_CGROUP_ZSTAT(mz, lru); | 539 | return MEM_CGROUP_ZSTAT(mz, lru); |
540 | } | ||
406 | 541 | ||
407 | return (nr_pages >> priority); | 542 | struct zone_reclaim_stat *mem_cgroup_get_reclaim_stat(struct mem_cgroup *memcg, |
543 | struct zone *zone) | ||
544 | { | ||
545 | int nid = zone->zone_pgdat->node_id; | ||
546 | int zid = zone_idx(zone); | ||
547 | struct mem_cgroup_per_zone *mz = mem_cgroup_zoneinfo(memcg, nid, zid); | ||
548 | |||
549 | return &mz->reclaim_stat; | ||
550 | } | ||
551 | |||
552 | struct zone_reclaim_stat * | ||
553 | mem_cgroup_get_reclaim_stat_from_page(struct page *page) | ||
554 | { | ||
555 | struct page_cgroup *pc; | ||
556 | struct mem_cgroup_per_zone *mz; | ||
557 | |||
558 | if (mem_cgroup_disabled()) | ||
559 | return NULL; | ||
560 | |||
561 | pc = lookup_page_cgroup(page); | ||
562 | mz = page_cgroup_zoneinfo(pc); | ||
563 | if (!mz) | ||
564 | return NULL; | ||
565 | |||
566 | return &mz->reclaim_stat; | ||
408 | } | 567 | } |
409 | 568 | ||
410 | unsigned long mem_cgroup_isolate_pages(unsigned long nr_to_scan, | 569 | unsigned long mem_cgroup_isolate_pages(unsigned long nr_to_scan, |
@@ -429,95 +588,281 @@ unsigned long mem_cgroup_isolate_pages(unsigned long nr_to_scan, | |||
429 | mz = mem_cgroup_zoneinfo(mem_cont, nid, zid); | 588 | mz = mem_cgroup_zoneinfo(mem_cont, nid, zid); |
430 | src = &mz->lists[lru]; | 589 | src = &mz->lists[lru]; |
431 | 590 | ||
432 | spin_lock(&mz->lru_lock); | ||
433 | scan = 0; | 591 | scan = 0; |
434 | list_for_each_entry_safe_reverse(pc, tmp, src, lru) { | 592 | list_for_each_entry_safe_reverse(pc, tmp, src, lru) { |
435 | if (scan >= nr_to_scan) | 593 | if (scan >= nr_to_scan) |
436 | break; | 594 | break; |
595 | |||
596 | page = pc->page; | ||
437 | if (unlikely(!PageCgroupUsed(pc))) | 597 | if (unlikely(!PageCgroupUsed(pc))) |
438 | continue; | 598 | continue; |
439 | page = pc->page; | ||
440 | |||
441 | if (unlikely(!PageLRU(page))) | 599 | if (unlikely(!PageLRU(page))) |
442 | continue; | 600 | continue; |
443 | 601 | ||
444 | /* | ||
445 | * TODO: play better with lumpy reclaim, grabbing anything. | ||
446 | */ | ||
447 | if (PageUnevictable(page) || | ||
448 | (PageActive(page) && !active) || | ||
449 | (!PageActive(page) && active)) { | ||
450 | __mem_cgroup_move_lists(pc, page_lru(page)); | ||
451 | continue; | ||
452 | } | ||
453 | |||
454 | scan++; | 602 | scan++; |
455 | list_move(&pc->lru, &pc_list); | ||
456 | |||
457 | if (__isolate_lru_page(page, mode, file) == 0) { | 603 | if (__isolate_lru_page(page, mode, file) == 0) { |
458 | list_move(&page->lru, dst); | 604 | list_move(&page->lru, dst); |
459 | nr_taken++; | 605 | nr_taken++; |
460 | } | 606 | } |
461 | } | 607 | } |
462 | 608 | ||
463 | list_splice(&pc_list, src); | ||
464 | spin_unlock(&mz->lru_lock); | ||
465 | |||
466 | *scanned = scan; | 609 | *scanned = scan; |
467 | return nr_taken; | 610 | return nr_taken; |
468 | } | 611 | } |
469 | 612 | ||
613 | #define mem_cgroup_from_res_counter(counter, member) \ | ||
614 | container_of(counter, struct mem_cgroup, member) | ||
615 | |||
470 | /* | 616 | /* |
471 | * Charge the memory controller for page usage. | 617 | * This routine finds the DFS walk successor. This routine should be |
472 | * Return | 618 | * called with hierarchy_mutex held |
473 | * 0 if the charge was successful | ||
474 | * < 0 if the cgroup is over its limit | ||
475 | */ | 619 | */ |
476 | static int mem_cgroup_charge_common(struct page *page, struct mm_struct *mm, | 620 | static struct mem_cgroup * |
477 | gfp_t gfp_mask, enum charge_type ctype, | 621 | mem_cgroup_get_next_node(struct mem_cgroup *curr, struct mem_cgroup *root_mem) |
478 | struct mem_cgroup *memcg) | ||
479 | { | 622 | { |
623 | struct cgroup *cgroup, *curr_cgroup, *root_cgroup; | ||
624 | |||
625 | curr_cgroup = curr->css.cgroup; | ||
626 | root_cgroup = root_mem->css.cgroup; | ||
627 | |||
628 | if (!list_empty(&curr_cgroup->children)) { | ||
629 | /* | ||
630 | * Walk down to children | ||
631 | */ | ||
632 | mem_cgroup_put(curr); | ||
633 | cgroup = list_entry(curr_cgroup->children.next, | ||
634 | struct cgroup, sibling); | ||
635 | curr = mem_cgroup_from_cont(cgroup); | ||
636 | mem_cgroup_get(curr); | ||
637 | goto done; | ||
638 | } | ||
639 | |||
640 | visit_parent: | ||
641 | if (curr_cgroup == root_cgroup) { | ||
642 | mem_cgroup_put(curr); | ||
643 | curr = root_mem; | ||
644 | mem_cgroup_get(curr); | ||
645 | goto done; | ||
646 | } | ||
647 | |||
648 | /* | ||
649 | * Goto next sibling | ||
650 | */ | ||
651 | if (curr_cgroup->sibling.next != &curr_cgroup->parent->children) { | ||
652 | mem_cgroup_put(curr); | ||
653 | cgroup = list_entry(curr_cgroup->sibling.next, struct cgroup, | ||
654 | sibling); | ||
655 | curr = mem_cgroup_from_cont(cgroup); | ||
656 | mem_cgroup_get(curr); | ||
657 | goto done; | ||
658 | } | ||
659 | |||
660 | /* | ||
661 | * Go up to next parent and next parent's sibling if need be | ||
662 | */ | ||
663 | curr_cgroup = curr_cgroup->parent; | ||
664 | goto visit_parent; | ||
665 | |||
666 | done: | ||
667 | root_mem->last_scanned_child = curr; | ||
668 | return curr; | ||
669 | } | ||
670 | |||
671 | /* | ||
672 | * Visit the first child (need not be the first child as per the ordering | ||
673 | * of the cgroup list, since we track last_scanned_child) of @mem and use | ||
674 | * that to reclaim free pages from. | ||
675 | */ | ||
676 | static struct mem_cgroup * | ||
677 | mem_cgroup_get_first_node(struct mem_cgroup *root_mem) | ||
678 | { | ||
679 | struct cgroup *cgroup; | ||
680 | struct mem_cgroup *ret; | ||
681 | bool obsolete; | ||
682 | |||
683 | obsolete = mem_cgroup_is_obsolete(root_mem->last_scanned_child); | ||
684 | |||
685 | /* | ||
686 | * Scan all children under the mem_cgroup mem | ||
687 | */ | ||
688 | mutex_lock(&mem_cgroup_subsys.hierarchy_mutex); | ||
689 | if (list_empty(&root_mem->css.cgroup->children)) { | ||
690 | ret = root_mem; | ||
691 | goto done; | ||
692 | } | ||
693 | |||
694 | if (!root_mem->last_scanned_child || obsolete) { | ||
695 | |||
696 | if (obsolete && root_mem->last_scanned_child) | ||
697 | mem_cgroup_put(root_mem->last_scanned_child); | ||
698 | |||
699 | cgroup = list_first_entry(&root_mem->css.cgroup->children, | ||
700 | struct cgroup, sibling); | ||
701 | ret = mem_cgroup_from_cont(cgroup); | ||
702 | mem_cgroup_get(ret); | ||
703 | } else | ||
704 | ret = mem_cgroup_get_next_node(root_mem->last_scanned_child, | ||
705 | root_mem); | ||
706 | |||
707 | done: | ||
708 | root_mem->last_scanned_child = ret; | ||
709 | mutex_unlock(&mem_cgroup_subsys.hierarchy_mutex); | ||
710 | return ret; | ||
711 | } | ||
712 | |||
713 | static bool mem_cgroup_check_under_limit(struct mem_cgroup *mem) | ||
714 | { | ||
715 | if (do_swap_account) { | ||
716 | if (res_counter_check_under_limit(&mem->res) && | ||
717 | res_counter_check_under_limit(&mem->memsw)) | ||
718 | return true; | ||
719 | } else | ||
720 | if (res_counter_check_under_limit(&mem->res)) | ||
721 | return true; | ||
722 | return false; | ||
723 | } | ||
724 | |||
725 | static unsigned int get_swappiness(struct mem_cgroup *memcg) | ||
726 | { | ||
727 | struct cgroup *cgrp = memcg->css.cgroup; | ||
728 | unsigned int swappiness; | ||
729 | |||
730 | /* root ? */ | ||
731 | if (cgrp->parent == NULL) | ||
732 | return vm_swappiness; | ||
733 | |||
734 | spin_lock(&memcg->reclaim_param_lock); | ||
735 | swappiness = memcg->swappiness; | ||
736 | spin_unlock(&memcg->reclaim_param_lock); | ||
737 | |||
738 | return swappiness; | ||
739 | } | ||
740 | |||
741 | /* | ||
742 | * Dance down the hierarchy if needed to reclaim memory. We remember the | ||
743 | * last child we reclaimed from, so that we don't end up penalizing | ||
744 | * one child extensively based on its position in the children list. | ||
745 | * | ||
746 | * root_mem is the original ancestor that we've been reclaim from. | ||
747 | */ | ||
748 | static int mem_cgroup_hierarchical_reclaim(struct mem_cgroup *root_mem, | ||
749 | gfp_t gfp_mask, bool noswap) | ||
750 | { | ||
751 | struct mem_cgroup *next_mem; | ||
752 | int ret = 0; | ||
753 | |||
754 | /* | ||
755 | * Reclaim unconditionally and don't check for return value. | ||
756 | * We need to reclaim in the current group and down the tree. | ||
757 | * One might think about checking for children before reclaiming, | ||
758 | * but there might be left over accounting, even after children | ||
759 | * have left. | ||
760 | */ | ||
761 | ret = try_to_free_mem_cgroup_pages(root_mem, gfp_mask, noswap, | ||
762 | get_swappiness(root_mem)); | ||
763 | if (mem_cgroup_check_under_limit(root_mem)) | ||
764 | return 0; | ||
765 | if (!root_mem->use_hierarchy) | ||
766 | return ret; | ||
767 | |||
768 | next_mem = mem_cgroup_get_first_node(root_mem); | ||
769 | |||
770 | while (next_mem != root_mem) { | ||
771 | if (mem_cgroup_is_obsolete(next_mem)) { | ||
772 | mem_cgroup_put(next_mem); | ||
773 | next_mem = mem_cgroup_get_first_node(root_mem); | ||
774 | continue; | ||
775 | } | ||
776 | ret = try_to_free_mem_cgroup_pages(next_mem, gfp_mask, noswap, | ||
777 | get_swappiness(next_mem)); | ||
778 | if (mem_cgroup_check_under_limit(root_mem)) | ||
779 | return 0; | ||
780 | mutex_lock(&mem_cgroup_subsys.hierarchy_mutex); | ||
781 | next_mem = mem_cgroup_get_next_node(next_mem, root_mem); | ||
782 | mutex_unlock(&mem_cgroup_subsys.hierarchy_mutex); | ||
783 | } | ||
784 | return ret; | ||
785 | } | ||
786 | |||
787 | bool mem_cgroup_oom_called(struct task_struct *task) | ||
788 | { | ||
789 | bool ret = false; | ||
480 | struct mem_cgroup *mem; | 790 | struct mem_cgroup *mem; |
481 | struct page_cgroup *pc; | 791 | struct mm_struct *mm; |
482 | unsigned long nr_retries = MEM_CGROUP_RECLAIM_RETRIES; | ||
483 | struct mem_cgroup_per_zone *mz; | ||
484 | unsigned long flags; | ||
485 | 792 | ||
486 | pc = lookup_page_cgroup(page); | 793 | rcu_read_lock(); |
487 | /* can happen at boot */ | 794 | mm = task->mm; |
488 | if (unlikely(!pc)) | 795 | if (!mm) |
796 | mm = &init_mm; | ||
797 | mem = mem_cgroup_from_task(rcu_dereference(mm->owner)); | ||
798 | if (mem && time_before(jiffies, mem->last_oom_jiffies + HZ/10)) | ||
799 | ret = true; | ||
800 | rcu_read_unlock(); | ||
801 | return ret; | ||
802 | } | ||
803 | /* | ||
804 | * Unlike exported interface, "oom" parameter is added. if oom==true, | ||
805 | * oom-killer can be invoked. | ||
806 | */ | ||
807 | static int __mem_cgroup_try_charge(struct mm_struct *mm, | ||
808 | gfp_t gfp_mask, struct mem_cgroup **memcg, | ||
809 | bool oom) | ||
810 | { | ||
811 | struct mem_cgroup *mem, *mem_over_limit; | ||
812 | int nr_retries = MEM_CGROUP_RECLAIM_RETRIES; | ||
813 | struct res_counter *fail_res; | ||
814 | |||
815 | if (unlikely(test_thread_flag(TIF_MEMDIE))) { | ||
816 | /* Don't account this! */ | ||
817 | *memcg = NULL; | ||
489 | return 0; | 818 | return 0; |
490 | prefetchw(pc); | 819 | } |
820 | |||
491 | /* | 821 | /* |
492 | * We always charge the cgroup the mm_struct belongs to. | 822 | * We always charge the cgroup the mm_struct belongs to. |
493 | * The mm_struct's mem_cgroup changes on task migration if the | 823 | * The mm_struct's mem_cgroup changes on task migration if the |
494 | * thread group leader migrates. It's possible that mm is not | 824 | * thread group leader migrates. It's possible that mm is not |
495 | * set, if so charge the init_mm (happens for pagecache usage). | 825 | * set, if so charge the init_mm (happens for pagecache usage). |
496 | */ | 826 | */ |
497 | 827 | mem = *memcg; | |
498 | if (likely(!memcg)) { | 828 | if (likely(!mem)) { |
499 | rcu_read_lock(); | 829 | mem = try_get_mem_cgroup_from_mm(mm); |
500 | mem = mem_cgroup_from_task(rcu_dereference(mm->owner)); | 830 | *memcg = mem; |
501 | if (unlikely(!mem)) { | ||
502 | rcu_read_unlock(); | ||
503 | return 0; | ||
504 | } | ||
505 | /* | ||
506 | * For every charge from the cgroup, increment reference count | ||
507 | */ | ||
508 | css_get(&mem->css); | ||
509 | rcu_read_unlock(); | ||
510 | } else { | 831 | } else { |
511 | mem = memcg; | 832 | css_get(&mem->css); |
512 | css_get(&memcg->css); | ||
513 | } | 833 | } |
834 | if (unlikely(!mem)) | ||
835 | return 0; | ||
836 | |||
837 | VM_BUG_ON(mem_cgroup_is_obsolete(mem)); | ||
838 | |||
839 | while (1) { | ||
840 | int ret; | ||
841 | bool noswap = false; | ||
842 | |||
843 | ret = res_counter_charge(&mem->res, PAGE_SIZE, &fail_res); | ||
844 | if (likely(!ret)) { | ||
845 | if (!do_swap_account) | ||
846 | break; | ||
847 | ret = res_counter_charge(&mem->memsw, PAGE_SIZE, | ||
848 | &fail_res); | ||
849 | if (likely(!ret)) | ||
850 | break; | ||
851 | /* mem+swap counter fails */ | ||
852 | res_counter_uncharge(&mem->res, PAGE_SIZE); | ||
853 | noswap = true; | ||
854 | mem_over_limit = mem_cgroup_from_res_counter(fail_res, | ||
855 | memsw); | ||
856 | } else | ||
857 | /* mem counter fails */ | ||
858 | mem_over_limit = mem_cgroup_from_res_counter(fail_res, | ||
859 | res); | ||
514 | 860 | ||
515 | while (unlikely(res_counter_charge(&mem->res, PAGE_SIZE))) { | ||
516 | if (!(gfp_mask & __GFP_WAIT)) | 861 | if (!(gfp_mask & __GFP_WAIT)) |
517 | goto out; | 862 | goto nomem; |
518 | 863 | ||
519 | if (try_to_free_mem_cgroup_pages(mem, gfp_mask)) | 864 | ret = mem_cgroup_hierarchical_reclaim(mem_over_limit, gfp_mask, |
520 | continue; | 865 | noswap); |
521 | 866 | ||
522 | /* | 867 | /* |
523 | * try_to_free_mem_cgroup_pages() might not give us a full | 868 | * try_to_free_mem_cgroup_pages() might not give us a full |
@@ -525,49 +870,214 @@ static int mem_cgroup_charge_common(struct page *page, struct mm_struct *mm, | |||
525 | * moved to swap cache or just unmapped from the cgroup. | 870 | * moved to swap cache or just unmapped from the cgroup. |
526 | * Check the limit again to see if the reclaim reduced the | 871 | * Check the limit again to see if the reclaim reduced the |
527 | * current usage of the cgroup before giving up | 872 | * current usage of the cgroup before giving up |
873 | * | ||
528 | */ | 874 | */ |
529 | if (res_counter_check_under_limit(&mem->res)) | 875 | if (mem_cgroup_check_under_limit(mem_over_limit)) |
530 | continue; | 876 | continue; |
531 | 877 | ||
532 | if (!nr_retries--) { | 878 | if (!nr_retries--) { |
533 | mem_cgroup_out_of_memory(mem, gfp_mask); | 879 | if (oom) { |
534 | goto out; | 880 | mutex_lock(&memcg_tasklist); |
881 | mem_cgroup_out_of_memory(mem_over_limit, gfp_mask); | ||
882 | mutex_unlock(&memcg_tasklist); | ||
883 | mem_over_limit->last_oom_jiffies = jiffies; | ||
884 | } | ||
885 | goto nomem; | ||
535 | } | 886 | } |
536 | } | 887 | } |
888 | return 0; | ||
889 | nomem: | ||
890 | css_put(&mem->css); | ||
891 | return -ENOMEM; | ||
892 | } | ||
537 | 893 | ||
894 | static struct mem_cgroup *try_get_mem_cgroup_from_swapcache(struct page *page) | ||
895 | { | ||
896 | struct mem_cgroup *mem; | ||
897 | swp_entry_t ent; | ||
898 | |||
899 | if (!PageSwapCache(page)) | ||
900 | return NULL; | ||
901 | |||
902 | ent.val = page_private(page); | ||
903 | mem = lookup_swap_cgroup(ent); | ||
904 | if (!mem) | ||
905 | return NULL; | ||
906 | if (!css_tryget(&mem->css)) | ||
907 | return NULL; | ||
908 | return mem; | ||
909 | } | ||
910 | |||
911 | /* | ||
912 | * commit a charge got by __mem_cgroup_try_charge() and makes page_cgroup to be | ||
913 | * USED state. If already USED, uncharge and return. | ||
914 | */ | ||
915 | |||
916 | static void __mem_cgroup_commit_charge(struct mem_cgroup *mem, | ||
917 | struct page_cgroup *pc, | ||
918 | enum charge_type ctype) | ||
919 | { | ||
920 | /* try_charge() can return NULL to *memcg, taking care of it. */ | ||
921 | if (!mem) | ||
922 | return; | ||
538 | 923 | ||
539 | lock_page_cgroup(pc); | 924 | lock_page_cgroup(pc); |
540 | if (unlikely(PageCgroupUsed(pc))) { | 925 | if (unlikely(PageCgroupUsed(pc))) { |
541 | unlock_page_cgroup(pc); | 926 | unlock_page_cgroup(pc); |
542 | res_counter_uncharge(&mem->res, PAGE_SIZE); | 927 | res_counter_uncharge(&mem->res, PAGE_SIZE); |
928 | if (do_swap_account) | ||
929 | res_counter_uncharge(&mem->memsw, PAGE_SIZE); | ||
543 | css_put(&mem->css); | 930 | css_put(&mem->css); |
544 | 931 | return; | |
545 | goto done; | ||
546 | } | 932 | } |
547 | pc->mem_cgroup = mem; | 933 | pc->mem_cgroup = mem; |
548 | /* | 934 | smp_wmb(); |
549 | * If a page is accounted as a page cache, insert to inactive list. | ||
550 | * If anon, insert to active list. | ||
551 | */ | ||
552 | pc->flags = pcg_default_flags[ctype]; | 935 | pc->flags = pcg_default_flags[ctype]; |
553 | 936 | ||
554 | mz = page_cgroup_zoneinfo(pc); | 937 | mem_cgroup_charge_statistics(mem, pc, true); |
555 | 938 | ||
556 | spin_lock_irqsave(&mz->lru_lock, flags); | ||
557 | __mem_cgroup_add_list(mz, pc); | ||
558 | spin_unlock_irqrestore(&mz->lru_lock, flags); | ||
559 | unlock_page_cgroup(pc); | 939 | unlock_page_cgroup(pc); |
940 | } | ||
560 | 941 | ||
561 | done: | 942 | /** |
562 | return 0; | 943 | * mem_cgroup_move_account - move account of the page |
944 | * @pc: page_cgroup of the page. | ||
945 | * @from: mem_cgroup which the page is moved from. | ||
946 | * @to: mem_cgroup which the page is moved to. @from != @to. | ||
947 | * | ||
948 | * The caller must confirm following. | ||
949 | * - page is not on LRU (isolate_page() is useful.) | ||
950 | * | ||
951 | * returns 0 at success, | ||
952 | * returns -EBUSY when lock is busy or "pc" is unstable. | ||
953 | * | ||
954 | * This function does "uncharge" from old cgroup but doesn't do "charge" to | ||
955 | * new cgroup. It should be done by a caller. | ||
956 | */ | ||
957 | |||
958 | static int mem_cgroup_move_account(struct page_cgroup *pc, | ||
959 | struct mem_cgroup *from, struct mem_cgroup *to) | ||
960 | { | ||
961 | struct mem_cgroup_per_zone *from_mz, *to_mz; | ||
962 | int nid, zid; | ||
963 | int ret = -EBUSY; | ||
964 | |||
965 | VM_BUG_ON(from == to); | ||
966 | VM_BUG_ON(PageLRU(pc->page)); | ||
967 | |||
968 | nid = page_cgroup_nid(pc); | ||
969 | zid = page_cgroup_zid(pc); | ||
970 | from_mz = mem_cgroup_zoneinfo(from, nid, zid); | ||
971 | to_mz = mem_cgroup_zoneinfo(to, nid, zid); | ||
972 | |||
973 | if (!trylock_page_cgroup(pc)) | ||
974 | return ret; | ||
975 | |||
976 | if (!PageCgroupUsed(pc)) | ||
977 | goto out; | ||
978 | |||
979 | if (pc->mem_cgroup != from) | ||
980 | goto out; | ||
981 | |||
982 | css_put(&from->css); | ||
983 | res_counter_uncharge(&from->res, PAGE_SIZE); | ||
984 | mem_cgroup_charge_statistics(from, pc, false); | ||
985 | if (do_swap_account) | ||
986 | res_counter_uncharge(&from->memsw, PAGE_SIZE); | ||
987 | pc->mem_cgroup = to; | ||
988 | mem_cgroup_charge_statistics(to, pc, true); | ||
989 | css_get(&to->css); | ||
990 | ret = 0; | ||
563 | out: | 991 | out: |
564 | css_put(&mem->css); | 992 | unlock_page_cgroup(pc); |
565 | return -ENOMEM; | 993 | return ret; |
994 | } | ||
995 | |||
996 | /* | ||
997 | * move charges to its parent. | ||
998 | */ | ||
999 | |||
1000 | static int mem_cgroup_move_parent(struct page_cgroup *pc, | ||
1001 | struct mem_cgroup *child, | ||
1002 | gfp_t gfp_mask) | ||
1003 | { | ||
1004 | struct page *page = pc->page; | ||
1005 | struct cgroup *cg = child->css.cgroup; | ||
1006 | struct cgroup *pcg = cg->parent; | ||
1007 | struct mem_cgroup *parent; | ||
1008 | int ret; | ||
1009 | |||
1010 | /* Is ROOT ? */ | ||
1011 | if (!pcg) | ||
1012 | return -EINVAL; | ||
1013 | |||
1014 | |||
1015 | parent = mem_cgroup_from_cont(pcg); | ||
1016 | |||
1017 | |||
1018 | ret = __mem_cgroup_try_charge(NULL, gfp_mask, &parent, false); | ||
1019 | if (ret || !parent) | ||
1020 | return ret; | ||
1021 | |||
1022 | if (!get_page_unless_zero(page)) | ||
1023 | return -EBUSY; | ||
1024 | |||
1025 | ret = isolate_lru_page(page); | ||
1026 | |||
1027 | if (ret) | ||
1028 | goto cancel; | ||
1029 | |||
1030 | ret = mem_cgroup_move_account(pc, child, parent); | ||
1031 | |||
1032 | /* drop extra refcnt by try_charge() (move_account increment one) */ | ||
1033 | css_put(&parent->css); | ||
1034 | putback_lru_page(page); | ||
1035 | if (!ret) { | ||
1036 | put_page(page); | ||
1037 | return 0; | ||
1038 | } | ||
1039 | /* uncharge if move fails */ | ||
1040 | cancel: | ||
1041 | res_counter_uncharge(&parent->res, PAGE_SIZE); | ||
1042 | if (do_swap_account) | ||
1043 | res_counter_uncharge(&parent->memsw, PAGE_SIZE); | ||
1044 | put_page(page); | ||
1045 | return ret; | ||
1046 | } | ||
1047 | |||
1048 | /* | ||
1049 | * Charge the memory controller for page usage. | ||
1050 | * Return | ||
1051 | * 0 if the charge was successful | ||
1052 | * < 0 if the cgroup is over its limit | ||
1053 | */ | ||
1054 | static int mem_cgroup_charge_common(struct page *page, struct mm_struct *mm, | ||
1055 | gfp_t gfp_mask, enum charge_type ctype, | ||
1056 | struct mem_cgroup *memcg) | ||
1057 | { | ||
1058 | struct mem_cgroup *mem; | ||
1059 | struct page_cgroup *pc; | ||
1060 | int ret; | ||
1061 | |||
1062 | pc = lookup_page_cgroup(page); | ||
1063 | /* can happen at boot */ | ||
1064 | if (unlikely(!pc)) | ||
1065 | return 0; | ||
1066 | prefetchw(pc); | ||
1067 | |||
1068 | mem = memcg; | ||
1069 | ret = __mem_cgroup_try_charge(mm, gfp_mask, &mem, true); | ||
1070 | if (ret || !mem) | ||
1071 | return ret; | ||
1072 | |||
1073 | __mem_cgroup_commit_charge(mem, pc, ctype); | ||
1074 | return 0; | ||
566 | } | 1075 | } |
567 | 1076 | ||
568 | int mem_cgroup_charge(struct page *page, struct mm_struct *mm, gfp_t gfp_mask) | 1077 | int mem_cgroup_newpage_charge(struct page *page, |
1078 | struct mm_struct *mm, gfp_t gfp_mask) | ||
569 | { | 1079 | { |
570 | if (mem_cgroup_subsys.disabled) | 1080 | if (mem_cgroup_disabled()) |
571 | return 0; | 1081 | return 0; |
572 | if (PageCompound(page)) | 1082 | if (PageCompound(page)) |
573 | return 0; | 1083 | return 0; |
@@ -589,7 +1099,10 @@ int mem_cgroup_charge(struct page *page, struct mm_struct *mm, gfp_t gfp_mask) | |||
589 | int mem_cgroup_cache_charge(struct page *page, struct mm_struct *mm, | 1099 | int mem_cgroup_cache_charge(struct page *page, struct mm_struct *mm, |
590 | gfp_t gfp_mask) | 1100 | gfp_t gfp_mask) |
591 | { | 1101 | { |
592 | if (mem_cgroup_subsys.disabled) | 1102 | struct mem_cgroup *mem = NULL; |
1103 | int ret; | ||
1104 | |||
1105 | if (mem_cgroup_disabled()) | ||
593 | return 0; | 1106 | return 0; |
594 | if (PageCompound(page)) | 1107 | if (PageCompound(page)) |
595 | return 0; | 1108 | return 0; |
@@ -601,6 +1114,8 @@ int mem_cgroup_cache_charge(struct page *page, struct mm_struct *mm, | |||
601 | * For GFP_NOWAIT case, the page may be pre-charged before calling | 1114 | * For GFP_NOWAIT case, the page may be pre-charged before calling |
602 | * add_to_page_cache(). (See shmem.c) check it here and avoid to call | 1115 | * add_to_page_cache(). (See shmem.c) check it here and avoid to call |
603 | * charge twice. (It works but has to pay a bit larger cost.) | 1116 | * charge twice. (It works but has to pay a bit larger cost.) |
1117 | * And when the page is SwapCache, it should take swap information | ||
1118 | * into account. This is under lock_page() now. | ||
604 | */ | 1119 | */ |
605 | if (!(gfp_mask & __GFP_WAIT)) { | 1120 | if (!(gfp_mask & __GFP_WAIT)) { |
606 | struct page_cgroup *pc; | 1121 | struct page_cgroup *pc; |
@@ -617,58 +1132,198 @@ int mem_cgroup_cache_charge(struct page *page, struct mm_struct *mm, | |||
617 | unlock_page_cgroup(pc); | 1132 | unlock_page_cgroup(pc); |
618 | } | 1133 | } |
619 | 1134 | ||
620 | if (unlikely(!mm)) | 1135 | if (do_swap_account && PageSwapCache(page)) { |
1136 | mem = try_get_mem_cgroup_from_swapcache(page); | ||
1137 | if (mem) | ||
1138 | mm = NULL; | ||
1139 | else | ||
1140 | mem = NULL; | ||
1141 | /* SwapCache may be still linked to LRU now. */ | ||
1142 | mem_cgroup_lru_del_before_commit_swapcache(page); | ||
1143 | } | ||
1144 | |||
1145 | if (unlikely(!mm && !mem)) | ||
621 | mm = &init_mm; | 1146 | mm = &init_mm; |
622 | 1147 | ||
623 | if (page_is_file_cache(page)) | 1148 | if (page_is_file_cache(page)) |
624 | return mem_cgroup_charge_common(page, mm, gfp_mask, | 1149 | return mem_cgroup_charge_common(page, mm, gfp_mask, |
625 | MEM_CGROUP_CHARGE_TYPE_CACHE, NULL); | 1150 | MEM_CGROUP_CHARGE_TYPE_CACHE, NULL); |
626 | else | 1151 | |
627 | return mem_cgroup_charge_common(page, mm, gfp_mask, | 1152 | ret = mem_cgroup_charge_common(page, mm, gfp_mask, |
628 | MEM_CGROUP_CHARGE_TYPE_SHMEM, NULL); | 1153 | MEM_CGROUP_CHARGE_TYPE_SHMEM, mem); |
1154 | if (mem) | ||
1155 | css_put(&mem->css); | ||
1156 | if (PageSwapCache(page)) | ||
1157 | mem_cgroup_lru_add_after_commit_swapcache(page); | ||
1158 | |||
1159 | if (do_swap_account && !ret && PageSwapCache(page)) { | ||
1160 | swp_entry_t ent = {.val = page_private(page)}; | ||
1161 | /* avoid double counting */ | ||
1162 | mem = swap_cgroup_record(ent, NULL); | ||
1163 | if (mem) { | ||
1164 | res_counter_uncharge(&mem->memsw, PAGE_SIZE); | ||
1165 | mem_cgroup_put(mem); | ||
1166 | } | ||
1167 | } | ||
1168 | return ret; | ||
1169 | } | ||
1170 | |||
1171 | /* | ||
1172 | * While swap-in, try_charge -> commit or cancel, the page is locked. | ||
1173 | * And when try_charge() successfully returns, one refcnt to memcg without | ||
1174 | * struct page_cgroup is aquired. This refcnt will be cumsumed by | ||
1175 | * "commit()" or removed by "cancel()" | ||
1176 | */ | ||
1177 | int mem_cgroup_try_charge_swapin(struct mm_struct *mm, | ||
1178 | struct page *page, | ||
1179 | gfp_t mask, struct mem_cgroup **ptr) | ||
1180 | { | ||
1181 | struct mem_cgroup *mem; | ||
1182 | int ret; | ||
1183 | |||
1184 | if (mem_cgroup_disabled()) | ||
1185 | return 0; | ||
1186 | |||
1187 | if (!do_swap_account) | ||
1188 | goto charge_cur_mm; | ||
1189 | /* | ||
1190 | * A racing thread's fault, or swapoff, may have already updated | ||
1191 | * the pte, and even removed page from swap cache: return success | ||
1192 | * to go on to do_swap_page()'s pte_same() test, which should fail. | ||
1193 | */ | ||
1194 | if (!PageSwapCache(page)) | ||
1195 | return 0; | ||
1196 | mem = try_get_mem_cgroup_from_swapcache(page); | ||
1197 | if (!mem) | ||
1198 | goto charge_cur_mm; | ||
1199 | *ptr = mem; | ||
1200 | ret = __mem_cgroup_try_charge(NULL, mask, ptr, true); | ||
1201 | /* drop extra refcnt from tryget */ | ||
1202 | css_put(&mem->css); | ||
1203 | return ret; | ||
1204 | charge_cur_mm: | ||
1205 | if (unlikely(!mm)) | ||
1206 | mm = &init_mm; | ||
1207 | return __mem_cgroup_try_charge(mm, mask, ptr, true); | ||
1208 | } | ||
1209 | |||
1210 | void mem_cgroup_commit_charge_swapin(struct page *page, struct mem_cgroup *ptr) | ||
1211 | { | ||
1212 | struct page_cgroup *pc; | ||
1213 | |||
1214 | if (mem_cgroup_disabled()) | ||
1215 | return; | ||
1216 | if (!ptr) | ||
1217 | return; | ||
1218 | pc = lookup_page_cgroup(page); | ||
1219 | mem_cgroup_lru_del_before_commit_swapcache(page); | ||
1220 | __mem_cgroup_commit_charge(ptr, pc, MEM_CGROUP_CHARGE_TYPE_MAPPED); | ||
1221 | mem_cgroup_lru_add_after_commit_swapcache(page); | ||
1222 | /* | ||
1223 | * Now swap is on-memory. This means this page may be | ||
1224 | * counted both as mem and swap....double count. | ||
1225 | * Fix it by uncharging from memsw. Basically, this SwapCache is stable | ||
1226 | * under lock_page(). But in do_swap_page()::memory.c, reuse_swap_page() | ||
1227 | * may call delete_from_swap_cache() before reach here. | ||
1228 | */ | ||
1229 | if (do_swap_account && PageSwapCache(page)) { | ||
1230 | swp_entry_t ent = {.val = page_private(page)}; | ||
1231 | struct mem_cgroup *memcg; | ||
1232 | memcg = swap_cgroup_record(ent, NULL); | ||
1233 | if (memcg) { | ||
1234 | res_counter_uncharge(&memcg->memsw, PAGE_SIZE); | ||
1235 | mem_cgroup_put(memcg); | ||
1236 | } | ||
1237 | |||
1238 | } | ||
1239 | /* add this page(page_cgroup) to the LRU we want. */ | ||
1240 | |||
629 | } | 1241 | } |
630 | 1242 | ||
1243 | void mem_cgroup_cancel_charge_swapin(struct mem_cgroup *mem) | ||
1244 | { | ||
1245 | if (mem_cgroup_disabled()) | ||
1246 | return; | ||
1247 | if (!mem) | ||
1248 | return; | ||
1249 | res_counter_uncharge(&mem->res, PAGE_SIZE); | ||
1250 | if (do_swap_account) | ||
1251 | res_counter_uncharge(&mem->memsw, PAGE_SIZE); | ||
1252 | css_put(&mem->css); | ||
1253 | } | ||
1254 | |||
1255 | |||
631 | /* | 1256 | /* |
632 | * uncharge if !page_mapped(page) | 1257 | * uncharge if !page_mapped(page) |
633 | */ | 1258 | */ |
634 | static void | 1259 | static struct mem_cgroup * |
635 | __mem_cgroup_uncharge_common(struct page *page, enum charge_type ctype) | 1260 | __mem_cgroup_uncharge_common(struct page *page, enum charge_type ctype) |
636 | { | 1261 | { |
637 | struct page_cgroup *pc; | 1262 | struct page_cgroup *pc; |
638 | struct mem_cgroup *mem; | 1263 | struct mem_cgroup *mem = NULL; |
639 | struct mem_cgroup_per_zone *mz; | 1264 | struct mem_cgroup_per_zone *mz; |
640 | unsigned long flags; | ||
641 | 1265 | ||
642 | if (mem_cgroup_subsys.disabled) | 1266 | if (mem_cgroup_disabled()) |
643 | return; | 1267 | return NULL; |
1268 | |||
1269 | if (PageSwapCache(page)) | ||
1270 | return NULL; | ||
644 | 1271 | ||
645 | /* | 1272 | /* |
646 | * Check if our page_cgroup is valid | 1273 | * Check if our page_cgroup is valid |
647 | */ | 1274 | */ |
648 | pc = lookup_page_cgroup(page); | 1275 | pc = lookup_page_cgroup(page); |
649 | if (unlikely(!pc || !PageCgroupUsed(pc))) | 1276 | if (unlikely(!pc || !PageCgroupUsed(pc))) |
650 | return; | 1277 | return NULL; |
651 | 1278 | ||
652 | lock_page_cgroup(pc); | 1279 | lock_page_cgroup(pc); |
653 | if ((ctype == MEM_CGROUP_CHARGE_TYPE_MAPPED && page_mapped(page)) | 1280 | |
654 | || !PageCgroupUsed(pc)) { | 1281 | mem = pc->mem_cgroup; |
655 | /* This happens at race in zap_pte_range() and do_swap_page()*/ | 1282 | |
656 | unlock_page_cgroup(pc); | 1283 | if (!PageCgroupUsed(pc)) |
657 | return; | 1284 | goto unlock_out; |
1285 | |||
1286 | switch (ctype) { | ||
1287 | case MEM_CGROUP_CHARGE_TYPE_MAPPED: | ||
1288 | if (page_mapped(page)) | ||
1289 | goto unlock_out; | ||
1290 | break; | ||
1291 | case MEM_CGROUP_CHARGE_TYPE_SWAPOUT: | ||
1292 | if (!PageAnon(page)) { /* Shared memory */ | ||
1293 | if (page->mapping && !page_is_file_cache(page)) | ||
1294 | goto unlock_out; | ||
1295 | } else if (page_mapped(page)) /* Anon */ | ||
1296 | goto unlock_out; | ||
1297 | break; | ||
1298 | default: | ||
1299 | break; | ||
658 | } | 1300 | } |
1301 | |||
1302 | res_counter_uncharge(&mem->res, PAGE_SIZE); | ||
1303 | if (do_swap_account && (ctype != MEM_CGROUP_CHARGE_TYPE_SWAPOUT)) | ||
1304 | res_counter_uncharge(&mem->memsw, PAGE_SIZE); | ||
1305 | |||
1306 | mem_cgroup_charge_statistics(mem, pc, false); | ||
659 | ClearPageCgroupUsed(pc); | 1307 | ClearPageCgroupUsed(pc); |
660 | mem = pc->mem_cgroup; | 1308 | /* |
1309 | * pc->mem_cgroup is not cleared here. It will be accessed when it's | ||
1310 | * freed from LRU. This is safe because uncharged page is expected not | ||
1311 | * to be reused (freed soon). Exception is SwapCache, it's handled by | ||
1312 | * special functions. | ||
1313 | */ | ||
661 | 1314 | ||
662 | mz = page_cgroup_zoneinfo(pc); | 1315 | mz = page_cgroup_zoneinfo(pc); |
663 | spin_lock_irqsave(&mz->lru_lock, flags); | ||
664 | __mem_cgroup_remove_list(mz, pc); | ||
665 | spin_unlock_irqrestore(&mz->lru_lock, flags); | ||
666 | unlock_page_cgroup(pc); | 1316 | unlock_page_cgroup(pc); |
667 | 1317 | ||
668 | res_counter_uncharge(&mem->res, PAGE_SIZE); | 1318 | /* at swapout, this memcg will be accessed to record to swap */ |
669 | css_put(&mem->css); | 1319 | if (ctype != MEM_CGROUP_CHARGE_TYPE_SWAPOUT) |
1320 | css_put(&mem->css); | ||
670 | 1321 | ||
671 | return; | 1322 | return mem; |
1323 | |||
1324 | unlock_out: | ||
1325 | unlock_page_cgroup(pc); | ||
1326 | return NULL; | ||
672 | } | 1327 | } |
673 | 1328 | ||
674 | void mem_cgroup_uncharge_page(struct page *page) | 1329 | void mem_cgroup_uncharge_page(struct page *page) |
@@ -689,16 +1344,55 @@ void mem_cgroup_uncharge_cache_page(struct page *page) | |||
689 | } | 1344 | } |
690 | 1345 | ||
691 | /* | 1346 | /* |
692 | * Before starting migration, account against new page. | 1347 | * called from __delete_from_swap_cache() and drop "page" account. |
1348 | * memcg information is recorded to swap_cgroup of "ent" | ||
1349 | */ | ||
1350 | void mem_cgroup_uncharge_swapcache(struct page *page, swp_entry_t ent) | ||
1351 | { | ||
1352 | struct mem_cgroup *memcg; | ||
1353 | |||
1354 | memcg = __mem_cgroup_uncharge_common(page, | ||
1355 | MEM_CGROUP_CHARGE_TYPE_SWAPOUT); | ||
1356 | /* record memcg information */ | ||
1357 | if (do_swap_account && memcg) { | ||
1358 | swap_cgroup_record(ent, memcg); | ||
1359 | mem_cgroup_get(memcg); | ||
1360 | } | ||
1361 | if (memcg) | ||
1362 | css_put(&memcg->css); | ||
1363 | } | ||
1364 | |||
1365 | #ifdef CONFIG_CGROUP_MEM_RES_CTLR_SWAP | ||
1366 | /* | ||
1367 | * called from swap_entry_free(). remove record in swap_cgroup and | ||
1368 | * uncharge "memsw" account. | ||
693 | */ | 1369 | */ |
694 | int mem_cgroup_prepare_migration(struct page *page, struct page *newpage) | 1370 | void mem_cgroup_uncharge_swap(swp_entry_t ent) |
1371 | { | ||
1372 | struct mem_cgroup *memcg; | ||
1373 | |||
1374 | if (!do_swap_account) | ||
1375 | return; | ||
1376 | |||
1377 | memcg = swap_cgroup_record(ent, NULL); | ||
1378 | if (memcg) { | ||
1379 | res_counter_uncharge(&memcg->memsw, PAGE_SIZE); | ||
1380 | mem_cgroup_put(memcg); | ||
1381 | } | ||
1382 | } | ||
1383 | #endif | ||
1384 | |||
1385 | /* | ||
1386 | * Before starting migration, account PAGE_SIZE to mem_cgroup that the old | ||
1387 | * page belongs to. | ||
1388 | */ | ||
1389 | int mem_cgroup_prepare_migration(struct page *page, struct mem_cgroup **ptr) | ||
695 | { | 1390 | { |
696 | struct page_cgroup *pc; | 1391 | struct page_cgroup *pc; |
697 | struct mem_cgroup *mem = NULL; | 1392 | struct mem_cgroup *mem = NULL; |
698 | enum charge_type ctype = MEM_CGROUP_CHARGE_TYPE_MAPPED; | ||
699 | int ret = 0; | 1393 | int ret = 0; |
700 | 1394 | ||
701 | if (mem_cgroup_subsys.disabled) | 1395 | if (mem_cgroup_disabled()) |
702 | return 0; | 1396 | return 0; |
703 | 1397 | ||
704 | pc = lookup_page_cgroup(page); | 1398 | pc = lookup_page_cgroup(page); |
@@ -706,41 +1400,67 @@ int mem_cgroup_prepare_migration(struct page *page, struct page *newpage) | |||
706 | if (PageCgroupUsed(pc)) { | 1400 | if (PageCgroupUsed(pc)) { |
707 | mem = pc->mem_cgroup; | 1401 | mem = pc->mem_cgroup; |
708 | css_get(&mem->css); | 1402 | css_get(&mem->css); |
709 | if (PageCgroupCache(pc)) { | ||
710 | if (page_is_file_cache(page)) | ||
711 | ctype = MEM_CGROUP_CHARGE_TYPE_CACHE; | ||
712 | else | ||
713 | ctype = MEM_CGROUP_CHARGE_TYPE_SHMEM; | ||
714 | } | ||
715 | } | 1403 | } |
716 | unlock_page_cgroup(pc); | 1404 | unlock_page_cgroup(pc); |
1405 | |||
717 | if (mem) { | 1406 | if (mem) { |
718 | ret = mem_cgroup_charge_common(newpage, NULL, GFP_KERNEL, | 1407 | ret = __mem_cgroup_try_charge(NULL, GFP_KERNEL, &mem, false); |
719 | ctype, mem); | ||
720 | css_put(&mem->css); | 1408 | css_put(&mem->css); |
721 | } | 1409 | } |
1410 | *ptr = mem; | ||
722 | return ret; | 1411 | return ret; |
723 | } | 1412 | } |
724 | 1413 | ||
725 | /* remove redundant charge if migration failed*/ | 1414 | /* remove redundant charge if migration failed*/ |
726 | void mem_cgroup_end_migration(struct page *newpage) | 1415 | void mem_cgroup_end_migration(struct mem_cgroup *mem, |
1416 | struct page *oldpage, struct page *newpage) | ||
727 | { | 1417 | { |
1418 | struct page *target, *unused; | ||
1419 | struct page_cgroup *pc; | ||
1420 | enum charge_type ctype; | ||
1421 | |||
1422 | if (!mem) | ||
1423 | return; | ||
1424 | |||
1425 | /* at migration success, oldpage->mapping is NULL. */ | ||
1426 | if (oldpage->mapping) { | ||
1427 | target = oldpage; | ||
1428 | unused = NULL; | ||
1429 | } else { | ||
1430 | target = newpage; | ||
1431 | unused = oldpage; | ||
1432 | } | ||
1433 | |||
1434 | if (PageAnon(target)) | ||
1435 | ctype = MEM_CGROUP_CHARGE_TYPE_MAPPED; | ||
1436 | else if (page_is_file_cache(target)) | ||
1437 | ctype = MEM_CGROUP_CHARGE_TYPE_CACHE; | ||
1438 | else | ||
1439 | ctype = MEM_CGROUP_CHARGE_TYPE_SHMEM; | ||
1440 | |||
1441 | /* unused page is not on radix-tree now. */ | ||
1442 | if (unused) | ||
1443 | __mem_cgroup_uncharge_common(unused, ctype); | ||
1444 | |||
1445 | pc = lookup_page_cgroup(target); | ||
728 | /* | 1446 | /* |
729 | * At success, page->mapping is not NULL. | 1447 | * __mem_cgroup_commit_charge() check PCG_USED bit of page_cgroup. |
730 | * special rollback care is necessary when | 1448 | * So, double-counting is effectively avoided. |
731 | * 1. at migration failure. (newpage->mapping is cleared in this case) | ||
732 | * 2. the newpage was moved but not remapped again because the task | ||
733 | * exits and the newpage is obsolete. In this case, the new page | ||
734 | * may be a swapcache. So, we just call mem_cgroup_uncharge_page() | ||
735 | * always for avoiding mess. The page_cgroup will be removed if | ||
736 | * unnecessary. File cache pages is still on radix-tree. Don't | ||
737 | * care it. | ||
738 | */ | 1449 | */ |
739 | if (!newpage->mapping) | 1450 | __mem_cgroup_commit_charge(mem, pc, ctype); |
740 | __mem_cgroup_uncharge_common(newpage, | 1451 | |
741 | MEM_CGROUP_CHARGE_TYPE_FORCE); | 1452 | /* |
742 | else if (PageAnon(newpage)) | 1453 | * Both of oldpage and newpage are still under lock_page(). |
743 | mem_cgroup_uncharge_page(newpage); | 1454 | * Then, we don't have to care about race in radix-tree. |
1455 | * But we have to be careful that this page is unmapped or not. | ||
1456 | * | ||
1457 | * There is a case for !page_mapped(). At the start of | ||
1458 | * migration, oldpage was mapped. But now, it's zapped. | ||
1459 | * But we know *target* page is not freed/reused under us. | ||
1460 | * mem_cgroup_uncharge_page() does all necessary checks. | ||
1461 | */ | ||
1462 | if (ctype == MEM_CGROUP_CHARGE_TYPE_MAPPED) | ||
1463 | mem_cgroup_uncharge_page(target); | ||
744 | } | 1464 | } |
745 | 1465 | ||
746 | /* | 1466 | /* |
@@ -748,29 +1468,26 @@ void mem_cgroup_end_migration(struct page *newpage) | |||
748 | * This is typically used for page reclaiming for shmem for reducing side | 1468 | * This is typically used for page reclaiming for shmem for reducing side |
749 | * effect of page allocation from shmem, which is used by some mem_cgroup. | 1469 | * effect of page allocation from shmem, which is used by some mem_cgroup. |
750 | */ | 1470 | */ |
751 | int mem_cgroup_shrink_usage(struct mm_struct *mm, gfp_t gfp_mask) | 1471 | int mem_cgroup_shrink_usage(struct page *page, |
1472 | struct mm_struct *mm, | ||
1473 | gfp_t gfp_mask) | ||
752 | { | 1474 | { |
753 | struct mem_cgroup *mem; | 1475 | struct mem_cgroup *mem = NULL; |
754 | int progress = 0; | 1476 | int progress = 0; |
755 | int retry = MEM_CGROUP_RECLAIM_RETRIES; | 1477 | int retry = MEM_CGROUP_RECLAIM_RETRIES; |
756 | 1478 | ||
757 | if (mem_cgroup_subsys.disabled) | 1479 | if (mem_cgroup_disabled()) |
758 | return 0; | 1480 | return 0; |
759 | if (!mm) | 1481 | if (page) |
1482 | mem = try_get_mem_cgroup_from_swapcache(page); | ||
1483 | if (!mem && mm) | ||
1484 | mem = try_get_mem_cgroup_from_mm(mm); | ||
1485 | if (unlikely(!mem)) | ||
760 | return 0; | 1486 | return 0; |
761 | 1487 | ||
762 | rcu_read_lock(); | ||
763 | mem = mem_cgroup_from_task(rcu_dereference(mm->owner)); | ||
764 | if (unlikely(!mem)) { | ||
765 | rcu_read_unlock(); | ||
766 | return 0; | ||
767 | } | ||
768 | css_get(&mem->css); | ||
769 | rcu_read_unlock(); | ||
770 | |||
771 | do { | 1488 | do { |
772 | progress = try_to_free_mem_cgroup_pages(mem, gfp_mask); | 1489 | progress = mem_cgroup_hierarchical_reclaim(mem, gfp_mask, true); |
773 | progress += res_counter_check_under_limit(&mem->res); | 1490 | progress += mem_cgroup_check_under_limit(mem); |
774 | } while (!progress && --retry); | 1491 | } while (!progress && --retry); |
775 | 1492 | ||
776 | css_put(&mem->css); | 1493 | css_put(&mem->css); |
@@ -779,117 +1496,295 @@ int mem_cgroup_shrink_usage(struct mm_struct *mm, gfp_t gfp_mask) | |||
779 | return 0; | 1496 | return 0; |
780 | } | 1497 | } |
781 | 1498 | ||
1499 | static DEFINE_MUTEX(set_limit_mutex); | ||
1500 | |||
782 | static int mem_cgroup_resize_limit(struct mem_cgroup *memcg, | 1501 | static int mem_cgroup_resize_limit(struct mem_cgroup *memcg, |
783 | unsigned long long val) | 1502 | unsigned long long val) |
784 | { | 1503 | { |
785 | 1504 | ||
786 | int retry_count = MEM_CGROUP_RECLAIM_RETRIES; | 1505 | int retry_count = MEM_CGROUP_RECLAIM_RETRIES; |
787 | int progress; | 1506 | int progress; |
1507 | u64 memswlimit; | ||
788 | int ret = 0; | 1508 | int ret = 0; |
789 | 1509 | ||
790 | while (res_counter_set_limit(&memcg->res, val)) { | 1510 | while (retry_count) { |
791 | if (signal_pending(current)) { | 1511 | if (signal_pending(current)) { |
792 | ret = -EINTR; | 1512 | ret = -EINTR; |
793 | break; | 1513 | break; |
794 | } | 1514 | } |
795 | if (!retry_count) { | 1515 | /* |
796 | ret = -EBUSY; | 1516 | * Rather than hide all in some function, I do this in |
1517 | * open coded manner. You see what this really does. | ||
1518 | * We have to guarantee mem->res.limit < mem->memsw.limit. | ||
1519 | */ | ||
1520 | mutex_lock(&set_limit_mutex); | ||
1521 | memswlimit = res_counter_read_u64(&memcg->memsw, RES_LIMIT); | ||
1522 | if (memswlimit < val) { | ||
1523 | ret = -EINVAL; | ||
1524 | mutex_unlock(&set_limit_mutex); | ||
797 | break; | 1525 | break; |
798 | } | 1526 | } |
799 | progress = try_to_free_mem_cgroup_pages(memcg, GFP_KERNEL); | 1527 | ret = res_counter_set_limit(&memcg->res, val); |
800 | if (!progress) | 1528 | mutex_unlock(&set_limit_mutex); |
801 | retry_count--; | 1529 | |
1530 | if (!ret) | ||
1531 | break; | ||
1532 | |||
1533 | progress = mem_cgroup_hierarchical_reclaim(memcg, GFP_KERNEL, | ||
1534 | false); | ||
1535 | if (!progress) retry_count--; | ||
802 | } | 1536 | } |
1537 | |||
803 | return ret; | 1538 | return ret; |
804 | } | 1539 | } |
805 | 1540 | ||
1541 | int mem_cgroup_resize_memsw_limit(struct mem_cgroup *memcg, | ||
1542 | unsigned long long val) | ||
1543 | { | ||
1544 | int retry_count = MEM_CGROUP_RECLAIM_RETRIES; | ||
1545 | u64 memlimit, oldusage, curusage; | ||
1546 | int ret; | ||
1547 | |||
1548 | if (!do_swap_account) | ||
1549 | return -EINVAL; | ||
1550 | |||
1551 | while (retry_count) { | ||
1552 | if (signal_pending(current)) { | ||
1553 | ret = -EINTR; | ||
1554 | break; | ||
1555 | } | ||
1556 | /* | ||
1557 | * Rather than hide all in some function, I do this in | ||
1558 | * open coded manner. You see what this really does. | ||
1559 | * We have to guarantee mem->res.limit < mem->memsw.limit. | ||
1560 | */ | ||
1561 | mutex_lock(&set_limit_mutex); | ||
1562 | memlimit = res_counter_read_u64(&memcg->res, RES_LIMIT); | ||
1563 | if (memlimit > val) { | ||
1564 | ret = -EINVAL; | ||
1565 | mutex_unlock(&set_limit_mutex); | ||
1566 | break; | ||
1567 | } | ||
1568 | ret = res_counter_set_limit(&memcg->memsw, val); | ||
1569 | mutex_unlock(&set_limit_mutex); | ||
1570 | |||
1571 | if (!ret) | ||
1572 | break; | ||
1573 | |||
1574 | oldusage = res_counter_read_u64(&memcg->memsw, RES_USAGE); | ||
1575 | mem_cgroup_hierarchical_reclaim(memcg, GFP_KERNEL, true); | ||
1576 | curusage = res_counter_read_u64(&memcg->memsw, RES_USAGE); | ||
1577 | if (curusage >= oldusage) | ||
1578 | retry_count--; | ||
1579 | } | ||
1580 | return ret; | ||
1581 | } | ||
806 | 1582 | ||
807 | /* | 1583 | /* |
808 | * This routine traverse page_cgroup in given list and drop them all. | 1584 | * This routine traverse page_cgroup in given list and drop them all. |
809 | * *And* this routine doesn't reclaim page itself, just removes page_cgroup. | 1585 | * *And* this routine doesn't reclaim page itself, just removes page_cgroup. |
810 | */ | 1586 | */ |
811 | #define FORCE_UNCHARGE_BATCH (128) | 1587 | static int mem_cgroup_force_empty_list(struct mem_cgroup *mem, |
812 | static void mem_cgroup_force_empty_list(struct mem_cgroup *mem, | 1588 | int node, int zid, enum lru_list lru) |
813 | struct mem_cgroup_per_zone *mz, | ||
814 | enum lru_list lru) | ||
815 | { | 1589 | { |
816 | struct page_cgroup *pc; | 1590 | struct zone *zone; |
817 | struct page *page; | 1591 | struct mem_cgroup_per_zone *mz; |
818 | int count = FORCE_UNCHARGE_BATCH; | 1592 | struct page_cgroup *pc, *busy; |
819 | unsigned long flags; | 1593 | unsigned long flags, loop; |
820 | struct list_head *list; | 1594 | struct list_head *list; |
1595 | int ret = 0; | ||
821 | 1596 | ||
1597 | zone = &NODE_DATA(node)->node_zones[zid]; | ||
1598 | mz = mem_cgroup_zoneinfo(mem, node, zid); | ||
822 | list = &mz->lists[lru]; | 1599 | list = &mz->lists[lru]; |
823 | 1600 | ||
824 | spin_lock_irqsave(&mz->lru_lock, flags); | 1601 | loop = MEM_CGROUP_ZSTAT(mz, lru); |
825 | while (!list_empty(list)) { | 1602 | /* give some margin against EBUSY etc...*/ |
826 | pc = list_entry(list->prev, struct page_cgroup, lru); | 1603 | loop += 256; |
827 | page = pc->page; | 1604 | busy = NULL; |
828 | if (!PageCgroupUsed(pc)) | 1605 | while (loop--) { |
829 | break; | 1606 | ret = 0; |
830 | get_page(page); | 1607 | spin_lock_irqsave(&zone->lru_lock, flags); |
831 | spin_unlock_irqrestore(&mz->lru_lock, flags); | 1608 | if (list_empty(list)) { |
832 | /* | 1609 | spin_unlock_irqrestore(&zone->lru_lock, flags); |
833 | * Check if this page is on LRU. !LRU page can be found | ||
834 | * if it's under page migration. | ||
835 | */ | ||
836 | if (PageLRU(page)) { | ||
837 | __mem_cgroup_uncharge_common(page, | ||
838 | MEM_CGROUP_CHARGE_TYPE_FORCE); | ||
839 | put_page(page); | ||
840 | if (--count <= 0) { | ||
841 | count = FORCE_UNCHARGE_BATCH; | ||
842 | cond_resched(); | ||
843 | } | ||
844 | } else { | ||
845 | spin_lock_irqsave(&mz->lru_lock, flags); | ||
846 | break; | 1610 | break; |
847 | } | 1611 | } |
848 | spin_lock_irqsave(&mz->lru_lock, flags); | 1612 | pc = list_entry(list->prev, struct page_cgroup, lru); |
1613 | if (busy == pc) { | ||
1614 | list_move(&pc->lru, list); | ||
1615 | busy = 0; | ||
1616 | spin_unlock_irqrestore(&zone->lru_lock, flags); | ||
1617 | continue; | ||
1618 | } | ||
1619 | spin_unlock_irqrestore(&zone->lru_lock, flags); | ||
1620 | |||
1621 | ret = mem_cgroup_move_parent(pc, mem, GFP_KERNEL); | ||
1622 | if (ret == -ENOMEM) | ||
1623 | break; | ||
1624 | |||
1625 | if (ret == -EBUSY || ret == -EINVAL) { | ||
1626 | /* found lock contention or "pc" is obsolete. */ | ||
1627 | busy = pc; | ||
1628 | cond_resched(); | ||
1629 | } else | ||
1630 | busy = NULL; | ||
849 | } | 1631 | } |
850 | spin_unlock_irqrestore(&mz->lru_lock, flags); | 1632 | |
1633 | if (!ret && !list_empty(list)) | ||
1634 | return -EBUSY; | ||
1635 | return ret; | ||
851 | } | 1636 | } |
852 | 1637 | ||
853 | /* | 1638 | /* |
854 | * make mem_cgroup's charge to be 0 if there is no task. | 1639 | * make mem_cgroup's charge to be 0 if there is no task. |
855 | * This enables deleting this mem_cgroup. | 1640 | * This enables deleting this mem_cgroup. |
856 | */ | 1641 | */ |
857 | static int mem_cgroup_force_empty(struct mem_cgroup *mem) | 1642 | static int mem_cgroup_force_empty(struct mem_cgroup *mem, bool free_all) |
858 | { | 1643 | { |
859 | int ret = -EBUSY; | 1644 | int ret; |
860 | int node, zid; | 1645 | int node, zid, shrink; |
1646 | int nr_retries = MEM_CGROUP_RECLAIM_RETRIES; | ||
1647 | struct cgroup *cgrp = mem->css.cgroup; | ||
861 | 1648 | ||
862 | css_get(&mem->css); | 1649 | css_get(&mem->css); |
863 | /* | 1650 | |
864 | * page reclaim code (kswapd etc..) will move pages between | 1651 | shrink = 0; |
865 | * active_list <-> inactive_list while we don't take a lock. | 1652 | /* should free all ? */ |
866 | * So, we have to do loop here until all lists are empty. | 1653 | if (free_all) |
867 | */ | 1654 | goto try_to_free; |
1655 | move_account: | ||
868 | while (mem->res.usage > 0) { | 1656 | while (mem->res.usage > 0) { |
869 | if (atomic_read(&mem->css.cgroup->count) > 0) | 1657 | ret = -EBUSY; |
1658 | if (cgroup_task_count(cgrp) || !list_empty(&cgrp->children)) | ||
1659 | goto out; | ||
1660 | ret = -EINTR; | ||
1661 | if (signal_pending(current)) | ||
870 | goto out; | 1662 | goto out; |
871 | /* This is for making all *used* pages to be on LRU. */ | 1663 | /* This is for making all *used* pages to be on LRU. */ |
872 | lru_add_drain_all(); | 1664 | lru_add_drain_all(); |
873 | for_each_node_state(node, N_POSSIBLE) | 1665 | ret = 0; |
874 | for (zid = 0; zid < MAX_NR_ZONES; zid++) { | 1666 | for_each_node_state(node, N_POSSIBLE) { |
875 | struct mem_cgroup_per_zone *mz; | 1667 | for (zid = 0; !ret && zid < MAX_NR_ZONES; zid++) { |
876 | enum lru_list l; | 1668 | enum lru_list l; |
877 | mz = mem_cgroup_zoneinfo(mem, node, zid); | 1669 | for_each_lru(l) { |
878 | for_each_lru(l) | 1670 | ret = mem_cgroup_force_empty_list(mem, |
879 | mem_cgroup_force_empty_list(mem, mz, l); | 1671 | node, zid, l); |
1672 | if (ret) | ||
1673 | break; | ||
1674 | } | ||
880 | } | 1675 | } |
1676 | if (ret) | ||
1677 | break; | ||
1678 | } | ||
1679 | /* it seems parent cgroup doesn't have enough mem */ | ||
1680 | if (ret == -ENOMEM) | ||
1681 | goto try_to_free; | ||
881 | cond_resched(); | 1682 | cond_resched(); |
882 | } | 1683 | } |
883 | ret = 0; | 1684 | ret = 0; |
884 | out: | 1685 | out: |
885 | css_put(&mem->css); | 1686 | css_put(&mem->css); |
886 | return ret; | 1687 | return ret; |
1688 | |||
1689 | try_to_free: | ||
1690 | /* returns EBUSY if there is a task or if we come here twice. */ | ||
1691 | if (cgroup_task_count(cgrp) || !list_empty(&cgrp->children) || shrink) { | ||
1692 | ret = -EBUSY; | ||
1693 | goto out; | ||
1694 | } | ||
1695 | /* we call try-to-free pages for make this cgroup empty */ | ||
1696 | lru_add_drain_all(); | ||
1697 | /* try to free all pages in this cgroup */ | ||
1698 | shrink = 1; | ||
1699 | while (nr_retries && mem->res.usage > 0) { | ||
1700 | int progress; | ||
1701 | |||
1702 | if (signal_pending(current)) { | ||
1703 | ret = -EINTR; | ||
1704 | goto out; | ||
1705 | } | ||
1706 | progress = try_to_free_mem_cgroup_pages(mem, GFP_KERNEL, | ||
1707 | false, get_swappiness(mem)); | ||
1708 | if (!progress) { | ||
1709 | nr_retries--; | ||
1710 | /* maybe some writeback is necessary */ | ||
1711 | congestion_wait(WRITE, HZ/10); | ||
1712 | } | ||
1713 | |||
1714 | } | ||
1715 | lru_add_drain(); | ||
1716 | /* try move_account...there may be some *locked* pages. */ | ||
1717 | if (mem->res.usage) | ||
1718 | goto move_account; | ||
1719 | ret = 0; | ||
1720 | goto out; | ||
1721 | } | ||
1722 | |||
1723 | int mem_cgroup_force_empty_write(struct cgroup *cont, unsigned int event) | ||
1724 | { | ||
1725 | return mem_cgroup_force_empty(mem_cgroup_from_cont(cont), true); | ||
1726 | } | ||
1727 | |||
1728 | |||
1729 | static u64 mem_cgroup_hierarchy_read(struct cgroup *cont, struct cftype *cft) | ||
1730 | { | ||
1731 | return mem_cgroup_from_cont(cont)->use_hierarchy; | ||
1732 | } | ||
1733 | |||
1734 | static int mem_cgroup_hierarchy_write(struct cgroup *cont, struct cftype *cft, | ||
1735 | u64 val) | ||
1736 | { | ||
1737 | int retval = 0; | ||
1738 | struct mem_cgroup *mem = mem_cgroup_from_cont(cont); | ||
1739 | struct cgroup *parent = cont->parent; | ||
1740 | struct mem_cgroup *parent_mem = NULL; | ||
1741 | |||
1742 | if (parent) | ||
1743 | parent_mem = mem_cgroup_from_cont(parent); | ||
1744 | |||
1745 | cgroup_lock(); | ||
1746 | /* | ||
1747 | * If parent's use_hiearchy is set, we can't make any modifications | ||
1748 | * in the child subtrees. If it is unset, then the change can | ||
1749 | * occur, provided the current cgroup has no children. | ||
1750 | * | ||
1751 | * For the root cgroup, parent_mem is NULL, we allow value to be | ||
1752 | * set if there are no children. | ||
1753 | */ | ||
1754 | if ((!parent_mem || !parent_mem->use_hierarchy) && | ||
1755 | (val == 1 || val == 0)) { | ||
1756 | if (list_empty(&cont->children)) | ||
1757 | mem->use_hierarchy = val; | ||
1758 | else | ||
1759 | retval = -EBUSY; | ||
1760 | } else | ||
1761 | retval = -EINVAL; | ||
1762 | cgroup_unlock(); | ||
1763 | |||
1764 | return retval; | ||
887 | } | 1765 | } |
888 | 1766 | ||
889 | static u64 mem_cgroup_read(struct cgroup *cont, struct cftype *cft) | 1767 | static u64 mem_cgroup_read(struct cgroup *cont, struct cftype *cft) |
890 | { | 1768 | { |
891 | return res_counter_read_u64(&mem_cgroup_from_cont(cont)->res, | 1769 | struct mem_cgroup *mem = mem_cgroup_from_cont(cont); |
892 | cft->private); | 1770 | u64 val = 0; |
1771 | int type, name; | ||
1772 | |||
1773 | type = MEMFILE_TYPE(cft->private); | ||
1774 | name = MEMFILE_ATTR(cft->private); | ||
1775 | switch (type) { | ||
1776 | case _MEM: | ||
1777 | val = res_counter_read_u64(&mem->res, name); | ||
1778 | break; | ||
1779 | case _MEMSWAP: | ||
1780 | if (do_swap_account) | ||
1781 | val = res_counter_read_u64(&mem->memsw, name); | ||
1782 | break; | ||
1783 | default: | ||
1784 | BUG(); | ||
1785 | break; | ||
1786 | } | ||
1787 | return val; | ||
893 | } | 1788 | } |
894 | /* | 1789 | /* |
895 | * The user of this function is... | 1790 | * The user of this function is... |
@@ -899,15 +1794,22 @@ static int mem_cgroup_write(struct cgroup *cont, struct cftype *cft, | |||
899 | const char *buffer) | 1794 | const char *buffer) |
900 | { | 1795 | { |
901 | struct mem_cgroup *memcg = mem_cgroup_from_cont(cont); | 1796 | struct mem_cgroup *memcg = mem_cgroup_from_cont(cont); |
1797 | int type, name; | ||
902 | unsigned long long val; | 1798 | unsigned long long val; |
903 | int ret; | 1799 | int ret; |
904 | 1800 | ||
905 | switch (cft->private) { | 1801 | type = MEMFILE_TYPE(cft->private); |
1802 | name = MEMFILE_ATTR(cft->private); | ||
1803 | switch (name) { | ||
906 | case RES_LIMIT: | 1804 | case RES_LIMIT: |
907 | /* This function does all necessary parse...reuse it */ | 1805 | /* This function does all necessary parse...reuse it */ |
908 | ret = res_counter_memparse_write_strategy(buffer, &val); | 1806 | ret = res_counter_memparse_write_strategy(buffer, &val); |
909 | if (!ret) | 1807 | if (ret) |
1808 | break; | ||
1809 | if (type == _MEM) | ||
910 | ret = mem_cgroup_resize_limit(memcg, val); | 1810 | ret = mem_cgroup_resize_limit(memcg, val); |
1811 | else | ||
1812 | ret = mem_cgroup_resize_memsw_limit(memcg, val); | ||
911 | break; | 1813 | break; |
912 | default: | 1814 | default: |
913 | ret = -EINVAL; /* should be BUG() ? */ | 1815 | ret = -EINVAL; /* should be BUG() ? */ |
@@ -916,27 +1818,59 @@ static int mem_cgroup_write(struct cgroup *cont, struct cftype *cft, | |||
916 | return ret; | 1818 | return ret; |
917 | } | 1819 | } |
918 | 1820 | ||
1821 | static void memcg_get_hierarchical_limit(struct mem_cgroup *memcg, | ||
1822 | unsigned long long *mem_limit, unsigned long long *memsw_limit) | ||
1823 | { | ||
1824 | struct cgroup *cgroup; | ||
1825 | unsigned long long min_limit, min_memsw_limit, tmp; | ||
1826 | |||
1827 | min_limit = res_counter_read_u64(&memcg->res, RES_LIMIT); | ||
1828 | min_memsw_limit = res_counter_read_u64(&memcg->memsw, RES_LIMIT); | ||
1829 | cgroup = memcg->css.cgroup; | ||
1830 | if (!memcg->use_hierarchy) | ||
1831 | goto out; | ||
1832 | |||
1833 | while (cgroup->parent) { | ||
1834 | cgroup = cgroup->parent; | ||
1835 | memcg = mem_cgroup_from_cont(cgroup); | ||
1836 | if (!memcg->use_hierarchy) | ||
1837 | break; | ||
1838 | tmp = res_counter_read_u64(&memcg->res, RES_LIMIT); | ||
1839 | min_limit = min(min_limit, tmp); | ||
1840 | tmp = res_counter_read_u64(&memcg->memsw, RES_LIMIT); | ||
1841 | min_memsw_limit = min(min_memsw_limit, tmp); | ||
1842 | } | ||
1843 | out: | ||
1844 | *mem_limit = min_limit; | ||
1845 | *memsw_limit = min_memsw_limit; | ||
1846 | return; | ||
1847 | } | ||
1848 | |||
919 | static int mem_cgroup_reset(struct cgroup *cont, unsigned int event) | 1849 | static int mem_cgroup_reset(struct cgroup *cont, unsigned int event) |
920 | { | 1850 | { |
921 | struct mem_cgroup *mem; | 1851 | struct mem_cgroup *mem; |
1852 | int type, name; | ||
922 | 1853 | ||
923 | mem = mem_cgroup_from_cont(cont); | 1854 | mem = mem_cgroup_from_cont(cont); |
924 | switch (event) { | 1855 | type = MEMFILE_TYPE(event); |
1856 | name = MEMFILE_ATTR(event); | ||
1857 | switch (name) { | ||
925 | case RES_MAX_USAGE: | 1858 | case RES_MAX_USAGE: |
926 | res_counter_reset_max(&mem->res); | 1859 | if (type == _MEM) |
1860 | res_counter_reset_max(&mem->res); | ||
1861 | else | ||
1862 | res_counter_reset_max(&mem->memsw); | ||
927 | break; | 1863 | break; |
928 | case RES_FAILCNT: | 1864 | case RES_FAILCNT: |
929 | res_counter_reset_failcnt(&mem->res); | 1865 | if (type == _MEM) |
1866 | res_counter_reset_failcnt(&mem->res); | ||
1867 | else | ||
1868 | res_counter_reset_failcnt(&mem->memsw); | ||
930 | break; | 1869 | break; |
931 | } | 1870 | } |
932 | return 0; | 1871 | return 0; |
933 | } | 1872 | } |
934 | 1873 | ||
935 | static int mem_force_empty_write(struct cgroup *cont, unsigned int event) | ||
936 | { | ||
937 | return mem_cgroup_force_empty(mem_cgroup_from_cont(cont)); | ||
938 | } | ||
939 | |||
940 | static const struct mem_cgroup_stat_desc { | 1874 | static const struct mem_cgroup_stat_desc { |
941 | const char *msg; | 1875 | const char *msg; |
942 | u64 unit; | 1876 | u64 unit; |
@@ -985,43 +1919,163 @@ static int mem_control_stat_show(struct cgroup *cont, struct cftype *cft, | |||
985 | cb->fill(cb, "unevictable", unevictable * PAGE_SIZE); | 1919 | cb->fill(cb, "unevictable", unevictable * PAGE_SIZE); |
986 | 1920 | ||
987 | } | 1921 | } |
1922 | { | ||
1923 | unsigned long long limit, memsw_limit; | ||
1924 | memcg_get_hierarchical_limit(mem_cont, &limit, &memsw_limit); | ||
1925 | cb->fill(cb, "hierarchical_memory_limit", limit); | ||
1926 | if (do_swap_account) | ||
1927 | cb->fill(cb, "hierarchical_memsw_limit", memsw_limit); | ||
1928 | } | ||
1929 | |||
1930 | #ifdef CONFIG_DEBUG_VM | ||
1931 | cb->fill(cb, "inactive_ratio", calc_inactive_ratio(mem_cont, NULL)); | ||
1932 | |||
1933 | { | ||
1934 | int nid, zid; | ||
1935 | struct mem_cgroup_per_zone *mz; | ||
1936 | unsigned long recent_rotated[2] = {0, 0}; | ||
1937 | unsigned long recent_scanned[2] = {0, 0}; | ||
1938 | |||
1939 | for_each_online_node(nid) | ||
1940 | for (zid = 0; zid < MAX_NR_ZONES; zid++) { | ||
1941 | mz = mem_cgroup_zoneinfo(mem_cont, nid, zid); | ||
1942 | |||
1943 | recent_rotated[0] += | ||
1944 | mz->reclaim_stat.recent_rotated[0]; | ||
1945 | recent_rotated[1] += | ||
1946 | mz->reclaim_stat.recent_rotated[1]; | ||
1947 | recent_scanned[0] += | ||
1948 | mz->reclaim_stat.recent_scanned[0]; | ||
1949 | recent_scanned[1] += | ||
1950 | mz->reclaim_stat.recent_scanned[1]; | ||
1951 | } | ||
1952 | cb->fill(cb, "recent_rotated_anon", recent_rotated[0]); | ||
1953 | cb->fill(cb, "recent_rotated_file", recent_rotated[1]); | ||
1954 | cb->fill(cb, "recent_scanned_anon", recent_scanned[0]); | ||
1955 | cb->fill(cb, "recent_scanned_file", recent_scanned[1]); | ||
1956 | } | ||
1957 | #endif | ||
1958 | |||
1959 | return 0; | ||
1960 | } | ||
1961 | |||
1962 | static u64 mem_cgroup_swappiness_read(struct cgroup *cgrp, struct cftype *cft) | ||
1963 | { | ||
1964 | struct mem_cgroup *memcg = mem_cgroup_from_cont(cgrp); | ||
1965 | |||
1966 | return get_swappiness(memcg); | ||
1967 | } | ||
1968 | |||
1969 | static int mem_cgroup_swappiness_write(struct cgroup *cgrp, struct cftype *cft, | ||
1970 | u64 val) | ||
1971 | { | ||
1972 | struct mem_cgroup *memcg = mem_cgroup_from_cont(cgrp); | ||
1973 | struct mem_cgroup *parent; | ||
1974 | if (val > 100) | ||
1975 | return -EINVAL; | ||
1976 | |||
1977 | if (cgrp->parent == NULL) | ||
1978 | return -EINVAL; | ||
1979 | |||
1980 | parent = mem_cgroup_from_cont(cgrp->parent); | ||
1981 | /* If under hierarchy, only empty-root can set this value */ | ||
1982 | if ((parent->use_hierarchy) || | ||
1983 | (memcg->use_hierarchy && !list_empty(&cgrp->children))) | ||
1984 | return -EINVAL; | ||
1985 | |||
1986 | spin_lock(&memcg->reclaim_param_lock); | ||
1987 | memcg->swappiness = val; | ||
1988 | spin_unlock(&memcg->reclaim_param_lock); | ||
1989 | |||
988 | return 0; | 1990 | return 0; |
989 | } | 1991 | } |
990 | 1992 | ||
1993 | |||
991 | static struct cftype mem_cgroup_files[] = { | 1994 | static struct cftype mem_cgroup_files[] = { |
992 | { | 1995 | { |
993 | .name = "usage_in_bytes", | 1996 | .name = "usage_in_bytes", |
994 | .private = RES_USAGE, | 1997 | .private = MEMFILE_PRIVATE(_MEM, RES_USAGE), |
995 | .read_u64 = mem_cgroup_read, | 1998 | .read_u64 = mem_cgroup_read, |
996 | }, | 1999 | }, |
997 | { | 2000 | { |
998 | .name = "max_usage_in_bytes", | 2001 | .name = "max_usage_in_bytes", |
999 | .private = RES_MAX_USAGE, | 2002 | .private = MEMFILE_PRIVATE(_MEM, RES_MAX_USAGE), |
1000 | .trigger = mem_cgroup_reset, | 2003 | .trigger = mem_cgroup_reset, |
1001 | .read_u64 = mem_cgroup_read, | 2004 | .read_u64 = mem_cgroup_read, |
1002 | }, | 2005 | }, |
1003 | { | 2006 | { |
1004 | .name = "limit_in_bytes", | 2007 | .name = "limit_in_bytes", |
1005 | .private = RES_LIMIT, | 2008 | .private = MEMFILE_PRIVATE(_MEM, RES_LIMIT), |
1006 | .write_string = mem_cgroup_write, | 2009 | .write_string = mem_cgroup_write, |
1007 | .read_u64 = mem_cgroup_read, | 2010 | .read_u64 = mem_cgroup_read, |
1008 | }, | 2011 | }, |
1009 | { | 2012 | { |
1010 | .name = "failcnt", | 2013 | .name = "failcnt", |
1011 | .private = RES_FAILCNT, | 2014 | .private = MEMFILE_PRIVATE(_MEM, RES_FAILCNT), |
1012 | .trigger = mem_cgroup_reset, | 2015 | .trigger = mem_cgroup_reset, |
1013 | .read_u64 = mem_cgroup_read, | 2016 | .read_u64 = mem_cgroup_read, |
1014 | }, | 2017 | }, |
1015 | { | 2018 | { |
2019 | .name = "stat", | ||
2020 | .read_map = mem_control_stat_show, | ||
2021 | }, | ||
2022 | { | ||
1016 | .name = "force_empty", | 2023 | .name = "force_empty", |
1017 | .trigger = mem_force_empty_write, | 2024 | .trigger = mem_cgroup_force_empty_write, |
1018 | }, | 2025 | }, |
1019 | { | 2026 | { |
1020 | .name = "stat", | 2027 | .name = "use_hierarchy", |
1021 | .read_map = mem_control_stat_show, | 2028 | .write_u64 = mem_cgroup_hierarchy_write, |
2029 | .read_u64 = mem_cgroup_hierarchy_read, | ||
2030 | }, | ||
2031 | { | ||
2032 | .name = "swappiness", | ||
2033 | .read_u64 = mem_cgroup_swappiness_read, | ||
2034 | .write_u64 = mem_cgroup_swappiness_write, | ||
1022 | }, | 2035 | }, |
1023 | }; | 2036 | }; |
1024 | 2037 | ||
2038 | #ifdef CONFIG_CGROUP_MEM_RES_CTLR_SWAP | ||
2039 | static struct cftype memsw_cgroup_files[] = { | ||
2040 | { | ||
2041 | .name = "memsw.usage_in_bytes", | ||
2042 | .private = MEMFILE_PRIVATE(_MEMSWAP, RES_USAGE), | ||
2043 | .read_u64 = mem_cgroup_read, | ||
2044 | }, | ||
2045 | { | ||
2046 | .name = "memsw.max_usage_in_bytes", | ||
2047 | .private = MEMFILE_PRIVATE(_MEMSWAP, RES_MAX_USAGE), | ||
2048 | .trigger = mem_cgroup_reset, | ||
2049 | .read_u64 = mem_cgroup_read, | ||
2050 | }, | ||
2051 | { | ||
2052 | .name = "memsw.limit_in_bytes", | ||
2053 | .private = MEMFILE_PRIVATE(_MEMSWAP, RES_LIMIT), | ||
2054 | .write_string = mem_cgroup_write, | ||
2055 | .read_u64 = mem_cgroup_read, | ||
2056 | }, | ||
2057 | { | ||
2058 | .name = "memsw.failcnt", | ||
2059 | .private = MEMFILE_PRIVATE(_MEMSWAP, RES_FAILCNT), | ||
2060 | .trigger = mem_cgroup_reset, | ||
2061 | .read_u64 = mem_cgroup_read, | ||
2062 | }, | ||
2063 | }; | ||
2064 | |||
2065 | static int register_memsw_files(struct cgroup *cont, struct cgroup_subsys *ss) | ||
2066 | { | ||
2067 | if (!do_swap_account) | ||
2068 | return 0; | ||
2069 | return cgroup_add_files(cont, ss, memsw_cgroup_files, | ||
2070 | ARRAY_SIZE(memsw_cgroup_files)); | ||
2071 | }; | ||
2072 | #else | ||
2073 | static int register_memsw_files(struct cgroup *cont, struct cgroup_subsys *ss) | ||
2074 | { | ||
2075 | return 0; | ||
2076 | } | ||
2077 | #endif | ||
2078 | |||
1025 | static int alloc_mem_cgroup_per_zone_info(struct mem_cgroup *mem, int node) | 2079 | static int alloc_mem_cgroup_per_zone_info(struct mem_cgroup *mem, int node) |
1026 | { | 2080 | { |
1027 | struct mem_cgroup_per_node *pn; | 2081 | struct mem_cgroup_per_node *pn; |
@@ -1047,7 +2101,6 @@ static int alloc_mem_cgroup_per_zone_info(struct mem_cgroup *mem, int node) | |||
1047 | 2101 | ||
1048 | for (zone = 0; zone < MAX_NR_ZONES; zone++) { | 2102 | for (zone = 0; zone < MAX_NR_ZONES; zone++) { |
1049 | mz = &pn->zoneinfo[zone]; | 2103 | mz = &pn->zoneinfo[zone]; |
1050 | spin_lock_init(&mz->lru_lock); | ||
1051 | for_each_lru(l) | 2104 | for_each_lru(l) |
1052 | INIT_LIST_HEAD(&mz->lists[l]); | 2105 | INIT_LIST_HEAD(&mz->lists[l]); |
1053 | } | 2106 | } |
@@ -1059,55 +2112,113 @@ static void free_mem_cgroup_per_zone_info(struct mem_cgroup *mem, int node) | |||
1059 | kfree(mem->info.nodeinfo[node]); | 2112 | kfree(mem->info.nodeinfo[node]); |
1060 | } | 2113 | } |
1061 | 2114 | ||
2115 | static int mem_cgroup_size(void) | ||
2116 | { | ||
2117 | int cpustat_size = nr_cpu_ids * sizeof(struct mem_cgroup_stat_cpu); | ||
2118 | return sizeof(struct mem_cgroup) + cpustat_size; | ||
2119 | } | ||
2120 | |||
1062 | static struct mem_cgroup *mem_cgroup_alloc(void) | 2121 | static struct mem_cgroup *mem_cgroup_alloc(void) |
1063 | { | 2122 | { |
1064 | struct mem_cgroup *mem; | 2123 | struct mem_cgroup *mem; |
2124 | int size = mem_cgroup_size(); | ||
1065 | 2125 | ||
1066 | if (sizeof(*mem) < PAGE_SIZE) | 2126 | if (size < PAGE_SIZE) |
1067 | mem = kmalloc(sizeof(*mem), GFP_KERNEL); | 2127 | mem = kmalloc(size, GFP_KERNEL); |
1068 | else | 2128 | else |
1069 | mem = vmalloc(sizeof(*mem)); | 2129 | mem = vmalloc(size); |
1070 | 2130 | ||
1071 | if (mem) | 2131 | if (mem) |
1072 | memset(mem, 0, sizeof(*mem)); | 2132 | memset(mem, 0, size); |
1073 | return mem; | 2133 | return mem; |
1074 | } | 2134 | } |
1075 | 2135 | ||
1076 | static void mem_cgroup_free(struct mem_cgroup *mem) | 2136 | /* |
2137 | * At destroying mem_cgroup, references from swap_cgroup can remain. | ||
2138 | * (scanning all at force_empty is too costly...) | ||
2139 | * | ||
2140 | * Instead of clearing all references at force_empty, we remember | ||
2141 | * the number of reference from swap_cgroup and free mem_cgroup when | ||
2142 | * it goes down to 0. | ||
2143 | * | ||
2144 | * Removal of cgroup itself succeeds regardless of refs from swap. | ||
2145 | */ | ||
2146 | |||
2147 | static void __mem_cgroup_free(struct mem_cgroup *mem) | ||
1077 | { | 2148 | { |
1078 | if (sizeof(*mem) < PAGE_SIZE) | 2149 | int node; |
2150 | |||
2151 | for_each_node_state(node, N_POSSIBLE) | ||
2152 | free_mem_cgroup_per_zone_info(mem, node); | ||
2153 | |||
2154 | if (mem_cgroup_size() < PAGE_SIZE) | ||
1079 | kfree(mem); | 2155 | kfree(mem); |
1080 | else | 2156 | else |
1081 | vfree(mem); | 2157 | vfree(mem); |
1082 | } | 2158 | } |
1083 | 2159 | ||
2160 | static void mem_cgroup_get(struct mem_cgroup *mem) | ||
2161 | { | ||
2162 | atomic_inc(&mem->refcnt); | ||
2163 | } | ||
2164 | |||
2165 | static void mem_cgroup_put(struct mem_cgroup *mem) | ||
2166 | { | ||
2167 | if (atomic_dec_and_test(&mem->refcnt)) | ||
2168 | __mem_cgroup_free(mem); | ||
2169 | } | ||
2170 | |||
2171 | |||
2172 | #ifdef CONFIG_CGROUP_MEM_RES_CTLR_SWAP | ||
2173 | static void __init enable_swap_cgroup(void) | ||
2174 | { | ||
2175 | if (!mem_cgroup_disabled() && really_do_swap_account) | ||
2176 | do_swap_account = 1; | ||
2177 | } | ||
2178 | #else | ||
2179 | static void __init enable_swap_cgroup(void) | ||
2180 | { | ||
2181 | } | ||
2182 | #endif | ||
1084 | 2183 | ||
1085 | static struct cgroup_subsys_state * | 2184 | static struct cgroup_subsys_state * |
1086 | mem_cgroup_create(struct cgroup_subsys *ss, struct cgroup *cont) | 2185 | mem_cgroup_create(struct cgroup_subsys *ss, struct cgroup *cont) |
1087 | { | 2186 | { |
1088 | struct mem_cgroup *mem; | 2187 | struct mem_cgroup *mem, *parent; |
1089 | int node; | 2188 | int node; |
1090 | 2189 | ||
1091 | if (unlikely((cont->parent) == NULL)) { | 2190 | mem = mem_cgroup_alloc(); |
1092 | mem = &init_mem_cgroup; | 2191 | if (!mem) |
1093 | } else { | 2192 | return ERR_PTR(-ENOMEM); |
1094 | mem = mem_cgroup_alloc(); | ||
1095 | if (!mem) | ||
1096 | return ERR_PTR(-ENOMEM); | ||
1097 | } | ||
1098 | |||
1099 | res_counter_init(&mem->res); | ||
1100 | 2193 | ||
1101 | for_each_node_state(node, N_POSSIBLE) | 2194 | for_each_node_state(node, N_POSSIBLE) |
1102 | if (alloc_mem_cgroup_per_zone_info(mem, node)) | 2195 | if (alloc_mem_cgroup_per_zone_info(mem, node)) |
1103 | goto free_out; | 2196 | goto free_out; |
2197 | /* root ? */ | ||
2198 | if (cont->parent == NULL) { | ||
2199 | enable_swap_cgroup(); | ||
2200 | parent = NULL; | ||
2201 | } else { | ||
2202 | parent = mem_cgroup_from_cont(cont->parent); | ||
2203 | mem->use_hierarchy = parent->use_hierarchy; | ||
2204 | } | ||
1104 | 2205 | ||
2206 | if (parent && parent->use_hierarchy) { | ||
2207 | res_counter_init(&mem->res, &parent->res); | ||
2208 | res_counter_init(&mem->memsw, &parent->memsw); | ||
2209 | } else { | ||
2210 | res_counter_init(&mem->res, NULL); | ||
2211 | res_counter_init(&mem->memsw, NULL); | ||
2212 | } | ||
2213 | mem->last_scanned_child = NULL; | ||
2214 | spin_lock_init(&mem->reclaim_param_lock); | ||
2215 | |||
2216 | if (parent) | ||
2217 | mem->swappiness = get_swappiness(parent); | ||
2218 | atomic_set(&mem->refcnt, 1); | ||
1105 | return &mem->css; | 2219 | return &mem->css; |
1106 | free_out: | 2220 | free_out: |
1107 | for_each_node_state(node, N_POSSIBLE) | 2221 | __mem_cgroup_free(mem); |
1108 | free_mem_cgroup_per_zone_info(mem, node); | ||
1109 | if (cont->parent != NULL) | ||
1110 | mem_cgroup_free(mem); | ||
1111 | return ERR_PTR(-ENOMEM); | 2222 | return ERR_PTR(-ENOMEM); |
1112 | } | 2223 | } |
1113 | 2224 | ||
@@ -1115,26 +2226,26 @@ static void mem_cgroup_pre_destroy(struct cgroup_subsys *ss, | |||
1115 | struct cgroup *cont) | 2226 | struct cgroup *cont) |
1116 | { | 2227 | { |
1117 | struct mem_cgroup *mem = mem_cgroup_from_cont(cont); | 2228 | struct mem_cgroup *mem = mem_cgroup_from_cont(cont); |
1118 | mem_cgroup_force_empty(mem); | 2229 | mem_cgroup_force_empty(mem, false); |
1119 | } | 2230 | } |
1120 | 2231 | ||
1121 | static void mem_cgroup_destroy(struct cgroup_subsys *ss, | 2232 | static void mem_cgroup_destroy(struct cgroup_subsys *ss, |
1122 | struct cgroup *cont) | 2233 | struct cgroup *cont) |
1123 | { | 2234 | { |
1124 | int node; | 2235 | mem_cgroup_put(mem_cgroup_from_cont(cont)); |
1125 | struct mem_cgroup *mem = mem_cgroup_from_cont(cont); | ||
1126 | |||
1127 | for_each_node_state(node, N_POSSIBLE) | ||
1128 | free_mem_cgroup_per_zone_info(mem, node); | ||
1129 | |||
1130 | mem_cgroup_free(mem_cgroup_from_cont(cont)); | ||
1131 | } | 2236 | } |
1132 | 2237 | ||
1133 | static int mem_cgroup_populate(struct cgroup_subsys *ss, | 2238 | static int mem_cgroup_populate(struct cgroup_subsys *ss, |
1134 | struct cgroup *cont) | 2239 | struct cgroup *cont) |
1135 | { | 2240 | { |
1136 | return cgroup_add_files(cont, ss, mem_cgroup_files, | 2241 | int ret; |
1137 | ARRAY_SIZE(mem_cgroup_files)); | 2242 | |
2243 | ret = cgroup_add_files(cont, ss, mem_cgroup_files, | ||
2244 | ARRAY_SIZE(mem_cgroup_files)); | ||
2245 | |||
2246 | if (!ret) | ||
2247 | ret = register_memsw_files(cont, ss); | ||
2248 | return ret; | ||
1138 | } | 2249 | } |
1139 | 2250 | ||
1140 | static void mem_cgroup_move_task(struct cgroup_subsys *ss, | 2251 | static void mem_cgroup_move_task(struct cgroup_subsys *ss, |
@@ -1142,25 +2253,12 @@ static void mem_cgroup_move_task(struct cgroup_subsys *ss, | |||
1142 | struct cgroup *old_cont, | 2253 | struct cgroup *old_cont, |
1143 | struct task_struct *p) | 2254 | struct task_struct *p) |
1144 | { | 2255 | { |
1145 | struct mm_struct *mm; | 2256 | mutex_lock(&memcg_tasklist); |
1146 | struct mem_cgroup *mem, *old_mem; | ||
1147 | |||
1148 | mm = get_task_mm(p); | ||
1149 | if (mm == NULL) | ||
1150 | return; | ||
1151 | |||
1152 | mem = mem_cgroup_from_cont(cont); | ||
1153 | old_mem = mem_cgroup_from_cont(old_cont); | ||
1154 | |||
1155 | /* | 2257 | /* |
1156 | * Only thread group leaders are allowed to migrate, the mm_struct is | 2258 | * FIXME: It's better to move charges of this process from old |
1157 | * in effect owned by the leader | 2259 | * memcg to new memcg. But it's just on TODO-List now. |
1158 | */ | 2260 | */ |
1159 | if (!thread_group_leader(p)) | 2261 | mutex_unlock(&memcg_tasklist); |
1160 | goto out; | ||
1161 | |||
1162 | out: | ||
1163 | mmput(mm); | ||
1164 | } | 2262 | } |
1165 | 2263 | ||
1166 | struct cgroup_subsys mem_cgroup_subsys = { | 2264 | struct cgroup_subsys mem_cgroup_subsys = { |
@@ -1173,3 +2271,13 @@ struct cgroup_subsys mem_cgroup_subsys = { | |||
1173 | .attach = mem_cgroup_move_task, | 2271 | .attach = mem_cgroup_move_task, |
1174 | .early_init = 0, | 2272 | .early_init = 0, |
1175 | }; | 2273 | }; |
2274 | |||
2275 | #ifdef CONFIG_CGROUP_MEM_RES_CTLR_SWAP | ||
2276 | |||
2277 | static int __init disable_swap_account(char *s) | ||
2278 | { | ||
2279 | really_do_swap_account = 0; | ||
2280 | return 1; | ||
2281 | } | ||
2282 | __setup("noswapaccount", disable_swap_account); | ||
2283 | #endif | ||
diff --git a/mm/memory.c b/mm/memory.c index 3f8fa06b963b..e009ce870859 100644 --- a/mm/memory.c +++ b/mm/memory.c | |||
@@ -2000,7 +2000,7 @@ gotten: | |||
2000 | cow_user_page(new_page, old_page, address, vma); | 2000 | cow_user_page(new_page, old_page, address, vma); |
2001 | __SetPageUptodate(new_page); | 2001 | __SetPageUptodate(new_page); |
2002 | 2002 | ||
2003 | if (mem_cgroup_charge(new_page, mm, GFP_KERNEL)) | 2003 | if (mem_cgroup_newpage_charge(new_page, mm, GFP_KERNEL)) |
2004 | goto oom_free_new; | 2004 | goto oom_free_new; |
2005 | 2005 | ||
2006 | /* | 2006 | /* |
@@ -2392,6 +2392,7 @@ static int do_swap_page(struct mm_struct *mm, struct vm_area_struct *vma, | |||
2392 | struct page *page; | 2392 | struct page *page; |
2393 | swp_entry_t entry; | 2393 | swp_entry_t entry; |
2394 | pte_t pte; | 2394 | pte_t pte; |
2395 | struct mem_cgroup *ptr = NULL; | ||
2395 | int ret = 0; | 2396 | int ret = 0; |
2396 | 2397 | ||
2397 | if (!pte_unmap_same(mm, pmd, page_table, orig_pte)) | 2398 | if (!pte_unmap_same(mm, pmd, page_table, orig_pte)) |
@@ -2430,7 +2431,7 @@ static int do_swap_page(struct mm_struct *mm, struct vm_area_struct *vma, | |||
2430 | lock_page(page); | 2431 | lock_page(page); |
2431 | delayacct_clear_flag(DELAYACCT_PF_SWAPIN); | 2432 | delayacct_clear_flag(DELAYACCT_PF_SWAPIN); |
2432 | 2433 | ||
2433 | if (mem_cgroup_charge(page, mm, GFP_KERNEL)) { | 2434 | if (mem_cgroup_try_charge_swapin(mm, page, GFP_KERNEL, &ptr)) { |
2434 | ret = VM_FAULT_OOM; | 2435 | ret = VM_FAULT_OOM; |
2435 | unlock_page(page); | 2436 | unlock_page(page); |
2436 | goto out; | 2437 | goto out; |
@@ -2448,7 +2449,19 @@ static int do_swap_page(struct mm_struct *mm, struct vm_area_struct *vma, | |||
2448 | goto out_nomap; | 2449 | goto out_nomap; |
2449 | } | 2450 | } |
2450 | 2451 | ||
2451 | /* The page isn't present yet, go ahead with the fault. */ | 2452 | /* |
2453 | * The page isn't present yet, go ahead with the fault. | ||
2454 | * | ||
2455 | * Be careful about the sequence of operations here. | ||
2456 | * To get its accounting right, reuse_swap_page() must be called | ||
2457 | * while the page is counted on swap but not yet in mapcount i.e. | ||
2458 | * before page_add_anon_rmap() and swap_free(); try_to_free_swap() | ||
2459 | * must be called after the swap_free(), or it will never succeed. | ||
2460 | * Because delete_from_swap_page() may be called by reuse_swap_page(), | ||
2461 | * mem_cgroup_commit_charge_swapin() may not be able to find swp_entry | ||
2462 | * in page->private. In this case, a record in swap_cgroup is silently | ||
2463 | * discarded at swap_free(). | ||
2464 | */ | ||
2452 | 2465 | ||
2453 | inc_mm_counter(mm, anon_rss); | 2466 | inc_mm_counter(mm, anon_rss); |
2454 | pte = mk_pte(page, vma->vm_page_prot); | 2467 | pte = mk_pte(page, vma->vm_page_prot); |
@@ -2456,10 +2469,11 @@ static int do_swap_page(struct mm_struct *mm, struct vm_area_struct *vma, | |||
2456 | pte = maybe_mkwrite(pte_mkdirty(pte), vma); | 2469 | pte = maybe_mkwrite(pte_mkdirty(pte), vma); |
2457 | write_access = 0; | 2470 | write_access = 0; |
2458 | } | 2471 | } |
2459 | |||
2460 | flush_icache_page(vma, page); | 2472 | flush_icache_page(vma, page); |
2461 | set_pte_at(mm, address, page_table, pte); | 2473 | set_pte_at(mm, address, page_table, pte); |
2462 | page_add_anon_rmap(page, vma, address); | 2474 | page_add_anon_rmap(page, vma, address); |
2475 | /* It's better to call commit-charge after rmap is established */ | ||
2476 | mem_cgroup_commit_charge_swapin(page, ptr); | ||
2463 | 2477 | ||
2464 | swap_free(entry); | 2478 | swap_free(entry); |
2465 | if (vm_swap_full() || (vma->vm_flags & VM_LOCKED) || PageMlocked(page)) | 2479 | if (vm_swap_full() || (vma->vm_flags & VM_LOCKED) || PageMlocked(page)) |
@@ -2480,7 +2494,7 @@ unlock: | |||
2480 | out: | 2494 | out: |
2481 | return ret; | 2495 | return ret; |
2482 | out_nomap: | 2496 | out_nomap: |
2483 | mem_cgroup_uncharge_page(page); | 2497 | mem_cgroup_cancel_charge_swapin(ptr); |
2484 | pte_unmap_unlock(page_table, ptl); | 2498 | pte_unmap_unlock(page_table, ptl); |
2485 | unlock_page(page); | 2499 | unlock_page(page); |
2486 | page_cache_release(page); | 2500 | page_cache_release(page); |
@@ -2510,7 +2524,7 @@ static int do_anonymous_page(struct mm_struct *mm, struct vm_area_struct *vma, | |||
2510 | goto oom; | 2524 | goto oom; |
2511 | __SetPageUptodate(page); | 2525 | __SetPageUptodate(page); |
2512 | 2526 | ||
2513 | if (mem_cgroup_charge(page, mm, GFP_KERNEL)) | 2527 | if (mem_cgroup_newpage_charge(page, mm, GFP_KERNEL)) |
2514 | goto oom_free_page; | 2528 | goto oom_free_page; |
2515 | 2529 | ||
2516 | entry = mk_pte(page, vma->vm_page_prot); | 2530 | entry = mk_pte(page, vma->vm_page_prot); |
@@ -2601,7 +2615,7 @@ static int __do_fault(struct mm_struct *mm, struct vm_area_struct *vma, | |||
2601 | ret = VM_FAULT_OOM; | 2615 | ret = VM_FAULT_OOM; |
2602 | goto out; | 2616 | goto out; |
2603 | } | 2617 | } |
2604 | if (mem_cgroup_charge(page, mm, GFP_KERNEL)) { | 2618 | if (mem_cgroup_newpage_charge(page, mm, GFP_KERNEL)) { |
2605 | ret = VM_FAULT_OOM; | 2619 | ret = VM_FAULT_OOM; |
2606 | page_cache_release(page); | 2620 | page_cache_release(page); |
2607 | goto out; | 2621 | goto out; |
diff --git a/mm/migrate.c b/mm/migrate.c index 55373983c9c6..a30ea5fcf9f1 100644 --- a/mm/migrate.c +++ b/mm/migrate.c | |||
@@ -121,20 +121,6 @@ static void remove_migration_pte(struct vm_area_struct *vma, | |||
121 | if (!is_migration_entry(entry) || migration_entry_to_page(entry) != old) | 121 | if (!is_migration_entry(entry) || migration_entry_to_page(entry) != old) |
122 | goto out; | 122 | goto out; |
123 | 123 | ||
124 | /* | ||
125 | * Yes, ignore the return value from a GFP_ATOMIC mem_cgroup_charge. | ||
126 | * Failure is not an option here: we're now expected to remove every | ||
127 | * migration pte, and will cause crashes otherwise. Normally this | ||
128 | * is not an issue: mem_cgroup_prepare_migration bumped up the old | ||
129 | * page_cgroup count for safety, that's now attached to the new page, | ||
130 | * so this charge should just be another incrementation of the count, | ||
131 | * to keep in balance with rmap.c's mem_cgroup_uncharging. But if | ||
132 | * there's been a force_empty, those reference counts may no longer | ||
133 | * be reliable, and this charge can actually fail: oh well, we don't | ||
134 | * make the situation any worse by proceeding as if it had succeeded. | ||
135 | */ | ||
136 | mem_cgroup_charge(new, mm, GFP_ATOMIC); | ||
137 | |||
138 | get_page(new); | 124 | get_page(new); |
139 | pte = pte_mkold(mk_pte(new, vma->vm_page_prot)); | 125 | pte = pte_mkold(mk_pte(new, vma->vm_page_prot)); |
140 | if (is_write_migration_entry(entry)) | 126 | if (is_write_migration_entry(entry)) |
@@ -378,9 +364,6 @@ static void migrate_page_copy(struct page *newpage, struct page *page) | |||
378 | anon = PageAnon(page); | 364 | anon = PageAnon(page); |
379 | page->mapping = NULL; | 365 | page->mapping = NULL; |
380 | 366 | ||
381 | if (!anon) /* This page was removed from radix-tree. */ | ||
382 | mem_cgroup_uncharge_cache_page(page); | ||
383 | |||
384 | /* | 367 | /* |
385 | * If any waiters have accumulated on the new page then | 368 | * If any waiters have accumulated on the new page then |
386 | * wake them up. | 369 | * wake them up. |
@@ -614,6 +597,7 @@ static int unmap_and_move(new_page_t get_new_page, unsigned long private, | |||
614 | struct page *newpage = get_new_page(page, private, &result); | 597 | struct page *newpage = get_new_page(page, private, &result); |
615 | int rcu_locked = 0; | 598 | int rcu_locked = 0; |
616 | int charge = 0; | 599 | int charge = 0; |
600 | struct mem_cgroup *mem; | ||
617 | 601 | ||
618 | if (!newpage) | 602 | if (!newpage) |
619 | return -ENOMEM; | 603 | return -ENOMEM; |
@@ -623,24 +607,26 @@ static int unmap_and_move(new_page_t get_new_page, unsigned long private, | |||
623 | goto move_newpage; | 607 | goto move_newpage; |
624 | } | 608 | } |
625 | 609 | ||
626 | charge = mem_cgroup_prepare_migration(page, newpage); | ||
627 | if (charge == -ENOMEM) { | ||
628 | rc = -ENOMEM; | ||
629 | goto move_newpage; | ||
630 | } | ||
631 | /* prepare cgroup just returns 0 or -ENOMEM */ | 610 | /* prepare cgroup just returns 0 or -ENOMEM */ |
632 | BUG_ON(charge); | ||
633 | |||
634 | rc = -EAGAIN; | 611 | rc = -EAGAIN; |
612 | |||
635 | if (!trylock_page(page)) { | 613 | if (!trylock_page(page)) { |
636 | if (!force) | 614 | if (!force) |
637 | goto move_newpage; | 615 | goto move_newpage; |
638 | lock_page(page); | 616 | lock_page(page); |
639 | } | 617 | } |
640 | 618 | ||
619 | /* charge against new page */ | ||
620 | charge = mem_cgroup_prepare_migration(page, &mem); | ||
621 | if (charge == -ENOMEM) { | ||
622 | rc = -ENOMEM; | ||
623 | goto unlock; | ||
624 | } | ||
625 | BUG_ON(charge); | ||
626 | |||
641 | if (PageWriteback(page)) { | 627 | if (PageWriteback(page)) { |
642 | if (!force) | 628 | if (!force) |
643 | goto unlock; | 629 | goto uncharge; |
644 | wait_on_page_writeback(page); | 630 | wait_on_page_writeback(page); |
645 | } | 631 | } |
646 | /* | 632 | /* |
@@ -693,7 +679,9 @@ static int unmap_and_move(new_page_t get_new_page, unsigned long private, | |||
693 | rcu_unlock: | 679 | rcu_unlock: |
694 | if (rcu_locked) | 680 | if (rcu_locked) |
695 | rcu_read_unlock(); | 681 | rcu_read_unlock(); |
696 | 682 | uncharge: | |
683 | if (!charge) | ||
684 | mem_cgroup_end_migration(mem, page, newpage); | ||
697 | unlock: | 685 | unlock: |
698 | unlock_page(page); | 686 | unlock_page(page); |
699 | 687 | ||
@@ -709,8 +697,6 @@ unlock: | |||
709 | } | 697 | } |
710 | 698 | ||
711 | move_newpage: | 699 | move_newpage: |
712 | if (!charge) | ||
713 | mem_cgroup_end_migration(newpage); | ||
714 | 700 | ||
715 | /* | 701 | /* |
716 | * Move the new page to the LRU. If migration was not successful | 702 | * Move the new page to the LRU. If migration was not successful |
diff --git a/mm/oom_kill.c b/mm/oom_kill.c index 6b9e758c98a5..40ba05061a4f 100644 --- a/mm/oom_kill.c +++ b/mm/oom_kill.c | |||
@@ -429,7 +429,6 @@ void mem_cgroup_out_of_memory(struct mem_cgroup *mem, gfp_t gfp_mask) | |||
429 | unsigned long points = 0; | 429 | unsigned long points = 0; |
430 | struct task_struct *p; | 430 | struct task_struct *p; |
431 | 431 | ||
432 | cgroup_lock(); | ||
433 | read_lock(&tasklist_lock); | 432 | read_lock(&tasklist_lock); |
434 | retry: | 433 | retry: |
435 | p = select_bad_process(&points, mem); | 434 | p = select_bad_process(&points, mem); |
@@ -444,7 +443,6 @@ retry: | |||
444 | goto retry; | 443 | goto retry; |
445 | out: | 444 | out: |
446 | read_unlock(&tasklist_lock); | 445 | read_unlock(&tasklist_lock); |
447 | cgroup_unlock(); | ||
448 | } | 446 | } |
449 | #endif | 447 | #endif |
450 | 448 | ||
@@ -560,6 +558,13 @@ void pagefault_out_of_memory(void) | |||
560 | /* Got some memory back in the last second. */ | 558 | /* Got some memory back in the last second. */ |
561 | return; | 559 | return; |
562 | 560 | ||
561 | /* | ||
562 | * If this is from memcg, oom-killer is already invoked. | ||
563 | * and not worth to go system-wide-oom. | ||
564 | */ | ||
565 | if (mem_cgroup_oom_called(current)) | ||
566 | goto rest_and_return; | ||
567 | |||
563 | if (sysctl_panic_on_oom) | 568 | if (sysctl_panic_on_oom) |
564 | panic("out of memory from page fault. panic_on_oom is selected.\n"); | 569 | panic("out of memory from page fault. panic_on_oom is selected.\n"); |
565 | 570 | ||
@@ -571,6 +576,7 @@ void pagefault_out_of_memory(void) | |||
571 | * Give "p" a good chance of killing itself before we | 576 | * Give "p" a good chance of killing itself before we |
572 | * retry to allocate memory. | 577 | * retry to allocate memory. |
573 | */ | 578 | */ |
579 | rest_and_return: | ||
574 | if (!test_thread_flag(TIF_MEMDIE)) | 580 | if (!test_thread_flag(TIF_MEMDIE)) |
575 | schedule_timeout_uninterruptible(1); | 581 | schedule_timeout_uninterruptible(1); |
576 | } | 582 | } |
diff --git a/mm/page_alloc.c b/mm/page_alloc.c index 7bf22e045318..5675b3073854 100644 --- a/mm/page_alloc.c +++ b/mm/page_alloc.c | |||
@@ -3523,10 +3523,10 @@ static void __paginginit free_area_init_core(struct pglist_data *pgdat, | |||
3523 | INIT_LIST_HEAD(&zone->lru[l].list); | 3523 | INIT_LIST_HEAD(&zone->lru[l].list); |
3524 | zone->lru[l].nr_scan = 0; | 3524 | zone->lru[l].nr_scan = 0; |
3525 | } | 3525 | } |
3526 | zone->recent_rotated[0] = 0; | 3526 | zone->reclaim_stat.recent_rotated[0] = 0; |
3527 | zone->recent_rotated[1] = 0; | 3527 | zone->reclaim_stat.recent_rotated[1] = 0; |
3528 | zone->recent_scanned[0] = 0; | 3528 | zone->reclaim_stat.recent_scanned[0] = 0; |
3529 | zone->recent_scanned[1] = 0; | 3529 | zone->reclaim_stat.recent_scanned[1] = 0; |
3530 | zap_zone_vm_stats(zone); | 3530 | zap_zone_vm_stats(zone); |
3531 | zone->flags = 0; | 3531 | zone->flags = 0; |
3532 | if (!size) | 3532 | if (!size) |
diff --git a/mm/page_cgroup.c b/mm/page_cgroup.c index d6507a660ed6..7006a11350c8 100644 --- a/mm/page_cgroup.c +++ b/mm/page_cgroup.c | |||
@@ -8,6 +8,7 @@ | |||
8 | #include <linux/memory.h> | 8 | #include <linux/memory.h> |
9 | #include <linux/vmalloc.h> | 9 | #include <linux/vmalloc.h> |
10 | #include <linux/cgroup.h> | 10 | #include <linux/cgroup.h> |
11 | #include <linux/swapops.h> | ||
11 | 12 | ||
12 | static void __meminit | 13 | static void __meminit |
13 | __init_page_cgroup(struct page_cgroup *pc, unsigned long pfn) | 14 | __init_page_cgroup(struct page_cgroup *pc, unsigned long pfn) |
@@ -15,6 +16,7 @@ __init_page_cgroup(struct page_cgroup *pc, unsigned long pfn) | |||
15 | pc->flags = 0; | 16 | pc->flags = 0; |
16 | pc->mem_cgroup = NULL; | 17 | pc->mem_cgroup = NULL; |
17 | pc->page = pfn_to_page(pfn); | 18 | pc->page = pfn_to_page(pfn); |
19 | INIT_LIST_HEAD(&pc->lru); | ||
18 | } | 20 | } |
19 | static unsigned long total_usage; | 21 | static unsigned long total_usage; |
20 | 22 | ||
@@ -72,7 +74,7 @@ void __init page_cgroup_init(void) | |||
72 | 74 | ||
73 | int nid, fail; | 75 | int nid, fail; |
74 | 76 | ||
75 | if (mem_cgroup_subsys.disabled) | 77 | if (mem_cgroup_disabled()) |
76 | return; | 78 | return; |
77 | 79 | ||
78 | for_each_online_node(nid) { | 80 | for_each_online_node(nid) { |
@@ -103,13 +105,11 @@ struct page_cgroup *lookup_page_cgroup(struct page *page) | |||
103 | /* __alloc_bootmem...() is protected by !slab_available() */ | 105 | /* __alloc_bootmem...() is protected by !slab_available() */ |
104 | static int __init_refok init_section_page_cgroup(unsigned long pfn) | 106 | static int __init_refok init_section_page_cgroup(unsigned long pfn) |
105 | { | 107 | { |
106 | struct mem_section *section; | 108 | struct mem_section *section = __pfn_to_section(pfn); |
107 | struct page_cgroup *base, *pc; | 109 | struct page_cgroup *base, *pc; |
108 | unsigned long table_size; | 110 | unsigned long table_size; |
109 | int nid, index; | 111 | int nid, index; |
110 | 112 | ||
111 | section = __pfn_to_section(pfn); | ||
112 | |||
113 | if (!section->page_cgroup) { | 113 | if (!section->page_cgroup) { |
114 | nid = page_to_nid(pfn_to_page(pfn)); | 114 | nid = page_to_nid(pfn_to_page(pfn)); |
115 | table_size = sizeof(struct page_cgroup) * PAGES_PER_SECTION; | 115 | table_size = sizeof(struct page_cgroup) * PAGES_PER_SECTION; |
@@ -145,7 +145,6 @@ static int __init_refok init_section_page_cgroup(unsigned long pfn) | |||
145 | __init_page_cgroup(pc, pfn + index); | 145 | __init_page_cgroup(pc, pfn + index); |
146 | } | 146 | } |
147 | 147 | ||
148 | section = __pfn_to_section(pfn); | ||
149 | section->page_cgroup = base - pfn; | 148 | section->page_cgroup = base - pfn; |
150 | total_usage += table_size; | 149 | total_usage += table_size; |
151 | return 0; | 150 | return 0; |
@@ -248,7 +247,7 @@ void __init page_cgroup_init(void) | |||
248 | unsigned long pfn; | 247 | unsigned long pfn; |
249 | int fail = 0; | 248 | int fail = 0; |
250 | 249 | ||
251 | if (mem_cgroup_subsys.disabled) | 250 | if (mem_cgroup_disabled()) |
252 | return; | 251 | return; |
253 | 252 | ||
254 | for (pfn = 0; !fail && pfn < max_pfn; pfn += PAGES_PER_SECTION) { | 253 | for (pfn = 0; !fail && pfn < max_pfn; pfn += PAGES_PER_SECTION) { |
@@ -273,3 +272,199 @@ void __meminit pgdat_page_cgroup_init(struct pglist_data *pgdat) | |||
273 | } | 272 | } |
274 | 273 | ||
275 | #endif | 274 | #endif |
275 | |||
276 | |||
277 | #ifdef CONFIG_CGROUP_MEM_RES_CTLR_SWAP | ||
278 | |||
279 | static DEFINE_MUTEX(swap_cgroup_mutex); | ||
280 | struct swap_cgroup_ctrl { | ||
281 | struct page **map; | ||
282 | unsigned long length; | ||
283 | }; | ||
284 | |||
285 | struct swap_cgroup_ctrl swap_cgroup_ctrl[MAX_SWAPFILES]; | ||
286 | |||
287 | /* | ||
288 | * This 8bytes seems big..maybe we can reduce this when we can use "id" for | ||
289 | * cgroup rather than pointer. | ||
290 | */ | ||
291 | struct swap_cgroup { | ||
292 | struct mem_cgroup *val; | ||
293 | }; | ||
294 | #define SC_PER_PAGE (PAGE_SIZE/sizeof(struct swap_cgroup)) | ||
295 | #define SC_POS_MASK (SC_PER_PAGE - 1) | ||
296 | |||
297 | /* | ||
298 | * SwapCgroup implements "lookup" and "exchange" operations. | ||
299 | * In typical usage, this swap_cgroup is accessed via memcg's charge/uncharge | ||
300 | * against SwapCache. At swap_free(), this is accessed directly from swap. | ||
301 | * | ||
302 | * This means, | ||
303 | * - we have no race in "exchange" when we're accessed via SwapCache because | ||
304 | * SwapCache(and its swp_entry) is under lock. | ||
305 | * - When called via swap_free(), there is no user of this entry and no race. | ||
306 | * Then, we don't need lock around "exchange". | ||
307 | * | ||
308 | * TODO: we can push these buffers out to HIGHMEM. | ||
309 | */ | ||
310 | |||
311 | /* | ||
312 | * allocate buffer for swap_cgroup. | ||
313 | */ | ||
314 | static int swap_cgroup_prepare(int type) | ||
315 | { | ||
316 | struct page *page; | ||
317 | struct swap_cgroup_ctrl *ctrl; | ||
318 | unsigned long idx, max; | ||
319 | |||
320 | if (!do_swap_account) | ||
321 | return 0; | ||
322 | ctrl = &swap_cgroup_ctrl[type]; | ||
323 | |||
324 | for (idx = 0; idx < ctrl->length; idx++) { | ||
325 | page = alloc_page(GFP_KERNEL | __GFP_ZERO); | ||
326 | if (!page) | ||
327 | goto not_enough_page; | ||
328 | ctrl->map[idx] = page; | ||
329 | } | ||
330 | return 0; | ||
331 | not_enough_page: | ||
332 | max = idx; | ||
333 | for (idx = 0; idx < max; idx++) | ||
334 | __free_page(ctrl->map[idx]); | ||
335 | |||
336 | return -ENOMEM; | ||
337 | } | ||
338 | |||
339 | /** | ||
340 | * swap_cgroup_record - record mem_cgroup for this swp_entry. | ||
341 | * @ent: swap entry to be recorded into | ||
342 | * @mem: mem_cgroup to be recorded | ||
343 | * | ||
344 | * Returns old value at success, NULL at failure. | ||
345 | * (Of course, old value can be NULL.) | ||
346 | */ | ||
347 | struct mem_cgroup *swap_cgroup_record(swp_entry_t ent, struct mem_cgroup *mem) | ||
348 | { | ||
349 | int type = swp_type(ent); | ||
350 | unsigned long offset = swp_offset(ent); | ||
351 | unsigned long idx = offset / SC_PER_PAGE; | ||
352 | unsigned long pos = offset & SC_POS_MASK; | ||
353 | struct swap_cgroup_ctrl *ctrl; | ||
354 | struct page *mappage; | ||
355 | struct swap_cgroup *sc; | ||
356 | struct mem_cgroup *old; | ||
357 | |||
358 | if (!do_swap_account) | ||
359 | return NULL; | ||
360 | |||
361 | ctrl = &swap_cgroup_ctrl[type]; | ||
362 | |||
363 | mappage = ctrl->map[idx]; | ||
364 | sc = page_address(mappage); | ||
365 | sc += pos; | ||
366 | old = sc->val; | ||
367 | sc->val = mem; | ||
368 | |||
369 | return old; | ||
370 | } | ||
371 | |||
372 | /** | ||
373 | * lookup_swap_cgroup - lookup mem_cgroup tied to swap entry | ||
374 | * @ent: swap entry to be looked up. | ||
375 | * | ||
376 | * Returns pointer to mem_cgroup at success. NULL at failure. | ||
377 | */ | ||
378 | struct mem_cgroup *lookup_swap_cgroup(swp_entry_t ent) | ||
379 | { | ||
380 | int type = swp_type(ent); | ||
381 | unsigned long offset = swp_offset(ent); | ||
382 | unsigned long idx = offset / SC_PER_PAGE; | ||
383 | unsigned long pos = offset & SC_POS_MASK; | ||
384 | struct swap_cgroup_ctrl *ctrl; | ||
385 | struct page *mappage; | ||
386 | struct swap_cgroup *sc; | ||
387 | struct mem_cgroup *ret; | ||
388 | |||
389 | if (!do_swap_account) | ||
390 | return NULL; | ||
391 | |||
392 | ctrl = &swap_cgroup_ctrl[type]; | ||
393 | mappage = ctrl->map[idx]; | ||
394 | sc = page_address(mappage); | ||
395 | sc += pos; | ||
396 | ret = sc->val; | ||
397 | return ret; | ||
398 | } | ||
399 | |||
400 | int swap_cgroup_swapon(int type, unsigned long max_pages) | ||
401 | { | ||
402 | void *array; | ||
403 | unsigned long array_size; | ||
404 | unsigned long length; | ||
405 | struct swap_cgroup_ctrl *ctrl; | ||
406 | |||
407 | if (!do_swap_account) | ||
408 | return 0; | ||
409 | |||
410 | length = ((max_pages/SC_PER_PAGE) + 1); | ||
411 | array_size = length * sizeof(void *); | ||
412 | |||
413 | array = vmalloc(array_size); | ||
414 | if (!array) | ||
415 | goto nomem; | ||
416 | |||
417 | memset(array, 0, array_size); | ||
418 | ctrl = &swap_cgroup_ctrl[type]; | ||
419 | mutex_lock(&swap_cgroup_mutex); | ||
420 | ctrl->length = length; | ||
421 | ctrl->map = array; | ||
422 | if (swap_cgroup_prepare(type)) { | ||
423 | /* memory shortage */ | ||
424 | ctrl->map = NULL; | ||
425 | ctrl->length = 0; | ||
426 | vfree(array); | ||
427 | mutex_unlock(&swap_cgroup_mutex); | ||
428 | goto nomem; | ||
429 | } | ||
430 | mutex_unlock(&swap_cgroup_mutex); | ||
431 | |||
432 | printk(KERN_INFO | ||
433 | "swap_cgroup: uses %ld bytes of vmalloc for pointer array space" | ||
434 | " and %ld bytes to hold mem_cgroup pointers on swap\n", | ||
435 | array_size, length * PAGE_SIZE); | ||
436 | printk(KERN_INFO | ||
437 | "swap_cgroup can be disabled by noswapaccount boot option.\n"); | ||
438 | |||
439 | return 0; | ||
440 | nomem: | ||
441 | printk(KERN_INFO "couldn't allocate enough memory for swap_cgroup.\n"); | ||
442 | printk(KERN_INFO | ||
443 | "swap_cgroup can be disabled by noswapaccount boot option\n"); | ||
444 | return -ENOMEM; | ||
445 | } | ||
446 | |||
447 | void swap_cgroup_swapoff(int type) | ||
448 | { | ||
449 | int i; | ||
450 | struct swap_cgroup_ctrl *ctrl; | ||
451 | |||
452 | if (!do_swap_account) | ||
453 | return; | ||
454 | |||
455 | mutex_lock(&swap_cgroup_mutex); | ||
456 | ctrl = &swap_cgroup_ctrl[type]; | ||
457 | if (ctrl->map) { | ||
458 | for (i = 0; i < ctrl->length; i++) { | ||
459 | struct page *page = ctrl->map[i]; | ||
460 | if (page) | ||
461 | __free_page(page); | ||
462 | } | ||
463 | vfree(ctrl->map); | ||
464 | ctrl->map = NULL; | ||
465 | ctrl->length = 0; | ||
466 | } | ||
467 | mutex_unlock(&swap_cgroup_mutex); | ||
468 | } | ||
469 | |||
470 | #endif | ||
diff --git a/mm/shmem.c b/mm/shmem.c index 5941f9801363..5d0de96c9789 100644 --- a/mm/shmem.c +++ b/mm/shmem.c | |||
@@ -928,7 +928,11 @@ found: | |||
928 | error = 1; | 928 | error = 1; |
929 | if (!inode) | 929 | if (!inode) |
930 | goto out; | 930 | goto out; |
931 | /* Precharge page using GFP_KERNEL while we can wait */ | 931 | /* |
932 | * Charge page using GFP_KERNEL while we can wait. | ||
933 | * Charged back to the user(not to caller) when swap account is used. | ||
934 | * add_to_page_cache() will be called with GFP_NOWAIT. | ||
935 | */ | ||
932 | error = mem_cgroup_cache_charge(page, current->mm, GFP_KERNEL); | 936 | error = mem_cgroup_cache_charge(page, current->mm, GFP_KERNEL); |
933 | if (error) | 937 | if (error) |
934 | goto out; | 938 | goto out; |
@@ -1320,15 +1324,19 @@ repeat: | |||
1320 | } else { | 1324 | } else { |
1321 | shmem_swp_unmap(entry); | 1325 | shmem_swp_unmap(entry); |
1322 | spin_unlock(&info->lock); | 1326 | spin_unlock(&info->lock); |
1323 | unlock_page(swappage); | ||
1324 | page_cache_release(swappage); | ||
1325 | if (error == -ENOMEM) { | 1327 | if (error == -ENOMEM) { |
1326 | /* allow reclaim from this memory cgroup */ | 1328 | /* allow reclaim from this memory cgroup */ |
1327 | error = mem_cgroup_shrink_usage(current->mm, | 1329 | error = mem_cgroup_shrink_usage(swappage, |
1330 | current->mm, | ||
1328 | gfp); | 1331 | gfp); |
1329 | if (error) | 1332 | if (error) { |
1333 | unlock_page(swappage); | ||
1334 | page_cache_release(swappage); | ||
1330 | goto failed; | 1335 | goto failed; |
1336 | } | ||
1331 | } | 1337 | } |
1338 | unlock_page(swappage); | ||
1339 | page_cache_release(swappage); | ||
1332 | goto repeat; | 1340 | goto repeat; |
1333 | } | 1341 | } |
1334 | } else if (sgp == SGP_READ && !filepage) { | 1342 | } else if (sgp == SGP_READ && !filepage) { |
@@ -1379,7 +1387,7 @@ repeat: | |||
1379 | 1387 | ||
1380 | /* Precharge page while we can wait, compensate after */ | 1388 | /* Precharge page while we can wait, compensate after */ |
1381 | error = mem_cgroup_cache_charge(filepage, current->mm, | 1389 | error = mem_cgroup_cache_charge(filepage, current->mm, |
1382 | gfp & ~__GFP_HIGHMEM); | 1390 | GFP_KERNEL); |
1383 | if (error) { | 1391 | if (error) { |
1384 | page_cache_release(filepage); | 1392 | page_cache_release(filepage); |
1385 | shmem_unacct_blocks(info->flags, 1); | 1393 | shmem_unacct_blocks(info->flags, 1); |
@@ -151,6 +151,26 @@ void rotate_reclaimable_page(struct page *page) | |||
151 | } | 151 | } |
152 | } | 152 | } |
153 | 153 | ||
154 | static void update_page_reclaim_stat(struct zone *zone, struct page *page, | ||
155 | int file, int rotated) | ||
156 | { | ||
157 | struct zone_reclaim_stat *reclaim_stat = &zone->reclaim_stat; | ||
158 | struct zone_reclaim_stat *memcg_reclaim_stat; | ||
159 | |||
160 | memcg_reclaim_stat = mem_cgroup_get_reclaim_stat_from_page(page); | ||
161 | |||
162 | reclaim_stat->recent_scanned[file]++; | ||
163 | if (rotated) | ||
164 | reclaim_stat->recent_rotated[file]++; | ||
165 | |||
166 | if (!memcg_reclaim_stat) | ||
167 | return; | ||
168 | |||
169 | memcg_reclaim_stat->recent_scanned[file]++; | ||
170 | if (rotated) | ||
171 | memcg_reclaim_stat->recent_rotated[file]++; | ||
172 | } | ||
173 | |||
154 | /* | 174 | /* |
155 | * FIXME: speed this up? | 175 | * FIXME: speed this up? |
156 | */ | 176 | */ |
@@ -168,10 +188,8 @@ void activate_page(struct page *page) | |||
168 | lru += LRU_ACTIVE; | 188 | lru += LRU_ACTIVE; |
169 | add_page_to_lru_list(zone, page, lru); | 189 | add_page_to_lru_list(zone, page, lru); |
170 | __count_vm_event(PGACTIVATE); | 190 | __count_vm_event(PGACTIVATE); |
171 | mem_cgroup_move_lists(page, lru); | ||
172 | 191 | ||
173 | zone->recent_rotated[!!file]++; | 192 | update_page_reclaim_stat(zone, page, !!file, 1); |
174 | zone->recent_scanned[!!file]++; | ||
175 | } | 193 | } |
176 | spin_unlock_irq(&zone->lru_lock); | 194 | spin_unlock_irq(&zone->lru_lock); |
177 | } | 195 | } |
@@ -386,12 +404,14 @@ void ____pagevec_lru_add(struct pagevec *pvec, enum lru_list lru) | |||
386 | { | 404 | { |
387 | int i; | 405 | int i; |
388 | struct zone *zone = NULL; | 406 | struct zone *zone = NULL; |
407 | |||
389 | VM_BUG_ON(is_unevictable_lru(lru)); | 408 | VM_BUG_ON(is_unevictable_lru(lru)); |
390 | 409 | ||
391 | for (i = 0; i < pagevec_count(pvec); i++) { | 410 | for (i = 0; i < pagevec_count(pvec); i++) { |
392 | struct page *page = pvec->pages[i]; | 411 | struct page *page = pvec->pages[i]; |
393 | struct zone *pagezone = page_zone(page); | 412 | struct zone *pagezone = page_zone(page); |
394 | int file; | 413 | int file; |
414 | int active; | ||
395 | 415 | ||
396 | if (pagezone != zone) { | 416 | if (pagezone != zone) { |
397 | if (zone) | 417 | if (zone) |
@@ -403,12 +423,11 @@ void ____pagevec_lru_add(struct pagevec *pvec, enum lru_list lru) | |||
403 | VM_BUG_ON(PageUnevictable(page)); | 423 | VM_BUG_ON(PageUnevictable(page)); |
404 | VM_BUG_ON(PageLRU(page)); | 424 | VM_BUG_ON(PageLRU(page)); |
405 | SetPageLRU(page); | 425 | SetPageLRU(page); |
426 | active = is_active_lru(lru); | ||
406 | file = is_file_lru(lru); | 427 | file = is_file_lru(lru); |
407 | zone->recent_scanned[file]++; | 428 | if (active) |
408 | if (is_active_lru(lru)) { | ||
409 | SetPageActive(page); | 429 | SetPageActive(page); |
410 | zone->recent_rotated[file]++; | 430 | update_page_reclaim_stat(zone, page, file, active); |
411 | } | ||
412 | add_page_to_lru_list(zone, page, lru); | 431 | add_page_to_lru_list(zone, page, lru); |
413 | } | 432 | } |
414 | if (zone) | 433 | if (zone) |
diff --git a/mm/swap_state.c b/mm/swap_state.c index 81c825f67a7f..3ecea98ecb45 100644 --- a/mm/swap_state.c +++ b/mm/swap_state.c | |||
@@ -17,6 +17,7 @@ | |||
17 | #include <linux/backing-dev.h> | 17 | #include <linux/backing-dev.h> |
18 | #include <linux/pagevec.h> | 18 | #include <linux/pagevec.h> |
19 | #include <linux/migrate.h> | 19 | #include <linux/migrate.h> |
20 | #include <linux/page_cgroup.h> | ||
20 | 21 | ||
21 | #include <asm/pgtable.h> | 22 | #include <asm/pgtable.h> |
22 | 23 | ||
@@ -108,6 +109,8 @@ int add_to_swap_cache(struct page *page, swp_entry_t entry, gfp_t gfp_mask) | |||
108 | */ | 109 | */ |
109 | void __delete_from_swap_cache(struct page *page) | 110 | void __delete_from_swap_cache(struct page *page) |
110 | { | 111 | { |
112 | swp_entry_t ent = {.val = page_private(page)}; | ||
113 | |||
111 | VM_BUG_ON(!PageLocked(page)); | 114 | VM_BUG_ON(!PageLocked(page)); |
112 | VM_BUG_ON(!PageSwapCache(page)); | 115 | VM_BUG_ON(!PageSwapCache(page)); |
113 | VM_BUG_ON(PageWriteback(page)); | 116 | VM_BUG_ON(PageWriteback(page)); |
@@ -118,6 +121,7 @@ void __delete_from_swap_cache(struct page *page) | |||
118 | total_swapcache_pages--; | 121 | total_swapcache_pages--; |
119 | __dec_zone_page_state(page, NR_FILE_PAGES); | 122 | __dec_zone_page_state(page, NR_FILE_PAGES); |
120 | INC_CACHE_INFO(del_total); | 123 | INC_CACHE_INFO(del_total); |
124 | mem_cgroup_uncharge_swapcache(page, ent); | ||
121 | } | 125 | } |
122 | 126 | ||
123 | /** | 127 | /** |
diff --git a/mm/swapfile.c b/mm/swapfile.c index eec5ca758a23..da422c47e2ee 100644 --- a/mm/swapfile.c +++ b/mm/swapfile.c | |||
@@ -33,6 +33,7 @@ | |||
33 | #include <asm/pgtable.h> | 33 | #include <asm/pgtable.h> |
34 | #include <asm/tlbflush.h> | 34 | #include <asm/tlbflush.h> |
35 | #include <linux/swapops.h> | 35 | #include <linux/swapops.h> |
36 | #include <linux/page_cgroup.h> | ||
36 | 37 | ||
37 | static DEFINE_SPINLOCK(swap_lock); | 38 | static DEFINE_SPINLOCK(swap_lock); |
38 | static unsigned int nr_swapfiles; | 39 | static unsigned int nr_swapfiles; |
@@ -470,8 +471,9 @@ out: | |||
470 | return NULL; | 471 | return NULL; |
471 | } | 472 | } |
472 | 473 | ||
473 | static int swap_entry_free(struct swap_info_struct *p, unsigned long offset) | 474 | static int swap_entry_free(struct swap_info_struct *p, swp_entry_t ent) |
474 | { | 475 | { |
476 | unsigned long offset = swp_offset(ent); | ||
475 | int count = p->swap_map[offset]; | 477 | int count = p->swap_map[offset]; |
476 | 478 | ||
477 | if (count < SWAP_MAP_MAX) { | 479 | if (count < SWAP_MAP_MAX) { |
@@ -486,6 +488,7 @@ static int swap_entry_free(struct swap_info_struct *p, unsigned long offset) | |||
486 | swap_list.next = p - swap_info; | 488 | swap_list.next = p - swap_info; |
487 | nr_swap_pages++; | 489 | nr_swap_pages++; |
488 | p->inuse_pages--; | 490 | p->inuse_pages--; |
491 | mem_cgroup_uncharge_swap(ent); | ||
489 | } | 492 | } |
490 | } | 493 | } |
491 | return count; | 494 | return count; |
@@ -501,7 +504,7 @@ void swap_free(swp_entry_t entry) | |||
501 | 504 | ||
502 | p = swap_info_get(entry); | 505 | p = swap_info_get(entry); |
503 | if (p) { | 506 | if (p) { |
504 | swap_entry_free(p, swp_offset(entry)); | 507 | swap_entry_free(p, entry); |
505 | spin_unlock(&swap_lock); | 508 | spin_unlock(&swap_lock); |
506 | } | 509 | } |
507 | } | 510 | } |
@@ -581,7 +584,7 @@ int free_swap_and_cache(swp_entry_t entry) | |||
581 | 584 | ||
582 | p = swap_info_get(entry); | 585 | p = swap_info_get(entry); |
583 | if (p) { | 586 | if (p) { |
584 | if (swap_entry_free(p, swp_offset(entry)) == 1) { | 587 | if (swap_entry_free(p, entry) == 1) { |
585 | page = find_get_page(&swapper_space, entry.val); | 588 | page = find_get_page(&swapper_space, entry.val); |
586 | if (page && !trylock_page(page)) { | 589 | if (page && !trylock_page(page)) { |
587 | page_cache_release(page); | 590 | page_cache_release(page); |
@@ -690,17 +693,18 @@ unsigned int count_swap_pages(int type, int free) | |||
690 | static int unuse_pte(struct vm_area_struct *vma, pmd_t *pmd, | 693 | static int unuse_pte(struct vm_area_struct *vma, pmd_t *pmd, |
691 | unsigned long addr, swp_entry_t entry, struct page *page) | 694 | unsigned long addr, swp_entry_t entry, struct page *page) |
692 | { | 695 | { |
696 | struct mem_cgroup *ptr = NULL; | ||
693 | spinlock_t *ptl; | 697 | spinlock_t *ptl; |
694 | pte_t *pte; | 698 | pte_t *pte; |
695 | int ret = 1; | 699 | int ret = 1; |
696 | 700 | ||
697 | if (mem_cgroup_charge(page, vma->vm_mm, GFP_KERNEL)) | 701 | if (mem_cgroup_try_charge_swapin(vma->vm_mm, page, GFP_KERNEL, &ptr)) |
698 | ret = -ENOMEM; | 702 | ret = -ENOMEM; |
699 | 703 | ||
700 | pte = pte_offset_map_lock(vma->vm_mm, pmd, addr, &ptl); | 704 | pte = pte_offset_map_lock(vma->vm_mm, pmd, addr, &ptl); |
701 | if (unlikely(!pte_same(*pte, swp_entry_to_pte(entry)))) { | 705 | if (unlikely(!pte_same(*pte, swp_entry_to_pte(entry)))) { |
702 | if (ret > 0) | 706 | if (ret > 0) |
703 | mem_cgroup_uncharge_page(page); | 707 | mem_cgroup_cancel_charge_swapin(ptr); |
704 | ret = 0; | 708 | ret = 0; |
705 | goto out; | 709 | goto out; |
706 | } | 710 | } |
@@ -710,6 +714,7 @@ static int unuse_pte(struct vm_area_struct *vma, pmd_t *pmd, | |||
710 | set_pte_at(vma->vm_mm, addr, pte, | 714 | set_pte_at(vma->vm_mm, addr, pte, |
711 | pte_mkold(mk_pte(page, vma->vm_page_prot))); | 715 | pte_mkold(mk_pte(page, vma->vm_page_prot))); |
712 | page_add_anon_rmap(page, vma, addr); | 716 | page_add_anon_rmap(page, vma, addr); |
717 | mem_cgroup_commit_charge_swapin(page, ptr); | ||
713 | swap_free(entry); | 718 | swap_free(entry); |
714 | /* | 719 | /* |
715 | * Move the page to the active list so it is not | 720 | * Move the page to the active list so it is not |
@@ -1492,6 +1497,9 @@ asmlinkage long sys_swapoff(const char __user * specialfile) | |||
1492 | spin_unlock(&swap_lock); | 1497 | spin_unlock(&swap_lock); |
1493 | mutex_unlock(&swapon_mutex); | 1498 | mutex_unlock(&swapon_mutex); |
1494 | vfree(swap_map); | 1499 | vfree(swap_map); |
1500 | /* Destroy swap account informatin */ | ||
1501 | swap_cgroup_swapoff(type); | ||
1502 | |||
1495 | inode = mapping->host; | 1503 | inode = mapping->host; |
1496 | if (S_ISBLK(inode->i_mode)) { | 1504 | if (S_ISBLK(inode->i_mode)) { |
1497 | struct block_device *bdev = I_BDEV(inode); | 1505 | struct block_device *bdev = I_BDEV(inode); |
@@ -1809,6 +1817,11 @@ asmlinkage long sys_swapon(const char __user * specialfile, int swap_flags) | |||
1809 | } | 1817 | } |
1810 | swap_map[page_nr] = SWAP_MAP_BAD; | 1818 | swap_map[page_nr] = SWAP_MAP_BAD; |
1811 | } | 1819 | } |
1820 | |||
1821 | error = swap_cgroup_swapon(type, maxpages); | ||
1822 | if (error) | ||
1823 | goto bad_swap; | ||
1824 | |||
1812 | nr_good_pages = swap_header->info.last_page - | 1825 | nr_good_pages = swap_header->info.last_page - |
1813 | swap_header->info.nr_badpages - | 1826 | swap_header->info.nr_badpages - |
1814 | 1 /* header page */; | 1827 | 1 /* header page */; |
@@ -1880,6 +1893,7 @@ bad_swap: | |||
1880 | bd_release(bdev); | 1893 | bd_release(bdev); |
1881 | } | 1894 | } |
1882 | destroy_swap_extents(p); | 1895 | destroy_swap_extents(p); |
1896 | swap_cgroup_swapoff(type); | ||
1883 | bad_swap_2: | 1897 | bad_swap_2: |
1884 | spin_lock(&swap_lock); | 1898 | spin_lock(&swap_lock); |
1885 | p->swap_file = NULL; | 1899 | p->swap_file = NULL; |
diff --git a/mm/vmscan.c b/mm/vmscan.c index b07c48b09a93..9a27c44aa327 100644 --- a/mm/vmscan.c +++ b/mm/vmscan.c | |||
@@ -125,11 +125,30 @@ static LIST_HEAD(shrinker_list); | |||
125 | static DECLARE_RWSEM(shrinker_rwsem); | 125 | static DECLARE_RWSEM(shrinker_rwsem); |
126 | 126 | ||
127 | #ifdef CONFIG_CGROUP_MEM_RES_CTLR | 127 | #ifdef CONFIG_CGROUP_MEM_RES_CTLR |
128 | #define scan_global_lru(sc) (!(sc)->mem_cgroup) | 128 | #define scanning_global_lru(sc) (!(sc)->mem_cgroup) |
129 | #else | 129 | #else |
130 | #define scan_global_lru(sc) (1) | 130 | #define scanning_global_lru(sc) (1) |
131 | #endif | 131 | #endif |
132 | 132 | ||
133 | static struct zone_reclaim_stat *get_reclaim_stat(struct zone *zone, | ||
134 | struct scan_control *sc) | ||
135 | { | ||
136 | if (!scanning_global_lru(sc)) | ||
137 | return mem_cgroup_get_reclaim_stat(sc->mem_cgroup, zone); | ||
138 | |||
139 | return &zone->reclaim_stat; | ||
140 | } | ||
141 | |||
142 | static unsigned long zone_nr_pages(struct zone *zone, struct scan_control *sc, | ||
143 | enum lru_list lru) | ||
144 | { | ||
145 | if (!scanning_global_lru(sc)) | ||
146 | return mem_cgroup_zone_nr_pages(sc->mem_cgroup, zone, lru); | ||
147 | |||
148 | return zone_page_state(zone, NR_LRU_BASE + lru); | ||
149 | } | ||
150 | |||
151 | |||
133 | /* | 152 | /* |
134 | * Add a shrinker callback to be called from the vm | 153 | * Add a shrinker callback to be called from the vm |
135 | */ | 154 | */ |
@@ -512,7 +531,6 @@ redo: | |||
512 | lru = LRU_UNEVICTABLE; | 531 | lru = LRU_UNEVICTABLE; |
513 | add_page_to_unevictable_list(page); | 532 | add_page_to_unevictable_list(page); |
514 | } | 533 | } |
515 | mem_cgroup_move_lists(page, lru); | ||
516 | 534 | ||
517 | /* | 535 | /* |
518 | * page's status can change while we move it among lru. If an evictable | 536 | * page's status can change while we move it among lru. If an evictable |
@@ -547,7 +565,6 @@ void putback_lru_page(struct page *page) | |||
547 | 565 | ||
548 | lru = !!TestClearPageActive(page) + page_is_file_cache(page); | 566 | lru = !!TestClearPageActive(page) + page_is_file_cache(page); |
549 | lru_cache_add_lru(page, lru); | 567 | lru_cache_add_lru(page, lru); |
550 | mem_cgroup_move_lists(page, lru); | ||
551 | put_page(page); | 568 | put_page(page); |
552 | } | 569 | } |
553 | #endif /* CONFIG_UNEVICTABLE_LRU */ | 570 | #endif /* CONFIG_UNEVICTABLE_LRU */ |
@@ -813,6 +830,7 @@ int __isolate_lru_page(struct page *page, int mode, int file) | |||
813 | return ret; | 830 | return ret; |
814 | 831 | ||
815 | ret = -EBUSY; | 832 | ret = -EBUSY; |
833 | |||
816 | if (likely(get_page_unless_zero(page))) { | 834 | if (likely(get_page_unless_zero(page))) { |
817 | /* | 835 | /* |
818 | * Be careful not to clear PageLRU until after we're | 836 | * Be careful not to clear PageLRU until after we're |
@@ -821,6 +839,7 @@ int __isolate_lru_page(struct page *page, int mode, int file) | |||
821 | */ | 839 | */ |
822 | ClearPageLRU(page); | 840 | ClearPageLRU(page); |
823 | ret = 0; | 841 | ret = 0; |
842 | mem_cgroup_del_lru(page); | ||
824 | } | 843 | } |
825 | 844 | ||
826 | return ret; | 845 | return ret; |
@@ -1029,6 +1048,7 @@ static unsigned long shrink_inactive_list(unsigned long max_scan, | |||
1029 | struct pagevec pvec; | 1048 | struct pagevec pvec; |
1030 | unsigned long nr_scanned = 0; | 1049 | unsigned long nr_scanned = 0; |
1031 | unsigned long nr_reclaimed = 0; | 1050 | unsigned long nr_reclaimed = 0; |
1051 | struct zone_reclaim_stat *reclaim_stat = get_reclaim_stat(zone, sc); | ||
1032 | 1052 | ||
1033 | pagevec_init(&pvec, 1); | 1053 | pagevec_init(&pvec, 1); |
1034 | 1054 | ||
@@ -1070,13 +1090,14 @@ static unsigned long shrink_inactive_list(unsigned long max_scan, | |||
1070 | __mod_zone_page_state(zone, NR_INACTIVE_ANON, | 1090 | __mod_zone_page_state(zone, NR_INACTIVE_ANON, |
1071 | -count[LRU_INACTIVE_ANON]); | 1091 | -count[LRU_INACTIVE_ANON]); |
1072 | 1092 | ||
1073 | if (scan_global_lru(sc)) { | 1093 | if (scanning_global_lru(sc)) |
1074 | zone->pages_scanned += nr_scan; | 1094 | zone->pages_scanned += nr_scan; |
1075 | zone->recent_scanned[0] += count[LRU_INACTIVE_ANON]; | 1095 | |
1076 | zone->recent_scanned[0] += count[LRU_ACTIVE_ANON]; | 1096 | reclaim_stat->recent_scanned[0] += count[LRU_INACTIVE_ANON]; |
1077 | zone->recent_scanned[1] += count[LRU_INACTIVE_FILE]; | 1097 | reclaim_stat->recent_scanned[0] += count[LRU_ACTIVE_ANON]; |
1078 | zone->recent_scanned[1] += count[LRU_ACTIVE_FILE]; | 1098 | reclaim_stat->recent_scanned[1] += count[LRU_INACTIVE_FILE]; |
1079 | } | 1099 | reclaim_stat->recent_scanned[1] += count[LRU_ACTIVE_FILE]; |
1100 | |||
1080 | spin_unlock_irq(&zone->lru_lock); | 1101 | spin_unlock_irq(&zone->lru_lock); |
1081 | 1102 | ||
1082 | nr_scanned += nr_scan; | 1103 | nr_scanned += nr_scan; |
@@ -1108,7 +1129,7 @@ static unsigned long shrink_inactive_list(unsigned long max_scan, | |||
1108 | if (current_is_kswapd()) { | 1129 | if (current_is_kswapd()) { |
1109 | __count_zone_vm_events(PGSCAN_KSWAPD, zone, nr_scan); | 1130 | __count_zone_vm_events(PGSCAN_KSWAPD, zone, nr_scan); |
1110 | __count_vm_events(KSWAPD_STEAL, nr_freed); | 1131 | __count_vm_events(KSWAPD_STEAL, nr_freed); |
1111 | } else if (scan_global_lru(sc)) | 1132 | } else if (scanning_global_lru(sc)) |
1112 | __count_zone_vm_events(PGSCAN_DIRECT, zone, nr_scan); | 1133 | __count_zone_vm_events(PGSCAN_DIRECT, zone, nr_scan); |
1113 | 1134 | ||
1114 | __count_zone_vm_events(PGSTEAL, zone, nr_freed); | 1135 | __count_zone_vm_events(PGSTEAL, zone, nr_freed); |
@@ -1134,10 +1155,9 @@ static unsigned long shrink_inactive_list(unsigned long max_scan, | |||
1134 | SetPageLRU(page); | 1155 | SetPageLRU(page); |
1135 | lru = page_lru(page); | 1156 | lru = page_lru(page); |
1136 | add_page_to_lru_list(zone, page, lru); | 1157 | add_page_to_lru_list(zone, page, lru); |
1137 | mem_cgroup_move_lists(page, lru); | 1158 | if (PageActive(page)) { |
1138 | if (PageActive(page) && scan_global_lru(sc)) { | ||
1139 | int file = !!page_is_file_cache(page); | 1159 | int file = !!page_is_file_cache(page); |
1140 | zone->recent_rotated[file]++; | 1160 | reclaim_stat->recent_rotated[file]++; |
1141 | } | 1161 | } |
1142 | if (!pagevec_add(&pvec, page)) { | 1162 | if (!pagevec_add(&pvec, page)) { |
1143 | spin_unlock_irq(&zone->lru_lock); | 1163 | spin_unlock_irq(&zone->lru_lock); |
@@ -1197,6 +1217,7 @@ static void shrink_active_list(unsigned long nr_pages, struct zone *zone, | |||
1197 | struct page *page; | 1217 | struct page *page; |
1198 | struct pagevec pvec; | 1218 | struct pagevec pvec; |
1199 | enum lru_list lru; | 1219 | enum lru_list lru; |
1220 | struct zone_reclaim_stat *reclaim_stat = get_reclaim_stat(zone, sc); | ||
1200 | 1221 | ||
1201 | lru_add_drain(); | 1222 | lru_add_drain(); |
1202 | spin_lock_irq(&zone->lru_lock); | 1223 | spin_lock_irq(&zone->lru_lock); |
@@ -1207,10 +1228,10 @@ static void shrink_active_list(unsigned long nr_pages, struct zone *zone, | |||
1207 | * zone->pages_scanned is used for detect zone's oom | 1228 | * zone->pages_scanned is used for detect zone's oom |
1208 | * mem_cgroup remembers nr_scan by itself. | 1229 | * mem_cgroup remembers nr_scan by itself. |
1209 | */ | 1230 | */ |
1210 | if (scan_global_lru(sc)) { | 1231 | if (scanning_global_lru(sc)) { |
1211 | zone->pages_scanned += pgscanned; | 1232 | zone->pages_scanned += pgscanned; |
1212 | zone->recent_scanned[!!file] += pgmoved; | ||
1213 | } | 1233 | } |
1234 | reclaim_stat->recent_scanned[!!file] += pgmoved; | ||
1214 | 1235 | ||
1215 | if (file) | 1236 | if (file) |
1216 | __mod_zone_page_state(zone, NR_ACTIVE_FILE, -pgmoved); | 1237 | __mod_zone_page_state(zone, NR_ACTIVE_FILE, -pgmoved); |
@@ -1251,8 +1272,7 @@ static void shrink_active_list(unsigned long nr_pages, struct zone *zone, | |||
1251 | * This helps balance scan pressure between file and anonymous | 1272 | * This helps balance scan pressure between file and anonymous |
1252 | * pages in get_scan_ratio. | 1273 | * pages in get_scan_ratio. |
1253 | */ | 1274 | */ |
1254 | if (scan_global_lru(sc)) | 1275 | reclaim_stat->recent_rotated[!!file] += pgmoved; |
1255 | zone->recent_rotated[!!file] += pgmoved; | ||
1256 | 1276 | ||
1257 | while (!list_empty(&l_inactive)) { | 1277 | while (!list_empty(&l_inactive)) { |
1258 | page = lru_to_page(&l_inactive); | 1278 | page = lru_to_page(&l_inactive); |
@@ -1263,7 +1283,7 @@ static void shrink_active_list(unsigned long nr_pages, struct zone *zone, | |||
1263 | ClearPageActive(page); | 1283 | ClearPageActive(page); |
1264 | 1284 | ||
1265 | list_move(&page->lru, &zone->lru[lru].list); | 1285 | list_move(&page->lru, &zone->lru[lru].list); |
1266 | mem_cgroup_move_lists(page, lru); | 1286 | mem_cgroup_add_lru_list(page, lru); |
1267 | pgmoved++; | 1287 | pgmoved++; |
1268 | if (!pagevec_add(&pvec, page)) { | 1288 | if (!pagevec_add(&pvec, page)) { |
1269 | __mod_zone_page_state(zone, NR_LRU_BASE + lru, pgmoved); | 1289 | __mod_zone_page_state(zone, NR_LRU_BASE + lru, pgmoved); |
@@ -1292,6 +1312,38 @@ static void shrink_active_list(unsigned long nr_pages, struct zone *zone, | |||
1292 | pagevec_release(&pvec); | 1312 | pagevec_release(&pvec); |
1293 | } | 1313 | } |
1294 | 1314 | ||
1315 | static int inactive_anon_is_low_global(struct zone *zone) | ||
1316 | { | ||
1317 | unsigned long active, inactive; | ||
1318 | |||
1319 | active = zone_page_state(zone, NR_ACTIVE_ANON); | ||
1320 | inactive = zone_page_state(zone, NR_INACTIVE_ANON); | ||
1321 | |||
1322 | if (inactive * zone->inactive_ratio < active) | ||
1323 | return 1; | ||
1324 | |||
1325 | return 0; | ||
1326 | } | ||
1327 | |||
1328 | /** | ||
1329 | * inactive_anon_is_low - check if anonymous pages need to be deactivated | ||
1330 | * @zone: zone to check | ||
1331 | * @sc: scan control of this context | ||
1332 | * | ||
1333 | * Returns true if the zone does not have enough inactive anon pages, | ||
1334 | * meaning some active anon pages need to be deactivated. | ||
1335 | */ | ||
1336 | static int inactive_anon_is_low(struct zone *zone, struct scan_control *sc) | ||
1337 | { | ||
1338 | int low; | ||
1339 | |||
1340 | if (scanning_global_lru(sc)) | ||
1341 | low = inactive_anon_is_low_global(zone); | ||
1342 | else | ||
1343 | low = mem_cgroup_inactive_anon_is_low(sc->mem_cgroup); | ||
1344 | return low; | ||
1345 | } | ||
1346 | |||
1295 | static unsigned long shrink_list(enum lru_list lru, unsigned long nr_to_scan, | 1347 | static unsigned long shrink_list(enum lru_list lru, unsigned long nr_to_scan, |
1296 | struct zone *zone, struct scan_control *sc, int priority) | 1348 | struct zone *zone, struct scan_control *sc, int priority) |
1297 | { | 1349 | { |
@@ -1302,8 +1354,7 @@ static unsigned long shrink_list(enum lru_list lru, unsigned long nr_to_scan, | |||
1302 | return 0; | 1354 | return 0; |
1303 | } | 1355 | } |
1304 | 1356 | ||
1305 | if (lru == LRU_ACTIVE_ANON && | 1357 | if (lru == LRU_ACTIVE_ANON && inactive_anon_is_low(zone, sc)) { |
1306 | (!scan_global_lru(sc) || inactive_anon_is_low(zone))) { | ||
1307 | shrink_active_list(nr_to_scan, zone, sc, priority, file); | 1358 | shrink_active_list(nr_to_scan, zone, sc, priority, file); |
1308 | return 0; | 1359 | return 0; |
1309 | } | 1360 | } |
@@ -1325,6 +1376,7 @@ static void get_scan_ratio(struct zone *zone, struct scan_control *sc, | |||
1325 | unsigned long anon, file, free; | 1376 | unsigned long anon, file, free; |
1326 | unsigned long anon_prio, file_prio; | 1377 | unsigned long anon_prio, file_prio; |
1327 | unsigned long ap, fp; | 1378 | unsigned long ap, fp; |
1379 | struct zone_reclaim_stat *reclaim_stat = get_reclaim_stat(zone, sc); | ||
1328 | 1380 | ||
1329 | /* If we have no swap space, do not bother scanning anon pages. */ | 1381 | /* If we have no swap space, do not bother scanning anon pages. */ |
1330 | if (nr_swap_pages <= 0) { | 1382 | if (nr_swap_pages <= 0) { |
@@ -1333,17 +1385,20 @@ static void get_scan_ratio(struct zone *zone, struct scan_control *sc, | |||
1333 | return; | 1385 | return; |
1334 | } | 1386 | } |
1335 | 1387 | ||
1336 | anon = zone_page_state(zone, NR_ACTIVE_ANON) + | 1388 | anon = zone_nr_pages(zone, sc, LRU_ACTIVE_ANON) + |
1337 | zone_page_state(zone, NR_INACTIVE_ANON); | 1389 | zone_nr_pages(zone, sc, LRU_INACTIVE_ANON); |
1338 | file = zone_page_state(zone, NR_ACTIVE_FILE) + | 1390 | file = zone_nr_pages(zone, sc, LRU_ACTIVE_FILE) + |
1339 | zone_page_state(zone, NR_INACTIVE_FILE); | 1391 | zone_nr_pages(zone, sc, LRU_INACTIVE_FILE); |
1340 | free = zone_page_state(zone, NR_FREE_PAGES); | 1392 | |
1341 | 1393 | if (scanning_global_lru(sc)) { | |
1342 | /* If we have very few page cache pages, force-scan anon pages. */ | 1394 | free = zone_page_state(zone, NR_FREE_PAGES); |
1343 | if (unlikely(file + free <= zone->pages_high)) { | 1395 | /* If we have very few page cache pages, |
1344 | percent[0] = 100; | 1396 | force-scan anon pages. */ |
1345 | percent[1] = 0; | 1397 | if (unlikely(file + free <= zone->pages_high)) { |
1346 | return; | 1398 | percent[0] = 100; |
1399 | percent[1] = 0; | ||
1400 | return; | ||
1401 | } | ||
1347 | } | 1402 | } |
1348 | 1403 | ||
1349 | /* | 1404 | /* |
@@ -1357,17 +1412,17 @@ static void get_scan_ratio(struct zone *zone, struct scan_control *sc, | |||
1357 | * | 1412 | * |
1358 | * anon in [0], file in [1] | 1413 | * anon in [0], file in [1] |
1359 | */ | 1414 | */ |
1360 | if (unlikely(zone->recent_scanned[0] > anon / 4)) { | 1415 | if (unlikely(reclaim_stat->recent_scanned[0] > anon / 4)) { |
1361 | spin_lock_irq(&zone->lru_lock); | 1416 | spin_lock_irq(&zone->lru_lock); |
1362 | zone->recent_scanned[0] /= 2; | 1417 | reclaim_stat->recent_scanned[0] /= 2; |
1363 | zone->recent_rotated[0] /= 2; | 1418 | reclaim_stat->recent_rotated[0] /= 2; |
1364 | spin_unlock_irq(&zone->lru_lock); | 1419 | spin_unlock_irq(&zone->lru_lock); |
1365 | } | 1420 | } |
1366 | 1421 | ||
1367 | if (unlikely(zone->recent_scanned[1] > file / 4)) { | 1422 | if (unlikely(reclaim_stat->recent_scanned[1] > file / 4)) { |
1368 | spin_lock_irq(&zone->lru_lock); | 1423 | spin_lock_irq(&zone->lru_lock); |
1369 | zone->recent_scanned[1] /= 2; | 1424 | reclaim_stat->recent_scanned[1] /= 2; |
1370 | zone->recent_rotated[1] /= 2; | 1425 | reclaim_stat->recent_rotated[1] /= 2; |
1371 | spin_unlock_irq(&zone->lru_lock); | 1426 | spin_unlock_irq(&zone->lru_lock); |
1372 | } | 1427 | } |
1373 | 1428 | ||
@@ -1383,11 +1438,11 @@ static void get_scan_ratio(struct zone *zone, struct scan_control *sc, | |||
1383 | * proportional to the fraction of recently scanned pages on | 1438 | * proportional to the fraction of recently scanned pages on |
1384 | * each list that were recently referenced and in active use. | 1439 | * each list that were recently referenced and in active use. |
1385 | */ | 1440 | */ |
1386 | ap = (anon_prio + 1) * (zone->recent_scanned[0] + 1); | 1441 | ap = (anon_prio + 1) * (reclaim_stat->recent_scanned[0] + 1); |
1387 | ap /= zone->recent_rotated[0] + 1; | 1442 | ap /= reclaim_stat->recent_rotated[0] + 1; |
1388 | 1443 | ||
1389 | fp = (file_prio + 1) * (zone->recent_scanned[1] + 1); | 1444 | fp = (file_prio + 1) * (reclaim_stat->recent_scanned[1] + 1); |
1390 | fp /= zone->recent_rotated[1] + 1; | 1445 | fp /= reclaim_stat->recent_rotated[1] + 1; |
1391 | 1446 | ||
1392 | /* Normalize to percentages */ | 1447 | /* Normalize to percentages */ |
1393 | percent[0] = 100 * ap / (ap + fp + 1); | 1448 | percent[0] = 100 * ap / (ap + fp + 1); |
@@ -1411,30 +1466,23 @@ static void shrink_zone(int priority, struct zone *zone, | |||
1411 | get_scan_ratio(zone, sc, percent); | 1466 | get_scan_ratio(zone, sc, percent); |
1412 | 1467 | ||
1413 | for_each_evictable_lru(l) { | 1468 | for_each_evictable_lru(l) { |
1414 | if (scan_global_lru(sc)) { | 1469 | int file = is_file_lru(l); |
1415 | int file = is_file_lru(l); | 1470 | int scan; |
1416 | int scan; | 1471 | |
1417 | 1472 | scan = zone_page_state(zone, NR_LRU_BASE + l); | |
1418 | scan = zone_page_state(zone, NR_LRU_BASE + l); | 1473 | if (priority) { |
1419 | if (priority) { | 1474 | scan >>= priority; |
1420 | scan >>= priority; | 1475 | scan = (scan * percent[file]) / 100; |
1421 | scan = (scan * percent[file]) / 100; | 1476 | } |
1422 | } | 1477 | if (scanning_global_lru(sc)) { |
1423 | zone->lru[l].nr_scan += scan; | 1478 | zone->lru[l].nr_scan += scan; |
1424 | nr[l] = zone->lru[l].nr_scan; | 1479 | nr[l] = zone->lru[l].nr_scan; |
1425 | if (nr[l] >= swap_cluster_max) | 1480 | if (nr[l] >= swap_cluster_max) |
1426 | zone->lru[l].nr_scan = 0; | 1481 | zone->lru[l].nr_scan = 0; |
1427 | else | 1482 | else |
1428 | nr[l] = 0; | 1483 | nr[l] = 0; |
1429 | } else { | 1484 | } else |
1430 | /* | 1485 | nr[l] = scan; |
1431 | * This reclaim occurs not because zone memory shortage | ||
1432 | * but because memory controller hits its limit. | ||
1433 | * Don't modify zone reclaim related data. | ||
1434 | */ | ||
1435 | nr[l] = mem_cgroup_calc_reclaim(sc->mem_cgroup, zone, | ||
1436 | priority, l); | ||
1437 | } | ||
1438 | } | 1486 | } |
1439 | 1487 | ||
1440 | while (nr[LRU_INACTIVE_ANON] || nr[LRU_ACTIVE_FILE] || | 1488 | while (nr[LRU_INACTIVE_ANON] || nr[LRU_ACTIVE_FILE] || |
@@ -1467,9 +1515,7 @@ static void shrink_zone(int priority, struct zone *zone, | |||
1467 | * Even if we did not try to evict anon pages at all, we want to | 1515 | * Even if we did not try to evict anon pages at all, we want to |
1468 | * rebalance the anon lru active/inactive ratio. | 1516 | * rebalance the anon lru active/inactive ratio. |
1469 | */ | 1517 | */ |
1470 | if (!scan_global_lru(sc) || inactive_anon_is_low(zone)) | 1518 | if (inactive_anon_is_low(zone, sc)) |
1471 | shrink_active_list(SWAP_CLUSTER_MAX, zone, sc, priority, 0); | ||
1472 | else if (!scan_global_lru(sc)) | ||
1473 | shrink_active_list(SWAP_CLUSTER_MAX, zone, sc, priority, 0); | 1519 | shrink_active_list(SWAP_CLUSTER_MAX, zone, sc, priority, 0); |
1474 | 1520 | ||
1475 | throttle_vm_writeout(sc->gfp_mask); | 1521 | throttle_vm_writeout(sc->gfp_mask); |
@@ -1504,7 +1550,7 @@ static void shrink_zones(int priority, struct zonelist *zonelist, | |||
1504 | * Take care memory controller reclaiming has small influence | 1550 | * Take care memory controller reclaiming has small influence |
1505 | * to global LRU. | 1551 | * to global LRU. |
1506 | */ | 1552 | */ |
1507 | if (scan_global_lru(sc)) { | 1553 | if (scanning_global_lru(sc)) { |
1508 | if (!cpuset_zone_allowed_hardwall(zone, GFP_KERNEL)) | 1554 | if (!cpuset_zone_allowed_hardwall(zone, GFP_KERNEL)) |
1509 | continue; | 1555 | continue; |
1510 | note_zone_scanning_priority(zone, priority); | 1556 | note_zone_scanning_priority(zone, priority); |
@@ -1557,12 +1603,12 @@ static unsigned long do_try_to_free_pages(struct zonelist *zonelist, | |||
1557 | 1603 | ||
1558 | delayacct_freepages_start(); | 1604 | delayacct_freepages_start(); |
1559 | 1605 | ||
1560 | if (scan_global_lru(sc)) | 1606 | if (scanning_global_lru(sc)) |
1561 | count_vm_event(ALLOCSTALL); | 1607 | count_vm_event(ALLOCSTALL); |
1562 | /* | 1608 | /* |
1563 | * mem_cgroup will not do shrink_slab. | 1609 | * mem_cgroup will not do shrink_slab. |
1564 | */ | 1610 | */ |
1565 | if (scan_global_lru(sc)) { | 1611 | if (scanning_global_lru(sc)) { |
1566 | for_each_zone_zonelist(zone, z, zonelist, high_zoneidx) { | 1612 | for_each_zone_zonelist(zone, z, zonelist, high_zoneidx) { |
1567 | 1613 | ||
1568 | if (!cpuset_zone_allowed_hardwall(zone, GFP_KERNEL)) | 1614 | if (!cpuset_zone_allowed_hardwall(zone, GFP_KERNEL)) |
@@ -1581,7 +1627,7 @@ static unsigned long do_try_to_free_pages(struct zonelist *zonelist, | |||
1581 | * Don't shrink slabs when reclaiming memory from | 1627 | * Don't shrink slabs when reclaiming memory from |
1582 | * over limit cgroups | 1628 | * over limit cgroups |
1583 | */ | 1629 | */ |
1584 | if (scan_global_lru(sc)) { | 1630 | if (scanning_global_lru(sc)) { |
1585 | shrink_slab(sc->nr_scanned, sc->gfp_mask, lru_pages); | 1631 | shrink_slab(sc->nr_scanned, sc->gfp_mask, lru_pages); |
1586 | if (reclaim_state) { | 1632 | if (reclaim_state) { |
1587 | sc->nr_reclaimed += reclaim_state->reclaimed_slab; | 1633 | sc->nr_reclaimed += reclaim_state->reclaimed_slab; |
@@ -1612,7 +1658,7 @@ static unsigned long do_try_to_free_pages(struct zonelist *zonelist, | |||
1612 | congestion_wait(WRITE, HZ/10); | 1658 | congestion_wait(WRITE, HZ/10); |
1613 | } | 1659 | } |
1614 | /* top priority shrink_zones still had more to do? don't OOM, then */ | 1660 | /* top priority shrink_zones still had more to do? don't OOM, then */ |
1615 | if (!sc->all_unreclaimable && scan_global_lru(sc)) | 1661 | if (!sc->all_unreclaimable && scanning_global_lru(sc)) |
1616 | ret = sc->nr_reclaimed; | 1662 | ret = sc->nr_reclaimed; |
1617 | out: | 1663 | out: |
1618 | /* | 1664 | /* |
@@ -1625,7 +1671,7 @@ out: | |||
1625 | if (priority < 0) | 1671 | if (priority < 0) |
1626 | priority = 0; | 1672 | priority = 0; |
1627 | 1673 | ||
1628 | if (scan_global_lru(sc)) { | 1674 | if (scanning_global_lru(sc)) { |
1629 | for_each_zone_zonelist(zone, z, zonelist, high_zoneidx) { | 1675 | for_each_zone_zonelist(zone, z, zonelist, high_zoneidx) { |
1630 | 1676 | ||
1631 | if (!cpuset_zone_allowed_hardwall(zone, GFP_KERNEL)) | 1677 | if (!cpuset_zone_allowed_hardwall(zone, GFP_KERNEL)) |
@@ -1661,19 +1707,24 @@ unsigned long try_to_free_pages(struct zonelist *zonelist, int order, | |||
1661 | #ifdef CONFIG_CGROUP_MEM_RES_CTLR | 1707 | #ifdef CONFIG_CGROUP_MEM_RES_CTLR |
1662 | 1708 | ||
1663 | unsigned long try_to_free_mem_cgroup_pages(struct mem_cgroup *mem_cont, | 1709 | unsigned long try_to_free_mem_cgroup_pages(struct mem_cgroup *mem_cont, |
1664 | gfp_t gfp_mask) | 1710 | gfp_t gfp_mask, |
1711 | bool noswap, | ||
1712 | unsigned int swappiness) | ||
1665 | { | 1713 | { |
1666 | struct scan_control sc = { | 1714 | struct scan_control sc = { |
1667 | .may_writepage = !laptop_mode, | 1715 | .may_writepage = !laptop_mode, |
1668 | .may_swap = 1, | 1716 | .may_swap = 1, |
1669 | .swap_cluster_max = SWAP_CLUSTER_MAX, | 1717 | .swap_cluster_max = SWAP_CLUSTER_MAX, |
1670 | .swappiness = vm_swappiness, | 1718 | .swappiness = swappiness, |
1671 | .order = 0, | 1719 | .order = 0, |
1672 | .mem_cgroup = mem_cont, | 1720 | .mem_cgroup = mem_cont, |
1673 | .isolate_pages = mem_cgroup_isolate_pages, | 1721 | .isolate_pages = mem_cgroup_isolate_pages, |
1674 | }; | 1722 | }; |
1675 | struct zonelist *zonelist; | 1723 | struct zonelist *zonelist; |
1676 | 1724 | ||
1725 | if (noswap) | ||
1726 | sc.may_swap = 0; | ||
1727 | |||
1677 | sc.gfp_mask = (gfp_mask & GFP_RECLAIM_MASK) | | 1728 | sc.gfp_mask = (gfp_mask & GFP_RECLAIM_MASK) | |
1678 | (GFP_HIGHUSER_MOVABLE & ~GFP_RECLAIM_MASK); | 1729 | (GFP_HIGHUSER_MOVABLE & ~GFP_RECLAIM_MASK); |
1679 | zonelist = NODE_DATA(numa_node_id())->node_zonelists; | 1730 | zonelist = NODE_DATA(numa_node_id())->node_zonelists; |
@@ -1761,7 +1812,7 @@ loop_again: | |||
1761 | * Do some background aging of the anon list, to give | 1812 | * Do some background aging of the anon list, to give |
1762 | * pages a chance to be referenced before reclaiming. | 1813 | * pages a chance to be referenced before reclaiming. |
1763 | */ | 1814 | */ |
1764 | if (inactive_anon_is_low(zone)) | 1815 | if (inactive_anon_is_low(zone, &sc)) |
1765 | shrink_active_list(SWAP_CLUSTER_MAX, zone, | 1816 | shrink_active_list(SWAP_CLUSTER_MAX, zone, |
1766 | &sc, priority, 0); | 1817 | &sc, priority, 0); |
1767 | 1818 | ||
@@ -2404,6 +2455,7 @@ retry: | |||
2404 | 2455 | ||
2405 | __dec_zone_state(zone, NR_UNEVICTABLE); | 2456 | __dec_zone_state(zone, NR_UNEVICTABLE); |
2406 | list_move(&page->lru, &zone->lru[l].list); | 2457 | list_move(&page->lru, &zone->lru[l].list); |
2458 | mem_cgroup_move_lists(page, LRU_UNEVICTABLE, l); | ||
2407 | __inc_zone_state(zone, NR_INACTIVE_ANON + l); | 2459 | __inc_zone_state(zone, NR_INACTIVE_ANON + l); |
2408 | __count_vm_event(UNEVICTABLE_PGRESCUED); | 2460 | __count_vm_event(UNEVICTABLE_PGRESCUED); |
2409 | } else { | 2461 | } else { |
@@ -2412,6 +2464,7 @@ retry: | |||
2412 | */ | 2464 | */ |
2413 | SetPageUnevictable(page); | 2465 | SetPageUnevictable(page); |
2414 | list_move(&page->lru, &zone->lru[LRU_UNEVICTABLE].list); | 2466 | list_move(&page->lru, &zone->lru[LRU_UNEVICTABLE].list); |
2467 | mem_cgroup_rotate_lru_list(page, LRU_UNEVICTABLE); | ||
2415 | if (page_evictable(page, NULL)) | 2468 | if (page_evictable(page, NULL)) |
2416 | goto retry; | 2469 | goto retry; |
2417 | } | 2470 | } |
diff --git a/net/ipv6/route.c b/net/ipv6/route.c index 76f06b94ab9f..c4a59824ac2c 100644 --- a/net/ipv6/route.c +++ b/net/ipv6/route.c | |||
@@ -2752,7 +2752,7 @@ int __init ip6_route_init(void) | |||
2752 | kmem_cache_create("ip6_dst_cache", sizeof(struct rt6_info), 0, | 2752 | kmem_cache_create("ip6_dst_cache", sizeof(struct rt6_info), 0, |
2753 | SLAB_HWCACHE_ALIGN, NULL); | 2753 | SLAB_HWCACHE_ALIGN, NULL); |
2754 | if (!ip6_dst_ops_template.kmem_cachep) | 2754 | if (!ip6_dst_ops_template.kmem_cachep) |
2755 | goto out;; | 2755 | goto out; |
2756 | 2756 | ||
2757 | ret = register_pernet_subsys(&ip6_route_net_ops); | 2757 | ret = register_pernet_subsys(&ip6_route_net_ops); |
2758 | if (ret) | 2758 | if (ret) |
diff --git a/net/ipv6/sysctl_net_ipv6.c b/net/ipv6/sysctl_net_ipv6.c index 9048fe7e7ea7..a031034720b4 100644 --- a/net/ipv6/sysctl_net_ipv6.c +++ b/net/ipv6/sysctl_net_ipv6.c | |||
@@ -128,7 +128,7 @@ static struct ctl_table_header *ip6_header; | |||
128 | 128 | ||
129 | int ipv6_sysctl_register(void) | 129 | int ipv6_sysctl_register(void) |
130 | { | 130 | { |
131 | int err = -ENOMEM;; | 131 | int err = -ENOMEM; |
132 | 132 | ||
133 | ip6_header = register_net_sysctl_rotable(net_ipv6_ctl_path, ipv6_table); | 133 | ip6_header = register_net_sysctl_rotable(net_ipv6_ctl_path, ipv6_table); |
134 | if (ip6_header == NULL) | 134 | if (ip6_header == NULL) |
diff --git a/net/sched/sch_sfq.c b/net/sched/sch_sfq.c index f3965df00559..33133d27b539 100644 --- a/net/sched/sch_sfq.c +++ b/net/sched/sch_sfq.c | |||
@@ -435,7 +435,7 @@ static int sfq_init(struct Qdisc *sch, struct nlattr *opt) | |||
435 | int i; | 435 | int i; |
436 | 436 | ||
437 | q->perturb_timer.function = sfq_perturbation; | 437 | q->perturb_timer.function = sfq_perturbation; |
438 | q->perturb_timer.data = (unsigned long)sch;; | 438 | q->perturb_timer.data = (unsigned long)sch; |
439 | init_timer_deferrable(&q->perturb_timer); | 439 | init_timer_deferrable(&q->perturb_timer); |
440 | 440 | ||
441 | for (i = 0; i < SFQ_HASH_DIVISOR; i++) | 441 | for (i = 0; i < SFQ_HASH_DIVISOR; i++) |
diff --git a/net/sctp/auth.c b/net/sctp/auth.c index 20c576f530fa..56935bbc1496 100644 --- a/net/sctp/auth.c +++ b/net/sctp/auth.c | |||
@@ -489,7 +489,7 @@ int sctp_auth_init_hmacs(struct sctp_endpoint *ep, gfp_t gfp) | |||
489 | return 0; | 489 | return 0; |
490 | 490 | ||
491 | out_err: | 491 | out_err: |
492 | /* Clean up any successfull allocations */ | 492 | /* Clean up any successful allocations */ |
493 | sctp_auth_destroy_hmacs(ep->auth_hmacs); | 493 | sctp_auth_destroy_hmacs(ep->auth_hmacs); |
494 | return -ENOMEM; | 494 | return -ENOMEM; |
495 | } | 495 | } |
diff --git a/security/device_cgroup.c b/security/device_cgroup.c index 5ba78701adc3..3aacd0fe7179 100644 --- a/security/device_cgroup.c +++ b/security/device_cgroup.c | |||
@@ -513,11 +513,14 @@ int devcgroup_inode_mknod(int mode, dev_t dev) | |||
513 | struct dev_cgroup *dev_cgroup; | 513 | struct dev_cgroup *dev_cgroup; |
514 | struct dev_whitelist_item *wh; | 514 | struct dev_whitelist_item *wh; |
515 | 515 | ||
516 | if (!S_ISBLK(mode) && !S_ISCHR(mode)) | ||
517 | return 0; | ||
518 | |||
516 | rcu_read_lock(); | 519 | rcu_read_lock(); |
517 | 520 | ||
518 | dev_cgroup = task_devcgroup(current); | 521 | dev_cgroup = task_devcgroup(current); |
519 | 522 | ||
520 | list_for_each_entry(wh, &dev_cgroup->whitelist, list) { | 523 | list_for_each_entry_rcu(wh, &dev_cgroup->whitelist, list) { |
521 | if (wh->type & DEV_ALL) | 524 | if (wh->type & DEV_ALL) |
522 | goto acc_check; | 525 | goto acc_check; |
523 | if ((wh->type & DEV_BLOCK) && !S_ISBLK(mode)) | 526 | if ((wh->type & DEV_BLOCK) && !S_ISBLK(mode)) |
diff --git a/security/smack/smackfs.c b/security/smack/smackfs.c index bf107a389ac1..71e2b914363e 100644 --- a/security/smack/smackfs.c +++ b/security/smack/smackfs.c | |||
@@ -569,7 +569,7 @@ static ssize_t smk_write_cipso(struct file *file, const char __user *buf, | |||
569 | if (skp == NULL) | 569 | if (skp == NULL) |
570 | goto out; | 570 | goto out; |
571 | 571 | ||
572 | rule += SMK_LABELLEN;; | 572 | rule += SMK_LABELLEN; |
573 | ret = sscanf(rule, "%d", &maplevel); | 573 | ret = sscanf(rule, "%d", &maplevel); |
574 | if (ret != 1 || maplevel > SMACK_CIPSO_MAXLEVEL) | 574 | if (ret != 1 || maplevel > SMACK_CIPSO_MAXLEVEL) |
575 | goto out; | 575 | goto out; |
diff --git a/sound/soc/au1x/dbdma2.c b/sound/soc/au1x/dbdma2.c index 74c823d60f91..bc8d654576c0 100644 --- a/sound/soc/au1x/dbdma2.c +++ b/sound/soc/au1x/dbdma2.c | |||
@@ -187,7 +187,7 @@ static int au1x_pcm_dbdma_realloc(struct au1xpsc_audio_dmadata *pcd, | |||
187 | au1x_pcm_dmatx_cb, (void *)pcd); | 187 | au1x_pcm_dmatx_cb, (void *)pcd); |
188 | 188 | ||
189 | if (!pcd->ddma_chan) | 189 | if (!pcd->ddma_chan) |
190 | return -ENOMEM;; | 190 | return -ENOMEM; |
191 | 191 | ||
192 | au1xxx_dbdma_set_devwidth(pcd->ddma_chan, msbits); | 192 | au1xxx_dbdma_set_devwidth(pcd->ddma_chan, msbits); |
193 | au1xxx_dbdma_ring_alloc(pcd->ddma_chan, 2); | 193 | au1xxx_dbdma_ring_alloc(pcd->ddma_chan, 2); |
diff --git a/sound/soc/davinci/davinci-pcm.c b/sound/soc/davinci/davinci-pcm.c index 74abc9b4f1cc..366049d8578c 100644 --- a/sound/soc/davinci/davinci-pcm.c +++ b/sound/soc/davinci/davinci-pcm.c | |||
@@ -212,7 +212,7 @@ davinci_pcm_pointer(struct snd_pcm_substream *substream) | |||
212 | if (substream->stream == SNDRV_PCM_STREAM_PLAYBACK) | 212 | if (substream->stream == SNDRV_PCM_STREAM_PLAYBACK) |
213 | count = src - runtime->dma_addr; | 213 | count = src - runtime->dma_addr; |
214 | else | 214 | else |
215 | count = dst - runtime->dma_addr;; | 215 | count = dst - runtime->dma_addr; |
216 | 216 | ||
217 | spin_unlock(&prtd->lock); | 217 | spin_unlock(&prtd->lock); |
218 | 218 | ||