Merge branch 'master' into upstream

1 files changed, 130 insertions, 54 deletions

diff --git a/mm/memory-failure.c b/mm/memory-failure.c
index 757f6b0accfe..46ab2c044b0e 100644
--- a/mm/memory-failure.c
+++ b/mm/memory-failure.c
@@ -7,21 +7,26 @@
  * Free Software Foundation.
  *
  * High level machine check handler. Handles pages reported by the
- * hardware as being corrupted usually due to a 2bit ECC memory or cache
+ * hardware as being corrupted usually due to a multi-bit ECC memory or cache
  * failure.
+ * 
+ * In addition there is a "soft offline" entry point that allows stop using
+ * not-yet-corrupted-by-suspicious pages without killing anything.
  *
  * Handles page cache pages in various states.  The tricky part
- * here is that we can access any page asynchronous to other VM
- * users, because memory failures could happen anytime and anywhere,
- * possibly violating some of their assumptions. This is why this code
- * has to be extremely careful. Generally it tries to use normal locking
- * rules, as in get the standard locks, even if that means the
- * error handling takes potentially a long time.
- *
- * The operation to map back from RMAP chains to processes has to walk
- * the complete process list and has non linear complexity with the number
- * mappings. In short it can be quite slow. But since memory corruptions
- * are rare we hope to get away with this.
+ * here is that we can access any page asynchronously in respect to 
+ * other VM users, because memory failures could happen anytime and 
+ * anywhere. This could violate some of their assumptions. This is why 
+ * this code has to be extremely careful. Generally it tries to use 
+ * normal locking rules, as in get the standard locks, even if that means 
+ * the error handling takes potentially a long time.
+ * 
+ * There are several operations here with exponential complexity because
+ * of unsuitable VM data structures. For example the operation to map back 
+ * from RMAP chains to processes has to walk the complete process list and 
+ * has non linear complexity with the number. But since memory corruptions
+ * are rare we hope to get away with this. This avoids impacting the core 
+ * VM.
  */
 
 /*
@@ -30,7 +35,6 @@
  * - kcore/oldmem/vmcore/mem/kmem check for hwpoison pages
  * - pass bad pages to kdump next kernel
  */
-#define DEBUG 1         /* remove me in 2.6.34 */
 #include <linux/kernel.h>
 #include <linux/mm.h>
 #include <linux/page-flags.h>
@@ -47,6 +51,7 @@
 #include <linux/slab.h>
 #include <linux/swapops.h>
 #include <linux/hugetlb.h>
+#include <linux/memory_hotplug.h>
 #include "internal.h"
 
 int sysctl_memory_failure_early_kill __read_mostly = 0;
@@ -78,7 +83,7 @@ static int hwpoison_filter_dev(struct page *p)
                 return 0;
 
         /*
-         * page_mapping() does not accept slab page
+         * page_mapping() does not accept slab pages.
          */
         if (PageSlab(p))
                 return -EINVAL;
@@ -268,7 +273,7 @@ struct to_kill {
         struct list_head nd;
         struct task_struct *tsk;
         unsigned long addr;
-        unsigned addr_valid:1;
+        char addr_valid;
 };
 
 /*
@@ -309,7 +314,7 @@ static void add_to_kill(struct task_struct *tsk, struct page *p,
          * a SIGKILL because the error is not contained anymore.
          */
         if (tk->addr == -EFAULT) {
-                pr_debug("MCE: Unable to find user space address %lx in %s\n",
+                pr_info("MCE: Unable to find user space address %lx in %s\n",
                         page_to_pfn(p), tsk->comm);
                 tk->addr_valid = 0;
         }
@@ -577,7 +582,7 @@ static int me_pagecache_clean(struct page *p, unsigned long pfn)
                                         pfn, err);
                 } else if (page_has_private(p) &&
                                 !try_to_release_page(p, GFP_NOIO)) {
-                        pr_debug("MCE %#lx: failed to release buffers\n", pfn);
+                        pr_info("MCE %#lx: failed to release buffers\n", pfn);
                 } else {
                         ret = RECOVERED;
                 }
@@ -693,11 +698,10 @@ static int me_swapcache_clean(struct page *p, unsigned long pfn)
  * Issues:
  * - Error on hugepage is contained in hugepage unit (not in raw page unit.)
  *   To narrow down kill region to one page, we need to break up pmd.
- * - To support soft-offlining for hugepage, we need to support hugepage
- *   migration.
  */
 static int me_huge_page(struct page *p, unsigned long pfn)
 {
+        int res = 0;
         struct page *hpage = compound_head(p);
         /*
          * We can safely recover from error on free or reserved (i.e.
@@ -710,8 +714,9 @@ static int me_huge_page(struct page *p, unsigned long pfn)
          * so there is no race between isolation and mapping/unmapping.
          */
         if (!(page_mapping(hpage) || PageAnon(hpage))) {
-                __isolate_hwpoisoned_huge_page(hpage);
-                return RECOVERED;
+                res = dequeue_hwpoisoned_huge_page(hpage);
+                if (!res)
+                        return RECOVERED;
         }
         return DELAYED;
 }
@@ -836,8 +841,6 @@ static int page_action(struct page_state *ps, struct page *p,
         return (result == RECOVERED || result == DELAYED) ? 0 : -EBUSY;
 }
 
-#define N_UNMAP_TRIES 5
-
 /*
  * Do all that is necessary to remove user space mappings. Unmap
  * the pages and send SIGBUS to the processes if the data was dirty.
@@ -849,7 +852,6 @@ static int hwpoison_user_mappings(struct page *p, unsigned long pfn,
         struct address_space *mapping;
         LIST_HEAD(tokill);
         int ret;
-        int i;
         int kill = 1;
         struct page *hpage = compound_head(p);
 
@@ -903,17 +905,7 @@ static int hwpoison_user_mappings(struct page *p, unsigned long pfn,
         if (kill)
                 collect_procs(hpage, &tokill);
 
-        /*
-         * try_to_unmap can fail temporarily due to races.
-         * Try a few times (RED-PEN better strategy?)
-         */
-        for (i = 0; i < N_UNMAP_TRIES; i++) {
-                ret = try_to_unmap(hpage, ttu);
-                if (ret == SWAP_SUCCESS)
-                        break;
-                pr_debug("MCE %#lx: try_to_unmap retry needed %d\n", pfn,  ret);
-        }
-
+        ret = try_to_unmap(hpage, ttu);
         if (ret != SWAP_SUCCESS)
                 printk(KERN_ERR "MCE %#lx: failed to unmap page (mapcount=%d)\n",
                                 pfn, page_mapcount(hpage));
@@ -981,7 +973,10 @@ int __memory_failure(unsigned long pfn, int trapno, int flags)
          * We need/can do nothing about count=0 pages.
          * 1) it's a free page, and therefore in safe hand:
          *    prep_new_page() will be the gate keeper.
-         * 2) it's part of a non-compound high order page.
+         * 2) it's a free hugepage, which is also safe:
+         *    an affected hugepage will be dequeued from hugepage freelist,
+         *    so there's no concern about reusing it ever after.
+         * 3) it's part of a non-compound high order page.
          *    Implies some kernel user: cannot stop them from
          *    R/W the page; let's pray that the page has been
          *    used and will be freed some time later.
@@ -993,6 +988,24 @@ int __memory_failure(unsigned long pfn, int trapno, int flags)
                 if (is_free_buddy_page(p)) {
                         action_result(pfn, "free buddy", DELAYED);
                         return 0;
+                } else if (PageHuge(hpage)) {
+                        /*
+                         * Check "just unpoisoned", "filter hit", and
+                         * "race with other subpage."
+                         */
+                        lock_page_nosync(hpage);
+                        if (!PageHWPoison(hpage)
+                            || (hwpoison_filter(p) && TestClearPageHWPoison(p))
+                            || (p != hpage && TestSetPageHWPoison(hpage))) {
+                                atomic_long_sub(nr_pages, &mce_bad_pages);
+                                return 0;
+                        }
+                        set_page_hwpoison_huge_page(hpage);
+                        res = dequeue_hwpoisoned_huge_page(hpage);
+                        action_result(pfn, "free huge",
+                                      res ? IGNORED : DELAYED);
+                        unlock_page(hpage);
+                        return res;
                 } else {
                         action_result(pfn, "high order kernel", IGNORED);
                         return -EBUSY;
@@ -1147,16 +1160,26 @@ int unpoison_memory(unsigned long pfn)
         page = compound_head(p);
 
         if (!PageHWPoison(p)) {
-                pr_debug("MCE: Page was already unpoisoned %#lx\n", pfn);
+                pr_info("MCE: Page was already unpoisoned %#lx\n", pfn);
                 return 0;
         }
 
         nr_pages = 1 << compound_order(page);
 
         if (!get_page_unless_zero(page)) {
+                /*
+                 * Since HWPoisoned hugepage should have non-zero refcount,
+                 * race between memory failure and unpoison seems to happen.
+                 * In such case unpoison fails and memory failure runs
+                 * to the end.
+                 */
+                if (PageHuge(page)) {
+                        pr_debug("MCE: Memory failure is now running on free hugepage %#lx\n", pfn);
+                        return 0;
+                }
                 if (TestClearPageHWPoison(p))
                         atomic_long_sub(nr_pages, &mce_bad_pages);
-                pr_debug("MCE: Software-unpoisoned free page %#lx\n", pfn);
+                pr_info("MCE: Software-unpoisoned free page %#lx\n", pfn);
                 return 0;
         }
 
@@ -1168,12 +1191,12 @@ int unpoison_memory(unsigned long pfn)
          * the free buddy page pool.
          */
         if (TestClearPageHWPoison(page)) {
-                pr_debug("MCE: Software-unpoisoned page %#lx\n", pfn);
+                pr_info("MCE: Software-unpoisoned page %#lx\n", pfn);
                 atomic_long_sub(nr_pages, &mce_bad_pages);
                 freeit = 1;
+                if (PageHuge(page))
+                        clear_page_hwpoison_huge_page(page);
         }
-        if (PageHuge(p))
-                clear_page_hwpoison_huge_page(page);
         unlock_page(page);
 
         put_page(page);
@@ -1187,7 +1210,11 @@ EXPORT_SYMBOL(unpoison_memory);
 static struct page *new_page(struct page *p, unsigned long private, int **x)
 {
         int nid = page_to_nid(p);
-        return alloc_pages_exact_node(nid, GFP_HIGHUSER_MOVABLE, 0);
+        if (PageHuge(p))
+                return alloc_huge_page_node(page_hstate(compound_head(p)),
+                                                   nid);
+        else
+                return alloc_pages_exact_node(nid, GFP_HIGHUSER_MOVABLE, 0);
 }
 
 /*
@@ -1204,25 +1231,31 @@ static int get_any_page(struct page *p, unsigned long pfn, int flags)
                 return 1;
 
         /*
-         * The lock_system_sleep prevents a race with memory hotplug,
-         * because the isolation assumes there's only a single user.
+         * The lock_memory_hotplug prevents a race with memory hotplug.
          * This is a big hammer, a better would be nicer.
          */
-        lock_system_sleep();
+        lock_memory_hotplug();
 
         /*
          * Isolate the page, so that it doesn't get reallocated if it
          * was free.
          */
         set_migratetype_isolate(p);
+        /*
+         * When the target page is a free hugepage, just remove it
+         * from free hugepage list.
+         */
         if (!get_page_unless_zero(compound_head(p))) {
-                if (is_free_buddy_page(p)) {
-                        pr_debug("get_any_page: %#lx free buddy page\n", pfn);
+                if (PageHuge(p)) {
+                        pr_info("get_any_page: %#lx free huge page\n", pfn);
+                        ret = dequeue_hwpoisoned_huge_page(compound_head(p));
+                } else if (is_free_buddy_page(p)) {
+                        pr_info("get_any_page: %#lx free buddy page\n", pfn);
                         /* Set hwpoison bit while page is still isolated */
                         SetPageHWPoison(p);
                         ret = 0;
                 } else {
-                        pr_debug("get_any_page: %#lx: unknown zero refcount page type %lx\n",
+                        pr_info("get_any_page: %#lx: unknown zero refcount page type %lx\n",
                                 pfn, p->flags);
                         ret = -EIO;
                 }
@@ -1231,7 +1264,47 @@ static int get_any_page(struct page *p, unsigned long pfn, int flags)
                 ret = 1;
         }
         unset_migratetype_isolate(p);
-        unlock_system_sleep();
+        unlock_memory_hotplug();
+        return ret;
+}
+
+static int soft_offline_huge_page(struct page *page, int flags)
+{
+        int ret;
+        unsigned long pfn = page_to_pfn(page);
+        struct page *hpage = compound_head(page);
+        LIST_HEAD(pagelist);
+
+        ret = get_any_page(page, pfn, flags);
+        if (ret < 0)
+                return ret;
+        if (ret == 0)
+                goto done;
+
+        if (PageHWPoison(hpage)) {
+                put_page(hpage);
+                pr_debug("soft offline: %#lx hugepage already poisoned\n", pfn);
+                return -EBUSY;
+        }
+
+        /* Keep page count to indicate a given hugepage is isolated. */
+
+        list_add(&hpage->lru, &pagelist);
+        ret = migrate_huge_pages(&pagelist, new_page, MPOL_MF_MOVE_ALL, 0);
+        if (ret) {
+                        putback_lru_pages(&pagelist);
+                pr_debug("soft offline: %#lx: migration failed %d, type %lx\n",
+                         pfn, ret, page->flags);
+                if (ret > 0)
+                        ret = -EIO;
+                return ret;
+        }
+done:
+        if (!PageHWPoison(hpage))
+                atomic_long_add(1 << compound_order(hpage), &mce_bad_pages);
+        set_page_hwpoison_huge_page(hpage);
+        dequeue_hwpoisoned_huge_page(hpage);
+        /* keep elevated page count for bad page */
         return ret;
 }
 
@@ -1262,6 +1335,9 @@ int soft_offline_page(struct page *page, int flags)
         int ret;
         unsigned long pfn = page_to_pfn(page);
 
+        if (PageHuge(page))
+                return soft_offline_huge_page(page, flags);
+
         ret = get_any_page(page, pfn, flags);
         if (ret < 0)
                 return ret;
@@ -1288,7 +1364,7 @@ int soft_offline_page(struct page *page, int flags)
                         goto done;
         }
         if (!PageLRU(page)) {
-                pr_debug("soft_offline: %#lx: unknown non LRU page type %lx\n",
+                pr_info("soft_offline: %#lx: unknown non LRU page type %lx\n",
                                 pfn, page->flags);
                 return -EIO;
         }
@@ -1302,7 +1378,7 @@ int soft_offline_page(struct page *page, int flags)
         if (PageHWPoison(page)) {
                 unlock_page(page);
                 put_page(page);
-                pr_debug("soft offline: %#lx page already poisoned\n", pfn);
+                pr_info("soft offline: %#lx page already poisoned\n", pfn);
                 return -EBUSY;
         }
 
@@ -1323,7 +1399,7 @@ int soft_offline_page(struct page *page, int flags)
         put_page(page);
         if (ret == 1) {
                 ret = 0;
-                pr_debug("soft_offline: %#lx: invalidated\n", pfn);
+                pr_info("soft_offline: %#lx: invalidated\n", pfn);
                 goto done;
         }
 
@@ -1339,13 +1415,13 @@ int soft_offline_page(struct page *page, int flags)
                 list_add(&page->lru, &pagelist);
                 ret = migrate_pages(&pagelist, new_page, MPOL_MF_MOVE_ALL, 0);
                 if (ret) {
-                        pr_debug("soft offline: %#lx: migration failed %d, type %lx\n",
+                        pr_info("soft offline: %#lx: migration failed %d, type %lx\n",
                                 pfn, ret, page->flags);
                         if (ret > 0)
                                 ret = -EIO;
                 }
         } else {
-                pr_debug("soft offline: %#lx: isolation failed: %d, page count %d, type %lx\n",
+                pr_info("soft offline: %#lx: isolation failed: %d, page count %d, type %lx\n",
                                 pfn, ret, page_count(page), page->flags);
         }
         if (ret)