aboutsummaryrefslogtreecommitdiffstats
path: root/drivers/md
diff options
context:
space:
mode:
authorNeilBrown <neilb@suse.de>2008-08-31 22:48:13 -0400
committerNeilBrown <neilb@suse.de>2008-08-31 22:48:13 -0400
commitb2d2c4ceaddc3098f19637a732f74b820a81a9e7 (patch)
tree5fb4150c33066f9fa70cc39853e5400ad6d54dca /drivers/md
parent271f5a9b8f8ae0db95de72779d115c9d0b9d3cc5 (diff)
Fix problem with waiting while holding rcu read lock in md/bitmap.c
A recent patch to protect the rdev list with rcu locking leaves us with a problem because we can sleep on memalloc while holding the rcu lock. The rcu lock is only needed while walking the linked list as uninteresting devices (failed or spares) can be removed at any time. So only take the rcu lock while actually walking the linked list. Take a refcount on the rdev during the time when we drop the lock and do the memalloc to start IO. When we return to the locked code, all the interesting devices on the list will not have moved, so we can simply use list_for_each_continue_rcu to pick up where we left off. Signed-off-by: NeilBrown <neilb@suse.de>
Diffstat (limited to 'drivers/md')
-rw-r--r--drivers/md/bitmap.c45
1 files changed, 38 insertions, 7 deletions
diff --git a/drivers/md/bitmap.c b/drivers/md/bitmap.c
index 7e65bad522cb..ac89a5deaca2 100644
--- a/drivers/md/bitmap.c
+++ b/drivers/md/bitmap.c
@@ -238,15 +238,47 @@ static struct page *read_sb_page(mddev_t *mddev, long offset, unsigned long inde
238 238
239} 239}
240 240
241static mdk_rdev_t *next_active_rdev(mdk_rdev_t *rdev, mddev_t *mddev)
242{
243 /* Iterate the disks of an mddev, using rcu to protect access to the
244 * linked list, and raising the refcount of devices we return to ensure
245 * they don't disappear while in use.
246 * As devices are only added or removed when raid_disk is < 0 and
247 * nr_pending is 0 and In_sync is clear, the entries we return will
248 * still be in the same position on the list when we re-enter
249 * list_for_each_continue_rcu.
250 */
251 struct list_head *pos;
252 rcu_read_lock();
253 if (rdev == NULL)
254 /* start at the beginning */
255 pos = &mddev->disks;
256 else {
257 /* release the previous rdev and start from there. */
258 rdev_dec_pending(rdev, mddev);
259 pos = &rdev->same_set;
260 }
261 list_for_each_continue_rcu(pos, &mddev->disks) {
262 rdev = list_entry(pos, mdk_rdev_t, same_set);
263 if (rdev->raid_disk >= 0 &&
264 test_bit(In_sync, &rdev->flags) &&
265 !test_bit(Faulty, &rdev->flags)) {
266 /* this is a usable devices */
267 atomic_inc(&rdev->nr_pending);
268 rcu_read_unlock();
269 return rdev;
270 }
271 }
272 rcu_read_unlock();
273 return NULL;
274}
275
241static int write_sb_page(struct bitmap *bitmap, struct page *page, int wait) 276static int write_sb_page(struct bitmap *bitmap, struct page *page, int wait)
242{ 277{
243 mdk_rdev_t *rdev; 278 mdk_rdev_t *rdev = NULL;
244 mddev_t *mddev = bitmap->mddev; 279 mddev_t *mddev = bitmap->mddev;
245 280
246 rcu_read_lock(); 281 while ((rdev = next_active_rdev(rdev, mddev)) != NULL) {
247 rdev_for_each_rcu(rdev, mddev)
248 if (test_bit(In_sync, &rdev->flags)
249 && !test_bit(Faulty, &rdev->flags)) {
250 int size = PAGE_SIZE; 282 int size = PAGE_SIZE;
251 if (page->index == bitmap->file_pages-1) 283 if (page->index == bitmap->file_pages-1)
252 size = roundup(bitmap->last_page_size, 284 size = roundup(bitmap->last_page_size,
@@ -281,8 +313,7 @@ static int write_sb_page(struct bitmap *bitmap, struct page *page, int wait)
281 + page->index * (PAGE_SIZE/512), 313 + page->index * (PAGE_SIZE/512),
282 size, 314 size,
283 page); 315 page);
284 } 316 }
285 rcu_read_unlock();
286 317
287 if (wait) 318 if (wait)
288 md_super_wait(mddev); 319 md_super_wait(mddev);