fix corruption of hibernation caused by reusing swap at saving image
in [Kernel]
|
Prev: linux-next: build failure after merge of the rr tree
Next: [S+Q3 13/23] slub: Move gfpflag masking out of the hotpath
From: KAMEZAWA Hiroyuki on 3 Aug 2010 22:40 On Wed, 4 Aug 2010 10:50:54 +0900 KAMEZAWA Hiroyuki <kamezawa.hiroyu(a)jp.fujitsu.com> wrote: > This patch is created against 2.6.35. CC'ed stable. > Thank you for all helps. > I'm sorry I now doubt this patch is wrong. What should be done is stop swap-reuse before hibernation_snapshot(). But, hmm, this one can't. Maybe what really happens is someone (kswapd?) reuses swap while swsusp_alloc() is on going. please allow me to retry.. Thanks, -Kame > = > From: KAMEZAWA Hiroyuki <kamezawa.hiroyu(a)jp.fujitsu.com> > > Since 2.6.31, swap_map[]'s refcounting was changed to show that > a used swap entry is just for swap-cache, can be reused. > Then, while scanning free entry in swap_map[], a swap entry may > be able to be reclaimed and reused. It was by the commit > c9e444103b5e7a5a3519f9913f59767f92e33baf. > > But this caused deta corruption at hibernation. Considering how > the image is saved, the calls of try_to_reclaim_swap() changes the > status of memory and there will be inconsitency between saved-memory-image's > swap_map[] / memmap / swapper_space because memory is saved per page with > swap-allocation per page. > > This patch is for avoiding bug by not reclaiming swap-entry at hibernation. > This is a quick fix for backporting. > > Cc: stable(a)kernel.org > Cc: Rafael J. Wysocki <rjw(a)sisk.pl> > Reported-by: Ondreg Zary <linux(a)rainbow-software.org> > Tested-by: Andrea Gelmini <andrea.gelmini(a)gmail.com> > Acked-by: Hugh Dickins <hughd(a)google.com> > Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu(a)jp.fujitsu.com> > --- > mm/swapfile.c | 6 ++++-- > 1 file changed, 4 insertions(+), 2 deletions(-) > > Index: linux-2.6.35.org/mm/swapfile.c > =================================================================== > --- linux-2.6.35.org.orig/mm/swapfile.c > +++ linux-2.6.35.org/mm/swapfile.c > @@ -318,8 +318,10 @@ checks: > if (offset > si->highest_bit) > scan_base = offset = si->lowest_bit; > > - /* reuse swap entry of cache-only swap if not busy. */ > - if (vm_swap_full() && si->swap_map[offset] == SWAP_HAS_CACHE) { > + /* reuse swap entry of cache-only swap if not hibernation. */ > + if (vm_swap_full() > + && usage == SWAP_HAS_CACHE > + && si->swap_map[offset] == SWAP_HAS_CACHE) { > int swap_was_freed; > spin_unlock(&swap_lock); > swap_was_freed = __try_to_reclaim_swap(si, offset); > > -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo(a)vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
From: KAMEZAWA Hiroyuki on 3 Aug 2010 23:00
On Wed, 4 Aug 2010 11:31:22 +0900 KAMEZAWA Hiroyuki <kamezawa.hiroyu(a)jp.fujitsu.com> wrote: > On Wed, 4 Aug 2010 10:50:54 +0900 > KAMEZAWA Hiroyuki <kamezawa.hiroyu(a)jp.fujitsu.com> wrote: > > > This patch is created against 2.6.35. CC'ed stable. > > Thank you for all helps. > > > > I'm sorry I now doubt this patch is wrong. Changed the description. Maybe the patch description/my understanding was wrong but the patch itself should work. == From: KAMEZAWA Hiroyuki <kamezawa.hiroyu(a)jp.fujitsu.com> Since 2.6.31, swap_map[]'s refcounting was changed to show that a used swap entry is just for swap-cache, can be reused. Then, while scanning free entry in swap_map[], a swap entry may be able to be reclaimed and reused. It was by the commit c9e444103b5e7a5a3519f9913f59767f92e33baf. But this caused deta corruption at resume. The scenario is - Assume a clean-swap cache, but mapped. - at hibernation_snapshot[], clean-swap-cache is saved as clean-swap-cache and swap_map[] is marked as SWAP_HAS_CACHE. - then, save_image() is called. And reuse SWAP_HAS_CACHE entry to save image, and break the contents. After resume. - the memory reclaim runs and finds clean-not-referenced-swap-cache and discards it because it's marked as clean. But here, the contents on disk and swap-cache is inconsistent. Here, memory is corrupted. This patch is for avoiding bug by not reclaiming swap-entry at hibernation. This is a quick fix for backporting. Cc: stable(a)kernel.org Cc: Rafael J. Wysocki <rjw(a)sisk.pl> Reported-by: Ondreg Zary <linux(a)rainbow-software.org> Tested-by: Andrea Gelmini <andrea.gelmini(a)gmail.com> Acked-by: Hugh Dickins <hughd(a)google.com> Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu(a)jp.fujitsu.com> --- mm/swapfile.c | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-) Index: linux-2.6.35.org/mm/swapfile.c =================================================================== --- linux-2.6.35.org.orig/mm/swapfile.c +++ linux-2.6.35.org/mm/swapfile.c @@ -318,8 +318,10 @@ checks: if (offset > si->highest_bit) scan_base = offset = si->lowest_bit; - /* reuse swap entry of cache-only swap if not busy. */ - if (vm_swap_full() && si->swap_map[offset] == SWAP_HAS_CACHE) { + /* reuse swap entry of cache-only swap if not hibernation. */ + if (vm_swap_full() + && usage == SWAP_HAS_CACHE + && si->swap_map[offset] == SWAP_HAS_CACHE) { int swap_was_freed; spin_unlock(&swap_lock); swap_was_freed = __try_to_reclaim_swap(si, offset); -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo(a)vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/ |