Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > linux.kernel > #1717733

[PATCH 4.12 09/41] md: fix test in md_write_start()

Show all headers | View raw

4.12-stable review patch.  If anyone has any objections, please let me know.

------------------

From: NeilBrown <neilb@suse.com>

commit 81fe48e9aa00bdd509bd3c37a76d1132da6b9f09 upstream.

md_write_start() needs to clear the in_sync flag is it is set, or if
there might be a race with set_in_sync() such that the later will
set it very soon.  In the later case it is sufficient to take the
spinlock to synchronize with set_in_sync(), and then set the flag
if needed.

The current test is incorrect.
It should be:
  if "flag is set" or "race is possible"

"flag is set" is trivially "mddev->in_sync".
"race is possible" should be tested by "mddev->sync_checkers".

If sync_checkers is 0, then there can be no race.  set_in_sync() will
wait in percpu_ref_switch_to_atomic_sync() for an RCU grace period,
and as md_write_start() holds the rcu_read_lock(), set_in_sync() will
be sure ot see the update to writes_pending.

If sync_checkers is > 0, there could be race.  If md_write_start()
happened entirely between
		if (!mddev->in_sync &&
		    percpu_ref_is_zero(&mddev->writes_pending)) {
and
			mddev->in_sync = 1;
in set_in_sync(), then it would not see that is_sync had been set,
and set_in_sync() would not see that writes_pending had been
incremented.

This bug means that in_sync is sometimes not set when it should be.
Consequently there is a small chance that the array will be marked as
"clean" when in fact it is inconsistent.

Fixes: 4ad23a976413 ("MD: use per-cpu counter for writes_pending")
Signed-off-by: NeilBrown <neilb@suse.com>
Signed-off-by: Shaohua Li <shli@fb.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

---
 drivers/md/md.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

--- a/drivers/md/md.c
+++ b/drivers/md/md.c
@@ -7979,7 +7979,7 @@ bool md_write_start(struct mddev *mddev,
 	if (mddev->safemode == 1)
 		mddev->safemode = 0;
 	/* sync_checkers is always 0 when writes_pending is in per-cpu mode */
-	if (mddev->in_sync || !mddev->sync_checkers) {
+	if (mddev->in_sync || mddev->sync_checkers) {
 		spin_lock(&mddev->lock);
 		if (mddev->in_sync) {
 			mddev->in_sync = 0;

Back to linux.kernel | Previous | Next — Previous in thread | Next in thread | Find similar | Unroll thread

Thread

[PATCH 4.12 00/41] 4.12.9-stable review Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-08-22 21:30 +0200
  [PATCH 4.12 28/41] perf/x86: Fix RDPMC vs. mm_struct tracking Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-08-22 21:30 +0200
  [PATCH 4.12 09/41] md: fix test in md_write_start() Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-08-22 21:30 +0200
  [PATCH 4.12 31/41] x86/elf: Remove the unnecessary ADDR_NO_RANDOMIZE checks Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-08-22 21:30 +0200
  [PATCH 4.12 19/41] mm/cma_debug.c: fix stack corruption due to sprintf usage Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-08-22 21:30 +0200
  [PATCH 4.12 21/41] mm/vmalloc.c: dont unconditonally use __GFP_HIGHMEM Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-08-22 21:30 +0200
  [PATCH 4.12 02/41] parisc: pci memory bar assignment fails with 64bit kernels on dino/cujo Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-08-22 21:30 +0200
  [PATCH 4.12 10/41] md: always clear ->safemode when md_check_recovery gets the mddev lock. Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-08-22 21:30 +0200
  [PATCH 4.12 25/41] blk-mq-pci: add a fallback when pci_irq_get_affinity returns NULL Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-08-22 21:30 +0200
  [PATCH 4.12 05/41] drm/i915: Perform an invalidate prior to executing golden renderstate Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-08-22 21:30 +0200
  [PATCH 4.12 22/41] mm: revert x86_64 and arm64 ELF_ET_DYN_BASE base changes Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-08-22 21:30 +0200
  [PATCH 4.12 01/41] audit: Fix use after free in audit_remove_watch_rule() Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-08-22 21:30 +0200
  [PATCH 4.12 27/41] xen-blkfront: use a right index when checking requests Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-08-22 21:30 +0200
  [PATCH 4.12 18/41] mm: fix double mmap_sem unlock on MMF_UNSTABLE enforced SIGBUS Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-08-22 21:30 +0200
  [PATCH 4.12 32/41] irqchip/atmel-aic: Fix unbalanced of_node_put() in aic_common_irq_fixup() Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-08-22 21:30 +0200
  [PATCH 4.12 29/41] x86/asm/64: Clear AC on NMI entries Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-08-22 21:30 +0200
  Re: [PATCH 4.12 00/41] 4.12.9-stable review Shuah Khan <shuahkh@osg.samsung.com> - 2017-08-23 02:40 +0200
    Re: [PATCH 4.12 00/41] 4.12.9-stable review Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-08-23 02:50 +0200

csiph-web