patches.lists.linux.dev archive mirror
 help / color / mirror / Atom feed
From: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
To: stable@vger.kernel.org
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>,
	patches@lists.linux.dev, Jan Kara <jack@suse.cz>,
	Zhang Yi <yi.zhang@huawei.com>, Theodore Tso <tytso@mit.edu>,
	Sasha Levin <sashal@kernel.org>
Subject: [PATCH 5.15 20/89] jbd2: fix a race when checking checkpoint buffer busy
Date: Mon, 28 Aug 2023 12:13:21 +0200	[thread overview]
Message-ID: <20230828101150.843521536@linuxfoundation.org> (raw)
In-Reply-To: <20230828101150.163430842@linuxfoundation.org>

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

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

From: Zhang Yi <yi.zhang@huawei.com>

[ Upstream commit 46f881b5b1758dc4a35fba4a643c10717d0cf427 ]

Before removing checkpoint buffer from the t_checkpoint_list, we have to
check both BH_Dirty and BH_Lock bits together to distinguish buffers
have not been or were being written back. But __cp_buffer_busy() checks
them separately, it first check lock state and then check dirty, the
window between these two checks could be raced by writing back
procedure, which locks buffer and clears buffer dirty before I/O
completes. So it cannot guarantee checkpointing buffers been written
back to disk if some error happens later. Finally, it may clean
checkpoint transactions and lead to inconsistent filesystem.

jbd2_journal_forget() and __journal_try_to_free_buffer() also have the
same problem (journal_unmap_buffer() escape from this issue since it's
running under the buffer lock), so fix them through introducing a new
helper to try holding the buffer lock and remove really clean buffer.

Link: https://bugzilla.kernel.org/show_bug.cgi?id=217490
Cc: stable@vger.kernel.org
Suggested-by: Jan Kara <jack@suse.cz>
Signed-off-by: Zhang Yi <yi.zhang@huawei.com>
Reviewed-by: Jan Kara <jack@suse.cz>
Link: https://lore.kernel.org/r/20230606135928.434610-6-yi.zhang@huaweicloud.com
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 fs/jbd2/checkpoint.c  | 38 +++++++++++++++++++++++++++++++++++---
 fs/jbd2/transaction.c | 17 +++++------------
 include/linux/jbd2.h  |  1 +
 3 files changed, 41 insertions(+), 15 deletions(-)

diff --git a/fs/jbd2/checkpoint.c b/fs/jbd2/checkpoint.c
index ab72aeb766a74..fc6989e7a8c51 100644
--- a/fs/jbd2/checkpoint.c
+++ b/fs/jbd2/checkpoint.c
@@ -376,11 +376,15 @@ static unsigned long journal_shrink_one_cp_list(struct journal_head *jh,
 		jh = next_jh;
 		next_jh = jh->b_cpnext;
 
-		if (!destroy && __cp_buffer_busy(jh))
-			continue;
+		if (destroy) {
+			ret = __jbd2_journal_remove_checkpoint(jh);
+		} else {
+			ret = jbd2_journal_try_remove_checkpoint(jh);
+			if (ret < 0)
+				continue;
+		}
 
 		nr_freed++;
-		ret = __jbd2_journal_remove_checkpoint(jh);
 		if (ret) {
 			*released = true;
 			break;
@@ -616,6 +620,34 @@ int __jbd2_journal_remove_checkpoint(struct journal_head *jh)
 	return 1;
 }
 
+/*
+ * Check the checkpoint buffer and try to remove it from the checkpoint
+ * list if it's clean. Returns -EBUSY if it is not clean, returns 1 if
+ * it frees the transaction, 0 otherwise.
+ *
+ * This function is called with j_list_lock held.
+ */
+int jbd2_journal_try_remove_checkpoint(struct journal_head *jh)
+{
+	struct buffer_head *bh = jh2bh(jh);
+
+	if (!trylock_buffer(bh))
+		return -EBUSY;
+	if (buffer_dirty(bh)) {
+		unlock_buffer(bh);
+		return -EBUSY;
+	}
+	unlock_buffer(bh);
+
+	/*
+	 * Buffer is clean and the IO has finished (we held the buffer
+	 * lock) so the checkpoint is done. We can safely remove the
+	 * buffer from this transaction.
+	 */
+	JBUFFER_TRACE(jh, "remove from checkpoint list");
+	return __jbd2_journal_remove_checkpoint(jh);
+}
+
 /*
  * journal_insert_checkpoint: put a committed buffer onto a checkpoint
  * list so that we know when it is safe to clean the transaction out of
diff --git a/fs/jbd2/transaction.c b/fs/jbd2/transaction.c
index ce4a5ccadeff4..62e68c5b8ec3d 100644
--- a/fs/jbd2/transaction.c
+++ b/fs/jbd2/transaction.c
@@ -1775,8 +1775,7 @@ int jbd2_journal_forget(handle_t *handle, struct buffer_head *bh)
 		 * Otherwise, if the buffer has been written to disk,
 		 * it is safe to remove the checkpoint and drop it.
 		 */
-		if (!buffer_dirty(bh)) {
-			__jbd2_journal_remove_checkpoint(jh);
+		if (jbd2_journal_try_remove_checkpoint(jh) >= 0) {
 			spin_unlock(&journal->j_list_lock);
 			goto drop;
 		}
@@ -2103,20 +2102,14 @@ __journal_try_to_free_buffer(journal_t *journal, struct buffer_head *bh)
 
 	jh = bh2jh(bh);
 
-	if (buffer_locked(bh) || buffer_dirty(bh))
-		goto out;
-
 	if (jh->b_next_transaction != NULL || jh->b_transaction != NULL)
-		goto out;
+		return;
 
 	spin_lock(&journal->j_list_lock);
-	if (jh->b_cp_transaction != NULL) {
-		/* written-back checkpointed metadata buffer */
-		JBUFFER_TRACE(jh, "remove from checkpoint list");
-		__jbd2_journal_remove_checkpoint(jh);
-	}
+	/* Remove written-back checkpointed metadata buffer */
+	if (jh->b_cp_transaction != NULL)
+		jbd2_journal_try_remove_checkpoint(jh);
 	spin_unlock(&journal->j_list_lock);
-out:
 	return;
 }
 
diff --git a/include/linux/jbd2.h b/include/linux/jbd2.h
index e6cfbcde96f29..ade8a6d7acff9 100644
--- a/include/linux/jbd2.h
+++ b/include/linux/jbd2.h
@@ -1441,6 +1441,7 @@ extern void jbd2_journal_commit_transaction(journal_t *);
 void __jbd2_journal_clean_checkpoint_list(journal_t *journal, bool destroy);
 unsigned long jbd2_journal_shrink_checkpoint_list(journal_t *journal, unsigned long *nr_to_scan);
 int __jbd2_journal_remove_checkpoint(struct journal_head *);
+int jbd2_journal_try_remove_checkpoint(struct journal_head *jh);
 void jbd2_journal_destroy_checkpoint(journal_t *journal);
 void __jbd2_journal_insert_checkpoint(struct journal_head *, transaction_t *);
 
-- 
2.40.1




  parent reply	other threads:[~2023-08-28 10:44 UTC|newest]

Thread overview: 102+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2023-08-28 10:13 [PATCH 5.15 00/89] 5.15.129-rc1 review Greg Kroah-Hartman
2023-08-28 10:13 ` [PATCH 5.15 01/89] objtool/x86: Fix SRSO mess Greg Kroah-Hartman
2023-08-28 10:13 ` [PATCH 5.15 02/89] NFSv4.2: fix error handling in nfs42_proc_getxattr Greg Kroah-Hartman
2023-08-28 10:13 ` [PATCH 5.15 03/89] NFSv4: fix out path in __nfs4_get_acl_uncached Greg Kroah-Hartman
2023-08-28 10:13 ` [PATCH 5.15 04/89] xprtrdma: Remap Receive buffers after a reconnect Greg Kroah-Hartman
2023-08-28 10:13 ` [PATCH 5.15 05/89] PCI: acpiphp: Reassign resources on bridge if necessary Greg Kroah-Hartman
2023-08-28 10:13 ` [PATCH 5.15 06/89] dlm: improve plock logging if interrupted Greg Kroah-Hartman
2023-08-28 10:13 ` [PATCH 5.15 07/89] dlm: replace usage of found with dedicated list iterator variable Greg Kroah-Hartman
2023-08-28 10:13 ` [PATCH 5.15 08/89] fs: dlm: add pid to debug log Greg Kroah-Hartman
2023-08-28 10:13 ` [PATCH 5.15 09/89] fs: dlm: change plock interrupted message to debug again Greg Kroah-Hartman
2023-08-28 10:13 ` [PATCH 5.15 10/89] fs: dlm: use dlm_plock_info for do_unlock_close Greg Kroah-Hartman
2023-08-28 10:13 ` [PATCH 5.15 11/89] fs: dlm: fix mismatch of plock results from userspace Greg Kroah-Hartman
2023-08-28 10:13 ` [PATCH 5.15 12/89] MIPS: cpu-features: Enable octeon_cache by cpu_type Greg Kroah-Hartman
2023-08-28 10:13 ` [PATCH 5.15 13/89] MIPS: cpu-features: Use boot_cpu_type for CPU type based features Greg Kroah-Hartman
2023-08-28 10:13 ` [PATCH 5.15 14/89] fbdev: Improve performance of sys_imageblit() Greg Kroah-Hartman
2023-08-28 10:13 ` [PATCH 5.15 15/89] fbdev: Fix sys_imageblit() for arbitrary image widths Greg Kroah-Hartman
2023-08-28 10:13 ` [PATCH 5.15 16/89] fbdev: fix potential OOB read in fast_imageblit() Greg Kroah-Hartman
2023-08-28 10:13 ` [PATCH 5.15 17/89] ALSA: pcm: Fix potential data race at PCM memory allocation helpers Greg Kroah-Hartman
2023-08-28 10:13 ` [PATCH 5.15 18/89] jbd2: remove t_checkpoint_io_list Greg Kroah-Hartman
2023-08-28 10:13 ` [PATCH 5.15 19/89] jbd2: remove journal_clean_one_cp_list() Greg Kroah-Hartman
2023-08-28 10:13 ` Greg Kroah-Hartman [this message]
2023-08-28 10:13 ` [PATCH 5.15 21/89] can: raw: fix receiver memory leak Greg Kroah-Hartman
2023-08-28 10:13 ` [PATCH 5.15 22/89] drm/amd/display: do not wait for mpc idle if tg is disabled Greg Kroah-Hartman
2023-08-28 10:13 ` [PATCH 5.15 23/89] drm/amd/display: check TG is non-null before checking if enabled Greg Kroah-Hartman
2023-08-28 10:13 ` [PATCH 5.15 24/89] can: raw: fix lockdep issue in raw_release() Greg Kroah-Hartman
2023-08-28 10:13 ` [PATCH 5.15 25/89] tracing: Fix cpu buffers unavailable due to record_disabled missed Greg Kroah-Hartman
2023-08-28 10:13 ` [PATCH 5.15 26/89] tracing: Fix memleak due to race between current_tracer and trace Greg Kroah-Hartman
2023-08-28 10:13 ` [PATCH 5.15 27/89] octeontx2-af: SDP: fix receive link config Greg Kroah-Hartman
2023-08-28 10:13 ` [PATCH 5.15 28/89] sock: annotate data-races around prot->memory_pressure Greg Kroah-Hartman
2023-08-28 10:13 ` [PATCH 5.15 29/89] dccp: annotate data-races in dccp_poll() Greg Kroah-Hartman
2023-08-28 10:13 ` [PATCH 5.15 30/89] ipvlan: Fix a reference count leak warning in ipvlan_ns_exit() Greg Kroah-Hartman
2023-08-28 10:13 ` [PATCH 5.15 31/89] net: bgmac: Fix return value check for fixed_phy_register() Greg Kroah-Hartman
2023-08-28 10:13 ` [PATCH 5.15 32/89] net: bcmgenet: " Greg Kroah-Hartman
2023-08-28 10:13 ` [PATCH 5.15 33/89] net: validate veth and vxcan peer ifindexes Greg Kroah-Hartman
2023-08-28 10:13 ` [PATCH 5.15 34/89] ice: fix receive buffer size miscalculation Greg Kroah-Hartman
2023-08-28 10:13 ` [PATCH 5.15 35/89] igb: Avoid starting unnecessary workqueues Greg Kroah-Hartman
2023-08-28 10:13 ` [PATCH 5.15 36/89] igc: Fix the typo in the PTM Control macro Greg Kroah-Hartman
2023-08-28 10:13 ` [PATCH 5.15 37/89] net/sched: fix a qdisc modification with ambiguous command request Greg Kroah-Hartman
2023-08-28 10:13 ` [PATCH 5.15 38/89] netfilter: nf_tables: flush pending destroy work before netlink notifier Greg Kroah-Hartman
2023-08-28 10:13 ` [PATCH 5.15 39/89] netfilter: nf_tables: fix out of memory error handling Greg Kroah-Hartman
2023-08-28 10:13 ` [PATCH 5.15 40/89] rtnetlink: return ENODEV when ifname does not exist and group is given Greg Kroah-Hartman
2023-08-28 10:13 ` [PATCH 5.15 41/89] rtnetlink: Reject negative ifindexes in RTM_NEWLINK Greg Kroah-Hartman
2023-08-28 10:13 ` [PATCH 5.15 42/89] net: remove bond_slave_has_mac_rcu() Greg Kroah-Hartman
2023-08-28 10:13 ` [PATCH 5.15 43/89] bonding: fix macvlan over alb bond support Greg Kroah-Hartman
2023-08-28 10:13 ` [PATCH 5.15 44/89] net/ncsi: make one oem_gma function for all mfr id Greg Kroah-Hartman
2023-08-28 10:13 ` [PATCH 5.15 45/89] net/ncsi: change from ndo_set_mac_address to dev_set_mac_address Greg Kroah-Hartman
2023-08-28 10:13 ` [PATCH 5.15 46/89] Revert "KVM: x86: enable TDP MMU by default" Greg Kroah-Hartman
2023-08-28 10:13 ` [PATCH 5.15 47/89] ibmveth: Use dcbf rather than dcbfl Greg Kroah-Hartman
2023-08-28 10:13 ` [PATCH 5.15 48/89] NFSv4: Fix dropped lock for racing OPEN and delegation return Greg Kroah-Hartman
2023-08-28 10:13 ` [PATCH 5.15 49/89] clk: Fix slab-out-of-bounds error in devm_clk_release() Greg Kroah-Hartman
2023-08-28 10:13 ` [PATCH 5.15 50/89] ALSA: ymfpci: Fix the missing snd_card_free() call at probe error Greg Kroah-Hartman
2023-08-28 10:13 ` [PATCH 5.15 51/89] mm: add a call to flush_cache_vmap() in vmap_pfn() Greg Kroah-Hartman
2023-08-28 10:13 ` [PATCH 5.15 52/89] NFS: Fix a use after free in nfs_direct_join_group() Greg Kroah-Hartman
2023-08-28 10:13 ` [PATCH 5.15 53/89] nfsd: Fix race to FREE_STATEID and cl_revoked Greg Kroah-Hartman
2023-08-28 10:13 ` [PATCH 5.15 54/89] selinux: set next pointer before attaching to list Greg Kroah-Hartman
2023-08-28 10:13 ` [PATCH 5.15 55/89] batman-adv: Trigger events for auto adjusted MTU Greg Kroah-Hartman
2023-08-28 10:13 ` [PATCH 5.15 56/89] batman-adv: Dont increase MTU when set by user Greg Kroah-Hartman
2023-08-28 10:13 ` [PATCH 5.15 57/89] batman-adv: Do not get eth header before batadv_check_management_packet Greg Kroah-Hartman
2023-08-28 10:13 ` [PATCH 5.15 58/89] batman-adv: Fix TT global entry leak when client roamed back Greg Kroah-Hartman
2023-08-28 10:14 ` [PATCH 5.15 59/89] batman-adv: Fix batadv_v_ogm_aggr_send memory leak Greg Kroah-Hartman
2023-08-28 10:14 ` [PATCH 5.15 60/89] batman-adv: Hold rtnl lock during MTU update via netlink Greg Kroah-Hartman
2023-08-28 10:14 ` [PATCH 5.15 61/89] lib/clz_ctz.c: Fix __clzdi2() and __ctzdi2() for 32-bit kernels Greg Kroah-Hartman
2023-08-28 10:14 ` [PATCH 5.15 62/89] radix tree: remove unused variable Greg Kroah-Hartman
2023-08-28 10:14 ` [PATCH 5.15 63/89] of: unittest: Fix EXPECT for parse_phandle_with_args_map() test Greg Kroah-Hartman
2023-08-28 10:14 ` [PATCH 5.15 64/89] of: dynamic: Refactor action prints to not use "%pOF" inside devtree_lock Greg Kroah-Hartman
2023-08-28 10:14 ` [PATCH 5.15 65/89] media: vcodec: Fix potential array out-of-bounds in encoder queue_setup Greg Kroah-Hartman
2023-08-28 10:14 ` [PATCH 5.15 66/89] PCI: acpiphp: Use pci_assign_unassigned_bridge_resources() only for non-root bus Greg Kroah-Hartman
2023-08-28 10:14 ` [PATCH 5.15 67/89] drm/vmwgfx: Fix shader stage validation Greg Kroah-Hartman
2023-08-28 10:14 ` [PATCH 5.15 68/89] drm/display/dp: Fix the DP DSC Receiver cap size Greg Kroah-Hartman
2023-08-28 10:14 ` [PATCH 5.15 69/89] x86/fpu: Invalidate FPU state correctly on exec() Greg Kroah-Hartman
2023-08-28 10:14 ` [PATCH 5.15 70/89] x86/fpu: Set X86_FEATURE_OSXSAVE feature after enabling OSXSAVE in CR4 Greg Kroah-Hartman
2023-08-28 10:14 ` [PATCH 5.15 71/89] nfs: use vfs setgid helper Greg Kroah-Hartman
2023-08-28 10:14 ` [PATCH 5.15 72/89] nfsd: " Greg Kroah-Hartman
2023-08-28 10:14 ` [PATCH 5.15 73/89] torture: Fix hang during kthread shutdown phase Greg Kroah-Hartman
2023-08-28 10:14 ` [PATCH 5.15 74/89] cgroup/cpuset: Rename functions dealing with DEADLINE accounting Greg Kroah-Hartman
2023-08-28 10:14 ` [PATCH 5.15 75/89] sched/cpuset: Bring back cpuset_mutex Greg Kroah-Hartman
2023-08-28 10:14 ` [PATCH 5.15 76/89] sched/cpuset: Keep track of SCHED_DEADLINE task in cpusets Greg Kroah-Hartman
2023-08-28 10:14 ` [PATCH 5.15 77/89] cgroup/cpuset: Iterate only if DEADLINE tasks are present Greg Kroah-Hartman
2023-08-28 10:14 ` [PATCH 5.15 78/89] sched/deadline: Create DL BW alloc, free & check overflow interface Greg Kroah-Hartman
2023-08-28 10:14 ` [PATCH 5.15 79/89] cgroup/cpuset: Free DL BW in case can_attach() fails Greg Kroah-Hartman
2023-08-28 10:14 ` [PATCH 5.15 80/89] drm/i915: Fix premature release of requests reusable memory Greg Kroah-Hartman
2023-08-28 10:14 ` [PATCH 5.15 81/89] can: raw: add missing refcount for memory leak fix Greg Kroah-Hartman
2023-08-28 10:14 ` [PATCH 5.15 82/89] scsi: snic: Fix double free in snic_tgt_create() Greg Kroah-Hartman
2023-08-28 10:14 ` [PATCH 5.15 83/89] scsi: core: raid_class: Remove raid_component_add() Greg Kroah-Hartman
2023-08-28 10:14 ` [PATCH 5.15 84/89] clk: Fix undefined reference to `clk_rate_exclusive_{get,put} Greg Kroah-Hartman
2023-08-28 10:14 ` [PATCH 5.15 85/89] pinctrl: renesas: rza2: Add lock around pinctrl_generic{{add,remove}_group,{add,remove}_function} Greg Kroah-Hartman
2023-08-28 10:14 ` [PATCH 5.15 86/89] dma-buf/sw_sync: Avoid recursive lock during fence signal Greg Kroah-Hartman
2023-08-28 10:14 ` [PATCH 5.15 87/89] mm: memory-failure: kill soft_offline_free_page() Greg Kroah-Hartman
2023-08-28 10:14 ` [PATCH 5.15 88/89] mm: memory-failure: fix unexpected return value in soft_offline_page() Greg Kroah-Hartman
2023-08-28 10:14 ` [PATCH 5.15 89/89] mm,ima,kexec,of: use memblock_free_late from ima_free_kexec_buffer Greg Kroah-Hartman
2023-08-29  1:57 ` [PATCH 5.15 00/89] 5.15.129-rc1 review SeongJae Park
2023-08-29  9:05 ` Naresh Kamboju
2023-08-29  9:36 ` Naresh Kamboju
2023-08-29 10:00 ` Harshit Mogalapalli
2023-08-29 11:50 ` Sudip Mukherjee (Codethink)
2023-08-29 14:13 ` Shuah Khan
2023-08-29 18:52 ` Florian Fainelli
2023-08-29 23:26 ` Ron Economos
2023-08-30  2:24 ` Guenter Roeck
2023-08-30 10:24 ` Jon Hunter
2023-08-30 13:19 ` Joel Fernandes
2023-08-30 16:32 ` Allen Pais

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20230828101150.843521536@linuxfoundation.org \
    --to=gregkh@linuxfoundation.org \
    --cc=jack@suse.cz \
    --cc=patches@lists.linux.dev \
    --cc=sashal@kernel.org \
    --cc=stable@vger.kernel.org \
    --cc=tytso@mit.edu \
    --cc=yi.zhang@huawei.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).