linux-fsdevel.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Zhang Yi <yi.zhang@huaweicloud.com>
To: linux-ext4@vger.kernel.org
Cc: linux-fsdevel@vger.kernel.org, linux-kernel@vger.kernel.org,
	tytso@mit.edu, adilger.kernel@dilger.ca, jack@suse.cz,
	yi.zhang@huawei.com, yi.zhang@huaweicloud.com,
	yizhang089@gmail.com, libaokun1@huawei.com, yangerkun@huawei.com
Subject: [PATCH v2 07/13] ext4: drop extent cache before splitting extent
Date: Fri, 21 Nov 2025 14:08:05 +0800	[thread overview]
Message-ID: <20251121060811.1685783-8-yi.zhang@huaweicloud.com> (raw)
In-Reply-To: <20251121060811.1685783-1-yi.zhang@huaweicloud.com>

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

When splitting an unwritten extent in the middle and converting it to
initialized in ext4_split_extent() with the EXT4_EXT_MAY_ZEROOUT and
EXT4_EXT_DATA_VALID2 flags set, it could leave a stale unwritten extent.

Assume we have an unwritten file and buffered write in the middle of it
without dioread_nolock enabled, it will allocate blocks as written
extent.

       0  A      B  N
       [UUUUUUUUUUUU] on-disk extent      U: unwritten extent
       [UUUUUUUUUUUU] extent status tree
       [--DDDDDDDD--]                     D: valid data
          |<-  ->| ----> this range needs to be initialized

ext4_split_extent() first try to split this extent at B with
EXT4_EXT_DATA_PARTIAL_VALID1 and EXT4_EXT_MAY_ZEROOUT flag set, but
ext4_split_extent_at() failed to split this extent due to temporary lack
of space. It zeroout B to N and leave the entire extent as unwritten.

       0  A      B  N
       [UUUUUUUUUUUU] on-disk extent
       [UUUUUUUUUUUU] extent status tree
       [--DDDDDDDDZZ]                     Z: zeroed data

ext4_split_extent() then try to split this extent at A with
EXT4_EXT_DATA_VALID2 flag set. This time, it split successfully and
leave
an written extent from A to N.

       0  A      B   N
       [UU|WWWWWWWWWW] on-disk extent     W: written extent
       [UU|UUUUUUUUUU] extent status tree
       [--|DDDDDDDDZZ]

Finally ext4_map_create_blocks() only insert extent A to B to the extent
status tree, and leave an stale unwritten extent in the status tree.

       0  A      B   N
       [UU|WWWWWWWWWW] on-disk extent     W: written extent
       [UU|WWWWWWWWUU] extent status tree
       [--|DDDDDDDDZZ]

Fix this issue by always remove cached extent status entry before
splitting extent.

Signed-off-by: Zhang Yi <yi.zhang@huawei.com>
---
 fs/ext4/extents.c | 6 ++++++
 1 file changed, 6 insertions(+)

diff --git a/fs/ext4/extents.c b/fs/ext4/extents.c
index 2b5aec3f8882..9bb80af4b5cf 100644
--- a/fs/ext4/extents.c
+++ b/fs/ext4/extents.c
@@ -3367,6 +3367,12 @@ static struct ext4_ext_path *ext4_split_extent(handle_t *handle,
 	ee_len = ext4_ext_get_actual_len(ex);
 	unwritten = ext4_ext_is_unwritten(ex);
 
+	/*
+	 * Drop extent cache to prevent stale unwritten extents remaining
+	 * after zeroing out.
+	 */
+	ext4_es_remove_extent(inode, ee_block, ee_len);
+
 	/* Do not cache extents that are in the process of being modified. */
 	flags |= EXT4_EX_NOCACHE;
 
-- 
2.46.1


  parent reply	other threads:[~2025-11-21  6:10 UTC|newest]

Thread overview: 52+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2025-11-21  6:07 [PATCH v2 00/13] ext4: replace ext4_es_insert_extent() when caching on-disk extents Zhang Yi
2025-11-21  6:07 ` [PATCH v2 01/13] ext4: cleanup zeroout in ext4_split_extent_at() Zhang Yi
2025-11-26 11:26   ` Ojaswin Mujoo
2025-11-27 12:02   ` Jan Kara
2025-11-28  2:22     ` Zhang Yi
2025-11-28  1:54   ` Baokun Li
2025-11-21  6:08 ` [PATCH v2 02/13] ext4: subdivide EXT4_EXT_DATA_VALID1 Zhang Yi
2025-11-26 11:27   ` Ojaswin Mujoo
2025-11-26 11:55     ` Ojaswin Mujoo
2025-11-27  6:09       ` Zhang Yi
2025-11-21  6:08 ` [PATCH v2 03/13] ext4: don't zero the entire extent if EXT4_EXT_DATA_PARTIAL_VALID1 Zhang Yi
2025-11-26 11:29   ` Ojaswin Mujoo
2025-11-27  6:13     ` Zhang Yi
2025-11-27 13:41   ` Jan Kara
2025-11-28  3:45     ` Zhang Yi
2025-11-28 10:58       ` Jan Kara
2025-11-28  7:28     ` Ojaswin Mujoo
2025-11-28 11:14       ` Jan Kara
2025-11-28 14:20         ` Ojaswin Mujoo
2025-11-28 19:52         ` Andreas Dilger
2025-11-29 18:41           ` Ojaswin Mujoo
2025-11-21  6:08 ` [PATCH v2 04/13] ext4: don't set EXT4_GET_BLOCKS_CONVERT when splitting before submitting I/O Zhang Yi
2025-11-26 11:50   ` Ojaswin Mujoo
2025-11-21  6:08 ` [PATCH v2 05/13] ext4: correct the mapping status if the extent has been zeroed Zhang Yi
2025-11-26 11:56   ` Ojaswin Mujoo
2025-11-21  6:08 ` [PATCH v2 06/13] ext4: don't cache extent during splitting extent Zhang Yi
2025-11-26 12:04   ` Ojaswin Mujoo
2025-11-27  7:01     ` Zhang Yi
2025-11-28  8:18       ` Ojaswin Mujoo
2025-11-21  6:08 ` Zhang Yi [this message]
2025-11-26 12:24   ` [PATCH v2 07/13] ext4: drop extent cache before " Ojaswin Mujoo
2025-11-27  7:27     ` Zhang Yi
2025-11-28  8:16       ` Ojaswin Mujoo
2025-11-29  1:36         ` Zhang Yi
2025-11-21  6:08 ` [PATCH v2 08/13] ext4: cleanup useless out tag in __es_remove_extent() Zhang Yi
2025-11-27 12:43   ` Jan Kara
2025-11-28  8:20   ` Ojaswin Mujoo
2025-11-21  6:08 ` [PATCH v2 09/13] ext4: make __es_remove_extent() check extent status Zhang Yi
2025-11-21  6:08 ` [PATCH v2 10/13] ext4: make ext4_es_cache_extent() support overwrite existing extents Zhang Yi
2025-11-21  6:08 ` [PATCH v2 11/13] ext4: adjust the debug info in ext4_es_cache_extent() Zhang Yi
2025-11-21  6:08 ` [PATCH v2 12/13] ext4: replace ext4_es_insert_extent() when caching on-disk extents Zhang Yi
2025-11-21  6:08 ` [PATCH v2 13/13] ext4: drop the TODO comment in ext4_es_insert_extent() Zhang Yi
2025-11-23 10:55 ` [PATCH v2 00/13] ext4: replace ext4_es_insert_extent() when caching on-disk extents Ojaswin Mujoo
2025-11-24  5:04   ` Zhang Yi
2025-11-24 12:50     ` Ojaswin Mujoo
2025-11-24 14:05       ` Zhang Yi
2025-11-27 12:24     ` Jan Kara
2025-11-28  4:37       ` Zhang Yi
2025-11-28 16:49 ` Theodore Tso
2025-11-29  1:32   ` Zhang Yi
2025-11-29  3:52     ` Theodore Tso
2025-11-29  4:44       ` Zhang Yi

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=20251121060811.1685783-8-yi.zhang@huaweicloud.com \
    --to=yi.zhang@huaweicloud.com \
    --cc=adilger.kernel@dilger.ca \
    --cc=jack@suse.cz \
    --cc=libaokun1@huawei.com \
    --cc=linux-ext4@vger.kernel.org \
    --cc=linux-fsdevel@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=tytso@mit.edu \
    --cc=yangerkun@huawei.com \
    --cc=yi.zhang@huawei.com \
    --cc=yizhang089@gmail.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).