From: Baokun Li <libaokun1@huawei.com>
To: Jan Kara <jack@suse.cz>, Matthew Wilcox <willy@infradead.org>,
Theodore Ts'o <tytso@mit.edu>
Cc: <linux-ext4@vger.kernel.org>, <adilger.kernel@dilger.ca>,
<ritesh.list@gmail.com>, <linux-kernel@vger.kernel.org>,
<jun.nie@linaro.org>, <ebiggers@kernel.org>,
<yi.zhang@huawei.com>, <yangerkun@huawei.com>,
<yukuai3@huawei.com>,
<syzbot+a158d886ca08a3fecca4@syzkaller.appspotmail.com>,
Baokun Li <libaokun1@huawei.com>
Subject: Re: [PATCH v2] ext4: fix race condition between buffer write and page_mkwrite
Date: Mon, 15 Apr 2024 12:28:01 +0800 [thread overview]
Message-ID: <49d5b109-7cc3-6717-b3c6-6858310aa3ba@huawei.com> (raw)
In-Reply-To: <20230605150855.7oaiplp7r57dcww3@quack3>
On 2023/6/5 23:08, Jan Kara wrote:
> On Mon 05-06-23 15:55:35, Matthew Wilcox wrote:
>> On Mon, Jun 05, 2023 at 02:21:41PM +0200, Jan Kara wrote:
>>> On Mon 05-06-23 11:16:55, Jan Kara wrote:
>>>> Yeah, I agree, that is also the conclusion I have arrived at when thinking
>>>> about this problem now. We should be able to just remove the conversion
>>>> from ext4_page_mkwrite() and rely on write(2) or truncate(2) doing it when
>>>> growing i_size.
>>> OK, thinking more about this and searching through the history, I've
>>> realized why the conversion is originally in ext4_page_mkwrite(). The
>>> problem is described in commit 7b4cc9787fe35b ("ext4: evict inline data
>>> when writing to memory map") but essentially it boils down to the fact that
>>> ext4 writeback code does not expect dirty page for a file with inline data
>>> because ext4_write_inline_data_end() should have copied the data into the
>>> inode and cleared the folio's dirty flag.
>>>
>>> Indeed messing with xattrs from the writeback path to copy page contents
>>> into inline data xattr would be ... interesting. Hum, out of good ideas for
>>> now :-|.
>> Is it so bad? Now that we don't have writepage in ext4, only
>> writepages, it seems like we have a considerably more benign locking
>> environment to work in.
> Well, yes, without ->writepage() it might be *possible*. But still rather
> ugly. The problem is that in ->writepages() i_size is not stable. Thus also
> whether the inode data is inline or not is not stable. We'd need inode_lock
> for that but that is not easily doable in the writeback path - inode lock
> would then become fs_reclaim unsafe...
>
> Honza
Hi Honza!
Hi Ted!
Hi Matthew!
Long time later came back to this, because while discussing another similar
ABBA problem with Hou Tao, he mentioned VM_FAULT_RETRY, and then I
thought that this could be used to solve this problem as well.
The general idea is that if we see a file with inline data in
ext4_page_mkwrite(),
we release the mmap_lock and grab the inode_lock to convert the inline data,
and then return VM_FAULT_RETRY to retry to get the mmap_lock.
The code implementation is as follows, do you have any thoughts?
diff --git a/fs/ext4/inode.c b/fs/ext4/inode.c
index 537803250ca9..e044c11c9cf6 100644
--- a/fs/ext4/inode.c
+++ b/fs/ext4/inode.c
@@ -6056,15 +6056,36 @@ vm_fault_t ext4_page_mkwrite(struct vm_fault *vmf)
if (unlikely(IS_IMMUTABLE(inode)))
return VM_FAULT_SIGBUS;
+ /*
+ * The ext4 writeback code does not expect dirty page for a file with
+ * inline data, so inline data needs to be converted. But it needs to
+ * hold the inode_lock when converting, and trying to lock the inode
+ * while holding the mmap_lock may result in an ABBA deadlock. So here
+ * we release the mmap_lock for conversion and retry after conversion.
+ * Only one retry is allowed to avoid endless loops.
+ * Acquire xattr_sem to avoid race with inline data conversion.
+ */
+ down_read(&EXT4_I(inode)->xattr_sem);
+ if (ext4_has_inline_data(inode)) {
+ up_read(&EXT4_I(inode)->xattr_sem);
+
+ if (!fault_flag_allow_retry_first(vmf->flags))
+ return VM_FAULT_SIGBUS;
+
+ release_fault_lock(vmf);
+
+ inode_lock(inode);
+ ext4_convert_inline_data(inode);
+ inode_unlock(inode);
+ return VM_FAULT_RETRY;
+ }
+ up_read(&EXT4_I(inode)->xattr_sem);
+
sb_start_pagefault(inode->i_sb);
file_update_time(vma->vm_file);
filemap_invalidate_lock_shared(mapping);
- err = ext4_convert_inline_data(inode);
- if (err)
- goto out_ret;
-
/*
* On data journalling we skip straight to the transaction handle:
* there's no delalloc; page truncated will be checked later; the
--
With Best Regards,
Baokun Li
.
next prev parent reply other threads:[~2024-04-15 4:28 UTC|newest]
Thread overview: 14+ messages / expand[flat|nested] mbox.gz Atom feed top
2023-05-30 13:44 [PATCH v2] ext4: fix race condition between buffer write and page_mkwrite Baokun Li
2023-05-31 13:15 ` Jan Kara
2023-06-04 3:04 ` Theodore Ts'o
2023-06-04 21:08 ` Theodore Ts'o
2023-06-05 1:58 ` Matthew Wilcox
2023-06-05 9:16 ` Jan Kara
2023-06-05 12:21 ` Jan Kara
2023-06-05 14:55 ` Matthew Wilcox
2023-06-05 15:08 ` Jan Kara
2023-06-06 12:19 ` Baokun Li
2024-04-15 4:28 ` Baokun Li [this message]
2024-04-15 12:34 ` Jan Kara
2024-04-15 14:07 ` Baokun Li
2023-06-05 2:17 ` Baokun Li
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=49d5b109-7cc3-6717-b3c6-6858310aa3ba@huawei.com \
--to=libaokun1@huawei.com \
--cc=adilger.kernel@dilger.ca \
--cc=ebiggers@kernel.org \
--cc=jack@suse.cz \
--cc=jun.nie@linaro.org \
--cc=linux-ext4@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=ritesh.list@gmail.com \
--cc=syzbot+a158d886ca08a3fecca4@syzkaller.appspotmail.com \
--cc=tytso@mit.edu \
--cc=willy@infradead.org \
--cc=yangerkun@huawei.com \
--cc=yi.zhang@huawei.com \
--cc=yukuai3@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).