[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <8e69aa15-9779-4696-98ab-f173666a87a7@kernel.org>
Date: Mon, 19 Feb 2024 11:13:18 +0800
From: Chao Yu <chao@...nel.org>
To: Jaegeuk Kim <jaegeuk@...nel.org>
Cc: linux-f2fs-devel@...ts.sourceforge.net, linux-kernel@...r.kernel.org
Subject: Re: [PATCH] f2fs: fix to return EIO when reading after device removal
On 2024/2/8 8:18, Jaegeuk Kim wrote:
> On 02/06, Chao Yu wrote:
>> generic/730 2s ... - output mismatch (see /media/fstests/results//generic/730.out.bad)
>> --- tests/generic/730.out 2023-08-07 01:39:51.055568499 +0000
>> +++ /media/fstests/results//generic/730.out.bad 2024-02-06 02:26:43.000000000 +0000
>> @@ -1,2 +1 @@
>> QA output created by 730
>> -cat: -: Input/output error
>> ...
>> (Run 'diff -u /media/fstests/tests/generic/730.out /media/fstests/results//generic/730.out.bad' to see the entire diff)
>> Ran: generic/730
>> Failures: generic/730
>> Failed 1 of 1 tests
>>
>> This patch adds a check condition in f2fs_file_read_iter() to
>> detect cp_error status after device removal, and retrurn -EIO
>> for such case.
>
> Can we check device removal?
We can use blk_queue_dying() to detect device removal, but, IMO, device
removal can almost not happen in Android, not sure for distros or server,
do you want to add this check condition into f2fs_cp_error()?
Thanks,
>
>>
>> Signed-off-by: Chao Yu <chao@...nel.org>
>> ---
>> fs/f2fs/file.c | 3 +++
>> 1 file changed, 3 insertions(+)
>>
>> diff --git a/fs/f2fs/file.c b/fs/f2fs/file.c
>> index 45b7e3610b0f..9e4386d4144c 100644
>> --- a/fs/f2fs/file.c
>> +++ b/fs/f2fs/file.c
>> @@ -4462,6 +4462,9 @@ static ssize_t f2fs_file_read_iter(struct kiocb *iocb, struct iov_iter *to)
>> const loff_t pos = iocb->ki_pos;
>> ssize_t ret;
>>
>> + if (unlikely(f2fs_cp_error(F2FS_I_SB(inode))))
>> + return -EIO;
>> +
>> if (!f2fs_is_compress_backend_ready(inode))
>> return -EOPNOTSUPP;
>>
>> --
>> 2.40.1
Powered by blists - more mailing lists