linux-kernel - Re: [PATCH V3 5/6] f2fs: ignore discard return value

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <820ffbc8-56cb-4f47-9112-2f4a79524025@nvidia.com>
Date: Wed, 26 Nov 2025 03:37:19 +0000
From: Chaitanya Kulkarni <chaitanyak@...dia.com>
To: Chao Yu <chao@...nel.org>
CC: "linux-block@...r.kernel.org" <linux-block@...r.kernel.org>,
	"jaegeuk@...nel.org" <jaegeuk@...nel.org>, Chaitanya Kulkarni
	<chaitanyak@...dia.com>, "hch@....de" <hch@....de>, "song@...nel.org"
	<song@...nel.org>, "axboe@...nel.dk" <axboe@...nel.dk>,
	"linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
	"dm-devel@...ts.linux.dev" <dm-devel@...ts.linux.dev>,
	"linux-raid@...r.kernel.org" <linux-raid@...r.kernel.org>,
	"linux-nvme@...ts.infradead.org" <linux-nvme@...ts.infradead.org>,
	"linux-f2fs-devel@...ts.sourceforge.net"
	<linux-f2fs-devel@...ts.sourceforge.net>, "cem@...nel.org" <cem@...nel.org>,
	"linux-xfs@...r.kernel.org" <linux-xfs@...r.kernel.org>, "sagi@...mberg.me"
	<sagi@...mberg.me>, "yukuai@...as.com" <yukuai@...as.com>,
	"bpf@...r.kernel.org" <bpf@...r.kernel.org>, "mpatocka@...hat.com"
	<mpatocka@...hat.com>, "Martin K . Petersen" <martin.petersen@...cle.com>,
	"agk@...hat.com" <agk@...hat.com>, Johannes Thumshirn
	<johannes.thumshirn@....com>, Chaitanya Kulkarni <ckulkarnilinux@...il.com>,
	"snitzer@...nel.org" <snitzer@...nel.org>
Subject: Re: [PATCH V3 5/6] f2fs: ignore discard return value

On 11/25/25 18:47, Chao Yu wrote:
> On 11/25/25 07:48, Chaitanya Kulkarni wrote:
>> __blkdev_issue_discard() always returns 0, making the error assignment
>> in __submit_discard_cmd() dead code.
>>
>> Initialize err to 0 and remove the error assignment from the
>> __blkdev_issue_discard() call to err. Move fault injection code into
>> already present if branch where err is set to -EIO.
>>
>> This preserves the fault injection behavior while removing dead error
>> handling.
>>
>> Reviewed-by: Martin K. Petersen <martin.petersen@...cle.com>
>> Reviewed-by: Johannes Thumshirn <johannes.thumshirn@....com>
>> Reviewed-by: Christoph Hellwig <hch@....de>
>> Signed-off-by: Chaitanya Kulkarni <ckulkarnilinux@...il.com>
>> ---
>>   fs/f2fs/segment.c | 10 +++-------
>>   1 file changed, 3 insertions(+), 7 deletions(-)
>>
>> diff --git a/fs/f2fs/segment.c b/fs/f2fs/segment.c
>> index b45eace879d7..22b736ec9c51 100644
>> --- a/fs/f2fs/segment.c
>> +++ b/fs/f2fs/segment.c
>> @@ -1343,15 +1343,9 @@ static int __submit_discard_cmd(struct f2fs_sb_info *sbi,
>>   
>>   		dc->di.len += len;
>>   
>> +		err = 0;
>>   		if (time_to_inject(sbi, FAULT_DISCARD)) {
>>   			err = -EIO;
>> -		} else {
>> -			err = __blkdev_issue_discard(bdev,
>> -					SECTOR_FROM_BLOCK(start),
>> -					SECTOR_FROM_BLOCK(len),
>> -					GFP_NOFS, &bio);
>> -		}
>> -		if (err) {
>>   			spin_lock_irqsave(&dc->lock, flags);
>>   			if (dc->state == D_PARTIAL)
>>   				dc->state = D_SUBMIT;
>> @@ -1360,6 +1354,8 @@ static int __submit_discard_cmd(struct f2fs_sb_info *sbi,
>>   			break;
>>   		}
>>   
>> +		__blkdev_issue_discard(bdev, SECTOR_FROM_BLOCK(start),
>> +				SECTOR_FROM_BLOCK(len), GFP_NOFS, &bio);
> Oh, wait, bio can be NULL? Then below f2fs_bug_on() will trigger panic or warning.
>
> Thanks,

That will happen without this patch also or not ?

Since __blkdev_issue_discard() is always returning 0 irrespective of bio
is null or not.

The following condition in original code will only execute when err is set to
-EIO and that will only happen when time_to_inject() -> true.
Original call to __blkdev_issue_discard() without this patch will always
return 0 even for bio == NULL after __blkdev_issue_discard().

This is what we are trying to fix so caller should not rely on
__blkdev_issue_discard() return value  :-

354                 if (err) {
1355                         spin_lock_irqsave(&dc->lock, flags);
1356                         if (dc->state == D_PARTIAL)
1357                                 dc->state = D_SUBMIT;
1358                         spin_unlock_irqrestore(&dc->lock, flags);
1359
1360                         break;
1361                 }

which will lead f2fs_bug_on() for bio == NULL even without this patch.

This patch is not changing exiting behavior, correct me if I'm wrong.


>
>>   		f2fs_bug_on(sbi, !bio);
>>   
>>   		/*

-ck