Linux SCSI subsystem development
 help / color / mirror / Atom feed
From: sashiko-bot@kernel.org
To: "Niklas Cassel" <cassel@kernel.org>
Cc: linux-scsi@vger.kernel.org
Subject: Re: [PATCH v3 1/6] scsi: scsi_debug: Report the residual of a write
Date: Thu, 17 Sep 2026 13:04:15 +0000	[thread overview]
Message-ID: <20260917130415.9A2A21F000FF@smtp.kernel.org> (raw)
In-Reply-To: <20260917125445.1376493-9-cassel@kernel.org>

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [High] scsi_debug: unsigned underflow in resp_write_scat() residual calculation
--

commit a487bfccdf1b47222ab79a838751982d68882fc7
Author: Niklas Cassel <cassel@kernel.org>
Subject: scsi: scsi_debug: Report the residual of a write

This patch updates the write paths in scsi_debug to report residual
bytes when a buffer provided by the initiator is larger than the logical
blocks consumed. The residual is now reported in resp_write_dt0(),
resp_write_scat(), and resp_atomic_write().

> diff --git a/drivers/scsi/scsi_debug.c b/drivers/scsi/scsi_debug.c
> index 6941809dfdb75..641dd6f937912 100644
> --- a/drivers/scsi/scsi_debug.c
> +++ b/drivers/scsi/scsi_debug.c
[ ... ]
> @@ -5352,6 +5354,7 @@ static int resp_write_scat(struct scsi_cmnd *scp,
>  		sg_off += num_by;
>  		cum_lb += num;
>  	}
> +	scsi_set_resid(scp, scsi_bufflen(scp) - sg_off);

[Severity: High]
Can this calculation result in an unsigned underflow if the initiator
provides a buffer shorter than what the CDB descriptors request?

In resp_write_scat(), sg_off tracks the data offset. For each LBA range
descriptor, sg_off is unconditionally incremented by num_by, which is the
size requested by the descriptor. This happens even if do_device_access()
copied less data due to hitting the end of the data-out buffer.

If the initiator provides a shorter buffer (scsi_bufflen) than what the
CDB descriptors request, sg_off will eventually exceed the buffer length.
The subtraction here would then underflow to a massive positive number.

The SCSI midlayer uses this residual to compute good_bytes in
scsi_io_completion(). Subtracting a massive residual from the expected
transfer length could cause good_bytes to wrap and become negative, which
might severely confuse the block layer I/O completion logic.

>  	ret = 0;
>  err_out_unlock:
>  	sdeb_meta_write_unlock(sip);

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260917125445.1376493-8-cassel@kernel.org?part=1

  reply	other threads:[~2026-09-17 13:04 UTC|newest]

Thread overview: 18+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-09-17 12:54 [PATCH v3 0/6] scsi: scsi_debug: fix zoned write validation Niklas Cassel
2026-09-17 12:54 ` [PATCH v3 1/6] scsi: scsi_debug: Report the residual of a write Niklas Cassel
2026-09-17 13:04   ` sashiko-bot [this message]
2026-09-17 13:06   ` Damien Le Moal
2026-09-17 12:54 ` [PATCH v3 2/6] scsi: scsi_debug: Do not write a partial physical block Niklas Cassel
2026-09-17 13:11   ` Damien Le Moal
2026-09-17 13:47     ` Niklas Cassel
2026-09-17 12:54 ` [PATCH v3 3/6] scsi: scsi_debug: Advance the write pointer over the data written Niklas Cassel
2026-09-17 13:13   ` Damien Le Moal
2026-09-17 12:54 ` [PATCH v3 4/6] scsi: scsi_debug: Enforce physical block alignment of zoned writes Niklas Cassel
2026-09-17 13:09   ` sashiko-bot
2026-09-17 12:54 ` [PATCH v3 5/6] scsi: scsi_debug: Map the region written by WRITE ATOMIC (16) Niklas Cassel
2026-09-17 13:12   ` sashiko-bot
2026-09-17 14:01     ` Niklas Cassel
2026-09-17 13:14   ` Damien Le Moal
2026-09-17 12:54 ` [PATCH v3 6/6] scsi: scsi_debug: Validate zone access for " Niklas Cassel
2026-09-17 13:15   ` Damien Le Moal
2026-09-17 13:05 ` [PATCH v3 0/6] scsi: scsi_debug: fix zoned write validation Damien Le Moal

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=20260917130415.9A2A21F000FF@smtp.kernel.org \
    --to=sashiko-bot@kernel.org \
    --cc=cassel@kernel.org \
    --cc=linux-scsi@vger.kernel.org \
    --cc=sashiko-reviews@lists.linux.dev \
    /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