[lvc-project] [PATCH 5.10.y] scsi: ufs: core: Improve SCSI abort handling

[Vasiliy Kovalev]

Hi Sasha,

On 4/21/26 16:19, Vasiliy Kovalev wrote:
> From: Bart Van Assche <bvanassche at acm.org>
> 
> commit 3ff1f6b6ba6f97f50862aa50e79959cc8ddc2566 upstream.
> 
> The following has been observed on a test setup:
> 
> WARNING: CPU: 4 PID: 250 at drivers/scsi/ufs/ufshcd.c:2737 ufshcd_queuecommand+0x468/0x65c
> Call trace:
>   ufshcd_queuecommand+0x468/0x65c
>   scsi_send_eh_cmnd+0x224/0x6a0
>   scsi_eh_test_devices+0x248/0x418
>   scsi_eh_ready_devs+0xc34/0xe58
>   scsi_error_handler+0x204/0x80c
>   kthread+0x150/0x1b4
>   ret_from_fork+0x10/0x30
> 
> That warning is triggered by the following statement:
> 
> 	WARN_ON(lrbp->cmd);
> 
> Fix this warning by clearing lrbp->cmd from the abort handler.
> 
> Link: https://lore.kernel.org/r/20211104181059.4129537-1-bvanassche@acm.org
> Fixes: 7a3e97b0dc4b ("[SCSI] ufshcd: UFS Host controller driver")
> Reviewed-by: Bean Huo <beanhuo at micron.com>
> Reviewed-by: Stanley Chu <stanley.chu at mediatek.com>
> Signed-off-by: Bart Van Assche <bvanassche at acm.org>
> Signed-off-by: Martin K. Petersen <martin.petersen at oracle.com>
> [ kovalev: bp to fix CVE-2021-47188; adapted placement of
>    lrbp->cmd = NULL for 5.10 function structure ]

Please drop this backport from the 5.10 queue — it is not needed.

After review feedback from Fedor Pchelkin, we verified that 5.10 is not
affected by this bug. The upstream commit 3ff1f6b6ba6f carries an
incorrect Fixes tag:

Fixes: 7a3e97b0dc4b ("[SCSI] ufshcd: UFS Host controller driver")

The actual regression was introduced by:

64180742605f ("scsi: ufs: Fix the SCSI abort handler")   [v5.15-rc1]

which restructured ufshcd_abort() and removed the 
__ufshcd_transfer_req_compl()
call from the successful abort path. Before that commit — and in 5.10 to
this day — __ufshcd_transfer_req_compl() is always called on the successful
path via the cleanup: label, and it clears lrbp->cmd. So the 
WARN_ON(lrbp->cmd)
in ufshcd_queuecommand() cannot trigger on 5.10, and the lrbp->cmd = NULL;
added by this patch would be dead code there.

64180742605f is not present in 5.10.y, therefore CVE-2021-47188 does not 
apply to 5.10.y.

Sorry for the noise.

> Signed-off-by: Vasiliy Kovalev <kovalev at altlinux.org>
> ---
>   drivers/scsi/ufs/ufshcd.c | 1 +
>   1 file changed, 1 insertion(+)
> 
> diff --git a/drivers/scsi/ufs/ufshcd.c b/drivers/scsi/ufs/ufshcd.c
> index c7bf0e6bc303..1b8072f47e7e 100644
> --- a/drivers/scsi/ufs/ufshcd.c
> +++ b/drivers/scsi/ufs/ufshcd.c
> @@ -6788,6 +6788,7 @@ static int ufshcd_abort(struct scsi_cmnd *cmd)
>   		__ufshcd_transfer_req_compl(hba, (1UL << tag));
>   		spin_unlock_irqrestore(host->host_lock, flags);
>   out:
> +		lrbp->cmd = NULL;
>   		err = SUCCESS;
>   	} else {
>   		dev_err(hba->dev, "%s: failed with err %d\n", __func__, err);

-- 
Thanks,
Vasiliy