linux-kernel - Re: [PATCH] scsi: core: Fix block I/O error of USB card reader during resume

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: <ceba3c96-5f32-9b5a-bacc-0eb942b9b90b@acm.org>
Date:   Wed, 17 Aug 2022 06:52:14 -0700
From:   Bart Van Assche <bvanassche@....org>
To:     Michael Wu <michael@...winnertech.com>, jejb@...ux.ibm.com,
        martin.petersen@...cle.com
Cc:     linux-scsi@...r.kernel.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH] scsi: core: Fix block I/O error of USB card reader during
 resume

On 8/17/22 01:34, Michael Wu wrote:
> When accessing storage device via an USB card reader, a block I/O error
> occurs during resume:
> 
> PM: suspend exit
> sd 0: 0:0:0: [sda] tag#0 UNKNOWN(0x2003) Result: hostbyte=0x00 driverbyte
> =0x08
> sd 0: 0:0:0: [sda] tag#0 Sense Key : 0x6 [current]
> sd 0: 0:0:0: [sda] tag#0 ASC=0x28 ASCQ=0x0
> sd 0: 0:0:0: [sda] tag#0 CDB: opcode=0x28 28 00 00 17 ce e1 00 00 f0 00
> blk_update_request: I/O error, dev sda, sector 1560289 op 0x0:(READ) flags
> 0x84700 phys_seg 19 prio class 0
> sd 0: 0:0:0: [sda] tag#0 device offline or changed
> 
> Fix it by changing the action in scsi_io_completion_action() from
> ACTION_FAIL to ACTION_RETRY by adding the condition `cmd->device->
> lockable`.
> 
> Signed-off-by: Michael Wu <michael@...winnertech.com>
> ---
>   drivers/scsi/scsi_lib.c | 3 ++-
>   1 file changed, 2 insertions(+), 1 deletion(-)
> 
> diff --git a/drivers/scsi/scsi_lib.c b/drivers/scsi/scsi_lib.c
> index 4dbd29ab1dcc..4bc480721947 100644
> --- a/drivers/scsi/scsi_lib.c
> +++ b/drivers/scsi/scsi_lib.c
> @@ -704,7 +704,8 @@ static void scsi_io_completion_action(struct scsi_cmnd *cmd, int result)
>   	} else if (sense_valid && sense_current) {
>   		switch (sshdr.sense_key) {
>   		case UNIT_ATTENTION:
> -			if (cmd->device->removable) {
> +			if (cmd->device->removable &&
> +			    cmd->device->lockable) {
>   				/* Detected disc change.  Set a bit
>   				 * and quietly refuse further access.
>   				 */

To me the above doesn't look like a good way to address this. I don't 
see why a device being lockable should control whether or not a unit 
attention results in a retry? Shouldn't the decision taken by 
scsi_io_completion_action() depend on the ASC and ASCQ codes rather than 
on whether a device is removable and/or lockable?

BTW, the code modified by the above patch is old. This is what I found 
in the 2002 version of scsi_lib.c:

if ((SCpnt->sense_buffer[0] & 0x7f) == 0x70
     && (SCpnt->sense_buffer[2] & 0xf) == UNIT_ATTENTION) {
         if (SCpnt->device->removable) {
                 /* detected disc change.  set a bit and quietly refuse
                  * further access.
                  */
                 SCpnt->device->changed = 1;
                 SCpnt = scsi_end_request(SCpnt, 0, this_count);
                 return;
         } else {
                 /*
                  * Must have been a power glitch, or a
                  * bus reset.  Could not have been a
                  * media change, so we just retry the
                  * request and see what happens.
                  */
                 scsi_queue_next_request(q, SCpnt);
                 return;
}

Bart.

Thanks,

Bart.