[PATCH] scsi: scsi_debug: Fix a warning in resp_write_scat()

[PATCH] scsi: scsi_debug: Fix a warning in resp_write_scat()

[Harshit Mogalapalli]

As 'lbdof_blen' is coming from user, if the size in kzalloc()
is >= MAX_ORDER then we hit a warning.

Call trace:

sg_ioctl
 sg_ioctl_common
   scsi_ioctl
    sg_scsi_ioctl
     blk_execute_rq
      blk_mq_sched_insert_request
       blk_mq_run_hw_queue
        __blk_mq_delay_run_hw_queue
         __blk_mq_run_hw_queue
          blk_mq_sched_dispatch_requests
           __blk_mq_sched_dispatch_requests
            blk_mq_dispatch_rq_list
             scsi_queue_rq
              scsi_dispatch_cmd
               scsi_debug_queuecommand
                schedule_resp
                 resp_write_scat

If you try to allocate a memory larger than(>=) MAX_ORDER, then kmalloc()
will definitely fail.  It creates a stack trace and messes up dmesg.
The user controls the size here so if they specify a too large size it
will fail.

Add __GFP_NOWARN in order to avoid too large allocation warning.
This is detected by static analysis using smatch.

Fixes: 481b5e5c7949 ("scsi: scsi_debug: add resp_write_scat function")
Signed-off-by: Harshit Mogalapalli <harshit.m.mogalapalli@oracle.com>
---
 drivers/scsi/scsi_debug.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/scsi/scsi_debug.c b/drivers/scsi/scsi_debug.c
index 697fc57bc711..273224d29ce9 100644
--- a/drivers/scsi/scsi_debug.c
+++ b/drivers/scsi/scsi_debug.c
@@ -3778,7 +3778,7 @@ static int resp_write_scat(struct scsi_cmnd *scp,
 		mk_sense_buffer(scp, ILLEGAL_REQUEST, INVALID_FIELD_IN_CDB, 0);
 		return illegal_condition_result;
 	}
-	lrdp = kzalloc(lbdof_blen, GFP_ATOMIC);
+	lrdp = kzalloc(lbdof_blen, GFP_ATOMIC | __GFP_NOWARN);
 	if (lrdp == NULL)
 		return SCSI_MLQUEUE_HOST_BUSY;
 	if (sdebug_verbose)
-- 
2.38.1

Re: [PATCH] scsi: scsi_debug: Fix a warning in resp_write_scat()

[Douglas Gilbert]

On 2022-11-11 05:05, Harshit Mogalapalli wrote:
> As 'lbdof_blen' is coming from user, if the size in kzalloc()
> is >= MAX_ORDER then we hit a warning.
> 
> Call trace:
> 
> sg_ioctl
>   sg_ioctl_common
>     scsi_ioctl
>      sg_scsi_ioctl
>       blk_execute_rq
>        blk_mq_sched_insert_request
>         blk_mq_run_hw_queue
>          __blk_mq_delay_run_hw_queue
>           __blk_mq_run_hw_queue
>            blk_mq_sched_dispatch_requests
>             __blk_mq_sched_dispatch_requests
>              blk_mq_dispatch_rq_list
>               scsi_queue_rq
>                scsi_dispatch_cmd
>                 scsi_debug_queuecommand
>                  schedule_resp
>                   resp_write_scat
> 
> If you try to allocate a memory larger than(>=) MAX_ORDER, then kmalloc()
> will definitely fail.  It creates a stack trace and messes up dmesg.
> The user controls the size here so if they specify a too large size it
> will fail.
> 
> Add __GFP_NOWARN in order to avoid too large allocation warning.
> This is detected by static analysis using smatch.
> 
> Fixes: 481b5e5c7949 ("scsi: scsi_debug: add resp_write_scat function")
> Signed-off-by: Harshit Mogalapalli <harshit.m.mogalapalli@oracle.com>

Acked-by: Douglas Gilbert <dgilbert@interlog.com>

Thanks.

> ---
>   drivers/scsi/scsi_debug.c | 2 +-
>   1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/drivers/scsi/scsi_debug.c b/drivers/scsi/scsi_debug.c
> index 697fc57bc711..273224d29ce9 100644
> --- a/drivers/scsi/scsi_debug.c
> +++ b/drivers/scsi/scsi_debug.c
> @@ -3778,7 +3778,7 @@ static int resp_write_scat(struct scsi_cmnd *scp,
>   		mk_sense_buffer(scp, ILLEGAL_REQUEST, INVALID_FIELD_IN_CDB, 0);
>   		return illegal_condition_result;
>   	}
> -	lrdp = kzalloc(lbdof_blen, GFP_ATOMIC);
> +	lrdp = kzalloc(lbdof_blen, GFP_ATOMIC | __GFP_NOWARN);
>   	if (lrdp == NULL)
>   		return SCSI_MLQUEUE_HOST_BUSY;
>   	if (sdebug_verbose)

Re: [PATCH] scsi: scsi_debug: Fix a warning in resp_write_scat()

[Harshit Mogalapalli]

Hi Doug,

On 11/11/22 10:53 pm, Douglas Gilbert wrote:
> On 2022-11-11 05:05, Harshit Mogalapalli wrote:
>> As 'lbdof_blen' is coming from user, if the size in kzalloc()
>> is >= MAX_ORDER then we hit a warning.
>>
>> Call trace:
>>
>> sg_ioctl
>>   sg_ioctl_common
>>     scsi_ioctl
>>      sg_scsi_ioctl
>>       blk_execute_rq
>>        blk_mq_sched_insert_request
>>         blk_mq_run_hw_queue
>>          __blk_mq_delay_run_hw_queue
>>           __blk_mq_run_hw_queue
>>            blk_mq_sched_dispatch_requests
>>             __blk_mq_sched_dispatch_requests
>>              blk_mq_dispatch_rq_list
>>               scsi_queue_rq
>>                scsi_dispatch_cmd
>>                 scsi_debug_queuecommand
>>                  schedule_resp
>>                   resp_write_scat
>>
>> If you try to allocate a memory larger than(>=) MAX_ORDER, then kmalloc()
>> will definitely fail.  It creates a stack trace and messes up dmesg.
>> The user controls the size here so if they specify a too large size it
>> will fail.
>>
>> Add __GFP_NOWARN in order to avoid too large allocation warning.
>> This is detected by static analysis using smatch.
>>
>> Fixes: 481b5e5c7949 ("scsi: scsi_debug: add resp_write_scat function")
>> Signed-off-by: Harshit Mogalapalli <harshit.m.mogalapalli@oracle.com>
> 
> Acked-by: Douglas Gilbert <dgilbert@interlog.com>
> 

Thank you very much for checking the patch.

We have two more similar cases in the scsi_debug.c file in resp_verify() 
and resp_report_zones() functions. These are also detected using static 
analysis with smatch.

I have sent out patches for those.

Thanks,
Harshit

> Thanks.
> 
>> ---
>>   drivers/scsi/scsi_debug.c | 2 +-
>>   1 file changed, 1 insertion(+), 1 deletion(-)
>>
>> diff --git a/drivers/scsi/scsi_debug.c b/drivers/scsi/scsi_debug.c
>> index 697fc57bc711..273224d29ce9 100644
>> --- a/drivers/scsi/scsi_debug.c
>> +++ b/drivers/scsi/scsi_debug.c
>> @@ -3778,7 +3778,7 @@ static int resp_write_scat(struct scsi_cmnd *scp,
>>           mk_sense_buffer(scp, ILLEGAL_REQUEST, INVALID_FIELD_IN_CDB, 0);
>>           return illegal_condition_result;
>>       }
>> -    lrdp = kzalloc(lbdof_blen, GFP_ATOMIC);
>> +    lrdp = kzalloc(lbdof_blen, GFP_ATOMIC | __GFP_NOWARN);
>>       if (lrdp == NULL)
>>           return SCSI_MLQUEUE_HOST_BUSY;
>>       if (sdebug_verbose)
> 

Re: [PATCH] scsi: scsi_debug: Fix a warning in resp_write_scat()

[Martin K. Petersen]

Harshit,

> As 'lbdof_blen' is coming from user, if the size in kzalloc() is >=
> MAX_ORDER then we hit a warning.

Applied to 6.2/scsi-staging, thanks!

-- 
Martin K. Petersen	Oracle Linux Engineering

Re: [PATCH] scsi: scsi_debug: Fix a warning in resp_write_scat()

[Martin K. Petersen]

On Fri, 11 Nov 2022 02:05:25 -0800, Harshit Mogalapalli wrote:

> As 'lbdof_blen' is coming from user, if the size in kzalloc()
> is >= MAX_ORDER then we hit a warning.
> 
> Call trace:
> 
> sg_ioctl
>  sg_ioctl_common
>    scsi_ioctl
>     sg_scsi_ioctl
>      blk_execute_rq
>       blk_mq_sched_insert_request
>        blk_mq_run_hw_queue
>         __blk_mq_delay_run_hw_queue
>          __blk_mq_run_hw_queue
>           blk_mq_sched_dispatch_requests
>            __blk_mq_sched_dispatch_requests
>             blk_mq_dispatch_rq_list
>              scsi_queue_rq
>               scsi_dispatch_cmd
>                scsi_debug_queuecommand
>                 schedule_resp
>                  resp_write_scat
> 
> [...]

Applied to 6.2/scsi-queue, thanks!

[1/1] scsi: scsi_debug: Fix a warning in resp_write_scat()
      https://git.kernel.org/mkp/scsi/c/216e179724c1

-- 
Martin K. Petersen	Oracle Linux Engineering