The return value from the call to scsi_execute_cmd() is int. However, the
return value is being assigned to an unsigned int variable 'the_result',
so making 'the_result' an int.

./drivers/scsi/sd.c:2333:6-16: WARNING: Unsigned expression compared with zero: the_result > 0.

Reported-by: Abaci Robot <abaci@linux.alibaba.com>
Closes: https://bugzilla.openanolis.cn/show_bug.cgi?id=9463
Signed-off-by: Jiapeng Chong <jiapeng.chong@linux.alibaba.com>
---
 drivers/scsi/sd.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/scsi/sd.c b/drivers/scsi/sd.c
index 979795dad62b..ade8c6cca295 100644
--- a/drivers/scsi/sd.c
+++ b/drivers/scsi/sd.c
@@ -2396,7 +2396,7 @@ sd_spinup_disk(struct scsi_disk *sdkp)
 	static const u8 cmd[10] = { TEST_UNIT_READY };
 	unsigned long spintime_expire = 0;
 	int spintime, sense_valid = 0;
-	unsigned int the_result;
+	int the_result;
 	struct scsi_sense_hdr sshdr;
 	struct scsi_failure failure_defs[] = {
 		/* Do not retry Medium Not Present */
-- 
2.20.1.7.g153144c

This title should really be more precise, like "Fix unsigned expression 
compared with zero in sd_spinup_disk()"

On 01/07/2024 10:06, Jiapeng Chong wrote:

+ Mike

> The return value from the call to scsi_execute_cmd() is int. However, the
> return value is being assigned to an unsigned int variable 'the_result',
> so making 'the_result' an int.

an "so making 'the_result' effectively an unsigned int", right?

> 
> ./drivers/scsi/sd.c:2333:6-16: WARNING: Unsigned expression compared with zero: the_result > 0.
> 
> Reported-by: Abaci Robot <abaci@linux.alibaba.com>
> Closes: https://urldefense.com/v3/__https://bugzilla.openanolis.cn/show_bug.cgi?id=9463__;!!ACWV5N9M2RV99hQ!MVk0tuPkuuYZTeD-oVg7RYLVba7HwHVUjWCL2CLavSPrJmuO4MSCUNQ0vqjbCSIKBN8eonwTlU4FxKD3vmHQQhN2YUhbLcnJ$
> Signed-off-by: Jiapeng Chong <jiapeng.chong@linux.alibaba.com>

Looks ok, but Mike can hopefully double-check:

Reviewed-by: John Garry <john.g.garry@oracle.com>

> ---
>   drivers/scsi/sd.c | 2 +-
>   1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/drivers/scsi/sd.c b/drivers/scsi/sd.c
> index 979795dad62b..ade8c6cca295 100644
> --- a/drivers/scsi/sd.c
> +++ b/drivers/scsi/sd.c
> @@ -2396,7 +2396,7 @@ sd_spinup_disk(struct scsi_disk *sdkp)
>   	static const u8 cmd[10] = { TEST_UNIT_READY };
>   	unsigned long spintime_expire = 0;
>   	int spintime, sense_valid = 0;
> -	unsigned int the_result;
> +	int the_result;
>   	struct scsi_sense_hdr sshdr;
>   	struct scsi_failure failure_defs[] = {
>   		/* Do not retry Medium Not Present */

On 7/1/24 2:06 AM, Jiapeng Chong wrote:
> The return value from the call to scsi_execute_cmd() is int. However, the
> return value is being assigned to an unsigned int variable 'the_result',
> so making 'the_result' an int.

Please explain the full effect of this patch in the patch description. I
think this patch causes a potential read of uninitialized data (sshdr)
to be skipped if scsi_execute_cmd() returns a negative value. Do you
agree with this?

Thanks,

Bart.

On 7/1/24 2:06 AM, Jiapeng Chong wrote:
> The return value from the call to scsi_execute_cmd() is int. However, the
> return value is being assigned to an unsigned int variable 'the_result',
> so making 'the_result' an int.
> 
> ./drivers/scsi/sd.c:2333:6-16: WARNING: Unsigned expression compared with zero: the_result > 0.

Since this is a bug fix, please add Cc: stable and Fixes: tags.

Thanks,

Bart.