[PATCH] scsi: qla2xxx: Fix an endian bug in fcpcmd_is_corrupted()

[PATCH] scsi: qla2xxx: Fix an endian bug in fcpcmd_is_corrupted()

[Dan Carpenter]

We should first do the le16_to_cpu endian conversion and then apply
the FCP_CMD_LENGTH_MASK mask.

Fixes: 5f35509db179 ("qla2xxx: Terminate exchange if corrupted")
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
---
I just happened to spot this when I was reviewing something unrelated.
I don't have the hardware to test it, so please review carefully.

diff --git a/drivers/scsi/qla2xxx/qla_target.h b/drivers/scsi/qla2xxx/qla_target.h
index fecf96f0225c..199d3ba1916d 100644
--- a/drivers/scsi/qla2xxx/qla_target.h
+++ b/drivers/scsi/qla2xxx/qla_target.h
@@ -374,8 +374,8 @@ struct atio_from_isp {
 static inline int fcpcmd_is_corrupted(struct atio *atio)
 {
 	if (atio->entry_type = ATIO_TYPE7 &&
-	    (le16_to_cpu(atio->attr_n_length & FCP_CMD_LENGTH_MASK) <
-	    FCP_CMD_LENGTH_MIN))
+	    ((le16_to_cpu(atio->attr_n_length) & FCP_CMD_LENGTH_MASK) <
+	     FCP_CMD_LENGTH_MIN))
 		return 1;
 	else
 		return 0;

Re: [PATCH] scsi: qla2xxx: Fix an endian bug in fcpcmd_is_corrupted()

[Tran, Quinn]
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Re: [PATCH] scsi: qla2xxx: Fix an endian bug in fcpcmd_is_corrupted()

[Madhani, Himanshu]

> On Sep 8, 2018, at 1:42 AM, Dan Carpenter <dan.carpenter@oracle.com> wrote:
> 
> External Email
> 
> We should first do the le16_to_cpu endian conversion and then apply
> the FCP_CMD_LENGTH_MASK mask.
> 
> Fixes: 5f35509db179 ("qla2xxx: Terminate exchange if corrupted")
> Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
> ---
> I just happened to spot this when I was reviewing something unrelated.
> I don't have the hardware to test it, so please review carefully.
> 
> diff --git a/drivers/scsi/qla2xxx/qla_target.h b/drivers/scsi/qla2xxx/qla_target.h
> index fecf96f0225c..199d3ba1916d 100644
> --- a/drivers/scsi/qla2xxx/qla_target.h
> +++ b/drivers/scsi/qla2xxx/qla_target.h
> @@ -374,8 +374,8 @@ struct atio_from_isp {
> static inline int fcpcmd_is_corrupted(struct atio *atio)
> {
>        if (atio->entry_type == ATIO_TYPE7 &&
> -           (le16_to_cpu(atio->attr_n_length & FCP_CMD_LENGTH_MASK) <
> -           FCP_CMD_LENGTH_MIN))
> +           ((le16_to_cpu(atio->attr_n_length) & FCP_CMD_LENGTH_MASK) <
> +            FCP_CMD_LENGTH_MIN))
>                return 1;
>        else
>                return 0;

Acked-By: Himanshu Madhani <himanshu.madhani@cavium.com>

Thanks,
- Himanshu

Re: [PATCH] scsi: qla2xxx: Fix an endian bug in fcpcmd_is_corrupted()

[Martin K. Petersen]

Dan,

> We should first do the le16_to_cpu endian conversion and then apply
> the FCP_CMD_LENGTH_MASK mask.

Applied to 4.19/scsi-fixes, thank you!

-- 
Martin K. Petersen	Oracle Linux Engineering