[PATCH] [scsi] lpfc: potential leak in lpfc_sli_hbqbuf_fill_hbqs

[PATCH] [scsi] lpfc: potential leak in lpfc_sli_hbqbuf_fill_hbqs

[Florin Malita]

If lpfc_hbq_alloc() fails, an hbq_buffer is leaked. Found by Coverity 
(1775).

Signed-off-by: Florin Malita <fmalita@gmail.com>
---

 drivers/scsi/lpfc/lpfc_sli.c |    4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

diff --git a/drivers/scsi/lpfc/lpfc_sli.c b/drivers/scsi/lpfc/lpfc_sli.c
index f4d5a6b..fe16b0c 100644
--- a/drivers/scsi/lpfc/lpfc_sli.c
+++ b/drivers/scsi/lpfc/lpfc_sli.c
@@ -622,8 +622,10 @@ lpfc_sli_hbqbuf_fill_hbqs(struct lpfc_hba *phba, uint32_t hbqno, uint32_t count)
 			return 1;
 		hbq_buffer->dbuf.virt = lpfc_hbq_alloc(phba, MEM_PRI,
 							&hbq_buffer->dbuf.phys);
-		if (hbq_buffer->dbuf.virt == NULL)
+		if (hbq_buffer->dbuf.virt == NULL) {
+			kfree(hbq_buffer);
 			return 1;
+		}
 		hbq_buffer->tag = (i | (hbqno << 16));
 		lpfc_sli_hbq_to_firmware(phba, hbqno, hbq_buffer);
 		lpfc_hbq_defs[hbqno]->buffer_count++;

Re: [PATCH] [scsi] lpfc: potential leak in lpfc_sli_hbqbuf_fill_hbqs

[James Smart]

Florin,

Thanks for the effort to create this patch.  Turns out we have already
fixed this, and have it included in our 8.2.2 patch set, which we will be
posting within the next week.

So James, don't pick this patch up, take the larger patch set instead.

Thanks.

-- james s

Florin Malita wrote:
> If lpfc_hbq_alloc() fails, an hbq_buffer is leaked. Found by Coverity 
> (1775).
> 
> Signed-off-by: Florin Malita <fmalita@gmail.com>
> ---
> 
> drivers/scsi/lpfc/lpfc_sli.c |    4 +++-
> 1 file changed, 3 insertions(+), 1 deletion(-)
> 
> diff --git a/drivers/scsi/lpfc/lpfc_sli.c b/drivers/scsi/lpfc/lpfc_sli.c
> index f4d5a6b..fe16b0c 100644
> --- a/drivers/scsi/lpfc/lpfc_sli.c
> +++ b/drivers/scsi/lpfc/lpfc_sli.c
> @@ -622,8 +622,10 @@ lpfc_sli_hbqbuf_fill_hbqs(struct lpfc_hba *phba, 
> uint32_t hbqno, uint32_t count)
>             return 1;
>         hbq_buffer->dbuf.virt = lpfc_hbq_alloc(phba, MEM_PRI,
>                             &hbq_buffer->dbuf.phys);
> -        if (hbq_buffer->dbuf.virt == NULL)
> +        if (hbq_buffer->dbuf.virt == NULL) {
> +            kfree(hbq_buffer);
>             return 1;
> +        }
>         hbq_buffer->tag = (i | (hbqno << 16));
>         lpfc_sli_hbq_to_firmware(phba, hbqno, hbq_buffer);
>         lpfc_hbq_defs[hbqno]->buffer_count++;
> 
> 

Re: [PATCH] [scsi] lpfc: potential leak in lpfc_sli_hbqbuf_fill_hbqs

[James Bottomley]

On Thu, 2007-07-19 at 13:44 -0400, James Smart wrote:
> Florin,
> 
> Thanks for the effort to create this patch.  Turns out we have already
> fixed this, and have it included in our 8.2.2 patch set, which we will be
> posting within the next week.

The 2.6.23 merge window is estimated to close on Sunday (22 July) ...

> So James, don't pick this patch up, take the larger patch set instead.

OK, will wait.

James