[PATCH 1/2] libfc: fix mem leak in fc_tm_done()

[PATCH 1/2] libfc: fix mem leak in fc_tm_done()

[Hillf Danton]

There seems frame should get freed against memory leakage.

Signed-off-by: Hillf Danton <dhillf-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>
---

--- a/drivers/scsi/libfc/fc_fcp.c	2010-09-13 07:07:38.000000000 +0800
+++ b/drivers/scsi/libfc/fc_fcp.c	2010-10-25 20:31:04.000000000 +0800
@@ -1294,6 +1294,7 @@ static void fc_tm_done(struct fc_seq *se
 	 */
 	if (!fsp->seq_ptr || !fsp->wait_for_comp) {
 		spin_unlock_bh(&fsp->scsi_pkt_lock);
+		fc_frame_free(fp);
 		return;
 	}

[PATCH 2/2] libfc: fix mem leak in fc_tm_done()

[Hillf Danton]

There seems fc_fcp_unlock_pkt should be issued against mem leak.

Signed-off-by: Hillf Danton <dhillf@gmail.com>
---

--- a/drivers/scsi/libfc/fc_fcp.c	2010-09-13 07:07:38.000000000 +0800
+++ b/drivers/scsi/libfc/fc_fcp.c	2010-10-25 20:32:18.000000000 +0800
@@ -1293,7 +1293,7 @@ static void fc_tm_done(struct fc_seq *se
 	 * raced with eh timeout handler.
 	 */
 	if (!fsp->seq_ptr || !fsp->wait_for_comp) {
-		spin_unlock_bh(&fsp->scsi_pkt_lock);
+		fc_fcp_unlock_pkt(fsp);
 		return;
 	}

Re: [Open-FCoE] [PATCH 1/2] libfc: fix mem leak in fc_tm_done()

[Robert Love]

On Mon, 2010-10-25 at 20:37 +0800, Hillf Danton wrote:
> There seems frame should get freed against memory leakage.
> 
> Signed-off-by: Hillf Danton <dhillf@gmail.com>
> ---
> 
> --- a/drivers/scsi/libfc/fc_fcp.c	2010-09-13 07:07:38.000000000 +0800
> +++ b/drivers/scsi/libfc/fc_fcp.c	2010-10-25 20:31:04.000000000 +0800
> @@ -1294,6 +1294,7 @@ static void fc_tm_done(struct fc_seq *se
>  	 */
>  	if (!fsp->seq_ptr || !fsp->wait_for_comp) {
>  		spin_unlock_bh(&fsp->scsi_pkt_lock);
> +		fc_frame_free(fp);
>  		return;
>  	}
> _______________________________________________
> devel mailing list
> devel@open-fcoe.org
> http://www.open-fcoe.org/mailman/listinfo/devel

Hi Hillf,

   I'm not sure how you're generating your patches, but it seems that
many of them (not sure if it's all) do not have a trailing whitespace
line. So, when I try to apply the patch either using 'patch' or 'git-am'
it sees the first line of the mailing list footer as part of the patch.
So the, "____________________" line is considered part of the patch and
it causes the application of the patch to fail. Here's what happens when
I try to apply the mbox for this patch.

rwlove@fritz:/var/testspace/open-fcoe-initiator$ git-am
~/Desktop/1-2-libfc-fix-mem-leak-in-fc_tm_done.patch 
Applying: libfc: fix mem leak in fc_tm_done()
fatal: corrupt patch at line 16
Patch failed at 0001 libfc: fix mem leak in fc_tm_done()
When you have resolved this problem run "git am --resolved".
If you would prefer to skip this patch, instead run "git am --skip".
To restore the original branch and stop patching run "git am --abort".
rwlove@fritz:/var/testspace/open-fcoe-initiator$ patch -p1 --dry-run
< .git/rebase-apply/0001 
patching file drivers/scsi/libfc/fc_fcp.c
patch: **** malformed patch at line 28:
_______________________________________________


   I'm not a patch formatting expert, but I think the lack of a
whitespace line at the end of the changes is the problem.

   I'd suggest mailing the patch to yourself first and then try applying
the mbox with 'git-am', if that works then everything should be fine.

   Also, I like to use stgit to generate and mail patches. Everyone has
their own development process, but I want to point that tool out to you
because it will take care of all of the formatting for you.

Thanks, //Rob

Re: [PATCH 1/2] libfc: fix mem leak in fc_tm_done()

[Robert Love]

On Mon, 2010-10-25 at 20:37 +0800, Hillf Danton wrote:
> There seems frame should get freed against memory leakage.
> 
> Signed-off-by: Hillf Danton <dhillf-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>
> ---
> 
> --- a/drivers/scsi/libfc/fc_fcp.c	2010-09-13 07:07:38.000000000 +0800
> +++ b/drivers/scsi/libfc/fc_fcp.c	2010-10-25 20:31:04.000000000 +0800
> @@ -1294,6 +1294,7 @@ static void fc_tm_done(struct fc_seq *se
>  	 */
>  	if (!fsp->seq_ptr || !fsp->wait_for_comp) {
>  		spin_unlock_bh(&fsp->scsi_pkt_lock);

I wonder if this should be a fc_fcp_unlock_pkt(fsp) since we get the
lock with the fc_fcp_lock_pkt(fsp) call above.

> +		fc_frame_free(fp);
>  		return;
>  	}
> _______________________________________________
> devel mailing list
> devel-s9riP+hp16TNLxjTenLetw@public.gmane.org
> http://www.open-fcoe.org/mailman/listinfo/devel

Re: [PATCH 1/2] libfc: fix mem leak in fc_tm_done()

[Mike Christie]

On 10/27/2010 08:24 PM, Robert Love wrote:
> On Mon, 2010-10-25 at 20:37 +0800, Hillf Danton wrote:
>> There seems frame should get freed against memory leakage.
>>
>> Signed-off-by: Hillf Danton<dhillf-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>
>> ---
>>
>> --- a/drivers/scsi/libfc/fc_fcp.c	2010-09-13 07:07:38.000000000 +0800
>> +++ b/drivers/scsi/libfc/fc_fcp.c	2010-10-25 20:31:04.000000000 +0800
>> @@ -1294,6 +1294,7 @@ static void fc_tm_done(struct fc_seq *se
>>   	 */
>>   	if (!fsp->seq_ptr || !fsp->wait_for_comp) {
>>   		spin_unlock_bh(&fsp->scsi_pkt_lock);
>
> I wonder if this should be a fc_fcp_unlock_pkt(fsp) since we get the
> lock with the fc_fcp_lock_pkt(fsp) call above.
>

Did I do that. I think you are right. Not sure what happened if it was me.

Re: [Open-FCoE] [PATCH 1/2] libfc: fix mem leak in fc_tm_done()

[Hillf Danton]

On Thu, Oct 28, 2010 at 9:55 AM, Mike Christie <michaelc@cs.wisc.edu> wrote:
> On 10/27/2010 08:24 PM, Robert Love wrote:
>>
>> On Mon, 2010-10-25 at 20:37 +0800, Hillf Danton wrote:
>>>
>>> There seems frame should get freed against memory leakage.
>>>
>>> Signed-off-by: Hillf Danton<dhillf@gmail.com>
>>> ---
>>>
>>> --- a/drivers/scsi/libfc/fc_fcp.c       2010-09-13 07:07:38.000000000
>>> +0800
>>> +++ b/drivers/scsi/libfc/fc_fcp.c       2010-10-25 20:31:04.000000000
>>> +0800
>>> @@ -1294,6 +1294,7 @@ static void fc_tm_done(struct fc_seq *se
>>>         */
>>>        if (!fsp->seq_ptr || !fsp->wait_for_comp) {
>>>                spin_unlock_bh(&fsp->scsi_pkt_lock);
>>
>> I wonder if this should be a fc_fcp_unlock_pkt(fsp) since we get the
>> lock with the fc_fcp_lock_pkt(fsp) call above.
>>
According to one question a patch, two were delivered, please check them.

Here I want frame get freed. //Hillf

>
> Did I do that. I think you are right. Not sure what happened if it was me.
>
--
To unsubscribe from this list: send the line "unsubscribe linux-scsi" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [Open-FCoE] [PATCH 1/2] libfc: fix mem leak in fc_tm_done()

[Hillf Danton]

On Thu, Oct 28, 2010 at 9:18 AM, Robert Love <robert.w.love@intel.com> wrote:
> On Mon, 2010-10-25 at 20:37 +0800, Hillf Danton wrote:
>> There seems frame should get freed against memory leakage.
>>
>> Signed-off-by: Hillf Danton <dhillf@gmail.com>
>> ---
>>
>> --- a/drivers/scsi/libfc/fc_fcp.c     2010-09-13 07:07:38.000000000 +0800
>> +++ b/drivers/scsi/libfc/fc_fcp.c     2010-10-25 20:31:04.000000000 +0800
>> @@ -1294,6 +1294,7 @@ static void fc_tm_done(struct fc_seq *se
>>        */
>>       if (!fsp->seq_ptr || !fsp->wait_for_comp) {
>>               spin_unlock_bh(&fsp->scsi_pkt_lock);
>> +             fc_frame_free(fp);
>>               return;
>>       }
>> _______________________________________________
>> devel mailing list
>> devel@open-fcoe.org
>> http://www.open-fcoe.org/mailman/listinfo/devel
>
> Hi Hillf,
>
>   I'm not sure how you're generating your patches, but it seems that
> many of them (not sure if it's all) do not have a trailing whitespace
> line. So, when I try to apply the patch either using 'patch' or 'git-am'
> it sees the first line of the mailing list footer as part of the patch.
> So the, "____________________" line is considered part of the patch and
> it causes the application of the patch to fail. Here's what happens when
> I try to apply the mbox for this patch.
>
> rwlove@fritz:/var/testspace/open-fcoe-initiator$ git-am
> ~/Desktop/1-2-libfc-fix-mem-leak-in-fc_tm_done.patch
> Applying: libfc: fix mem leak in fc_tm_done()
> fatal: corrupt patch at line 16
> Patch failed at 0001 libfc: fix mem leak in fc_tm_done()
> When you have resolved this problem run "git am --resolved".
> If you would prefer to skip this patch, instead run "git am --skip".
> To restore the original branch and stop patching run "git am --abort".
> rwlove@fritz:/var/testspace/open-fcoe-initiator$ patch -p1 --dry-run
> < .git/rebase-apply/0001
> patching file drivers/scsi/libfc/fc_fcp.c
> patch: **** malformed patch at line 28:
> _______________________________________________
>
>
>   I'm not a patch formatting expert, but I think the lack of a
> whitespace line at the end of the changes is the problem.
>
>   I'd suggest mailing the patch to yourself first and then try applying
> the mbox with 'git-am', if that works then everything should be fine.
>
>   Also, I like to use stgit to generate and mail patches. Everyone has
> their own development process, but I want to point that tool out to you
> because it will take care of all of the formatting for you.
>
> Thanks, //Rob

Sorry, get you in trouble again. More care will be taken. //Hillf
--
To unsubscribe from this list: send the line "unsubscribe linux-scsi" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

RE: [Open-FCoE] [PATCH 1/2] libfc: fix mem leak in fc_tm_done()

[Zou, Yi]

> On Thu, Oct 28, 2010 at 9:55 AM, Mike Christie <michaelc@cs.wisc.edu>
> wrote:
> > On 10/27/2010 08:24 PM, Robert Love wrote:
> >>
> >> On Mon, 2010-10-25 at 20:37 +0800, Hillf Danton wrote:
> >>>
> >>> There seems frame should get freed against memory leakage.
> >>>
> >>> Signed-off-by: Hillf Danton<dhillf@gmail.com>
> >>> ---
> >>>
> >>> --- a/drivers/scsi/libfc/fc_fcp.c       2010-09-13 07:07:38.000000000
> >>> +0800
> >>> +++ b/drivers/scsi/libfc/fc_fcp.c       2010-10-25 20:31:04.000000000
> >>> +0800
> >>> @@ -1294,6 +1294,7 @@ static void fc_tm_done(struct fc_seq *se
> >>>         */
> >>>        if (!fsp->seq_ptr || !fsp->wait_for_comp) {
> >>>                spin_unlock_bh(&fsp->scsi_pkt_lock);
> >>
> >> I wonder if this should be a fc_fcp_unlock_pkt(fsp) since we get the
> >> lock with the fc_fcp_lock_pkt(fsp) call above.
> >>
> According to one question a patch, two were delivered, please check them.
> 
> Here I want frame get freed. //Hillf
> 
> >
> > Did I do that. I think you are right. Not sure what happened if it was
> me.
> >

I would think fc_frame_free() be called on all the if () { ..return } cases, the
caller won't do that from fc_exch_recv(), also, frame should be freed after unlock
the fsp:
{
	if (IS_ERR(fp)) 
		goto out;
	if (fc_fcp_lock_pkt(fsp))
		goto out;
	if (!fsp->seq_ptr || !fsp->wait_for_comp)
		goto out_unlock;

	...

out_unlock:
	fc_fcp_unlock_pkt(fsp);
out:
	fc_frame_free(fp);
	       
}

yi