回复: Re: [PATCH v2 rdma-rc] RDMA/bnxt_re: Fix a potential memory leak in destroy_gsi_sqp

回复: Re: [PATCH v2 rdma-rc] RDMA/bnxt_re: Fix a potential memory leak in destroy_gsi_sqp

[代彦龙]

[-- Attachment #1: Type: text/html, Size: 7654 bytes --]

Re: åå¤: Re: [PATCH v2 rdma-rc] RDMA/bnxt_re: Fix a potential memory leak in destroy_gsi_sqp

[yanjun.zhu]

On 9/25/25 7:39 PM, ä»£å½¦é¾ wrote:
> Hello Kalesh, Selvin, List,
> 
> Gentle ping on this patch.
> I just wanted to follow up and make sure the v2 version was received 
> correctly.
> 
> For easy reference, the patch is available on lore.kernel.org here:
> https://lore.kernel.org/all/20250924061444.11288-1-daiyanlong <https:// 
> lore.kernel.org/all/20250924061444.11288-1-daiyanlong>@kylinos.cn/
> 
> Please let me know if there's anything else needed from my side, or if 
> you've had a chance to review the technical changes.
> Thank you for your time and consideration.
> 
> Best regards,
> YanLong Dai
> 

Hi, YanLong

In regions where Chinese is not supported, the email may appear garbled. 
We recommend replacing any Chinese content in the email with the 
corresponding English to ensure clarity.

Thanks a lot.
Yanjun.Zhu

> 
> 
> 
> *主   题：*Re: [PATCH v2 rdma-rc] RDMA/bnxt_re: Fix a potential memory 
> leak in destroy_gsi_sqp
> *日   期：*2025年09月24日14:33
> *发件人：*代彦龙
> *收件人：*Kalesh Anakkur Purayil,代彦龙
> *抄送人：*jgg,leon,linux-kernel,linux-rdma,selvin.xavier,dyl_wlc
> 
> Hello Selvin, Kalesh, List,
> 
> Thank you so much for your patience and guidance throughout this 
> process. I truly appreciate you taking the time to review my patches and 
> provide detailed feedback - it has been a great learning experience.
> 
> I have incorporated all of your suggestions in this v2 version:
> - Added the "rdma-rc" target tree prefix in the subject line
> - Used proper version numbering (v2)
> - Included the changelog below the '---' line as recommended
> 
> The updated patch is attached. I believe it now follows all the required 
> guidelines. Please let me know if there are any further issues or 
> adjustments needed.
> 
> The patch is also available on lore.kernel.org here:
> https://lore.kernel.org/all/20250924061444.11288-1-daiyanlong <https:// 
> lore.kernel.org/all/20250924061444.11288-1-daiyanlong>@kylinos.cn/
> 
> Best regards,
> YanLong Dai
> 
> ---
> 
> 
> 
> 
> *主   题：*Re: [PATCH] RDMA/bnxt_re: Fix a potential memory leak in 
> destroy_gsi_sqp
> *日   期：*2025年09月24日13:01
> *发件人：*Kalesh Anakkur Purayil
> *收件人：*Kalesh Anakkur Purayil
> *抄送人：*jgg,leon,linux-kernel,linux-rdma,selvin.xavier,dyl_wlc
> 
> Hi YanLong,
> Few generic guidelines.
> 1. You should select a target tree in the subject prefix and specify a 
> revision number. Since this is a bug fix, the target tree should be 
> "rdma-rc".
> 2. When you send an updated version of the patch, please mention version 
> number. Also, mention the changes made in each version. i.e. under --- 
> you can add extra info that will not be included in the actual commit, 
> e.g. changes between each version of patches.
> One comment in line.
> 
> On Mon, Sep 22, 2025 at 7:53 AM YanLong Dai <daiyanlong@kylinos.cn 
> <mailto:daiyanlong@kylinos.cn>> wrote:
> 
>     From: daiyanlong <daiyanlong@kylinos.cn <mailto:daiyanlong@kylinos.cn>>
> 
>     The current error handling path in bnxt_re_destroy_gsi_sqp() could lead
>     to a resource leak. When bnxt_qplib_destroy_qp() fails, the function
>     jumps to the 'fail' label and returns immediately, skipping the call
>     to bnxt_qplib_free_qp_res().
> 
>     Continue the resource teardown even if bnxt_qplib_destroy_qp() fails,
>     which aligns with the driver's general error handling strategy and
>     prevents the potential leak.
> 
>     Fixes: 8dae419f9ec73 ("RDMA/bnxt_re: Refactor queue pair creation code")
>     [Kalesh] Blank line is not needed between Fixes tag and SOB tag
>     Signed-off-by: daiyanlong <daiyanlong@kylinos.cn
>     <mailto:daiyanlong@kylinos.cn>>
>     ---
>       drivers/infiniband/hw/bnxt_re/ib_verbs.c | 7 ++-----
>       1 file changed, 2 insertions(+), 5 deletions(-)
> 
>     diff --git a/drivers/infiniband/hw/bnxt_re/ib_verbs.c b/drivers/
>     infiniband/hw/bnxt_re/ib_verbs.c
>     index 260dc67b8b87..15d3f5d5c0ee 100644
>     --- a/drivers/infiniband/hw/bnxt_re/ib_verbs.c
>     +++ b/drivers/infiniband/hw/bnxt_re/ib_verbs.c
>     @@ -931,10 +931,9 @@ static int bnxt_re_destroy_gsi_sqp(struct
>     bnxt_re_qp *qp)
> 
>              ibdev_dbg(&rdev->ibdev, "Destroy the shadow QP\n");
>              rc = bnxt_qplib_destroy_qp(&rdev->qplib_res, &gsi_sqp-
>      >qplib_qp);
>     -       if (rc) {
>     +       if (rc)
>                      ibdev_err(&rdev->ibdev, "Destroy Shadow QP failed");
>     -               goto fail;
>     -       }
>     +
>              bnxt_qplib_free_qp_res(&rdev->qplib_res, &gsi_sqp->qplib_qp);
> 
>              /* remove from active qp list */
>     @@ -951,8 +950,6 @@ static int bnxt_re_destroy_gsi_sqp(struct
>     bnxt_re_qp *qp)
>              rdev->gsi_ctx.sqp_tbl = NULL;
> 
>              return 0;
>     -fail:
>     -       return rc;
>       }
> 
>       /* Queue Pairs */
>     -- 
>     2.43.0
> 
> 
> -- 
> Regards,
> Kalesh AP
> 
> 
> ---
> 
> 
> ---
> 

Re:[PATCH v2 rdma-rc] RDMA/bnxt_re: Fix a potential memory leak in destroy_gsi_sqp

[YanLong Dai]

Hi Kalesh, Yanjun

Thank you both for your time and reviews.

On Wed, Sep 24, 2025 at 10:31:34 +0530, Kalesh wrote:
>Hi YanLong,
>
>Few generic guidelines.
>
>1. You should select a target tree in the subject prefix and specify a revision number. Since this is a bug fix, the target tree should be "rdma-rc".
>2. When you send an updated version of the patch, please mention version number. Also, mention the changes made in each version. i.e. under --- you can add extra info that will not be included in the actual commit, e.g. changes between each version of patches.
>
>One comment in line.

Thank you so much for your patience and guidance throughout this process.

I have incorporated all of your suggestions in this v2 version:
- Added the "rdma-rc" target tree prefix in the subject line
- Used proper version numbering (v2)
- Included the changelog below the '---' line as recommended

Gentle ping on this patch.
For easy reference, the patch is available on lore.kernel.org here: 
https://lore.kernel.org/all/20250924061444.11288-1-daiyanlong@kylinos.cn/


---


On Fri, Sep 26, 2025 at 09:11:11AM -0700, Yanjun wrote:
> Hi, YanLong
> 
> In regions where Chinese is not supported, the email may appear garbled. 
> We recommend replacing any Chinese content in the email with the 
> corresponding English to ensure clarity.
> 
> Thanks a lot.
> Yanjun.Zhu

Thank you for the reminder.
I have replaced all the Chinese content in my previous email with English to ensure clarity and avoid any encoding issues.


Best regards,
YanLong Dai

Re: [PATCH v2 rdma-rc] RDMA/bnxt_re: Fix a potential memory leak in destroy_gsi_sqp

[Zhu Yanjun]

在 2025/9/27 18:11, YanLong Dai 写道:
> Hi Kalesh, Yanjun
>
> Thank you both for your time and reviews.
>
> On Wed, Sep 24, 2025 at 10:31:34 +0530, Kalesh wrote:
>> Hi YanLong,
>>
>> Few generic guidelines.
>>
>> 1. You should select a target tree in the subject prefix and specify a revision number. Since this is a bug fix, the target tree should be "rdma-rc".
>> 2. When you send an updated version of the patch, please mention version number. Also, mention the changes made in each version. i.e. under --- you can add extra info that will not be included in the actual commit, e.g. changes between each version of patches.
>>
>> One comment in line.
> Thank you so much for your patience and guidance throughout this process.
>
> I have incorporated all of your suggestions in this v2 version:
> - Added the "rdma-rc" target tree prefix in the subject line
> - Used proper version numbering (v2)
> - Included the changelog below the '---' line as recommended
>
> Gentle ping on this patch.
> For easy reference, the patch is available on lore.kernel.org here:
> https://lore.kernel.org/all/20250924061444.11288-1-daiyanlong@kylinos.cn/
>
>
> ---
>
>
> On Fri, Sep 26, 2025 at 09:11:11AM -0700, Yanjun wrote:
>> Hi, YanLong
>>
>> In regions where Chinese is not supported, the email may appear garbled.
>> We recommend replacing any Chinese content in the email with the
>> corresponding English to ensure clarity.
>>
>> Thanks a lot.
>> Yanjun.Zhu
> Thank you for the reminder.
> I have replaced all the Chinese content in my previous email with English to ensure clarity and avoid any encoding issues.

Thanks a lot. The mail looks better. I am fine with this.

Yanjun.Zhu

>
>
> Best regards,
> YanLong Dai
>
-- 
Best Regards,
Yanjun.Zhu

Re: [PATCH v2 rdma-rc] RDMA/bnxt_re: Fix a potential memory leak in destroy_gsi_sqp

[代彦龙]

[-- Attachment #1: Type: text/html, Size: 5773 bytes --]

[PATCH v2 rdma-rc] RDMA/bnxt_re: Fix a potential memory leak in destroy_gsi_sqp

[YanLong Dai]

The current error handling path in bnxt_re_destroy_gsi_sqp() could lead
to a resource leak. When bnxt_qplib_destroy_qp() fails, the function
jumps to the 'fail' label and returns immediately, skipping the call
to bnxt_qplib_free_qp_res().

Continue the resource teardown even if bnxt_qplib_destroy_qp() fails,
which aligns with the driver's general error handling strategy and
prevents the potential leak.

Fixes: 8dae419f9ec73 ("RDMA/bnxt_re: Refactor queue pair creation code")

Signed-off-by: YanLong Dai <daiyanlong@kylinos.cn>

---
 v2:
 - Updated subject prefix to 'rdma-rc'
---
 drivers/infiniband/hw/bnxt_re/ib_verbs.c | 7 ++-----
 1 file changed, 2 insertions(+), 5 deletions(-)

diff --git a/drivers/infiniband/hw/bnxt_re/ib_verbs.c b/drivers/infiniband/hw/bnxt_re/ib_verbs.c
index 260dc67b8b87..15d3f5d5c0ee 100644
--- a/drivers/infiniband/hw/bnxt_re/ib_verbs.c
+++ b/drivers/infiniband/hw/bnxt_re/ib_verbs.c
@@ -931,10 +931,9 @@ static int bnxt_re_destroy_gsi_sqp(struct bnxt_re_qp *qp)
 
 	ibdev_dbg(&rdev->ibdev, "Destroy the shadow QP\n");
 	rc = bnxt_qplib_destroy_qp(&rdev->qplib_res, &gsi_sqp->qplib_qp);
-	if (rc) {
+	if (rc)
 		ibdev_err(&rdev->ibdev, "Destroy Shadow QP failed");
-		goto fail;
-	}
+
 	bnxt_qplib_free_qp_res(&rdev->qplib_res, &gsi_sqp->qplib_qp);
 
 	/* remove from active qp list */
@@ -951,8 +950,6 @@ static int bnxt_re_destroy_gsi_sqp(struct bnxt_re_qp *qp)
 	rdev->gsi_ctx.sqp_tbl = NULL;
 
 	return 0;
-fail:
-	return rc;
 }
 
 /* Queue Pairs */
-- 
2.43.0

Re: [PATCH v2 rdma-rc] RDMA/bnxt_re: Fix a potential memory leak in destroy_gsi_sqp

[Leon Romanovsky]

On Wed, 24 Sep 2025 14:14:44 +0800, YanLong Dai wrote:
> The current error handling path in bnxt_re_destroy_gsi_sqp() could lead
> to a resource leak. When bnxt_qplib_destroy_qp() fails, the function
> jumps to the 'fail' label and returns immediately, skipping the call
> to bnxt_qplib_free_qp_res().
> 
> Continue the resource teardown even if bnxt_qplib_destroy_qp() fails,
> which aligns with the driver's general error handling strategy and
> prevents the potential leak.
> 
> [...]

Applied, thanks!

[1/1] RDMA/bnxt_re: Fix a potential memory leak in destroy_gsi_sqp
      https://git.kernel.org/rdma/rdma/c/88de89f184661e

Best regards,
-- 
Leon Romanovsky <leon@kernel.org>