[PATCH 1/4] net/mlx5: release representor interrupt handler

[PATCH 1/4] net/mlx5: release representor interrupt handler

[Bing Zhao]

From: Roi Dayan <roid@nvidia.com>

Looks like commit the cited commit removed the representor
interrupt handler cleanup by mistake. Fix it.

Fixes: 5cf0707fc7e9 ("net/mlx5: remove Rx queue data list from device")
Cc: xuemingl@nvidia.com

Signed-off-by: Roi Dayan <roid@nvidia.com>
---
 drivers/net/mlx5/mlx5.c | 5 +++++
 1 file changed, 5 insertions(+)

diff --git a/drivers/net/mlx5/mlx5.c b/drivers/net/mlx5/mlx5.c
index 666216c21a..591420c77f 100644
--- a/drivers/net/mlx5/mlx5.c
+++ b/drivers/net/mlx5/mlx5.c
@@ -2389,6 +2389,11 @@ mlx5_dev_close(struct rte_eth_dev *dev)
 	/* Free the eCPRI flex parser resource. */
 	mlx5_flex_parser_ecpri_release(dev);
 	mlx5_flex_item_port_cleanup(dev);
+	if (priv->representor) {
+		/* Each representor has a dedicated interrupts handler */
+		rte_intr_instance_free(dev->intr_handle);
+		dev->intr_handle = NULL;
+	}
 	mlx5_indirect_list_handles_release(dev);
 #ifdef HAVE_MLX5_HWS_SUPPORT
 	mlx5_nta_sample_context_free(dev);
-- 
2.34.1

RE: [PATCH 1/4] net/mlx5: release representor interrupt handler

[Suanming Mou]

Hi,

> -----Original Message-----
> From: Bing Zhao <bingz@nvidia.com>
> Sent: Wednesday, November 12, 2025 3:40 PM
> To: Slava Ovsiienko <viacheslavo@nvidia.com>; dev@dpdk.org; Raslan
> Darawsheh <rasland@nvidia.com>
> Cc: Ori Kam <orika@nvidia.com>; Dariusz Sosnowski
> <dsosnowski@nvidia.com>; Suanming Mou <suanmingm@nvidia.com>;
> Matan Azrad <matan@nvidia.com>; NBU-Contact-Thomas Monjalon
> (EXTERNAL) <thomas@monjalon.net>; Bing Zhao <bingz@nvidia.com>; Roi
> Dayan <roid@nvidia.com>; Xueming Li <xuemingl@nvidia.com>
> Subject: [PATCH 1/4] net/mlx5: release representor interrupt handler
> 
> From: Roi Dayan <roid@nvidia.com>
> 
> Looks like commit the cited commit removed the representor interrupt
> handler cleanup by mistake. Fix it.
> 
> Fixes: 5cf0707fc7e9 ("net/mlx5: remove Rx queue data list from device")
> Cc: xuemingl@nvidia.com
> 
> Signed-off-by: Roi Dayan <roid@nvidia.com>
Acked-by: Suanming Mou <suanmingm@nvidia.com>

Thanks

Re: [PATCH 1/4] net/mlx5: release representor interrupt handler

[Raslan Darawsheh]

Hi,


On 12/11/2025 9:40 AM, Bing Zhao wrote:
> From: Roi Dayan <roid@nvidia.com>
> 
> Looks like commit the cited commit removed the representor
> interrupt handler cleanup by mistake. Fix it.
> 
> Fixes: 5cf0707fc7e9 ("net/mlx5: remove Rx queue data list from device")
> Cc: xuemingl@nvidia.com
> 
> Signed-off-by: Roi Dayan <roid@nvidia.com>

Series applied to next-net-mlx,

Kindest regards
Raslan Darawsheh

Re: [PATCH 1/4] net/mlx5: release representor interrupt handler

[Kevin Traynor]

On 12/11/2025 07:40, Bing Zhao wrote:
> From: Roi Dayan <roid@nvidia.com>
> 
> Looks like commit the cited commit removed the representor
> interrupt handler cleanup by mistake. Fix it.
> 
> Fixes: 5cf0707fc7e9 ("net/mlx5: remove Rx queue data list from device")
> Cc: xuemingl@nvidia.com
> 

Hi,

Some of this series (see below) is fixing issues in older releases, but
does not include stable tag to indicate the fixes should be backported.

Just checking if this is deliberate, or an oversight and they should be
backported ?

thanks,
Kevin.

25.11 dbaed15366 - F net/mlx5: release representor interrupt handler
(Fixes issue in 21.11)

25.11 aef94343d3 - F common/mlx5: release unused mempool entries (Fixes
issue in  21.11 (partially fixed in 8947eebc999e @ 21.11))

25.11 8d1fe10768 - F net/mlx5/hws: fix buddy memory allocation (Fixes
issue in 22.11)


> Signed-off-by: Roi Dayan <roid@nvidia.com>
> ---
>  drivers/net/mlx5/mlx5.c | 5 +++++
>  1 file changed, 5 insertions(+)
> 
> diff --git a/drivers/net/mlx5/mlx5.c b/drivers/net/mlx5/mlx5.c
> index 666216c21a..591420c77f 100644
> --- a/drivers/net/mlx5/mlx5.c
> +++ b/drivers/net/mlx5/mlx5.c
> @@ -2389,6 +2389,11 @@ mlx5_dev_close(struct rte_eth_dev *dev)
>  	/* Free the eCPRI flex parser resource. */
>  	mlx5_flex_parser_ecpri_release(dev);
>  	mlx5_flex_item_port_cleanup(dev);
> +	if (priv->representor) {
> +		/* Each representor has a dedicated interrupts handler */
> +		rte_intr_instance_free(dev->intr_handle);
> +		dev->intr_handle = NULL;
> +	}
>  	mlx5_indirect_list_handles_release(dev);
>  #ifdef HAVE_MLX5_HWS_SUPPORT
>  	mlx5_nta_sample_context_free(dev);

Re: [PATCH 1/4] net/mlx5: release representor interrupt handler

[Thomas Monjalon]

20/11/2025 18:08, Kevin Traynor:
> On 12/11/2025 07:40, Bing Zhao wrote:
> > From: Roi Dayan <roid@nvidia.com>
> > 
> > Looks like commit the cited commit removed the representor
> > interrupt handler cleanup by mistake. Fix it.
> > 
> > Fixes: 5cf0707fc7e9 ("net/mlx5: remove Rx queue data list from device")
> > Cc: xuemingl@nvidia.com
> > 
> 
> Hi,
> 
> Some of this series (see below) is fixing issues in older releases, but
> does not include stable tag to indicate the fixes should be backported.
> 
> Just checking if this is deliberate, or an oversight and they should be
> backported ?
> 
> thanks,
> Kevin.
> 
> 25.11 dbaed15366 - F net/mlx5: release representor interrupt handler
> (Fixes issue in 21.11)
> 
> 25.11 aef94343d3 - F common/mlx5: release unused mempool entries (Fixes
> issue in  21.11 (partially fixed in 8947eebc999e @ 21.11))
> 
> 25.11 8d1fe10768 - F net/mlx5/hws: fix buddy memory allocation (Fixes
> issue in 22.11)

I think it is an oversight.
Usually I fix it myself, but this time I missed some.

Raslan, please run devtools/check-git-log.sh for each patch,
it is raising such issue.

Re: [PATCH 1/4] net/mlx5: release representor interrupt handler

[Kevin Traynor]

On 20/11/2025 17:39, Thomas Monjalon wrote:
> 20/11/2025 18:08, Kevin Traynor:
>> On 12/11/2025 07:40, Bing Zhao wrote:
>>> From: Roi Dayan <roid@nvidia.com>
>>>
>>> Looks like commit the cited commit removed the representor
>>> interrupt handler cleanup by mistake. Fix it.
>>>
>>> Fixes: 5cf0707fc7e9 ("net/mlx5: remove Rx queue data list from device")
>>> Cc: xuemingl@nvidia.com
>>>
>>
>> Hi,
>>
>> Some of this series (see below) is fixing issues in older releases, but
>> does not include stable tag to indicate the fixes should be backported.
>>
>> Just checking if this is deliberate, or an oversight and they should be
>> backported ?
>>
>> thanks,
>> Kevin.
>>
>> 25.11 dbaed15366 - F net/mlx5: release representor interrupt handler
>> (Fixes issue in 21.11)
>>
>> 25.11 aef94343d3 - F common/mlx5: release unused mempool entries (Fixes
>> issue in  21.11 (partially fixed in 8947eebc999e @ 21.11))
>>
>> 25.11 8d1fe10768 - F net/mlx5/hws: fix buddy memory allocation (Fixes
>> issue in 22.11)
> 
> I think it is an oversight.
> Usually I fix it myself, but this time I missed some.
> 

No problem, they all apply cleanly (24.11). Just checking builds now.

> Raslan, please run devtools/check-git-log.sh for each patch,
> it is raising such issue.
> 
>