Stable branches missing ed7a01fd3fd7

Stable branches missing ed7a01fd3fd7

[Håkon Bugge]

Hi Greg,


Commit 00313983cda6 ("RDMA/nldev: provide detailed CM_ID information") is in the following stable releases:

  stable/linux-4.17.y
  stable/linux-4.18.y
  stable/linux-4.19.y
  stable/linux-4.20.y
  stable/linux-5.0.y
  stable/linux-5.1.y
  stable/linux-5.2.y
  stable/linux-5.3.y
  stable/master

It has a potential for a big leak of task_struct's, and if the case is hit, the number of task_struct entries in /proc/slabinfo increases rapidly.

The fix, ed7a01fd3fd7 ("RDMA/restrack: Release task struct which was hold by CM_ID object"), is in the following stable releases:

  stable/linux-4.20.y
  stable/linux-5.0.y
  stable/linux-5.1.y
  stable/linux-5.2.y
  stable/linux-5.3.y
  stable/master

Hence, this commit needs to be included in 4-17..4.19.


Thxs, Håkon

Re: Stable branches missing ed7a01fd3fd7

[Greg Kroah-Hartman]

On Wed, Sep 18, 2019 at 11:17:59AM +0200, Håkon Bugge wrote:
> Hi Greg,
> 
> 
> Commit 00313983cda6 ("RDMA/nldev: provide detailed CM_ID information") is in the following stable releases:
> 
>   stable/linux-4.17.y
>   stable/linux-4.18.y
>   stable/linux-4.19.y
>   stable/linux-4.20.y
>   stable/linux-5.0.y
>   stable/linux-5.1.y
>   stable/linux-5.2.y
>   stable/linux-5.3.y
>   stable/master

It was part of the 4.17 release, so yes, of course it is in all later
releases.

> It has a potential for a big leak of task_struct's, and if the case is hit, the number of task_struct entries in /proc/slabinfo increases rapidly.
> 
> The fix, ed7a01fd3fd7 ("RDMA/restrack: Release task struct which was hold by CM_ID object"), is in the following stable releases:
> 
>   stable/linux-4.20.y
>   stable/linux-5.0.y
>   stable/linux-5.1.y
>   stable/linux-5.2.y
>   stable/linux-5.3.y
>   stable/master

It was part of the 4.20 release, so yes, it will be in all releases
newer than that.

> Hence, this commit needs to be included in 4-17..4.19.

Given there is only one "active" kernel branch you are looking at here,
that means it should only go into 4.19.y, right?

thanks,

greg k-h

Re: Stable branches missing ed7a01fd3fd7

[Håkon Bugge]

> On 18 Sep 2019, at 13:10, Greg Kroah-Hartman <gregkh@linuxfoundation.org> wrote:
> 
> On Wed, Sep 18, 2019 at 11:17:59AM +0200, Håkon Bugge wrote:
>> Hi Greg,
>> 
>> 
>> Commit 00313983cda6 ("RDMA/nldev: provide detailed CM_ID information") is in the following stable releases:
>> 
>>  stable/linux-4.17.y
>>  stable/linux-4.18.y
>>  stable/linux-4.19.y
>>  stable/linux-4.20.y
>>  stable/linux-5.0.y
>>  stable/linux-5.1.y
>>  stable/linux-5.2.y
>>  stable/linux-5.3.y
>>  stable/master
> 
> It was part of the 4.17 release, so yes, of course it is in all later
> releases.
> 
>> It has a potential for a big leak of task_struct's, and if the case is hit, the number of task_struct entries in /proc/slabinfo increases rapidly.
>> 
>> The fix, ed7a01fd3fd7 ("RDMA/restrack: Release task struct which was hold by CM_ID object"), is in the following stable releases:
>> 
>>  stable/linux-4.20.y
>>  stable/linux-5.0.y
>>  stable/linux-5.1.y
>>  stable/linux-5.2.y
>>  stable/linux-5.3.y
>>  stable/master
> 
> It was part of the 4.20 release, so yes, it will be in all releases
> newer than that.
> 
>> Hence, this commit needs to be included in 4-17..4.19.
> 
> Given there is only one "active" kernel branch you are looking at here,
> that means it should only go into 4.19.y, right?

Yes, you are right. I should have read the collateral first, kernel.org/releases.html :-)


Thxs, Håkon