Re: Two net_sched fixes for stable

[David Miller]

From: Cong Wang 
Date: Tue, 6 Feb 2018 21:11:15 -0800

> On Tue, Feb 6, 2018 at 12:20 PM, David Miller  wrote:
>> From: Cong Wang 
>>>
>>> Please let me know how you want to handle this for 4.14.
>>
>> Ok, I sent this off for 4.15 -stable but I need you to do the
>> 4.14 backport.
>>
> 
> OK. I assume you mean I should send the backports directly
> to stable. I will do it.

Yes, please.  You can CC: me.

Re: Two net_sched fixes for stable

[Cong Wang]

On Tue, Feb 6, 2018 at 12:20 PM, David Miller  wrote:
> From: Cong Wang 
>>
>> Please let me know how you want to handle this for 4.14.
>
> Ok, I sent this off for 4.15 -stable but I need you to do the
> 4.14 backport.
>

OK. I assume you mean I should send the backports directly
to stable. I will do it.

Thanks.

Re: Two net_sched fixes for stable

[David Miller]

From: Cong Wang 
Date: Tue, 6 Feb 2018 11:25:10 -0800

> Can you queue the following commits for stable? They fix a
> sleep-in-atomic warning reported by Roland.
> 
> commit efbf78973978b0d25af59bc26c8013a942af6e64
> Author: Cong Wang 
> Date:   Mon Dec 4 10:48:18 2017 -0800
> 
> net_sched: get rid of rcu_barrier() in tcf_block_put_ext()
> 
> commit df45bf84e4f5a48f23d4b1a07d21d566e8b587b2
> Author: Jiri Pirko 
> Date:   Fri Dec 8 19:27:27 2017 +0100
> 
> net: sched: fix use-after-free in tcf_block_put_ext
> 
> This problem was introduced by:
> 
> commit e2ef75445340ca7ec2c4558f84ae6c8c5d650fc8
> Author: Cong Wang 
> Date:   Mon Sep 11 16:33:31 2017 -0700
> 
> net_sched: fix reference counting of tc filter chain
> 
> so should apply to 4.14 too.
> 
> For 4.15, they can just apply nearly cleanly. However, it is not
> straight-forward to cherry pick them to 4.14 due to some big changes
> between 4.14 and 4.15.
> 
> Please let me know how you want to handle this for 4.14.

Ok, I sent this off for 4.15 -stable but I need you to do the
4.14 backport.

Thank you.

Two net_sched fixes for stable

[Cong Wang]

Hi, David

Can you queue the following commits for stable? They fix a
sleep-in-atomic warning reported by Roland.


commit efbf78973978b0d25af59bc26c8013a942af6e64
Author: Cong Wang 
Date:   Mon Dec 4 10:48:18 2017 -0800

net_sched: get rid of rcu_barrier() in tcf_block_put_ext()

commit df45bf84e4f5a48f23d4b1a07d21d566e8b587b2
Author: Jiri Pirko 
Date:   Fri Dec 8 19:27:27 2017 +0100

net: sched: fix use-after-free in tcf_block_put_ext


This problem was introduced by:

commit e2ef75445340ca7ec2c4558f84ae6c8c5d650fc8
Author: Cong Wang 
Date:   Mon Sep 11 16:33:31 2017 -0700

net_sched: fix reference counting of tc filter chain

so should apply to 4.14 too.

For 4.15, they can just apply nearly cleanly. However, it is not
straight-forward to cherry pick them to 4.14 due to some big changes
between 4.14 and 4.15.

Please let me know how you want to handle this for 4.14.

Thanks.