Subject: Re: [PATCH 5/9] signal: Protect parent child relationships by childs siglock

On 2022-04-26 17:52:07 [-0500], Eric W. Biederman wrote:
> The functions ptrace_stop and do_signal_stop have to drop siglock
> and grab tasklist_lock because the parent/child relation ship
> is guarded by siglock and not siglock.

"is guarded by tasklist_lock and not siglock." ?

> Simplify things by guarding the parent/child relationship
> with siglock. For the most part this just requires a little bit
> of code motion. In a couple of places more locking was needed.
>
>
> Signed-off-by: "Eric W. Biederman" <[email protected]>

Sebastian


2022-04-27 14:07:02

by Eric W. Biederman

[permalink] [raw]
Subject: Re: [PATCH 5/9] signal: Protect parent child relationships by childs siglock

Sebastian Andrzej Siewior <[email protected]> writes:

> On 2022-04-26 17:52:07 [-0500], Eric W. Biederman wrote:
>> The functions ptrace_stop and do_signal_stop have to drop siglock
>> and grab tasklist_lock because the parent/child relation ship
>> is guarded by siglock and not siglock.
>
> "is guarded by tasklist_lock and not siglock." ?

Yes. Thank you. I will fix that.

>> Simplify things by guarding the parent/child relationship
>> with siglock. For the most part this just requires a little bit
>> of code motion. In a couple of places more locking was needed.
>>
>>
>> Signed-off-by: "Eric W. Biederman" <[email protected]>
>
> Sebastian

Eric