On 04/01, Eric W. Biederman wrote:
>
> These patches only handle the case when sighand_struct is not
> shared between different multi-threaded processes.
Ah, I didn't read 0/2 when looked at 2/2, so at least you documented this.
> The general
> case is solvable but that is a quite a bit more code
I spent a little more time with this and only waiting until the killed
thread are zombies (and not reaped as we do today) really looks like
the right fix.
Oleg the following two patches work on top of your PTRACE_EVENT_EXIT
change and probably need a little more cleanup until they are ready
for s
2 matches
Mail list logo