[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20240911201402.GB16757@redhat.com>
Date: Wed, 11 Sep 2024 22:14:02 +0200
From: Oleg Nesterov <oleg@...hat.com>
To: Linus Torvalds <torvalds@...ux-foundation.org>
Cc: Roman Kisel <romank@...ux.microsoft.com>, akpm@...ux-foundation.org,
apais@...rosoft.com, benhill@...rosoft.com, ebiederm@...ssion.com,
linux-kernel@...r.kernel.org, ssengar@...rosoft.com,
sunilmut@...rosoft.com, vdso@...bites.dev, workingjubilee@...il.com
Subject: Re: [PATCH 1/1] ptrace: Get tracer PID without reliance on the proc
FS
On 09/11, Linus Torvalds wrote:
>
> On Wed, 11 Sept 2024 at 12:54, Oleg Nesterov <oleg@...hat.com> wrote:
> >
> > - please try to make your changelog more convincing. And in particular,
> > please explain why !!current->ptrace is not enough and this feature
> > needs the tracer's pid.
>
> Oleg, I realize you like the simpler patch that only has that
^^^^^^^^
No, no, I don't!!! ;)
> "!!current->ptrace", but my point is that even that simpler patch is
> simply WRONG, WRONG, WRONG.
and I agree, agree, agree.
> There is simply no valid situation where a "I have a tracer" is a good
> thing to test for.
Yes, yes, and that is why I added you/Eric to this discussion.
I just tried to play fair. I just thought that I can't simply "nack" this
change, because I can't explain why I didn't like the whole idea.
Oleg.
Powered by blists - more mailing lists