[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20250506-zertifikat-teint-d866c715291a@brauner>
Date: Tue, 6 May 2025 09:39:22 +0200
From: Christian Brauner <brauner@...nel.org>
To: Jann Horn <jannh@...gle.com>
Cc: Kuniyuki Iwashima <kuniyu@...zon.com>, alexander@...alicyn.com,
bluca@...ian.org, daan.j.demeyer@...il.com, davem@...emloft.net,
david@...dahead.eu, edumazet@...gle.com, horms@...nel.org, jack@...e.cz,
kuba@...nel.org, lennart@...ttering.net, linux-fsdevel@...r.kernel.org,
linux-kernel@...r.kernel.org, me@...dnzj.com, netdev@...r.kernel.org, oleg@...hat.com,
pabeni@...hat.com, viro@...iv.linux.org.uk, zbyszek@...waw.pl
Subject: Re: [PATCH RFC v3 08/10] net, pidfs, coredump: only allow
coredumping tasks to connect to coredump socket
> ("a kernel socket" is not necessarily the same as "a kernel socket
> intended for core dumping")
Indeed. The usermodehelper is a kernel protocol. Here it's the task with
its own credentials that's connecting to a userspace socket. Which makes
this very elegant because it's just userspace IPC. No one is running
around with kernel credentials anywhere.
Powered by blists - more mailing lists