[<prev] [next>] [thread-next>] [day] [month] [year] [list]
Message-Id: <20250403-work-pidfd-fixes-v1-0-a123b6ed6716@kernel.org>
Date: Thu, 03 Apr 2025 16:09:00 +0200
From: Christian Brauner <brauner@...nel.org>
To: Oleg Nesterov <oleg@...hat.com>
Cc: linux-fsdevel@...r.kernel.org, Jeff Layton <jlayton@...nel.org>,
Lennart Poettering <lennart@...ttering.net>,
Daan De Meyer <daan.j.demeyer@...il.com>, Mike Yuan <me@...dnzj.com>,
linux-kernel@...r.kernel.org, Christian Brauner <brauner@...nel.org>
Subject: [PATCH RFC 0/4] pidfd: improve uapi when task isn't found
Oleg,
We currently report EINVAL whenever a struct pid has no tasked attached
anymore thereby conflating two concepts:
(1) The task has already been reaped.
(2) The caller requested a pidfd for a thread-group leader but the pid
actually references a struct pid that isn't used as a thread-group
leader.
This is causing issues for non-threaded workloads as in [1] where they
expect ESRCH to be reported, not EINVAL. I think that's a very resonable
assumption.
This patch tries to allow userspace to distinguish between (1) and (2).
This is racy of course but that shouldn't matter.
Christian
Signed-off-by: Christian Brauner <brauner@...nel.org>
---
Christian Brauner (4):
selftests/pidfd: adapt to recent changes
pidfd: remove unneeded NULL check from pidfd_prepare()
pidfd: improve uapi when task isn't found
selftest/pidfd: add test for thread-group leader pidfd open for thread
kernel/fork.c | 31 ++++++++++++++++++++++---
tools/testing/selftests/pidfd/pidfd_info_test.c | 13 ++++++-----
2 files changed, 35 insertions(+), 9 deletions(-)
---
base-commit: a2cc6ff5ec8f91bc463fd3b0c26b61166a07eb11
change-id: 20250403-work-pidfd-fixes-54c5b13ee0ee
Powered by blists - more mailing lists