]> git.ipfire.org Git - thirdparty/kernel/stable-queue.git/commitdiff
drop queue-5.4/posix-cpu-timers-fix-race-between-handle_posix_cpu_timers-and-posix_cp...
authorGreg Kroah-Hartman <gregkh@linuxfoundation.org>
Tue, 17 Jun 2025 14:48:00 +0000 (16:48 +0200)
committerGreg Kroah-Hartman <gregkh@linuxfoundation.org>
Tue, 17 Jun 2025 14:48:00 +0000 (16:48 +0200)
Broke the build

queue-5.4/posix-cpu-timers-fix-race-between-handle_posix_cpu_timers-and-posix_cpu_timer_del.patch [deleted file]
queue-5.4/series

diff --git a/queue-5.4/posix-cpu-timers-fix-race-between-handle_posix_cpu_timers-and-posix_cpu_timer_del.patch b/queue-5.4/posix-cpu-timers-fix-race-between-handle_posix_cpu_timers-and-posix_cpu_timer_del.patch
deleted file mode 100644 (file)
index baaaeed..0000000
+++ /dev/null
@@ -1,55 +0,0 @@
-From f90fff1e152dedf52b932240ebbd670d83330eca Mon Sep 17 00:00:00 2001
-From: Oleg Nesterov <oleg@redhat.com>
-Date: Fri, 13 Jun 2025 19:26:50 +0200
-Subject: posix-cpu-timers: fix race between handle_posix_cpu_timers() and posix_cpu_timer_del()
-MIME-Version: 1.0
-Content-Type: text/plain; charset=UTF-8
-Content-Transfer-Encoding: 8bit
-
-From: Oleg Nesterov <oleg@redhat.com>
-
-commit f90fff1e152dedf52b932240ebbd670d83330eca upstream.
-
-If an exiting non-autoreaping task has already passed exit_notify() and
-calls handle_posix_cpu_timers() from IRQ, it can be reaped by its parent
-or debugger right after unlock_task_sighand().
-
-If a concurrent posix_cpu_timer_del() runs at that moment, it won't be
-able to detect timer->it.cpu.firing != 0: cpu_timer_task_rcu() and/or
-lock_task_sighand() will fail.
-
-Add the tsk->exit_state check into run_posix_cpu_timers() to fix this.
-
-This fix is not needed if CONFIG_POSIX_CPU_TIMERS_TASK_WORK=y, because
-exit_task_work() is called before exit_notify(). But the check still
-makes sense, task_work_add(&tsk->posix_cputimers_work.work) will fail
-anyway in this case.
-
-Cc: stable@vger.kernel.org
-Reported-by: BenoĆ®t Sevens <bsevens@google.com>
-Fixes: 0bdd2ed4138e ("sched: run_posix_cpu_timers: Don't check ->exit_state, use lock_task_sighand()")
-Signed-off-by: Oleg Nesterov <oleg@redhat.com>
-Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
-Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
----
- kernel/time/posix-cpu-timers.c |    9 +++++++++
- 1 file changed, 9 insertions(+)
-
---- a/kernel/time/posix-cpu-timers.c
-+++ b/kernel/time/posix-cpu-timers.c
-@@ -1321,6 +1321,15 @@ static int posix_cpu_nsleep(const clocki
-       int error;
-       /*
-+       * Ensure that release_task(tsk) can't happen while
-+       * handle_posix_cpu_timers() is running. Otherwise, a concurrent
-+       * posix_cpu_timer_del() may fail to lock_task_sighand(tsk) and
-+       * miss timer->it.cpu.firing != 0.
-+       */
-+      if (tsk->exit_state)
-+              return;
-+
-+      /*
-        * Diagnose required errors first.
-        */
-       if (CPUCLOCK_PERTHREAD(which_clock) &&
index dbf75ba4425575299dacfd682d09cff346121139..ea7fac27f02ae4d71b11b4def01b9413c5d317c1 100644 (file)
@@ -84,7 +84,6 @@ net_sched-prio-fix-a-race-in-prio_tune.patch
 net_sched-red-fix-a-race-in-__red_change.patch
 net_sched-tbf-fix-a-race-in-tbf_change.patch
 net-mdio-c22-is-now-optional-eopnotsupp-if-not-provi.patch
-posix-cpu-timers-fix-race-between-handle_posix_cpu_timers-and-posix_cpu_timer_del.patch
 x86-boot-compressed-prefer-cc-option-for-cflags-additions.patch
 mips-move-wa-msoft-float-check-from-as-option-to-cc-option.patch
 kbuild-update-assembler-calls-to-use-proper-flags-and-language-target.patch