This is a note to let you know that I've just added the patch titled exit: Use READ_ONCE() for all oops/warn limit reads to the 6.1-stable tree which can be found at: http://www.kernel.org/git/?p=linux/kernel/git/stable/stable-queue.git;a=summary The filename of the patch is: exit-use-read_once-for-all-oops-warn-limit-reads.patch and it can be found in the queue-6.1 subdirectory. If you, or anyone else, feels it should not be added to the stable tree, please let <stable@xxxxxxxxxxxxxxx> know about it. >From 7535b832c6399b5ebfc5b53af5c51dd915ee2538 Mon Sep 17 00:00:00 2001 From: Kees Cook <keescook@xxxxxxxxxxxx> Date: Fri, 16 Dec 2022 12:26:57 -0800 Subject: exit: Use READ_ONCE() for all oops/warn limit reads From: Kees Cook <keescook@xxxxxxxxxxxx> commit 7535b832c6399b5ebfc5b53af5c51dd915ee2538 upstream. Use a temporary variable to take full advantage of READ_ONCE() behavior. Without this, the report (and even the test) might be out of sync with the initial test. Reported-by: Peter Zijlstra <peterz@xxxxxxxxxxxxx> Link: https://lore.kernel.org/lkml/Y5x7GXeluFmZ8E0E@xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx Fixes: 9fc9e278a5c0 ("panic: Introduce warn_limit") Fixes: d4ccd54d28d3 ("exit: Put an upper limit on how often we can oops") Cc: "Eric W. Biederman" <ebiederm@xxxxxxxxxxxx> Cc: Jann Horn <jannh@xxxxxxxxxx> Cc: Arnd Bergmann <arnd@xxxxxxxx> Cc: Petr Mladek <pmladek@xxxxxxxx> Cc: Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx> Cc: Luis Chamberlain <mcgrof@xxxxxxxxxx> Cc: Marco Elver <elver@xxxxxxxxxx> Cc: tangmeng <tangmeng@xxxxxxxxxxxxx> Cc: Sebastian Andrzej Siewior <bigeasy@xxxxxxxxxxxxx> Cc: Tiezhu Yang <yangtiezhu@xxxxxxxxxxx> Signed-off-by: Kees Cook <keescook@xxxxxxxxxxxx> Signed-off-by: Greg Kroah-Hartman <gregkh@xxxxxxxxxxxxxxxxxxx> --- kernel/exit.c | 6 ++++-- kernel/panic.c | 7 +++++-- 2 files changed, 9 insertions(+), 4 deletions(-) --- a/kernel/exit.c +++ b/kernel/exit.c @@ -931,6 +931,7 @@ void __noreturn make_task_dead(int signr * Then do everything else. */ struct task_struct *tsk = current; + unsigned int limit; if (unlikely(in_interrupt())) panic("Aiee, killing interrupt handler!"); @@ -954,8 +955,9 @@ void __noreturn make_task_dead(int signr * To make sure this can't happen, place an upper bound on how often the * kernel may oops without panic(). */ - if (atomic_inc_return(&oops_count) >= READ_ONCE(oops_limit) && oops_limit) - panic("Oopsed too often (kernel.oops_limit is %d)", oops_limit); + limit = READ_ONCE(oops_limit); + if (atomic_inc_return(&oops_count) >= limit && limit) + panic("Oopsed too often (kernel.oops_limit is %d)", limit); /* * We're taking recursive faults here in make_task_dead. Safest is to just --- a/kernel/panic.c +++ b/kernel/panic.c @@ -231,12 +231,15 @@ static void panic_print_sys_info(bool co void check_panic_on_warn(const char *origin) { + unsigned int limit; + if (panic_on_warn) panic("%s: panic_on_warn set ...\n", origin); - if (atomic_inc_return(&warn_count) >= READ_ONCE(warn_limit) && warn_limit) + limit = READ_ONCE(warn_limit); + if (atomic_inc_return(&warn_count) >= limit && limit) panic("%s: system warned too often (kernel.warn_limit is %d)", - origin, warn_limit); + origin, limit); } /** Patches currently in stable-queue which might be from keescook@xxxxxxxxxxxx are queue-6.1/panic-consolidate-open-coded-panic_on_warn-checks.patch queue-6.1/exit-put-an-upper-limit-on-how-often-we-can-oops.patch queue-6.1/panic-introduce-warn_limit.patch queue-6.1/exit-allow-oops_limit-to-be-disabled.patch queue-6.1/panic-separate-sysctl-logic-from-config_smp.patch queue-6.1/exit-use-read_once-for-all-oops-warn-limit-reads.patch queue-6.1/exit-expose-oops_count-to-sysfs.patch queue-6.1/panic-expose-warn_count-to-sysfs.patch queue-6.1/docs-fix-path-paste-o-for-sys-kernel-warn_count.patch