On Mon, Jan 06, 2025 at 11:41:20AM +0100, gregkh@linuxfoundation.org wrote:
This is a note to let you know that I've just added the patch titled
vmstat: disable vmstat_work on vmstat_cpu_down_prep()to the 6.12-stable tree which can be found at: http://www.kernel.org/git/?p=linux/kernel/git/stable/stable-queue.git%3Ba=su...
The filename of the patch is: vmstat-disable-vmstat_work-on-vmstat_cpu_down_prep.patch and it can be found in the queue-6.12 subdirectory.
If you, or anyone else, feels it should not be added to the stable tree, please let stable@vger.kernel.org know about it.
Hi, could you hold off on adding this for now? It's broken [1] and needs to be fixed. Once a follow-up fix is ready, I'll make sure to notify you.
[1] https://lore.kernel.org/linux-mm/7ed97096-859e-46d0-8f27-16a2298a8914@lucife...
Thanks.
-Koichiro Den
From adcfb264c3ed51fbbf5068ddf10d309a63683868 Mon Sep 17 00:00:00 2001 From: Koichiro Den koichiro.den@canonical.com Date: Sat, 21 Dec 2024 12:33:20 +0900 Subject: vmstat: disable vmstat_work on vmstat_cpu_down_prep()
From: Koichiro Den koichiro.den@canonical.com
commit adcfb264c3ed51fbbf5068ddf10d309a63683868 upstream.
Even after mm/vmstat:online teardown, shepherd may still queue work for the dying cpu until the cpu is removed from online mask. While it's quite rare, this means that after unbind_workers() unbinds a per-cpu kworker, it potentially runs vmstat_update for the dying CPU on an irrelevant cpu before entering atomic AP states. When CONFIG_DEBUG_PREEMPT=y, it results in the following error with the backtrace.
BUG: using smp_processor_id() in preemptible [00000000] code: \ kworker/7:3/1702 caller is refresh_cpu_vm_stats+0x235/0x5f0 CPU: 0 UID: 0 PID: 1702 Comm: kworker/7:3 Tainted: G Tainted: [N]=TEST Workqueue: mm_percpu_wq vmstat_update Call Trace:
<TASK> dump_stack_lvl+0x8d/0xb0 check_preemption_disabled+0xce/0xe0 refresh_cpu_vm_stats+0x235/0x5f0 vmstat_update+0x17/0xa0 process_one_work+0x869/0x1aa0 worker_thread+0x5e5/0x1100 kthread+0x29e/0x380 ret_from_fork+0x2d/0x70 ret_from_fork_asm+0x1a/0x30 </TASK>
So, for mm/vmstat:online, disable vmstat_work reliably on teardown and symmetrically enable it on startup.
Link: https://lkml.kernel.org/r/20241221033321.4154409-1-koichiro.den@canonical.co... Signed-off-by: Koichiro Den koichiro.den@canonical.com Cc: Sebastian Andrzej Siewior bigeasy@linutronix.de Cc: stable@vger.kernel.org Signed-off-by: Andrew Morton akpm@linux-foundation.org Signed-off-by: Greg Kroah-Hartman gregkh@linuxfoundation.org
mm/vmstat.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-)
--- a/mm/vmstat.c +++ b/mm/vmstat.c @@ -2139,13 +2139,14 @@ static int vmstat_cpu_online(unsigned in if (!node_state(cpu_to_node(cpu), N_CPU)) { node_set_state(cpu_to_node(cpu), N_CPU); }
- enable_delayed_work(&per_cpu(vmstat_work, cpu));
return 0; } static int vmstat_cpu_down_prep(unsigned int cpu) {
- cancel_delayed_work_sync(&per_cpu(vmstat_work, cpu));
- disable_delayed_work_sync(&per_cpu(vmstat_work, cpu)); return 0;
}
Patches currently in stable-queue which might be from koichiro.den@canonical.com are
queue-6.12/vmstat-disable-vmstat_work-on-vmstat_cpu_down_prep.patch