Re: [Linux 2.6.29-rc2] BUG: using smp_processor_id() in preemptible

[Andrew Morton <akpm@linux-foundation.org>]

On Sat, 17 Jan 2009 09:58:47 +0100 "Maciej Rutecki" <maciej.rutecki@gmail.com> wrote:

> During suspend to ram:
> [  131.287012] BUG: using smp_processor_id() in preemptible [00000000]
> code: suspend_to_ram./2958
> [  131.287012] caller is retrigger_next_event+0x13/0xb0
> [  131.287012] Pid: 2958, comm: suspend_to_ram. Not tainted 2.6.29-rc2 #1
> [  131.287012] Call Trace:
> [  131.287012]  [<c025b41f>] debug_smp_processor_id+0xbf/0xd0
> [  131.287012]  [<c01473b3>] retrigger_next_event+0x13/0xb0
> [  131.287012]  [<c01489b7>] raw_notifier_call_chain+0x17/0x20
> [  131.287012]  [<c014b938>] timekeeping_resume+0xe8/0x110
> [  131.287012]  [<c02cc651>] __sysdev_resume+0x11/0x50
> [  131.287012]  [<c02cc6d7>] sysdev_resume+0x47/0x80
> [  131.287012]  [<c02d2478>] device_power_up+0x8/0x10
> [  131.287012]  [<c0159c7b>] suspend_devices_and_enter+0xeb/0x160
> [  131.287012]  [<c015a47f>] freeze_processes+0x4f/0xa0
> [  131.287012]  [<c0159e3d>] enter_state+0xfd/0x150
> [  131.287012]  [<c0159f1c>] state_store+0x8c/0xd0
> [  131.287012]  [<c0159e90>] state_store+0x0/0xd0
> [  131.287012]  [<c0250d54>] kobj_attr_store+0x24/0x30
> [  131.287012]  [<c01e25d1>] sysfs_write_file+0xa1/0x110
> [  131.287012]  [<c01e2530>] sysfs_write_file+0x0/0x110
> [  131.287012]  [<c019bf48>] vfs_write+0xa8/0x140
> [  131.287012]  [<c019c0b1>] sys_write+0x41/0x80
> [  131.287012]  [<c010348d>] sysenter_do_call+0x12/0x21
> [  131.287012]  [<c03b0000>] init_cyrix+0x72/0x444
> [  131.287012] BUG: using smp_processor_id() in preemptible [00000000]
> code: suspend_to_ram./2958

Coming out of the hrtimer_hres_active() call, I assume.

Is this new behaviour, or were 2.6.28 and earlier doing it?

If it's new behaviour then I'm not sure why, but it triggers my Rusty
nerve.

afaict we should

a) remove the wrong "Called with interrupts disabled via
   on_each_cpu()" comment from retrigger_next_event(), and replace it
   with something useful and truthful.

b) change hres_timers_resume() to do local_irq_save/restore


Does this fix it?

--- a/kernel/hrtimer.c~a
+++ a/kernel/hrtimer.c
@@ -562,7 +562,7 @@ static int hrtimer_reprogram(struct hrti
 /*
  * Retrigger next event is called after clock was set
  *
- * Called with interrupts disabled via on_each_cpu()
+ * Called with interrupts disabled.
  */
 static void retrigger_next_event(void *arg)
 {
@@ -614,8 +614,12 @@ void clock_was_set(void)
  */
 void hres_timers_resume(void)
 {
+	unsigned long flags;
+
 	/* Retrigger the CPU local events: */
+	local_irq_save(flags);
 	retrigger_next_event(NULL);
+	local_irq_restore(flags);
 }
 
 /*
_