Re: [x86-tip] panic during cpu_up

[Dhaval Giani <dhaval@linux.vnet.ibm.com>]

> > 
> > The other suspect would be acpi idle:
> > 
> >         /* Fall back to the default idle loop */
> >         pm_idle = pm_idle_save;
> > 

Yep, you are right.

> > Could you try latest tip/master or the debug patch below? It should show 
> > where the NULL comes from, without crashing.
> > 
> 
> Sure, will give it a run sometime soon.
> 

------------[ cut here ]------------
WARNING: at drivers/acpi/processor_idle.c:1316
acpi_processor_cst_has_changed+0x64/0xc0()
Modules linked in:
Pid: 4433, comm: bash Not tainted 2.6.26-rc8-tip #15
 [<c0127a0a>] warn_on_slowpath+0x41/0x60
 [<c0393a00>] ? __cpufreq_set_policy+0x135/0x1bf
 [<c0141c28>] ? mark_held_locks+0x46/0x61
 [<c0141d7d>] ? trace_hardirqs_on+0xb/0xd
 [<c0141d4a>] ? trace_hardirqs_on_caller+0xe9/0x111
 [<c0141d7d>] ? trace_hardirqs_on+0xb/0xd
 [<c027fc1d>] ? acpi_processor_get_platform_limit+0x9f/0xac
 [<c027f285>] acpi_processor_cst_has_changed+0x64/0xc0
 [<c027cadd>] acpi_cpu_soft_notify+0x2a/0x39
 [<c041193c>] notifier_call_chain+0x32/0x64
 [<c013b87f>] __raw_notifier_call_chain+0xe/0x10
 [<c013b88d>] raw_notifier_call_chain+0xc/0xe
 [<c040cb33>] _cpu_up+0xb3/0xdc
 [<c040cb9e>] cpu_up+0x42/0x52
 [<c03fab35>] store_online+0x39/0x5d
 [<c03faafc>] ? store_online+0x0/0x5d
 [<c02c1e6c>] sysdev_store+0x20/0x25
 [<c01b2190>] flush_write_buffer+0x3e/0x53
 [<c01b21e3>] sysfs_write_file+0x3e/0x5d
 [<c017e281>] vfs_write+0x8d/0x105
 [<c017e394>] sys_write+0x3b/0x60
 [<c0103931>] sysenter_past_esp+0x6a/0xa5
 =======================
---[ end trace 987615de2ae7cfa5 ]---

Which corresponds to 

        /* Fall back to the default idle loop */
        WARN_ON_ONCE(!pm_idle_save);
        if (pm_idle_save)
                pm_idle = pm_idle_save;
        synchronize_sched();    /* Relies on interrupts forcing exit from idle. */


-- 
regards,
Dhaval