[meta-intel] [meta-fri2] Reboot hangs

[Andrei Gherzan]

Hello,

On a master, core-image-minimal I get a very strange behavior while
rebooting the board:

1. If board is booted without eth cable, reboot works as expected.
2. If board is booted with eth cable, reboot hangs at "The system is going
down for reboot NOW!" on serial console and at "Sending all preocesses the
TERM signal..." on console.
3. If board is booted without eth cable and I plug in the cable before
reboot, the board hangs with:
The system is going down for reboot NOW!
root at fri2:~# irq 23: nobody cared (try booting with the "irqpoll" option)
CPU: 0 PID: 0 Comm: swapper/0 Not tainted 3.10.17-yocto-standard #2
Hardware name: Intel Corp. Fish River Island II, BIOS Tiano-TunnelCreek
11/19/2010
 ed9ef3c0 ed9ef3c0 ed809ec4 c16872c4 ed809ee4 c10a1d8a c1841650 00000017
 00000000 f8128522 ed9ef3c0 00000017 ed809f08 c10a214d 07eef203 00000017
 c103a341 00000000 ed9ef3c0 ed9ef410 00000000 ed809f4c c10a00a9 ed809f50
Call Trace:
 [<c16872c4>] dump_stack+0x16/0x18
 [<c10a1d8a>] __report_bad_irq+0x2a/0xd0
 [<c10a214d>] note_interrupt+0xfd/0x1a0
 [<c103a341>] ? __do_softirq+0x81/0x210
 [<c10a00a9>] handle_irq_event_percpu+0xf9/0x230
 [<c13d6a5f>] ? credit_entropy_bits.part.11+0xef/0x1e0
 [<c10a0211>] handle_irq_event+0x31/0x50
 [<c10a2b90>] handle_fasteoi_irq+0x50/0xe0
 [<c109f844>] generic_handle_irq+0x24/0x40
 [<c135533e>] pch_gpio_handler+0x4e/0x70
 [<c10a0005>] handle_irq_event_percpu+0x55/0x230
 [<c168eadc>] ? add_preempt_count+0xc/0xb0
 [<c168eb8c>] ? sub_preempt_count+0xc/0xb0
 [<c168b6b7>] ? _raw_spin_unlock_irqrestore+0x17/0x40
 [<c10a0211>] handle_irq_event+0x31/0x50
 [<c10a2b40>] ? unmask_irq+0x30/0x30
 [<c10a2b90>] handle_fasteoi_irq+0x50/0xe0
 <IRQ>  [<c169250d>] ? do_IRQ+0x3d/0xb0
 [<c103a5d8>] ? irq_exit+0x58/0x90
 [<c16925da>] ? smp_apic_timer_interrupt+0x5a/0x8b
 [<c16923ec>] ? common_interrupt+0x2c/0x34
 [<c104007b>] ? ftrace_raw_output_timer_class+0x1b/0x50
 [<c103a341>] ? __do_softirq+0x81/0x210
 [<c103a2c0>] ? __tasklet_hrtimer_trampoline+0x40/0x40
 <IRQ>  [<c103a5ee>] ? irq_exit+0x6e/0x90
 [<c16925da>] ? smp_apic_timer_interrupt+0x5a/0x8b
 [<c168c001>] ? apic_timer_interrupt+0x2d/0x34
 [<c107007b>] ? print_cfs_rq+0x4cb/0x1870
 [<c14d8f10>] ? cpuidle_enter_state+0x40/0xc0
 [<c14d9033>] ? cpuidle_idle_call+0xa3/0x220
 [<c1008a8d>] ? arch_cpu_idle+0xd/0x30
 [<c10755d6>] ? cpu_startup_entry+0x106/0x1f0
 [<c1676591>] ? rest_init+0x71/0x80
 [<c1970a18>] ? start_kernel+0x338/0x33f
 [<c1970514>] ? repair_env_string+0x51/0x51
 [<c1970358>] ? i386_start_kernel+0x12e/0x131
handlers:
[<c1541910>] azx_interrupt
Disabling IRQ #23

All these three use-cases seem to be 100% reproducible on my board. Any
hints on this?

-- 
*ag*
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.yoctoproject.org/pipermail/meta-intel/attachments/20131205/37c545ec/attachment-0001.html>