2.5.66-bk12: acpi_power_off: sleeping function called from illegal context

2.5.66-bk12: acpi_power_off: sleeping function called from illegal context

[Felipe Alfaro Solana]

Hi,

I'm seeing this oops on my P4 box running 2.5.66-bk12 + ACPI when
shutting it down (copied by hand):

Power down.
acpi_power_off called
Debug: sleeping function called from illegal context at
include/asm/semaphore.h: 119
Call  Trace:
 [<c012088a>] __might_sleep+0x5f/0x75
 [<c01ffb10>] acpi_os_wait_semaphore+0xc5/0xea
 [<c021440c>] acpi_ut_acquire_mutex+0x51/0x73
 [<c020ab4f>] acpi_set_register+0x34/0x15d
 [<c020b1f6>] acpi_enter_sleep_state+0x77/0x1ab
 [<c0215d7d>] acpi_power_off+0x21/0x23
 [<c011a3e5>] machine_power_off+0x10/0x13
 [<c0135f7b>] sys_reboot+0x332/0x741
 [<c011e010>] schedule+0x210/0x6d7
 [<c0130daf>] group_send_sig_info+0x2af/0x6b6
 [<c011e50d>] preempt_schedule+0x36/0x50
 [<c0131384>] kill_proc_info+0x60/0x62
 [<c0134346>] sys_kill+0x4d/0x51
 [<c016e76b>] __fput+0xaf/0xfb
 [<c016cabc>] filp_close+0x160/0x226
 [<c01850cb>] sys_ioctl+0x197/0x3e8
 [<c010af29>] sysenter_past_esp+0x52/0x71

This is not 100% reproducible. When using ACPI, the machine is *never*
able to properly shut it off (only APM is able to turn the machine off).

________________________________________________________________________
Linux Registered User #287198

Re: 2.5.66-bk12: acpi_power_off: sleeping function called from illegal context

[Zwane Mwaikambo]

On Sun, 7 Apr 2003, Felipe Alfaro Solana wrote:

> Power down.
> acpi_power_off called
> Debug: sleeping function called from illegal context at
> include/asm/semaphore.h: 119
> Call  Trace:
>  [<c012088a>] __might_sleep+0x5f/0x75
>  [<c01ffb10>] acpi_os_wait_semaphore+0xc5/0xea
>  [<c021440c>] acpi_ut_acquire_mutex+0x51/0x73
>  [<c020ab4f>] acpi_set_register+0x34/0x15d
>  [<c020b1f6>] acpi_enter_sleep_state+0x77/0x1ab
>  [<c0215d7d>] acpi_power_off+0x21/0x23
>  [<c011a3e5>] machine_power_off+0x10/0x13
>  [<c0135f7b>] sys_reboot+0x332/0x741
>  [<c011e010>] schedule+0x210/0x6d7
>  [<c0130daf>] group_send_sig_info+0x2af/0x6b6
>  [<c011e50d>] preempt_schedule+0x36/0x50
>  [<c0131384>] kill_proc_info+0x60/0x62
>  [<c0134346>] sys_kill+0x4d/0x51
>  [<c016e76b>] __fput+0xaf/0xfb
>  [<c016cabc>] filp_close+0x160/0x226
>  [<c01850cb>] sys_ioctl+0x197/0x3e8
>  [<c010af29>] sysenter_past_esp+0x52/0x71

You probably need this;

Index: linux-2.5.66/drivers/acpi/osl.c
===================================================================
RCS file: /build/cvsroot/linux-2.5.66/drivers/acpi/osl.c,v
retrieving revision 1.1.1.1
diff -u -p -B -r1.1.1.1 osl.c
--- linux-2.5.66/drivers/acpi/osl.c	24 Mar 2003 23:39:26 -0000	1.1.1.1
+++ linux-2.5.66/drivers/acpi/osl.c	6 Apr 2003 22:46:04 -0000
@@ -750,7 +750,7 @@ acpi_os_wait_semaphore(
 
 	ACPI_DEBUG_PRINT ((ACPI_DB_MUTEX, "Waiting for semaphore[%p|%d|%d]\n", handle, units, timeout));
 
-	if (in_atomic())
+	if (in_atomic() || irqs_disabled())
 		timeout = 0;
 
 	switch (timeout)
-- 
function.linuxpower.ca

Re: 2.5.66-bk12: acpi_power_off: sleeping function called from illegal context

[Andrew Morton]

Zwane Mwaikambo <zwane@linuxpower.ca> wrote:
>
> -	if (in_atomic())
> +	if (in_atomic() || irqs_disabled())
>  		timeout = 0;

Andy, why does the ACPI code have this test?

Is it to determine whether a caller of this functon is currently holding a
spinlock?  If so then it will only work on a preemptible kernel.

A non-preempt kernel will not increment preempt_count() when it takes a
spinlock and ACPI could mistakenly schedule away and cause a system deadlock.

Re: 2.5.66-bk12: acpi_power_off: sleeping function called from illegal context

[Felipe Alfaro Solana]

On Mon, 2003-04-07 at 00:48, Zwane Mwaikambo wrote:
> On Sun, 7 Apr 2003, Felipe Alfaro Solana wrote:
> 
> > Power down.
> > acpi_power_off called
> > Debug: sleeping function called from illegal context at
> > include/asm/semaphore.h: 119
> > Call  Trace:
> >  [<c012088a>] __might_sleep+0x5f/0x75
> >  [<c01ffb10>] acpi_os_wait_semaphore+0xc5/0xea
> >  [<c021440c>] acpi_ut_acquire_mutex+0x51/0x73
> >  [<c020ab4f>] acpi_set_register+0x34/0x15d
> >  [<c020b1f6>] acpi_enter_sleep_state+0x77/0x1ab
> >  [<c0215d7d>] acpi_power_off+0x21/0x23
> >  [<c011a3e5>] machine_power_off+0x10/0x13
> >  [<c0135f7b>] sys_reboot+0x332/0x741
> >  [<c011e010>] schedule+0x210/0x6d7
> >  [<c0130daf>] group_send_sig_info+0x2af/0x6b6
> >  [<c011e50d>] preempt_schedule+0x36/0x50
> >  [<c0131384>] kill_proc_info+0x60/0x62
> >  [<c0134346>] sys_kill+0x4d/0x51
> >  [<c016e76b>] __fput+0xaf/0xfb
> >  [<c016cabc>] filp_close+0x160/0x226
> >  [<c01850cb>] sys_ioctl+0x197/0x3e8
> >  [<c010af29>] sysenter_past_esp+0x52/0x71
> 
> You probably need this;
> 
> Index: linux-2.5.66/drivers/acpi/osl.c
> ===================================================================
> RCS file: /build/cvsroot/linux-2.5.66/drivers/acpi/osl.c,v
> retrieving revision 1.1.1.1
> diff -u -p -B -r1.1.1.1 osl.c
> --- linux-2.5.66/drivers/acpi/osl.c	24 Mar 2003 23:39:26 -0000	1.1.1.1
> +++ linux-2.5.66/drivers/acpi/osl.c	6 Apr 2003 22:46:04 -0000
> @@ -750,7 +750,7 @@ acpi_os_wait_semaphore(
>  
>  	ACPI_DEBUG_PRINT ((ACPI_DB_MUTEX, "Waiting for semaphore[%p|%d|%d]\n", handle, units, timeout));
>  
> -	if (in_atomic())
> +	if (in_atomic() || irqs_disabled())
>  		timeout = 0;
>  
>  	switch (timeout)

OK, now it doesn't cause an oops, but the machine doesn't power off: it
stays powered on showing:

acpi_power_off

and I must manually cycle the power.
This 2.5 ACPI stuff is causing me great deal of pain... I can never get
it to work realiably.

________________________________________________________________________
Linux Registered User #287198