2009-09-08 18:57:43

by Alex Bennee

[permalink] [raw]
Subject: Machine resumes straight after suspend to RAM (or disk)

Hi,

I thought I'd enable suspend and resume on my desktop machine so I
could save myself some time in the morning when I restart the machine.
The machine is fairly standard Intel Core Duo based system (lspci
attached). However when I execute pm-suspend (or Gentoo's own
hibernate-script) the system does suspend but resumes immediately. To
add insult to injury the ethernet doesn't come back properly and needs
a reboot to reset it.

I've played about a little with the ACPI debug stuff but it's easy to
overflow the dmesg buffer. The relevant section around the
suspend/resume looks like:

[ 1025.178019] Disabling non-boot CPUs ...
[ 1025.182071] kvm: disabling virtualization on CPU1
[ 1025.227729] CPU 1 is now offline
[ 1025.227731] SMP alternatives: switching to UP code
[ 1025.232627] CPU0 attaching NULL sched-domain.
[ 1025.232630] CPU1 attaching NULL sched-domain.
[ 1025.232634] CPU0 attaching NULL sched-domain.
[ 1025.232771] CPU1 is down
[ 1025.233264] acpi_suspend_enter acpi_state=3
[ 1025.233276] evgpeblk-0144 [ffff8800b70d8000] [05] ev_walk_gpe_list
: ----Entry
[ 1025.233286] evgpeblk-0178 [ffff8800b70d8000] [05] ev_walk_gpe_list
: ----Exit- AE_OK
[ 1025.233290] evgpeblk-0144 [ffff8800b70d8000] [05] ev_walk_gpe_list
: ----Entry
[ 1025.233299] evgpeblk-0178 [ffff8800b70d8000] [05] ev_walk_gpe_list
: ----Exit- AE_OK
[ 1025.233303] evgpeblk-0144 [ffff8800b70d8000] [05] ev_walk_gpe_list
: ----Entry
[ 1025.233311] evgpeblk-0178 [ffff8800b70d8000] [05] ev_walk_gpe_list
: ----Exit- AE_OK
[ 1025.233315] evgpeblk-0144 [ffff8800b70d8000] [05] ev_walk_gpe_list
: ----Entry
[ 1025.233319] evgpeblk-0178 [ffff8800b70d8000] [05] ev_walk_gpe_list
: ----Exit- AE_OK
[ 1025.233319] evxfevnt-0074 [ffff8800b70d8000] [05] enable
: ----Entry
[ 1025.233319] evxfevnt-0086 [ffff8800b70d8000] [05] enable
: System is already in ACPI mode
[ 1025.233319] evxfevnt-0101 [ffff8800b70d8000] [05] enable
: ----Exit- AE_OK
[ 1025.233319] evxfevnt-0397 [ffff8800b70d8000] [05] clear_event
: ----Entry
[ 1025.233319] evxfevnt-0413 [ffff8800b70d8000] [05] clear_event
: ----Exit- AE_OK
[ 1025.233319] evxfevnt-0831 [ffff8800b70d8000] [05] disable_all_gpes
: ----Entry
[ 1025.233319] evgpeblk-0144 [ffff8800b70d8000] [07] ev_walk_gpe_list
: ----Entry
[ 1025.233319] evgpeblk-0178 [ffff8800b70d8000] [07] ev_walk_gpe_list
: ----Exit- AE_OK
[ 1025.233319] evgpeblk-0144 [ffff8800b70d8000] [07] ev_walk_gpe_list
: ----Entry
[ 1025.233319] evgpeblk-0178 [ffff8800b70d8000] [07] ev_walk_gpe_list
: ----Exit- AE_OK
[ 1025.233319] evxfevnt-0841 [ffff8800b70d8000] [05] disable_all_gpes
: ----Exit- AE_OK
[ 1025.233319] Back to C!
[ 1025.233319] CPU0: Thermal LVT vector (0xfa) already installed
[ 1025.233319] Enabling non-boot CPUs ...
[ 1025.233319] SMP alternatives: switching to SMP code
[ 1025.237548] Booting processor 1 APIC 0x1 ip 0x6000
[ 1025.232536] Initializing CPU#1
[ 1025.232536] Calibrating delay using timer specific routine..
4245.78 BogoMIPS (lpj=8491564)

Unfortunately the ACPI stuff is all Greek to me. If anyone wants
additional debug information then please let me know.

So any ideas?

--
Alex, homepage: http://www.bennee.com/~alex/
http://www.half-llama.co.uk


Attachments:
lspci.out (6.32 kB)

2009-09-08 19:37:21

by Rafael J. Wysocki

[permalink] [raw]
Subject: Re: Machine resumes straight after suspend to RAM (or disk)

On Tuesday 08 September 2009, Alex Bennee wrote:
> Hi,
>
> I thought I'd enable suspend and resume on my desktop machine so I
> could save myself some time in the morning when I restart the machine.
> The machine is fairly standard Intel Core Duo based system (lspci
> attached). However when I execute pm-suspend (or Gentoo's own
> hibernate-script) the system does suspend but resumes immediately. To
> add insult to injury the ethernet doesn't come back properly and needs
> a reboot to reset it.
>
> I've played about a little with the ACPI debug stuff but it's easy to
> overflow the dmesg buffer. The relevant section around the
> suspend/resume looks like:
>
> [ 1025.178019] Disabling non-boot CPUs ...
> [ 1025.182071] kvm: disabling virtualization on CPU1
> [ 1025.227729] CPU 1 is now offline
> [ 1025.227731] SMP alternatives: switching to UP code
> [ 1025.232627] CPU0 attaching NULL sched-domain.
> [ 1025.232630] CPU1 attaching NULL sched-domain.
> [ 1025.232634] CPU0 attaching NULL sched-domain.
> [ 1025.232771] CPU1 is down
> [ 1025.233264] acpi_suspend_enter acpi_state=3
> [ 1025.233276] evgpeblk-0144 [ffff8800b70d8000] [05] ev_walk_gpe_list
> : ----Entry
> [ 1025.233286] evgpeblk-0178 [ffff8800b70d8000] [05] ev_walk_gpe_list
> : ----Exit- AE_OK
> [ 1025.233290] evgpeblk-0144 [ffff8800b70d8000] [05] ev_walk_gpe_list
> : ----Entry
> [ 1025.233299] evgpeblk-0178 [ffff8800b70d8000] [05] ev_walk_gpe_list
> : ----Exit- AE_OK
> [ 1025.233303] evgpeblk-0144 [ffff8800b70d8000] [05] ev_walk_gpe_list
> : ----Entry
> [ 1025.233311] evgpeblk-0178 [ffff8800b70d8000] [05] ev_walk_gpe_list
> : ----Exit- AE_OK
> [ 1025.233315] evgpeblk-0144 [ffff8800b70d8000] [05] ev_walk_gpe_list
> : ----Entry
> [ 1025.233319] evgpeblk-0178 [ffff8800b70d8000] [05] ev_walk_gpe_list
> : ----Exit- AE_OK
> [ 1025.233319] evxfevnt-0074 [ffff8800b70d8000] [05] enable
> : ----Entry
> [ 1025.233319] evxfevnt-0086 [ffff8800b70d8000] [05] enable
> : System is already in ACPI mode
> [ 1025.233319] evxfevnt-0101 [ffff8800b70d8000] [05] enable
> : ----Exit- AE_OK
> [ 1025.233319] evxfevnt-0397 [ffff8800b70d8000] [05] clear_event
> : ----Entry
> [ 1025.233319] evxfevnt-0413 [ffff8800b70d8000] [05] clear_event
> : ----Exit- AE_OK
> [ 1025.233319] evxfevnt-0831 [ffff8800b70d8000] [05] disable_all_gpes
> : ----Entry
> [ 1025.233319] evgpeblk-0144 [ffff8800b70d8000] [07] ev_walk_gpe_list
> : ----Entry
> [ 1025.233319] evgpeblk-0178 [ffff8800b70d8000] [07] ev_walk_gpe_list
> : ----Exit- AE_OK
> [ 1025.233319] evgpeblk-0144 [ffff8800b70d8000] [07] ev_walk_gpe_list
> : ----Entry
> [ 1025.233319] evgpeblk-0178 [ffff8800b70d8000] [07] ev_walk_gpe_list
> : ----Exit- AE_OK
> [ 1025.233319] evxfevnt-0841 [ffff8800b70d8000] [05] disable_all_gpes
> : ----Exit- AE_OK
> [ 1025.233319] Back to C!
> [ 1025.233319] CPU0: Thermal LVT vector (0xfa) already installed
> [ 1025.233319] Enabling non-boot CPUs ...
> [ 1025.233319] SMP alternatives: switching to SMP code
> [ 1025.237548] Booting processor 1 APIC 0x1 ip 0x6000
> [ 1025.232536] Initializing CPU#1
> [ 1025.232536] Calibrating delay using timer specific routine..
> 4245.78 BogoMIPS (lpj=8491564)
>
> Unfortunately the ACPI stuff is all Greek to me. If anyone wants
> additional debug information then please let me know.
>
> So any ideas?

What's in /proc/acpi/wakeup after a clean boot?

Rafael

2009-09-08 20:05:25

by Alex Bennee

[permalink] [raw]
Subject: Re: Machine resumes straight after suspend to RAM (or disk)

2009/9/8 Rafael J. Wysocki <[email protected]>:
> On Tuesday 08 September 2009, Alex Bennee wrote:
>> Hi,
>>
>> I thought I'd enable suspend and resume on my desktop machine so I
>> could save myself some time in the morning when I restart the machine.
>> The machine is fairly standard Intel Core Duo based system (lspci
>> attached). However when I execute pm-suspend (or Gentoo's own
>> hibernate-script) the system does suspend but resumes immediately.
<snip>
>>
>> So any ideas?
>
> What's in /proc/acpi/wakeup after a clean boot?

21:03 alex@danny/x86_64 [~] >cat /proc/acpi/wakeup
Device S-state Status Sysfs node
P0P1 S4 disabled
P0P3 S4 disabled pci:0000:00:1e.0
P0P4 S4 disabled pci:0000:00:1c.0
P0P5 S4 disabled
P0P6 S4 disabled
P0P7 S4 disabled
P0P8 S4 disabled
P0P9 S4 disabled
PS2K S4 disabled pnp:00:09
PS2M S4 disabled pnp:00:0a
MC97 S4 disabled
USB1 S4 disabled pci:0000:00:1d.0
USB2 S4 disabled pci:0000:00:1d.1
USB3 S4 disabled pci:0000:00:1d.2
USB4 S4 disabled pci:0000:00:1d.3
EUSB S4 disabled pci:0000:00:1d.7


--
Alex, homepage: http://www.bennee.com/~alex/
http://www.half-llama.co.uk

2009-09-08 20:47:33

by Maxim Levitsky

[permalink] [raw]
Subject: Re: Machine resumes straight after suspend to RAM (or disk)

On Tue, 2009-09-08 at 19:57 +0100, Alex Bennee wrote:
> Hi,
>
> I thought I'd enable suspend and resume on my desktop machine so I
> could save myself some time in the morning when I restart the machine.
> The machine is fairly standard Intel Core Duo based system (lspci
> attached). However when I execute pm-suspend (or Gentoo's own
> hibernate-script) the system does suspend but resumes immediately. To
> add insult to injury the ethernet doesn't come back properly and needs
> a reboot to reset it.
>
> I've played about a little with the ACPI debug stuff but it's easy to
> overflow the dmesg buffer. The relevant section around the
> suspend/resume looks like:

Try to unload ehci_hcd before suspend.

Best regards,
Maxim Levitsky

2009-09-09 07:04:10

by Alex Bennee

[permalink] [raw]
Subject: Re: Machine resumes straight after suspend to RAM (or disk)

2009/9/8 Maxim Levitsky <[email protected]>:
> On Tue, 2009-09-08 at 19:57 +0100, Alex Bennee wrote:
< snip>
>> The machine is fairly standard Intel Core Duo based system (lspci
>> attached). However when I execute pm-suspend (or Gentoo's own
>> hibernate-script) the system does suspend but resumes immediately. To
>> add insult to injury the ethernet doesn't come back properly and needs
>> a reboot to reset it.
>>
<snip>
>
> Try to unload ehci_hcd before suspend.

That worked. Is it known why ehci_hcd causes a problem?

Unfortunately the network card never recovers but I suspect
that is another problem.

--
Alex, homepage: http://www.bennee.com/~alex/
http://www.half-llama.co.uk