>
> Hmm. Perhaps it's possible to test this patch on the affected boxes:
> https://bugzilla.kernel.org/attachment.cgi?id=45192
Still no change with the patch..
--
Pardus Linux
http://www.pardus.org.tr/eng
On Thursday, February 03, 2011, Ozan Çağlayan wrote:
>
> >
> > Hmm. Perhaps it's possible to test this patch on the affected boxes:
> > https://bugzilla.kernel.org/attachment.cgi?id=45192
>
>
> Still no change with the patch..
I guess we see a few different problems here.
Can you remind me, please, what the known workarounds are?
Rafael
On 04.02.2011 00:04, Rafael J. Wysocki wrote:
>
> I guess we see a few different problems here.
>
> Can you remind me, please, what the known workarounds are?
OK in summary:
ASUS X71S - Doesn't boot without nolapic
ASUS F50SV - Doesn't boot without nolapic (latest tried and working 2.6.31)
ASUS N61JQ - X out of Y shutdown hangs (Cured with nolapic too)
Note that nolapic is the finer grained fix. acpi=off fixes the issue too.
Doesn't seem to be related to idle as processor.max_cstates=0 didn't help.
This is 2.6.37, intel_idle enabled with 2 upstream patches mentioned
before + the mm fix.
--
Pardus Linux
http://www.pardus.org.tr/eng
On Thursday, February 03, 2011, Ozan Çağlayan wrote:
> On 04.02.2011 00:04, Rafael J. Wysocki wrote:
>
> >
> > I guess we see a few different problems here.
> >
> > Can you remind me, please, what the known workarounds are?
>
> OK in summary:
>
> ASUS X71S - Doesn't boot without nolapic
> ASUS F50SV - Doesn't boot without nolapic (latest tried and working 2.6.31)
> ASUS N61JQ - X out of Y shutdown hangs (Cured with nolapic too)
>
> Note that nolapic is the finer grained fix. acpi=off fixes the issue too.
Hmm. Are those boxes uniprocessor running SMP kernels?
Rafael
Cuma 04 Şubat 2011 günü (saat 00:51:31) Rafael J. Wysocki şunları yazmıştı:
> Hmm. Are those boxes uniprocessor running SMP kernels?
The kernel is SMP.
ASUS X61S is
CPU0: Intel(R) Core(TM)2 Duo CPU P8700 @ 2.53GHz stepping 0a
F50SV has 2 CPU's too according to dmesg:
[ 0.000000] Processor #0 (Bootup-CPU)
[ 0.000000] Processor #1
[ 0.000000] I/O APIC #2 Version 20 at 0xFEC00000.
[ 0.000000] IOAPIC[0]: apic_id 2, version 20, address 0xfec00000, GSI 0-23
[ 0.000000] Processors: 2
[ 0.000000] SMP: Allowing 2 CPUs, 0 hotplug CPUs
ASUS N51JQ (occasional shutdown hang) has:
CPU0: Intel(R) Core(TM) i7 CPU Q 740 @ 1.73GHz stepping 05
[ 0.132435] Booting Node 0, Processors #1 #2 #3 #4 #5 #6 #7 Ok.
[ 0.768809] Brought up 8 CPUs
---
Ozan Çağlayan
TUBITAK/UEKAE - Pardus Linux
http://www.pardus.org.tr/eng
On Friday, February 04, 2011, Ozan Çağlayan wrote:
> Cuma 04 Şubat 2011 günü (saat 00:51:31) Rafael J. Wysocki şunları yazmıştı:
>
> > Hmm. Are those boxes uniprocessor running SMP kernels?
>
> The kernel is SMP.
>
> ASUS X61S is
> CPU0: Intel(R) Core(TM)2 Duo CPU P8700 @ 2.53GHz stepping 0a
>
> F50SV has 2 CPU's too according to dmesg:
> [ 0.000000] Processor #0 (Bootup-CPU)
> [ 0.000000] Processor #1
> [ 0.000000] I/O APIC #2 Version 20 at 0xFEC00000.
> [ 0.000000] IOAPIC[0]: apic_id 2, version 20, address 0xfec00000, GSI 0-23
> [ 0.000000] Processors: 2
> [ 0.000000] SMP: Allowing 2 CPUs, 0 hotplug CPUs
>
> ASUS N51JQ (occasional shutdown hang) has:
> CPU0: Intel(R) Core(TM) i7 CPU Q 740 @ 1.73GHz stepping 05
> [ 0.132435] Booting Node 0, Processors #1 #2 #3 #4 #5 #6 #7 Ok.
> [ 0.768809] Brought up 8 CPUs
OK, so it's not the case I was thinking of. Ohwell.
At this point I guess it's best to open a bug entry in the kernel Bugzilla,
against ACPI (although ACPI may be a red herring), and put some system
information (ie. dmesg outputs from working kernels, if any, outputs of
acpidump) -- or links to where it can be found.
Thanks,
Rafael
On 04.02.2011 22:36, Rafael J. Wysocki wrote:
>
> At this point I guess it's best to open a bug entry in the kernel Bugzilla,
> against ACPI (although ACPI may be a red herring), and put some system
> information (ie. dmesg outputs from working kernels, if any, outputs of
> acpidump) -- or links to where it can be found.
Well I've already filed two bugs:
https://bugzilla.kernel.org/show_bug.cgi?id=26502
https://bugzilla.kernel.org/show_bug.cgi?id=27412
--
Pardus Linux
http://www.pardus.org.tr/eng
On Saturday, February 05, 2011, Ozan Çağlayan wrote:
> On 04.02.2011 22:36, Rafael J. Wysocki wrote:
>
> >
> > At this point I guess it's best to open a bug entry in the kernel Bugzilla,
> > against ACPI (although ACPI may be a red herring), and put some system
> > information (ie. dmesg outputs from working kernels, if any, outputs of
> > acpidump) -- or links to where it can be found.
>
> Well I've already filed two bugs:
> https://bugzilla.kernel.org/show_bug.cgi?id=26502
> https://bugzilla.kernel.org/show_bug.cgi?id=27412
OK
So the conclusion from this thread is that disabling cpuidle doesn't help, right?
Did you carry out the !CONFIG_NO_HZ tests?
Rafael