Uhhuh - Dazed and confused, but trying to continue :)

Uhhuh - Dazed and confused, but trying to continue :)

[poma]

Salutem

This happened only on thaw from S4 aka hibernate.
What should be "strange power saving mode" these messages relate!?

[  208.252986] Uhhuh. NMI received for unknown reason 3d on CPU 0.
[  208.252991] Do you have a strange power saving mode enabled?
[  208.252992] Dazed and confused, but trying to continue

3.18.3-200.fc21.x86_64


poma

Re: Uhhuh - Dazed and confused, but trying to continue :)

[Don Zickus]

On Sun, Jan 18, 2015 at 11:50:38AM +0100, poma wrote:
> 
> Salutem
> 
> This happened only on thaw from S4 aka hibernate.
> What should be "strange power saving mode" these messages relate!?
> 
> [  208.252986] Uhhuh. NMI received for unknown reason 3d on CPU 0.
> [  208.252991] Do you have a strange power saving mode enabled?
> [  208.252992] Dazed and confused, but trying to continue
> 
> 3.18.3-200.fc21.x86_64

This looks to be an external unknown NMI (unless you can reproduce on
something other than cpu0).

I would have to know what hardware you have and see the dmesg log after
the resume from hibernate.

You can open a bz and assign it to me.

Just attach the dmesg log, lspci and lspci -t output to the bz.

Though this is most likely a pci device firmware problem.  But I can try
to narrow it down.

Cheers,
Don

Re: Uhhuh - Dazed and confused, but trying to continue :)

[poma]

On 19.01.2015 14:49, Don Zickus wrote:
> On Sun, Jan 18, 2015 at 11:50:38AM +0100, poma wrote:
>>
>> Salutem
>>
>> This happened only on thaw from S4 aka hibernate.
>> What should be "strange power saving mode" these messages relate!?
>>
>> [  208.252986] Uhhuh. NMI received for unknown reason 3d on CPU 0.
>> [  208.252991] Do you have a strange power saving mode enabled?
>> [  208.252992] Dazed and confused, but trying to continue
>>
>> 3.18.3-200.fc21.x86_64
> 
> This looks to be an external unknown NMI (unless you can reproduce on
> something other than cpu0).
> 
> I would have to know what hardware you have and see the dmesg log after
> the resume from hibernate.
> 
> You can open a bz and assign it to me.
> 
> Just attach the dmesg log, lspci and lspci -t output to the bz.
> 
> Though this is most likely a pci device firmware problem.  But I can try
> to narrow it down.
> 
> Cheers,
> Don
> 

Thank you, in the meantime, I've found what causes non-maskable interrupt, [23]d on CPU 0, on resuming S4,
on AMD Phenom(tm) II X4 955 Processor.

/etc/default/tlp
...
# Kernel NMI Watchdog
# 0=disable (default, saves power) / 1=enable (for kernel debugging only)
NMI_WATCHDOG=0

Toggling to 1, messages no longer appear,
NMI_WATCHDOG=1

$ cat /proc/sys/kernel/nmi_watchdog
1
$ sysctl kernel.watchdog
kernel.watchdog = 1

Isn't it brilliant.

Re: Uhhuh - Dazed and confused, but trying to continue :)

[Don Zickus]

On Wed, Jan 21, 2015 at 10:01:32AM +0100, poma wrote:
> On 19.01.2015 14:49, Don Zickus wrote:
> 
> Thank you, in the meantime, I've found what causes non-maskable interrupt, [23]d on CPU 0, on resuming S4,
> on AMD Phenom(tm) II X4 955 Processor.
> 
> /etc/default/tlp
> ...
> # Kernel NMI Watchdog
> # 0=disable (default, saves power) / 1=enable (for kernel debugging only)
> NMI_WATCHDOG=0
> 
> Toggling to 1, messages no longer appear,
> NMI_WATCHDOG=1
> 
> $ cat /proc/sys/kernel/nmi_watchdog
> 1
> $ sysctl kernel.watchdog
> kernel.watchdog = 1
> 
> Isn't it brilliant.


Hehe.  Ok.  All that does is swallow your unknown NMI.  But if that makes
you happy, I am fine with that. :-)

Cheers,
Don

Re: Uhhuh - Dazed and confused, but trying to continue :)

[poma]

On 21.01.2015 15:24, Don Zickus wrote:
> On Wed, Jan 21, 2015 at 10:01:32AM +0100, poma wrote:
>> On 19.01.2015 14:49, Don Zickus wrote:
>>
>> Thank you, in the meantime, I've found what causes non-maskable interrupt, [23]d on CPU 0, on resuming S4,
>> on AMD Phenom(tm) II X4 955 Processor.
>>
>> /etc/default/tlp
>> ...
>> # Kernel NMI Watchdog
>> # 0=disable (default, saves power) / 1=enable (for kernel debugging only)
>> NMI_WATCHDOG=0
>>
>> Toggling to 1, messages no longer appear,
>> NMI_WATCHDOG=1
>>
>> $ cat /proc/sys/kernel/nmi_watchdog
>> 1
>> $ sysctl kernel.watchdog
>> kernel.watchdog = 1
>>
>> Isn't it brilliant.
> 
> 
> Hehe.  Ok.  All that does is swallow your unknown NMI.  But if that makes
> you happy, I am fine with that. :-)
> 
> Cheers,
> Don
> 

I don't want to bother you for nothing, as long as nothing crashes, it can pass.
Thanks again.