Re: [linux-pm] linux-next: Tree for Feb 8 [ smp|cpufreq: WARNING: at kernel/smp.c:245 smp_call_function_single ]

[Sedat Dilek <sedat.dilek@gmail.com>]

On Fri, Feb 8, 2013 at 5:28 PM, Sedat Dilek <sedat.dilek@gmail.com> wrote:
> On Fri, Feb 8, 2013 at 4:21 PM, Sedat Dilek <sedat.dilek@gmail.com> wrote:
>> On Fri, Feb 8, 2013 at 3:56 PM, Viresh Kumar <viresh.kumar@linaro.org> wrote:
>>> On Fri, Feb 8, 2013 at 8:18 PM, Sedat Dilek <sedat.dilek@gmail.com> wrote:
>>>> Nah, I pulled in latest pm-next where this commit is new...
>>>>
>>>> commit 8d5666f3456f2fd4a4e5dced228475b829851e53
>>>> "ACPI: Unbind ACPI drv when probe failed"
>>>>
>>>> ...building with it.
>>>>
>>>> Same to you, say concretely which commit is fixing what...
>>>>
>>>> Pull-N-B-Happy was never my strategy... I want to understand what went
>>>> wrong and have stolen my time.
>>>
>>> I don't have any pointers to broken tree and so can't point you to the culprit,
>>> but it was this patch:
>>>
>>> http://git.linaro.org/gitweb?p=people/vireshk/linux.git;a=commit;h=e034e731f4d9d18ad0401f033f485a3096796c58
>>>
>>> minus
>>>
>>> the patch i sent you as attachment.
>>>
>>> There were some locking introduced around init/exit of cpufreq_driver, which
>>> caused some drivers to break. Its fixed now in the above commit.
>>
>> Hmm, this "high-patch-maths" is not user-friendly!
>>
>> I will pull-in your tree into Linux-Next (next-20130208) and see if it
>> applies cleanly.
>>
>> - Sedat -
>
> No, it did NOT apply cleanly and I merged your tree like this.
> To me it does not look like your changes from the patch you sent me
> are included?
>
> - Sedat -

Looks like I did it correct - no WARNINGs.

# diff -uprN /boot/config-3.8.0-rc6-next20130208-1-iniza-small
/boot/config-3.8.0-rc6-next20130208-6-iniza-small
--- /boot/config-3.8.0-rc6-next20130208-1-iniza-small   2013-02-08
09:34:59.000000000 +0100
+++ /boot/config-3.8.0-rc6-next20130208-6-iniza-small   2013-02-08
17:36:02.000000000 +0100
@@ -601,6 +601,7 @@ CONFIG_CPU_FREQ_GOV_CONSERVATIVE=y
 #
 # x86 CPU frequency scaling drivers
 #
+CONFIG_X86_INTEL_PSTATE=y
 CONFIG_X86_PCC_CPUFREQ=y
 CONFIG_X86_ACPI_CPUFREQ=y
 CONFIG_X86_ACPI_CPUFREQ_CPB=y

# dmesg | grep -i pstate
[    0.387256] Intel pstate controlling: cpu 0
[    0.387273] Intel pstate controlling: cpu 1
[    0.387286] Intel pstate controlling: cpu 2
[    0.387303] Intel pstate controlling: cpu 3
[  151.315937] Intel pstate controlling: cpu 1
[  151.329400] Intel pstate controlling: cpu 2
[  151.342778] Intel pstate controlling: cpu 3

So, suspend-resume is good here.

- Sedat -