Re: [LKP] [mailbox] afd0b1fb22: kmsg.Error_parsing_PCC_subspaces_from_PCCT

From: David Arcari
Date: Tue Oct 02 2018 - 08:06:24 EST



Hi,

Can someone help me interpret this email? I was expecting this to be about a
dmesg change which is expected, but the content doesn't seem to indicate that.

For the record, the commit shouldn't really introduce any functional changes.
It merely avoids calling kcalloc with count==-EINVAL. In both cases, an error
should be returned from acpi_pcc_probe to pcc_init, which should in turn always
return -ENODEV.

So with the commit in place, this would be logged:

[ 3.957345] ACPI: [PCCT:0x00] Invalid zero length
[ 3.962340] Error parsing PCC subspaces from PCCT

instead of:

[ 3.957345] ACPI: [PCCT:0x00] Invalid zero length
[ 3.962340] Could not allocate space for PCC mbox channels


On 09/29/2018 11:17 PM, kernel test robot wrote:
> FYI, we noticed the following commit (built with gcc-7):
>
> commit: afd0b1fb22269f48d68fdf269891c653818c8047 ("mailbox: PCC: handle parse error")
> https://git.kernel.org/cgit/linux/kernel/git/next/linux-next.git master
>
> in testcase: ftq
> with following parameters:
>
> nr_task: 100%
> samples: 6000ss
> test: cache
> freq: 20
> ucode: 0x200004d
> cpufreq_governor: performance
>
> test-description: The FTQ benchmarks measure hardware and software interference or 'noise' on a node from the applications perspective.
> test-url: https://github.com/rminnich/ftq
>
>
> on test machine: 104 threads Intel(R) Xeon(R) Platinum 8170 CPU @ 2.10GHz with 192G memory
>
> caused below changes (please refer to attached dmesg/kmsg for entire log/backtrace):
>
>
>
>
> 2018-09-21 23:31:26 ./ftq_cache -f 20 -n 6000 -t 104 -a 360448
> ticks per ns 2.09508

This ^^^ appears to be the actual error, but I don't understand what the
expected result should be or what this really means.

Thanks,

-Dave


>
>
>
> To reproduce:
>
> git clone https://github.com/intel/lkp-tests.git
> cd lkp-tests
> bin/lkp install job.yaml # job file is attached in this email
> bin/lkp run job.yaml
>
>
>
> Thanks,
> Rong Chen
>