Re: [PM] bfcc1e67ff: kernel-selftests.breakpoints.step_after_suspend_test.fail

From: Florian Fainelli <f.fainelli@gmail.com>
To: "Rafael J. Wysocki" <rafael.j.wysocki@intel.com>
Cc: LKML <linux-kernel@vger.kernel.org>,
	Linux Memory Management List <linux-mm@kvack.org>,
	lkp@lists.01.org, lkp@intel.com,
	kernel test robot <oliver.sang@intel.com>
Subject: Re: [PM] bfcc1e67ff: kernel-selftests.breakpoints.step_after_suspend_test.fail
Date: Thu, 14 Oct 2021 14:55:42 -0700	[thread overview]
Message-ID: <6755cf07-fa5a-cbb7-c076-57c162a08c99@gmail.com> (raw)
In-Reply-To: <e526de16-5efd-6474-20e1-3f96a2e3c524@intel.com>

On 10/14/21 12:23 PM, Rafael J. Wysocki wrote:
> On 10/14/2021 6:26 PM, Florian Fainelli wrote:
>> On 10/14/21 12:57 AM, kernel test robot wrote:
>>>
>>> Greeting,
>>>
>>> FYI, we noticed the following commit (built with gcc-9):
>>>
>>> commit: bfcc1e67ff1e4aa8bfe2ca57f99390fc284c799d ("PM: sleep: Do not
>>> assume that "mem" is always present")
>>> https://git.kernel.org/cgit/linux/kernel/git/next/linux-next.git master
>>>
>>>
>>> in testcase: kernel-selftests
>>> version: kernel-selftests-x86_64-c8c9111a-1_20210929
>>> with following parameters:
>>>
>>>     group: group-00
>>>     ucode: 0x11
>>>
>>> test-description: The kernel contains a set of "self tests" under the
>>> tools/testing/selftests/ directory. These are intended to be small
>>> unit tests to exercise individual code paths in the kernel.
>>> test-url: https://www.kernel.org/doc/Documentation/kselftest.txt
>>>
>>>
>>> on test machine: 288 threads 2 sockets Intel(R) Xeon Phi(TM) CPU 7295
>>> @ 1.50GHz with 80G memory
>>>
>>> caused below changes (please refer to attached dmesg/kmsg for entire
>>> log/backtrace):
>>>
>>>
>>>
>>>
>>> If you fix the issue, kindly add following tag
>>> Reported-by: kernel test robot <oliver.sang@intel.com>
>> Thanks for your report. Assuming that the code responsible for
>> registering the suspend operations is drivers/acpi/sleep.c for your
>> platform, and that acpi_sleep_suspend_setup() iterated over all possible
>> sleep states, your platform must somehow be returning that ACPI_STATE_S3
>> is not a supported state somehow?
>>
>> Rafael have you ever encountered something like that?
> 
> Yes, there are systems with ACPI that don't support S3.

OK and do you know what happens when we enter suspend with "mem" in
those cases? Do we immediately return because ultimately the firmware
does not support ACPI S3?

I will rework
tools/testing/selftests/breakpoints/step_after_suspend_test.c not to
assume that "mem" is always available, since it may not be.
-- 
Florian