APPLIED(E)/cmt: [SRU][OEM-OSP1/E/unstable][PATCH 0/1] make the SND_SOC_SOF_BROADWELL_SUPPORT exclusive with SND_SOC_INTEL_HASWELL

[Stefan Bader]

On 04.12.19 15:17, Hui Wang wrote:
> 
> On 2019/12/4 下午6:39, Stefan Bader wrote:
>> On 02.12.19 09:06, Khaled Elmously wrote:
>>> The bug is not targeted to Eoan - and I'm unable to target it myself for some
>>> reason.
>> Also note, that by not testing this change it caused build failures because:
>> 1) The config get adjusted to no longer build snd-sof-intel-bdw, including
>>     this change in a config update patch also clarifies that this is what is
>>     meant to happen. This part does not cause the build to fail but it will
>>     delay creating a kernel for the cycle. Here it is a problem because without
>>     going into details it is unclear which of the two drivers should be kept.
>> 2) The module _must_ be removed from the ABI modules list, otherwise the
>>     build will fail (this change can go with the config update). The annoying
>>     part here is that this happens rather late in the build process, so one
>>     can loose a lot of time.
> 
> Oh, I never thought that. If it couldn't be merged into this cycle, I will
> prepare the abi patch for the next cycle.
> 
> I always build the kernel with "skipabi=true skipconfig=true skipmodule=true
> skipretpoline=true", so this problem didn't expose.


There I would strongly suggest using cranky-test-build or at least never use
those options to check for being ready for submitting. I don't think it saves
you time but instead hides all problems which a change might introduce.

> 
> 
> Thanks,
> 
> Hui.
> 
> 
>> For now I added this change but it could happen that patches causing build
>> failures just get unapplied and will miss a cycle.
>>
>> -Stefan
>>
>>> On 2019-11-20 11:04:22 , Hui Wang wrote:
>>>> BugLink: https://bugs.launchpad.net/bugs/1846539
>>>>
>>>> [Impact]
>>>> With the current default kernel config, the SND_SOC_SOF_BROADWELL_SUPPORT
>>>> is set to Y, then it will make the legacy intel soc driver fail, and
>>>> moreover the current sof driver can't support broadwell well, as a
>>>> result, many broadwell computers with I2S codec can't support sound anymore.
>>>>
>>>> [Fix]
>>>> Intel submitted a patch, it will make those two kconfig options exclusive,
>>>> and the legacy soc's option has higher priority, if both of them are
>>>> enabled, sof's option will be disabled.
>>>>
>>>> [Test Case]
>>>> Build the kernel and check the debian/build/$buidling_dir/.config,  the
>>>> SND_SOC_SOF_BROADWELL_SUPPORT is disabled.
>>>>
>>>> [Regression Risk]
>>>> Low, so far, the sof driver can't support broadwell machine well, so
>>>> there is no machine use this driver yet.
>>>>
>>>>
>>>> Pierre-Louis Bossart (1):
>>>>    UBUNTU: SAUCE: ASoC: SOF: Intel: Broadwell: clarify mutual exclusion
>>>>      with legacy driver
>>>>
>>>>   sound/soc/intel/Kconfig     |  3 +++
>>>>   sound/soc/sof/intel/Kconfig | 10 +++++++++-
>>>>   2 files changed, 12 insertions(+), 1 deletion(-)
>>>>
>>>> -- 
>>>> 2.17.1
>>>>
>>>>
>>>> -- 
>>>> kernel-team mailing list
>>>> kernel-team at lists.ubuntu.com
>>>> https://lists.ubuntu.com/mailman/listinfo/kernel-team
>>


-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 833 bytes
Desc: OpenPGP digital signature
URL: <https://lists.ubuntu.com/archives/kernel-team/attachments/20191204/d1c1daea/attachment.sig>