Re: Build regressions/improvements in v6.3-rc3 (drm/msm/)

[Randy Dunlap]

Hiya,

On 3/21/23 00:34, Geert Uytterhoeven wrote:
> Hi Randy,
> 
> On Tue, Mar 21, 2023 at 6:38 AM Randy Dunlap <rdunlap@xxxxxxxxxxxxx> wrote:
>> On 3/20/23 01:21, Geert Uytterhoeven wrote:
>>> Below is the list of build error/warning regressions/improvements in
>>> v6.3-rc3[1] compared to v6.2[2].
>>>
>>> Summarized:
>>>   - build errors: +9/-14
>>>   - build warnings: +4/-1447
>>>
>>> JFYI, when comparing v6.3-rc3[1] to v6.3-rc2[3], the summaries are:
>>>   - build errors: +0/-1
>>>   - build warnings: +0/-0
>>>
>>> Happy fixing! ;-)
>>>
>>> Thanks to the linux-next team for providing the build service.
>>>
>>> [1] http://kisskb.ellerman.id.au/kisskb/branch/linus/head/e8d018dd0257f744ca50a729e3d042cf2ec9da65/ (all 152 configs)
>>> [2] http://kisskb.ellerman.id.au/kisskb/branch/linus/head/c9c3395d5e3dcc6daee66c6908354d47bf98cb0c/ (all 152 configs)
>>> [3] http://kisskb.ellerman.id.au/kisskb/branch/linus/head/eeac8ede17557680855031c6f305ece2378af326/ (all 152 configs)
>>>
>>>
>>> *** ERRORS ***
>>>
>>> 9 error regressions:
>>
>>>   + /kisskb/src/drivers/gpu/drm/msm/msm_mdss.c: error: case label does not reduce to an integer constant:  => 300:2, 299:2, 296:2
>>
>>
>> Are these due to the sign bit being set after a shift?
>> It looks that way since it is only reported for such values.
> 
> Yep.
> 
>> From the reports on the build server, it only happens when building with gcc5.
>> I don't have the ability to build with gcc5 or I would test it.
> 
> I deliberately installed gcc-5.5.0-nolibc/aarch64-linux to reproduce it
> (gcc5 on x86 didn't reproduce).

Yes, I installed that same compiler. When I tried to use it, I got:

/opt/crosstool/gcc-5.5.0-nolibc/aarch64-linux/bin/aarch64-linux-gcc: unknown C compiler
scripts/Kconfig.include:44: Sorry, this C compiler is not supported.

and I didn't try to track down what that means or what is causing it...

>> @Rob and other drm/msm people, what do you think about this?
>> (or is this already fixed somewhere but not yet in linux-next?)
> 
> Thanks, I posted a similar fix two weeks ago:
> https://lore.kernel.org/all/20230306090633.65918-1-geert+renesas@xxxxxxxxx

OK, I replied to that one.

Thanks.

-- 
~Randy