linux-next: build failure after merge of the drm tree

linux-next: build failure after merge of the drm tree

[Stephen Rothwell]

[-- Attachment #1: Type: text/plain, Size: 1341 bytes --]

Hi all,

After merging the drm tree, today's linux-next build (powerpc
allyesconfig) failed like this:

drivers/gpu/drm/ttm/ttm_pool.c:73:29: error: variably modified 'global_write_combined' at file scope
   73 | static struct ttm_pool_type global_write_combined[TTM_DIM_ORDER];
      |                             ^~~~~~~~~~~~~~~~~~~~~
drivers/gpu/drm/ttm/ttm_pool.c:74:29: error: variably modified 'global_uncached' at file scope
   74 | static struct ttm_pool_type global_uncached[TTM_DIM_ORDER];
      |                             ^~~~~~~~~~~~~~~
drivers/gpu/drm/ttm/ttm_pool.c:76:29: error: variably modified 'global_dma32_write_combined' at file scope
   76 | static struct ttm_pool_type global_dma32_write_combined[TTM_DIM_ORDER];
      |                             ^~~~~~~~~~~~~~~~~~~~~~~~~~~
drivers/gpu/drm/ttm/ttm_pool.c:77:29: error: variably modified 'global_dma32_uncached' at file scope
   77 | static struct ttm_pool_type global_dma32_uncached[TTM_DIM_ORDER];
      |                             ^~~~~~~~~~~~~~~~~~~~~

Caused by commit

  322458c2bb1a ("drm/ttm: Reduce the number of used allocation orders for TTM pages")

PMD_SHIFT is not necessarily a constant on ppc (see
arch/powerpc/include/asm/book3s/64/pgtable.h).

I have reverted that commit for today.

-- 
Cheers,
Stephen Rothwell

[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 488 bytes --]

build failure from drm/ttm commit now in mainline (was: linux-next: build failure after merge of the drm tree)

[Thorsten Leemhuis]

Lo!

Sometimes the regression tracker runs into regressions himself... :-D

On 11.04.23 08:47, Stephen Rothwell wrote:
> 
> After merging the drm tree, today's linux-next build (powerpc
> allyesconfig) failed like this:
> 
> drivers/gpu/drm/ttm/ttm_pool.c:73:29: error: variably modified 'global_write_combined' at file scope
>    73 | static struct ttm_pool_type global_write_combined[TTM_DIM_ORDER];
>       |                             ^~~~~~~~~~~~~~~~~~~~~
> drivers/gpu/drm/ttm/ttm_pool.c:74:29: error: variably modified 'global_uncached' at file scope
>    74 | static struct ttm_pool_type global_uncached[TTM_DIM_ORDER];
>       |                             ^~~~~~~~~~~~~~~
> drivers/gpu/drm/ttm/ttm_pool.c:76:29: error: variably modified 'global_dma32_write_combined' at file scope
>    76 | static struct ttm_pool_type global_dma32_write_combined[TTM_DIM_ORDER];
>       |                             ^~~~~~~~~~~~~~~~~~~~~~~~~~~
> drivers/gpu/drm/ttm/ttm_pool.c:77:29: error: variably modified 'global_dma32_uncached' at file scope
>    77 | static struct ttm_pool_type global_dma32_uncached[TTM_DIM_ORDER];
>       |                             ^~~~~~~~~~~~~~~~~~~~~
> 
> Caused by commit
> 
>   322458c2bb1a ("drm/ttm: Reduce the number of used allocation orders for TTM pages")
> 
> PMD_SHIFT is not necessarily a constant on ppc (see
> arch/powerpc/include/asm/book3s/64/pgtable.h).
> 
> I have reverted that commit for today.

Did anyone look into this? I today ran into what looks like the same
compiler error when building a mainline snapshot using a Fedora rawhide
config for ppc64le:

```
 drivers/gpu/drm/ttm/ttm_pool.c:73:29: error: variably modified
'global_write_combined' at file scope
    73 | static struct ttm_pool_type global_write_combined[TTM_DIM_ORDER];
       |                             ^~~~~~~~~~~~~~~~~~~~~
 drivers/gpu/drm/ttm/ttm_pool.c:74:29: error: variably modified
'global_uncached' at file scope
    74 | static struct ttm_pool_type global_uncached[TTM_DIM_ORDER];
       |                             ^~~~~~~~~~~~~~~
 drivers/gpu/drm/ttm/ttm_pool.c:76:29: error: variably modified
'global_dma32_write_combined' at file scope
    76 | static struct ttm_pool_type
global_dma32_write_combined[TTM_DIM_ORDER];
       |                             ^~~~~~~~~~~~~~~~~~~~~~~~~~~
 drivers/gpu/drm/ttm/ttm_pool.c:77:29: error: variably modified
'global_dma32_uncached' at file scope
    77 | static struct ttm_pool_type global_dma32_uncached[TTM_DIM_ORDER];
       |                             ^~~~~~~~~~~~~~~~~~~~~
```

Full build log:

https://copr-be.cloud.fedoraproject.org/results/@kernel-vanilla/mainline/fedora-37-ppc64le/05850588-mainline-mainline-releases/build.log.gz

Ciao, Thorsten

Re: build failure from drm/ttm commit now in mainline (was: linux-next: build failure after merge of the drm tree)

[Michael Ellerman]

Thorsten Leemhuis <regressions@leemhuis.info> writes:
> Lo!
>
> Sometimes the regression tracker runs into regressions himself... :-D
>
> On 11.04.23 08:47, Stephen Rothwell wrote:
>> 
>> After merging the drm tree, today's linux-next build (powerpc
>> allyesconfig) failed like this:
>> 
>> drivers/gpu/drm/ttm/ttm_pool.c:73:29: error: variably modified 'global_write_combined' at file scope
>>    73 | static struct ttm_pool_type global_write_combined[TTM_DIM_ORDER];
>>       |                             ^~~~~~~~~~~~~~~~~~~~~
>> drivers/gpu/drm/ttm/ttm_pool.c:74:29: error: variably modified 'global_uncached' at file scope
>>    74 | static struct ttm_pool_type global_uncached[TTM_DIM_ORDER];
>>       |                             ^~~~~~~~~~~~~~~
>> drivers/gpu/drm/ttm/ttm_pool.c:76:29: error: variably modified 'global_dma32_write_combined' at file scope
>>    76 | static struct ttm_pool_type global_dma32_write_combined[TTM_DIM_ORDER];
>>       |                             ^~~~~~~~~~~~~~~~~~~~~~~~~~~
>> drivers/gpu/drm/ttm/ttm_pool.c:77:29: error: variably modified 'global_dma32_uncached' at file scope
>>    77 | static struct ttm_pool_type global_dma32_uncached[TTM_DIM_ORDER];
>>       |                             ^~~~~~~~~~~~~~~~~~~~~
>> 
>> Caused by commit
>> 
>>   322458c2bb1a ("drm/ttm: Reduce the number of used allocation orders for TTM pages")
>> 
>> PMD_SHIFT is not necessarily a constant on ppc (see
>> arch/powerpc/include/asm/book3s/64/pgtable.h).
>> 
>> I have reverted that commit for today.
>
> Did anyone look into this?

Christian sent a revert:
  56e51681246e ("drm/ttm: revert "Reduce the number of used allocation orders for TTM pages"")

Which has been in linux-next since the 21st. But I guess it missed being
in the initial DRM pull request.

See also other thread:
  https://lore.kernel.org/all/ZEljCgVFnDl%2FC+l3@debian/

Where Christian says the revert is "on the way" to Linus.

cheers