Re: [PATCH] drm: hisilicon: depends on MMU

[Javier Martinez Canillas <javierm@redhat.com>]

On 6/19/22 16:05, Javier Martinez Canillas wrote:
> Hello Randy,
> 
> On 5/31/22 04:55, Randy Dunlap wrote:
>> Prevent a kconfig warning when MMU is not enabled by making
>> DRM_HISI_HIBMC depend on MMU.
>>
>> WARNING: unmet direct dependencies detected for DRM_TTM
>>   Depends on [n]: HAS_IOMEM [=y] && DRM [=m] && MMU [=n]
>>   Selected by [m]:
>>   - DRM_TTM_HELPER [=m] && HAS_IOMEM [=y] && DRM [=m]
>>   - DRM_HISI_HIBMC [=m] && HAS_IOMEM [=y] && DRM [=m] && PCI [=y] && (ARM64 || COMPILE_TEST [=y])
>>
>> Fixes: 4f7f1973b0c8 ("drm/vram: fix Kconfig")

Actually this fixes isn't the correct one AFAICT, it should probably be
either acad3fe650a5 ("drm/hisilicon: Removed the dependency on the mmu")
or a0f25a6bb319 ("drm/hisilicon/hibmc: Allow to be built if COMPILE_TEST
is enabled"). And those two commits should be mentioned in the description.

Maybe we can append the following two paragraphs to your commit message ?

The Kconfig symbol depended on MMU but that was dropped by commit acad3fe650a5
("drm/hisilicon: Removed the dependency on the mmu") because it already had as
a dependency ARM64 that already selects MMU.

But later, commit a0f25a6bb319 ("drm/hisilicon/hibmc: Allow to be built if
COMPILE_TEST is enabled") allowed the driver to be built for non-ARM64 when
the COMPILE_TEST is set but that could lead to unmet direct dependencies and
linking errors.

If you agree with the changes, I can do it before applying. Please let me know.

-- 
Best regards,

Javier Martinez Canillas
Linux Engineering
Red Hat