Re: [PATCH v2] drm/hisilicon: fix build error without fbdev emulation

[Arnd Bergmann]

On Wed, Jul 26, 2017 at 7:55 AM, Daniel Vetter <daniel.vetter@xxxxxxxx> wrote:
> On Tue, Jul 25, 2017 at 8:05 PM, Arnd Bergmann <arnd@xxxxxxxx> wrote:
>> We cannot reference priv->fbdev outside of the #ifdef:
>>
>> drivers/net/virtio_net.c:1881:12: error: 'virtnet_restore_up' defined but not used [-Werror=unused-function]
>>  static int virtnet_restore_up(struct virtio_device *vdev)
>> drivers/net/virtio_net.c:1859:13: error: 'virtnet_freeze_down' defined but not used [-Werror=unused-function]
>>  static void virtnet_freeze_down(struct virtio_device *vdev)
>>
>> As the #ifdef is a bit annoying here, this removes it entirely
>> and uses an IS_ENABLED() check in it place where needed.
>>
>> Fixes: b4dd9f1ffaba ("drm/hisilicon: Remove custom FB helper deferred setup")
>> Signed-off-by: Arnd Bergmann <arnd@xxxxxxxx>
>
> I guess I wasn't clear enough, but you don't even need the IS_ENABLED.
> The cma_fini/init functions themselves don't get no-opped out (I guess
> we could fix that), but the underlying fb helper functions they call
> do, so this is all perfectly fine to call unconditionally. And that's
> what all other drivers do. Should I edit while applying, or do you
> want to respin?

Please just edit as you like then, I think that's quicker.

The version I sent was meant to have smaller object code as well, and
I didn't think we could rely on drm_fb_cma_helper.c being built without
CONFIG_DRM_FBDEV_EMULATION, but I see now that they are
independent as you say, so making them unconditional indeed gives
the simplest code.

Thanks!

        Arnd