On Tue, 2018-07-17 at 15:21 +0200, Philipp Zabel wrote:
On Tue, 2018-07-17 at 15:11 +0300, Leonard Crestez wrote:
This has been unused since commit 44b460cfe554 ("drm: imx: remove struct imx_drm_crtc and imx_drm_crtc_helper_funcs")
Signed-off-by: Leonard Crestez leonard.crestez@nxp.com
Applied to imx-drm/next, thank you.
drivers/gpu/drm/imx/imx-drm-core.c | 1 - 1 file changed, 1 deletion(-)
Looking at the imx_drm_device struct it is only used to save the drm_atomic_helper_suspend state. It seems like this could be replaced with drm_mode_config_helper_suspend/resume and the entire struct removed.
The only difference between imx_drm_suspend/resume and drm_mode_config_helper_suspend/resume is that the latter also suspends the fb_helper. This would be an improvement, right?
This looks about right to me. We currently don't call the fbdev suspend/resume notifiers at all. The correct call would be drm_fbdev_cma_set_suspend_unlocked instead of drm_fb_helper_set_suspend_unlocked, though.
Since drm_fb_cma_fbdev_init registers its fb_helper with the drm_device via drm_fb_helper_init, both (currently) do exactly the same thing.
Is drm_fbdev_cma_set_suspend expected to ever do anything extra on top of drm_fb_helper_set_suspend? Since it only handles memory management my answer would be "no".
Grepping for drm_fbdev_cma_set_suspend I only find two users: rcar-du and hdlcd. The suspend/resume code in both of those drivers looks entirely equivalent to drm_mode_config_helper_suspend/resume.
So maybe drm_fbdev_cma_set_suspend itself should be dropped in favor of drm_mode_config_helper_suspend everywhere.