Hello, This small patch series came out of the review of [1].
I noticed that the DU driver enforces suspend/resume ordering with the CMMs but not with the VSPs, which patch 1/4 fixes. This in turn made me notice that the driver leaks the device_link instances to the CMM. Patches 2/4 and 3/4 refactor CMM handling a bit, to prepare for 4/4 that fixes the leak. [1] https://lore.kernel.org/r/02669d4630e04fe24c17dd2576ec8b27ded458f0.1765541401.git.tommaso.merciai...@bp.renesas.com Laurent Pinchart (4): drm: rcar-du: Ensure correct suspend/resume ordering with VSP drm: rcar-du: Store CMM device pointer instead of platform_device drm: rcar-du: Use __free() to simplify device_node handling drm: rcar-du: Don't leak device_link to CMM drivers/gpu/drm/renesas/rcar-du/rcar_cmm.c | 26 +++++----- drivers/gpu/drm/renesas/rcar-du/rcar_cmm.h | 18 +++---- .../gpu/drm/renesas/rcar-du/rcar_du_crtc.c | 16 +++--- .../gpu/drm/renesas/rcar-du/rcar_du_crtc.h | 3 +- drivers/gpu/drm/renesas/rcar-du/rcar_du_drv.h | 8 ++- drivers/gpu/drm/renesas/rcar-du/rcar_du_kms.c | 52 +++++++++---------- drivers/gpu/drm/renesas/rcar-du/rcar_du_vsp.c | 16 ++++++ drivers/gpu/drm/renesas/rcar-du/rcar_du_vsp.h | 2 + 8 files changed, 81 insertions(+), 60 deletions(-) base-commit: 3f071d00fc9478344f3231f585a5e39ceb6c63e6 -- Regards, Laurent Pinchart
