kernel/linux.git/drivers/gpu, branch v6.3.6

drm/i915: Fix PIPEDMC disabling for a bigjoiner configuration

2023-06-05T07:29:37+00:00

[ Upstream commit 45dfbd992923f4df174db4e23b96fca7e30d73e2 ] For a bigjoiner configuration display->crtc_disable() will be called first for the slave CRTCs and then for the master CRTC. However slave CRTCs will be actually disabled only after the master CRTC is disabled (from the encoder disable hooks called with the master CRTC state). Hence the slave PIPEDMCs can be disabled only after the master CRTC is disabled, make this so. intel_encoders_post_pll_disable() must be called only for the master CRTC, as for the other two encoder disable hooks. While at it fix this up as well. This didn't cause a problem, since intel_encoders_post_pll_disable() will call the corresponding hook only for an encoder/connector connected to the given CRTC, however slave CRTCs will have no associated encoder/connector. Fixes: 3af2ff0840be ("drm/i915: Enable a PIPEDMC whenever its corresponding pipe is enabled") Cc: Rodrigo Vivi Cc: Ville Syrjälä Reviewed-by: Ville Syrjälä Signed-off-by: Imre Deak Link: https://patchwork.freedesktop.org/patch/msgid/20230510103131.1618266-2-imre.deak@intel.com (cherry picked from commit 7eeef32719f6af935a1554813e6bc206446339cd) Signed-off-by: Joonas Lahtinen Signed-off-by: Sasha Levin

drm/i915: Disable DPLLs before disconnecting the TC PHY

2023-06-05T07:29:37+00:00

[ Upstream commit b108bdd0e22a402bd3e4a6391acbb6aefad31a9e ] Bspec requires disabling the DPLLs on TC ports before disconnecting the port's PHY. Add a post_pll_disable encoder hook and move the call to disconnect the port's PHY from the post_disable hook to the new hook. Reviewed-by: Mika Kahola Signed-off-by: Imre Deak Link: https://patchwork.freedesktop.org/patch/msgid/20230323142035.1432621-28-imre.deak@intel.com Stable-dep-of: 45dfbd992923 ("drm/i915: Fix PIPEDMC disabling for a bigjoiner configuration") Signed-off-by: Sasha Levin

drm/i915: Move shared DPLL disabling into CRTC disable hook

2023-06-05T07:29:36+00:00

[ Upstream commit 3acac2d06a7e0f0b182b86b25bb8a2e9b3300406 ] The spec requires disabling the PLL on TC ports before disconnecting the port's PHY. Prepare for that by moving the PLL disabling to the CRTC disable hook, while disconnecting the PHY will be moved to the post_pll_disable() encoder hook in the next patch. v2: Move the call from intel_crtc_disable_noatomic() as well. Reviewed-by: Mika Kahola # v1 Signed-off-by: Imre Deak Link: https://patchwork.freedesktop.org/patch/msgid/20230323142035.1432621-27-imre.deak@intel.com Stable-dep-of: 45dfbd992923 ("drm/i915: Fix PIPEDMC disabling for a bigjoiner configuration") Signed-off-by: Sasha Levin

drm: fix drmm_mutex_init()

2023-05-30T13:17:24+00:00

commit c21f11d182c2180d8b90eaff84f574cfa845b250 upstream. In mutex_init() lockdep identifies a lock by defining a special static key for each lock class. However if we wrap the macro in a function, like in drmm_mutex_init(), we end up generating: int drmm_mutex_init(struct drm_device *dev, struct mutex *lock) { static struct lock_class_key __key; __mutex_init((lock), "lock", &__key); .... } The static __key here is what lockdep uses to identify the lock class, however since this is just a normal function the key here will be created once, where all callers then use the same key. In effect the mutex->depmap.key will be the same pointer for different drmm_mutex_init() callers. This then results in impossible lockdep splats since lockdep thinks completely unrelated locks are the same lock class. To fix this turn drmm_mutex_init() into a macro such that it generates a different "static struct lock_class_key __key" for each invocation, which looks to be inline with what mutex_init() wants. v2: - Revamp the commit message with clearer explanation of the issue. - Rather export __drmm_mutex_release() than static inline. Reported-by: Thomas Hellström Reported-by: Sarah Walker Fixes: e13f13e039dc ("drm: Add DRM-managed mutex_init()") Cc: Stanislaw Gruszka Cc: Boris Brezillon Cc: Thomas Zimmermann Cc: Jocelyn Falempe Cc: Daniel Vetter Cc: dri-devel@lists.freedesktop.org Signed-off-by: Matthew Auld Reviewed-by: Boris Brezillon Reviewed-by: Stanislaw Gruszka Reviewed-by: Lucas De Marchi Acked-by: Thomas Zimmermann Signed-off-by: Thomas Zimmermann Link: https://patchwork.freedesktop.org/patch/msgid/20230519090733.489019-1-matthew.auld@intel.com Signed-off-by: Greg Kroah-Hartman

drm/amd/amdgpu: limit one queue per gang

2023-05-30T13:17:23+00:00

commit 5ee33d905f89c18d4b33da6e5eefdae6060502df upstream. Limit one queue per gang in mes self test, due to mes schq fw change. Signed-off-by: Jack Xiao Reviewed-by: Hawking Zhang Signed-off-by: Alex Deucher Cc: Mario Limonciello Signed-off-by: Greg Kroah-Hartman

drm/amd/pm: Fix output of pp_od_clk_voltage

2023-05-30T13:17:23+00:00

commit 40baba5693b9af586dc1063af603d05a79e57a6b upstream. Printing the other clock types should not be conditioned on being able to print OD_SCLK. Some GPUs currently have limited capability of only printing a subset of these. Since this condition was introduced in v5.18-rc1, reading from `pp_od_clk_voltage` has been returning empty on the Asus ROG Strix G15 (2021). Fixes: 79c65f3fcbb1 ("drm/amd/pm: do not expose power implementation details to amdgpu_pm.c") Reviewed-by: Evan Quan Signed-off-by: Jonatas Esteves Signed-off-by: Alex Deucher Cc: stable@vger.kernel.org Signed-off-by: Greg Kroah-Hartman

drm/amd/pm: add missing NotifyPowerSource message mapping for SMU13.0.7

2023-05-30T13:17:23+00:00

commit 0d2dd02d74e6377268f56b90261de0fae8f0d2cb upstream. Otherwise, the power source switching will fail due to message unavailable. Fixes: bf4823267a81 ("drm/amd/pm: fix possible power mode mismatch between driver and PMFW") Signed-off-by: Evan Quan Reviewed-by: Guchun Chen Signed-off-by: Alex Deucher Cc: stable@vger.kernel.org Signed-off-by: Greg Kroah-Hartman

drm/amdgpu: don't enable secure display on incompatible platforms

2023-05-30T13:17:23+00:00

commit 7fc602dbfd548045862df096910b7d21e6d300bf upstream. [why] [drm] psp gfx command LOAD_TA(0x1) failed and response status is (0x7) [drm] psp gfx command INVOKE_CMD(0x3) failed and response status is (0x4) amdgpu 0000:04:00.0: amdgpu: Secure display: Generic Failure. [how] don't enable secure display on incompatible platforms Suggested-by: Aaron Liu Signed-off-by: Jesse zhang Reviewed-by: Aaron Liu Signed-off-by: Alex Deucher Cc: stable@vger.kernel.org Signed-off-by: Greg Kroah-Hartman

drm/radeon: reintroduce radeon_dp_work_func content

2023-05-30T13:17:23+00:00

commit a34fc1bcd2c4d8b09dcfc0b95ac65bca1e579bd7 upstream. Put back the radeon_dp_work_func logic. It seems that handling DP RX interrupts is necessary to make some panels work. This was removed with the MST support, but it regresses some systems so add it back. While we are here, add the proper mutex locking. Link: https://gitlab.freedesktop.org/drm/amd/-/issues/2567 Fixes: 01ad1d9c2888 ("drm/radeon: Drop legacy MST support") Reviewed-by: Lyude Paul Signed-off-by: Alex Deucher Cc: Lyude Paul Cc: stable@vger.kernel.org Signed-off-by: Greg Kroah-Hartman

drm/mgag200: Fix gamma lut not initialized.

2023-05-30T13:17:23+00:00

commit ad81e23426a651eb89a4b306e1c4169e6308c124 upstream. When mgag200 switched from simple KMS to regular atomic helpers, the initialization of the gamma settings was lost. This leads to a black screen, if the bios/uefi doesn't use the same pixel color depth. v2: rebase on top of drm-misc-fixes, and add Cc stable tag. Link: https://bugzilla.redhat.com/show_bug.cgi?id=2171155 Fixes: 1baf9127c482 ("drm/mgag200: Replace simple-KMS with regular atomic helpers") Cc: Tested-by: Phil Oester Reviewed-by: Thomas Zimmermann Signed-off-by: Jocelyn Falempe Link: https://patchwork.freedesktop.org/patch/msgid/20230510131034.284078-1-jfalempe@redhat.com Signed-off-by: Greg Kroah-Hartman