aboutsummaryrefslogtreecommitdiffstats
path: root/drivers/gpu/drm
AgeCommit message (Collapse)AuthorFilesLines
2015-01-26drm/i915: Init PPGTT before context enableDavid Woodhouse1-13/+6
Commit 82460d972 ("drm/i915: Rework ppgtt init to no require an aliasing ppgtt") introduced a regression on Broadwell, triggering the following IOMMU fault at startup: vgaarb: device changed decodes: PCI:0000:00:02.0,olddecodes=io+mem,decodes=io+mem:owns=io+mem dmar: DRHD: handling fault status reg 2 dmar: DMAR:[DMA Write] Request device [00:02.0] fault addr 880000 DMAR:[fault reason 23] Unknown fbcon: inteldrmfb (fb0) is primary device Further commentary from Daniel: I sugggested this change to David after staring at the offending patch for a while. I have no idea and theory whatsoever why this would upset the gpu less than the other way round. But it seems to work. David promised to chase hw people a bit more to get a more meaningful answer. Wrt the comment that this deletes: I've done some digging and afaict loading context before ppgtt enable was once required before our recent restructuring of the context/ppgtt init code: Before that context sw setup (i.e. allocating the default context) and hw setup was smashed together. Also the setup of the default context was the bit that actually allocated the aliasing ppgtt structures. Which is the reason for the context before ppgtt depency. Or was, since with all the untangling there's no no real depency any more (functional, who knows what the hw is doing), so the comment is just stale. Signed-off-by: David Woodhouse <David.Woodhouse@intel.com> Cc: stable@vger.kernel.org Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch> Signed-off-by: Jani Nikula <jani.nikula@intel.com>
2015-01-26drm/plane-helper: Fix transitional helper kerneldocsMatt Roper1-2/+2
drm_plane_helper_{update,disable} are not specific to primary planes; fix some copy/paste summaries to avoid confusion. Cc: dri-devel@lists.freedesktop.org Signed-off-by: Matt Roper <matthew.d.roper@intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2015-01-26drm/plane-helper: Skip prepare_fb/cleanup_fb when newfb==oldfbMatt Roper1-1/+9
When commiting a plane update where the framebuffer doesn't change, we can skip the prepare_fb/cleanup_fb steps. This also allows us to avoid an unnecessary vblank wait at the end of the operation when we're just moving a plane and not changing its image (e.g., for a cursor). At the moment, i915 is the only upstream driver using the transitional plane helpers, and thus the only driver affected by this change. Note that this replicates a corresponding change in the atomic helpers implemented in commit ab58e3384b9f9863bfd029b458ff337d381bf6d2 Author: Daniel Vetter <daniel.vetter@ffwll.ch> Date: Mon Nov 24 20:42:42 2014 +0100 drm/atomic-helper: Skip vblank waits for unchanged fbs Reported-by: Jeremiah Mahler <jmmahler@gmail.com> Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=88540 Signed-off-by: Matt Roper <matthew.d.roper@intel.com> Tested-by: Tested-by: Rodrigo Vivi <rodrigo.vivi@intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
2015-01-25drm/exynos: fimd: check error status for drm_iommu_attach_deviceAjay Kumar1-2/+15
check error status for drm_iommu_attach_device() and make sure it propagates till the caller. Signed-off-by: Ajay Kumar <ajaykumar.rs@samsung.com> Signed-off-by: Inki Dae <daeinki@gmail.com>
2015-01-25drm/exynos: create exynos_check_plane()Gustavo Padovan3-20/+36
Split update plane in two parts, an initial check part that can fail and the update part that can't fail. This is a important step for the upcoming atomic modesetting support. Signed-off-by: Gustavo Padovan <gustavo.padovan@collabora.co.uk> Signed-off-by: Inki Dae <inki.dae@samsung.com>
2015-01-25drm/exynos: remove mode_set() ops from exynos_crtcGustavo Padovan1-3/+0
It is no longer used anywhere. Signed-off-by: Gustavo Padovan <gustavo.padovan@collabora.co.uk> Signed-off-by: Inki Dae <inki.dae@samsung.com>
2015-01-25drm/exynos: don't duplicate drm_display_mode in fimd contextGustavo Padovan2-16/+1
We can safely use the mode stored in the crtc. Signed-off-by: Gustavo Padovan <gustavo.padovan@collabora.co.uk> Signed-off-by: Inki Dae <inki.dae@samsung.com>
2015-01-25drm/exynos: remove struct exynos_drm_managerGustavo Padovan7-244/+211
exynos_drm_manager was just a redundant struct to represent the crtc as well. In this commit we merge exynos_drm_manager into exynos_drm_crtc to remove an unnecessary level of indirection easing the understand of the flow on exynos. Signed-off-by: Gustavo Padovan <gustavo.padovan@collabora.co.uk> Signed-off-by: Inki Dae <inki.dae@samsung.com>
2015-01-25drm/exynos: remove drm_dev from struct exynos_drm_managerGustavo Padovan6-12/+13
manager-drm_dev is only accessed by exynos_drm_crtc_create() so this patch pass drm_dev as argument on exynos_drm_crtc_create() and remove it from struct exynos_drm_manager. Signed-off-by: Gustavo Padovan <gustavo.padovan@collabora.co.uk> Signed-off-by: Inki Dae <inki.dae@samsung.com>
2015-01-25drm/exynos: move 'type' from manager to crtc structGustavo Padovan6-14/+17
'type' is now part of the struct exynos_drm_crtc. This is just another step in the struct exynos_drm_manager removal. Signed-off-by: Gustavo Padovan <gustavo.padovan@collabora.co.uk> Signed-off-by: Inki Dae <inki.dae@samsung.com>
2015-01-25drm/exynos: remove pipe member of struct exynos_drm_managerGustavo Padovan5-6/+4
It is not longer used. This is part of the process of removing struct exynos_drm_manager entirely. Signed-off-by: Gustavo Padovan <gustavo.padovan@collabora.co.uk> Signed-off-by: Inki Dae <inki.dae@samsung.com>
2015-01-25drm/exynos: add pipe param to exynos_drm_crtc_create()Gustavo Padovan5-8/+8
Get the pipe value from a parameter instead of getting it from manager->pipe. We are removing manager->pipe. Signed-off-by: Gustavo Padovan <gustavo.padovan@collabora.co.uk> Signed-off-by: Inki Dae <inki.dae@samsung.com>
2015-01-25drm/exynos: rename base object of struct exynos_drm_crtc to 'base'Gustavo Padovan2-6/+5
'base' is more widely used name in the drm subsystem for the base object. Signed-off-by: Gustavo Padovan <gustavo.padovan@collabora.co.uk> Signed-off-by: Inki Dae <inki.dae@samsung.com>
2015-01-25drm/exynos: remove exynos_drm_crtc_mode_set_commit()Gustavo Padovan1-7/+1
This was just as extra chain in the call stack. We just rename it to _set_base() and let it do everything alone. Signed-off-by: Gustavo Padovan <gustavo.padovan@collabora.co.uk> Signed-off-by: Inki Dae <inki.dae@samsung.com>
2015-01-25drm/exynos: call exynos_update_plane() directly on page flipsGustavo Padovan1-2/+6
Avoid an extra call to exynos_drm_crtc_mode_set_commit() that only calls exynos_update_plane(). Signed-off-by: Gustavo Padovan <gustavo.padovan@collabora.co.uk> Signed-off-by: Inki Dae <inki.dae@samsung.com>
2015-01-25drm/exynos: unify plane update on exynos_update_plane()Gustavo Padovan3-16/+8
We can safely use the exynos_update_plane() to update the plane framebuffer for both the overlay and primary planes. Note that this patch removes a call to manager->ops->commit() in exynos_drm_crtc_mode_set_commit(). The commit() call is used only by the fimd driver to set underlying timings and need only in full modeset operations. For plane update only win_commit is needed. Signed-off-by: Gustavo Padovan <gustavo.padovan@collabora.co.uk> Signed-off-by: Inki Dae <inki.dae@samsung.com>
2015-01-25drm/exynos: remove exynos_plane_commit() wrapperGustavo Padovan3-12/+8
It's doing nothing but calling exynos_crtc->ops->win_commit(), so let's call this directly to avoid extra layers of abstraction. Signed-off-by: Gustavo Padovan <gustavo.padovan@collabora.co.uk> Signed-off-by: Inki Dae <inki.dae@samsung.com>
2015-01-25drm/exynos: don't do any DPMS operation while updating planesGustavo Padovan1-1/+7
DPMS only makes sense when the mode changes, for plane update changes do not perform any dpms operation. This move places the win_commit() and commit() calls directly in the code instead of calling exynos_drm_crtc_commit() thus avoiding DPMS operations. Signed-off-by: Gustavo Padovan <gustavo.padovan@collabora.co.uk> Signed-off-by: Inki Dae <inki.dae@samsung.com>
2015-01-25drm/exynos: Don't touch DPMS when updating overlay planesGustavo Padovan1-1/+0
DPMS settings should only be changed by a full modeset. exynos_plane_update() should only care about updating the planes itself and nothing else. Signed-off-by: Gustavo Padovan <gustavo.padovan@collabora.co.uk> Signed-off-by: Inki Dae <inki.dae@samsung.com>
2015-01-25drm/exynos/vidi: remove useless ops->commit()Gustavo Padovan1-12/+0
vidi_commit does nothing, remove it and its callers. Signed-off-by: Gustavo Padovan <gustavo.padovan@collabora.co.uk> Signed-off-by: Inki Dae <inki.dae@samsung.com>
2015-01-25drm/exynos/fimd: don't initialize 'ret' variable in fimd_probe()Gustavo Padovan1-1/+1
We set it in the beginning of the function, thus no need to set it at initialization. Signed-off-by: Gustavo Padovan <gustavo.padovan@collabora.co.uk> Signed-off-by: Inki Dae <inki.dae@samsung.com>
2015-01-25drm/exynos: remove struct exynos_drm_overlayGustavo Padovan6-116/+111
struct exynos_drm_overlay has no practical advantage nor serves as important piece of the exynos API design. The only place it was used was inside the struct exynos_plane which was just causing a extra access overhead. Users had to access the overlay first and just then get the plane information it contains. This patch merges struct exynos_drm_overlay into struct exynos_plane. It also renames struct exynos_plane to struct exynos_drm_plane. The rational is to cut one step to access plane information. Signed-off-by: Gustavo Padovan <gustavo.padovan@collabora.co.uk> Signed-off-by: Inki Dae <inki.dae@samsung.com>
2015-01-25drm/exynos: remove exynos_drm_crtc_plane_* wrappersGustavo Padovan2-37/+15
This functions were doing nothing but calling a manager op function, so remove them and call the manager directly. Signed-off-by: Gustavo Padovan <gustavo.padovan@collabora.co.uk> Signed-off-by: Inki Dae <inki.dae@samsung.com>
2015-01-25drm/exynos: expose struct exynos_drm_crtcGustavo Padovan2-30/+30
Let other pieces of the driver access struct exynos_drm_crtc as well. struct exynos_drm_manager will be merged into struct exynos_drm_crtc, in the sense we will move all its members to exynos_drm_crtc, so to start this conversion exynos_drm_crtc need to be exposed as well. Signed-off-by: Gustavo Padovan <gustavo.padovan@collabora.co.uk> Signed-off-by: Inki Dae <inki.dae@samsung.com>
2015-01-25drm/exynos: move to_exynos_crtc() macro to main headerGustavo Padovan2-3/+3
With this change we allow other pieces of the code to use this macro. Signed-off-by: Gustavo Padovan <gustavo.padovan@collabora.co.uk> Signed-off-by: Inki Dae <inki.dae@samsung.com>
2015-01-22drm/radeon: Remove rdev->gart.pages_addr arrayMichel Dänzer6-20/+2
radeon_vm_map_gart can use rdev->gart.pages_entry instead. Also move the masking of the page address to radeon_vm_map_gart from its callers. Reviewed-by: Christian König <christian.koenig@amd.com> Signed-off-by: Michel Dänzer <michel.daenzer@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2015-01-22drm/radeon: Restore GART table contents after pinning it in VRAM v3Michel Dänzer1-0/+13
The GART table BO has to be moved out of VRAM for suspend/resume. Any updates to the GART table during that time were silently dropped without this change. This caused GPU lockups on resume in some cases, see the bug reports referenced below. This might also make GPU reset more robust in some cases, as we no longer rely on the GART table in VRAM being preserved across the GPU lockup/reset. v2: Add logic to radeon_gart_table_vram_pin directly instead of reinstating radeon_gart_restore v3: Move code after assignment of rdev->gart.table_addr so that the GART TLB flush can work as intended, add code comment explaining why we're doing this Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=85204 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=86267 Reviewed-by: Christian König <christian.koenig@amd.com> Cc: stable@vger.kernel.org Signed-off-by: Michel Dänzer <michel.daenzer@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2015-01-22drm/radeon: Split off gart_get_page_entry ASIC hook from set_page_entryMichel Dänzer9-37/+100
get_page_entry calculates the GART page table entry, which is just written to the GART page table by set_page_entry. This is a prerequisite for the following fix. Reviewed-by: Christian König <christian.koenig@amd.com> Cc: stable@vger.kernel.org Signed-off-by: Michel Dänzer <michel.daenzer@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2015-01-22drm/amdkfd: change amdkfd version to 0.7.1Oded Gabbay1-2/+2
Signed-off-by: Oded Gabbay <oded.gabbay@amd.com>
2015-01-22drm/radeon: cik_sdma_ctx_switch_enable() can be statickbuild test robot1-1/+1
drivers/gpu/drm/radeon/cik_sdma.c:293:6: sparse: symbol 'cik_sdma_ctx_switch_enable' was not declared. Should it be static? Signed-off-by: Fengguang Wu <fengguang.wu@intel.com> Signed-off-by: Oded Gabbay <oded.gabbay@amd.com>
2015-01-22drm/amdkfd: Fix sparse errorsOded Gabbay3-37/+27
Signed-off-by: Oded Gabbay <oded.gabbay@amd.com> Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
2015-01-22drm/radeon: make MMU_NOTIFIER optionalRob Clark3-2/+11
In cases where MMU_NOTIFIER is not available, userptr will not be available. Similar to i915, although not making an exception for CAP_SYS_ADMIN. The proposed userspace patches for userptr seem to handle the fall- back properly, so a userptr-less kernel should not be a problem. Signed-off-by: Rob Clark <robdclark@gmail.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2015-01-22drm/radeon: use NULL rather then 0 in audio detectAlex Deucher1-1/+1
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2015-01-22drm/radeon: whitespace clean up in radeon_audio.cAlex Deucher1-33/+33
Clean up some formatting in radeon_audio.c to be more consistent with the rest of the driver. Reviewed-by: Christian König <christian.koenig@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2015-01-22radeon/audio: enable DP audioSlava Grigorev5-0/+147
Reviewed-by: Christian König <christian.koenig@amd.com> Signed-off-by: Slava Grigorev <slava.grigorev@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2015-01-22radeon/audio: moved audio caps programming to audio_hotplug() functionSlava Grigorev1-35/+43
Reviewed-by: Christian König <christian.koenig@amd.com> Signed-off-by: Slava Grigorev <slava.grigorev@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2015-01-22radeon/audio: applied audio_dpms() and audio_mode_set() callsSlava Grigorev5-26/+27
Reviewed-by: Christian König <christian.koenig@amd.com> Signed-off-by: Slava Grigorev <slava.grigorev@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2015-01-22radeon/audio: consolidate audio_mode_set() functionsSlava Grigorev8-162/+66
Reviewed-by: Christian König <christian.koenig@amd.com> Signed-off-by: Slava Grigorev <slava.grigorev@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2015-01-22radeon/audio: removed unnecessary debug settingsSlava Grigorev3-18/+0
Reviewed-by: Christian König <christian.koenig@amd.com> Signed-off-by: Slava Grigorev <slava.grigorev@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2015-01-22radeon/audio: moved mute programming to a separate functionSlava Grigorev5-11/+58
Reviewed-by: Christian König <christian.koenig@amd.com> Signed-off-by: Slava Grigorev <slava.grigorev@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2015-01-22radeon/audio: moved audio packet programming to a separate functionSlava Grigorev5-112/+134
Reviewed-by: Christian König <christian.koenig@amd.com> Signed-off-by: Slava Grigorev <slava.grigorev@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2015-01-22radeon/audio: set_avi_packet() function cleanupSlava Grigorev5-79/+59
Reviewed-by: Christian König <christian.koenig@amd.com> Signed-off-by: Slava Grigorev <slava.grigorev@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2015-01-22radeon/audio: removed unnecessary CRC control programingSlava Grigorev2-6/+0
Reviewed-by: Christian König <christian.koenig@amd.com> Signed-off-by: Slava Grigorev <slava.grigorev@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2015-01-22radeon: moved HDMI color depth programming to a separate functionSlava Grigorev5-34/+62
Reviewed-by: Christian König <christian.koenig@amd.com> Signed-off-by: Slava Grigorev <slava.grigorev@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2015-01-22radeon/audio: moved VBI packet programming to separate functionsAlex Deucher5-21/+45
Reviewed-by: Christian König <christian.koenig@amd.com> Signed-off-by: Slava Grigorev <slava.grigorev@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2015-01-22radeon/audio: consolidate update_acr() functions (v2)Slava Grigorev5-133/+183
V2: fix missing dce6 callback Signed-off-by: Slava Grigorev <slava.grigorev@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2015-01-22radeon/audio: consolidate update_avi_infoframe() functionsSlava Grigorev5-22/+34
Reviewed-by: Christian König <christian.koenig@amd.com> Signed-off-by: Slava Grigorev <slava.grigorev@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2015-01-22radeon/audio: consolidate audio_set_dto() functionsSlava Grigorev8-99/+225
Reviewed-by: Christian König <christian.koenig@amd.com> Signed-off-by: Slava Grigorev <slava.grigorev@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2015-01-22radeon/audio: consolidate audio_fini() functionsSlava Grigorev12-41/+26
Reviewed-by: Christian König <christian.koenig@amd.com> Signed-off-by: Slava Grigorev <slava.grigorev@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2015-01-22radeon/audio: consolidate audio_enable() functionsSlava Grigorev6-26/+28
Reviewed-by: Christian König <christian.koenig@amd.com> Signed-off-by: Slava Grigorev <slava.grigorev@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Copyright © 1996 – 2026 Jason A. Donenfeld. All Rights Reverse Engineered.