- 06 Feb, 2019 25 commits
-
-
Chiawen Huang authored
[Why] When querying HPD via GPIO flow, it will create a new gpio object then free in the end of query. There is a irql issue for HPD querying at ISR level. [How] Therefore, creating the HPD gpio object in dc_link and set it as unlcok in default. 1. reducing unnecessary malloc/free when HPD querying. 2. reducing init GPIO flow. 3. add lock/unlock to prevent multi gpio service running. Signed-off-by:
Chiawen Huang <chiawen.huang@amd.com> Reviewed-by:
Tony Cheng <Tony.Cheng@amd.com> Acked-by:
Bhawanpreet Lakha <Bhawanpreet.Lakha@amd.com> Signed-off-by:
Alex Deucher <alexander.deucher@amd.com>
-
Steven Chiu authored
Signed-off-by:
Steven Chiu <steven.chiu@amd.com> Reviewed-by:
Aric Cyr <Aric.Cyr@amd.com> Acked-by:
-
Wenjing Liu authored
[why] is_blanked is not a general indicator of if a pipe is synced for all asics. plane state is more accurate and applicable for all asics. [how] Remove is_blanked call and add checking plane_state against NULL instead. Signed-off-by:
Wenjing Liu <Wenjing.Liu@amd.com> Reviewed-by:
-
Krunoslav Kovac authored
[Why] On HW that doesn't have input LUT, we may combine degamma with OS ramp Problem here is that it assumes DGAM is inverse of PQ or SRGB. It doesn't handle linear case, it would default to sRGB and always enable DGAM.. [How] Add handling for linear case. Also check for null ramp and instead of blowing up, assume it's identity. Signed-off-by:
Krunoslav Kovac <Krunoslav.Kovac@amd.com> Reviewed-by:
Anthony Koo <Anthony.Koo@amd.com> Acked-by:
-
Ilya Bakoulin authored
[Why] Setting monitor refresh rate below freesync range would cause the monitor to go blank indefinitely with freesync enabled [How] Set vrr_supported and ignore_msa_timing_param according to whether the refresh rate is above or below the minimum freesync frequency. Signed-off-by:
Ilya Bakoulin <Ilya.Bakoulin@amd.com> Reviewed-by:
Nicholas Kazlauskas <Nicholas.Kazlauskas@amd.com> Acked-by:
-
Paul Hsieh authored
[Why] Second screen to clone/extend mode, driver will send ABM pipe command to DMCU. Change mode from clone/extend to second screen only, driver send ABM level command to disable ABM but this command will not clear ABM pipe data. At this time, change second screen to PC screen only, driver will send first command "ABM_LEVEL", it will turn on ABM with incorrect ABM pile so that DMCU is blocking. [How] When driver try to disable ABM, change command from "ABM LEVEL" to "ABM PIPE" so that it will clear ABM pile data. Signed-off-by:
Paul Hsieh <Paul.Hsieh@amd.com> Reviewed-by:
Charlene Liu <Charlene.Liu@amd.com> Acked-by:
-
Anthony Koo authored
[Why] The problem is that we accidentally stopped loading some of the IRAM bytes used for the backlight ramping mechanism. This happened when we started reserving some region of IRAM as DMCU FW write only. [How] This change will define a start+end region for the IRAM read only region. So the parameters needed for the backlight operation will be loaded since it will be defined outside of the read only region. Signed-off-by:
-
Xiaodong Yan authored
[Why] Different panel need different time from backlight disable to end of valid video data, if the time is too short, panel will flash when dpms off [How] Add monitor patch to control the time from backlight disable to end of valid video data, Signed-off-by:
Xiaodong Yan <Xiaodong.Yan@amd.com> Reviewed-by:
-
Nicholas Kazlauskas authored
[Why] This provides little debug value and creates a lot of dmesg noise. [How] Remove it. Signed-off-by:
David Francis <David.Francis@amd.com> Acked-by:
-
Nicholas Kazlauskas authored
[Why] The stream_update struct is left unitialized but DC will access its fields. This usually results in global state validation occur during any atomic commit with state->allow_modeset = true. [How] Initialize the struct to zero for every stream we check. Signed-off-by:
-
Nicholas Kazlauskas authored
[Why] The indenting for this function is a few levels too deep and can be simplified a fair bit. This patch is in preparation for functional changes that fix update type determination to occur less frequently and more accurately. [How] Place checks early and exit/continue when possible. This isn't a functional change. Signed-off-by:
-
Nicholas Kazlauskas authored
[Why] We were always passing the first surface on the stream status for flip updates when we should be using the surface associated with the plane. [How] Use the dc_plane_state from the plane that's being updated. FreeSync should also only keep track of updates from the primary plane, so the check needed to be updated. The acrtc->stream state doesn't need to be checked for NULL before updating FreeSync either since there needs to be a stream to be inside this function as a prerequisite. Signed-off-by:
-
Nicholas Kazlauskas authored
[Why] Most surface updates weren't propagated onto the surface during dc_commit_updates_for_stream. This makes it more difficult for DC to determine the actual surface update type required. [How] Use copy_surface_update_to_plane to propagate the changes. The FreeSync surface timing information update for BTR has been moved out of amdgpu_dm.c into this function as well. Signed-off-by:
Sun peng Li <Sunpeng.Li@amd.com> Acked-by:
-
John Barberiz authored
[Why] "IRQ_HPD Pulse Length Test" DP compliance test fails. Test complains that certain DPCD registers are not read within 100 ms. [How] msleep is inaccurate for small values. Used udelay instead for accuracy. Signed-off-by:
John Barberiz <John.Barberiz@amd.com> Reviewed-by:
-
Christian König authored
No functional change. Signed-off-by:
Christian König <christian.koenig@amd.com> Reviewed-by:
Felix Kuehling <Felix.Kuehling@amd.com> Signed-off-by:
-
Christian König authored
Otherwise we open up the possibility to use uninitialized memory. Signed-off-by:
-
Shirish S authored
[What] FBC fails to get enabled when switched between LINEAR(console/VT) and non-LINEAR(GUI) based rendering due to default value of tiling info stored in the current_state which is used for deciding whether or not to turn FBC on or off. [How] Use context structure's tiling information which is coherant with the screen updates. Signed-off-by:
Shirish S <shirish.s@amd.com> Reviewed-by:
Roman Li <Roman.Li@amd.com> Signed-off-by:
-
Nicholas Kazlauskas authored
[Why] When CONFIG_DEBUG_FS isn't defined then amdgpu_dm_crtc_set_crc_source is NULL. This causes a compilation error since it's being called unconditionally. [How] Guard the call based on CONFIG_DEBUG_FS - CRC capture isn't supported without this. Cc: Leo Li <sunpeng.li@amd.com> Cc: Harry Wentland <harry.wentland@amd.com> Fixes: 43a6a02e ("drm/amd/display: Re-enable CRC capture following modeset") Signed-off-by:
Harry Wentland <harry.wentland@amd.com> Signed-off-by:
-
Marek Olšák authored
I'm not increasing the DRM version because GDS isn't totally without bugs yet. v2: update emit_ib_size Signed-off-by:
Marek Olšák <marek.olsak@amd.com> Acked-by:
-
Andrey Grodzovsky authored
New chunk for dependency on start of job's execution instead on the end. This is used for GPU deadlock prevention when userspace uses mid-IB fences to wait for mid-IB work on other rings. v2: Fix typo in AMDGPU_CHUNK_ID_SCHEDULED_DEPENDENCIES v3: Bump KMS version v4: put old fence AFTER acquiring the scheduled fence. Signed-off-by:
Andrey Grodzovsky <andrey.grodzovsky@amd.com> Suggested-by:
-
Kenneth Feng authored
update soc boot and max level,then uclk isn't stuck at minimum. Bug: https://bugs.freedesktop.org/show_bug.cgi?id=109462Signed-off-by:
Kenneth Feng <kenneth.feng@amd.com> Acked-by:
-
Nathan Chancellor authored
Clang warns: drivers/gpu/drm/amd/amdgpu/../display/amdgpu_dm/amdgpu_dm.c:2314:38: warning: suggest braces around initialization of subobject [-Wmissing-braces] struct dc_surface_dcc_cap output = {0}; ^ {} Previous efforts to fix this type of warning by adding or removing braces have been met with some pushback in favor of using memset [1][2]. Do that here, mirroring commit 05794eff ("drm/amdgpu/gmc: fix compiler errors [-Werror,-Wmissing-braces] (V2)") in this tree. [1]: https://lore.kernel.org/lkml/022e41c0-8465-dc7a-a45c-64187ecd9684@amd.com/ [2]: https://lore.kernel.org/lkml/20181128.215241.702406654469517539.davem@davemloft.net/ Fixes: 7df7e505 ("drm/amd/display: Set requested plane state DCC params for GFX9") Signed-off-by:Nathan Chancellor <natechancellor@gmail.com> Signed-off-by:
-
Nathan Chancellor authored
Clang warns: drivers/gpu/drm/amd/amdgpu/../display/amdgpu_dm/amdgpu_dm.c:5089:60: warning: suggest braces around initialization of subobject [-Wmissing-braces] struct dc_surface_update dummy_updates[MAX_SURFACES] = { 0 }; ^ {} Previous efforts to fix this type of warning by adding or removing braces have been met with some pushback in favor of using memset [1][2]. Do that here, mirroring commit 05794eff ("drm/amdgpu/gmc: fix compiler errors [-Werror,-Wmissing-braces] (V2)") in this tree. [1]: https://lore.kernel.org/lkml/022e41c0-8465-dc7a-a45c-64187ecd9684@amd.com/ [2]: https://lore.kernel.org/lkml/20181128.215241.702406654469517539.davem@davemloft.net/ Fixes: 02d6a6fc ("drm/amd/display: Simplify underscan and ABM commit") Signed-off-by: -
Nathan Chancellor authored
Clang warns: drivers/gpu/drm/amd/amdgpu/../display/dc/core/dc_link_dp.c:50:57: warning: suggest braces around initialization of subobject [-Wmissing-braces] union training_aux_rd_interval training_rd_interval = {0}; ^ {} 1 warning generated. Previous efforts to fix this type of warning by adding or removing braces have been met with some pushback in favor of using memset [1][2]. Do that here, mirroring commit 05794eff ("drm/amdgpu/gmc: fix compiler errors [-Werror,-Wmissing-braces] (V2)") in this tree. [1]: https://lore.kernel.org/lkml/022e41c0-8465-dc7a-a45c-64187ecd9684@amd.com/ [2]: https://lore.kernel.org/lkml/20181128.215241.702406654469517539.davem@davemloft.net/ Fixes: 3cec4176 ("drm/amd/display: Fix use of uninitialized union") Signed-off-by: -
Colin Ian King authored
There is a spelling mistake in a dev_err message. Fix it. Signed-off-by:
Colin Ian King <colin.king@canonical.com> Signed-off-by:
-
- 01 Feb, 2019 1 commit
-
-
Marek Olšák authored
- move all adjustments into one place - specify GDS/GWS/OA alignment in basic units of the heaps - it looks like GDS alignment was 1 instead of 4 Signed-off-by:
-
- 29 Jan, 2019 5 commits
-
-
Christian König authored
kptr is not used any more. Signed-off-by:
Chunming Zhou <david1.zhou@amd.com> Signed-off-by:
-
shaoyunl authored
Reduce the repeated node and hive information during XGMI initialization Signed-off-by:
shaoyunl <shaoyun.liu@amd.com> Acked-by:
-
Alex Deucher authored
Revert "drm/amd/display: add -msse2 to prevent Clang from emitting libcalls to undefined SW FP routines" This reverts commit 10117450. Causes a crash. Bug: https://bugs.freedesktop.org/show_bug.cgi?id=109487Reviewed-by:
-
wentalou authored
sriov need to restrict max_pfn below AMDGPU_GMC_HOLE. access the hole results in a range fault interrupt IIRC. Signed-off-by:
Wentao Lou <Wentao.Lou@amd.com> Reviewed-by:
-
wentalou authored
After removing unnecessary VM size calculations, vm_manager.max_pfn would reach 0x10,0000,0000 max_pfn << AMDGPU_GPU_PAGE_SHIFT exceeding AMDGPU_GMC_HOLE_START would cause GPU reset. Signed-off-by:
Monk Liu <monk.liu@amd.com> Signed-off-by:
-
- 28 Jan, 2019 9 commits
-
-
Eric Huang authored
It is to solve RDMA performance issue. Signed-off-by:
Eric Huang <JinhuiEric.Huang@amd.com> Acked-by:
-
Nicholas Kazlauskas authored
[Why] The flip and full structures were allocated but never freed. [How] Free them at the end of the function. There's a small behavioral change here with the function returning early if the allocation fails but we wouldn't should be doing anything in that case anyway. Fixes: c00e0cc0fdc0 ("drm/amd/display: Call into DC once per multiplane flip") Fixes: ea39594e0855 ("drm/amd/display: Perform plane updates only when needed") Signed-off-by:Michel Dänzer <michel.daenzer@amd.com> Reviewed-by:
-
Fatemeh Darbehani authored
[Why] Enhanced sync need to use vertical_interrupt1. [How] Add vertical_interrupt1 source to irq manger, Implment setup vline interrupt interface. Signed-off-by:
Fatemeh Darbehani <fatemeh.darbehani@amd.com> Reviewed-by:
Jun Lei <Jun.Lei@amd.com> Acked-by:
-
Nicholas Kazlauskas authored
[Why] During any modeset the CRTC stream is removed and a new stream is added. This new stream doesn't carry over CRC capture state if it was previously set. [How] Re-program the stream for CRC capture. The existing DRM callback can be re-used here for the most part - the only modification needed is additional locking now that it's called from within commit tail. Signed-off-by:
-
Nicholas Kazlauskas authored
[Why] In order to read CRC events when CRC capture is enabled the vblank interrput handler needs to be running for the CRTC. The handler is enabled while there is an active vblank reference. When running IGT tests there will often be no active vblank reference but the test expects to read a CRC value. This is valid usage (and works on i915 since they have a CRC interrupt handler) so the reference to the vblank should be grabbed while capture is active. This issue was found running: igt@kms_plane_multiple@atomic-pipe-b-tiling-none The pipe-b is the only one in the initial commit and was not previously active so no vblank reference is grabbed. The vblank interrupt is not enabled and the test times out. [How] Keep a reference to the vblank as long as CRC capture is enabled. If userspace never explicitly disables it then the reference is also dropped when removing the CRTC from the context (stream = NULL). Signed-off-by:
-
Martin Tsai authored
[Why] On current design, driver cannot handle the interrupt for down reply when link training is processing. The DOWN REQ send before link training will keep in the pending DOWN REP state in the queue. It makes the next DOWN REQ be queued until time out. [How] To add a polling sequence before clear payload allocation table to make sure the pending DOWN REP can be handled. Signed-off-by:
Martin Tsai <martin.tsai@amd.com> Reviewed-by:
-
Thomas Zimmermann authored
Both functions are obsolete and all calls have been replaced by ttm_bo_get and ttm_bo_put. Signed-off-by:
Thomas Zimmermann <tzimmermann@suse.de> Reviewed-by:
Huang Rui <ray.huang@amd.com> Signed-off-by:
-
Thomas Zimmermann authored
The function ttm_bo_put releases a reference to a TTM buffer object. The function's name is more aligned to the Linux kernel convention of naming ref-counting function _get and _put. A call to ttm_bo_unref takes the address of the TTM BO object's pointer and clears the pointer's value to NULL. This is not necessary in most cases and sometimes even worked around by the calling code. A call to ttm_bo_put only releases the reference without clearing the pointer. The current behaviour of cleaning the pointer is kept in the calling code, but should be removed if not required in a later patch. Signed-off-by:
-
Thomas Zimmermann authored
The function ttm_bo_put releases a reference to a TTM buffer object. The function's name is more aligned to the Linux kernel convention of naming ref-counting function _get and _put. A call to ttm_bo_unref takes the address of the TTM BO object's pointer and clears the pointer's value to NULL. This is not necessary in most cases and sometimes even worked around by the calling code. A call to ttm_bo_put only releases the reference without clearing the pointer. In places where is might be necessary, the current behaviour of cleaning the pointer is kept. Signed-off-by:
-
