- 19 Mar, 2019 40 commits
-
-
xinhui pan authored
Add ras fw part, xgmi and ras fw are combined together in ta binary. Reading the data from the info is not implemented yet. v2: squash in "drm/amdgpu: fix NULL pointer when ta is missing" Signed-off-by:
xinhui pan <xinhui.pan@amd.com> Reviewed-by:
Alex Deucher <alexander.deucher@amd.com> Signed-off-by:
-
Kevin Wang authored
each asic may be has different smc if version, so move its to asic file to implement. Signed-off-by:
Kevin Wang <kevin1.wang@amd.com> Reviewed-by:
Huang Rui <ray.huang@amd.com> Signed-off-by:
-
Kevin Wang authored
simplify macro of MSG_MAP for sw-smu Signed-off-by:
-
Oak Zeng authored
Use function parameter mc as the second parameter of amdgpu_gmc_vram_location, so codes look more consistent. Signed-off-by:
Oak Zeng <Oak.Zeng@amd.com> Reviewed-by:
Christian Konig <christian.koenig@amd.com> Signed-off-by:
-
Nicholas Kazlauskas authored
[Why] For new DC planes the correct plane address fields are filled based on whether the plane had a graphics or video format. However, when we perform stream and plane updates using DC we only ever fill in the graphics format fields. This causing corruption and hangs when using video surface formats like NV12 for planes. [How] Use the same logic everywhere we update dc_plane_address - always fill in the correct fields based on the surface format type. There are 3 places this is done: - Atomic check, during DC plane creation - Atomic commit, during plane prepare_fb - Atomic commit tail, during amdgpu_dm_commit_planes We use the fill_plane_tiling_attributes in all 3 locations and it already needs the address to update DCC attributes, so the surface address update logic can be moved into this helper. Signed-off-by:
Nicholas Kazlauskas <nicholas.kazlauskas@amd.com> Reviewed-by:
Leo Li <sunpeng.li@amd.com> Signed-off-by:
-
Andrey Grodzovsky authored
Problem: Using SDMA for TLB invalidation in certain ASICs exposed a problem of IB pool not being ready while SDMA already up on Init and already shutt down while SDMA still running on Fini. This caused IB allocation failure. Temproary fix was commited into a bringup branch but this is the generic fix. Fix: Init IB pool rigth after GMC is ready but before SDMA is ready. Do th opposite for Fini. v2: Remove restriction on SDMA early init and move amdgpu_ib_pool_fini Reviewed-by:
Andrey Grodzovsky <andrey.grodzovsky@amd.com> Signed-off-by:
-
David Francis authored
[Why] Cursor updates must acquire the pipe control lock to prevent vupdate from triggering in the middle of cursor programming. On DCN1 the pipe control lock prevents pageflips from occurring. This means that a cursor update right before vupdate can delay a pending pageflip [How] If the time until the next vupdate is less than a conservative estimate of the cursor programming time, wait until the vupdate has passed before locking. Signed-off-by:
David Francis <David.Francis@amd.com> Reviewed-by:
Harry Wentland <Harry.Wentland@amd.com> Acked-by:
-
Wenjing Liu authored
[why] We will not retry when EDID read failure using i2c over aux [how] treat i2c over aux failure the same as defer Signed-off-by:
Wenjing Liu <Wenjing.Liu@amd.com> Reviewed-by:
-
Jun Lei authored
[why] "reference clock" is a very overloaded variable in DC and causes confusion as there are multiple sources of reference clock, which may be different values incorrect input values to DML will cause DCHUB to be programmed improperly and lead to hard to debug underflow issues [how] instead of using ref clock everywhere, specify WHICH ref clock: - xtalin - dccg refclk - dchub refclk these are all distinct values which may not be equal Signed-off-by:
Jun Lei <Jun.Lei@amd.com> Reviewed-by:
Yongqiang Sun <yongqiang.sun@amd.com> Acked-by:
-
Nicholas Kazlauskas authored
[Why] The DRM overlay planes DM exposes support RGBA formats but are currently forced as fully opaque over whatever they overlay. [How] Expose DRM blending mode and alpha properties to userspace. The overlays exposed support per-pixel pre-multiplied alpha along with global plane opacity. Signed-off-by:
-
Charlene Liu authored
Signed-off-by:
Charlene Liu <charlene.liu@amd.com> Reviewed-by:
-
Jun Lei authored
[why] Actual breakdown of DPM level varies by SKU (for the same family) DC needs some ability to ammend pre-silicon numbers Signed-off-by:
Tony Cheng <Tony.Cheng@amd.com> Acked-by:
-
Aric Cyr authored
Signed-off-by:
Aric Cyr <aric.cyr@amd.com> Reviewed-by:
-
Josip Pavic authored
[Why] Increased power savings are desired for ABM 2.2. [How] Reduce the minimum reduction level, the deviation gain and the contrast factor to allow for more aggressive operation of the algorithm. Signed-off-by:
Josip Pavic <Josip.Pavic@amd.com> Reviewed-by:
Anthony Koo <Anthony.Koo@amd.com> Acked-by:
-
Eric Bernstein authored
HDMI has TMDS and FRL signal types. Be specific about what is used. Signed-off-by:
Eric Bernstein <eric.bernstein@amd.com> Reviewed-by:
Nevenko Stupar <Nevenko.Stupar@amd.com> Reviewed-by:
-
Wenjing Liu authored
[why] Stream update will adjust both info packets and stream params, need to make sure all things are applied togather. [how] add pipe lock during stream update Signed-off-by:
-
Ken Chalmers authored
[Why] At 24 Hz, a frame is 41.7 ms, so a 30 ms wait can (and does often) timeout. [How] Bump timeout from 30 ms to 50 ms. Signed-off-by:
Ken Chalmers <ken.chalmers@amd.com> Reviewed-by:
-
Yongqiang Sun authored
[Why] field toggle write is actual field sequence write with the same field name. [How] Use REG_UPDATE_SEQ_2 for both sequence write and toggle. Rename REG_UPDATE_1by1_3 to REG_UPDATE_SEQ_3. Signed-off-by:
-
SivapiriyanKumarasamy authored
[WHY] We have new bios capabilities enabling s0i2 entry on SMU interrupt. We want this interrupt to be fired on PSR transitions such that we enter s0i2 when entering PSR active. [HOW] Add code to send the SMU interrupt with the appropriate staticscreen flag when entering and exting PSR. Protect this code with a config flag since it currently impacts BL PWM. Signed-off-by:
SivapiriyanKumarasamy <sivapiriyan.kumarasamy@amd.com> Reviewed-by:
-
Yongqiang Sun authored
[Why] Current reg update and reg set use same functions and only delta is update reads reg value and call update function. [How] Refactor reg update and reg set functions. 1.Implement different functions for reg update and reg set. 2.Wrap same process to a help function, both reg update and reg set will call it. Signed-off-by:
-
Eric Bernstein authored
Cross a TODO item off the list. Cleanup SIGNAL_TYPE_HDMI_FRL, it's not currently supported. Signed-off-by:
Dmytro Laktyushkin <Dmytro.Laktyushkin@amd.com> Acked-by:
-
Dmytro Laktyushkin authored
Get rid of DV style dml init in favour of the cleaner DC style. Signed-off-by:
-
Evan Quan authored
Applied vdci flush workaround for Vega20 BACO. Signed-off-by:
Evan Quan <evan.quan@amd.com> Acked-by:
-
shaoyunl authored
Driver vote low to high pstate switch whenever there is an outstanding XGMI mapping request. Driver vote high to low pstate when all the outstanding XGMI mapping is terminated. Signed-off-by:
shaoyunl <shaoyun.liu@amd.com> Reviewed-by:
-
shaoyunl authored
Adjust vram base offset for XGMI mapping when update the PT entry so the address will fall into correct XGMI aperture for peer device Signed-off-by:
-
Andrey Grodzovsky authored
For each device a file xgmi_device_id is created. On the first device a subdirectory named xgmi_hive_info is created, It contains a file named hive_id and symlinks named node 1-4 linking to each device in the hive. v2: Return error codes instead of '-1' and few misspellings. Signed-off-by:
-
Nicholas Kazlauskas authored
[Why] Can happen on ASICs with 6 planes, but this isn't a bug since we haven't written outside the array. [How] Use <= instead of <. Cc: Leo Li <sunpeng.li@amd.com> Signed-off-by:
Michel Dänzer <michel.daenzer@amd.com> Reviewed-by:
-
Nicholas Kazlauskas authored
[Why] New DRM versions manage locking for private objects for us, so this is no longer needed. This also prevents a WARN_ON from occurring when the private object is duplicated during the forced atomic commit that occurs from the HPD handler. The HPD handler calls drm_modeset_lock_all before the forced commit and if the private object is duplicated then the DEBUG_LOCKS_WARN_ON(ww_ctx->done_acquire) warning will be triggered since we're trying to lock something when everything should have already been locked. [How] Drop the lock and let DRM manage this. Cc: Harry Wentland <harry.wentland@amd.com> Signed-off-by:
-
Tom St Denis authored
Signed-off-by:
Tom St Denis <tom.stdenis@amd.com> Acked-by:
-
Christian König authored
Only GMC9 supports true huge pages, but we can still free invalid mappings on GMC8. Signed-off-by:
Felix Kuehling <Felix.Kuehling@amd.com> Acked-by:
-
Christian König authored
Not needed any more since we now free PDs/PTs on demand. Signed-off-by:
-
Christian König authored
When something is unmapped we now free the affected PDs/PTs again. Signed-off-by:
-
Christian König authored
Let's start to allocate VM PDs/PTs on demand instead of pre-allocating them during mapping. Signed-off-by:
-
Christian König authored
Instead of providing it from outside figure out the ats status in the function itself from the data structures. v2: simplify finding the right level v3: partially revert changes from v2, more cleanup and split code into more functions. Signed-off-by: -
Christian König authored
This way we only deal with the real BO in here. v2: use a do { ... } while loop instead v3: fix NULL pointer in v2 Signed-off-by: -
Likun Gao authored
The operation of mutex_unlock smu->mutex should be done when forced level is larger than NUM_LINK_LEVELS in the function of force_clk_levels. Reported-by:
Julia Lawall <julia.lawall@lip6.fr> Signed-off-by:
Likun Gao <Likun.Gao@amd.com> Reviewed-by:
-
kbuild test robot authored
Simplify the code a bit by using kmemdup instead of kzalloc and memcpy. Generated by: scripts/coccinelle/api/memdup.cocci CC: Likun Gao <Likun.Gao@amd.com> Reviewed-by:
kbuild test robot <fengguang.wu@intel.com> Signed-off-by:
-
Colin Ian King authored
An earlier commit replaced ttm_bo_wait with amdgpu_bo_sync_wait and removed the error return assignment to variable ret. Fix this by adding the assignment back. Also break line to clean up checkpatch overly long line warning. Detected by CoverityScan, CID#1477327 ("Logically dead code") Fixes: c60cd590 ("drm/amdgpu: Replace ttm_bo_wait with amdgpu_bo_sync_wait") Signed-off-by:Colin Ian King <colin.king@canonical.com> Signed-off-by:
-
Christian König authored
Same patch we alredy did for Vega10. Just re-route page faults to a separate ring to avoid drowning in interrupts. Signed-off-by:
-
Christian König authored
Page faults can easily overwhelm the interrupt handler. So to make sure that we never lose valuable interrupts on the primary ring we re-route page faults to IH ring 1. Signed-off-by:
-
