- 20 May, 2021 40 commits
-
-
Chengming Gui authored
Add support for beige_goby cp/rlc firmware Signed-off-by:
Chengming Gui <Jack.Gui@amd.com> Reviewed-by:
Hawking Zhang <Hawking.Zhang@amd.com> Reviewed-by:
Tao Zhou <tao.zhou1@amd.com> Signed-off-by:
Alex Deucher <alexander.deucher@amd.com>
-
Chengming Gui authored
Same as navi series Signed-off-by:
-
Chengming Gui authored
Use direct load for beige_goby Signed-off-by:
-
Chengming Gui authored
Add chip type for beige_goby v2: fix enum count (Alex) Signed-off-by:
-
Aric Cyr authored
Signed-off-by:
Aric Cyr <aric.cyr@amd.com> Acked-by:
Stylon Wang <stylon.wang@amd.com> Tested-by:
Daniel Wheeler <daniel.wheeler@amd.com> Signed-off-by:
-
Anthony Koo authored
Signed-off-by:
Anthony Koo <Anthony.Koo@amd.com> Acked-by:
-
Wyatt Wood authored
[Why] Visual confirm will indicate if driver is programming the surface address. Refactor is required because much of the visual confirm logic is buried deep in the mpcc files. In addition, visual confirm is not updated during fast updates. [How] In order to have visual confirm for driver flips, visual confirm needs to be updated on every frame, including fast updates. Add a new hw sequencer interface update_visual_confirm_color, and a new mpc function pointer set_bg_color. v2: drop unused variable (Alex) Signed-off-by:
Wyatt Wood <wyatt.wood@amd.com> Reviewed-by:
-
Nikola Cornij authored
[why] As per spec, DCN3.x can do 6:1 downscaling and DCN2.x can do 4:1. The max downscaling limit value for DCN2.x is 250, which means it's calculated as 1000 / 4 = 250. For DCN3.x this then gives 1000 / 6 = 167. [how] Set maximum downscaling limit to 167 for DCN3.x Signed-off-by:
Nikola Cornij <nikola.cornij@amd.com> Reviewed-by:
Charlene Liu <Charlene.Liu@amd.com> Reviewed-by:
Harry Wentland <Harry.Wentland@amd.com> Acked-by:
-
Wayne Lin authored
[Why] amdgpu_dm_crtc_set_crc_source() will call amdgpu_dm_crtc_configure_crc_source() to enable/disable CRC generation. However, configuration will be deferred to stream enabled. If stream is not enabled, current flow will still try to get/put vblank refcount. [How] Return EINVAL to skip actions on vblank refcount when stream is not enabled. Signed-off-by:
Wayne Lin <Wayne.Lin@amd.com> Reviewed-by:
-
Zhan Liu authored
[Why] eDP version and DPCD revision are different. Per VESA spec, "The DPCD revision for eDP v1.4 is 13h". SUPPORTED_LINK_RATES is valid since eDP v1.4 (DPCD_REV_13). [How] Correct DPCD_REV for eDP v1.4. Signed-off-by:
Zhan Liu <zhan.liu@amd.com> Reviewed-by:
-
Bhawanpreet Lakha authored
[Why] In some small modes (<200pixels) the stutter period is really big and will cause overflow. In these cases we shouldnt try to enable MALL as it will exceeds range of hysteresis timer (this can be seen in some IGT tests where the plane size is small) [How] Compare the stutter_period with the frame time and if we will overflow there is no point in trying to enable MALL (and see the ASSERT) so we early exist in this case Signed-off-by:
Bhawanpreet Lakha <Bhawanpreet.Lakha@amd.com> Reviewed-by:
Jun Lei <Jun.Lei@amd.com> Acked-by:
-
Wayne Lin authored
[Why] Once set ROI and do suspend/resume, current flow will not enable OTG_CRC_CTL again due to we'll defer crc configuration when stream is enabled. [How] Remove current suspend/resume function and have logic implemented into amdgpu_dm_atomic_commit_tail() Signed-off-by:
-
Wenjing Liu authored
[why] Some lttpr configuration steps are exclusive to 8b/10b channel coding mode. We need to take channel conding into account. Signed-off-by:
Wenjing Liu <wenjing.liu@amd.com> Reviewed-by:
George Shen <George.Shen@amd.com> Acked-by:
Wesley Chalmers <Wesley.Chalmers@amd.com> Tested-by:
-
Wenjing Liu authored
[why] The function's name doesn't represent what it actaully does. The function implements necessary steps for our hardware to transition from link training mode back to video idle mode. Therefore, rename the function as dp_transition_to_video_idle so everyone can understand it. Signed-off-by:
-
Wenjing Liu authored
[how] Rename initialize_training_settings to decide_training_settings. Call get link encoding format and decide training settings based on current channel coding. Signed-off-by:
-
Wenjing Liu authored
[how] Implement a function that determines link encoding format based on the link settings passed in. Signed-off-by:
-
Chris Park authored
[Why] Active DP dongles return no EDID when dongle is connected, but VGA display is taken out. Current driver behavior does not remove the active display when this happens, and this is a gap between dongle DTP and dongle behavior. [How] For active DP dongles and non-DP scenario, disconnect sink on detection when no EDID is read due to timeout. Signed-off-by:
Chris Park <Chris.Park@amd.com> Reviewed-by:
Nicholas Kazlauskas <Nicholas.Kazlauskas@amd.com> Acked-by:
-
George Shen authored
[Why] Improve readability and maintainability of code. [How] Refactor test pattern size calculation out of function call parameter and store value in variable. Signed-off-by:
-
Rodrigo Siqueira authored
During the investigation on how to add visual confirmation on top of the planes used by DCN, it becomes evident that the lack of information in the code makes this work unnecessarily complicated. This commit introduces a set of documentation related to the RECOUT operation in order to make it easy for developers to navigate this set of functions. Signed-off-by:
Rodrigo Siqueira <Rodrigo.Siqueira@amd.com> Reviewed-by:
-
Rodrigo Siqueira authored
Signed-off-by:
-
Rodrigo Siqueira authored
To improve the code readability, this commit removes a set of commented and not used functions for a long time. Notice that now we have the amdgpu_dm_dtn_log, which prints all the relevant information that we need. Signed-off-by:
-
Rodrigo Siqueira authored
This commit introduces kernel documentation to some essential functions related to power gate control over planes. It also adds a macro to make one part of the code easy to understand. Signed-off-by:
-
Wyatt Wood authored
[Why] Need to get current DMUB time. [How] Add get_current_time interface to dmub_srv. v2: drop whitespace changes (Alex) Signed-off-by:
-
Hugo Hu authored
Previous patch caused crash and had been reverted. This patch addresses the issue without regression. [Why] 1. Driver use umachannelnumber to calculate watermarks for stutter. In asymmetric memory config, the actual bandwidth is less than dual-channel. The bandwidth should be the same as single-channel. 2. We found single rank dimm need additional delay time for stutter. [How] Get information from each DIMM. Treat memory config as a single-channel for asymmetric memory in bandwidth calculating. Add additional delay time for single rank dimm. Signed-off-by:
Hugo Hu <hugo.hu@amd.com> Reviewed-by:
Sung Lee <Sung.Lee@amd.com> Acked-by:
-
Lijo Lazar authored
When determinism mode is disabled on aldebaran, max GFX clock will be reset to default max frequency value. Signed-off-by:
Lijo Lazar <lijo.lazar@amd.com> Reviewed-by:
Kevin Wang <kevin1.wang@amd.com> Signed-off-by:
-
Lijo Lazar authored
v1: Use the current and custom pstate frequencies to track the current and user-set min/max values in manual and determinism mode. Previously, only actual_* value was used to track the currrent and user requested value. The value will get reassigned whenever user requests a new value with pp_od_clk_voltage node. Hence it will show incorrect values when user requests an invalid value or tries a partial request without committing the values. Separating out to custom and current variable fixes such issues. v2: Remove redundant if-else check Signed-off-by:
-
Lijo Lazar authored
Add custom member for user requested custom frequency, level mask or min/max frequencies. Add curr member to keep track of the current active values. Signed-off-by:
-
John Clements authored
Only clear RAS error counters if perestent EDC harvesting is not supported Reviewed-by:
John Clements <john.clements@amd.com> Signed-off-by:
-
Changfeng authored
There is problem with 3DCGCG firmware and it will cause compute test hang on picasso/raven1. It needs to disable 3DCGCG in driver to avoid compute hang. Signed-off-by:
Changfeng <Changfeng.Zhu@amd.com> Reviewed-by:
Huang Rui <ray.huang@amd.com> Signed-off-by:
-
Peng Ju Zhou authored
there are 2 hubs to flush in the gmc, to make it easier to debug when hub flush failed, refine the logs. Signed-off-by:
Peng Ju Zhou <PengJu.Zhou@amd.com> Reviewed-by:
-
Peng Ju Zhou authored
KMD should not the program these registers, so skip them in the SRIOV environment. Signed-off-by:
-
Yi Li authored
When PAGE_SIZE is larger than AMDGPU_PAGE_SIZE, the number of GPU TLB entries which need to update in amdgpu_map_buffer() should be multiplied by AMDGPU_GPU_PAGES_IN_CPU_PAGE (PAGE_SIZE / AMDGPU_PAGE_SIZE). Reviewed-by:
Christian König <christian.koenig@amd.com> Signed-off-by:
Yi Li <liyi@loongson.cn> Signed-off-by:
Huacai Chen <chenhuacai@loongson.cn> Signed-off-by:
-
Philip Yang authored
Need do a heavy-weight TLB flush to make sure we have no more dirty data in the cache for the unmapped pages. Define enum TLB_FLUSH_TYPE, add flush_type parameter to amdgpu_amdkfd_flush_gpu_tlb_pasid. Signed-off-by:
Philip Yang <Philip.Yang@amd.com> Reviewed-by:
Felix Kuehling <Felix.Kuehling@amd.com> Reviewed-by:
-
Philip Yang authored
This reverts commit 1704ac8e. After "drm/amdgpu: flush TLB if valid PDE turns into PTE" is checked in, this workaround is not needed. Signed-off-by:
-
Jiawei Gu authored
Initialize unique_id from PF2VF under virtualization. V2: skip smu_get_unique_id() under virtualization Signed-off-by:
Jiawei Gu <Jiawei.Gu@amd.com> Acked-by:
-
Philip Yang authored
Mapping huge page, 2MB aligned address with 2MB size, uses PDE0 as PTE. If previously valid PDE0, PDE0.V=1 and PDE0.P=0 turns into PTE, this requires TLB flush, otherwise page table walker will not read updated PDE0. Change page table update mapping to return table_freed flag to indicate the previously valid PDE may have turned into a PTE if page table is freed. Signed-off-by:
-
Christian König authored
Imported BOs don't have a pagelist any more. Signed-off-by:
-
Christian König authored
Not used outside of that file. Signed-off-by:
Nirmoy Das <nirmoy.das@amd.com> Signed-off-by:
-
Christian König authored
Drop the last user of drm_mm_node. Signed-off-by:
-
Christian König authored
This avoids reallocating scanout BOs on first pin in a lot of cases. Signed-off-by:
-
