- 20 May, 2024 32 commits
-
-
Lijo Lazar authored
Add documentation about the newly added pm_policy node in sysfs. Signed-off-by:
Lijo Lazar <lijo.lazar@amd.com> Acked-by:
Alex Deucher <alexander.deucher@amd.com> Reviewed-by:
Asad Kamal <asad.kamal@amd.com> Signed-off-by:
-
Asad Kamal authored
Remove gpu_metrics_v1_6 usage for SMUv13.0.6 temporarily and use gpu_metrics_v1_5 until tool support is ready for it. This reverts commit 74ab6253. Signed-off-by:
-
Lijo Lazar authored
Remove unused callback to set PLPD policy and its implementation from arcturus, aldebaran and SMUv13.0.6 SOCs. Signed-off-by:
Hawking Zhang <Hawking.Zhang@amd.com> Reviewed-by:
-
Tao Zhou authored
Avoid overflow issue. Signed-off-by:
Tao Zhou <tao.zhou1@amd.com> Reviewed-by:
Yang Wang <kevinyang.wang@amd.com> Signed-off-by:
-
Li Ma authored
MinGfxClk in Dpmclocks_t is wrong. According to pmfw, dropping an uint8_t spare in Dpmclocks_t. Signed-off-by:
Li Ma <li.ma@amd.com> Reviewed-by:
Yifan Zhang <yifan1.zhang@amd.com> Signed-off-by:
-
Victor Skvortsov authored
The guest recovery sequence is buggy in Fatal Error when both FLR & KFD reset workitems are queued at the same time. In addition, FLR guest recovery sequence is out of order when PF/VF communication breaks due to a GPU fatal error As a temporary work around, perform a KFD style reset (Initiate reset request from the guest) inside the pf2vf thread on FED. Signed-off-by:
Victor Skvortsov <victor.skvortsov@amd.com> Reviewed-by:
Zhigang Luo <zhigang.luo@amd.com> Signed-off-by:
-
Victor Skvortsov authored
Runtime KIQ interface to read/write registers in VF may take longer than expected for BM environment. Extend the timeout. Signed-off-by:
-
Aric Cyr authored
This version pairs with DMUB FW Release 0.0.218.0 for dcn314/315/316, dcn35/351, dcn401 and brings along the following: - Fix powerpc compilation - Fix TBT+TypeC Daisy-chain lightup - Fix ODM combine setup - Fix OTC underflow on dcn35 - Fix DVI config for dcn401 - Add ips status info to debugfs - Add 3DLUT DMA load trigger - Modify clock programming to support DPM - Disable dcn401 idle optimizations Acked-by:
Roman Li <roman.li@amd.com> Signed-off-by:
Aric Cyr <aric.cyr@amd.com> Tested-by:
Daniel Wheeler <daniel.wheeler@amd.com> Signed-off-by:
-
Roman Li authored
[Why] Compilation errors while compiling without CONFIG_DRM_AMD_DC_FP: "undefined reference to `dc_bandwidth_in_kbps_from_timing'" [How] Fix Makefile to move dsc files out of DC_FP guard. Fixes: 00c39110 ("drm/amd/display: Add misc DC changes for DCN401") Reviewed-by:
Aurabindo Pillai <aurabindo.pillai@amd.com> Signed-off-by:
-
Cruise authored
[Why] When the link BW is smaller than the request BW, the DP LT just kept running and fallback to lower link config. DP LT just aborted if is_hpd_pending bit is high. But is_hpd_pending bit indicates a new HPD event received. It doesn't mean the HPD is low. [How] Abort the DP LT if the link BW is smaller than the request BW. Remove checking is_hpd_pending bit in DP LT. Reviewed-by:
Wenjing Liu <wenjing.liu@amd.com> Acked-by:
Cruise <cruise.hung@amd.com> Tested-by:
-
Dillon Varone authored
[WHY&HOW] At the time of block sequence construction, the exact reference DPP/DISP clock is not yet known, so the clock should be passed by reference to the DTO programming function. Reviewed-by:
Alvin Lee <alvin.lee2@amd.com> Acked-by:
Dillon Varone <dillon.varone@amd.com> Tested-by:
-
Dillon Varone authored
[WHY&HOW] Disable to improve stability for now. Reviewed-by:
-
Alvin Lee authored
[Description] There is a corner case where we're in an ODM config that has recout.x != 0. In these scenarios we have to take into account the extra offset in the ODM adjustment for cursor. Reviewed-by:
-
Wenjing Liu authored
[why] A recent change for ODM combine refactor contains a typo which causes ODM combine mode programmed incorrectly. Reviewed-by:
George Shen <george.shen@amd.com> Acked-by:
-
Roman Li authored
[Why] Disable idle optimization for each atomic commit is unnecessary, and can lead to a potential race condition. [How] Remove idle optimization check from amdgpu_dm_atomic_commit_tail() Fixes: 196107eb ("drm/amd/display: Add IPS checks before dcn register access") Cc: stable@vger.kernel.org Reviewed-by:
Hamza Mahfooz <hamza.mahfooz@amd.com> Acked-by:
-
Nicholas Susanto authored
[Why] Missing check for when there is new pipe configuration but both cur_pipe and new_pipe are both populated causing update_state of DSC for that instance not being updated correctly. This causes some display mode changes to cause underflow since DSCCLK is still gated when the display requires DSC. [How] Added another condition in the new pipe addition branch that checks if there is a new pipe configuration and if it is not the same as cur_pipe. cur_pipe does not necessarily have to be NULL to go in this branch. Reviewed-by:
Nicholas Kazlauskas <nicholas.kazlauskas@amd.com> Acked-by:
Nicholas Susanto <nicholas.susanto@amd.com> Tested-by:
-
Roman Li authored
[Why] For debugging and testing purposes. [How] If IPS is supported create ips_status debugfs entry. Usage: cat /sys/kernel/debug/dri/0/amdgpu_dm_ips_status Reviewed-by:
Jerry Zuo <jerry.zuo@amd.com> Acked-by:
-
Roman Li authored
[Why] Reset the shared dmub firmware region on dmub hw init to start with known state. [How] Memset the shared region to 0 in dmub_hw_init(). Suggested-by:
-
Ilya Bakoulin authored
[Why/How] Need to be able to trigger a DMA load to update 3DLUT contents in MPC. Adding a HWSS function to serve as the trigger. Reviewed-by:
Krunoslav Kovac <krunoslav.kovac@amd.com> Acked-by:
Ilya Bakoulin <ilya.bakoulin@amd.com> Tested-by:
-
Chris Park authored
[Why] DML 2.1 allocates two types of memory in its ctx structure but does not destroy them, causing memory leak whenever DML 2.1 instance is created and destroyed. [How] Deallocate two instances of allocated memory whenever DML 2.1 is destroyed. Reviewed-by:
Rodrigo Siqueira <Rodrigo.Siqueira@amd.com> Signed-off-by:
Chris Park <chris.park@amd.com> Tested-by:
-
Hersen Wu authored
[Why] Coverity reports NULL_RETURN warning. [How] Add pointer NULL check. Reviewed-by:
Hersen Wu <hersenxs.wu@amd.com> Tested-by:
-
Kenneth Feng authored
update driver-if interface for smu 14.0.2/3 Signed-off-by:
Kenneth Feng <kenneth.feng@amd.com> Reviewed-by:
-
Rodrigo Siqueira authored
Acked-by:
-
Dillon Varone authored
Need to select DTBCLK and DPREFCLK as DTBCLK_p source according to hardware guidance. Reviewed-by:
-
Harikrishna Revalla authored
[why] Cleaning up the code refactor requires hubbub to be in its own component. [how] Move all DCN401 files under newly created hubbub folder and fixing the makefiles. Reviewed-by:
Harikrishna Revalla <Harikrishna.Revalla@amd.com> Tested-by:
-
Samson Tam authored
[Why] Enable adaptive scaler support for DCN401 [How] - Enable build flag for SPL - Set prefer_easf flag to true - Apply light linear scaling policy based on transfer function and pixel format. Choose between linear or non-linear scaling - Set matrix_mode based on pixel format - Disable ring estimator - Add missing EASF register defines, masks, and writes - Disable EASF if scale ratio or number of taps is unsupported and when bypassing the scaler - Add debug flags and registry keys for debugging SPL and EASF - Add support for Visual Confirm with EASF Reviewed-by:
Samson Tam <Samson.Tam@amd.com> Tested-by:
-
Alex Hung authored
This fixes indentations and adjust spaces for better readability and code styles. Reviewed-by:
Alex Hung <alex.hung@amd.com> Tested-by:
-
Chris Park authored
[Why] DVI is TMDS signal like HDMI but without audio. Current signal check does not correctly reflect DVI clock programming. [How] Define a new signal check for TMDS that includes DVI to HDMI TMDS programming. Reviewed-by:
-
Alex Hung authored
The comparisons intend to be DCN401 inclusive, and fix it by adding equal signs. Reviewed-by:
-
Wenjing Liu authored
The functions are missing. These two functions are required to support MST. Reviewed-by:
-
Rodrigo Siqueira authored
The function that commits planes calls the same set of functions twice, and in the case of the FAMs utilization, it is not desired to call the dmub, hwss_build and hwss_execute. This commit just removes the unnecessary calls to those functions. Acked-by:
-
Wenjing Liu authored
Move dsc functions from dc.c to dc_dsc.c. Co-developed-by:
-
- 17 May, 2024 8 commits
-
-
Yang Wang authored
fix ACA no query result after gpu reset. Signed-off-by:
-
Lijo Lazar authored
Replace the legacy interface with amdgpu_dpm_set_pm_policy to set XGMI PLPD mode. Also, xgmi_plpd_policy sysfs node is not used by any client. Remove that as well. Signed-off-by:
-
Lijo Lazar authored
On arcturus, allow changing xgmi plpd policy through 'pm_policy/xgmi_plpd' sysfs interface. Signed-off-by:
-
Yang Wang authored
modify the lock type to 'spinlock' to avoid schedule issue in interrupt context. Signed-off-by:
-
Lijo Lazar authored
On aldebaran, allow changing xgmi plpd policy through 'pm_policy/xgmi_plpd' sysfs interface. Signed-off-by:
-
Lijo Lazar authored
On SOCs with SMU v13.0.6, allow changing xgmi plpd policy through 'pm_policy/xgmi_plpd' sysfs interface. Signed-off-by:
-
Lijo Lazar authored
Add support to set XGMI PLPD policy levels through 'pm_policy/xgmi_plpd' sysfs node. Signed-off-by:
-
Lijo Lazar authored
Add support to select pstate policy in SOCs with SMUv13.0.6 Signed-off-by:
-
