- 02 Dec, 2021 1 commit
-
-
Stanley.Yang authored
this is a workaround due to get ecc info failed during gpu recovery [ 700.236122] amdgpu 0000:09:00.0: amdgpu: Failed to export SMU ecc table! [ 700.236128] amdgpu 0000:09:00.0: amdgpu: GPU reset begin! [ 704.331171] amdgpu: qcm fence wait loop timeout expired [ 704.331194] amdgpu: The cp might be in an unrecoverable state due to an unsuccessful queues preemption [ 704.332445] amdgpu 0000:09:00.0: amdgpu: GPU reset begin! [ 704.332448] amdgpu 0000:09:00.0: amdgpu: Bailing on TDR for s_job:ffffffffffffffff, as another already in progress [ 704.332456] amdgpu: Pasid 0x8000 destroy queue 0 failed, ret -62 [ 710.360924] amdgpu 0000:09:00.0: amdgpu: SMU: I'm not done with your previous command: SMN_C2PMSG_66:0x00000013 SMN_C2PMSG_82:0x00000007 [ 710.360964] amdgpu 0000:09:00.0: amdgpu: Failed to disable smu features. [ 710.361002] amdgpu 0000:09:00.0: amdgpu: Fail to disable dpm features! [ 710.361014] [drm:amdgpu_device_ip_suspend_phase2 [amdgpu]] *ERROR* suspend of IP block <smu> failed -62 Signed-off-by:
Stanley.Yang <Stanley.Yang@amd.com> Reviewed-by:
Hawking Zhang <Hawking.Zhang@amd.com> Signed-off-by:
Alex Deucher <alexander.deucher@amd.com>
-
- 01 Dec, 2021 39 commits
-
-
Yann Dirson authored
amdgpu_ucode_get_load_type() does not interpret this parameter as documented. It is ignored for many ASIC types (which presumably only support one load_type), and when not ignored it is only used to force direct loading instead of PSP loading. SMU loading is only available for ASICs for which the parameter is ignored. Signed-off-by:
Yann Dirson <ydirson@free.fr> Signed-off-by:
-
Philip Yang authored
Refactor userptr and pin_bo path to make it less confusing, move err_pin_bo label up to remove mem from process_info kfd_bo_list. Signed-off-by:
Philip Yang <Philip.Yang@amd.com> Reviewed-by:
Felix Kuehling <Felix.Kuehling@amd.com> Signed-off-by:
-
Philip Yang authored
process_info->lock is used to protect kfd_bo_list, vm_list_head, n_vms and userptr valid/inval list, svm_range_restore_work and svm_range_set_attr don't access those, so do not need to take process_info lock. This will avoid potential circular locking issue. Signed-off-by:
-
Graham Sider authored
With device_info initialization being handled in kfd_device_info_init, these structs may be removed. Also add comments to help matching IP versions to asic names. Signed-off-by:
Graham Sider <Graham.Sider@amd.com> Reviewed-by:
-
Graham Sider authored
Initializes kfd->device_info given either asic_type (enum) if GFX version is less than GFX9, or GC IP version if greater. Also takes in vf and the target compiler gfx version. Uses SDMA version to determine num_sdma_queues_per_engine. Convert device_info to a non-pointer member of kfd, change references accordingly. Change unsupported asic condition to only probe f2g, move device_info initialization post-switch. Signed-off-by:
-
Graham Sider authored
device_info->asic_name and amdgpu_asic_name[adev->asic_type] both provide asic name strings, with the only difference being casing. Remove asic_name from device_info and replace sysfs entry with lowercase amdgpu_asic_name[]. Ensures string is null-terminated so that this doesn't break if dev->node_props.name ever gets set anywhere else. Signed-off-by:
-
shaoyunl authored
This change revert previous commits: 9f4f2c1a ("drm/amd/amdgpu: fix the kfd pre_reset sequence in sriov") 271fd38c ("drm/amdgpu: move kfd post_reset out of reset_sriov function") This change moves the amdgpu_amdkfd_pre_reset to an earlier place in amdgpu_device_reset_sriov, presumably to address the sequence issue that the first patch was originally meant to fix. Some register access(GRBM_GFX_CNTL) only be allowed on full access mode. Move kfd_pre_reset and kfd_post_reset back inside reset_sriov function. Fixes: 9f4f2c1a ("drm/amd/amdgpu: fix the kfd pre_reset sequence in sriov") Fixes: 271fd38c ("drm/amdgpu: move kfd post_reset out of reset_sriov function") Signed-off-by:
shaoyunl <shaoyun.liu@amd.com> Reviewed-by:
-
Perry Yuan authored
[Why] IGT bypass test will set crc source as DPRX,and display DM didn`t check connection type, it run the test on the HDMI connector ,then the kernel will be crashed because aux->transfer is set null for HDMI connection. This patch will skip the invalid connection test and fix kernel crash issue. [How] Check the connector type while setting the pipe crc source as DPRX or auto,if the type is not DP or eDP, the crtc crc source will not be set and report error code to IGT test,IGT will show the this subtest as no valid crtc/connector combinations found. 116.779714] [IGT] amd_bypass: starting subtest 8bpc-bypass-mode [ 117.730996] BUG: kernel NULL pointer dereference, address: 0000000000000000 [ 117.731001] #PF: supervisor instruction fetch in kernel mode [ 117.731003] #PF: error_code(0x0010) - not-present page [ 117.731004] PGD 0 P4D 0 [ 117.731006] Oops: 0010 [#1] SMP NOPTI [ 117.731009] CPU: 11 PID: 2428 Comm: amd_bypass Tainted: G OE 5.11.0-34-generic #36~20.04.1-Ubuntu [ 117.731011] Hardware name: AMD CZN/, BIOS AB.FD 09/07/2021 [ 117.731012] RIP: 0010:0x0 [ 117.731015] Code: Unable to access opcode bytes at RIP 0xffffffffffffffd6. [ 117.731016] RSP: 0018:ffffa8d64225bab8 EFLAGS: 00010246 [ 117.731017] RAX: 0000000000000000 RBX: 0000000000000020 RCX: ffffa8d64225bb5e [ 117.731018] RDX: ffff93151d921880 RSI: ffffa8d64225bac8 RDI: ffff931511a1a9d8 [ 117.731022] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 [ 117.731023] CR2: ffffffffffffffd6 CR3: 000000010d5a4000 CR4: 0000000000750ee0 [ 117.731023] PKRU: 55555554 [ 117.731024] Call Trace: [ 117.731027] drm_dp_dpcd_access+0x72/0x110 [drm_kms_helper] [ 117.731036] drm_dp_dpcd_read+0xb7/0xf0 [drm_kms_helper] [ 117.731040] drm_dp_start_crc+0x38/0xb0 [drm_kms_helper] [ 117.731047] amdgpu_dm_crtc_set_crc_source+0x1ae/0x3e0 [amdgpu] [ 117.731149] crtc_crc_open+0x174/0x220 [drm] [ 117.731162] full_proxy_open+0x168/0x1f0 [ 117.731165] ? open_proxy_open+0x100/0x100 BugLink: https://gitlab.freedesktop.org/drm/amd/-/issues/1546Reviewed-by:
Harry Wentland <harry.wentland@amd.com> Reviewed-by:
Rodrigo Siqueira <Rodrigo.Siqueira@amd.com> Signed-off-by:
Perry Yuan <Perry.Yuan@amd.com> Signed-off-by:
-
Philip Yang authored
drm_gem_object_put calls release_notify callback to free the mem structure and unreserve_mem_limit, move it down after the last access of mem and make it conditional call. Signed-off-by:
-
Philip Yang authored
To silence the following Smatch static checker warning: drivers/gpu/drm/amd/amdgpu/../amdkfd/kfd_svm.c:2615 svm_range_restore_pages() warn: missing error code here? 'get_task_mm()' failed. 'r' = '0' Signed-off-by:
Dan Carpenter <dan.carpenter@oracle.com> Reviewed-by:
-
Aric Cyr authored
This version brings along the following: - FW promotion to 0.0.95 - DSC fixes for supported Docks - Fixes eDP display issue - Vendor LTTR workarounds - Fixes Tiled display audio issue Signed-off-by:
Aric Cyr <aric.cyr@amd.com> Acked-by:
Bhawanpreet.Lakha <Bhawanpreet.Lakha@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:
-
George Shen authored
[Why] Certain display configurations require an extra delay before reading lane status with certain LTTPR. [How] Add temporary workaround to force AUX RD interval to 16ms for CR and EQ. Needs to be refactored later. Reviewed-by:
Jun Lei <Jun.Lei@amd.com> Acked-by:
George Shen <George.Shen@amd.com> Tested-by:
-
Michael Strauss authored
[WHY] Allow changing DET size with debug flag for testing purposes Reviewed-by:
Nicholas Kazlauskas <Nicholas.Kazlauskas@amd.com> Acked-by:
Michael Strauss <michael.strauss@amd.com> Tested-by:
-
George Shen authored
[Why] Vendor specific link rate toggle workaround should not be applied outside of link training scenario, such as when outputting PHY test pattern. [How] Add flag to indicate when link training is occuring. Make workaround a no-op when flag is not set. Reviewed-by:
-
George Shen authored
[Why] Certain LTTPR require special workarounds in order to comply with DP specifications. [How] Implement vendor specific sequences via DPCD writes to vendor-specific LTTPR registers. Reviewed-by:
Wenjing Liu <Wenjing.Liu@amd.com> Acked-by:
-
Mikita Lipski authored
[why] Adding a debugfs to show PSR capability of the panel and the driver. To use: cat /sys/kernel/debug/dri/0/eDP-X/psr_capability Expected output: Sink support: no/yes Driver support: no/yes [how] Adding psr_capability to eDP connectors. Reviewed-by:
Mikita Lipski <mikita.lipski@amd.com> Tested-by:
-
Brandon Syu authored
[WHY] Driver use primary eDP link stream as the condition to enable fastboot or not. In dual eDP case, set second screen only, this would cause slave eDP timing to be used on primary one. [HOW] Add a check for unmatched eDP link stream. Reviewed-by:
Brandon Syu <Brandon.Syu@amd.com> Tested-by:
-
Jimmy Kizito authored
[Why] Certain USB4 docks do not seem to be able to handle disabling DSC once it has been enabled on an MST stream. This can result in blank displays. [How] As a work around, always enable DSC on docks exhibiting this issue. The flag to indicate the use of DSC for MST streams on a USB4 dock is set during detection of the dock and only cleared when the USB4 dock is disconnected. Reviewed-by:
Jimmy Kizito <Jimmy.Kizito@amd.com> Tested-by:
-
Evgenii Krasnikov authored
[HOW&WHY] Add function to be used for early eDP power on Reviewed-by:
Evgenii Krasnikov <Evgenii.Krasnikov@amd.com> Tested-by:
-
Mustapha Ghaddar authored
[WHY] It seems like after a series of plug/unplugs we end up in a situation where tiled display doesnt support Audio. [HOW] The issue seems to be related to when we check streams changed after an HPD, we should be checking the audio_struct as well to see if any of its values changed. Reviewed-by:
Mustapha Ghaddar <mustapha.ghaddar@amd.com> Tested-by:
-
Meenakshikumar Somasundaram authored
[Why] During otg sync trigger, plane states are used to decide whether the otg is already synchronized or not. There are scenarions when otgs are disabled without plane state getting disabled and in such case the otg is excluded from synchronization. [How] Introduced pipe_idx_syncd in pipe_ctx that tracks each otgs master pipe. When a otg is disabled/enabled, pipe_idx_syncd is reset to itself. On sync trigger, pipe_idx_syncd is checked to decide whether a otg is already synchronized and the otg is further included or excluded from synchronization. Reviewed-by:
meenakshikumar somasundaram <meenakshikumar.somasundaram@amd.com> Tested-by:
-
Shen, George authored
[Why] VS and PE requested by repeater should not persist for the sink. [How] Clear DPCD lane settings after repeater link training finishes. Reviewed-by:
Wesley Chalmers <wesley.chalmers@amd.com> Acked-by:
-
Dale Zhao authored
[Why] Using the hdmi_disable option doesnt disable 6GB bandwidth [How] Add debug.hdmi20_disable flage when checking 6GB enable or not. Reviewed-by:
Chris Park <Chris.Park@amd.com> Reviewed-by:
Dale Zhao <dale.zhao@amd.com> Tested-by:
-
Nicholas Kazlauskas authored
[Why] When trying to lightup two 4k60 non-DSC displays behind a branch device that supports DSC we can't lightup both at once due to bandwidth limitations - each requires 48 VCPI slots but we only have 63. [How] The workaround already exists in the code but is guarded by a CONFIG that cannot be set by the user and shouldn't need to be. Check for specific branch device IDs to device whether to enable the workaround for multiple display scenarios. Reviewed-by:
Hersen Wu <hersenxs.wu@amd.com> Acked-by:
-
Stylon Wang authored
[Why + How] Enable P010 for SDR video applications. Reviewed-by:
Stylon Wang <stylon.wang@amd.com> Tested-by:
-
Fangzhi Zuo authored
Change since v1: add brief description 1. Add hdmi frl pcon support to existing asic family. 2. Determine pcon frl capability based on pcon dpcd. 3. pcon frl is taken into consideration into mode validation. v2: squash in warning fix (Alex) Signed-off-by:
Fangzhi Zuo <Jerry.Zuo@amd.com> Reviewed-by:
-
Fangzhi Zuo authored
1. Parse DSC caps from PCON DPCD 2. Policy determins if DSC is decoded at PCON 3. Enable/disable DSC at PCON Signed-off-by:
-
Lijo Lazar authored
On aldebaran, RLC also controls GFXCLK. Skip halting RLC during GFX IP suspend and keep it running till PMFW disables all DPMs. [ 578.019986] amdgpu 0000:23:00.0: amdgpu: GPU reset begin! [ 583.245566] amdgpu 0000:23:00.0: amdgpu: Failed to disable smu features. [ 583.245621] amdgpu 0000:23:00.0: amdgpu: Fail to disable dpm features! [ 583.245639] [drm:amdgpu_device_ip_suspend_phase2 [amdgpu]] *ERROR* suspend of IP block <smu> failed -62 [ 583.248504] [drm] free PSP TMR buffer Signed-off-by:Lijo Lazar <lijo.lazar@amd.com> Reviewed-by:
-
Lijo Lazar authored
HW_ID_MAX considers HWID of all IPs, far more than what amdgpu uses. amdgpu tracks only the IPs defined by amd_hw_ip_block_type whose max is MAX_HWIP. Signed-off-by:
Guchun Chen <guchun.chen@amd.com> Signed-off-by:
-
Guchun Chen authored
There is no base reg offset or ip_version set for SDMA2 and SDMA3 on SIENNA_CICHLID, so add them. Signed-off-by:
Kevin Wang <kevinyang.wang@amd.com> Signed-off-by:
-
Guchun Chen authored
>> drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c:503:6: warning: no previous prototype for function 'release_psp_cmd_buf' [-Wmissing-prototypes] void release_psp_cmd_buf(struct psp_context *psp) ^ drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c:503:1: note: declare 'static' if the function is not intended to be used outside of this translation unit void release_psp_cmd_buf(struct psp_context *psp) ^ static 1 warning generated. Reported-by:kernel test robot <lkp@intel.com> Signed-off-by:
-
Darren Powell authored
After modifying navi10 in previous commit, extend the changes so all implementations of pp_hwmgr_func->get_power_profile_mode and pptable_funcs->get_power_profile_mode use amdgpu_pp_profile_name == Test == LOGFILE=pp_profile_strings.test.log AMDGPU_PCI_ADDR=`lspci -nn | grep "VGA\|Display" | cut -d " " -f 1` AMDGPU_HWMON=`ls -la /sys/class/hwmon | grep $AMDGPU_PCI_ADDR | awk '{print $9}'` HWMON_DIR=/sys/class/hwmon/${AMDGPU_HWMON} lspci -nn | grep "VGA\|Display" > $LOGFILE FILES="pp_power_profile_mode " for f in $FILES do echo === $f === >> $LOGFILE cat $HWMON_DIR/device/$f >> $LOGFILE done cat $LOGFILE Signed-off-by:Darren Powell <darren.powell@amd.com> Reviewed-by:
-
Darren Powell authored
== Description == All the power profile modes use the same strings (or a subset of) Creating a public array of the strings will allow sharing rather than duplicating for each chip First patch only implements change for navi10, followup with other chips == Changes == Create a declaration of the public array in kgd_pp_interface.h Define the public array in amdgpu_pm.c Modify the implementaiton of navi10_get_power_profile_mode to use new array == Test == LOGFILE=pp_profile_strings.test.log AMDGPU_PCI_ADDR=`lspci -nn | grep "VGA\|Display" | cut -d " " -f 1` AMDGPU_HWMON=`ls -la /sys/class/hwmon | grep $AMDGPU_PCI_ADDR | awk '{print $9}'` HWMON_DIR=/sys/class/hwmon/${AMDGPU_HWMON} lspci -nn | grep "VGA\|Display" > $LOGFILE FILES="pp_power_profile_mode " for f in $FILES do echo === $f === >> $LOGFILE cat $HWMON_DIR/device/$f >> $LOGFILE done cat $LOGFILE Signed-off-by: -
Philip Yang authored
IH ring1 is used to process GPU retry fault, overflow is enabled to drain retry fault because we want receive other interrupts while handling retry fault to recover range. There is no overflow flag set when wptr pass rptr. Use timestamp of rptr and wptr to handle overflow and drain retry fault. If fault timestamp goes backward, the fault is filtered and should not be processed. Drain fault is finished if processed_timestamp is equal to or larger than checkpoint timestamp. Add amdgpu_ih_functions interface decode_iv_ts for different chips to get timestamp from IV entry with different iv size and timestamp offset. amdgpu_ih_decode_iv_ts_helper is used for vega10, vega20, navi10. Signed-off-by:
Christian König <christian.koenig@amd.com> Signed-off-by:
-
Stanley.Yang authored
v2: still need call ras_disable_all_featrures to handle ras initilization failure case. Function amdgpu_device_fini_hw is called before amdgpu_device_fini_sw, so ras ta will unload before send ras disable command, ras dsiable operation must before hw fini. Signed-off-by: -
Lijo Lazar authored
v1: Ideally power gate/ungate requests shouldn't come when smu block is uninitialized. Add a WARN message to check the origins if such a thing ever happens. v2: Use dev_WARN to log device info (Felix/Guchun). Signed-off-by:
-
Flora Cui authored
since vkms support atomic KMS interface Signed-off-by:
Flora Cui <flora.cui@amd.com> Reviewed-by:
Alex Deucher <aleander.deucher@amd.com> Reviewed-by:
-
Flora Cui authored
otherwise adev->mode_info.crtcs[] is NULL Signed-off-by:
-
