drm/amd/display: Enable double buffer for OTG_BLANK

[Why]
Currently if seamless boot is enabled, we will skip double buffer enable
for OTG_BLANK. However, we need the double buffer enable in order to
block global sync signals when OTG becomes blanked (for PSR). Blocking
global sync signals prevent pipe from requesting data.

[How]
Move tg_init before seamless boot check.

Signed-off-by: Alvin Lee <alvin.lee2@amd.com>
Reviewed-by: Jun Lei <Jun.Lei@amd.com>
Acked-by: Harry Wentland <harry.wentland@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Add monitor patch for AUO dpcd issue

[Why]
dpcd cap mismatch in 2200 vs base

[How]
Add monitor patch which using based caps to overwrite 2200

Signed-off-by: Lewis Huang <Lewis.Huang@amd.com>
Reviewed-by: Tony Cheng <Tony.Cheng@amd.com>
Acked-by: Harry Wentland <harry.wentland@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: wait for test pattern after when all pipes are programmed

[why]
Currently we wait for test pattern after each pipe is programmed.  For
ODM combined scenario it will cause test pattern is shown on only half
screen for 1 frame. This is not desirable.

[how]
No wait between odm pipe programming, only wait after all pipes are
programmed.

Signed-off-by: Wenjing Liu <Wenjing.Liu@amd.com>
Reviewed-by: Tony Cheng <Tony.Cheng@amd.com>
Acked-by: Harry Wentland <harry.wentland@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: make PSR static screen entry within 30 ms

[Why]
With different refresh rate panels, the PSR entry/exit time is
different since it is dependent on 2 frame entry time today

[How]
Make static screen num frame entry time to be calculated
such that entry time is within 30 ms instead of fixed num
frames.

Signed-off-by: Anthony Koo <Anthony.Koo@amd.com>
Reviewed-by: Tony Cheng <Tony.Cheng@amd.com>
Acked-by: Aric Cyr <Aric.Cyr@amd.com>
Acked-by: Harry Wentland <harry.wentland@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: 3.2.67

Signed-off-by: Aric Cyr <aric.cyr@amd.com>
Reviewed-by: Aric Cyr <Aric.Cyr@amd.com>
Acked-by: Harry Wentland <harry.wentland@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: remove psr state condition when psr exit case

[Why]
DMCU need to wait a vblank to handle psr enter command. When psr exit
coming before vblank, the psr exit command will be skip because current
psr state is disable.

[How]
remove psr state condition when psr exit case

Signed-off-by: Lewis Huang <Lewis.Huang@amd.com>
Reviewed-by: Tony Cheng <Tony.Cheng@amd.com>
Acked-by: Harry Wentland <harry.wentland@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: implement fw-driver interface for abm 2.4

[Why]
IRAM definition needed for versions of DMCU containing ABM 2.4

[How]
Pass ABM 2.3 IRAM definition, which is compatible with ABM 2.4, to DMCU
when ABM 2.4 FW is detected

Signed-off-by: Josip Pavic <Josip.Pavic@amd.com>
Reviewed-by: Aric Cyr <Aric.Cyr@amd.com>
Acked-by: Harry Wentland <harry.wentland@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Add default switch case for DCC

Signed-off-by: Chris Park <Chris.Park@amd.com>
Reviewed-by: Aric Cyr <Aric.Cyr@amd.com>
Reviewed-by: Jordan Lazare <Jordan.Lazare@amd.com>
Acked-by: Harry Wentland <harry.wentland@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Use SMU ClockTable Values for DML Calculations

[Why]
DML Initialization was previously done on dcn21_resource initialization.
This meant that DML soc struct was populated with hardcoded values.

[How]
Move DML initialization to after updating bounding box, to use clock table
values from SMU.

Signed-off-by: Sung Lee <sung.lee@amd.com>
Reviewed-by: Tony Cheng <Tony.Cheng@amd.com>
Acked-by: Harry Wentland <harry.wentland@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: rename _lvp to l_vp

Signed-off-by: Charlene Liu <charlene.liu@amd.com>
Reviewed-by: Charlene Liu <Charlene.Liu@amd.com>
Acked-by: Harry Wentland <harry.wentland@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Add w/a to reset PHY before link training in verify_link_cap

[Why]
PHY will sometimes be in bad state on hotplugging display from USB-C
dongle.

[How]
Add additional calls to disable and then enable PHY before link training
starts during verify_link_cap.

Signed-off-by: George Shen <george.shen@amd.com>
Reviewed-by: Tony Cheng <Tony.Cheng@amd.com>
Acked-by: Harry Wentland <harry.wentland@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Fix DMUB PSR command IDs

[Why]
The DMUB PSR command IDs do not have the correct values.

[How]
Fix the command IDs and cleanup the formatting.

Signed-off-by: Wyatt Wood <wyatt.wood@amd.com>
Reviewed-by: Aric Cyr <Aric.Cyr@amd.com>
Acked-by: Harry Wentland <harry.wentland@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Refactor surface flip programming

Rework surface programming for RN to separate preparing parameters and
register programming.

Signed-off-by: Yongqiang Sun <yongqiang.sun@amd.com>
Reviewed-by: Tony Cheng <Tony.Cheng@amd.com>
Acked-by: Harry Wentland <harry.wentland@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Indirect reg read macro with shift and mask

[Why]
Recent double buffering changes for dcn2 use IX_REG_READ.
However, this macro returns the full register value, with the need to
manually shift and mask it to retrieve field data.

[How]
Create new IX_REG_GET macro that handles shift and mask.
Use this for double buffering reads instead of IX_REG_READ.

Signed-off-by: Noah Abradjian <noah.abradjian@amd.com>
Reviewed-by: Tony Cheng <Tony.Cheng@amd.com>
Acked-by: Harry Wentland <harry.wentland@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: DMCUB FW Changes to support PSR

[Why]
Moving PSR from DMCU to DMCUB.

[How]
Cleanup psr spec files and add PSR hw programming files.
No functionality is included in this change.

Signed-off-by: Wyatt Wood <wyatt.wood@amd.com>
Reviewed-by: Tony Cheng <Tony.Cheng@amd.com>
Acked-by: Harry Wentland <harry.wentland@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: programing surface flip by dmcub.

Programming surface flip addresses via dmcub uC for optimizing the data
flush.

Signed-off-by: Yongqiang Sun <yongqiang.sun@amd.com>
Reviewed-by: Tony Cheng <Tony.Cheng@amd.com>
Acked-by: Harry Wentland <harry.wentland@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Double buffer dcn2 Gamut Remap

[Why]
When rapidly adjusting color temperature, screen tearing was observed.
This was due to overwritten values in gamut remap registers.
This issue was solved for OCSC and ICSC by alternating between "A" and
"B" registers to double buffer the writes.

[How]
Create new set_gamut_remap and program_gamut_remap for dcn20.
Alternate which registers are written to by switching modes each time.
Also fixes ICSC mode reg read to use proper data offset.

Signed-off-by: Noah Abradjian <noah.abradjian@amd.com>
Reviewed-by: Tony Cheng <Tony.Cheng@amd.com>
Acked-by: Harry Wentland <harry.wentland@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Driverside changes to support PSR in DMCUB

[Why]
Moving PSR from DMCU to DMCUB.

[How]
Add driverside PSR changes required to send inbox messages to fw.
These changes are non-functional until the psr structure allocation
is uncommented.

Signed-off-by: Wyatt Wood <wyatt.wood@amd.com>
Reviewed-by: Nicholas Kazlauskas <Nicholas.Kazlauskas@amd.com>
Acked-by: Harry Wentland <harry.wentland@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Fix double buffering in dcn2 ICSC

[Why]
When rapidly adjusting video brightness, screen tearing was observed.
This was due to overwritten values in ICSC registers. In dcn10, this issue had been
fixed by implementing double buffering via alternating ICSC modes.
However, the second register set used in dcn1 doesn't exist in dcn2.

[How]
Create new program_input_csc for dcn20.
Use ICSC_B registers instead of COMA registers as second set.

Signed-off-by: Noah Abradjian <noah.abradjian@amd.com>
Reviewed-by: Tony Cheng <Tony.Cheng@amd.com>
Acked-by: Harry Wentland <harry.wentland@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: 3.2.66

Signed-off-by: Aric Cyr <aric.cyr@amd.com>
Reviewed-by: Aric Cyr <Aric.Cyr@amd.com>
Acked-by: Harry Wentland <harry.wentland@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: expand dml structs

Add more fields to support upcoming dml versions

Signed-off-by: Dmytro Laktyushkin <Dmytro.Laktyushkin@amd.com>
Reviewed-by: Jun Lei <Jun.Lei@amd.com>
Acked-by: Harry Wentland <harry.wentland@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: stop doing unnecessary detection when going to D3

[Why]
Don't want to start HW discovery unless we have lost power,
as doing rediscovery otherwise is both unnecessary and time consuming.
Before this change it takes 40 seconds to go in to suspend, after it
takes 27 seconds

[How]
Accelerated mode gets cleared if we lose power. Only do detection if
this register is cleared

Signed-off-by: Joseph Gravenor <joseph.gravenor@amd.com>
Reviewed-by: Eric Yang <eric.yang2@amd.com>
Acked-by: Harry Wentland <harry.wentland@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: wait for update when setting dpg test pattern

Test pattern should be applied to hardware when exiting set test pattern
function.

Signed-off-by: Wenjing Liu <Wenjing.Liu@amd.com>
Reviewed-by: Nikola Cornij <Nikola.Cornij@amd.com>
Acked-by: Harry Wentland <harry.wentland@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Update HDMI hang w/a to apply to all TMDS signals

[WHY]
48mhz turn off feature does not work on HDMI and DVI, but the feauture
was only blocked on HDMI, this change will apply the same wa on DVI

[HOW]
Apply workaround for all TMDS signal types (HDMI, DVI single/dual link)

Signed-off-by: Michael Strauss <michael.strauss@amd.com>
Reviewed-by: Eric Yang <eric.yang2@amd.com>
Acked-by: Harry Wentland <harry.wentland@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Don't always set pstate true if dummy latency = 0

[Why]
If dummy pstate latency is 0 we should report the actual
pstate support, and not that pstate is always supported.

[How]
Don't always program pstate support true

Signed-off-by: Alvin Lee <alvin.lee2@amd.com>
Reviewed-by: Jun Lei <Jun.Lei@amd.com>
Acked-by: Harry Wentland <harry.wentland@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: store lttpr mode with dpcd

Make sure that lttpr_caps has the mode set to repeater.

Signed-off-by: abdoulaye berthe <abdoulaye.berthe@amd.com>
Reviewed-by: Wenjing Liu <Wenjing.Liu@amd.com>
Acked-by: Harry Wentland <harry.wentland@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Add double buffering to dcn20 OCSC

[Why]
When rapidly adjusting colour properties (e.g. brightness), screen tearing was observed.
This was due to overwritten values in OCSC registers. In dcn10, this issue had been fixed by
implementing double buffering by alternating OCSC modes.

[How]
Alternate which OCSC registers are used by switching modes each time.
This double buffers the CSC writes.

Signed-off-by: Noah Abradjian <noah.abradjian@amd.com>
Reviewed-by: Tony Cheng <Tony.Cheng@amd.com>
Acked-by: Harry Wentland <harry.wentland@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Soft reset DMUIF during DMUB reset

[Why]
We need to ensure that the DMUIF in MMHUBBUB is also in reset so we
aren't generating requests while the DMCUB is in reset.

[How]
Set DMUIF_SOFT_RESET=1 on reset and DMUIF_SOFT_RESET=0 on reset
release.

Signed-off-by: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com>
Reviewed-by: Tony Cheng <Tony.Cheng@amd.com>
Acked-by: Harry Wentland <harry.wentland@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Read inst_fb data back during DMUB loading

[Why]
The inst firmware isn't necessarily fully flushed to framebuffer
memory and the DMCUB can hang due to invalid inst being parsed.

[How]
Like the fix to flush ringbuffer commands before updating the inbox
write pointer we need to read back inst memory before writing the
CW0 registers.

Add a helper to read 64-byte chunks to avoid a large temporary buffer.
Read the remaining leftover bytes if the inst_fb isn't fully aligned.

Signed-off-by: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com>
Reviewed-by: Tony Cheng <Tony.Cheng@amd.com>
Acked-by: Harry Wentland <harry.wentland@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Flush framebuffer data before passing to DMCUB

[Why]
There's a data race that can occur between when we update the
inbox write pointer vs when the memory for the command actually gets
flushed from the map to the framebuffer.

DMCUB can read stale or partially invalid data when this race occurs.

[How]
Before updating the write pointer we can read back all pending commands
to ensure that we stall for the writes to be flushed to framebuffer.

We don't need to worry about choosing HDP vs VM flush with this
mechanism.

Drop the dmub_srv_cmd_submit() while we're updating this to work
correctly since nothing was actually using this API and the caller
should be explicit about the API flow here - by doing this on execute
we can give some extra time for the flush to finish while
preparing other commands.

We should try to avoid writing single commands
because of this overhead.

Signed-off-by: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com>
Reviewed-by: Tony Cheng <Tony.Cheng@amd.com>
Acked-by: Harry Wentland <harry.wentland@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Adding forgotten hubbub func

[why]
While doing seamless boot I made some changes
to dcn2 hubbub functions, missed a link

[how]
link hubbub1 func to hubbub2 usage. It has already
been successfully linked in dcn1 and 3.

Signed-off-by: Martin Leung <martin.leung@amd.com>
Reviewed-by: Alvin Lee <Alvin.Lee2@amd.com>
Acked-by: Anthony Koo <Anthony.Koo@amd.com>
Acked-by: Harry Wentland <harry.wentland@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Disable secondary link for certain monitors

[why]
If the specific monitor supports DSC, the secondary link should be
disabled, and the other way around, too: if either that monitor or
our ASIC doesn't support DSC, the secodary link should be enabled.

[how]
Add a monitor patch and disable secondary link if that monitor
is detected and if ASIC supports DSC, or otherwise enable secondary
link.

Signed-off-by: Nikola Cornij <nikola.cornij@amd.com>
Reviewed-by: Joshua Aberback <Joshua.Aberback@amd.com>
Acked-by: Harry Wentland <harry.wentland@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Fix 300Hz Freesync bug

Needed to reprogram vblank_start in dml properly in order to get the
correct dlg params to program VTG.

Signed-off-by: Alvin Lee <alvin.lee2@amd.com>
Reviewed-by: Martin Leung <Martin.Leung@amd.com>
Acked-by: Harry Wentland <harry.wentland@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: fix chroma vp wa corner case

[Why]
Previous implementation we may have residual chroma address offset
if transition from wa enable -> wa disable.

[How]
Clear address offset cache when viewport updates. Also update the
vp size check condition to account for rotation angle

Signed-off-by: Eric Yang <Eric.Yang2@amd.com>
Reviewed-by: Tony Cheng <Tony.Cheng@amd.com>
Acked-by: Harry Wentland <harry.wentland@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: 3.2.65

Signed-off-by: Aric Cyr <aric.cyr@amd.com>
Reviewed-by: Aric Cyr <Aric.Cyr@amd.com>
Acked-by: Harry Wentland <harry.wentland@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Clear state after exiting fixed active VRR state

[why]
Upon exiting a fixed active VRR state, the state isn't cleared. This
leads to the variable VRR range to be calculated incorrectly.

[how]
Set fixed active state to false when updating vrr params

Signed-off-by: Amanda Liu <amanda.liu@amd.com>
Reviewed-by: Anthony Koo <Anthony.Koo@amd.com>
Acked-by: Harry Wentland <harry.wentland@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/include: Add OCSC registers

Add registers for handling Post Gamma Color Blending (OCSC), which is
useful for conversion from RGB->YUV for HDMI.

Reviewed-by: Leo Li <sunpeng.li@amd.com>
Acked-by: Harry Wentland <harry.wentland@amd.com>
Signed-off-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/scheduler: fix documentation by replacing rq_list with sched_list

This also replaces old artifacts with a correct one in drm_sched_entity_init()
declaration

Signed-off-by: Nirmoy Das <nirmoy.das@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: add arcturus to gpu recovery check code path

support check if dirver should try gpu recovery for
arcturus

Signed-off-by: Hawking Zhang <Hawking.Zhang@amd.com>
Reviewed-by: Guchun Chen <guchun.chen@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: check if driver should try recovery in ras recovery path

To allow the flexibilty for user to disable gpu recovery
in RAS recovery path by module parameter amdgpu_gpu_recovery

Signed-off-by: Hawking Zhang <Hawking.Zhang@amd.com>
Reviewed-by: Guchun Chen <guchun.chen@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/powerplay: a quick fix for the deadlock issue below

NFO: task ocltst:2028 blocked for more than 120 seconds.
     Tainted: G           OE     5.0.0-37-generic #40~18.04.1-Ubuntu
echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message.
cltst          D    0  2028   2026 0x00000000
all Trace:
__schedule+0x2c0/0x870
schedule+0x2c/0x70
schedule_preempt_disabled+0xe/0x10
__mutex_lock.isra.9+0x26d/0x4e0
__mutex_lock_slowpath+0x13/0x20
? __mutex_lock_slowpath+0x13/0x20
mutex_lock+0x2f/0x40
amdgpu_dpm_set_powergating_by_smu+0x64/0xe0 [amdgpu]
gfx_v8_0_enable_gfx_static_mg_power_gating+0x3c/0x70 [amdgpu]
gfx_v8_0_set_powergating_state+0x66/0x260 [amdgpu]
amdgpu_device_ip_set_powergating_state+0x62/0xb0 [amdgpu]
pp_dpm_force_performance_level+0xe7/0x100 [amdgpu]
amdgpu_set_dpm_forced_performance_level+0x129/0x330 [amdgpu]

Fixes: a64c9e15e624 ("drm/amd/powerplay: cleanup the interfaces for powergate setting through SMU")
Signed-off-by: Evan Quan <evan.quan@amd.com>
Reported-by: Rui Teng <Rui.Teng@amd.com>
Acked-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: only set cp active field for kiq queue

The mec ucode will set the CP_HQD_ACTIVE bit while the queue is mapped by
MAP_QUEUES packet. So we only need set cp active field for kiq queue.

Signed-off-by: Huang Rui <ray.huang@amd.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdkfd: Add a message when SW scheduler is used

SW scheduler is previously called non HW scheduler, or non HWS. This
message is useful when triaging issues from dmesg.

Signed-off-by: Yong Zhao <Yong.Zhao@amd.com>
Acked-by: Huang Rui <ray.huang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu/pm: clean up return types

count is size_t so don't use negative values.

Reviewed-by: Evan Quan <evan.quan@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/scheduler: improve job distribution with multiple queues

This patch uses score based logic to select a new rq for better
loadbalance between multiple rq/scheds instead of num_jobs.

Below are test results after running amdgpu_test from mesa drm

Before this patch:

sched_name     num of many times it got scheduled
=========      ==================================
sdma0          314
sdma1          32
comp_1.0.0     56
comp_1.0.1     0
comp_1.1.0     0
comp_1.1.1     0
comp_1.2.0     0
comp_1.2.1     0
comp_1.3.0     0
comp_1.3.1     0
After this patch:

sched_name     num of many times it got scheduled
=========      ==================================
sdma0          216
sdma1          185
comp_1.0.0     39
comp_1.0.1     9
comp_1.1.0     12
comp_1.1.1     0
comp_1.2.0     12
comp_1.2.1     0
comp_1.3.0     12
comp_1.3.1     0

Signed-off-by: Nirmoy Das <nirmoy.das@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu/vcn2.5: implement indirect DPG SRAM mode

Implement indirect DPG SRAM mode for vcn2.5

Signed-off-by: James Zhu <James.Zhu@amd.com>
Reviewed-by: Leo Liu <leo.liu@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu/vcn2.5: add dpg pause mode

Add dpg pause mode support for vcn2.5

Signed-off-by: James Zhu <James.Zhu@amd.com>
Reviewed-by: Leo Liu <leo.liu@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu/vcn2.5: add DPG mode start and stop

Add DPG mode start and stop functions for vcn2.5

v2: Correct firmware ucode index in vcn_v2_5_mc_resume_dpg_mode

Signed-off-by: James Zhu <James.Zhu@amd.com>
Reviewed-by: Leo Liu <leo.liu@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu/vcn: move macro from vcn2.0 to share amdgpu_vcn (v2)

Move macro from vcn2.0 to amdgpu_vcn to share with vcn2.5

v2: squash in macro fix

Signed-off-by: James Zhu <James.Zhu@amd.com>
Reviewed-by: Leo Liu <leo.liu@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu/vcn: support multiple instance direct SRAM read and write (v2)

Add multiple instance direct SRAM read and write support for vcn2.5

v2: squash in indexing fix

Signed-off-by: James Zhu <James.Zhu@amd.com>
Reviewed-by: Leo Liu <leo.liu@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu/vcn: support multiple-instance dpg pause mode

Add multiple-instance dpg pause mode support for VCN2.5

Signed-off-by: James Zhu <James.Zhu@amd.com>
Reviewed-by: Leo Liu <leo.liu@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu/sriov skip the update of SMU_TABLE_ACTIVITY_MONITOR_COEFF

There's no need to dump ACTIVITY_MONITOR_COEFF under VF.
Therefore, Skip the update of SMU_TABLE_ACTIVITY_MONITOR_COEFF
under SRIOV VF.

Signed-off-by: Jack Zhang <Jack.Zhang1@amd.com>
Reviewed-by: Evan Quan <evan.quan@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: fix modprobe failure of the secondary GPU when GDDR6 training enabled(V5)

[why]
In dual GPUs scenario, stolen_size is assigned to zero on the secondary GPU,
since there is no pre-OS console using that memory. Then the bottom region of
VRAM was allocated as GTT, unfortunately a small region of bottom VRAM was
encroached by UMC firmware during GDDR6 BIST training, this cause page fault.

[how]
Forcing stolen_size to 3MB, then the bottom region of VRAM was
allocated as stolen memory, GTT corruption avoid.

Reviewed-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Feifei Xu <Feifei.Xu@amd.com>
Signed-off-by: Tianci.Yin <tianci.yin@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu/gfx10: update gfx golden settings for navi14

remove registers: mmSPI_CONFIG_CNTL
add registers: mmSPI_CONFIG_CNTL_1

Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Tianci.Yin <tianci.yin@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu/gfx10: update gfx golden settings

remove registers: mmSPI_CONFIG_CNTL
add registers: mmSPI_CONFIG_CNTL_1

Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Tianci.Yin <tianci.yin@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: check rlc_g firmware pointer is valid before using it

In SRIOV, rlc_g firmware is loaded by host, guest driver won't load it which will
cause the rlc_fw pointer is null

Signed-off-by: shaoyunl <shaoyun.liu@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: drop amdgpu_job.owner

Entirely unused.

Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: error out on entity with no run queue

Disabled HW IP's entity initialized with NULL rq. We should not
process any submit request from userspace for a disabled HW IP.

Signed-off-by: Nirmoy Das <nirmoy.das@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdkfd: use map_queues for hiq on gfx v10 as well

To align with gfx v9, we use the map_queues packet to load hiq MQD.

Signed-off-by: Huang Rui <ray.huang@amd.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdkfd: use kiq to load the mqd of hiq queue for gfx v9 (v6)

There is an issue that CP will check the HIQ queue to be configured and mapped
with KIQ ring, otherwise, it will be unable to read back the secure buffer while
the gfxoff is enabled even with trusted IP blocks.

v1 -> v2:
- Fix to remove surplus set_resources packets.
- Fill the whole configuration in MQD.
- Change the author as Aaron because he addressed the key point of this issue.
- Add kiq ring lock.

v2 -> v3:
- Free the lock while in error return case.
- Remove the programming only needed by the queue is unmapped.

v3 -> v4:
- Remove doorbell programming because it's used for restarting queue.
- Remove CP scheduler programming because map_queue packet will handle this.

v4 -> v5:
- Remove cp_hqd_active because mec ucode will enable it while use map_queues.
- Revise goto out_unlock.
- Correct the right doorbell offset for HIQ that kfd driver assigned in the
  packet.

v5 -> v6:
- Merge Arcturus fix into this patch because it will get oops in Arcturus
  platform.

Reported-by: Lisa Saturday <Lisa.Saturday@amd.com>
Signed-off-by: Aaron Liu <aaron.liu@amd.com>
Signed-off-by: Huang Rui <ray.huang@amd.com>
Reviewed-and-Tested-by: Aaron Liu <aaron.liu@amd.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: flush TLB functions removal from kfd2kgd interface

[Why]
kfd2kgd interface will be deprecated. This removal only covers TLB
invalidation for now. They have been replaced in amdgpu_amdkfd API.

[How]
TLB invalidate functions removed from the different amdkfd_gfx_v*
versions.

Signed-off-by: Alex Sierra <alex.sierra@amd.com>
Acked-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: GPU TLB flush API moved to amdgpu_amdkfd

[Why]
TLB flush method has been deprecated using kfd2kgd interface.
This implementation is now on the amdgpu_amdkfd API.

[How]
TLB flush functions now implemented in amdgpu_amdkfd.

Signed-off-by: Alex Sierra <alex.sierra@amd.com>
Acked-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: export function to flush TLB via pasid

This can be used directly from amdgpu and amdkfd to invalidate
TLB through pasid.
It supports gmc v7, v8, v9 and v10.

Signed-off-by: Alex Sierra <alex.sierra@amd.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: replace kcq enable/disable functions on gfx_v9

[Why]
There are HW-indpendent functions that enables and disables kcq. These functions use
the kiq_pm4_funcs implementation.

[How]
Local kcq enable and disable functions removed and replace it by the generic kcq
enable under amdgpu_gfx

Signed-off-by: Alex Sierra <alex.sierra@amd.com>
Acked-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: implement tlbs invalidate on gfx9 gfx10

tlbs invalidate pointer function added to kiq_pm4_funcs struct.
This way, tlb flush can be done through kiq member.
TLBs invalidatation implemented for gfx9 and gfx10.

Signed-off-by: Alex Sierra <alex.sierra@amd.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: kiq pm4 function implementation for gfx_v9

Functions implemented from kiq_pm4_funcs struct members
for gfx_v9 version.

Signed-off-by: Alex Sierra <alex.sierra@amd.com>
Acked-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: Avoid reclaim fs while eviction lock

[Why]
Avoid reclaim filesystem while eviction lock is held called from
MMU notifier.

[How]
Setting PF_MEMALLOC_NOFS flags while eviction mutex is locked.
Using memalloc_nofs_save / memalloc_nofs_restore API.

Signed-off-by: Alex Sierra <alex.sierra@amd.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu/debugfs: properly handle runtime pm

If driver debugfs files are accessed, power up the GPU
when necessary.

Reviewed-by: Evan Quan <evan.quan@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu/pm: properly handle runtime pm

If power management sysfs or debugfs files are accessed,
power up the GPU when necessary.

Reviewed-by: Evan Quan <evan.quan@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: add header file for macro SZ_1M

Fixes: 4dee6e4ca50a ("drm/amdgpu: use linux size macro to simplify ONE_Kib & One_Mib")
Signed-off-by: Flora Cui <flora.cui@amd.com>
Reviewed-by: Kevin Wang <kevin1.wang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: remove unnecessary conversion to bool

The conversion to bool is not needed, remove it.

Reviewed-by: Harry Wentland <harry.wentland@amd.com>
Signed-off-by: Chen Zhou <chenzhou10@huawei.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Reorder detect_edp_sink_caps before link settings read.

read_current_link_settings_on_detect() on eDP 1.4+ may use the
edp_supported_link_rates table which is set up by
detect_edp_sink_caps(), so that function needs to be called first.

Reviewed-by: Harry Wentland <harry.wentland@amd.com>
Signed-off-by: Mario Kleiner <mario.kleiner.de@gmail.com>
Cc: Martin Leung <martin.leung@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu/psp: declare navi1x ta firmware

So that it gets included in the initrd.  At the moment
this is optional firmware that contains support for HDCP.

Reviewed-by: Evan Quan <evan.quan@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: Match TC hash settings to DF settings (v2)

On Arcturus, data fabric hashing is set by the VBIOS, and
affects which addresses map to which memory channels. The
gfx core's caches also need to know this mapping, but the
hash settings for these these caches is set by the driver.

This change queries the DF to understand how the VBIOS
configured DF, then matches the TC hash configuration bits
to do the same thing.

v2: squash in warning fix

Signed-off-by: Joseph Greathouse <Joseph.Greathouse@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: add defines for DF and TCP Hashing

On Arcturus, we need TC channel hashing, which is set by the
driver, to match DF hashing, which is set by VBIOS. To match
these, we plan to query the DF information and then properly
set the TC configuration bits to match them.

This patch adds the required fields to register definitions
in preparation for a future patch which will use them.

Signed-off-by: Joseph Greathouse <Joseph.Greathouse@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: Create generic DF struct in adev

The only data fabric information the adev struct currently
contains is a function pointer table. In the near future,
we will be adding some cached DF information into adev. As
such, this patch creates a new amdgpu_df struct for adev.
Right now, it only containst the old function pointer table,
but new stuff will be added soon.

Signed-off-by: Joseph Greathouse <Joseph.Greathouse@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: add Pollock IDs, fix Pollock & Dali clk mgr construct

[WHY]
Only a single voltage level should be available to Pollock (min level)
Pollock & Dali get misidentified as Renoir, use wrong clk mgr constructor

[HOW]
Add provided Pollock IDs to ASIC Rev. ID list.
Create new Pollock ASIC RID check, fix RV2 & Dali ASIC checks.
Check RID and set max voltage level to 0 if Pollock is detected.
Work around broken ASICREV_IS_RENOIR, IS_RAVEN2, etc. checks by
performing Dali/Pollock checks before they can be misidentified as RN.

Signed-off-by: Michael Strauss <michael.strauss@amd.com>
Signed-off-by: Bhawanpreet Lakha <Bhawanpreet.Lakha@amd.com>
Reviewed-by: Huang Rui <ray.huang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: preserve RSMU UMC index mode state

between UMC RAS err register access restore previous RSMU UMC index mode state

Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: John Clements <john.clements@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: disable XGMI TA unload for arcturus

in event of GPU reset, XGMI TA unload causes unrecoverable GPU hang

Acked-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: John Clements <john.clements@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: update goldensetting for renoir

Update mmSDMA0_UTCL1_WATERMK golden setting for renoir.

Signed-off-by: Aaron Liu <aaron.liu@amd.com>
Reviewed-by: Huang Rui <ray.huang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/powerplay: update SMU12_DRIVER_IF_VERSION to 11

This patch updates SMU12_DRIVER_IF_VERSION to 11.

Signed-off-by: Aaron Liu <aaron.liu@amd.com>
Reviewed-by: Huang Rui <ray.huang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu/gmc10: free stolen memory in late_init

We don't need to store the pre-OS console memory after
the driver has loaded so free it.

Reviewed-by: Huang Rui <ray.huang@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu/gmc10: remove dead code

Leftover from bring up.  We look up the actual pre-OS memory usage
value later in the same function.

Reviewed-by: Huang Rui <ray.huang@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu/display: set gpu vm flag for renoir

It won't get used unless the driver allows the gtt domain for
display buffers which is controlled elsewhere.

Reviewed-by: Harry Wentland <harry.wentland@amd.com>
Acked-by: Huang Rui <ray.huang@amd.com>
Acked-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: enable S/G display on PCO and RV2 (v2)

It should work on all Raven variants, but some users have
reported issues with original Raven with IOMMU enabled.
So far there have been no issues observed with PCO or RV2.

v2: split out the dm init and domain changes into separate
    patches.

Acked-by: Harry Wentland <harry.wentland@amd.com>
Acked-by: Huang Rui <ray.huang@amd.com>
Acked-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu/display: set gpu vm flag for all asics which support it

It won't get used unless the driver allows the gtt domain for
display buffers which is controlled elsewhere.

Acked-by: Huang Rui <ray.huang@amd.com>
Acked-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Harry Wentland <harry.wentland@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu/gfx9: remove unused sdma headers

All of the sdma stuff these were used for moves to
the sdma code, so remove them.

Reviewed-by: Huang Rui <ray.huang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/amdgpu: add missing umc_6_1_2_sh_mask.h header file (v2)

(v2): Fix preprocessor tag

Signed-off-by: Tom St Denis <tom.stdenis@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: check sdma ras funcs pointer before accessing

sdma ras funcs are not supported by ASIC prior
to vega20

Signed-off-by: Hawking Zhang <Hawking.Zhang@amd.com>
Reviewed-by: Le Ma <Le.Ma@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: calculate MCUMC_ADDRT0 per asic's UMC offset

Hardcoded offset is not friendly. And another benifit of this
patch is to keep read and write access to this register be
consistent with other similar UMC regsiters  in this file.

Signed-off-by: Guchun Chen <guchun.chen@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: add MCUMC_ADDRT0 offset to ip header file

Both are needed on vega20 and arcturus chip.

Signed-off-by: Guchun Chen <guchun.chen@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu/powerplay: fix warning in smu_v11_0.c

Cast to make min() happy.  The values are well within
range.

Reviewed-by: Evan Quan <evan.quan@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/powerplay: sw ctf for arcturus

change the sw ctf setting to smu_v11_0_set_thermal_range()
since software_shutdown_temp shares the same definition and
name in all the smu11 project.

Signed-off-by: Kenneth Feng <kenneth.feng@amd.com>
Reviewed-by: Evan Quan <evan.quan@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu/sriov: workaround on rev_id for Navi12 under sriov

guest vm gets 0xffffffff when reading RCC_DEV0_EPF0_STRAP0,
as a consequence, the rev_id and external_rev_id are wrong.

workaround it by hardcoding the rev_id to 0, which is the default value.

v2. add comment in the code

Signed-off-by: Tiecheng Zhou <Tiecheng.Zhou@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Don't disable DP PHY when link loss happens

[Why]
There is a use case that link loss happens accidentally,
and we need to recover that link loss as soon as possible.
Under this circumstance, we will perform link training,
and try to recover the link that's just lost.

However, if link PHY is disabled before link training
happens, then DP display will never come back again.

Also, please note that dropping this disable_phy function
call won't break USB-C hotplug functionality.
(This line of code was firstly introduced associated with
a patch to fix USB-C hotplug issue)

[How]
Don't disable DP transmitter and its encoder before link
training happens, even if link loss is detected.

Signed-off-by: Zhan Liu <zhan.liu@amd.com>
Reviewed-by: Harry Wentland <harry.wentland@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: read sdma edc counter to clear the counters

SDMA edc counter registers were added in gfx edc counters
array. When querying gfx error counter in that array, there
is no way to differentiate sdma instance number for different
asic and then results to NULL pointer access when trying to
read sdma register base address for instances greater
than 2 on Vega20.
In addition, this also results to wrong gfx error counters
since it actually added sdma edc counters.
Therefore, sdma edc counter registers should be separated
from gfx edc counter regsiter array and only get initialized
when driver tries to enable sdma ras.

Signed-off-by: Hawking Zhang <Hawking.Zhang@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: add ras_late_init and ras_fini for sdma v4

move ras_late_init and ras_fini to sdma_ras_funcs table

Signed-off-by: Hawking Zhang <Hawking.Zhang@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: support error reporting for sdma ip block

invoke sdma query_ras_error_count to get sdma single
bit error count

Signed-off-by: Hawking Zhang <Hawking.Zhang@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: add query_ras_error_count function for sdma v4

query_ras_error_count function will be invoked to query
single bit error count detected in sdma ip block

Signed-off-by: Hawking Zhang <Hawking.Zhang@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: enable VCN2.5 IP block for Arcturus

With default PSP FW loading

Signed-off-by: Leo Liu <leo.liu@amd.com>
Reviewed-by: James Zhu <James.Zhu@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>