drm/amdgpu: Fix a printing message

The printing message "PSP loading VCN firmware" is mis-leading because
people might think driver is loading VCN firmware. Actually when this
message is printed, driver is just preparing some VCN ucode, not loading
VCN firmware yet. The actual VCN firmware loading will be in the PSP block
hw_init. Fix the printing message

Signed-off-by: Oak Zeng <Oak.Zeng@amd.com>
Reviewed-by: Christian Konig <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: Add error message when programing registers fails

Squash in warning fix (Alex)

Signed-off-by: Roy Sun <Roy.Sun@amd.com>
Reviewed-by: Zhou pengju <pengju.zhou@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: Change the imprecise function name

The callback functions are used for SRIOV read/write instead
of just for rlcg read/write

Signed-off-by: Roy Sun <Roy.Sun@amd.com>
Reviewed-by: Zhou pengju <pengju.zhou@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu - Corrected the video codecs array name for yellow carp

Signed-off-by: Veerabadhran Gopalakrishnan <veerabadhran.gopalakrishnan@amd.com>
Reviewed-by: James Zhu <James.Zhu@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdkfd: report pcie bandwidth to the kfd

Similar to xGMI reporting the min/max bandwidth between direct peers, PCIe
will report the min/max bandwidth to the KFD.

Signed-off-by: Jonathan Kim <jonathan.kim@amd.com>
Reviewed-by: Felix Kuehling <felix.kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdkfd: report xgmi bandwidth between direct peers to the kfd

Report the min/max bandwidth in megabytes to the kfd for direct
xgmi connections only.  Indirect peers will report 0 since
indirect route is unknown.

Signed-off-by: Jonathan Kim <jonathan.kim@amd.com>
Reviewed-by: Felix Kuehling <felix.kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: add psp command to get num xgmi links between direct peers

The TA can now be invoked to provide the number of xgmi links connecting
a direct source and destination peer.
Non-direct peers will report zero links.

Signed-off-by: Jonathan Kim <jonathan.kim@amd.com>
Reviewed-by: Felix Kuehling <felix.kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: Fix documentaion for amdgpu_bo_add_to_shadow_list

make htmldocs complaints about parameter for amdgpu_bo_add_to_shadow_list

./drivers/gpu/drm/amd/amdgpu/amdgpu_object.c:739: warning: Excess function parameter 'bo' description in 'amdgpu_bo_add_to_shadow_list'
./drivers/gpu/drm/amd/amdgpu/amdgpu_object.c:739: warning: Function parameter or member 'vmbo' not described in 'amdgpu_bo_add_to_shadow_list'
./drivers/gpu/drm/amd/amdgpu/amdgpu_object.c:739: warning: Excess function parameter 'bo' description in 'amdgpu_bo_add_to_shadow_list'

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

drm/amd/pm: Support board calibration on aldebaran

Add support for board power calibration on Aldebaran.
Board calibration is done after DC offset calibration.

Signed-off-by: Lijo Lazar <lijo.lazar@amd.com>
Reviewed-by: Kevin Wang <kevin1.wang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: change zstate allow msg condition

[Why]
PMFW message which previously thought to only control Z9 controls both
Z9 and Z10. Also HW design team requested that Z9 must only be supported
on eDP due to content protection interop.

[How]
Change zstate support condition to match updated policy

Reviewed-by: Nicholas Kazlauskas <Nicholas.Kazlauskas@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Eric Yang <Eric.Yang2@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: 3.2.145

DC version 3.2.145 brings improvements in multiple areas. In summary, we
highlight:

- Code improvements for passive
- Cursor manipulation enhancements
- Expand debug in some areas
- Fix problems in DML
- Other minor code refactors

Reviewed-by: Anson Jacob <Anson.Jacob@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Aric Cyr <aric.cyr@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: [FW Promotion] Release 0.0.75

- Add reserved bits for future feature development
- Fix issue with mismatch with type const
- Replaced problematic code with old memcpy and casted problematic
  pointers to unsigned char pointers

Reviewed-by: Aric Cyr <aric.cyr@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Anthony Koo <Anthony.Koo@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Refine condition for cursor visibility

[why]
There's a special case where upper plane is not the main plane. If it owns
the cursor, it will be invisible in the majority of the screen.

[How]
The condition for disabling cursor is changed:
- check if upper viewport completely covers current. This was the
previous change that doesn't handle all scenarios with pipe splitting.
- if not, show the cursor only if it's not scaled or no upper pipe.

Reviewed-by: Aric Cyr <aric.cyr@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Krunoslav Kovac <Krunoslav.Kovac@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Populate dtbclk entries for dcn3.02/3.03

[Why]
Populate dtbclk values from bwparams for dcn302, dcn303.

[How]
dtbclk values are fetched from bandwidthparams for all DPM levels and
for DPM levels where smu returns 0, previous level values are reported.

Reviewed-by: Roman Li <Roman.Li@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Bindu Ramamurthy <bindu.r@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: add workaround for riommu invalidation request hang

[Why]
When an riommu invalidation request come at the same time as a pipe is
disabled there can be a case where DCN cannot ACK the request if only
one VMID is setup in the inuse list.

[How]
Setup a second unused VMID will work around the issue.

Reviewed-by: Jun Lei <jun.lei@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Eric Yang <Eric.Yang2@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Line Buffer changes

DCN 3x increased Line buffer size for DCHUB latency hiding, from 4 lines
of 4K resolution lines to 5 lines of 4K resolution lines. All Line
Buffer can be used as extended memory for P State change latency hiding.
The maximum number of lines is increased to 32 lines. Finally,
LB_MEMORY_CONFIG_1 (LB memory piece 1) and LB_MEMORY _CONFIG_2 (LB
memory piece 2) are not affected, no change in size, only 3 pieces is
affected, i.e., when all 3 pieces are used in both LB_MEMORY_CONFIG_0
and LB_MEMORY_CONFIG_3 (for 4:2:0) modes.

Reviewed-by: Jun Lei <jun.lei@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Nevenko Stupar <Nevenko.Stupar@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Remove MALL function from DCN3.1

[why]
DCN31 doesn't have MALL in DMUB so to avoid sending unknown commands to
DMUB just remove the function pointer.

[how]
Remove apply_idle_power_optimizations from function pointers structure
for DCN31

Reviewed-by: Nicholas Kazlauskas <Nicholas.Kazlauskas@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Mikita Lipski <mikita.lipski@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: DCN2X Prefer ODM over bottom pipe to find second pipe

[WHY]
When finding a second pipe for pipe split, currently will look for
bottom pipe in context first to decide the second pipe. This causes
issues in 2 plane to 1 plane transitions like fullscreen video where
bottom pipe no longer exists in the new configuration.

[HOW]
If previous context had an ODM pipe, use that to find the secondary pipe
first before looking at bottom pipe.

Reviewed-by: Dmytro Laktyushkin <Dmytro.Laktyushkin@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: sunglee <sunglee@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Only set default brightness for OLED

[Why]
We used to unconditionally set backlight path as AUX for panels capable
of backlight adjustment via DPCD in set default brightness.

[How]
This should be limited to OLED panel only since we control backlight via
PWM path for SDR mode in LCD HDR panel.

Reviewed-by: Krunoslav Kovac <krunoslav.kovac@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Camille Cho <Camille.Cho@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Update bounding box for DCN3.1

[Why & How]
We're missing a default value for dram_channel_width_bytes in the
DCN3.1 SOC bounding box and we don't currently have the interface in
place to query the actual value from VBIOS.

Put in a hardcoded default until we have the interface in place.

Reviewed-by: Eric Yang <eric.yang2@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Query VCO frequency from register for DCN3.1

[Why]
Hardcoding the VCO frequency isn't correct since we don't own or control
the value.

In the case where the hardcode is also missing we can't lightup display.

[How]
Query from the CLK register instead. Update the DFS frequency to be able
to compute the VCO frequency.

Reviewed-by: Eric Yang <eric.yang2@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Populate socclk entries for dcn3.02/3.03

[Why]
Initialize socclk entries in bandwidth params for dcn302, dcn303.

[How]
Fetch the sockclk values from smu for the DPM levels and for the DPM
levels where smu returns 0, previous level values are reported.

Reviewed-by: Roman Li <Roman.Li@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Bindu Ramamurthy <bindu.r@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Fix max vstartup calculation for modes with borders

[Why]
Vertical and horizontal borders in timings are treated as increasing the
active area - vblank and hblank actually shrink.

Our input into DML does not include these borders so it incorrectly
assumes it has more time than available for vstartup and tmdl
calculations for some modes with borders.

An example of such a timing would be 640x480@72Hz:

h_total: 832
h_border_left: 8
h_addressable: 640
h_border_right: 8
h_front_porch: 16
h_sync_width: 40
v_total: 520
v_border_top: 8
v_addressable: 480
v_border_bottom: 8
v_front_porch: 1
v_sync_width: 3
pix_clk_100hz: 315000

[How]
Include borders as part of destination vactive/hactive.

This change DCN20+ so it has wide impact, but the destination vactive
and hactive are only really used for vstartup calculation anyway.

Most modes do not have vertical or horizontal borders.

Reviewed-by: Dmytro Laktyushkin <Dmytro.Laktyushkin@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Enable eDP ILR on DCN2.1

[WHY]
Enable feature for 21.40

Reviewed-by: Sung Lee <Sung.Lee@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Michael Strauss <michael.strauss@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: 3.2.144

Reviewed-by: Shahin Khayyer <Shahin.Khayyer@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Aric Cyr <aric.cyr@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Fix comparison error in dcn21 DML

[why]
A comparison error made it possible to not iterate through all the
specified prefetch modes.

[how]
Correct "<" to "<="

Reviewed-by: Dmytro Laktyushkin <Dmytro.Laktyushkin@amd.com>
Reviewed-by: Yongqiang Sun <Yongqiang.Sun@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Victor Lu <victorchengchi.lu@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Fixed hardware power down bypass during headless boot

[Why]
During headless boot, DIG may be on which causes HW/SW discrepancies.
To avoid this we power down hardware on boot if DIG is turned on. With
introduction of multiple eDP, hardware power down is being bypassed
under certain conditions.

[How]
Fixed hardware power down bypass, and ensured hardware will power down
if DIG is on and seamless boot is not enabled.

Reviewed-by: Nicholas Kazlauskas <Nicholas.Kazlauskas@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Jake Wang <haonan.wang2@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Add copyright notice to new files

Reviewed-by: Shahin Khayyer <Shahin.Khayyer@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Wesley Chalmers <Wesley.Chalmers@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Reduce delay when sink device not able to ACK 00340h write

[Why]
Theoretically, per DP 1.4a spec, sink device needs to AUX_ACK 00340h
write. However, due to hardware limitation, some sink devices have no
00340h dpcd address at all. This results in sink side fails to reply
ACK, and consequently cause source side keep retrying DPCD write on DPCD
00340h. This results in significant delay when DPCD 00340h write is
triggered (e.g. at S3 resume).

[How]
Check whether sink device could ACK on DPCD 00340h write on boot. If
sink device fails to ACK, then remember that, so we won't write to DPCD
00340h later on.

There will be a drm.debug KMS level message to inform user once a 00340h
DPCD write is skipped on purpose.

Reviewed-by: Nikola Cornij <Nikola.Cornij@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Zhan Liu <zhan.liu@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: add debug print for DCC validation failure

[Why&How]
Print a debug message when dcc validation fails in the display driver.
Most DCC enablement related errors are from userspace. Adding a debug
print in case of a failure from display driver will aid quicker triage.

Reviewed-by: Harry Wentland <Harry.Wentland@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Aurabindo Pillai <aurabindo.pillai@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: 3.2.143

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

drm/amd/display: [FW Promotion] Release 0.0.73

- Add reserved bits for future feature development
- Fix issue with mismatch with type const

Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Anthony Koo <Anthony.Koo@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: remove compbuf size wait

This tends to take miliseconds in certain scenarios and we'd rather not
wait that long. Due to how this interacts with det size update and
locking waiting should not be necessary as compbuf updates before
unlock.

Add a watch for config error instead as that is something we actually do
care about.

Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Dmytro Laktyushkin <Dmytro.Laktyushkin@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Extend dmub_cmd_psr_copy_settings_data struct

Reserve padding bytes for new feature implementation

Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Ian Chen <ian.chen@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: implement workaround for riommu related hang

[Why]
During S4/S5/reboot, sometimes riommu invalidation request arrive too
early, DCN may be unable to respond to the invalidation request
resulting in pstate hang.

[How]
VBIOS will force allow pstate for riommu invalidation and driver will
clear it after powering down display pipes.

Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Eric Yang <Eric.Yang2@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: log additional register state for debug

[Why & How]
Extend existing state collection functions to add some additional
registers useful for debug, and add state collection function for DC
hubbub

Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Josip Pavic <Josip.Pavic@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Prevent Diags from entering S2

[why]
Lowering clocks when entering S2 Idle state causes DMUB to hang with
Diags.

[how]
Do not enter S2 optimization with Diags on dcn301 to prevent DMUB hang.

Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Mikita Lipski <mikita.lipski@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Fix timer_per_pixel unit error

[why]
The units of the time_per_pixel variable were incorrect, this had to be
changed for the code to properly function.

[how]
The change was very straightforward, only required one line of code to
be changed where the calculation was done.

Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Oliver Logush <oliver.logush@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Assume active upper layer owns the HW cursor

[why]
The current logic checks if there's an upper pipe whose viewport
completely covers the current pipe viewport.
This fails in pipe splitting case as you can have layer 1 pipe that
crosses the two layer 0 pipes where it's contained in both, but neither
covers it completely, hence we allow the cursor on both layers.

[How]
Instead of trying to "sum up" rectangles from the higher level pipes
which could leave gaps and would not work generically, we will assume if
there's an upper layer that is active, it will control the HW cursor.

Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Krunoslav Kovac <Krunoslav.Kovac@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: reset dpcd_cap.dpcd_rev for passive dongle.

[why]
currently dc has never reset this dpcd_cap.dpcd_rev.

[how]
ideally  we should reset this before redo detection.
change the passive dongle only for now to reduce the impact.

Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Charlene Liu <Charlene.Liu@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Fix 10bit 4K display on CIK GPUs

Commit 72a7cf0aec0c ("drm/amd/display: Keep linebuffer pixel depth at
30bpp for DCE-11.0.") doesn't seems to have fixed 10bit 4K rendering over
DisplayPort for CIK GPUs. On my machine with a HAWAII GPU I get a broken
image that looks like it has an effective resolution of 1920x1080 but
scaled up in an irregular way. Reverting the commit or applying this
patch fixes the problem on v5.14-rc1.

Fixes: 72a7cf0aec0c ("drm/amd/display: Keep linebuffer pixel depth at 30bpp for DCE-11.0.")
Acked-by: Mario Kleiner <mario.kleiner.de@gmail.com>
Reviewed-by: Harry Wentland <harry.wentland@amd.com>
Signed-off-by: Liviu Dudau <liviu@dudau.co.uk>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu/ttm: optimize vram access in amdgpu_ttm_access_memory()

1. using vram aper to access vram if possible
2. avoid MM_INDEX/MM_DATA is not working when mmio protect feature is
enabled.

Signed-off-by: Kevin Wang <kevin1.wang@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu/ttm: replace duplicate code with exiting function

using exiting function to replace duplicate code blocks in
amdgpu_ttm_vram_write().

Signed-off-by: Kevin Wang <kevin1.wang@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: split amdgpu_device_access_vram() into two small parts

split amdgpu_device_access_vram()
1. amdgpu_device_mm_access(): using MM_INDEX/MM_DATA to access vram
2. amdgpu_device_aper_access(): using vram aperature to access vram (option)

Signed-off-by: Kevin Wang <kevin1.wang@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu/display: make a const array common_rates static, makes object smaller

Don't populate the const array common_rates on the stack but instead it
static. Makes the object code smaller by 80 bytes:

Before:
   text    data     bss     dec     hex filename
 268019   98322     256  366597   59805 ../display/amdgpu_dm/amdgpu_dm.o

After:
   text    data     bss     dec     hex filename
 267843   98418     256  366517   597b5 ../display/amdgpu_dm/amdgpu_dm.o

Reduction of 80 bytes

(gcc version 10.3.0)

Signed-off-by: Colin Ian King <colin.king@canonical.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/pm: update DRIVER_IF_VERSION for beige_goby

Update the version to 0xD for beige_goby.

Signed-off-by: Tao Zhou <tao.zhou1@amd.com>
Reviewed-by: Jack Gui <Jack.Gui@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: update gc golden setting for dimgrey_cavefish

Update gc_10_3_4 golden setting.

Signed-off-by: Tao Zhou <tao.zhou1@amd.com>
Reviewed-by: Guchun Chen <guchun.chen@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: update golden setting for sienna_cichlid

Update GFX golden setting for sienna_cichlid.

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

drm/amdgpu: update the golden setting for vangogh

This patch is to update the golden setting for vangogh.

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

drm/amd/display: avoid printing ERROR for unknown CEA parse(v2)

For the unknown CEA parse case on DMUB-enabled ASICs, dmesg will
print an error message like below, this will be captured by
automation tools as it has the word like ERROR during boot up
and treated as a false error, as it does not break bootup process.
So use DRM_WARN printing for this.

[drm:amdgpu_dm_update_freesync_caps [amdgpu]] *ERROR* Unknown EDID CEA parser results

v2: Use DRM_WARN to print such info.

Signed-off-by: Guchun Chen <guchun.chen@amd.com>
Reviewed-by: Stylon Wang <stylon.wang@amd.com>
Reviewed-by: Harry Wentland <harry.wentland@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: Switch to LFB for USBC PD FW in psp v13

Add USBC PD FW implementation here to be used with relevant ASICs.

Signed-off-by: Andrey Grodzovsky <andrey.grodzovsky@amd.com>
Reviewed-by: Alexander Deucher <Alexander.Deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: Switch to VRAM buffer for USBC PD FW.

System memory-based implementation for updating the
USBCPD is deprecated for so switching
to LFB based implementation for all the ASICs.

Signed-off-by: Andrey Grodzovsky <andrey.grodzovsky@amd.com>
Reviewed-by: Alexander Deucher <Alexander.Deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

amdgpu/nv.c - Optimize code for video codec support structure

Optimized the code for codec info structure initialization

Signed-off-by: Veerabadhran Gopalakrishnan <veerabadhran.gopalakrishnan@amd.com>
Reviewed-by: James Zhu <James.Zhu@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

amdgpu/nv.c - Added video codec support for Yellow Carp

Added the supported codecs in the video capabilities query.

Signed-off-by: Veerabadhran Gopalakrishnan <veerabadhran.gopalakrishnan@amd.com>
Reviewed-by: James Zhu <James.Zhu@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: add another Renoir DID

Add new PCI device id.

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

drm/amdgpu: Resolve bug in UMC 6.7 error offset calculation

Use correct channel and instance values

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/amd/pm: Add waiting for response of mode-reset message for yellow carp

Remove mdelay process and use smu_cmn_send_smc_msg_with_param to send
mode-reset message to SMC.

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

Revert "drm/amdkfd: Add heavy-weight TLB flush after unmapping"

This reverts commit 1098d658bef05e5fee634aab0b6a1fa590cfca24.

Reason for revert: it causes regressions on several Asics.

Signed-off-by: Eric Huang <jinhuieric.huang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

Revert "drm/amdgpu: Add table_freed parameter to amdgpu_vm_bo_update"

This reverts commit 075e8080c1a7571563171a07fa9ce47c4bc80044.

Reason for revert: the related commit is reverted.

Signed-off-by: Eric Huang <jinhuieric.huang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

Revert "drm/amdkfd: Make TLB flush conditional on mapping"

This reverts commit 31f33243788dcbae8bd2819ed83923a73f7dfd30.

Reason for revert: it causes regressions on several Asics.

Signed-off-by: Eric Huang <jinhuieric.huang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

Revert "drm/amdgpu: Fix warning of Function parameter or member not described"

This reverts commit 7a68d188d1c4a9d947369acaa19040a58baaaeda.

Reason for revert: the related commit is reverted.

Signed-off-by: Eric Huang <jinhuieric.huang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

Revert "drm/amdkfd: Add memory sync before TLB flush on unmap"

This reverts commit 3be4dca197010d1328df8b11febc8c40491be498.

Reason for revert: it causes regressions on several Asics.

Signed-off-by: Eric Huang <jinhuieric.huang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

Revert "drm/amdkfd: Only apply TLB flush optimization on ALdebaran"

This reverts commit 51627f03804173a64d23828bc9e4d8474451814f.

Reason for revert: it causes regression on Aldebaran.

Signed-off-by: Eric Huang <jinhuieric.huang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/pm: Fix BACO state setting for Beige_Goby

Correct BACO state setting for Beige_Goby

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

drm/amdgpu: Restore msix after FLR

After FLR, the msix will be cleared, so need to re-enable it.

Signed-off-by: Peng Ju Zhou <PengJu.Zhou@amd.com>
Signed-off-by: Emily.Deng <Emily.Deng@amd.com>
Reviewed-by: Lijo Lazar <lijo.lazar@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdkfd: Allow CPU access for all VRAM BOs

The thunk needs to mmap all BOs for CPU access to allow the debugger to
access them. Invisible ones are mapped with PROT_NONE.

Fixes: 71df0368e9b6 ("drm/amdgpu: Implement mmap as GEM object function")
Signed-off-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu/display - only update eDP's backlight level when necessary

[Why]
The original logic is to update eDP's backlight level
on every amdgpu dm atomic commit, which causes excessive
DMUB write. As a result, when playing game or moving window
around, DMUB timeout and system lagging are observed.

[How]
We only need to update eDP's backlight level when current level
doesn't match requested level.

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

drm/amdgpu: initialize umc ras function

support umc ras function initialization for aldebaran

v2: squash in compile fix

Signed-off-by: John Clements <john.clements@amd.com>
Signed-off-by: Stanley.Yang <Stanley.Yang@amd.com>
Reviewed-by: Guchun Chen <guchun.chen@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdkfd: handle fault counters on invalid address

prange is NULL if vm fault retry on invalid address, for this case, can
not use prange to get pdd, use adev to get gpuidx and then get pdd
instead, then increase pdd vm fault counter.

Signed-off-by: Philip Yang <Philip.Yang@amd.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: Correct the irq numbers for virtual crtc

The irq number should be decided by num_crtc, and the num_crtc could change
by parameter.

Signed-off-by: Emily Deng <Emily.Deng@amd.com>
Reviewed by: Monk Liu <monk.liu@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: update header file name

Update the register header file name.

Signed-off-by: Xiaomeng Hou <Xiaomeng.Hou@amd.com>
Reviewed-by: Aaron Liu <aaron.liu@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/pm: drop smu_v13_0_1.c|h files for yellow carp

Since there's nothing special in smu implementation for yellow carp,
it's better to reuse the common smu_v13_0 interfaces and drop the
specific smu_v13_0_1.c|h files.

v2: remove the duplicate register offset and shift mask header files as
well.

Signed-off-by: Xiaomeng Hou <Xiaomeng.Hou@amd.com>
Reviewed-by: Lijo Lazar <lijo.lazar@amd.com>
Reviewed-by: Kevin Wang <kevin1.wang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/pm: bump DRIVER_IF_VERSION for Sienna Cichlid

To suppress the annoying warning about version mismatch.

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

drm/amd/pm: update the gpu metrics data retrieving for Sienna Cichlid

Due to the structure layout change: "uint32_t ThrottlerStatus" -> "
uint8_t  ThrottlingPercentage[THROTTLER_COUNT]".

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

drm/amd/pm: new SmuMetrics data structure for Sienna Cichlid

Due to the structure layout change: "uint32_t ThrottlerStatus" -> "
uint8_t  ThrottlingPercentage[THROTTLER_COUNT]".

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

drm/amdgpu: return -EFAULT if copy_to_user() fails

If copy_to_user() fails then this should return -EFAULT instead of
-EINVAL.

Fixes: c65b0805e77919 ("drm/amdgpu: RAS EEPROM table is now in debugfs")
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Reviewed-by: Luben Tuikov <luben.tuikov@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: unlock on error in amdgpu_ras_debugfs_table_read()

This error path needs to unlock before returning.  While we're at it,
the correct error code from copy_to_user() failure is -EFAULT, not
-EINVAL.

Fixes: c65b0805e77919 ("drm/amdgpu: RAS EEPROM table is now in debugfs")
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Reviewed-by: Luben Tuikov <luben.tuikov@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: Fix signedness bug in __amdgpu_eeprom_xfer()

The i2c_transfer() function returns negatives or else the number of
messages transferred.  This code does not work because ARRAY_SIZE()
is type size_t and so that means negative values of "r" are type
promoted to high positive values which are greater than the ARRAY_SIZE().

Fix this by changing the < to != which works regardless of type
promotion.

Fixes: 746b584762e452 ("drm/amdgpu: Fixes to the AMDGPU EEPROM driver")
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Reviewed-by: Luben Tuikov <luben.tuikov@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: fix a signedness bug in __verify_ras_table_checksum()

If amdgpu_eeprom_read() returns a negative error code then the error
handling checks:

if (res < buf_size) {

The problem is that "buf_size" is a u32 so negative values are type
promoted to a high positive values and the condition is false.  Fix
this by changing the type of "buf_size" to int.

Fixes: 63d4c081a556a1 ("drm/amdgpu: Optimize EEPROM RAS table I/O")
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Reviewed-by: Luben Tuikov <luben.tuikov@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: increase max EDID size to 2k

[Why]
EDID CTS requires at least 2k (16 blocks) to be readable.

[How]
Increase EDID buffer size to 2k

Signed-off-by: Aric Cyr <aric.cyr@amd.com>
Reviewed-by: Anthony Koo <Anthony.Koo@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Round KHz up when calculating clock requests

[Why]
When requesting clocks from SMU which takes MHz inputs, DC will round
down KHz when converting to MHz, thus potentially requesting too low a
clock value.

[How]
Round up (ceil) when converting KHz to MHz for clock requests to SMU.

Signed-off-by: Aric Cyr <aric.cyr@amd.com>
Reviewed-by: Dmytro Laktyushkin <Dmytro.Laktyushkin@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: 3.2.142

DC version 3.2.142 brings improvements in multiple areas. In summary, we
highlight:

- Freesync improvements
- Remove unnecessary assert
- Firmware release 0.0.72
- Improve the EDID manipulation and DML calculations

Signed-off-by: Aric Cyr <aric.cyr@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: [FW Promotion] Release 0.0.72

- Updated SCR definition for FW boot options for Separate DCN init
  for DMUB FW loaded in VBL

Signed-off-by: Anthony Koo <Anthony.Koo@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Adjust types and formatting for future development

Type adjustments and formatting fixes.

Signed-off-by: Alvin Lee <Alvin.Lee2@amd.com>
Reviewed-by: Dmytro Laktyushkin <Dmytro.Laktyushkin@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: remove faulty assert

Signed-off-by: Dmytro Laktyushkin <Dmytro.Laktyushkin@amd.com>
Reviewed-by: Wenjing Liu <Wenjing.Liu@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

Revert "drm/amd/display: Always write repeater mode regardless of LTTPR"

This reverts commit 2b7605d73b97e2fa28e0817242e66ca968d2a7cb

Some displays are not lighting up when put in LTTPR Transparent Mode

Signed-off-by: Wesley Chalmers <Wesley.Chalmers@amd.com>
Reviewed-by: Jun Lei <Jun.Lei@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Fix updating infoframe for DCN3.1 eDP

[Why]
We're only treating TMDS as a valid target for infoframe updates which
results in PSR being unable to transition from state 4 to state 5.

[How]
Also allow infoframe updates for DCN3.1 - following how we handle
this path for earlier ASIC as well.

Signed-off-by: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com>
Reviewed-by: Eric Yang <eric.yang2@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Add Freesync HDMI support to DM with DMUB

[Why]
Changes in DM needed to support Freesync HDMI on DMUB.

[How]
Change implementation to parse CEA blocks in case of DMUB-enabled ASICs.

Signed-off-by: Stylon Wang <stylon.wang@amd.com>
Reviewed-by: Nicholas Kazlauskas <Nicholas.Kazlauskas@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Add null checks

Added NULL checks before two problematic statements

Signed-off-by: Wang <anguwang@amd.com>
Reviewed-by: Martin Leung <Martin.Leung@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: DMUB Outbound Interrupt Process-X86

[Why]
dmub would notify x86 response time violation by GPINT_DATAOUT

[How]
1. Use GPINT_DATAOUT to trigger x86 interrupt
2. Register GPINT_DATAOUT interrupt handler.
3. Trigger ACR while GPINT_DATAOUT occurred.

Signed-off-by: Chun-Liang Chang <Chun-Liang.Chang@amd.com>
Reviewed-by: Jun Lei <Jun.Lei@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: isolate link training setting override to its own function

There is a difference between our default behavior and override
behavior. For default behavior we need to decide link training settings
within specs' limitation and mandates.
For override behavior we do not need to follow all these requirements.
We are isolating override decision to its own function to maintain the
integrity of our specs compliant default behavior.

Signed-off-by: Wenjing Liu <wenjing.liu@amd.com>
Reviewed-by: George Shen <George.Shen@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: Return error if no RAS

In amdgpu_ras_query_error_count() return an error
if the device doesn't support RAS. This prevents
that function from having to always set the values
of the integer pointers (if set), and thus
prevents function side effects--always to have to
set values of integers if integer pointers set,
regardless of whether RAS is supported or
not--with this change this side effect is
mitigated.

Also, if no pointers are set, don't count, since
we've no way of reporting the counts.

Also, give this function a kernel-doc.

Cc: Alexander Deucher <Alexander.Deucher@amd.com>
Cc: John Clements <john.clements@amd.com>
Cc: Hawking Zhang <Hawking.Zhang@amd.com>
Reported-by: Tom Rix <trix@redhat.com>
Fixes: a46751fbcde505 ("drm/amdgpu: Fix RAS function interface")
Signed-off-by: Luben Tuikov <luben.tuikov@amd.com>
Reviewed-by: Alexander Deucher <Alexander.Deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: SRIOV flr_work should take write_lock

[Why]
If flr_work takes read_lock, then other threads who takes
read_lock can access hardware when host is doing vf flr.

[How]
flr_work should take write_lock to avoid this case.

Signed-off-by: Jingwen Chen <Jingwen.Chen2@amd.com>
Reviewed-by: Monk Liu <monk.liu@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: The I2C IP doesn't support 0 writes/reads

The I2C IP doesn't support writes or reads of 0 bytes.

In order for a START/STOP transaction to take
place on the bus, the data written/read has to be
at least one byte.

That is, you cannot generate a write with 0 bytes,
just to get the ACK from a device, just so you can
probe that device if it is on the bus and so to
discover all devices on the bus--you'll have to
read at least one byte. Writes of 0 bytes generate
no START/STOP on this I2C IP--the bus is not
engaged at all.

Set the I2C_AQ_NO_ZERO_LEN to the existing I2C
quirk tables for Aldebaran, Arcturus, Navi10 and
Sienna Cichlid, and add a quirk table to the I2C
driver which drives the bus when the SMU
doesn't--for instance on Vega20.

Cc: Alexander Deucher <Alexander.Deucher@amd.com>
Cc: Andrey Grodzovsky <Andrey.Grodzovsky@amd.com>
Cc: Lijo Lazar <Lijo.Lazar@amd.com>
Cc: John Clements <john.clements@amd.com>
Cc: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Luben Tuikov <luben.tuikov@amd.com>
Reviewed-by: Lijo Lazar <Lijo.Lazar@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/pm: Add I2C quirk table to Aldebaran

Add I2C quirk table to Aldebaran.

Cc: Alexander Deucher <Alexander.Deucher@amd.com>
Cc: Andrey Grodzovsky <Andrey.Grodzovsky@amd.com>
Cc: Lijo Lazar <Lijo.Lazar@amd.com>
Cc: John Clements <john.clements@amd.com>
Cc: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Luben Tuikov <luben.tuikov@amd.com>
Reviewed-by: Lijo Lazar <Lijo.Lazar@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: Read clock counter via MMIO to reduce delay (v5)

[Why]
GPU timing counters are read via KIQ under sriov, which will introduce
a delay.

[How]
It could be directly read by MMIO.

v2: Add additional check to prevent carryover issue.
v3: Only check for carryover for once to prevent performance issue.
v4: Add comments of the rough frequency where carryover happens.
v5: Remove mutex and gfxoff ctrl unused with current timing registers.

Signed-off-by: YuBiao Wang <YuBiao.Wang@amd.com>
Acked-by: Horace Chen <horace.chen@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.co>
Reviewed-by: Monk Liu <monk.liu@amd.com>
Reviewed-by: Lijo Lazar <lijo.lazar@amd.com>
Reviewed-by: Luben Tuikov <luben.tuikov@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdkfd: Only apply TLB flush optimization on ALdebaran

It is based on reverting two patches back.
  drm/amdkfd: Make TLB flush conditional on mapping
  drm/amdgpu: Add table_freed parameter to amdgpu_vm_bo_update

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

drm/amdgpu: separate out vm pasid assignment

Use new helper function amdgpu_vm_set_pasid() to
assign vm pasid value. This also ensures that we don't free
a pasid from vm code as pasids are allocated somewhere else.

Signed-off-by: Nirmoy Das <nirmoy.das@amd.com>
Acked-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: use xarray for storing pasid in vm

Replace idr with xarray as we actually need hash functionality.
Cleanup code related to vm pasid by adding helper function.

Signed-off-by: Nirmoy Das <nirmoy.das@amd.com>
Acked-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: Avoid printing of stack contents on firmware load error

In case when psp_init_asd_microcode() fails to load ASD microcode file,
psp_v12_0_init_microcode() tries to print the firmware filename that
failed to load before bailing out.

This is wrong because:

- the firmware filename it would want it print is an incorrect one as
  psp_init_asd_microcode() and psp_v12_0_init_microcode() are loading
  different filenames
- it tries to print fw_name, but that's not yet been initialized by that
  time, so it prints random stack contents, e.g.

    amdgpu 0000:04:00.0: Direct firmware load for amdgpu/renoir_asd.bin failed with error -2
    amdgpu 0000:04:00.0: amdgpu: fail to initialize asd microcode
    amdgpu 0000:04:00.0: amdgpu: psp v12.0: Failed to load firmware "\xfeTO\x8e\xff\xff"

Fix that by bailing out immediately, instead of priting the bogus error
message.

Reported-by: Vojtech Pavlik <vojtech@ucw.cz>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>