drm/amdgpu: add the IP discovery IP versions for HW INFO data

Use the former pad element to store the IP versions from the
IP discovery table.  This allows userspace to get the IP
version from the kernel to better align with hardware IP
versions.

Proposed mesa patch:
https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/17411/diffs?commit_id=c8a63590dfd0d64e6e6a634dcfed993f135dd075

Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Drop FPU flags from dcn32_clk_mgr

We are working to isolate FPU operations inside the DML folder, and the
file dcn32_clk_mgr has some of these operations. This commit moves the
FPU operations inside the clock manager and creates the dcn32_fpu file
to aggregate those operations. Note that there is no functional change
ere, just moving code from one part to another.

Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Reviewed-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/amd/display: Revert "drm/amd/display: disable idle optimizations"

This reverts commit e7ef5569e71bf3fec01ea513c27c6081c0dbbc64.

Idle optimization was disabled due to SMU and firmware bugs. Enable it
back for DCN32. DCN321 has them enabled already

Fixes: 6a640b95b061 ("drm/amd/display: disable idle optimizations")
Signed-off-by: Aurabindo Pillai <aurabindo.pillai@amd.com>
Reviewed-and-tested-by: Evan Quan <evan.quan@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Update DML logic for unbounded req handling

[why]
Unbounded request logic in resource/DML has some issues where unbounded
request is being enabled incorrectly. SW today enables unbounded request
unconditionally in hardware, on the assumption that HW can always
support it in single pipe scenarios.

This worked until now because the same assumption is made in DML. A new
DML update is needed to fix a bug, where there are single pipe scenarios
where unbounded cannot be enabled, and this change in DML needs to be
ported in, and dcn32 resource logic fixed.

[how]
First, dcn32_resource should program unbounded req in HW according to
unbounded req enablement output from DML, as opposed to DML input.

Second, port in DML update which disables unbounded req in some
scenarios to fix an issue with poor stutter performance

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

drm/amd/display: Update Cursor Attribute MALL cache

[Why]
Cursor size can update without MALL cache update.
Update the register on cursor attribute as well.

[How]
Update cursor MALL cache on cursor attribute update.

Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Reviewed-by: Alvin Lee <Alvin.Lee2@amd.com>
Acked-by: Alan Liu <HaoPing.Liu@amd.com>
Signed-off-by: Chris Park <chris.park@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: Fix comment typo

The double `to' is duplicated in the comment, remove one.

Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Jason Wang <wangborong@cdjrlc.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/radeon: Fix comment typo

The double `have' is duplicated in line 696, remove one.

Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Jason Wang <wangborong@cdjrlc.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: add dm ip block for dcn 3.1.4

Adding dm ip block to enable display on dcn 3.1.4.

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

drm/amd/display: Disable dmcu fw loading for dcn314

[Why]
DCN 3.1.4 uses dmub not dmcu. Attempt to identify
dmcu firmware for dcn314 results in dm init error:
"Unsupported ASIC type"

[How]
Add dcn314 to the list of asics that don't require dmcu

Signed-off-by: Roman Li <roman.li@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Harry Wentland <harry.wentland@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Loop through all pipes for DET allocation

[Why & How]
There are cases where the pipes populated are not all at the top
of the pipes list under context. Loop through all pipes for DET
allocation instead of just the number of populated ones, even if
some unpopulated pipes are iterated through unnecessarily.

Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Reviewed-by: Alvin Lee <Alvin.Lee2@amd.com>
Acked-by: Alan Liu <HaoPing.Liu@amd.com>
Signed-off-by: Taimur Hassan <Syed.Hassan@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Disable GPUVM in IP resource configuration

[Why]
VM enabled in IP configuration causes UCLK not
reaching DPM0. The expectation for VM enable should
be that KMD will indicate to DAL when VM is enabled,
then DAL will set the bit accordingly

[How]
Set gpuvm_enable to zero in DCN3_20 and DCN3_21 resource.

Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Reviewed-by: Martin Leung <Martin.Leung@amd.com>
Acked-by: Alan Liu <HaoPing.Liu@amd.com>
Signed-off-by: Vladimir Stempen <vladimir.stempen@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

Documentation/gpu: Add GFXOFF section

Add a GFXOFF section at "GPU Power Controls" file, explaining what it is
and how userspace can interact with it.

v2: minor tweaks to the documenation (Alex)

Signed-off-by: André Almeida <andrealmeid@igalia.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/debugfs: Expose GFXOFF state to userspace

GFXOFF has two different "state" values: one to define if the GPU is
allowed/disallowed to enter GFXOFF, usually called state; and another
one to define if currently GFXOFF is being used, usually called status.
Even when GFXOFF is allowed, GPU firmware can decide to not used it
accordingly to the GPU load.

Userspace can allow/disallow GPUs to enter into GFXOFF via debugfs. The
kernel maintains a counter of requests for GFXOFF (gfx_off_req_count)
that should be decreased to allow GFXOFF and increased to disallow.

The issue with this interface is that userspace can't be sure if GFXOFF
is currently allowed. Even by checking amdgpu_gfxoff file, one might get
an ambiguous 2, that means that GPU is currently out of GFXOFF, but that
can be either because it's currently disallowed or because it's allowed
but given the current GPU load it's enabled. Then, userspace needs to
rely on the fact that GFXOFF is enabled by default on boot and to track
this information.

To make userspace life easier and GFXOFF more reliable, return the
current state of GFXOFF to userspace when reading amdgpu_gfxoff with the
same semantics of writing: 0 means not allowed, not 0 means allowed.

Expose the current status of GFXOFF through a new file,
amdgpu_gfxoff_status.

Signed-off-by: André Almeida <andrealmeid@igalia.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: move dcn31_update_soc_for_wm_a func to dml fpu folder

Although dcn31_update_soc_for_wm_a() is only called in dml/dcn31/dcn31_fpu by
dc->res_pool->funcs->update_soc_for_wm_a(dc, context), it's declared in
dcn31_resource that is not FPU protected. Move this function to dcn31_fpu
file as part of the work to isolate FPU code.

Signed-off-by: Melissa Wen <mwen@igalia.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Remove unused MaxUsedBW variable

Remove the variable MaxUsedBW from the function
DISPCLKDPPCLKDCFCLKDeepSleepPrefetchParametersWatermarksAndPerformanceCalculation.
As a side-effect, the variables MaxPerPlaneVActiveWRBandwidth and
WRBandwidth are also removed.

This was pointed by clang with the following warning:

drivers/gpu/drm/amd/amdgpu/../display/dc/dml/dcn30/display_mode_vba_30.c:3043:10:
warning: variable 'MaxUsedBW' set but not used [-Wunused-but-set-variable]
                double MaxUsedBW = 0;
                       ^
1 warning generated.

Reviewed-by: André Almeida <andrealmeid@igalia.com>
Signed-off-by: Maíra Canal <mairacanal@riseup.net>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Remove unused variables from dcn10_stream_encoder

The variable regval from the function enc1_update_generic_info_packet
and the variables dynamic_range_rgb and dynamic_range_ycbcr from the
function enc1_stream_encoder_dp_set_stream_attribute are not currently
used.

This was pointed by clang with the following warnings:

drivers/gpu/drm/amd/amdgpu/../display/dc/dcn10/dcn10_stream_encoder.c:62:11:
warning: variable 'regval' set but not used [-Wunused-but-set-variable]
        uint32_t regval;
                 ^
drivers/gpu/drm/amd/amdgpu/../display/dc/dcn10/dcn10_stream_encoder.c:262:10:
warning: variable 'dynamic_range_rgb' set but not used [-Wunused-but-set-variable]
        uint8_t dynamic_range_rgb = 0; /*full range*/
                ^
drivers/gpu/drm/amd/amdgpu/../display/dc/dcn10/dcn10_stream_encoder.c:263:10:
warning: variable 'dynamic_range_ycbcr' set but not used [-Wunused-but-set-variable]
        uint8_t dynamic_range_ycbcr = 1; /*bt709*/
                ^
3 warnings generated.

Reviewed-by: André Almeida <andrealmeid@igalia.com>
Signed-off-by: Maíra Canal <mairacanal@riseup.net>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Remove unused variables from dml_rq_dlg_get_dlg_params

Remove the variables dispclk_delay_subtotal and dppclk_delay_subtotal from
the function dml_rq_dlg_get_dlg_params.

This was pointed by clang with the following warning:

drivers/gpu/drm/amd/amdgpu/../display/dc/dml/dcn31/display_rq_dlg_calc_31.c:920:15:
warning: variable 'dispclk_delay_subtotal' set but not used [-Wunused-but-set-variable]
        unsigned int dispclk_delay_subtotal;
                     ^
drivers/gpu/drm/amd/amdgpu/../display/dc/dml/dcn31/display_rq_dlg_calc_31.c:919:15:
warning: variable 'dppclk_delay_subtotal' set but not used [-Wunused-but-set-variable]
        unsigned int dppclk_delay_subtotal;
                     ^
2 warnings generated.

Reviewed-by: André Almeida <andrealmeid@igalia.com>
Signed-off-by: Maíra Canal <mairacanal@riseup.net>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Remove unused NumberOfStates variable

Remove the unused unsigned int NumberOfStates from the file, which was
declared but never hooked up.

This was pointed by clang with the following warning:

drivers/gpu/drm/amd/amdgpu/../display/dc/dml/dcn32/display_mode_vba_32.c:33:27:
warning: unused variable 'NumberOfStates' [-Wunused-const-variable]
static const unsigned int NumberOfStates = DC__VOLTAGE_STATES;
                          ^
1 warning generated.

Reviewed-by: André Almeida <andrealmeid@igalia.com>
Signed-off-by: Maíra Canal <mairacanal@riseup.net>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Remove unused dml32_CalculatedoublePipeDPPCLKAndSCLThroughput function

Remove dml32_CalculatedoublePipeDPPCLKAndSCLThroughput function, which is not used in
the codebase.

This was pointed by clang with the following warning:

drivers/gpu/drm/amd/amdgpu/../display/dc/dml/dcn32/display_mode_vba_util_32.c:393:6:
warning: no previous prototype for function
'dml32_CalculatedoublePipeDPPCLKAndSCLThroughput' [-Wmissing-prototypes]
void dml32_CalculatedoublePipeDPPCLKAndSCLThroughput(
     ^
drivers/gpu/drm/amd/amdgpu/../display/dc/dml/dcn32/display_mode_vba_util_32.c:393:1:
note: declare 'static' if the function is not intended to be used outside of
this translation unit
void dml32_CalculatedoublePipeDPPCLKAndSCLThroughput(
^
static
1 warning generated.

Reviewed-by: André Almeida <andrealmeid@igalia.com>
Signed-off-by: Maíra Canal <mairacanal@riseup.net>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Remove unused clk_src variable

Remove the variable clk_src from the function dcn3_get_pix_clk_dividers.

This was pointed by clang with the following warning:

drivers/gpu/drm/amd/amdgpu/../display/dc/dce/dce_clock_source.c:1279:25: warning:
variable 'clk_src' set but not used [-Wunused-but-set-variable]
        struct dce110_clk_src *clk_src;
                               ^
1 warning generated.

Reviewed-by: André Almeida <andrealmeid@igalia.com>
Signed-off-by: Maíra Canal <mairacanal@riseup.net>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Update de-tile override to anticipate pipe splitting

[Why]
For certain MPO configurations, DML will split a pipe after DET buffer has
already been allocated by driver, resulting in allocation of more DET
segments than the configurable return buffer has, causing underflow.

[How]
Determine during DET override calculation whether or not a pipe will be
split later on by DML, and distribute DET segments based on expected
number of pipes.

Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Reviewed-by: Dmytro Laktyushkin <Dmytro.Laktyushkin@amd.com>
Acked-by: Alan Liu <HaoPing.Liu@amd.com>
Signed-off-by: Taimur Hassan <Syed.Hassan@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Don't set dram clock change requirement for SubVP

[Description]
In general cases we want to keep the dram clock change requirement (we
prefer configs that support MCLK switch). Only override to false for
SubVP.

Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Acked-by: Alan Liu <HaoPing.Liu@amd.com>
Signed-off-by: Alvin Lee <Alvin.Lee2@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Fix hard hang if DSC is disabled

We want to calculate the DTB clock values when DSC is enabled; however,
this is not the current behavior implemented in DCN32. Right now, DML is
trying to calculate DSC values even if DSC is disabled; as a result, we
can have a hard hang due to wrong clock calculation. This commit fixes
this issue by moving the calculation after the DSC check.

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

drm/amd/display: remove number of DSC slices override in DML

[why]
Number of DSC slices is an input to DML with high dependency
on display specific capability. This isn't something DML can decide
on its own. DML has to use the original number of DSC slices input
to DML during validation without modification. Otherwise the
computed DSC delay will not reflect the current configuration
and therefore causes validation failures.

[how]
Remove DML override for number of DSC slices parameter.

Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Reviewed-by: Alvin Lee <Alvin.Lee2@amd.com>
Acked-by: Alan Liu <HaoPing.Liu@amd.com>
Signed-off-by: Wenjing Liu <wenjing.liu@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Create a file dedicated for CRTC

[Why]
The amdgpu_dm file contains most of the code that works as an interface
between DRM API and DC. As a result, this file becomes very large since
it comprises multiple abstractions such as CRTC manipulation.

[How]
This commit extracts the CRTC code to its specific file named
amdgpu_dm_crtc. This change does not change anything inside the
functions; the only exception is converting some static functions to a
global function.

v2: fix ifdef merge mix up (Alex)

Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Reviewed-by: Harry Wentland <Harry.Wentland@amd.com>
Acked-by: Alan Liu <HaoPing.Liu@amd.com>
Signed-off-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Create a file dedicated to planes

[Why]
The amdgpu_dm file contains most of the code that works as an interface
between DRM API and DC. As a result, this file becomes very large since
it comprises multiple abstractions such as plane manipulation.

[How]
This commit extracts the plane code to its specific file named
amdgpu_dm_plane. This change does not change anything inside the
functions; the only exception is converting some static functions to a
global function.

Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Reviewed-by: Harry Wentland <Harry.Wentland@amd.com>
Acked-by: Alan Liu <HaoPing.Liu@amd.com>
Signed-off-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Add tags for indicating mst progress status

[Why & How]
In order to leverage igt tool to maintain mst feature, expose new
debugfs entry "mst_progress_status".

In our dm flow, record down the result of each phase of mst and user
can examine the mst result by checking whether each phase get completed
successfully.

Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Reviewed-by: Hersen Wu <hersenxs.wu@amd.com>
Acked-by: Alan Liu <HaoPing.Liu@amd.com>
Signed-off-by: Wayne Lin <Wayne.Lin@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Add is_mst_connector debugfs entry

[Why & How]
Add "is_mst_connector" debugfs entry to help distinguish whether
a connector is in a mst topology or not.

Access it with the following command:
    cat /sys/kernel/debug/dri/0/DP-X/is_mst_connector

Result:
- "root" stands for the root connector of the topology
- "branch" stands for branch device of the topology
- "end" stands for leaf node connector of the topology
- "no" stands for the connector is not a device of a mst topology

Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Reviewed-by: Hersen Wu <hersenxs.wu@amd.com>
Acked-by: Alan Liu <HaoPing.Liu@amd.com>
Signed-off-by: Wayne Lin <Wayne.Lin@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: fix trigger_hotplug to support mst case

[Why & How]
Correct few problems below to have debugfs trigger_hotplug entry
supports mst case

* Adjust the place for acquiring the hpd_lock. We'll also access
  dc_link when simulate unplug
* When detect the connector is a mst root, call
  reset_cur_dp_mst_topology() to simulate unplug
* Don't support hotplug caused by CSN message since we can't change
  mst topology info directly. We can't simulate that
* Clean up redundant code

Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Reviewed-by: Hersen Wu <hersenxs.wu@amd.com>
Acked-by: Alan Liu <HaoPing.Liu@amd.com>
Signed-off-by: Wayne Lin <Wayne.Lin@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Expose function reset_cur_dp_mst_topology

[Why & How]
Need to leverage this function out of dc_link.c. Change it to public.

Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Reviewed-by: Hersen Wu <hersenxs.wu@amd.com>
Acked-by: Alan Liu <HaoPing.Liu@amd.com>
Signed-off-by: Wayne Lin <Wayne.Lin@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Update in dml

Update DML to configure drr_display in vba struct.

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

drm/amd/display: Remove unused variable

Remove an unused variable "remove_disconnect_edp" which was a workaround
bit.

Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Acked-by: Alan Liu <HaoPing.Liu@amd.com>
Signed-off-by: Jun Lei <Jun.Lei@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Support vertical interrupt 0 for all dcn ASIC

[Why]
When CONFIG_DRM_AMD_SECURE_DISPLAY is enabled, it will try
to register vertical interrupt 0 for specific task.

Currently, only dcn10 have defined relevant info for vertical interrupt
0. If we enable CONFIG_DRM_AMD_SECURE_DISPLAY for other dcn ASIC, will
get DC_IRQ_SOURCE_INVALID while calling dc_interrupt_to_irq_source() and
cause pointer errors.

[How]
Add support of vertical interrupt 0 for all dcn ASIC.

Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Acked-by: Alan Liu <HaoPing.Liu@amd.com>
Signed-off-by: Wayne Lin <Wayne.Lin@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/pm: enable gfx ulv and gpo on smu_v13_0_7

enable gfx ulv and gpo on smu_v13_0_7

Signed-off-by: Kenneth Feng <kenneth.feng@amd.com>
Acked-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Change get_pipe_idx function scope

Turn previously global function into a static function as it is not used
outside the file.

Reviewed-by: André Almeida <andrealmeid@igalia.com>
Signed-off-by: Maíra Canal <mairacanal@riseup.net>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: Write masked value to control register

On the dce_v6_0 and dce_v8_0 hpd tear down callback, the tmp variable
should be written into the control register instead of 0.

Reviewed-by: André Almeida <andrealmeid@igalia.com>
Signed-off-by: Maíra Canal <mairacanal@riseup.net>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: fix scratch register access method in SRIOV

The scratch register should be accessed through MMIO instead of RLCG
in SRIOV, since it being used in RLCG register access function.

Fixes: d54762cc3e6a ("drm/amdgpu: nuke dynamic gfx scratch reg allocation")
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Gavin Wan <Gavin.Wan@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: remove acc_size from reserve/unreserve mem

TTM used to track the "acc_size" of all BOs internally. We needed to
keep track of it in our memory reservation to avoid TTM running out
of memory in its own accounting. However, that "acc_size" accounting
has since been removed from TTM. Therefore we don't really need to
track it any more.

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

drm/amdgpu: Protect the amdgpu_bo_list list with a mutex v2

Protect the struct amdgpu_bo_list with a mutex. This is used during command
submission in order to avoid buffer object corruption as recorded in
the link below.

v2 (chk): Keep the mutex looked for the whole CS to avoid using the
  list from multiple CS threads at the same time.

Suggested-by: Christian König <christian.koenig@amd.com>
Cc: Alex Deucher <Alexander.Deucher@amd.com>
Cc: Andrey Grodzovsky <Andrey.Grodzovsky@amd.com>
Cc: Vitaly Prosyak <Vitaly.Prosyak@amd.com>
Link: https://gitlab.freedesktop.org/drm/amd/-/issues/2048
Signed-off-by: Luben Tuikov <luben.tuikov@amd.com>
Signed-off-by: Christian König <christian.koenig@amd.com>
Tested-by: Luben Tuikov <luben.tuikov@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: add comment to HW_IP_VCN_ENC type

From VCN4, AMDGPU_HW_IP_VCN_ENC is re-used to support
both encoding and decoding jobs.

Link: https://gitlab.freedesktop.org/mesa/drm/-/merge_requests/245/commits
Reviewed-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Leo Liu <leo.liu@amd.com>
Signed-off-by: Ruijing Dong <ruijing.dong@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/pm: enable mode1 reset for smu_v13_0_7

enable mode1 reset for smu_v13_0_7 since it's missing.

Signed-off-by: Kenneth Feng <kenneth.feng@amd.com>
Reviewed-by: Yang Wang <kevinyang.wang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: correct the PSP_BL_CMD enum

To match with the enum defined in trusted os

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

drm/amdgpu: drop runpm from amdgpu_device structure

It's redundant, as now switching to rpm_mode to indicate
runtime power management mode.

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

drm/amdgpu: drop runtime pm disablement quirk on several sienna cichlid cards

This quirk is not needed any more as it's fixed by bypassing
SMU FW reloading in runtime resume.

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

drm/amdgpu: skip SMU FW reloading in runpm BACO case

SMU is always alive, so it's fine to skip SMU FW reloading
when runpm resumed from BACO, this can avoid some race issues
when resuming SMU.

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

drm/amdgpu: introduce runtime pm mode

It can benefit code consistency in future.

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

drm/amd/display: Remove unnecessary NULL check in commit_planes_for_stream()

Smatch complains that:

    drivers/gpu/drm/amd/amdgpu/../display/dc/core/dc.c:3369 commit_planes_for_stream()
    warn: variable dereferenced before check 'stream' (see line 3114)

The 'stream' pointer cannot be NULL and the check can be removed.

Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: Clarify asics naming in Kconfig options

Clarify which architecture those asics acronyms refers to.

Signed-off-by: André Almeida <andrealmeid@igalia.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Fix spelling mistake "supporing" -> "supporting"

There is a spelling mistake in a dml_print message. Fix it.

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

drm/amd/display: Enable building new display engine with KCOV enabled

The new display engine uses floating point math, which is not supported
by KCOV. Commit 9d1d02ff3678 ("drm/amd/display: Don't build DCN1 when kcov
is enabled") tried to work around the problem by disabling
CONFIG_DRM_AMD_DC_DCN if KCOV_INSTRUMENT_ALL and KCOV_ENABLE_COMPARISONS
are enabled. The result is that KCOV can not be enabled on systems which
require this display engine. A much simpler and less invasive solution is
to disable KCOV selectively when compiling the display enagine while
keeping it enabled for the rest of the kernel.

Fixes: 9d1d02ff3678 ("drm/amd/display: Don't build DCN1 when kcov is enabled")
Cc: Arnd Bergmann <arnd@arndb.de>
Cc: Leo Li <sunpeng.li@amd.com>
Reviewed-by: Harry Wentland <harry.wentland@amd.com>
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: restore original stable pstate on ctx fini

Save the original stable pstate on ctx init and restore
it on ctx fini so that we restore a manually selected
stable pstate on ctx exit.

v2: fix init order (Alex)
v3: don't add new variable to ctx struct (Evan)

Fixes: c65b364c52ba ("drm/amdgpu/ctx: only reset stable pstate if the user changed it (v2)")
Reviewed-by: Evan Quan <evan.quan@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: use the same HDP flush registers for all nbio 2.3.x

Align RDNA2.x with other asics.  One HDP bit per SDMA instance,
aligned with firmware.  This is effectively a revert of
commit 369b7d04baf3 ("drm/amdgpu/nbio2.3: don't use GPU_HDP_FLUSH bit 12").
On further discussions with the relevant hardware teams,
re-align the bits for SDMA.

Fixes: 369b7d04baf3 ("drm/amdgpu/nbio2.3: don't use GPU_HDP_FLUSH bit 12")
Reviewed-by: Kent Russell <kent.russell@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: use the same HDP flush registers for all nbio 7.4.x

Align aldebaran with all other asics.  One HDP bit per
SDMA instance, aligned with firmware.  This is effectively
a revert of
commit a0f9f8546668 ("drm/amdgpu/nbio7.4: don't use GPU_HDP_FLUSH bit 12").
On further discussions with the relevant hardware teams,
re-align the bits for SDMA.

Fixes: a0f9f8546668 ("drm/amdgpu/nbio7.4: don't use GPU_HDP_FLUSH bit 12")
Reviewed-by: Kent Russell <kent.russell@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: create I2S platform devices for Jadeite platform

Jadeite platform uses I2S MICSP instance.
Create platform devices for DMA controller and I2S controller for
Jadeite platform.

Signed-off-by: Vijendar Mukunda <Vijendar.Mukunda@amd.com>
Acked-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: add dmi check for jadeite platform

DMI check is required to distinguish Jadeite platform from
Stoney base variant.
Add DMI check logic for Jadeite platform.

Signed-off-by: Vijendar Mukunda <Vijendar.Mukunda@amd.com>
Acked-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: Call trace info was found in dmesg when loading amdgpu

In the case of SRIOV, the register smnMp1_PMI_3_FIFO will get an invalid
value which will cause the "shift out of bound". In Ubuntu22.04, this
issue will be checked an related call trace will be reported in dmesg.

Signed-off-by: lin cao <lin.cao@amd.com>
Reviewed-by: Jingwen Chen <Jingwen.Chen2@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: fix for coding style issues

Fixed below checkpatch warnings and errors

drivers/gpu/drm/amd/amdgpu/amdgpu_acp.c:131: CHECK: Comparison to NULL could be written "apd"
drivers/gpu/drm/amd/amdgpu/amdgpu_acp.c:150: CHECK: Comparison to NULL could be written "apd"
drivers/gpu/drm/amd/amdgpu/amdgpu_acp.c:196: CHECK: Prefer kernel type 'u64' over 'uint64_t'
drivers/gpu/drm/amd/amdgpu/amdgpu_acp.c:224: CHECK: Please don't use multiple blank lines
drivers/gpu/drm/amd/amdgpu/amdgpu_acp.c:226: CHECK: Comparison to NULL could be written "!adev->acp.acp_genpd"
drivers/gpu/drm/amd/amdgpu/amdgpu_acp.c:233: CHECK: Please don't use multiple blank lines
drivers/gpu/drm/amd/amdgpu/amdgpu_acp.c:239: CHECK: Alignment should match open parenthesis
drivers/gpu/drm/amd/amdgpu/amdgpu_acp.c:241: CHECK: Comparison to NULL could be written "!adev->acp.acp_cell"
drivers/gpu/drm/amd/amdgpu/amdgpu_acp.c:247: CHECK: Comparison to NULL could be written "!adev->acp.acp_res"
drivers/gpu/drm/amd/amdgpu/amdgpu_acp.c:253: CHECK: Comparison to NULL could be written "!i2s_pdata"
drivers/gpu/drm/amd/amdgpu/amdgpu_acp.c:350: CHECK: Alignment should match open parenthesis
drivers/gpu/drm/amd/amdgpu/amdgpu_acp.c:550: ERROR: that open brace { should be on the previous line

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

drm/amdgpu: add umc ras functions for umc v8_10_0

1. Support query umc ras error counter.
2. Support ras umc ue error address remapping.

Signed-off-by: YiPeng Chai <YiPeng.Chai@amd.com>
Reviewed-by: Alexander Deucher <Alexander.Deucher@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Reviewed-by: Tao Zhou <tao.zhou1@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: add umc v8_10_0 ip headers

Add umc v8_10_0 register offset and shift masks
header files

Signed-off-by: YiPeng Chai <YiPeng.Chai@amd.com>
Reviewed-by: Alexander Deucher <Alexander.Deucher@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Reviewed-by: Tao Zhou <tao.zhou1@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: Get rid of amdgpu_job->external_hw_fence

This is a follow-up cleanup to [1]. See bellow refcount balancing
for calling amdgpu_job_submit_direct after this cleanup as far
as I calculated.

amdgpu_fence_emit
dma_fence_init 1
dma_fence_get(fence) 2
rcu_assign_pointer(*ptr, dma_fence_get(fence) 3

---> amdgpu_job_submit_direct completes before fence signaled
amdgpu_sa_bo_free
(*sa_bo)->fence = dma_fence_get(fence) 4

amdgpu_job_free
dma_fence_put 3

amdgpu_vcn_enc_get_destroy_msg
*fence = dma_fence_get(f) 4
dma_fence_put(f); 3

amdgpu_vcn_enc_ring_test_ib
dma_fence_put(fence) 2

amdgpu_fence_process
dma_fence_put 1

amdgpu_sa_bo_remove_locked
dma_fence_put 0

---> amdgpu_job_submit_direct completes after fence signaled
amdgpu_fence_process
dma_fence_put 2

amdgpu_job_free
dma_fence_put 1

amdgpu_vcn_enc_get_destroy_msg
*fence = dma_fence_get(f) 2
dma_fence_put(f); 1

amdgpu_vcn_enc_ring_test_ib
dma_fence_put(fence) 0

[1] - https://patchwork.kernel.org/project/dri-devel/cover/20220624180955.485440-1-andrey.grodzovsky@amd.com/

Signed-off-by: Andrey Grodzovsky <andrey.grodzovsky@amd.com>
Suggested-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: limiting AV1 to first instance on VCN4 decode

AV1 is only supported on first instance.

Signed-off-by: Sonny Jiang <sonny.jiang@amd.com>
Reviewed-by: James Zhu <James.Zhu@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

Merge tag 'drm-misc-next-fixes-2022-07-14' of git://anongit.freedesktop.org/drm/drm-misc into drm-next

Short summary of fixes:

 - dma-buf: revert change to fence handling
 - mgag200: fix PCI register initialization

Signed-off-by: Dave Airlie <airlied@redhat.com>
From: Thomas Zimmermann <tzimmermann@suse.de>
Link: https://patchwork.freedesktop.org/patch/msgid/YtAjgcWC1zkNOGWa@linux-uq9g

Merge tag 'amd-drm-next-5.20-2022-07-14' of https://gitlab.freedesktop.org/agd5f/linux into drm-next

amd-drm-next-5.20-2022-07-14:

amdgpu:
- DCN3.2 updates
- DC SubVP support
- DP MST fixes
- Audio fixes
- DC code cleanup
- SMU13 updates
- Adjust GART size on newer APUs for S/G display
- Soft reset for GFX 11
- Soft reset for SDMA 6
- Add gfxoff status query for vangogh
- Improve BO domain pinning
- Fix timestamps for cursor only commits
- MES fixes
- DCN 3.1.4 support
- Misc fixes
- Misc code cleanup

amdkfd:
- Simplify GPUVM validation
- Unified memory for CWSR save/restore area
- fix possible list corruption on queue failure

radeon:
- Fix bogus power of two warning

UAPI:
- Unified memory for CWSR save/restore area for KFD
  Proposed userspace: https://lists.freedesktop.org/archives/amd-gfx/2022-June/080952.html

Signed-off-by: Dave Airlie <airlied@redhat.com>
From: Alex Deucher <alexander.deucher@amd.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220714214716.8203-1-alexander.deucher@amd.com

drm/mgag200: Don't read-back PCI option register before writing

Remove the read operation from mgag200_init_pci_options(). It was
incorrectly added while refactoring the code. Reading the PCI option
register clears the register's new value and subsequently leads to
re-writing the old value.

Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
Reviewed-by: Jocelyn Falempe <jfalempe@redhat.com>
Fixes: ce19021fd99a ("drm/mgag200: Move PCI-option setup into model-specific code")
Cc: Thomas Zimmermann <tzimmermann@suse.de>
Cc: Jocelyn Falempe <jfalempe@redhat.com>
Cc: Dave Airlie <airlied@redhat.com>
Cc: dri-devel@lists.freedesktop.org
Link: https://patchwork.freedesktop.org/patch/msgid/20220708072114.13960-1-tzimmermann@suse.de

dma-buf: revert "return only unsignaled fences in dma_fence_unwrap_for_each v3"

This reverts commit 8f61973718485f3e89bc4f408f929048b7b47c83.

It turned out that this is not correct. Especially the sync_file info
IOCTL needs to see even signaled fences to correctly report back their
status to userspace.

Instead add the filter in the merge function again where it makes sense.

Signed-off-by: Christian König <christian.koenig@amd.com>
Tested-by: Karolina Drobnik <karolina.drobnik@intel.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20220712102849.1562-1-christian.koenig@amd.com

Merge drm/drm-next into drm-misc-next-fixes

Backmerging from drm/drm-next for the final fixes that will go
into v5.20.

Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>

drm/amd/display: remove duplicate dcn314 includes

Several headers were included twice.  Fix that.

Reported-by: kernel test robot <yujie.liu@intel.com>
Reviewed-by: Harry Wentland <harry.wentland@amd.com>
Reviewed-by: André Almeida <andrealmeid@igalia.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Enable DCN314 in DM

Add support for DCN 3.1.4 in Display Manager

Signed-off-by: Roman Li <roman.li@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Add DMUB support for DCN314

Initialize DMUB for DCN 3.1.4.
Use same funcs as DCN31.

Signed-off-by: Roman Li <roman.li@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Enable DCN314 in DC

Add support for DCN 3.1.4 in Display Core

Signed-off-by: Roman Li <roman.li@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Add DCN314 version identifiers

DCN 3.1.4 version and family ids

Signed-off-by: Roman Li <roman.li@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Add DCN314 DML calculation support

Display mode library for DCN 3.1.4

v2: squash in checkpatch fix (Alex)

Signed-off-by: Roman Li <roman.li@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Add DCN314 DC resources

Display Core support for DCN 3.1.4

v2:(squash)fix non-x86 in dc/dcn314/Makefile
Properly handle PPC as well. (Alex)
v3: minor cleanup (Alex)
v4: fix comment (Alex)

Signed-off-by: Roman Li <roman.li@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Add DCN314 clock manager

Clock and SMU interfaces for DCN 3.1.4

Signed-off-by: Roman Li <roman.li@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Add DCN314 IRQ services

IRQ services to support DCN 3.1.4 interrupts.

v2: make to_dal_irq_source_dcn314 static (Alex)

Signed-off-by: Roman Li <roman.li@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: Add reg headers for DCN314

Register headers for the following IPs:
- DCN  3.1.4
- DPCS 3.1.4

v2:(squash) clean up (Alex)

Signed-off-by: Roman Li <roman.li@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Ensure valid event timestamp for cursor-only commits

Requires enabling the vblank machinery for them.

Bug: https://gitlab.freedesktop.org/drm/amd/-/issues/2030
Acked-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Harry Wentland <harry.wentland@amd.com>
Signed-off-by: Michel Dänzer <mdaenzer@redhat.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: Check BO's requested pinning domains against its preferred_domains

When pinning a buffer, we should check to see if there are any
additional restrictions imposed by bo->preferred_domains. This will
prevent the BO from being moved to an invalid domain when pinning.

For example, this can happen if the user requests to create a BO in GTT
domain for display scanout. amdgpu_dm will allow pinning to either VRAM
or GTT domains, since DCN can scanout from either or. However, in
amdgpu_bo_pin_restricted(), pinning to VRAM is preferred if there is
adequate carveout. This can lead to pinning to VRAM despite the user
requesting GTT placement for the BO.

v2: Allow the kernel to override the domain, which can happen when
    exporting a BO to a V4L camera (for example).

Signed-off-by: Leo Li <sunpeng.li@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Cc: stable@vger.kernel.org

drm/amd/display: attempt to fix the logic in commit_planes_for_stream()

The indentation is screwed up.  I'm not sure quite how the logic
should flow.  Someone more familiar with this code should
verify this.

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

drm/amd/pm: Implement get GFXOFF status for vangogh

Implement function to get current GFXOFF status for vangogh.

Signed-off-by: André Almeida <andrealmeid@igalia.com>
Acked-by: Evan Quan <evan.quan@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: correct check of coverage blend mode

Check the value of per_pixel_alpha to decide whether the Coverage pixel
blend mode is applicable or not.

Fixes: 76818cdd11a2 ("drm/amd/display: add Coverage blend mode for overlay plane")
Reported-by: kernel test robot <lkp@intel.com>
Reported-by: Dan Carpenter <dan.carpenter@oracle.com>
Reviewed-by: Harry Wentland <harry.wentland@amd.com>
Signed-off-by: Melissa Wen <mwen@igalia.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/pm: Prevent divide by zero

divide error: 0000 [#1] SMP PTI
CPU: 3 PID: 78925 Comm: tee Not tainted 5.15.50-1-lts #1
Hardware name: MSI MS-7A59/Z270 SLI PLUS (MS-7A59), BIOS 1.90 01/30/2018
RIP: 0010:smu_v11_0_set_fan_speed_rpm+0x11/0x110 [amdgpu]

Speed is user-configurable through a file.
I accidentally set it to zero, and the driver crashed.

Reviewed-by: Evan Quan <evan.quan@amd.com>
Reviewed-by: André Almeida <andrealmeid@igalia.com>
Signed-off-by: Yefim Barashkin <mr.b34r@kolabnow.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Only use depth 36 bpp linebuffers on DCN display engines.

Various DCE versions had trouble with 36 bpp lb depth, requiring fixes,
last time in commit 353ca0fa5630 ("drm/amd/display: Fix 10bit 4K display
on CIK GPUs") for DCE-8. So far >= DCE-11.2 was considered ok, but now I
found out that on DCE-11.2 it causes dithering when there shouldn't be
any, so identity pixel passthrough with identity gamma LUTs doesn't work
when it should. This breaks various important neuroscience applications,
as reported to me by scientific users of Polaris cards under Ubuntu 22.04
with Linux 5.15, and confirmed by testing it myself on DCE-11.2.

Lets only use depth 36 for DCN engines, where my testing showed that it
is both necessary for high color precision output, e.g., RGBA16 fb's,
and not harmful, as far as more than one year in real-world use showed.

DCE engines seem to work fine for high precision output at 30 bpp, so
this ("famous last words") depth 30 should hopefully fix all known problems
without introducing new ones.

Successfully retested on DCE-11.2 Polaris and DCN-1.0 Raven Ridge on
top of Linux 5.19.0-rc2 + drm-next.

Fixes: 353ca0fa5630 ("drm/amd/display: Fix 10bit 4K display on CIK GPUs")
Signed-off-by: Mario Kleiner <mario.kleiner.de@gmail.com>
Tested-by: Mario Kleiner <mario.kleiner.de@gmail.com>
Cc: stable@vger.kernel.org # 5.14.0
Cc: Alex Deucher <alexander.deucher@amd.com>
Cc: Harry Wentland <harry.wentland@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: make some dc_dmub_srv functions static

Not used outside of dc_dmub_srv.c.

Reported-by: kernel test robot <lkp@intel.com>
Reviewed-by: André Almeida <andrealmeid@igalia.com>
Reviewed-by: Harry Wentland <harry.wentland@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: fix file permissions on some files

Drop execute.

Bug: https://gitlab.freedesktop.org/drm/amd/-/issues/2085
Reviewed-by: Guchun Chen <guchun.chen@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: Fix acronym typo in glossary

The initialism of RunList Controller is RLC, not RCL

Signed-off-by: Kent Russell <kent.russell@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: 3.2.194

This version brings along following fixes:

- Fixes for MST, MPO, PSRSU, DP 2.0, Freesync and others
- Add register offsets of NBI and DCN.
- Improvement of ALPM
- Removing assert statement for Linux DM
- Re-implementing ARGB16161616 pixel format

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

drm/amd/display: Fix lag when moving windowed MPO across display using ODM 2:1 combine

[Why]
With single display odm 2:1 policy, when moving windowed MPO across
 the display, we experience a momentary lag when we move between the
 centre of the display and the right half of the display.  This is
 caused by the MPO pipe being reallocated when it crosses this
 boundary

[How]
Handle two cases:
1. if the head pipe has a MPO pipe already allocated in the old
 context, then use that pipe if it is available in the current
 context
2. if the head pipe is on the left side, check the right side to
 see if it has a MPO pipe already allocated.  If so, don't use
 that pipe if it is selected as the idle pipe in the current
 context
Add new function pointer called .acquire_idle_pipe_for_head_pipe
 that will pass in the head pipe and handle case 1
Add find_idle_secondary_pipe_check_mpo() to handle case 2
 if we don't hit case 1.

In dc_add_plane_to_context(), start with head pipe and check
 case 1 and 2 in call acquire_free_pipe_for_head().
If we are on the right side of the display, check case 1
 again by passing in right side pipe as the new head in
 call acquire_free_pipe_for_head().

Reviewed-by: Dmytro Laktyushkin <Dmytro.Laktyushkin@amd.com>
Reviewed-by: Ariel Bernstein <Eric.Bernstein@amd.com>
Acked-by: Solomon Chiu <solomon.chiu@amd.com>
Signed-off-by: Samson Tam <Samson.Tam@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Add NBIO reg offsets to DC

[Why&How]
Add a field to store the NBIO IP offset for use with runtime offset
calculation

Reviewed-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Acked-by: Solomon Chiu <solomon.chiu@amd.com>
Signed-off-by: Aurabindo Pillai <aurabindo.pillai@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Add DCN reg offsets to DC

[Why&How]
Add a field to store the DCN IP offset for use with runtime offset
calculation

This offset is indexed using reg*_BASE_IDX for the corresponding
group of registers. For example, address of DIG_BE_CNTL instance 0 is
calculated like: dcn_reg_offsets[regDIG0_DIG_BE_CNTL_BASE_IDX] +
regDIG0_DIG_BE_CNTL.

{dcn,nbio}_reg_offsets are used only for the ASICs for which runtime
initializaion of offsets are enabled through the modified SR* macros
that contain an additional REG_STRUCT element in the macro definition.

DCN3.5+ will fail dc_create() if {dcn,nbio}_reg_offsets are null. They
are applicable starting with DCN32/321 and are not used for ASICs
upstreamed before them. ASICs before DCN32/321 will not contain any
computation that involves {dcn,nbio}_reg_offsets. For them, the
address/offset computation is done during compile time.

This is evident from the BASE_INNER definition for compile time vs run
time initialization:

Compile time init: #define BASE_INNER(seg) DCN_BASE__INST0_SEG ## seg
Run time init:     #define BASE_INNER(seg) ctx->dcn_reg_offsets[seg]

BASE_INNER macro is local to each dcnxx_resource.c and hence different
ASICs can have either runtime or compile time initialization of offsets.

The computation of offset is done for registers all at once during
driver load and hence it does not introduce any performance overhead
during normal operation.

Reviewed-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Acked-by: Solomon Chiu <solomon.chiu@amd.com>
Signed-off-by: Harry Wentland <harry.wentland@amd.com>
Signed-off-by: Aurabindo Pillai <aurabindo.pillai@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: add system info table log

[why]
insert log for debug use.

v2: squash in fix (Alex)

Reviewed-by: Nicholas Kazlauskas <Nicholas.Kazlauskas@amd.com>
Acked-by: Solomon Chiu <solomon.chiu@amd.com>
Signed-off-by: Charlene Liu <Charlene.Liu@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Grab dc_lock before detecting link

[Why & How]
There is chance we change dc state while calling dc_link_detect().
As the result of that, grab the dm.dc_lock before detecting link.

Reviewed-by: Hersen Wu <hersen.wu@amd.com>
Acked-by: Solomon Chiu <solomon.chiu@amd.com>
Signed-off-by: Wayne Lin <Wayne.Lin@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Re-implementing ARGB16161616 pixel format as 22

[Why]
ABGR16161616 colour format was added to dcn10/20/30, and set
any ARGB16161616 to the same value as it (26). As such, the
HDR10 Green Point y value was too far off of the EDID stated
value for DisplayPort.

[How]
Added back the pixel format as 22 for ARGB16161616 for
dcn10/20/30.

Reviewed-by: Reza Amini <reza.amini@amd.com>
Acked-by: Solomon Chiu <solomon.chiu@amd.com>
Signed-off-by: Ethan Wellenreiter <Ethan.Wellenreiter@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: 3.2.193

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

drm/amd/display: update DML1 logic for unbounded req handling

[why]
Unbounded request logic in resource/DML has some issues where
unbounded request is being enabled incorrectly.  SW today enables
unbounded request unconditionally in hardware, on the assumption
that HW can always support it in single pipe scenarios.

This worked until now because the same assumption is made in DML.
A new DML update is needed to fix a bug, where there are single
pipe scenarios where unbounded cannot be enabled, and this change
in DML needs to be ported in, and dcn32 resource logic fixed.

[how]
First, dcn32_resource should program unbounded req in HW according
to unbounded req enablement output from DML, as opposed to DML input

Second, port in DML1 update which disables unbounded req in some
scenarios to fix an issue with poor stutter performance

Reviewed-by: Nevenko Stupar <Nevenko.Stupar@amd.com>
Reviewed-by: Dmytro Laktyushkin <Dmytro.Laktyushkin@amd.com>
Acked-by: Solomon Chiu <solomon.chiu@amd.com>
Signed-off-by: Jun Lei <jun.lei@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu/gfx11: add aggregated doorbell support

Port aggregated doorbell support to gfx11.

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

drm/amdgpu/sdma6: add aggregated doorbell support

Port aggregated doorbell support to sdma6.

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

drm/amdgpu/mes: ring aggregatged doorbell when mes queue is unmapped

Ring aggregated doorbel to make unmapped queue scheduled in mes firmware.

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

drm/amdgpu/mes11: initialize aggregated doorbell

Allocate and enable aggregated doorbell.

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

drm/amdgpu/mes: init aggregated doorbell

Allocate and enable aggregated doorbell.

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