drm/amdgpu: add psp block for navy_flounder

Add psp and smu block for navy_flounder with
psp firmware load type.

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

drm/amdgpu: add psp support for navy_flounder

Currently skip ASD FW loading and ih reroute per
sienna_cichlid.

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

drm/amdgpu: add smu block for navy_flounder

Add SMU block for navy_flounder with direct
firmware load type.

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

drm/amdgpu/powerplay: add smu support for navy_flounder

Now navy_flounder will reuse the smu11 driver_if header and ppt
functions for sienna_cichlid. Later navy_flounder can maintain
its own version if the compatibility is broken.

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

drm/amdgpu: add gmc cg support for navy_flounder

The athub version used for navy_flounder is v2.1.

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

drm/amdgpu: force pa_sc_tile_steering_override to 0 for navy_flounder

pa_sc_tile_steering_override is only programmable for
gfx10.0/10.1/10.2, and navy_flounder has the same gfx10.3 IP
with sienna_cichlid.

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

drm/amdgpu: configure navy_flounder gfx according to gfx 10.3

The gfx version of navy_flounder is 10.3, identical to
sienna_cichlid, follow the way of sienna_cichlid.

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

drm/amdgpu: add virtual display support for navy_flounder.

Virtual display support for bring up and virtualization.

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

drm/amdgpu: add sdma ip block for navy_flounder

Navy_Flounder has the same sdma IP version with
sienna_cichlid, and it has 2 sdma controllers.

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

drm/amdgpu: add gfx ip block for navy_flounder

since navy_flounder has similar gc IP version with
sienna_cichlid, follow its setting for the moment.

Signed-off-by: Jiansong Chen <Jiansong.Chen@amd.com>
Reviewed-by: Tao Zhou <Tao.Zhou1@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: add ih ip block for navy_flounder

navy_flounder has the same osssys IP verison with
sienna_cichlid, follow its setting.

Signed-off-by: Jiansong Chen <Jiansong.Chen@amd.com>
Reviewed-by: Tao Zhou <Tao.Zhou1@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: add gmc ip block for navy_flounder

navy_flounder has similar gc IP version with sienna_cichlid,
follow its setting for the moment.

Signed-off-by: Jiansong Chen <Jiansong.Chen@amd.com>
Reviewed-by: Tao Zhou <Tao.Zhou1@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: add common ip block for navy_flounder

Signed-off-by: Jiansong Chen <Jiansong.Chen@amd.com>
Reviewed-by: Tao Zhou <Tao.Zhou1@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: add support on mmhub for navy_flounder

navy_flounder has the same mmhub IP version with sienna_cichlid,
follow its setting.

Signed-off-by: Jiansong Chen <Jiansong.Chen@amd.com>
Reviewed-by: Tao Zhou <Tao.Zhou1@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: initialize IP offset for navy_flounder

since navy_flounder has the same ip offset with sienna_cichlid,
follow sienna_cichlid setting for the moment.

Signed-off-by: Jiansong Chen <Jiansong.Chen@amd.com>
Reviewed-by: Tao Zhou <Tao.Zhou1@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu/soc15: add support for navy_flounder

Add soc support.

Signed-off-by: Jiansong Chen <Jiansong.Chen@amd.com>
Reviewed-by: Tao Zhou <Tao.Zhou1@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu/gfx10: add clockgating support for navy_flounder

Same as navi10.

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

drm/amdgpu/gmc10: add navy_flounder support

Same as navi10.

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

drm/amdgpu/gfx10: add support for navy_flounder firmware

Declare the gfx/compute firmwares.

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

drm/amdgpu: set asic family and ip blocks for navy_flounder

Add the asic family and IP blocks for navy flounder.

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

drm/amdgpu: set fw load type for navy_flounder

Currently navy_flounder only supports backdoor loading type.
Will switch to psp load type when psp is ready.

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

drm/amdgpu: add navy_flounder gpu info firmware

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

drm/amdgpu: add navy_flounder asic type

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

drm/amdgpu: expand to add multiple trap event irq id

Sienna_cichlid has four sdma instances, but other chips don't.
So we need expand to add multiple trap event irq id in sdma
v5.2.

Signed-off-by: Huang Rui <ray.huang@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>

drm/amd/sriov skip vcn powergating and dec_ring_test

1.Skip decode_ring test in VF, because VCN in SRIOV does not
support direct register read/write.

2.Skip powergating configuration in hw fini because
VCN3.0 SRIOV doesn't support powergating.

V2: delete unneccessary white lines and refine implementation.

Signed-off-by: Jack Zhang <Jack.Zhang1@amd.com>
Reviewed-by: Leo Liu <leo.liu@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: correct ta header v2 ucode init start address

resolve bug calculating fw start address within binary

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

drm/amd/sriov porting sriov cap to vcn3.0

1.In early_init and for sriov, hardcode
  harvest_config=0, enc_num=1

2.sw_init/fini
  alloc & free mm_table for sriov
  doorbell setting for sriov

3.hw_init/fini
  Under sriov, add start_sriov to config mmsch
  Skip ring_test to avoid mmio in VF, but need to initialize wptr for vcn rings.

4.Implementation for vcn_v3_0_start_sriov

V2:Clean-up some uneccessary funciton declaration.

Signed-off-by: Jack Zhang <Jack.Zhang1@amd.com>
Reviewed-by: Leo Liu <leo.liu@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/sriov add mmsch_v3 interface

For VCN3.0 SRIOV, Guest driver needs to communicate with mmsch
to set the World Switch for MM appropriately. This patch add
the interface for mmsch_v3.0.

Signed-off-by: Jack Zhang <Jack.Zhang1@amd.com>
Reviewed-by: Dennis Li <Dennis.Li@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: optimize rlcg write for gfx_v10

For gfx10 boards, except for nv12, other boards take mmio write
rather than rlcg write

Signed-off-by: Jack Zhang <Jack.Zhang1@amd.com>
Acked-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/sriov skip jped ip block and close pgcg flags

For SIENNA_CICHLID SRIOV, jpeg and pgcp is not supported.

Signed-off-by: Jack Zhang <Jack.Zhang1@amd.com>
Acked-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/powerplay: drop unused code around thermal range setting

Leftover of previous cleanups.

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

drm/amd/powerplay: maximum the code sharing on thermal irq setting

Put the common code in smu_v11_0.c instead of having one copy each.

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

drm/amd/powerplay: sort the call flow on temperature ranges retrieving

This can help to maintain clear code layer.

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

drm/amd/powerplay: cache the software_shutdown_temp

As it's needed in the succeeding thermal irq setting.

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

drm/amd/powerplay: correct Sienna Cichlid temperature limit settings

These are needed for temp1/2/3 related hwmon interfaces.

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

drm/amd/powerplay: correct Navi1X temperature limit settings

These are needed for temp1/2/3 related hwmon interfaces.

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

drm/amd/powerplay: correct the supported pcie GenSpeed and LaneCount

The LCLK dpm table setup should be performed in .update_pcie_parameters().
Otherwise, the updated GenSpeed and LaneCount information will be lost.

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

drm/amd/powerplay: drop unnecessary wrapper around pcie parameters setting

This can also help to maintain clear code layer.

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

drm/amd/powerplay: drop unused APIs and parameters

Leftover of previous performance level setting cleanups.

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

drm/amd/powerplay: drop smu_v12_0.c unnecessary wrapper V2

By moving the implemention to renoir_ppt.c considering
it's really ASIC specific.

V2: fix compile warnings below
drivers/gpu/drm/amd/amdgpu/../powerplay/renoir_ppt.h:40:25: warning: array subscript is above array bounds [-Warray-bounds]
    freq = table->FClocks[dpm_level].Freq; \
drivers/gpu/drm/amd/amdgpu/../powerplay/renoir_ppt.c:195:2: note: in expansion of macro ‘GET_DPM_CUR_FREQ’
  GET_DPM_CUR_FREQ(clk_table, clk_type, dpm_level, *freq);
  ^~~~~~~~~~~~~~~~
drivers/gpu/drm/amd/amdgpu/../powerplay/renoir_ppt.h:46:25: warning: array subscript is above array bounds [-Warray-bounds]
    freq = table->FClocks[dpm_level].Freq;  \
drivers/gpu/drm/amd/amdgpu/../powerplay/renoir_ppt.c:195:2: note: in expansion of macro ‘GET_DPM_CUR_FREQ’
  GET_DPM_CUR_FREQ(clk_table, clk_type, dpm_level, *freq);

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

drm/amd/powerplay: drop unnecessary wrappers

By calling the target APIs directly.

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

drm/amd/powerplay: apply gfxoff disablement/enablement for all SMU11 ASICs

Before and after setting gfx clock soft max/min frequency.

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

drm/amd/powerplay: drop Sienna Cichlid specific set_soft_freq_limited_range

Use the common smu_v11_0_set_soft_freq_limited_range.

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

drm/amd/powerplay: drop unnecessary Sienna Cichlid specific APIs

As a common performance level setting API is used. Then these
ASIC specific APIs are not needed any more.

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

drm/amd/powerplay: drop unnecessary Navi1x specific APIs

As a common performance level setting API is used. Then these
ASIC specific APIs are not needed any more.

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

drm/amd/powerplay: drop unnecessary Arcturus specific APIs

As a common performance level setting API is used. Then these
ASIC specific APIs are not needed any more.

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

drm/amd/powerplay: update the common API for performance level setting

So that it can be more widely shared around SMU v11 ASICs.

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

drm/amd/powerplay: update UMD pstate clock settings

Preparing for coming code sharing around performance level
setting.

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

drm/amd/powerplay: add new UMD pstate data structure

This is used to cache the clock frequencies for all UMD pstates.
So that we do not need to calculate from scratch on every UMD
pstate switch.

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

drm/amd/powerplay: update Sienna Cichlid default dpm table setup

Cache all clocks levels for every dpm table. They are needed
by other APIs.

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

drm/amd/powerplay: update Navi10 default dpm table setup

Cache all clocks levels for every dpm table. They are needed
by other APIs.

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

drm/amd/powerplay: update Arcturus default dpm table setting

Preparing for coming code sharing around performance level
setting.

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

drm/amd/powerplay: implement a common set dpm table API for smu V11

Maximum the code sharing around smu V11.

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

drm/amd/powerplay: add more members for dpm table

These members can help to cache the clock frequencies for all
dpm levels. Then simplifying the code for dpm level switching
is possible.

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

drm/amdgpu: add module parameter choose reset mode

Default value is auto, doesn't change
original reset method logic.

v2: change to use parameter reset_method
v3: add warn msg if specified mode isn't supported

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

drm/amdgpu: enable mode1 reset

For sienna cichlid, add mode1 reset path for sGPU.

v2: hiding MP0/MP1 mode1 reset under AMD_RESET_METHOD_MODE1
v3: split emergency restart logic to a new patch

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

drm/amdgpu: RAS emergency restart logic refine

If we are in RAS triggered situation and
BACO isn't support, emergency restart is needed,
and this code is only needed for some specific
cases(vega20 with given smu fw version).

After we add smu mode1 reset for sienna cichlid, we
need to share AMD_RESET_METHOD_MODE1 with psp mode1 reset,
so in amdgpu_device_gpu_recover, we need differentiate
which mode1 reset we are using, then decide if it's
a full reset and then decide if emergency restart is needed,
the logic will become much more complex.

After discussion with Hawking, move emergency restart logic
to an independent function.

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

drm/amd/powerplay: add SMU mode1 reset

From PM FW 58.26.0 for sienna cichlid, SMU mode1 reset
is support, driver sends PPSMC_MSG_Mode1Reset message
to PM FW could trigger this reset.

v2: add mode1 reset dpm interface
v3: change maro name

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

drm/amdkfd: fix kernel-doc and cleanup

 - fix some styling issues
 - fixes for kernel-doc type

Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Rajneesh Bhardwaj <rajneesh.bhardwaj@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: remove an unnecessary NULL check

The "stream" pointer can't be NULL and it's dereferenced on the line
before so the check doesn't make sense.

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

drm/amdgpu/sdma5: fix wptr overwritten in ->get_wptr()

"u64 *wptr" points to the the wptr value in write back buffer and
"*wptr = (*wptr) >> 2;" results in the value being overwritten each time
when ->get_wptr() is called.

umr uses /sys/kernel/debug/dri/0/amdgpu_ring_sdma0 to get rptr/wptr and
decode ring content and it is affected by this issue.

fix and simplify the logic similar as sdma_v4_0_ring_get_wptr().

v2: fix for sdma5.2 as well

Suggested-by: Le Ma <le.ma@amd.com>
Signed-off-by: Xiaojie Yuan <xiaojie.yuan@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: use ARRAY_SIZE() to add amdgpu debugfs files

to easily add new debugfs file w/o changing the hardcoded list count.

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

drm/amdgpu/powerplay: Modify SMC message name for setting power profile mode

I consulted Cai Land(Chuntian.Cai@amd.com), he told me corresponding smc
message name to fSMC_MSG_SetWorkloadMask() is
"PPSMC_MSG_ActiveProcessNotify" in firmware code of Renoir.

Strange though it may seem, but it's a fact.

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

drm/amd/display: Allow for vblank enabled with no active planes

[Why]
CRC capture doesn't work when the active plane count is 0 since we
currently tie both vblank and pageflip interrupts to active_plane_count
greater than 0.

[How]
The frontend is what generates the vblank interrupts while the backend
is what generates pageflip interrupts. Both have a requirement for
the CRTC to be active, so control the overall interrupt state based
on that instead.

Pageflip interrupts need to be enabled based on active plane count, but
we actually rely on power gating to take care of disabling the interrupt
for us on pipes that can be power gated.

For pipes that can't be power gated it's still fine to leave it enabled
since the interrupt only triggers after the address has been written
to that particular pipe - which we won't be doing without an active
plane.

The issue we had before with this setup was that we couldn't force
the state back on. We were essentially manipulating the refcount
to enable or disable as needed in a two pass approach.

However, there is a function that solves this problem more elegantly:
amdgpu_irq_update() will unconditionally call the set based on what it
thinks the current enablement state is.

This leaves two future TODO items for our IRQ handling:
- Disabling IRQs in commit tail instead of atomic commit
- Mapping the pageflip interrupt to VUPDATE or something that's tied to
  the frontend instead of the backend since the mapping to CRTC is not
  correct

Signed-off-by: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com>
Reviewed-by: Bhawanpreet Lakha <Bhawanpreet.Lakha@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Add missing DCN30 registers and fields for OTG_CRC_CNTL2

[Why]
When enabling the debugfs for CRC capture we hit assertions caused by
register address and field masks and shifts missing.

[How]
We want these registers programmed, so add in the SRI/SF entries for
this register.

Signed-off-by: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com>
Reviewed-by: Bhawanpreet Lakha <Bhawanpreet.Lakha@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: handle failed allocation during stream construction

[Why]
Failing to allocate a transfer function during stream construction leads
to a null pointer dereference

[How]
Handle the failed allocation by failing the stream construction

Cc: stable@vger.kernel.org
Signed-off-by: Josip Pavic <Josip.Pavic@amd.com>
Reviewed-by: Aric Cyr <Aric.Cyr@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Reuse parsing code of debugfs write buffer

[why]
Move code for parsing debugfs input into an array of int parameters by
specifying the max number of expected parameters

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

drm/amd/display: 3.2.94

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

drm/amd/display: [FW Promotion] Release 0.0.24

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

drm/amd/display: fix dcn3 p_state_change_support validation (v2)

Our validation is a known mess with actual validation mixed with
topology configuration. This change makes sure topolgical validation is
completed before any topology changes are made so we do not run into
issues where we merge and split a pipe over the course of a single call.

v2: Squash in compilation fix

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

drm/amd/display: p-state warning occurs while changing resolution from 120hz to 60hz

[Why]
new calculated dispclk, dppclk are stored in
context->bw_ctx.bw.dcn.clk.dispclk_khz, dppclk_khz. Current dispclk,
dppclk are from dc->clk_mgr->clks.dispclk_khz. dcn_validate_bandwidth
compute new dispclk, dppclk. dispclk will put in use after
optimize_bandwidth when ramp_up_dispclk_with_dpp is called. There are
two places for dppclk be put in use. One location is the same as the
location as dispclk. Another is within update_dchubp_dpp which happens
between pre_bandwidth and optimize_bandwidth. dppclk updated within
update_dchubp_dpp will cause new clock values of dispclk and dppclk not
be in use at the same time. when clocks are decreased, this may cause
dppclk is lower than current configuration and let pipe stuck. for
example, eDP + external dp, change resolution of DP from 1920x1080x144hz
to 1280x960x60hz.

before change: dispclk = 337889 dppclk = 337889
change mode, dcn_validate_bandwidth calculate
             dispclk = 143122 dppclk = 143122
update_dchubp_dpp be executed before dispclk be updated,
dispclk = 337889, but dppclk use new value dispclk /2 =
168944. this will cause pipe pstate warning issue.

[How]
between pre_bandwidth and optimize_bandwidth, while dispclk is going to
be decreased, keep dppclk = dispclk

Signed-off-by: Hersen Wu <hersenxs.wu@amd.com>
Reviewed-by: Nicholas Kazlauskas <Nicholas.Kazlauskas@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: OLED panel backlight adjust not work with external display connected

[Why]
amdgpu_dm->backlight_caps is for single eDP only. the caps are upddated
for very connector. Real eDP caps will be overwritten by other external
display. For OLED panel, caps->aux_support is set to 1 for OLED pnael.
after external connected, caps+.aux_support is set to 0. This causes
OLED backlight adjustment not work.

[How]
within update_conector_ext_caps, backlight caps will be updated only for
eDP connector.

Cc: stable@vger.kernel.org
Signed-off-by: hersen wu <hersenxs.wu@amd.com>
Reviewed-by: Nicholas Kazlauskas <Nicholas.Kazlauskas@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: reduce sr_xxx_time by 3 us when ppt disable

[Why]
when ppt disabled, the watermark doesn't get fine tune causing
underflow.

[How]
It is a temporary solution to reduce sr_xxx_time by 3 us when ppt
disable.

Signed-off-by: Chiawen Huang <chiawen.huang@amd.com>
Reviewed-by: Tony Cheng <Tony.Cheng@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Power down hardware if set mode is not called before timeout

[WHY]
In headless systems, if set mode is not called, hardware will not be
powered down on boot, causing HW/SW discrepancies.  Powering down
hardware on boot will ensure SW state is accurate.

[HOW]
Set a timer callback on boot for 10 seconds. If set mode is not called
within that time, power down hardware. Otherwise, do not power down.

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

drm/amd/display: update dml var

Update *DynamicMetadata variables for providing more flexibility.

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

drm/amdgpu/powerplay: Target power profile mode should be the second parameter of renoir_set_power_profile_mode

A small mistake

Signed-off-by: chen gong <curry.gong@amd.com>
Acked-by: Evan Quan <evan.quan@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu/display: create fake mst encoders ahead of time (v4)

Prevents a warning in the MST create connector case.

v2: create global fake encoders rather per connector fake encoders
to avoid running out of encoder indices.

v3: use the actual number of crtcs on the asic rather than the max
to conserve encoders.

v4: v3 plus missing hunk I forgot to git add.

Bug: https://gitlab.freedesktop.org/drm/amd/-/issues/1108
Fixes: c6385e503aeaf9 ("drm/amdgpu: drop legacy drm load and unload callbacks")
Reviewed-by: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: fix spelling mistake "Falied" -> "Failed"

There is a spelling mistake in a DRM_ERROR error message. Fix it.

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

drm/amdgpu: fix preemption unit test

Remove signaled jobs from job list and ensure the
job was indeed preempted.

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/gfx10: fix race condition for kiq

During preemption test for gfx10, it uses kiq to trigger
gfx preemption, which would result in race condition
with flushing TLB for kiq.

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

drm/amdgpu: minor cleanup of phase1 suspend code

Cleanup of phase1 suspend code to reduce unnecessary indentation.

Signed-off-by: Nirmoy Das <nirmoy.das@amd.com>
Acked-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/powerplay: put dpm frequency setting common code in smu_v11_0.c

As designed the common code shared among all smu v11 ASCIs go to
smu_v11_0.c. This helps to maintain clear code layers.

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

drm/amd/powerplay: revise calling chain on retrieving frequency range

This helps to maintain clear code layers and drop unnecessary
parameter.

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

drm/amd/powerplay: revise calling chain on setting soft limit

This helps to maintain clear code layers and drop unnecessary
parameter.

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

drm/amd/powerplay: put setting hard limit common code in smu_v11_0.c

As designed the common code shared among all smu v11 ASCIs go to
smu_v11_0.c. This helps to maintain clear code layers.

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

drm/amd/powerplay: drop unused code and wrapper around clock retrieving

Clean code by dropping unnecessary ones.

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

drm/amd/powerplay: unshare the code for retrieving current clock frequency

By moving them into asic specific file. Since they are really ASIC
specific.

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

drm/amd/powerplay: bypass wrapper on retrieving current clock frequency

By calling the target API(asic specific) directly.

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

drm/amd/powerplay: drop unnecessary wrappers around clock retrieving

By calling the target APIs directly.

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

drm/amd/powerplay: update driver if file for sienna_cichlid

Update sienna_cichlid driver if header and related files.

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

drm/amd/powerplay: correct the .get_workload_type() pointer

This seemed a typo.

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

Revert "drm/amd/display: add mechanism to skip DCN init"

To avoid s3 faild at the first cycle on renoir platform, it needs to
revert this patch:
drm/amd/display: add mechanism to skip DCN init

Signed-off-by: changfeng <Changfeng.Zhu@amd.com>
Ackedy-by: Huang Rui <ray.huang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Fix CSC remap matrix not being applied on dcn30

[Why]
DCN3 has two gamut remap matrices. When using CSC adjustment the CM
remap is set to bypass and MPCC remap is used. However to bypass CM
some state in the context is modified and not restored correctly
resulting in subsequent calls to disable MPCC remap as well.

[How]
Fix logic for save/restore of remap enable flag when programming MPCC
remap matrix.

Signed-off-by: Aric Cyr <aric.cyr@amd.com>
Signed-off-by: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com>
Reviewed-by: Bhawanpreet Lakha <Bhawanpreet.Lakha@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Use VUPDATE_NO_LOCK instead of VUPDATE for dcn30

[Why]
Soft hangs occur when FreeSync is engaged since we utilize VUPDATE
(which doesn't fire when holding the pipe lock) to send back vblank
events when FreeSync is active.

[How]
The alternative (working) interrupt source for this mechanism is
VUPDATE_NO_LOCK. We already use this all other DCN revisions so align
dcn30 with those as well.

Signed-off-by: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com>
Reviewed-by: Bhawanpreet Lakha <Bhawanpreet.Lakha@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: enable gpu recovery for sienna cichlid

Enable gpu recovery for sienna cichlid by default to trigger
gpu recovery once needed.

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: remove unnecessary logic of ASIC check

Remove some unused ASIC check logic.
Remove some definition of amdgpu_device which only used by
the removed ASIC check logic.(V2)

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: don't do soft recovery if gpu_recovery=0

It's impossible to debug shader hangs with soft recovery.

Signed-off-by: Marek Olšák <marek.olsak@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>

drm/radeon: fix double free

clang static analysis flags this error

drivers/gpu/drm/radeon/ci_dpm.c:5652:9: warning: Use of memory after it is freed [unix.Malloc]
                kfree(rdev->pm.dpm.ps[i].ps_priv);
                      ^~~~~~~~~~~~~~~~~~~~~~~~~~
drivers/gpu/drm/radeon/ci_dpm.c:5654:2: warning: Attempt to free released memory [unix.Malloc]
        kfree(rdev->pm.dpm.ps);
        ^~~~~~~~~~~~~~~~~~~~~~

problem is reported in ci_dpm_fini, with these code blocks.

for (i = 0; i < rdev->pm.dpm.num_ps; i++) {
kfree(rdev->pm.dpm.ps[i].ps_priv);
}
kfree(rdev->pm.dpm.ps);

The first free happens in ci_parse_power_table where it cleans up locally
on a failure.  ci_dpm_fini also does a cleanup.

ret = ci_parse_power_table(rdev);
if (ret) {
ci_dpm_fini(rdev);
return ret;
}

So remove the cleanup in ci_parse_power_table and
move the num_ps calculation to inside the loop so ci_dpm_fini
will know how many array elements to free.

Fixes: cc8dbbb4f62a ("drm/radeon: add dpm support for CI dGPUs (v2)")

Signed-off-by: Tom Rix <trix@redhat.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

gpu/drm: Remove debug info about CPU address

When I update the latest kernel, I see the following "____ptrval____" boot
messages.

[    1.872600] radeon 0000:01:05.0: fence driver on ring 0 use gpu addr 0x0000000048000c00 and cpu addr 0x(____ptrval____)
[    1.879095] radeon 0000:01:05.0: fence driver on ring 5 use gpu addr 0x0000000040056038 and cpu addr 0x(____ptrval____)

Both radeon_fence_driver_start_ring() and amdgpu_fence_driver_start_ring()
have the similar issue, there exists the following two methods to solve it:
(1) Use "%pK" instead of "%p" so that the CPU address can be printed when
the kptr_restrict sysctl is set to 1.
(2) Just completely drop the CPU address suggested by Christian, because
the CPU address was useful in the past, but isn't any more. We now have a
debugfs file to read the current fence values.

Since the CPU address is not much useful, just remove the debug info about
CPU address.

Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Tiezhu Yang <yangtiezhu@loongson.cn>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Add missing CW4 programming for DCN30

[Why]
To support inbox1 in CW4 we need to actually program CW4 instead of
region 4 for newer firmware.

This is done correctly on DCN20/DCN21 but this code wasn't added to
DCN30.

[How]
Copy over the missing code. It doesn't need address translation since
DCN30 uses virtual addressing.

Signed-off-by: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com>
Reviewed-by: Bhawanpreet Lakha <Bhawanpreet.Lakha@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>