drm/amd/display: Don't apply FIFO resync W/A if rdivider = 0

[Description]
It is not valid to set the WDIVIDER value to 0, so do not
re-write to DISPCLK_WDIVIDER if the current value is 0
(i.e., it is at it's initial value and we have not made any
requests to change DISPCLK yet).

Reviewed-by: Saaem Rizvi <syedsaaem.rizvi@amd.com>
Acked-by: Alex Hung <alex.hung@amd.com>
Signed-off-by: Alvin Lee <alvin.lee2@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Prevent invalid pipe connections

[Description]
 - Prevent ODM pipe connections between pipes that are not part
   of the same plane when adding a plane to context
 - Re-attach child pipes of ODM slice about to be disconnected
   to prevent any lost pipes with invalid tops/bottoms
 - We cannot split if head_pipe is not in ODM. Preventing this
   avoids creating an invalid context with an invalid pipe.

Reviewed-by: Alvin Lee <alvin.lee2@amd.com>
Acked-by: Alex Hung <alex.hung@amd.com>
Signed-off-by: Ethan Bitnun <ethan.bitnun@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: Update ring scheduler info as needed

Not all rings have scheduler associated. Only update scheduler data for
rings with scheduler. It could result in out of bound access as total
rings are more than those associated with particular IPs.

Signed-off-by: Lijo Lazar <lijo.lazar@amd.com>
Reviewed-by: James Zhu <James.Zhu@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: Enabling FW workaround through shared memory for VCN4_0_2

This patch will enable VCN FW workaround using
DRM KEY INJECT WORKAROUND method,
which is helping in fixing the secure playback.

Signed-off-by: sguttula <Suresh.Guttula@amd.com>
Reviewed-by: Leo Liu <leo.liiu@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Add VESA SCR case for default aux backlight

[How & Why]
When determining default aux backlight level, read from
DPCD address 0x734 for VESA SCR on OLED.

Reviewed-by: Felipe Clark <felipe.clark@amd.com>
Acked-by: Alex Hung <alex.hung@amd.com>
Signed-off-by: Iswara Nagulendran <iswara.nagulendran@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/radeon: Prefer dev_* variant over printk

Changed from pr_err/info to dev_* variants so that
we get better debug info when there are multiple GPUs
in the system.

Fixes: 09edeb4c1680 ("drm/radeon: Prefer pr_err/_info over printk")
Suggested-by: Alex Deucher <alexander.deucher@amd.com>
Cc: Christian König <christian.koenig@amd.com>
Cc: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Srinivasan Shanmugam <srinivasan.shanmugam@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Evan Quan <evan.quan@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/radeon: Fix style issues in radeon _encoders.c & _gart.c

Conform to Linux kernel coding style.

Fixes the following & other checks in radeon_encoders.c & radeon_gart.c:

WARNING: Missing a blank line after declarations
WARNING: Block comments use * on subsequent lines
WARNING: Prefer 'unsigned int' to bare use of 'unsigned'
WARNING: braces {} are not necessary for single statement blocks

Cc: Christian König <christian.koenig@amd.com>
Cc: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Srinivasan Shanmugam <srinivasan.shanmugam@amd.com>
Reviewed-by: Guchun Chen <guchun.chen@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/amdgpu: Fix warnings in amdgpu/amdgpu_display.c

Fixes the below checkpatch.pl warnings:

WARNING: Block comments use * on subsequent lines
WARNING: Block comments use a trailing */ on a separate line
WARNING: suspect code indent for conditional statements (8, 12)
WARNING: braces {} are not necessary for single statement blocks

Cc: Christian König <christian.koenig@amd.com>
Cc: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Srinivasan Shanmugam <srinivasan.shanmugam@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/radeon: Avoid externs & do not initialize globals to 0 in radeon_drv.c

Fixes the following:

 - WARNING: externs should be avoided in .c files
 - ERROR: do not initialise globals to 0
 - WARNING: Missing a blank line after declarations

Cc: Christian König <christian.koenig@amd.com>
Cc: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Srinivasan Shanmugam <srinivasan.shanmugam@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Guard DCN31 PHYD32CLK logic against chip family

[Why]
Current yellow carp B0 PHYD32CLK logic is incorrectly applied to other
ASICs.

[How]
Add guard to check chip family is yellow carp before applying logic.

Reviewed-by: Hansen Dsouza <hansen.dsouza@amd.com>
Acked-by: Alex Hung <alex.hung@amd.com>
Signed-off-by: George Shen <george.shen@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Correct grammar mistakes

[Why]
There are grammer mistakes in comments

[How]
Correct grammar mistakes

Reviewed-by: Anthony Koo <anthony.koo@amd.com>
Acked-by: Alex Hung <alex.hung@amd.com>
Signed-off-by: Reza Amini <reza.amini@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: Return -ENOMEM when there is no memory in 'amdgpu_gfx_mqd_sw_init'

Return -ENOMEM, when there is no sufficient dynamically allocated memory
to create MQD backup for ring

Cc: Christian König <christian.koenig@amd.com>
Cc: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Srinivasan Shanmugam <srinivasan.shanmugam@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: Fix do not add new typedefs in amdgpu_fw_attestation.c

Fixes the following to align to coding style:

WARNING: do not add new typedefs
+typedef struct FW_ATT_DB_HEADER

WARNING: do not add new typedefs
+typedef struct FW_ATT_RECORD

WARNING: Symbolic permissions 'S_IRUSR' are not preferred. Consider using octal permissions '0400'.
+                           S_IRUSR,

ERROR: "(foo*)" should be "(foo *)"
WARNING: please, no space before tabs

Cc: Christian König <christian.koenig@amd.com>
Cc: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Srinivasan Shanmugam <srinivasan.shanmugam@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: Prefer #if IS_ENABLED over #if defined in amdgpu_drv.c

Adhere to linux coding style

Fixes the following:

WARNING: Prefer IS_ENABLED(<FOO>) to CONFIG_<FOO> || CONFIG_<FOO>_MODULE
+#if defined(CONFIG_DRM_RADEON) || defined(CONFIG_DRM_RADEON_MODULE)

WARNING: Prefer IS_ENABLED(<FOO>) to CONFIG_<FOO> || CONFIG_<FOO>_MODULE
+#if defined(CONFIG_DRM_RADEON) || defined(CONFIG_DRM_RADEON_MODULE)

Cc: Christian König <christian.koenig@amd.com>
Cc: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Srinivasan Shanmugam <srinivasan.shanmugam@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdkfd: enable cooperative groups for gfx11

MES can concurrently schedule queues on the device that require
exclusive device access if marked exclusively_scheduled without the
requirement of GWS.  Similar to the F32 HWS, MES will manage
quality of service for these queues.
Use this for cooperative groups since cooperative groups are device
occupancy limited.

Since some GFX11 devices can only be debugged with partial CUs, do not
allow the debugging of cooperative groups on these devices as the CU
occupancy limit will change on attach.

In addition, zero initialize the MES add queue submission vector for MES
initialization tests as we do not want these to be cooperative
dispatches.

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: set sw state to gfxoff after SR-IOV reset

[Why]
Current SR-IOV will not set GC to off state, while it is a real
GC hard reset. Whthout GFX off flag, driver may do gfxhub invalidation
before firmware load and gfxhub gart enable. This operation may cause
CP to become busy because GC is not in the right state for invalidation.

[How]
Add a function for SR-IOV to clean up some sw state before recover. Set
adev->gfx.is_poweron to false to prevent gfxhub invalidation before gfx
firmware autoload complete.

Signed-off-by: Horace Chen <horace.chen@amd.com>
Reviewed-by: HaiJun Chang <HaiJun.Chang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/smu: use AverageGfxclkFrequency* to replace previous GFX Curr Clock

Report current GFX clock also from average clock value as the original
CurrClock data is not valid/accurate any more as per FW team

Signed-off-by: Jane Jian <Jane.Jian@amd.com>
Reviewed-by: Lijo Lazar <lijo.lazar@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: Fix one kernel-doc comment

Use colon to separate parameter name from their specific meaning.
silence the warning:

drivers/gpu/drm/amd/amdgpu/amdgpu_vm_pt.c:793: warning: Function parameter or member 'adev' not described in 'amdgpu_vm_pte_update_noretry_flags'

Reviewed-by: Randy Dunlap <rdunlap@infradead.org>
Signed-off-by: Yang Li <yang.lee@linux.alibaba.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd: Fix an error handling mistake in psp_sw_init()

If the second call to amdgpu_bo_create_kernel() fails, the memory
allocated from the first call should be cleared.  If the third call
fails, the memory from the second call should be cleared.

Fixes: b95b5391684b ("drm/amdgpu/psp: move PSP memory alloc from hw_init to sw_init")
Signed-off-by: Mario Limonciello <mario.limonciello@amd.com>
Reviewed-by: Lijo Lazar <lijo.lazar@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: Fix infinite loop in gfxhub_v1_2_xcc_gart_enable (v2)

An instance of for_each_inst() was not changed to match its new
behaviour and is causing a loop.

v2: remove tmp_mask variable

Fixes: b579ea632fca ("drm/amdgpu: Modify for_each_inst macro")
Signed-off-by: Victor Lu <victorchengchi.lu@amd.com>
Acked-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: Program xcp_ctl registers as needed

XCP_CTL register is expected to be programmed by firmware. Under certain
conditions FW may not have programmed it correctly. As a workaround,
program it when FW has not programmed the right values.

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

drm/amdkfd: fix trap handling work around for debugging

Update the list of devices that require the cwsr trap handling
workaround for debugging use cases.

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

drm/amd/display: Allow building DC with clang on RISC-V

clang on RISC-V appears to be unaffected by the bug causing excessive
stack usage in calculate_bandwidth(). clang 16 with -fstack-usage
reports a 304 byte stack frame size with CONFIG_ARCH_RV32I, and 512
bytes with CONFIG_ARCH_RV64I.

Reviewed-by: Nathan Chancellor <nathan@kernel.org>
Tested-by: Nathan Chancellor <nathan@kernel.org>
Signed-off-by: Samuel Holland <samuel.holland@sifive.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: remove an unused file

[Why&How]
Internal subvp state is not referenced in driver code, so it can be removed.

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

drm/radeon: Prefer pr_err/_info over printk

Fixes the following checkpatch.pl:

WARNING: printk() should include KERN_<LEVEL> facility level

Cc: Christian König <christian.koenig@amd.com>
Cc: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Srinivasan Shanmugam <srinivasan.shanmugam@amd.com>
Reviewed-by: Guchun Chen <guchun.chen@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: allow secure submission on VCN4 ring

This patch will enable secure decode playback on VCN4_0_2

Signed-off-by: sguttula <Suresh.Guttula@amd.com>
Reviewed-by: Leo Liu <leo.liiu@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd: Avoid reading the VBIOS part number twice

The VBIOS part number is read both in amdgpu_atom_parse() as well
as in atom_get_vbios_pn() and stored twice in the `struct atom_context`
structure. Remove the first unnecessary read and move the `pr_info`
line from that read into the second.

v2: squash in unused variable removal

Signed-off-by: Mario Limonciello <mario.limonciello@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: use a macro to define no xcp partition case

~0 as no xcp partition is used in several places, so improve its
definition by a macro for code consistency.

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

drm/amdgpu/vm: use the same xcp_id from root PD

Other PDs/PTs allocation should just use the same xcp_id as that
stored in root PD.

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

drm/amdgpu: fix slab-out-of-bounds issue in amdgpu_vm_pt_create

Recent code set xcp_id stored from file private data when opening
device to amdgpu bo for accounting memory usage etc, but not all
VMs are attached to this fpriv structure like the vm cases in
amdgpu_mes_self_test, otherwise, KASAN will complain below out
of bound access. And more importantly, VM code should not touch
fpriv structure, so drop fpriv code handling from amdgpu_vm_pt.

[   77.292314] BUG: KASAN: slab-out-of-bounds in amdgpu_vm_pt_create+0x17e/0x4b0 [amdgpu]
[   77.293845] Read of size 4 at addr ffff888102c48a48 by task modprobe/1069
[   77.294146] Call Trace:
[   77.294178]  <TASK>
[   77.294208]  dump_stack_lvl+0x49/0x63
[   77.294260]  print_report+0x16f/0x4a6
[   77.294307]  ? amdgpu_vm_pt_create+0x17e/0x4b0 [amdgpu]
[   77.295979]  ? kasan_complete_mode_report_info+0x3c/0x200
[   77.296057]  ? amdgpu_vm_pt_create+0x17e/0x4b0 [amdgpu]
[   77.297556]  kasan_report+0xb4/0x130
[   77.297609]  ? amdgpu_vm_pt_create+0x17e/0x4b0 [amdgpu]
[   77.299202]  __asan_load4+0x6f/0x90
[   77.299272]  amdgpu_vm_pt_create+0x17e/0x4b0 [amdgpu]
[   77.300796]  ? amdgpu_init+0x6e/0x1000 [amdgpu]
[   77.302222]  ? amdgpu_vm_pt_clear+0x750/0x750 [amdgpu]
[   77.303721]  ? preempt_count_sub+0x18/0xc0
[   77.303786]  amdgpu_vm_init+0x39e/0x870 [amdgpu]
[   77.305186]  ? amdgpu_vm_wait_idle+0x90/0x90 [amdgpu]
[   77.306683]  ? kasan_set_track+0x25/0x30
[   77.306737]  ? kasan_save_alloc_info+0x1b/0x30
[   77.306795]  ? __kasan_kmalloc+0x87/0xa0
[   77.306852]  amdgpu_mes_self_test+0x169/0x620 [amdgpu]

v2: without specifying xcp partition for PD/PT bo, the xcp id is -1.

Link: https://gitlab.freedesktop.org/drm/amd/-/issues/2686
Fixes: 3ebfd221c1a8 ("drm/amdkfd: Store xcp partition id to amdgpu bo")
Signed-off-by: Guchun Chen <guchun.chen@amd.com>
Tested-by: Mikhail Gavrilov <mikhail.v.gavrilov@gmail.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: Allocate root PD on correct partition

file_priv needs to be setup firstly, otherwise, root PD
will always be allocated on partition 0, even if opening
the device from other partitions.

Fixes: 3ebfd221c1a8 ("drm/amdkfd: Store xcp partition id to amdgpu bo")
Signed-off-by: Guchun Chen <guchun.chen@amd.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: Add RLCG interface driver implementation for gfx v9.4.3 (v3)

Add RLCG interface support for gfx v9.4.3 and multiple XCCs.
Do not enable it yet.

v2: Fix amdgpu_rlcg_reg_access_ctrl init, add support for multiple XCCs
    in amdgpu_mm_wreg_mmio_rlc

v3: Use GET_INST() when indexing amdgpu_rlcg_reg_access_ctrl

Signed-off-by: Victor Lu <victorchengchi.lu@amd.com>
Reviewed-by: Zhigang Luo <zhigang.luo@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Promote DAL to 3.2.243

This version brings along following fixes:
- Update 128b/132b downspread factor to 0.3%
- Add helpers to get DMUB FW boot options
- Initialize necessary uninitialized variables
- Add stream overhead in BW calculations for 128b/132b
- Add link encoding to timing BW calculation parameters
- Prevent vtotal from being set to 0
- Fix race condition when turning off an output alone
- Keep PHY active for DP displays on DCN31
- Fix ASIC check in aux timeout workaround
- ABM pause toggle
- Add missing triggers for full updates
- Disable MPC split by default on special asic
- Add additional refresh rate conditions for SubVP cases
- Fix DP2 link training failure with RCO
- Reenable all root clock gating options
- Cache backlight_millinits in link structure and setting brightness accordingly
- Refine to decide the verified link setting
- Update SW cursor fallback for subvp high refresh

Acked-by: Alan Liu <haoping.liu@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: Add missing triggers for full updates

[Description]
- Full update was missed for the following cases:
- Idle optimization is enabled
- Plane is not in current context
- Also don't clear surface updates at end of
  commit_plane_for_stream_fast as they are cleared at
  the beginning of each flip (only stream updates need
  to be cleared in case there is no stream update in
  the next flip)

Reviewed-by: Samson Tam <samson.tam@amd.com>
Acked-by: Alan Liu <haoping.liu@amd.com>
Signed-off-by: Alvin Lee <alvin.lee2@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: ABM pause toggle

[why]
Allow ABM states to be transferred across display
adapters for smooth display transitions.

[how]
We call DMUB to pause and get ABM states. We
transfer data to other gpu, and deliver data and
ask ABM to un-pause.

Reviewed-by: Harry Vanzylldejong <harry.vanzylldejong@amd.com>
Acked-by: Alan Liu <haoping.liu@amd.com>
Signed-off-by: Reza Amini <reza.amini@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Fix ASIC check in aux timeout workaround

[Why]
Aux write was meant to be ASIC specific, and is
causing compliance failures on newer parts.

[How]
Make workaround specific to single ASIC.

Reviewed-by: Michael Strauss <michael.strauss@amd.com>
Acked-by: Alan Liu <haoping.liu@amd.com>
Signed-off-by: Taimur Hassan <syed.hassan@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Keep PHY active for DP displays on DCN31

[Why & How]
Port of a change that went into DCN314 to keep the PHY enabled
when we have a connected and active DP display.

The PHY can hang if PHY refclk is disabled inadvertently.

Cc: Mario Limonciello <mario.limonciello@amd.com>
Cc: Alex Deucher <alexander.deucher@amd.com>
Cc: stable@vger.kernel.org
Reviewed-by: Josip Pavic <josip.pavic@amd.com>
Acked-by: Alan Liu <haoping.liu@amd.com>
Signed-off-by: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Prevent vtotal from being set to 0

[Why]
In dcn314 DML the destination pipe vtotal was being set
to the crtc adjustment vtotal_min value even in cases
where that value is 0.

[How]
Only set vtotal to the crtc adjustment vtotal_min value
in cases where the value is non-zero.

Cc: Mario Limonciello <mario.limonciello@amd.com>
Cc: Alex Deucher <alexander.deucher@amd.com>
Cc: stable@vger.kernel.org
Reviewed-by: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com>
Acked-by: Alan Liu <haoping.liu@amd.com>
Signed-off-by: Daniel Miess <daniel.miess@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Add link encoding to timing BW calculation parameters

[Why]
There certain cases where the timing BW is dependent on the type of link
encoding in use. Thus to calculate the correct BW required for a given
timing, the link encoding should be added as a parameter.

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

drm/amd/display: Add stream overhead in BW calculations for 128b/132b

[Why]
Current BW calculations do not account for the additional padding added
for uncompressed pixel-to-symbol packing.

This results in X.Y being too low for 128b/132b SST streams in certain
scenarios. If X.Y is too low, end user can observe image corruption.

[How]
Add function to calculate stream overhead to timing BW calculation for
128b/132b SST cases.

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

drm/amd/display: Initialize necessary uninitialized variables

This commit initializes uninitialized variables.
For some compilers uninitialized variable warnings are treated as Error.

Reviewed-by: Wenjing Liu <wenjing.liu@amd.com>
Acked-by: Alan Liu <haoping.liu@amd.com>
Signed-off-by: Meera Patel <meera.patel@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Add helpers to get DMUB FW boot options

[Why & How]
To query the bits and print them out for debug purposes.

Reviewed-by: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com>
Acked-by: Alan Liu <haoping.liu@amd.com>
Signed-off-by: Cruise Hung <cruise.hung@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Promote DAL to 3.2.242

This version brings along following fixes:
- Setup stream encoder before link enable for TMDS
- Use is_dig_enable function instead of dcn10 hardcode
- Re-add aux intercept disable delay generically for 2+ LTTPRs
- Add polling method to handle MST reply packet
- Remove unused function
- Improve the include of header file
- Add missing static
- Hardcode vco_freq for dcn316
- Reduce stack size
- Skip enabling DMCUB when using emulation
- Check TG is non-null before checking if enabled
- Update scaler recout data for visual confirm
- Skip querying caps when DMCUB emulation is in use

Acked-by: Alan Liu <haoping.liu@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: Disable MPC split by default on special asic

[WHY]
All of pipes will be used when the MPC split enable on the dcn
which just has 2 pipes. Then MPO enter will trigger the minimal
transition which need programe dcn from 2 pipes MPC split to 2
pipes MPO. This action will cause lag if happen frequently.

[HOW]
Disable the MPC split for the platform which dcn resource is limited

Cc: Mario Limonciello <mario.limonciello@amd.com>
Cc: Alex Deucher <alexander.deucher@amd.com>
Cc: stable@vger.kernel.org
Reviewed-by: Alvin Lee <alvin.lee2@amd.com>
Acked-by: Alan Liu <haoping.liu@amd.com>
Signed-off-by: Zhikai Zhai <zhikai.zhai@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Update 128b/132b downspread factor to 0.3%

[Why]
Updating downspread factor to 0.3% to add additional margin to account
for potential link rate deviations (up to 300ppm as per the DP spec).

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

drm/amd/display: Fix race condition when turning off an output alone

[Why]
When 2 threads are doing commit_tail parallelly, one thread could
commit new streams to dc state but another thread remove it from dc
right away.

[How]
If we don't have new dm state change from commit_check, then we should
not call dc_commit_streams() in commit_tail. A new function
amdgpu_dm_commit_streams() is introduced to refator dc_commit_stream()
adjacent code and fix this issue.

Reviewed-by: Harry Wentland <harry.wentland@amd.com>
Acked-by: Alan Liu <haoping.liu@amd.com>
Signed-off-by: Alan Liu <haoping.liu@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: add additional refresh rate conditions for SubVP cases

[Why]
Want to further constrain these refresh rate conditions for SubVP:
- SubVP + SubVP: either both <= 120Hz or both between 120-165Hz
- SubVP + DRR: SubVP <= 120Hz
- SubVP + VBlank: SubVP <= 120Hz

[How]
Add constraints in subvp_subvp_admissable(),
dcn32_subvp_drr_admissable() and dcn32_subvp_vblank_admissable()

Reviewed-by: Alvin Lee <alvin.lee2@amd.com>
Acked-by: Alan Liu <haoping.liu@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: refine to decide the verified link setting

[WHY]
We return false by default when link training fail at link loss.
It will cause we get a fail verified link caps directly.

[HOW]
Record the highest verified link caps. Use the recorded value as
the verified link caps if it is not successful in the last attempt
to avoid to use the lowest link setting.

Reviewed-by: Wenjing Liu <wenjing.liu@amd.com>
Acked-by: Alan Liu <haoping.liu@amd.com>
Signed-off-by: Zhikai Zhai <zhikai.zhai@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Fix DP2 link training failure with RCO

[Why]
When RCO is enabled for symclk32_le we get failures during
DP2 link traing compliance tests.

[How]
Break out symclk32_le RCO into a separate function that is
called for hpo when link is enabled/disabled.

Reviewed-by: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com>
Acked-by: Alan Liu <haoping.liu@amd.com>
Signed-off-by: Daniel Miess <daniel.miess@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Cache backlight_millinits in link structure and setting brightness accordingly

[Why]
Need to save the cached backlight level so that display lights up using
appropriate brightness level instead of the default brightness when
waking up from s0i3.

[How]
Adding a backlight level cache in link structure. Also instead on
calling set_default_brightness_aux, check if cached values exists.

Reviewed-by: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com>
Reviewed-by: Wenjing Liu <wenjing.liu@amd.com>
Acked-by: Alan Liu <haoping.liu@amd.com>
Signed-off-by: Nicholas Susanto <nicholas.susanto@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Reenable all root clock gating options

[Why & How]
Re-enable all RCO options now that all known issues with
RCO have been addressed

Reviewed-by: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com>
Acked-by: Alan Liu <haoping.liu@amd.com>
Signed-off-by: Daniel Miess <daniel.miess@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Update SW cursor fallback for subvp high refresh

[Description]
Since SubVP high refresh is now enabled, we must
fallback to SW cursor under if we're in a SubVP
high refresh config

Reviewed-by: Samson Tam <samson.tam@amd.com>
Acked-by: Alan Liu <haoping.liu@amd.com>
Signed-off-by: Alvin Lee <alvin.lee2@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Skip querying caps when DMCUB emulation is in use

[Why]
Workaround to avoid accessing DMCUB state too early if the emulator
is in use - we don't support any of the features the caps are querying
with emulation anyway.

[How]
Guard the query if emulation is in use.

Reviewed-by: Charlene Liu <charlene.liu@amd.com>
Acked-by: Alan Liu <haoping.liu@amd.com>
Signed-off-by: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Update scaler recout data for visual confirm

[Why]
Visual confirm color is not as expected for Autoa Color Management
feature test.

[How]
Calculate scaler recout data when visual confirm enabled to update
the visual confirm bar on the display.

Reviewed-by: Aric Cyr <aric.cyr@amd.com>
Acked-by: Alan Liu <haoping.liu@amd.com>
Signed-off-by: Leo Ma <hanghong.ma@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: check TG is non-null before checking if enabled

[Why & How]
If there is no TG allocation we can dereference a NULL pointer when
checking if the TG is enabled.

Cc: Mario Limonciello <mario.limonciello@amd.com>
Cc: Alex Deucher <alexander.deucher@amd.com>
Cc: stable@vger.kernel.org
Reviewed-by: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com>
Acked-by: Alan Liu <haoping.liu@amd.com>
Signed-off-by: Taimur Hassan <syed.hassan@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Skip enabling DMCUB when using emulation

[Why & How]
Shouldn't be touching path for HW DMCUB when emulating.

Reviewed-by: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com>
Acked-by: Alan Liu <haoping.liu@amd.com>
Signed-off-by: Taimur Hassan <syed.hassan@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Reduce stack size

Reduce stack size pointed by clang:

amdgpu_dm/amdgpu_dm.c:8655:13: error: stack frame size (1048) exceeds limit (1024) in 'amdgpu_dm_atomic_commit_tail' [-Werror,-Wframe-larger-than]

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

drm/amd/display: Hardcode vco_freq for dcn316

There is no need to calculate the VCO frequency. In our internal branch
we've hard-coded this for a while, so it's well-tested. This also allows
us to remove the now unused code for calculating the VCO frequency.

Reviewed-by: Harry Wentland <harry.wentland@amd.com>
Acked-by: Alan Liu <haoping.liu@amd.com>
Signed-off-by: Alan Liu <haoping.liu@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Add missing static

After enable DRM_AMDGPU_WERROR, clang highlight multiple functions that
need to have `static`, and this commit address those issues and also
improve the indents.

Reviewed-by: Hamza Mahfooz <hamza.mahfooz@amd.com>
Acked-by: Alan Liu <haoping.liu@amd.com>
Signed-off-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 unused function

The get_engine_type is never used in the code, for this reason, this
commit drops this function.

Reviewed-by: Hamza Mahfooz <hamza.mahfooz@amd.com>
Acked-by: Alan Liu <haoping.liu@amd.com>
Signed-off-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: Improve the include of header file

This commit improves the include of some header files to make them align
with other includes.

Reviewed-by: Hamza Mahfooz <hamza.mahfooz@amd.com>
Acked-by: Alan Liu <haoping.liu@amd.com>
Signed-off-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 polling method to handle MST reply packet

[Why]
Specific TBT4 dock doesn't send out short HPD to notify source
that IRQ event DOWN_REP_MSG_RDY is set. Which violates the spec
and cause source can't send out streams to mst sinks.

[How]
To cover this misbehavior, add an additional polling method to detect
DOWN_REP_MSG_RDY is set. HPD driven handling method is still kept.
Just hook up our handler to drm mgr->cbs->poll_hpd_irq().

Cc: Mario Limonciello <mario.limonciello@amd.com>
Cc: Alex Deucher <alexander.deucher@amd.com>
Cc: stable@vger.kernel.org
Reviewed-by: Jerry Zuo <jerry.zuo@amd.com>
Acked-by: Alan Liu <haoping.liu@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-add aux intercept disable delay generically for 2+ LTTPRs

[WHY]
Issue not display generic, required for multiple 2+ LTTPR link configurations.

[HOW]
Revert monitor patch change and remove delay for single LTTPR case

Reviewed-by: George Shen <george.shen@amd.com>
Acked-by: Alan Liu <haoping.liu@amd.com>
Signed-off-by: Michael Strauss <michael.strauss@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Use is_dig_enable function instead of dcn10 hardcode

[Why]
This can differ depending on ASIC and we can end up skipping all
transmitter control if we're relying on the wrong bit as a shortcut
for whether link is on/off.

[How]
The ASIC table itself provides the correct DIG check, use that instead
of the dcn10 hardcode.

Reviewed-by: Syed Hassan <syed.hassan@amd.com>
Acked-by: Alan Liu <haoping.liu@amd.com>
Signed-off-by: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Setup stream encoder before link enable for TMDS

[Why]
HDMI spec requires TMDS clock to be not more than 340MHz. Stream encoder ensure
this requirement but driver enable stream encoder later than PHY. So PHY will
output full speed TMDS clock first.

[How]
Enable stream encoder first in TMDS case.

Reviewed-by: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com>
Acked-by: Alan Liu <haoping.liu@amd.com>
Signed-off-by: JinZe Xu <jinze.xu@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: create a new file for doorbell manager

This patch:
- creates a new file for doorbell management.
- moves doorbell code from amdgpu_device.c to this file.

V2:
 - remove doc from function declaration (Christian)
 - remove 'device' from function names to make it consistent (Alex)
 - add SPDX license identifier (Luben)

V3:
 - change license to MIT license(Christian)

Cc: Alex Deucher <alexander.deucher@amd.com>
Cc: Christian Koenig <christian.koenig@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Shashank Sharma <shashank.sharma@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: Add dcdebugmask option to enable DPIA trace

[Why & How]
It's useful to be able to enable DPIA trace with dcdebugmask
option, especially to debug DPIA issues involved in transition
of system power states.

This patch adds an option to amdgpu.dcdebugmask to be picked up
by amdgpu DM to enable DPIA trace.

Signed-off-by: Stylon Wang <stylon.wang@amd.com>
Reviewed-by: Aurabindo Pillai <aurabindo.pillai@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: Allow the initramfs generator to include psp_13_0_6_ta

Allow the initramfs generator to automatically include psp_13_0_6_ta
firmware to initramfs.

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

drm/amdgpu: Disable RAS by default on APU flatform

Disable RAS feature by default for aqua vanjaram on APU platform.

Changed from V1:
Splite Disable RAS by default on APU platform into a
separated patch.

Changed from V2:
Avoid to modify global variable amdgpu_ras_enable.

Signed-off-by: Stanley.Yang <Stanley.Yang@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: Enable aqua vanjaram RAS

Enable RAS for aqua vanjaram.

Changed from V1:
Split the change in amdgpu_ras_asic_supported into a
separated patch.

Changed from V2:
Avoid to modify global variable amdgpu_ras_enable.

Signed-off-by: Stanley.Yang <Stanley.Yang@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: Avoid possiblity of kernel crash in 'gmc_v8_0, gmc_v7_0_init_microcode()'

If the function 'gmc_v8_0_ or gmc_v7_0_init_microcode()' fails, the
driver will just fail to load, hence return -EINVAL rather having BUG(),
fixes WARNING: Do not crash the kernel unless it is absolutely
unavoidable--use WARN_ON_ONCE() plus recovery code (if feasible) instead
of BUG() or variants

Fixes: 2f77b5931f68 ("drm/amdgpu: Fix error & warnings in gmc_v8_0.c")
Fixes: 0cfc1d683046 ("drm/amdgpu: Fix errors & warnings in gmc_ v6_0, v7_0.c")
Suggested-by: Alex Deucher <alexander.deucher@amd.com>
Cc: Christian König <christian.koenig@amd.com>
Cc: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Srinivasan Shanmugam <srinivasan.shanmugam@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Eliminate warnings in amdgpu_dm_helpers.c

Fix the following warnings reported by checkpatch:

WARNING: Block comments use a trailing */ on a separate line
WARNING: Prefer using '"%s...", __func__' to using 'execute_synaptics_rc_command', this function's name, in a string
WARNING: Prefer using '"%s...", __func__' to using 'apply_synaptics_fifo_reset_wa', this function's name, in a string
WARNING: braces {} are not necessary for single statement blocks

Cc: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Cc: Aurabindo Pillai <aurabindo.pillai@amd.com>
Signed-off-by: Srinivasan Shanmugam <srinivasan.shanmugam@amd.com>
Reviewed-by: Aurabindo Pillai <aurabindo.pillai@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Clean up style problems in amdgpu_dm_hdcp.c

Conform to Linux kernel coding style.

And promote sysfs entry for set/get srm to kdoc.

Suggested-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Cc: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Cc: Aurabindo Pillai <aurabindo.pillai@amd.com>
Signed-off-by: Srinivasan Shanmugam <srinivasan.shanmugam@amd.com>
Reviewed-by: Aurabindo Pillai <aurabindo.pillai@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Implement zpos property

Despite certain GPUs supporting multiple overlay planes already in
AMDGPU, the driver did not expose the zpos property which is required
for userspace to take advantage of multiple overlay planes in any
meaningful way.

The driver was already hooked up to normalized_zpos, but was just
missing the exposure of it.

Cc: Harry Wentland <harry.wentland@amd.com>
Cc: Melissa Wen <mwen@igalia.com>
Cc: Simon Ser <contact@emersion.fr>
Reviewed-by: Harry Wentland <harry.wentland@amd.com>
Signed-off-by: Joshua Ashton <joshua@froggi.es>
Signed-off-by: Hamza Mahfooz <hamza.mahfooz@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu/pm: make mclk consistent for smu 13.0.7

Use current uclk to be consistent with other dGPUs.

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

drm/amdgpu/pm: make gfxclock consistent for sienna cichlid

Use average gfxclock for consistency with other dGPUs.

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

drm/amd/display: Enable 3 plane for DCN 3.01

Steam Deck/Gamescope wants to take advantage of more planes which is
possible on VanGogh but was not previously exposed.

Cc: Harry Wentland <harry.wentland@amd.com>
Cc: Melissa Wen <mwen@igalia.com>
Cc: Simon Ser <contact@emersion.fr>
Reviewed-by: Harry Wentland <harry.wentland@amd.com>
Signed-off-by: Joshua Ashton <joshua@froggi.es>
Signed-off-by: Hamza Mahfooz <hamza.mahfooz@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Expose more formats for overlay planes on DCN

DCN planes are universal and therefore overlay planes can use the same
formats as primary planes, unlike DCE.

Gamescope/Steam Deck would like to take advantage of this functionality
for partial composition which in some cases in our pipeline, can contain
negative values in some instances.

Cc: Harry Wentland <harry.wentland@amd.com>
Cc: Melissa Wen <mwen@igalia.com>
Cc: Simon Ser <contact@emersion.fr>
Reviewed-by: Harry Wentland <harry.wentland@amd.com>
Signed-off-by: Joshua Ashton <joshua@froggi.es>
Signed-off-by: Hamza Mahfooz <hamza.mahfooz@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

Revert "drm/amdgpu:update kernel vcn ring test"

VCN FW depncencies revert it to unblock others

This reverts commit f3fa86f5c778e258cd5c01bb420d4639bb393bd0.

Signed-off-by: Saleemkhan Jamadar <saleemkhan.jamadar@amd.com>
Acked-by: Veerabadhran Gopalakrishnan <Veerabadhran.Gopalakrishnan@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

Revert "drm/amdgpu: update kernel vcn ring test"

VCN FW depncencies revert it to unlock others

This reverts commit 3ebfa943b8451e4675d023b3f387911702ebee17.

Signed-off-by: Saleemkhan Jamadar <saleemkhan.jamadar@amd.com>
Acked-by: Veerabadhran Gopalakrishnan <Veerabadhran.Gopalakrishnan@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu/vkms: relax timer deactivation by hrtimer_try_to_cancel

In below thousands of screen rotation loop tests with virtual display
enabled, a CPU hard lockup issue may happen, leading system to unresponsive
and crash.

do {
xrandr --output Virtual --rotate inverted
xrandr --output Virtual --rotate right
xrandr --output Virtual --rotate left
xrandr --output Virtual --rotate normal
} while (1);

NMI watchdog: Watchdog detected hard LOCKUP on cpu 1

? hrtimer_run_softirq+0x140/0x140
? store_vblank+0xe0/0xe0 [drm]
hrtimer_cancel+0x15/0x30
amdgpu_vkms_disable_vblank+0x15/0x30 [amdgpu]
drm_vblank_disable_and_save+0x185/0x1f0 [drm]
drm_crtc_vblank_off+0x159/0x4c0 [drm]
? record_print_text.cold+0x11/0x11
? wait_for_completion_timeout+0x232/0x280
? drm_crtc_wait_one_vblank+0x40/0x40 [drm]
? bit_wait_io_timeout+0xe0/0xe0
? wait_for_completion_interruptible+0x1d7/0x320
? mutex_unlock+0x81/0xd0
amdgpu_vkms_crtc_atomic_disable

It's caused by a stuck in lock dependency in such scenario on different
CPUs.

CPU1                                             CPU2
drm_crtc_vblank_off                              hrtimer_interrupt
    grab event_lock (irq disabled)                   __hrtimer_run_queues
        grab vbl_lock/vblank_time_block                  amdgpu_vkms_vblank_simulate
            amdgpu_vkms_disable_vblank                       drm_handle_vblank
                hrtimer_cancel                                         grab dev->event_lock

So CPU1 stucks in hrtimer_cancel as timer callback is running endless on
current clock base, as that timer queue on CPU2 has no chance to finish it
because of failing to hold the lock. So NMI watchdog will throw the errors
after its threshold, and all later CPUs are impacted/blocked.

So use hrtimer_try_to_cancel to fix this, as disable_vblank callback
does not need to wait the handler to finish. And also it's not necessary
to check the return value of hrtimer_try_to_cancel, because even if it's
-1 which means current timer callback is running, it will be reprogrammed
in hrtimer_start with calling enable_vblank to make it works.

v2: only re-arm timer when vblank is enabled (Christian) and add a Fixes
tag as well

v3: drop warn printing (Christian)

v4: drop superfluous check of blank->enabled in timer function, as it's
guaranteed in drm_handle_vblank (Christian)

Fixes: 84ec374bd580 ("drm/amdgpu: create amdgpu_vkms (v4)")
Cc: stable@vger.kernel.org
Suggested-by: Christian König <christian.koenig@amd.com>
Signed-off-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/amd/display: add DCN301 specific logic for OTG programming

[Why&How]
DCN301 does not have FAMS hence the workaround needed on other DCN3x
variants related to OTG min/max selector programming is not applicable for it.
Hence isolate it and have it use the old sequence without workaround.

Fixes: 1598fc576420 ("drm/amd/display: Program OTG vtotal min/max selectors unconditionally for DCN1+")
Reviewed-by: Swapnil Patel <swapnil.patel@amd.com>
Signed-off-by: Aurabindo Pillai <aurabindo.pillai@amd.com>
Tested-by: Guilherme G. Piccoli <gpiccoli@igalia.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: export some optc function for reuse

[Why&How]
Make a few functions non static so that they can be reused for other
asic. This is in preparation for separating out OTG programming sequence
for DCN301

Fixes: 1598fc576420 ("drm/amd/display: Program OTG vtotal min/max selectors unconditionally for DCN1+")
Reviewed-by: Swapnil Patel <swapnil.patel@amd.com>
Signed-off-by: Aurabindo Pillai <aurabindo.pillai@amd.com>
Tested-by: Guilherme G. Piccoli <gpiccoli@igalia.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd: Use amdgpu_device_pcie_dynamic_switching_supported() for SMU7

SMU7 does a check if the dGPU is inserted into a Rocket Lake system,
to turn off DPM.  Extend this check to all systems that have problems
with dynamic switching by using the
amdgpu_device_pcie_dynamic_switching_supported() helper.

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

drm/amdgpu: Fix error & warnings in gmc_v8_0.c

Fix below checkpatch error & warnings:

ERROR: trailing statements should be on next line
+       default: BUG();

WARNING: braces {} are not necessary for single statement blocks
WARNING: braces {} are not necessary for any arm of this statement
WARNING: Block comments should align the * on each line

Cc: Christian König <christian.koenig@amd.com>
Cc: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Srinivasan Shanmugam <srinivasan.shanmugam@amd.com>
Acked-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: dc.h: eliminate kernel-doc warnings

Quash 175 kernel-doc warnings in dc.h by unmarking 2 struct
comments as containing kernel-doc notation and by spelling one
struct field correctly in a kernel-doc comment.

Fixes: 1682bd1a6b5f ("drm/amd/display: Expand kernel doc for DC")
Fixes: ea76895ffab1 ("drm/amd/display: Document pipe split policy")
Fixes: f6ae69f49fcf ("drm/amd/display: Include surface of unaffected streams")
Reviewed-by: Harry Wentland <harry.wentland@amd.com>
Signed-off-by: Randy Dunlap <rdunlap@infradead.org>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Cc: Harry Wentland <harry.wentland@amd.com>
Cc: Leo Li <sunpeng.li@amd.com>
Cc: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Cc: Aurabindo Pillai <aurabindo.pillai@amd.com>
Cc: Alex Deucher <alexander.deucher@amd.com>
Cc: amd-gfx@lists.freedesktop.org
Cc: Alex Deucher <alexander.deucher@amd.com>
Cc: Christian König <christian.koenig@amd.com>
Cc: "Pan, Xinhui" <Xinhui.Pan@amd.com>
Cc: dri-devel@lists.freedesktop.org

drm/amdgpu: Rename to amdgpu_vm_tlb_seq_struct

Rename struct amdgpu_vm_tlb_seq_cb {...} to struct amdgpu_vm_tlb_seq_struct
{...}, so as to not conflict with documentation processing tools. Of course, C
has no problem with this.

Cc: Randy Dunlap <rdunlap@infradead.org>
Cc: Alex Deucher <Alexander.Deucher@amd.com>
Cc: Christian König <christian.koenig@amd.com>
Link: https://lore.kernel.org/r/b5ebc891-ee63-1638-0377-7b512d34b823@infradead.org
Signed-off-by: Luben Tuikov <luben.tuikov@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Randy Dunlap <rdunlap@infradead.org>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdkfd: Fix stack size in 'amdgpu_amdkfd_unmap_hiq'

Dynamically allocate large local variable instead of putting it onto the
stack to avoid exceeding the stack size:
drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c: In function ‘amdgpu_amdkfd_unmap_hiq’:
drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c:868:1: warning: the frame size of 1280 bytes is larger than 1024 bytes [-Wframe-larger-than=]

Reported-by: kernel test robot <lkp@intel.com>
Closes: https://lore.kernel.org/oe-kbuild-all/202307080505.V12qS0oz-lkp@intel.com
Suggested-by: Guchun Chen <guchun.chen@amd.com>
Cc: Felix Kuehling <Felix.Kuehling@amd.com>
Cc: Christian König <christian.koenig@amd.com>
Cc: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Srinivasan Shanmugam <srinivasan.shanmugam@amd.com>
Acked-by: Guchun Chen <guchun.chen@amd.com>
Reviewed-by: Mukul Joshi <mukul.joshi@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdkfd: report dispatch id always saved in ttmps after gc9.4.2

The feature to save the dispatch ID in trap temporaries 6 & 7 on context
save is unconditionally enabled during MQD initialization.

Now that TTMPs are always setup regardless of debug mode for GC 9.4.3, we
should report that the dispatch ID is always available for debug/trap
handling.

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/amd: Align SMU11 SMU_MSG_OverridePcieParameters implementation with SMU13

SMU13 overrides dynamic PCIe lane width and dynamic speed by when on
certain hosts. commit 38e4ced80479 ("drm/amd/pm: conditionally disable
pcie lane switching for some sienna_cichlid SKUs") worked around this
issue by setting up certain SKUs to set up certain limits, but the same
fundamental problem with those hosts affects all SMU11 implmentations
as well, so align the SMU11 and SMU13 driver handling.

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

drm/amd: Move helper for dynamic speed switch check out of smu13

This helper is used for checking if the connected host supports
the feature, it can be moved into generic code to be used by other
smu implementations as well.

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

drm/amdgpu: update kernel vcn ring test

add session context buffer to decoder ring test for vcn v1 to v3.

v3 - correct the cmd for sesssion ctx buf
v2 - add the buffer into IB (Leo liu)

Signed-off-by: Saleemkhan Jamadar <saleemkhan.jamadar@amd.com>
Reviewed-by: Leo Liu <leo.liu@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu:update kernel vcn ring test

add session context buffer to decoder ring test.

v5 - clear the session ct buffer (Christian)
v4 - data type, explain change of ib size change (Christian)
v3 - indent and  v2 changes correction. (Christian)
v2 - put the buffer at the end of the IB (Christian)

Signed-off-by: Saleemkhan Jamadar <saleemkhan.jamadar@amd.com>
Acked-by: Leo Liu <leo.liu@amd.com>
Acked-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: only accept async flips for fast updates

Up until now, amdgpu was silently degrading to vsync when
user-space requested an async flip but the hardware didn't support
it.

The hardware doesn't support immediate flips when the update changes
the FB pitch, the DCC state, the rotation, enables or disables CRTCs
or planes, etc. This is reflected in the dm_crtc_state.update_type
field: UPDATE_TYPE_FAST means that immediate flip is supported.

Silently degrading async flips to vsync is not the expected behavior
from a uAPI point-of-view. Xorg expects async flips to fail if
unsupported, to be able to fall back to a blit. i915 already behaves
this way.

This patch aligns amdgpu with uAPI expectations and returns a failure
when an async flip is not possible.

Signed-off-by: Simon Ser <contact@emersion.fr>
Reviewed-by: André Almeida <andrealmeid@igalia.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Harry Wentland <harry.wentland@amd.com>
Signed-off-by: André Almeida <andrealmeid@igalia.com>
Signed-off-by: Hamza Mahfooz <hamza.mahfooz@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/pm: conditionally disable pcie lane/speed switching for SMU13

Intel platforms such as Sapphire Rapids and Raptor Lake don't support
dynamic pcie lane or speed switching.

This limitation seems to carry over from one generation to another.
To be safer, disable dynamic pcie lane width and speed switching when
running on an Intel platform.

Link: https://edc.intel.com/content/www/us/en/design/products/platforms/details/raptor-lake-s/13th-generation-core-processors-datasheet-volume-1-of-2/005/pci-express-support/
Link: https://gitlab.freedesktop.org/drm/amd/-/issues/2663
Co-developed-by: Evan Quan <evan.quan@amd.com>
Signed-off-by: Evan Quan <evan.quan@amd.com>
Signed-off-by: Mario Limonciello <mario.limonciello@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: add watchdog timer enablement for gfx_v9_4_3

Configure SQ watchdog timer setting.

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

drm/amdkfd: Update CWSR grace period for GFX9.4.3

For GFX9.4.3, setup a reduced default CWSR grace period equal to
1000 cycles instead of 64000 cycles.

Signed-off-by: Mukul Joshi <mukul.joshi@amd.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/radeon: ERROR: "(foo*)" should be "(foo *)"

Fix five occurrences of the checkpatch.pl error:
ERROR: "(foo*)" should be "(foo *)"

Signed-off-by: Ran Sun <sunran001@208suo.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/radeon: ERROR: that open brace { should be on the previous line

Fix eleven occurrences of the checkpatch.pl error:
ERROR: that open brace { should be on the previous line

Signed-off-by: Ran Sun <sunran001@208suo.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/radeon: ERROR: "(foo*)" should be "(foo *)"

Fix four occurrences of the checkpatch.pl error:
ERROR: "(foo*)" should be "(foo *)"

Signed-off-by: Ran Sun <sunran001@208suo.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>