drm/amd: Use `amdgpu_ucode_*` helpers for GFX10

The `amdgpu_ucode_request` helper will ensure that the return code for
missing firmware is -ENODEV so that early_init can fail.

The `amdgpu_ucode_release` helper is for symmetry on unload.

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

drm/amd: Load GFX9 microcode during early_init

If GFX9 microcode is required but not available during early init, the
firmware framebuffer will have already been released and the screen will
freeze.

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

drm/amd: Use `amdgpu_ucode_*` helpers for GFX9

The `amdgpu_ucode_request` helper will ensure that the return code for
missing firmware is -ENODEV so that early_init can fail.

The `amdgpu_ucode_release` helper will provide symmetry on unload.

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

drm/amd: Remove superfluous assignment for `adev->mes.adev`

`amdgpu_mes_init` already sets `adev->mes.adev`, so there is no need
to also set it in the IP specific versions.

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

drm/amd: Use `amdgpu_ucode_*` helpers for MES

The `amdgpu_ucode_request` helper will ensure that the return code for
missing firmware is -ENODEV so that early_init can fail.

The `amdgpu_ucode_release` helper provides symmetry for releasing firmware.

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

drm/amd: Load MES microcode during early_init

Add an early_init phase to MES for fetching and validating microcode
from the filesystem.

If MES microcode is required but not available during early init, the
firmware framebuffer will have already been released and the screen will
freeze.

Move the request for MES microcode into the early_init phase
so that if it's not available, early_init will fail.

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

drm/amd/display: Remove unneeded semicolon

./drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c:7431:3-4: Unneeded semicolon
./drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c:7485:4-5: Unneeded semicolon
./drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c:7489:3-4: Unneeded semicolon

Link: https://bugzilla.openanolis.cn/show_bug.cgi?id=3635
Reported-by: Abaci Robot <abaci@linux.alibaba.com>
Signed-off-by: Yang Li <yang.lee@linux.alibaba.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Remove redundant assignment to variable dc

Smatch report warning as follows:

Line 53679: drivers/gpu/drm/amd/display/dc/core/dc_stream.c:402
dc_stream_set_cursor_position() warn: variable dereferenced before
check 'stream'

The value of 'dc' has been assigned after check whether 'stream' is
NULL. Fix it by remove redundant assignment.

Reviewed-by: Harry Wentland <harry.wentland@amd.com>
Signed-off-by: Yi Yang <yiyang13@huawei.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: fix PSR-SU/DSC interoperability support

Currently, there are issues with enabling PSR-SU + DSC. This stems from
the fact that DSC imposes a slice height on transmitted video data and
we are not conforming to that slice height in PSR-SU regions. So, pass
slice_height into su_y_granularity to feed the DSC slice height into
PSR-SU code.

Acked-by: Harry Wentland <harry.wentland@amd.com>
Signed-off-by: Hamza Mahfooz <hamza.mahfooz@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd: Load VCN microcode during early_init

Simplifies the code so that all VCN versions will get the firmware
name from `amdgpu_ucode_ip_version_decode` and then use this filename
to load microcode as part of the early_init process.

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

drm/amd: Use `amdgpu_ucode_*` helpers for VCN

The `amdgpu_ucode_request` helper will ensure that the return code for
missing firmware is -ENODEV so that early_init can fail.

The `amdgpu_ucode_release` helper is for symmetry.

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

drm/amd: Make SDMA firmware load failures less noisy.

When firmware is missing we get failures at every step.
```
[    3.855086] amdgpu 0000:04:00.0: Direct firmware load for amdgpu/green_sardine_sdma.bin failed with error -2
[    3.855087] [drm:amdgpu_sdma_init_microcode [amdgpu]] *ERROR* SDMA: Failed to init firmware "amdgpu/green_sardine_sdma.bin"
[    3.855398] [drm:sdma_v4_0_early_init [amdgpu]] *ERROR* Failed to load sdma firmware!
```
Realistically we don't need all of these, a user can tell from the first one
that request_firmware emitted what happened. Drop the others.

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

drm/amd: Convert SDMA to use `amdgpu_ucode_ip_version_decode`

Simplifies the code so that all SDMA versions will get the firmware
name from `amdgpu_ucode_ip_version_decode`.

v2: squash in fix from Srinivasan

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

drm/amd: Use `amdgpu_ucode_request` helper for SDMA

The `amdgpu_ucode_request` helper will ensure that the return code for
missing firmware is -ENODEV so that early_init can fail.

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

drm/amd: Add a new helper for loading/validating microcode

All microcode runs a basic validation after it's been loaded. Each
IP block as part of init will run both.

Introduce a wrapper for request_firmware and amdgpu_ucode_validate.
This wrapper will also remap any error codes from request_firmware
to -ENODEV.  This is so that early_init will fail if firmware couldn't
be loaded instead of the IP block being disabled.

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

drm/amd: Convert SMUv13 microcode to use `amdgpu_ucode_ip_version_decode`

The special case for the one dGPU has been moved into
`amdgpu_ucode_ip_version_decode`, so simplify this code.

Signed-off-by: Mario Limonciello <mario.limonciello@amd.com>
Acked-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Lijo Lazar <lijo.lazar@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd: Convert SMUv11 microcode to use `amdgpu_ucode_ip_version_decode`

Remove the special casing from SMU v11 code. No intended functional
changes.

Signed-off-by: Mario Limonciello <mario.limonciello@amd.com>
Acked-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Lijo Lazar <lijo.lazar@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd: Add a legacy mapping to "amdgpu_ucode_ip_version_decode"

This will allow other parts of the driver that currently special
case firmware file names to before IP version style naming to just
have a single call to `amdgpu_ucode_ip_version_decode`.

Signed-off-by: Mario Limonciello <mario.limonciello@amd.com>
Acked-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Lijo Lazar <lijo.lazar@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd: Delay removal of the firmware framebuffer

Removing the firmware framebuffer from the driver means that even
if the driver doesn't support the IP blocks in a GPU it will no
longer be functional after the driver fails to initialize.

This change will ensure that unsupported IP blocks at least cause
the driver to work with the EFI framebuffer.

Cc: stable@vger.kernel.org
Suggested-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Lijo Lazar <lijo.lazar@amd.com>
Signed-off-by: Mario Limonciello <mario.limonciello@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: Fix potential NULL dereference

Fix potential NULL dereference, in the case when "man", the resource manager
might be NULL, when/if we print debug information.

Cc: Alex Deucher <Alexander.Deucher@amd.com>
Cc: Christian König <christian.koenig@amd.com>
Cc: AMD Graphics <amd-gfx@lists.freedesktop.org>
Cc: Dan Carpenter <error27@gmail.com>
Cc: kernel test robot <lkp@intel.com>
Fixes: 7554886daa31ea ("drm/amdgpu: Fix size validation for non-exclusive domains (v4)")
Signed-off-by: Luben Tuikov <luben.tuikov@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: update ta_secureDisplay_if.h to v27.00.00.08

1. Rename securedisplay_cmd to ta_securedisplay_cmd.
2. Rename ta_securedisplay_max_phy to ta_securedisplay_phy_ID.
3. update securedisplay_cmd to ta_securedisplay_cmd

Signed-off-by: Aaron Liu <aaron.liu@amd.com>
Signed-off-by: Shane Xiao <shane.xiao@amd.com>
Reviewed-by: Alan Liu <HaoPing.Liu@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/pm: add the missing mapping for PPT feature on SMU13.0.0 and 13.0.7

Then we are able to set a new ppt limit via the hwmon interface(power1_cap).

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/pm: correct the reference clock for fan speed(rpm) calculation

Correct the reference clock as 25Mhz for SMU13 fan speed calculation.

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/display: Remove redundant logs from DSC code

[Why & How]
Remove redundant log in DSC that just add additional blank prints

Signed-off-by: Praful Swarnakar <praful.swarnakar@amd.com>
Signed-off-by: Hamza Mahfooz <hamza.mahfooz@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: fix array-bounds errors in dc_stream_remove_writeback()

The following errors occurred when using gcc 7.5.0-3ubuntu1~18.04:
drivers/gpu/drm/amd/amdgpu/../display/dc/core/dc_stream.c: In function ‘dc_stream_remove_writeback’:
drivers/gpu/drm/amd/amdgpu/../display/dc/core/dc_stream.c:543:55: warning: array subscript is above array bounds [-Warray-bounds]
     stream->writeback_info[j] = stream->writeback_info[i];
                                 ~~~~~~~~~~~~~~~~~~~~~~^~~
Add a check to make sure that num_wb_info won't overflowing the writeback_info buffer.

Fixes: 6fbefb84a98e ("drm/amd/display: Add DC core changes for DCN2")
Signed-off-by: Wen Yang <wenyang.linux@foxmail.com>
Cc: Aurabindo Pillai <aurabindo.pillai@amd.com>
Cc: Hamza Mahfooz <hamza.mahfooz@amd.com>
Cc: Guenter Roeck <linux@roeck-us.net>
Cc: Alex Deucher <alexander.deucher@amd.com>
Cc: Harry Wentland <harry.wentland@amd.com>
Cc: Leo Li <sunpeng.li@amd.com>
Cc: amd-gfx@lists.freedesktop.org
Cc: dri-devel@lists.freedesktop.org
Cc: linux-kernel@vger.kernel.org
Signed-off-by: Hamza Mahfooz <hamza.mahfooz@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: drop unnecessary NULL checks in debugfs

pipe_ctx pointer cannot be NULL when getting the address of
an element of the pipe_ctx array. Moreover, the MAX_PIPES is
defined as 6, so pipe_ctx is not NULL after the loop either.

Detected using the static analysis tool - Svace.

Signed-off-by: Alexey Kodanev <aleksei.kodanev@bell-sw.com>
Signed-off-by: Hamza Mahfooz <hamza.mahfooz@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: allow query error counters for specific IP block

amdgpu_ras_block_late_init will be invoked in IP
specific ras_late_init call as a common helper for
all the IP blocks.

However, when amdgpu_ras_block_late_init call
amdgpu_ras_query_error_count to query ras error
counters, amdgpu_ras_query_error_count queries
all the IP blocks that support ras query interface.

This results to wrong error counters cached in
software copies when there are ras errors detected
at time zero or warm reset procedure. i.e., in
sdma_ras_late_init phase, it counts on sdma/mmhub
errors, while, in mmhub_ras_late_init phase, it
still counts on sdma/mmhub errors.

The change updates amdgpu_ras_query_error_count
interface to allow query specific ip error counter.
It introduces a new input parameter: query_info. if
query_info is NULL,  it means query all the IP blocks,
otherwise, only query the ip block specified by
query_info.

Signed-off-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: Retry DDC probing on DVI on failure if we got an HPD interrupt

HPD signals on DVI ports can be fired off before the pins required for
DDC probing actually make contact, due to the pins for HPD making
contact first. This results in a HPD signal being asserted but DDC
probing failing, resulting in hotplugging occasionally failing.

Rescheduling the hotplug work for a second when we run into an HPD
signal with a failing DDC probe usually gives enough time for the rest
of the connector's pins to make contact, and fixes this issue.

Signed-off-by: xurui <xurui@kylinos.cn>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdkfd: simplify cases

A number of the gfx8 cases were the same.  Clean them
up.

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

drm/amd/display: fix dc/core/dc.c kernel-doc

Fix all kernel-doc warnings in dc/core/dc.c:

dc.c:385: warning: missing initial short description on line:
 *  dc_stream_adjust_vmin_vmax:
dc.c:392: warning: contents before sections
dc.c:399: warning: No description found for return value of 'dc_stream_adjust_vmin_vmax'
dc.c:434: warning: Excess function parameter 'adjust' description in 'dc_stream_get_last_used_drr_vtotal'
dc.c:434: warning: No description found for return value of 'dc_stream_get_last_used_drr_vtotal'
dc.c:574: warning: No description found for return value of 'dc_stream_configure_crc'
dc.c:1746: warning: No description found for return value of 'dc_commit_state_no_check'
dc.c:4991: warning: This comment starts with '/**', but isn't a kernel-doc comment. Refer Documentation/doc-guide/kernel-doc.rst
 * dc_extended_blank_supported 0 Decide whether extended blank is supported
dc.c:4991: warning: missing initial short description on line:
 * dc_extended_blank_supported 0 Decide whether extended blank is supported
dc.c:4723: warning: Function parameter or member 'dc' not described in 'dc_enable_dmub_outbox'
dc.c:4926: warning: Function parameter or member 'dc' not described in 'dc_process_dmub_dpia_hpd_int_enable'
dc.c:4926: warning: Function parameter or member 'hpd_int_enable' not described in 'dc_process_dmub_dpia_hpd_int_enable'
12 warnings

Signed-off-by: Randy Dunlap <rdunlap@infradead.org>
Reported-by: kernel test robot <lkp@intel.com>
Cc: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Cc: Alex Deucher <alexander.deucher@amd.com>
Cc: Hamza Mahfooz <hamza.mahfooz@amd.com>
Cc: Harry Wentland <harry.wentland@amd.com>
Cc: Leo Li <sunpeng.li@amd.com>
Cc: Christian König <christian.koenig@amd.com>
Cc: "Pan, Xinhui" <Xinhui.Pan@amd.com>
Cc: amd-gfx@lists.freedesktop.org
Signed-off-by: Hamza Mahfooz <hamza.mahfooz@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

Revert "drm/amd/display: Enable Freesync Video Mode by default"

This reverts commit de05abe6b9d0fe08f65d744f7f75a4cba4df27ad.

The bug referenced below was bisected to this commit. There has been no
activity toward fixing it in 3 months, so let's revert for now.

Bug: https://gitlab.freedesktop.org/drm/amd/-/issues/2162
Signed-off-by: Michel Dänzer <mdaenzer@redhat.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Remove the unused function dmub_outbox_irq_info_funcs

The function dmub_outbox_irq_info_funcs is defined in the
irq_service_dcn201.c file, but not called elsewhere, so remove this
unused function.

drivers/gpu/drm/amd/amdgpu/../display/dc/irq/dcn201/irq_service_dcn201.c:139:43: warning: unused variable 'dmub_outbox_irq_info_funcs'.

Link: https://bugzilla.openanolis.cn/show_bug.cgi?id=3520
Reported-by: Abaci Robot <abaci@linux.alibaba.com>
Signed-off-by: Jiapeng Chong <jiapeng.chong@linux.alibaba.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Fix dsc mismatch of acquire and validation of dsc engine

[Why]
We skip dsc_validation on pipes that are underlays, but in the
acquire_dsc code we don't have this check.

In certain conditions (when underlay pipe index is lower) we will assign
the dsc resource to the underlay pipe and skip the base pipe.

Now during dsc_validation we will skip the underlay pipe (this has the
dsc resource) but try to validate the base pipe(this doesn't have a dsc
resource) due to this mismatch we hit a NULLPTR

[How]
In the acquire_dsc add a check for underlay pipe so we
don't acquire a dsc resource for this pipe. This will match the
acquire/validation conditions.

Reviewed-by: Wenjing Liu <Wenjing.Liu@amd.com>
Reviewed-by: Hersen Wu <Hersenxs.Wu@amd.com>
Acked-by: Praful Swarnakar <Praful.Swarnakar@amd.com>
Signed-off-by: Bhawanpreet Lakha <Bhawanpreet.Lakha@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: remove enable ras cmd call trace

[Why]
    [   41.285804] RIP: 0010:amdgpu_ras_feature_enable+0x15c/0x310 [amdgpu]
    [   41.285945] Code: 48 89 c1 48 c7 c2 b9 f2 88 c1 48 c7 c0 c0 f2 88 c1 49 8b 3c 24 48 0f 44 d0 48 c7 c6 98 33 80 c1 e8 5f 52 75 d9 e9 fa fe ff ff <0f> 0b e9 66 ff ff ff 48 8b 3d 86 8c 0f da ba 00 04 00 00 be c0 0d
    [   41.285946] RSP: 0018:ffffbccdc72efc90 EFLAGS: 00010246
    [   41.285948] RAX: 0000000000000004 RBX: ffff931897406980 RCX: 0000000000000002
    [   41.285949] RDX: 0000000000000dc0 RSI: 0000000000000002 RDI: ffff931500042b00
    [   41.285950] RBP: ffffbccdc72efcc0 R08: 0000000000000002 R09: ffff931885b87000
    [   41.285951] R10: 0000000000ffff10 R11: 0000000000000001 R12: ffff931893e20000
    [   41.285952] R13: 0000000000000001 R14: ffff931885b87000 R15: 0000000000000000
    [   41.285953] FS:  0000000000000000(0000) GS:ffff931c6f200000(0000) knlGS:0000000000000000
    [   41.285954] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
    [   41.285955] CR2: 000055dd6f532008 CR3: 000000061b010006 CR4: 00000000003706e0
    [   41.285956] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
    [   41.285957] DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
    [   41.285958] Call Trace:
    [   41.285959]  <TASK>
    [   41.285963]  ? gfx_v11_0_early_init+0x250/0x250 [amdgpu]
    [   41.286117]  gfx_v11_0_late_init+0x8c/0xb0 [amdgpu]
    [   41.286271]  amdgpu_device_ip_late_init+0x8d/0x3c0 [amdgpu]
    [   41.286401]  amdgpu_device_init.cold+0x1677/0x1fda [amdgpu]
    [   41.286616]  ? pci_bus_read_config_word+0x4a/0x70
    [   41.286621]  ? do_pci_enable_device+0xdb/0x110
    [   41.286625]  amdgpu_driver_load_kms+0x1a/0x160 [amdgpu]
    [   41.286762]  amdgpu_pci_probe+0x18d/0x3a0 [amdgpu]
    [   41.286898]  local_pci_probe+0x4b/0x90
    [   41.286901]  work_for_cpu_fn+0x1a/0x30
    [   41.286903]  process_one_work+0x22b/0x3d0
    [   41.286905]  worker_thread+0x223/0x420
    [   41.286907]  ? process_one_work+0x3d0/0x3d0
    [   41.286908]  kthread+0x12a/0x150
    [   41.286911]  ? set_kthread_struct+0x50/0x50
    [   41.286913]  ret_from_fork+0x22/0x30

[How]
    For specific asic, only mem ecc is enabled, sram ecc is not enabled,
    but it still need to send ras enable cmd to gfx block to support
    poison mode, so add check posion mode.

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

drm/amdgpu: correct umc poison mode set value

For GFX 11.0.3, Due to security policy, there is no way to check UcFatalEn
field of UMCCH0_0_GeccCtrl to identify UMC poison mode. This is workaround
force set umc poison mode as 1 for GFX 11.0.3

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

drm/amd/display: 3.2.217

Acked-by: Aurabindo Pillai <aurabindo.pillai@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: Reorder dc_state fields to optimize clearing the struct

[why & how]
By moving bw_ctx field to the end of the dc_state the state can be
cleared more efficiently without resulting in large DML memcpy
operations, resulting in better mode enumeration performance on some
platforms.

Acked-by: Aurabindo Pillai <aurabindo.pillai@amd.com>
Signed-off-by: Aric Cyr <aric.cyr@amd.com>
Reviewed-by: Nevenko Stupar <Nevenko.Stupar@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Adding braces to prepare for future changes to behavior of if block

[Why & How]
For certain features, there will be more implementations needed in the if-block.
Braces are added as part of the preparation.

Acked-by: Aurabindo Pillai <aurabindo.pillai@amd.com>
Signed-off-by: Leo Chen <sancchen@amd.com>
Reviewed-by: Charlene Liu <Charlene.Liu@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: set ignore msa parameter only if freesync is enabled

[Why&How]

ignore_msa_timing_param is used by SubVP logic to determine if SubVP
+ DRR is possible. Linux does not support freesync on multi display
config, which results in incorrect assumption of VRR support if we
set this parameter when VRR is supported, but not enabled.

Signed-off-by: Aurabindo Pillai <aurabindo.pillai@amd.com>
Reviewed-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: allow hpo and dio encoder switching during dp retrain test

[why]
During DP2.1 LL CTS if test equipment requests to change between
DP2.1 and DP1.4 link rates, we need to swap between HPO and DIO
encoders by remapping encoder resource.

[how]
Add a function dc resource to update encoder resources and toggle
dpms state for all enabled stream associated witht the link under test.

Acked-by: Aurabindo Pillai <aurabindo.pillai@amd.com>
Signed-off-by: Wenjing Liu <wenjing.liu@amd.com>
Reviewed-by: Jun Lei <Jun.Lei@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: update pixel rate div in enable stream

[why]
Pixel rate div depends on the type of encoder
that we are enabling stream with. If we swap between
HPO and DIO encoder at the time we call enable stream
for the new encoder, we must reprogram pixel rate div
based on the new encoder type.

Acked-by: Aurabindo Pillai <aurabindo.pillai@amd.com>
Signed-off-by: Wenjing Liu <wenjing.liu@amd.com>
Reviewed-by: Jun Lei <Jun.Lei@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: move dccg programming from link hwss hpo dp to hwss

[why] dccg clock programming shouldn't be part of link hwss programming
sequence. The scope of link hwss is limited to encoder and phy
programming.

Acked-by: Aurabindo Pillai <aurabindo.pillai@amd.com>
Signed-off-by: Wenjing Liu <wenjing.liu@amd.com>
Reviewed-by: Jun Lei <Jun.Lei@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: fix dc_get_edp_link_panel_inst to only consider links with panels

This function is meant to be used on multi-edp systems and only makes sense
if only links with connected panels are considered.

Acked-by: Aurabindo Pillai <aurabindo.pillai@amd.com>
Signed-off-by: Dmytro Laktyushkin <Dmytro.Laktyushkin@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Defer DIG FIFO disable after VID stream enable

[Why]
On some monitors we see a brief flash of corruption during the
monitor disable sequence caused by FIFO being disabled in the middle
of an active DP stream.

[How]
Wait until DP vid stream is disabled before turning off the FIFO.

The FIFO reset on DP unblank should take care of clearing any FIFO
error, if any.

Acked-by: Aurabindo Pillai <aurabindo.pillai@amd.com>
Signed-off-by: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com>
Reviewed-by: Syed 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: patch cases with unknown plane state to prevent warning

[Why]
DCN301 resource function is missing function pointer to
handle cases with unknown plane state.
This causes assertion when global state is validated while
using swizzle parameter as “DC_UNKNOWN”

[How]
Add function pointer to handle and patch cases when plane
state is unknown.

Acked-by: Aurabindo Pillai <aurabindo.pillai@amd.com>
Signed-off-by: Swapnil Patel <Swapnil.Patel@amd.com>
Reviewed-by: Sung joon Kim <Sungjoon.Kim@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Turn on phantom OTG before disabling phantom pipe

[Description]
- Proper phantom pipe disable sequence was missing in
  commit_planes_for_stream
- If disabling phantom pipe, turn on phantom OTG first, and turn
  off the phantom OTG after the plane is disabled
- Also update sequence for enabling / disabling phantom streams
  (apply_ctx_to_hw). When enabling phantom pipes, enable before
  doing front end programming for phantom pipes. If disabling
  phantom pipes, disable after front end programming (i.e. after
  phantom plane disable)
- TODO: Still need to properly handle transition case when a phantom
  pipe is transitioned directly into a real pipe (need to fully disable
  the phantom pipe first)

Acked-by: Aurabindo Pillai <aurabindo.pillai@amd.com>
Signed-off-by: Alvin Lee <Alvin.Lee2@amd.com>
Reviewed-by: Jun Lei <Jun.Lei@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Improvements in secure display

[Why]
- Need error message when failing to allocating secure_display_ctx.
- Need to check if secure display context in psp is initialized or not
before using it.

[How]
- Add error message when memory allocation fail.
- Add check before accessing psp secure display context.

Acked-by: Aurabindo Pillai <aurabindo.pillai@amd.com>
Signed-off-by: Alan Liu <HaoPing.Liu@amd.com>
Reviewed-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: Uninitialized variables causing 4k60 UCLK to stay at DPM1 and not DPM0

[Why]
SwathSizePerSurfaceY[] and SwathSizePerSurfaceC[] values are uninitialized
 because we are using += instead of = operator.

[How]
Assign values in loop with = operator.

Acked-by: Aurabindo Pillai <aurabindo.pillai@amd.com>
Signed-off-by: Samson Tam <samson.tam@amd.com>
Reviewed-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: phase3 mst hdcp for multiple displays

[Why]
multiple display hdcp are enabled within event_property_validate,
event_property_update by looping all displays on mst hub. when
one of display on mst hub in unplugged or disabled, hdcp are
disabled for all displays on mst hub within hdcp_reset_display
by looping all displays of mst link. for displays still active,
their encryption status are off. kernel driver will not run hdcp
authentication again. therefore, hdcp are not enabled automatically.

[How]
within is_content_protection_different, check drm_crtc_state changes
of all displays on mst hub, if need, triger hdcp_update_display to
re-run hdcp authentication.

Acked-by: Aurabindo Pillai <aurabindo.pillai@amd.com>
Signed-off-by: hersen wu <hersenxs.wu@amd.com>
Reviewed-by: Bhawanpreet Lakha <Bhawanpreet.Lakha@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: allow zero as vram limit

This allows testing the driver without any VRAM.

Signed-off-by: Christian König <christian.koenig@amd.com>
Acked-by: Luben Tuikov <luben.tuikov@amd.com>
Acked-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: cleanup visible vram size handling

Centralize the limit handling and validation in one place instead
of spreading that around in different hw generations.

Signed-off-by: Christian König <christian.koenig@amd.com>
Acked-by: Luben Tuikov <luben.tuikov@amd.com>
Acked-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: rename vram_scratch into mem_scratch

Rename vram_scratch into mem_scratch and allow allocating it into GTT as
well.

The only problem with that is that we won't have a default page for the
system aperture any more.

Signed-off-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Luben Tuikov <luben.tuikov@amd.com>
Acked-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: use VRAM|GTT for a bunch of kernel allocations

Technically all of those can use GTT as well, no need to force things
into VRAM.

Signed-off-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Luben Tuikov <luben.tuikov@amd.com>
Acked-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: enable VCN DPG for GC IP v11.0.4

Enable VCN Dynamic Power Gating control for GC IP v11.0.4.

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

drm/amdgpu: adjust the sequence to check soft reset

1.Drop soft reset check when do should recover gpu check.
  (As it will skip gpu reset operation if some ip is hang but
   not support soft reset)
2.Check soft reset status before do soft reset when pre asic reset.
  a. If check soft reset return true, it means: some ip is hang and
     it also support soft reset, will try soft reset first.
  b. If check soft reset return false, it means:
       I.  All the ip are not hang, will skip gpu reset.
       II. Some ip is hang but not support soft reset, will skip soft
           reset and retry with full reset later.

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/amdkfd: Fix kernel warning during topology setup

This patch fixes the following kernel warning seen during
driver load by correctly initializing the p2plink attr before
creating the sysfs file:

[  +0.002865] ------------[ cut here ]------------
[  +0.002327] kobject: '(null)' (0000000056260cfb): is not initialized, yet kobject_put() is being called.
[  +0.004780] WARNING: CPU: 32 PID: 1006 at lib/kobject.c:718 kobject_put+0xaa/0x1c0
[  +0.001361] Call Trace:
[  +0.001234]  <TASK>
[  +0.001067]  kfd_remove_sysfs_node_entry+0x24a/0x2d0 [amdgpu]
[  +0.003147]  kfd_topology_update_sysfs+0x3d/0x750 [amdgpu]
[  +0.002890]  kfd_topology_add_device+0xbd7/0xc70 [amdgpu]
[  +0.002844]  ? lock_release+0x13c/0x2e0
[  +0.001936]  ? smu_cmn_send_smc_msg_with_param+0x1e8/0x2d0 [amdgpu]
[  +0.003313]  ? amdgpu_dpm_get_mclk+0x54/0x60 [amdgpu]
[  +0.002703]  kgd2kfd_device_init.cold+0x39f/0x4ed [amdgpu]
[  +0.002930]  amdgpu_amdkfd_device_init+0x13d/0x1f0 [amdgpu]
[  +0.002944]  amdgpu_device_init.cold+0x1464/0x17b4 [amdgpu]
[  +0.002970]  ? pci_bus_read_config_word+0x43/0x80
[  +0.002380]  amdgpu_driver_load_kms+0x15/0x100 [amdgpu]
[  +0.002744]  amdgpu_pci_probe+0x147/0x370 [amdgpu]
[  +0.002522]  local_pci_probe+0x40/0x80
[  +0.001896]  work_for_cpu_fn+0x10/0x20
[  +0.001892]  process_one_work+0x26e/0x5a0
[  +0.002029]  worker_thread+0x1fd/0x3e0
[  +0.001890]  ? process_one_work+0x5a0/0x5a0
[  +0.002115]  kthread+0xea/0x110
[  +0.001618]  ? kthread_complete_and_exit+0x20/0x20
[  +0.002422]  ret_from_fork+0x1f/0x30
[  +0.001808]  </TASK>
[  +0.001103] irq event stamp: 59837
[  +0.001718] hardirqs last  enabled at (59849): [<ffffffffb30fab12>] __up_console_sem+0x52/0x60
[  +0.004414] hardirqs last disabled at (59860): [<ffffffffb30faaf7>] __up_console_sem+0x37/0x60
[  +0.004414] softirqs last  enabled at (59654): [<ffffffffb307d9c7>] irq_exit_rcu+0xd7/0x130
[  +0.004205] softirqs last disabled at (59649): [<ffffffffb307d9c7>] irq_exit_rcu+0xd7/0x130
[  +0.004203] ---[ end trace 0000000000000000 ]---

Fixes: 0f28cca87e9a ("drm/amdkfd: Extend KFD device topology to surface peer-to-peer links")
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/amdgpu: skip mes self test after s0i3 resume for MES IP v11.0

MES is part of gfxoff and MES suspend and resume are skipped for S0i3.
But the mes_self_test call path is still in the amdgpu_device_ip_late_init.
it's should also be skipped for s0ix as no hardware re-initialization
happened.

Besides, mes_self_test will free the BO that triggers a lot of warning
messages while in the suspend state.

[   81.656085] WARNING: CPU: 2 PID: 1550 at drivers/gpu/drm/amd/amdgpu/amdgpu_object.c:425 amdgpu_bo_free_kernel+0xfc/0x110 [amdgpu]
[   81.679435] Call Trace:
[   81.679726]  <TASK>
[   81.679981]  amdgpu_mes_remove_hw_queue+0x17a/0x230 [amdgpu]
[   81.680857]  amdgpu_mes_self_test+0x390/0x430 [amdgpu]
[   81.681665]  mes_v11_0_late_init+0x37/0x50 [amdgpu]
[   81.682423]  amdgpu_device_ip_late_init+0x53/0x280 [amdgpu]
[   81.683257]  amdgpu_device_resume+0xae/0x2a0 [amdgpu]
[   81.684043]  amdgpu_pmops_resume+0x37/0x70 [amdgpu]
[   81.684818]  pci_pm_resume+0x5c/0xa0
[   81.685247]  ? pci_pm_thaw+0x90/0x90
[   81.685658]  dpm_run_callback+0x4e/0x160
[   81.686110]  device_resume+0xad/0x210
[   81.686529]  async_resume+0x1e/0x40
[   81.686931]  async_run_entry_fn+0x33/0x120
[   81.687405]  process_one_work+0x21d/0x3f0
[   81.687869]  worker_thread+0x4a/0x3c0
[   81.688293]  ? process_one_work+0x3f0/0x3f0
[   81.688777]  kthread+0xff/0x130
[   81.689157]  ? kthread_complete_and_exit+0x20/0x20
[   81.689707]  ret_from_fork+0x22/0x30
[   81.690118]  </TASK>
[   81.690380] ---[ end trace 0000000000000000 ]---

v2: make the comment clean and use adev->in_s0ix instead of
adev->suspend

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

drm/amd/display: fix some coding style issues

Fix the following checkpatch checks in amdgpu_dm.c

CHECK: Prefer kernel type 'u8' over 'uint8_t'
CHECK: Prefer kernel type 'u32' over 'uint32_t'
CHECK: Prefer kernel type 'u64' over 'uint64_t'
CHECK: Prefer kernel type 's32' over 'int32_t'

Signed-off-by: Srinivasan Shanmugam <srinivasan.shanmugam@amd.com>
Reviewed-by: Harry Wentland <harry.wentland@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/pm: correct the fan speed retrieving in PWM for some SMU13 asics

For SMU 13.0.0 and 13.0.7, the output from PMFW is in percent. Driver
need to convert that into correct PMW(255) based.

Signed-off-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: save restore hdcp state when display is unplugged from mst hub

[Why]
connector hdcp properties are lost after display is
unplgged from mst hub. connector is destroyed with
dm_dp_mst_connector_destroy. when display is plugged
back, hdcp is not desired and it wouldnt be enabled.

[How]
save hdcp properties into hdcp_work within
amdgpu_dm_atomic_commit_tail. If the same display is
plugged back with same display index, its hdcp
properties will be retrieved from hdcp_work within
dm_dp_mst_get_modes.

Acked-by: Aurabindo Pillai <aurabindo.pillai@amd.com>
Signed-off-by: hersen wu <hersenxs.wu@amd.com>
Reviewed-by: Bhawanpreet Lakha <Bhawanpreet.Lakha@amd.com>
Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/pm: bump SMU13.0.0 driver_if header to version 0x34

To fit the latest PMFW and suppress the warning emerged on driver loading.

Signed-off-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/pm: add support for WINDOW3D profile mode on SMU13.0.0

Add the support for WINDOW3D profile mode as for other profile modes.

Signed-off-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/amdgpu: skip MES for S0ix as well since it's part of GFX

It's also part of gfxoff.

Cc: stable@vger.kernel.org # 6.0, 6.1
Reviewed-by: Mario Limonciello <mario.limonciello@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

Revert "drm/amdgpu: force exit gfxoff on sdma resume for rmb s0ix"

This reverts commit e5d59cfa330523e47cba62a496864acc3948fc27.

This is no longer needed since we no longer suspend SDMA during
S0ix.

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

Revert "drm/amdgpu: disallow gfxoff until GC IP blocks complete s2idle resume"

This reverts commit f543d28687480fad06b708bc6e0b0b6ec953b078.

This is no longer needed since we no longer touch SDMA 5.x for s0i3.

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

drm/amdgpu: for S0ix, skip SDMA 5.x+ suspend/resume

SDMA 5.x is part of the GFX block so it's controlled via
GFXOFF.  Skip suspend as it should be handled the same
as GFX.

v2: drop SDMA 4.x.  That requires special handling.

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

drm/amdgpu: don't mess with SDMA clock or powergating in S0ix

It's handled by GFXOFF for SDMA 5.x and SMU saves the state on
SDMA 4.x.

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

drm/amdgpu/gmc11: don't touch gfxhub registers during S0ix

gfxhub registers are part of gfx IP and should not need to be
changed.  Doing so without disabling gfxoff can hang the gfx IP.

v2: add comments explaining why we can skip the interrupt
    control for S0i3

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

drm/amdgpu/gmc10: don't touch gfxhub registers during S0ix

gfxhub registers are part of gfx IP and should not need to be
changed.  Doing so without disabling gfxoff can hang the gfx IP.

v2: add comments explaining why we can skip the interrupt
    control for S0i3

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

drm/amdgpu/gmc9: don't touch gfxhub registers during S0ix

gfxhub registers are part of gfx IP and should not need to be
changed.  Doing so without disabling gfxoff can hang the gfx IP.

v2: add comments explaining why we can skip the interrupt
    control for S0i3

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

drm/amd/pm: avoid large variable on kernel stack

The activity_monitor_external[] array is too big to fit on the
kernel stack, resulting in this warning with clang:

drivers/gpu/drm/amd/amdgpu/../pm/swsmu/smu13/smu_v13_0_7_ppt.c:1438:12: error: stack frame size (1040) exceeds limit (1024) in 'smu_v13_0_7_get_power_profile_mode' [-Werror,-Wframe-larger-than]

Use dynamic allocation instead. It should also be possible to
have single element here instead of the array, but this seems
easier.

v2: fix up argument to sizeof() (Alex)

Fixes: 334682ae8151 ("drm/amd/pm: enable workload type change on smu_v13_0_7")
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: fix duplicate assignments

The .set_odm_combine callback pointer was added twice, causing
a harmless -Wextra warning:

drivers/gpu/drm/amd/amdgpu/../display/dc/dcn314/dcn314_optc.c:258:36: error: initialized field overwritten [-Werror=override-init]
  258 |                 .set_odm_combine = optc314_set_odm_combine,
      |                                    ^~~~~~~~~~~~~~~~~~~~~~~
drivers/gpu/drm/amd/amdgpu/../display/dc/dcn314/dcn314_optc.c:258:36: note: (near initialization for 'dcn314_tg_funcs.set_odm_combine')

Fixes: 5ade1b951dec ("drm/amd/display: Add OTG/ODM functions")
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdkfd: Fix double release compute pasid

If kfd_process_device_init_vm returns failure after vm is converted to
compute vm and vm->pasid set to compute pasid, KFD will not take
pdd->drm_file reference. As a result, drm close file handler maybe
called to release the compute pasid before KFD process destroy worker to
release the same pasid and set vm->pasid to zero, this generates below
WARNING backtrace and NULL pointer access.

Add helper amdgpu_amdkfd_gpuvm_set_vm_pasid and call it at the last step
of kfd_process_device_init_vm, to ensure vm pasid is the original pasid
if acquiring vm failed or is the compute pasid with pdd->drm_file
reference taken to avoid double release same pasid.

 amdgpu: Failed to create process VM object
 ida_free called for id=32770 which is not allocated.
 WARNING: CPU: 57 PID: 72542 at ../lib/idr.c:522 ida_free+0x96/0x140
 RIP: 0010:ida_free+0x96/0x140
 Call Trace:
  amdgpu_pasid_free_delayed+0xe1/0x2a0 [amdgpu]
  amdgpu_driver_postclose_kms+0x2d8/0x340 [amdgpu]
  drm_file_free.part.13+0x216/0x270 [drm]
  drm_close_helper.isra.14+0x60/0x70 [drm]
  drm_release+0x6e/0xf0 [drm]
  __fput+0xcc/0x280
  ____fput+0xe/0x20
  task_work_run+0x96/0xc0
  do_exit+0x3d0/0xc10

 BUG: kernel NULL pointer dereference, address: 0000000000000000
 RIP: 0010:ida_free+0x76/0x140
 Call Trace:
  amdgpu_pasid_free_delayed+0xe1/0x2a0 [amdgpu]
  amdgpu_driver_postclose_kms+0x2d8/0x340 [amdgpu]
  drm_file_free.part.13+0x216/0x270 [drm]
  drm_close_helper.isra.14+0x60/0x70 [drm]
  drm_release+0x6e/0xf0 [drm]
  __fput+0xcc/0x280
  ____fput+0xe/0x20
  task_work_run+0x96/0xc0
  do_exit+0x3d0/0xc10

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

drm/amdkfd: Fix kfd_process_device_init_vm error handling

Should only destroy the ib_mem and let process cleanup worker to free
the outstanding BOs. Reset the pointer in pdd->qpd structure, to avoid
NULL pointer access in process destroy worker.

 BUG: kernel NULL pointer dereference, address: 0000000000000010
 Call Trace:
  amdgpu_amdkfd_gpuvm_unmap_gtt_bo_from_kernel+0x46/0xb0 [amdgpu]
  kfd_process_device_destroy_cwsr_dgpu+0x40/0x70 [amdgpu]
  kfd_process_destroy_pdds+0x71/0x190 [amdgpu]
  kfd_process_wq_release+0x2a2/0x3b0 [amdgpu]
  process_one_work+0x2a1/0x600
  worker_thread+0x39/0x3d0

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

drm/amdgpu: Add poison mode query for df v4_3

Add poison mode query support on df v4_3.

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: Add df v4_3 headers

Add df v4_3 header files.

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: bump minor version number for DEV_INFO and SENSOR IOCTLs update

Update AMDGPU_INFO_DEV_INFO IOCTL for minimum engine and memory clock.
And update AMDGPU_INFO_SENSOR IOCTL for PEAK_PSTATE engine and memory clock.
User applications can better utilize these IOCTLs to get needed informations.

Increase the minor version number to indicate that the new flags are
available.

Proposed mesa patch:
https://gitlab.freedesktop.org/mesa/drm/-/merge_requests/278

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: expose the minimum shader/memory clock frequency

Otherwise, some UMD tools will treate them as 0 at default while
actually they are not.

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: expose peak profiling mode shader/memory clocks

Expose those informations to UMD who need them as for standard
profiling mode.

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/pm: fulfill powerplay peak profiling mode shader/memory clock settings

Enable peak profiling mode shader/memory clock reporting for powerplay
framework.

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/pm: fulfill swsmu peak profiling mode shader/memory clock settings

Enable peak profiling mode shader/memory clocks reporting for swsmu
framework.

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/pm: drop unused SMU v13 API

The API is not in use. And it's unlikely to be used in
the future either.

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/display: Add missing Adaptive Sync DPCD definitions

The missing DPCD definitions from DP2.0 spec is as follows:

DOWNSPREAD_CTRL (107h):
FIXED_VTOTAL_AS_SDP_EN_IN_PR_ACTIVE (bit 6)
For sink devices that support Adaptive-Sync operation
and Panel Replay

DPRX_FEATURE_ENUMERATION_LIST_CONT_1 (2214h):
ADAPTIVE_SYNC_SDP_SUPPORTED (bit 0)
Bit to check sink device has Adaptive-Sync capability
AS_SDP_FIRST_HALF_LINE_OR_3840_PIXEL_CYCLE_WINDOW_NOT_SUPPORTED (bit 1)
A sink device that clears this bit will generate VSync pulse
leading edge of the HDMI output on the line count at which
Adaptive-Sync SDP is received as long as source device transmits
Adaptive-Sync SDP either in first line or first 3840 pixel cycles
of the line whichever occurs first.
VSC_EXT_SDP_FRAMEWORK_VERSION_1_SUPPORTED (bit 4)
Bit to check sink device has SDP framework version 1 capability

Signed-off-by: Sung Joon Kim <Sungjoon.Kim@amd.com>
Reviewed-by: Harry Wentland <harry.wentland@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/radeon: Replace 1-element arrays with flexible-array members

One-element arrays are deprecated, and we are replacing them with
flexible array members instead. So, replace one-element array with
flexible-array member in structs _ATOM_DISPLAY_OBJECT_PATH,
_ATOM_DISPLAY_OBJECT_PATH_TABLE, _ATOM_OBJECT_TABLE, GOP_VBIOS_CONTENT
_ATOM_GPIO_VOLTAGE_OBJECT_V3 and refactor the rest of the code accordingly.

It's worth mentioning that doing a build before/after this patch
results in no binary output differences.

This helps with the ongoing efforts to tighten the FORTIFY_SOURCE
routines on memcpy() and help us make progress towards globally
enabling -fstrict-flex-arrays=3 [1].

Link: https://github.com/KSPP/linux/issues/79
Link: https://github.com/KSPP/linux/issues/239
Link: https://gcc.gnu.org/bugzilla/show_bug.cgi?id=101836
Signed-off-by: Paulo Miguel Almeida <paulo.miguel.almeida.rodenas@gmail.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: define RAS query poison mode function

1. no need to query poison mode on SRIOV guest side, host can handle it.
2. define the function to simplify code.

v2: rename amdgpu_ras_poison_mode_query to amdgpu_ras_query_poison_mode.

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/amdgpu: update VCN/JPEG RAS setting

Support VCN/JPEG RAS in both bare metal and SRIOV environment.

v2: update commit description.

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/amdgpu: skip RAS error injection in SRIOV

Injection on guest is not allowed.

v2: return directly in SRIOV environment.

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/amdgpu: add VCN poison consumption handler for SRIOV

Inform host and let host handle consumption interrupt.

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/amdgpu: add RAS poison consumption handler for SRIOV

Send message to PF if VF receives RAS poison consumption interrupt.

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/amdgpu: add RAS poison consumption handler for NV SRIOV

Send handling request to host.

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/amdgpu: add RAS poison consumption handler for AI SRIOV

Send message to host and host will handle it.

v2: split the patch into two parts, one is for mxgpu ai and another one
is for common poison consumption handler.

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/amd/display: Block FPO / SubVP (DRR) on HDMI VRR configs

[Description]
- Current policy does not support HDMI VRR by default, so we
  cannot enable FPO / SubVP (DRR) cases

Reviewed-by: Nevenko Stupar <Nevenko.Stupar@amd.com>
Reviewed-by: Jun Lei <Jun.Lei@amd.com>
Acked-by: Jasdeep Dhillon <jdhillon@amd.com>
Signed-off-by: Alvin Lee <Alvin.Lee2@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: 3.2.216

This version brings along following fixes:

-Fix array index out of bound error
-Speed up DML fast vadlaite
-Implement multiple secure display
-MST HDCP for multiple display
-Add DPIA notification
-Add support for three new square pattern variant

Reviewed-by: Bhawanpreet Lakha <bhawanpreet.lakha@amd.com>
Acked-by: Jasdeep Dhillon <jdhillon@amd.com>
Signed-off-by: Aric Cyr <aric.cyr@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Demote Error Level When ODM Transition Supported

[Why && How]
On dcn32, HW supports odm transition in fast update. Hence this
error message is considered false positive. Downgrade the error level
to avoid catching unnecessary attention.

Reviewed-by: Dillon Varone <Dillon.Varone@amd.com>
Acked-by: Jasdeep Dhillon <jdhillon@amd.com>
Signed-off-by: Fangzhi Zuo <Jerry.Zuo@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: add support for three new square pattern variants from DP2.1 specs

[why]
DP2.1 specs has brought 3 new variants of sqaure patterns with different
pre-shoot and de-emphasis equalization requirements. The commit adds
logic to identify these variants and apply corresponding eqaulization
requirements into hardware lane settings.

Reviewed-by: George Shen <George.Shen@amd.com>
Acked-by: Jasdeep Dhillon <jdhillon@amd.com>
Signed-off-by: Wenjing Liu <wenjing.liu@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: run subvp validation with supported vlevel

[WHY]
Subvp portion validation currently assumes that if vlevel provided does not
support pstate, then none will, and so subvp is not used.

[HOW]
After get vlevel, use lowest vlevel that supports pstate if it
exists, and use that for subvp validation.

Reviewed-by: Alvin Lee <Alvin.Lee2@amd.com>
Reviewed-by: Jun Lei <Jun.Lei@amd.com>
Acked-by: Jasdeep Dhillon <jdhillon@amd.com>
Signed-off-by: Dillon Varone <Dillon.Varone@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Check for PSR in no memory request case

[Why]
When we have a PSR display, we will not be requesting data from memory anymore.
So we report back true for no memory request case.

[How]
Check for PSR by checking PSR version in link settings

Reviewed-by: Alvin Lee <Alvin.Lee2@amd.com>
Acked-by: Jasdeep Dhillon <jdhillon@amd.com>
Signed-off-by: Samson Tam <Samson.Tam@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Clear link res when merging a pipe split

[Description]
- When merging a pipe that was previously pipe split, we need
  to also clear the link resources or the next stream/plane that
  uses the pipe may have an incorrect link resource state

Reviewed-by: Wenjing Liu <Wenjing.Liu@amd.com>
Reviewed-by: Nevenko Stupar <Nevenko.Stupar@amd.com>
Acked-by: Jasdeep Dhillon <jdhillon@amd.com>
Signed-off-by: Alvin Lee <Alvin.Lee2@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Add DPIA NOTIFICATION logic

[WHY]
Adding the new DPIA NOTIFY packets from DMUB
As per the design with Cruise to account for
250ms response delay otherwise

[HOW]
Added th DPIA NOTIFY logic as per DMUB logic

Reviewed-by: Nicholas Kazlauskas <Nicholas.Kazlauskas@amd.com>
Acked-by: Jasdeep Dhillon <jdhillon@amd.com>
Signed-off-by: Mustapha Ghaddar <mghaddar@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Clear MST topology if it fails to resume

[Why]
In case of failure to resume MST topology after suspend, an emtpty
mst tree prevents further mst hub detection on the same connector.
That causes the issue with MST hub hotplug after it's been unplug in
suspend.

[How]
Stop topology manager on the connector after detecting DM_MST failure.

Reviewed-by: Wayne Lin <Wayne.Lin@amd.com>
Acked-by: Jasdeep Dhillon <jdhillon@amd.com>
Signed-off-by: Roman Li <roman.li@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>