drm/amdgpu: drm_device to amdgpu_device by inline-f (v2)

Get the amdgpu_device from the DRM device by use
of an inline function, drm_to_adev(). The inline
function resolves a pointer to struct drm_device
to a pointer to struct amdgpu_device.

v2: Use a typed visible static inline function
    instead of an invisible macro.

Signed-off-by: Luben Tuikov <luben.tuikov@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: enable HDP clock gatting

Enabe HDP SD/DS clock gatting in Renoir series.

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

drm/amdgpu: enable ATHUB clock gatting

Enable ATHUB clock gatting set in Renoir series.

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

drm/amd/pm: set VCN pg per instances

When deciding whether to set pg for vcn1, instances
number is more generic than chip name.

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

drm/amdgpu: annotate a false positive recursive locking

Re-apply commit 72e14ebf9fc09e33b28b70f00a2ed9821c198633

[  584.110304] ============================================
[  584.110590] WARNING: possible recursive locking detected
[  584.110876] 5.6.0-deli-v5.6-2848-g3f3109b0e75f #1 Tainted: G           OE
[  584.111164] --------------------------------------------
[  584.111456] kworker/38:1/553 is trying to acquire lock:
[  584.111721] ffff9b15ff0a47a0 (&adev->reset_sem){++++}, at: amdgpu_device_gpu_recover+0x262/0x1030 [amdgpu]
[  584.112112]
               but task is already holding lock:
[  584.112673] ffff9b1603d247a0 (&adev->reset_sem){++++}, at: amdgpu_device_gpu_recover+0x262/0x1030 [amdgpu]
[  584.113068]
               other info that might help us debug this:
[  584.113689]  Possible unsafe locking scenario:

[  584.114350]        CPU0
[  584.114685]        ----
[  584.115014]   lock(&adev->reset_sem);
[  584.115349]   lock(&adev->reset_sem);
[  584.115678]
                *** DEADLOCK ***

[  584.116624]  May be due to missing lock nesting notation

[  584.117284] 4 locks held by kworker/38:1/553:
[  584.117616]  #0: ffff9ad635c1d348 ((wq_completion)events){+.+.}, at: process_one_work+0x21f/0x630
[  584.117967]  #1: ffffac708e1c3e58 ((work_completion)(&con->recovery_work)){+.+.}, at: process_one_work+0x21f/0x630
[  584.118358]  #2: ffffffffc1c2a5d0 (&tmp->hive_lock){+.+.}, at: amdgpu_device_gpu_recover+0xae/0x1030 [amdgpu]
[  584.118786]  #3: ffff9b1603d247a0 (&adev->reset_sem){++++}, at: amdgpu_device_gpu_recover+0x262/0x1030 [amdgpu]
[  584.119222]
               stack backtrace:
[  584.119990] CPU: 38 PID: 553 Comm: kworker/38:1 Kdump: loaded Tainted: G           OE     5.6.0-deli-v5.6-2848-g3f3109b0e75f #1
[  584.120782] Hardware name: Supermicro SYS-7049GP-TRT/X11DPG-QT, BIOS 3.1 05/23/2019
[  584.121223] Workqueue: events amdgpu_ras_do_recovery [amdgpu]
[  584.121638] Call Trace:
[  584.122050]  dump_stack+0x98/0xd5
[  584.122499]  __lock_acquire+0x1139/0x16e0
[  584.122931]  ? trace_hardirqs_on+0x3b/0xf0
[  584.123358]  ? cancel_delayed_work+0xa6/0xc0
[  584.123771]  lock_acquire+0xb8/0x1c0
[  584.124197]  ? amdgpu_device_gpu_recover+0x262/0x1030 [amdgpu]
[  584.124599]  down_write+0x49/0x120
[  584.125032]  ? amdgpu_device_gpu_recover+0x262/0x1030 [amdgpu]
[  584.125472]  amdgpu_device_gpu_recover+0x262/0x1030 [amdgpu]
[  584.125910]  ? amdgpu_ras_error_query+0x1b8/0x2a0 [amdgpu]
[  584.126367]  amdgpu_ras_do_recovery+0x159/0x190 [amdgpu]
[  584.126789]  process_one_work+0x29e/0x630
[  584.127208]  worker_thread+0x3c/0x3f0
[  584.127621]  ? __kthread_parkme+0x61/0x90
[  584.128014]  kthread+0x12f/0x150
[  584.128402]  ? process_one_work+0x630/0x630
[  584.128790]  ? kthread_park+0x90/0x90
[  584.129174]  ret_from_fork+0x3a/0x50

Each adev has owned lock_class_key to avoid false positive
recursive locking.

v2:
1. register adev->lock_key into lockdep, otherwise lockdep will
report the below warning

[ 1216.705820] BUG: key ffff890183b647d0 has not been registered!
[ 1216.705924] ------------[ cut here ]------------
[ 1216.705972] DEBUG_LOCKS_WARN_ON(1)
[ 1216.705997] WARNING: CPU: 20 PID: 541 at kernel/locking/lockdep.c:3743 lockdep_init_map+0x150/0x210

v3:
change to use down_write_nest_lock to annotate the false dead-lock
warning.

Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Signed-off-by: Dennis Li <Dennis.Li@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: refine create and release logic of hive info

Change to dynamically create and release hive info object,
which help driver support more hives in the future.

v2:
Change to save hive object pointer in adev, to avoid locking
xgmi_mutex every time when calling amdgpu_get_xgmi_hive.

v3:
1. Change type of hive object pointer in adev from void* to
amdgpu_hive_info*.
2. remove unnecessary variable initialization.

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

drm/amdgpu: refine message print for devices of hive

Using dev_xxx instead of DRM_xxx/pr_xxx to indicate which device
of a hive is the message for.

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

drm/amdgpu: fix the nullptr issue when reenter GPU recovery

in single gpu system, if driver reenter gpu recovery,
amdgpu_device_lock_adev will return false, but hive is
nullptr now.

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

drm/amdgpu: change reset lock from mutex to rw_semaphore

clients don't need reset-lock for synchronization when no
GPU recovery.

v2:
change to return the return value of down_read_killable.

v3:
if GPU recovery begin, VF ignore FLR notification.

Reviewed-by: Monk Liu <monk.liu@amd.com>
Acked-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Dennis Li <Dennis.Li@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/pm: enable run_btc callback for sienna_cichlid

DC BTC support for sienna_cichlid is added, it provides
the DC tolerance and aging measurements.

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

drivers: gpu: amd: Initialize amdgpu_dm_backlight_caps object to 0 in amdgpu_dm_update_backlight_caps

In `amdgpu_dm_update_backlight_caps()`, there is a local
`amdgpu_dm_backlight_caps` object that is filled in by
`amdgpu_acpi_get_backlight_caps()`. However, this object is
uninitialized before the call and hence the subsequent check for
aux_support can fail since it is not initialized by
`amdgpu_acpi_get_backlight_caps()` as well. This change initializes
this local `amdgpu_dm_backlight_caps` object to 0.

Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Furquan Shaikh <furquan@google.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/pm: Remove unnecessary cast

In init_powerplay_table_information() the value returned from kmalloc()
is cast unnecessarily. Remove cast.

Issue identified with Coccinelle.

Signed-off-by: Alex Dewar <alex.dewar90@gmail.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/powerplay: remove duplicate include

Remove asic_reg/nbio/nbio_6_1_offset.h which is included more than once

Reviewed-by: Evan Quan <evan.quan@amd.com>
Reported-by: Hulk Robot <hulkci@huawei.com>
Signed-off-by: Wang Hai <wanghai38@huawei.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: remove unintended executable mode

Besides the intended change, commit 4cc1178e166a ("drm/amdgpu: replace DRM
prefix with PCI device info for gfx/mmhub") also set the source files
mmhub_v1_0.c and gfx_v9_4.c to be executable, i.e., changed fromold mode
644 to new mode 755.

Commit 241b2ec9317e ("drm/amd/display: Add dcn30 Headers (v2)") added the
four header files {dpcs,dcn}_3_0_0_{offset,sh_mask}.h as executable, i.e.,
mode 755.

Set to the usual modes for source and headers files and clean up those
mistakes. No functional change.

Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Lukas Bulwahn <lukas.bulwahn@gmail.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: refine codes to avoid reentering GPU recovery

if other threads have holden the reset lock, recovery will
fail to try_lock. Therefore we introduce atomic hive->in_reset
and adev->in_gpu_reset, to avoid reentering GPU recovery.

v2:
drop "? true : false" in the definition of amdgpu_in_reset

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

drm/amd/display: Reject overlay plane configurations in multi-display scenarios

[Why]
These aren't stable on some platform configurations when driving
multiple displays, especially on higher resolution.

In particular the delay in asserting p-state and validating from
x86 outweights any power or performance benefit from the hardware
composition.

Under some configurations this will manifest itself as extreme stutter
or unresponsiveness especially when combined with cursor movement.

[How]
Disable these for now. Exposing overlays to userspace doesn't guarantee
that they'll be able to use them in any and all configurations and it's
part of the DRM contract to have userspace gracefully handle validation
failures when they occur.

Valdiation occurs as part of DC and this in particular affects RV, so
disable this in dcn10_global_validation.

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

drm/amdkfd: sparse: Fix warning in reading SDMA counters

Add __user annotation to fix related sparse warning while reading
SDMA counters from userland.
Also, rework the read SDMA counters function by removing redundant
checks.

Reported-by: kernel test robot <lkp@intel.com>
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/amd/display: use correct scale for actual_brightness

Documentation for sysfs backlight level interface requires that
values in both 'brightness' and 'actual_brightness' files are
interpreted to be in range from 0 to the value given in the
'max_brightness' file.

With amdgpu, max_brightness gives 255, and values written by the user
into 'brightness' are internally rescaled to a wider range. However,
reading from 'actual_brightness' gives the raw register value without
inverse rescaling. This causes issues for various userspace tools such
as PowerTop and systemd that expect the value to be in the correct
range.

Introduce a helper to retrieve internal backlight range. Use it to
reimplement 'convert_brightness' as 'convert_brightness_from_user' and
introduce 'convert_brightness_to_user'.

Bug: https://bugzilla.kernel.org/show_bug.cgi?id=203905
Bug: https://gitlab.freedesktop.org/drm/amd/-/issues/1242
Cc: Alex Deucher <alexander.deucher@amd.com>
Cc: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alexander Monakov <amonakov@ispras.ru>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: should check error using DC_OK

core_link_read_dpcd returns only DC_OK(1) and DC_ERROR_UNEXPECTED(-1),
the caller should check error using DC_OK instead of checking against 0

Signed-off-by: Tong Zhang <ztong0001@gmail.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: fix potential integer overflow when shifting 32 bit variable bl_pwm

The 32 bit unsigned integer bl_pwm is being shifted using 32 bit arithmetic
and then being assigned to a 64 bit unsigned integer.  There is a potential
for a 32 bit overflow so cast bl_pwm to enforce a 64 bit shift operation
to avoid this.

Addresses-Coverity: ("unintentional integer overflow")
Fixes: 3ba01817365c ("drm/amd/display: Move panel_cntl specific register from abm to panel_cntl.")
Signed-off-by: Colin Ian King <colin.king@canonical.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu/pm: only hide average power on SI and pre-RENOIR APUs

We can get this on RENOIR and newer via the SMU metrics
table.

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

drm/amdgpu/swsmu: implement power metrics for RENOIR

Grab the data from the SMU metrics table.

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

drm/amdgpu/swsmu: implement voltage metrics for RENOIR

Grab the data from the SMU metrics table.

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

drm/amdgpu/pm: remove duplicate check

FAMILY_KV is APUs and we already check for APUs.

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

drm/amdgpu/jpeg: remove redundant check when it returns

Fix warning from kernel test robot
v2: remove the local variable as well

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

drm/amdgpu: Limit the error info print rate

Use function printk_ratelimit to limit the print rate.

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

drm/amdgpu: Fix repeatly flr issue

Only for no job running test case need to do recover in
flr notification.
For having job in mirror list, then let guest driver to
hit job timeout, and then do recover.

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

drm/amd/pm: add SMU11 common deep sleep control interface

Considering the same logic can be applied to Arcturus, Navi1X
and Sienna Cichlid.

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: disable/enable deep sleep features on UMD pstate enter/exit

Add deep sleep disablement/enablement on UMD pstate entering/exiting.

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: add SMU11 common gfx ulv control interface

Considering the same logic can be applied to Arcturus, Navi1X
and Sienna Cichlid.

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: disable/enable gfx ulv on UMD pstate enter/exit

Add gfx ulv disablement/enablement on UMD pstate entering/exiting.

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: update driver if version for navy_flounder

It's in accordance with pmfw 65.7.0 for navy_flounder.

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

Revert "drm/amdgpu: disable gfxoff for navy_flounder"

This reverts commit ba4e049e63b607ac2e0c070b1406826390d5047e.
Newly released sdma fw (51.52) provides a fix for the issue.

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

drm/scheduler: Remove priority macro INVALID (v2)

Remove DRM_SCHED_PRIORITY_INVALID. We no longer
carry around an invalid priority and cut it off
at the source.

Backwards compatibility behaviour of AMDGPU CTX
IOCTL passing in garbage for context priority
from user space and then mapping that to
DRM_SCHED_PRIORITY_NORMAL is preserved.

v2: Revert "res"  --> "r" and
           "prio" --> "priority".

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/scheduler: Scheduler priority fixes (v2)

Remove DRM_SCHED_PRIORITY_LOW, as it was used
in only one place.

Rename and separate by a line
DRM_SCHED_PRIORITY_MAX to DRM_SCHED_PRIORITY_COUNT
as it represents a (total) count of said
priorities and it is used as such in loops
throughout the code. (0-based indexing is the
the count number.)

Remove redundant word HIGH in priority names,
and rename *KERNEL* to *HIGH*, as it really
means that, high.

v2: Add back KERNEL and remove SW and HW,
    in lieu of a single HIGH between NORMAL and KERNEL.

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/amd/display: Add dsc_to_stream_resource for dcn3

Without this, enabling dsc will cause a nullptr

Reviewed-by: Mikita Lipski <Mikita.Lipski@amd.com>
Signed-off-by: Bhawanpreet Lakha <Bhawanpreet.Lakha@amd.com>
Acked-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdkfd: Initialize SDMA activity counter to 0

To prevent reporting erroneous SDMA usage, initialize SDMA
activity counter to 0 before using.

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/amd/display: Add DSC_DBG_EN shift/mask for dcn3

This field is not defined for DCN3

Signed-off-by: Bhawanpreet Lakha <Bhawanpreet.Lakha@amd.com>
Reviewed-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: [FW Promotion] Release 0.0.29

[Header Changes]
 - Add command for panel power seq control

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

drm/amd/display: Mask DPCD field to avoid reading unsupported field value

[Why]
DSC will be disabled if DPCD 0006F[6:3] is set to a non-zero value
because bits 6:3 are not currently supported. When 6:3 is populated, an
unsupported INCREMENT OF bits_per_pixel value is read (DPCD 0006F[2:0])

[How]
Mask the INCREMENT OF bits_per_pixel field so that values in the
unsupported field are ignored.

Signed-off-by: Jaehyun Chung <jaehyun.chung@amd.com>
Reviewed-by: Wenjing Liu <Wenjing.Liu@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: fix pow() crashing when given base 0

[Why&How]
pow(a,x) is implemented as exp(x*log(a)). log(0) will crash.
So return 0^x = 0, unless x=0, convention seems to be 0^0 = 1.

Cc: stable@vger.kernel.org
Signed-off-by: Krunoslav Kovac <Krunoslav.Kovac@amd.com>
Reviewed-by: Anthony Koo <Anthony.Koo@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Reset scrambling on Test Pattern

[Why]
Programming is missing the sequence where for eDP the scrambling is
reset when testing for eye diagram test pattern.

[How]
Include the required register in the definition

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

drm/amd/display: Add extra T3 delay

[Why]
For some special panel, it will have an invalid HPD high cause driver
DPCD read/write fail.

[How]
Add extra T3 delay as a monitor patch in dce110_edp_wait_for_hpd_ready

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

drm/amd/display: Fix DSC force enable on SST

[why]
Previously when force enabling DSC on SST display we unknowingly
supressed lane count, which caused DSC to be enabled automatically.

[how]
By adding an additional flag to force enable DSC in dc_dsc.c DSC can
always be enabled with debugfs dsc_clock_en forced to 1

Cc: stable@vger.kernel.org
Signed-off-by: Eryk Brol <eryk.brol@amd.com>
Signed-off-by: Mikita Lipski <mikita.lipski@amd.com>
Reviewed-by: Wenjing Liu <Wenjing.Liu@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: fix dcn3 wide timing dsc validation

Wide timing DSC requires odm. Since spreadsheet is missing this dsc
validation we have to modify DML vba code ourselves.

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

drm/amd/display: Screen corruption on dual displays (DP+USB-C)

[why]
Current pipe merge and split logic only supports cases where new
dc_state is allocated and relies on dc->current_state to gather
information from previous dc_state.

Calls to validate_bandwidth on UPDATE_TYPE_MED would cause an issue
because there is no new dc_state allocated, and data in
dc->current_state would be overwritten during pipe merge.

[how]
Only allow validate_bandwidth when new dc_state space is created.

Signed-off-by: Qingqing Zhuo <qingqing.zhuo@amd.com>
Reviewed-by: Nicholas Kazlauskas <Nicholas.Kazlauskas@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Add read backlight interface

[Why & How]
Add read backlight interface to get ABM data.

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

drm/amd/display: PSR eDP p-state warning occurs intermittently after unplug DP

[Why]
with eDP + DP, each display use one pipe. after DP unplugged, eDP switch
from one pipe to two pipes -- pipe split.  dpp1_cm_set_regamma_pwl will
be executed too. The duration from switch single pipe to dual pipes is a
little long which could let eDP enter PSR mode. upon two pipes for eDP
are setup, eDP PHY is disabled. front pipe is not really running to
fetch data from frame buffer. i.e., dchubp is not in normal working
status. execution of hubbub1_wm_change_req_wa may cause p-state warning.

[How]
disable eDP PSR before dc_commit_state. psr is disabled when execute
hubbub1_wm_change_req_wa.

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

drm/amd/display: Fix DFPstate hang due to view port changed

[Why]
Place the cursor in the center of screen between two pipes then
adjusting the viewport but cursour doesn't update cause DFPstate hang.

[How]
If viewport changed, update cursor as well.

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

drm/amd/display: Change ABM config init interface

[Why & How]
change abm config init interface to support multiple ABMs.

Signed-off-by: Yongqiang Sun <yongqiang.sun@amd.com>
Reviewed-by: Chris Park <Chris.Park@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: create a function to decide mst link settings

[why]
create a dedicated function to make mst link settings decision, so that
the policy's decision is made in a unified place.

Signed-off-by: Wenjing Liu <wenjing.liu@amd.com>
Reviewed-by: Tony Cheng <Tony.Cheng@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Assign correct left shift

[Why]
Reading for DP alt registers return incorrect values due to LE_SF
definition missing.

[How]
Define correct LE_SF or DP alt registers.

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

drm/amd/display: Call DMUB for eDP power control

[Why]
If DMUB is used, LVTMA VBIOS call can be used to control eDP instead of
tranditional transmitter control.  Interface is agreed with VBIOS for
eDP to use this new path to program LVTMA registers.

[How]
Create DAL interface to send DMUB command for LVTMA as currently
implemented in VBIOS.

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

drm/amd/display: 3.2.98

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

drm/amd/display: [FW Promotion] Release 0.0.28

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

drm/amdgpu: add condition check for trace_amdgpu_cs()

v1:
add trace event enabled check to avoid nop loop when submit multi ibs
in amdgpu_cs_ioctl() function.

v2:
add a new wrapper function to trace all amdgpu cs ibs.

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

drm/amdgpu: fix amdgpu_bo_release_notify() comment error

fix amdgpu_bo_release_notify() comment error.

Signed-off-by: Kevin Wang <kevin1.wang@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Dennis Li <Dennis.Li@amd.com>
Acked-by: Nirmoy Das <nirmoy.das@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdkfd: fix the wrong sdma instance query for renoir

Renoir only has one sdma instance, it will get failed once query the
sdma1 registers. So use switch-case instead of static register array.

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

drm/amdgpu: drop log message in amdgpu_dpm_baco_reset()

The caller does this now for all reset types.  This is now
a duplicate call.

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

drm/amdgpu: note what type of reset we are using

When we reset the GPU, note what type of reset will be
used.  This makes debugging different reset scenarios
more clear as the driver may use different reset
methods depending on conditions on the system.

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

drm/amdgpu: print where we get the vbios image from

ACPI, ROM, PCI BAR, etc.

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

drm/amdgpu: parse ta firmware for navy_flounder

Use the same case as sienna_cichlid

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

drm/amdgpu/vcn3.0: only SIENNA_CICHLID need specify instance for dec/enc

Only SIENNA_CICHLID(VCN3) has 2 unsymmetrical instances, there're less
codecs on instance 1, we use 0 for decode and 1 for encode.

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

drm/amd/pm: optimize the power related source code layout

The target is to provide a clear entry point(for power routines).
Also this can help to maintain a clear view about the frameworks
used on different ASICs. Hopefully all these can make power part
more friendly to play with.

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

drm/amd/powerplay: put those exposed power interfaces in amdgpu_dpm.c

As other power interfaces.

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

drm/amd/powerplay: optimize i2c bus access implementation

The caller needs not care about the internal details how the powerplay
API implemented.

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

drm/amd/powerplay: drop unnecessary pp_funcs checker

It's redundant. Also, the callers should not care about
the implementation details.

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

drm/amd/powerplay: optimize amdgpu_dpm_set_clockgating_by_smu() implementation

Cover the implementation details from outside(of power part).

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

drm/amd/pm: drop redundant MEM_TYPE_* macros

As these are already defined in amdgpu_atombios.h. Otherwise, we may
hit "redefined" compile warning.

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

drm/amd/powerplay: suppress the kernel test robot warning

Suppress the warning below:
   In file included from drivers/gpu/drm/amd/amdgpu/../powerplay/smu_cmn.c:
>> drivers/gpu/drm/amd/powerplay/smu_cmn.c:485:9: warning: Identical condition 'ret', second condition is always false [identicalConditionAfterEarlyExit]
    return ret;
           ^
   drivers/gpu/drm/amd/powerplay/smu_cmn.c:477:6: note: first condition
    if (ret)
        ^
   drivers/gpu/drm/amd/powerplay/smu_cmn.c:485:9: note: second condition
    return ret;
           ^

Signed-off-by: Evan Quan <evan.quan@amd.com>
Reported-by: kernel test robot <lkp@intel.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: guard ras debugfs creation/removal based on CONFIG_DEBUG_FS

It can avoid potential build warn/error when
CONFIG_DEBUG_FS is not set.

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

drm/amdgpu: fix NULL pointer access issue when unloading driver

When unloading driver by "modprobe -r amdgpu", one NULL pointer
dereference bug occurs in ras debugfs releasing. The cause is the
duplicated debugfs_remove, as drm debugfs_root dir has been cleaned
up already by drm_minor_unregister.

BUG: kernel NULL pointer dereference, address: 00000000000000a0
PGD 0 P4D 0
Oops: 0002 [#1] SMP PTI
CPU: 11 PID: 1526 Comm: modprobe Tainted: G           OE     5.6.0-guchchen #1
Hardware name: System manufacturer System Product Name/TUF Z370-PLUS GAMING II, BIOS 0411 09/21/2018
RIP: 0010:down_write+0x15/0x40
Code: eb de e8 7e 17 72 ff cc cc cc cc cc cc cc cc cc cc cc cc cc cc 0f 1f 44 00 00 53 48 89 fb e8 92
d8 ff ff 31 c0 ba 01 00 00 00 <f0> 48 0f b1 13 75 0f 65 48 8b 04 25 c0 8b 01 00 48 89 43 08 5b c3
RSP: 0018:ffffb1590386fcd0 EFLAGS: 00010246
RAX: 0000000000000000 RBX: 00000000000000a0 RCX: 0000000000000000
RDX: 0000000000000001 RSI: ffffffff85b2fcc2 RDI: 00000000000000a0
RBP: ffffb1590386fd30 R08: ffffffff85b2fcc2 R09: 000000000002b3c0
R10: ffff97a330618c40 R11: 00000000000005f6 R12: ffff97a3481beb40
R13: 00000000000000a0 R14: ffff97a3481beb40 R15: 0000000000000000
FS:  00007fb11a717540(0000) GS:ffff97a376cc0000(0000) knlGS:0000000000000000
CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
CR2: 00000000000000a0 CR3: 00000004066d6006 CR4: 00000000003606e0
DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
Call Trace:
 simple_recursive_removal+0x63/0x370
 ? debugfs_remove+0x60/0x60
 debugfs_remove+0x40/0x60
 amdgpu_ras_fini+0x82/0x230 [amdgpu]
 ? __kernfs_remove.part.17+0x101/0x1f0
 ? kernfs_name_hash+0x12/0x80
 amdgpu_device_fini+0x1c0/0x580 [amdgpu]
 amdgpu_driver_unload_kms+0x3e/0x70 [amdgpu]
 amdgpu_pci_remove+0x36/0x60 [amdgpu]
 pci_device_remove+0x3b/0xb0
 device_release_driver_internal+0xe5/0x1c0
 driver_detach+0x46/0x90
 bus_remove_driver+0x58/0xd0
 pci_unregister_driver+0x29/0x90
 amdgpu_exit+0x11/0x25 [amdgpu]
 __x64_sys_delete_module+0x13d/0x210
 do_syscall_64+0x5f/0x250
 entry_SYSCALL_64_after_hwframe+0x44/0xa9

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

drm/amdgpu: revert "fix system hang issue during GPU reset"

The whole approach wasn't thought through till the end.

We already had a reset lock like this in the past and it caused the same problems like this one.

Completely revert the patch for now and add individual trylock protection to the hardware access functions as necessary.

This reverts commit df9c8d1aa278c435c30a69b8f2418b4a52fcb929.

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

drm/amd/powerplay: enable Sienna Cichlid mgpu fan boost feature

Support Sienna Cichlid mgpu fan boost enablement.

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

drm/amd/powerplay: enable Navi1X mgpu fan boost feature(V2)

Support Navi1X mgpu fan boost enablement.

V2: rich the comment and correct the revision id check

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

drm/amd/powerplay: enable swSMU mgpu fan boost support

Enable mgpu fan boost feature on swSMU routines.

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

drm/amd/powerplay: optimize the interface for mgpu fan boost enablement

Cover the implementation details from outside(of power). Also preparing
for expanding this to swSMU.

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

drm/amdgpu: fix uninit-value in arcturus_log_thermal_throttling_event()

when function arcturus_get_smu_metrics_data() call failed,
it will cause the variable "throttler_status" isn't initialized before use.

warning:
powerplay/arcturus_ppt.c:2268:24: warning: ‘throttler_status’ may be used uninitialized in this function [-Wmaybe-uninitialized]
 2268 |   if (throttler_status & logging_label[throttler_idx].feature_mask) {

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

drm/amdgpu: disable gfxoff for navy_flounder

gfxoff is temporarily disabled for navy_flounder,
since at present the feature has broken some basic
amdgpu test.

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

drm/amd/powerplay: bump NAVI12 driver if version

To fit the latest SMU firmware.

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/powerplay: maximum the code sharing around metrics table retrieving

Instead of having one copy in each ASIC.

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

drm/amd/powerplay: update the metrics table cache interval as 1ms

To make the setting same as Arcturus/Navi1x/Sienna_Cichlid.

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: Use function pointer for some mmhub functions

Add more function pointers to amdgpu_mmhub_funcs. ASIC specific
implementation of most mmhub functions are called from a general
function pointer, instead of calling different function for
different ASIC. Simplify the code by deleting duplicate functions

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

drm/amdgpu: pass NULL pointer instead of 0

Fixes: c030f2e4166c3f55 ("drm/amdgpu: add amdgpu_ras.c to support ras (v2)")
Signed-off-by: Nirmoy Das <nirmoy.das@amd.com>
Reviewed-by: Guchun Chen <guchun.chen@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: annotate a false positive recursive locking

[  584.110304] ============================================
[  584.110590] WARNING: possible recursive locking detected
[  584.110876] 5.6.0-deli-v5.6-2848-g3f3109b0e75f #1 Tainted: G           OE
[  584.111164] --------------------------------------------
[  584.111456] kworker/38:1/553 is trying to acquire lock:
[  584.111721] ffff9b15ff0a47a0 (&adev->reset_sem){++++}, at: amdgpu_device_gpu_recover+0x262/0x1030 [amdgpu]
[  584.112112]
               but task is already holding lock:
[  584.112673] ffff9b1603d247a0 (&adev->reset_sem){++++}, at: amdgpu_device_gpu_recover+0x262/0x1030 [amdgpu]
[  584.113068]
               other info that might help us debug this:
[  584.113689]  Possible unsafe locking scenario:

[  584.114350]        CPU0
[  584.114685]        ----
[  584.115014]   lock(&adev->reset_sem);
[  584.115349]   lock(&adev->reset_sem);
[  584.115678]
                *** DEADLOCK ***

[  584.116624]  May be due to missing lock nesting notation

[  584.117284] 4 locks held by kworker/38:1/553:
[  584.117616]  #0: ffff9ad635c1d348 ((wq_completion)events){+.+.}, at: process_one_work+0x21f/0x630
[  584.117967]  #1: ffffac708e1c3e58 ((work_completion)(&con->recovery_work)){+.+.}, at: process_one_work+0x21f/0x630
[  584.118358]  #2: ffffffffc1c2a5d0 (&tmp->hive_lock){+.+.}, at: amdgpu_device_gpu_recover+0xae/0x1030 [amdgpu]
[  584.118786]  #3: ffff9b1603d247a0 (&adev->reset_sem){++++}, at: amdgpu_device_gpu_recover+0x262/0x1030 [amdgpu]
[  584.119222]
               stack backtrace:
[  584.119990] CPU: 38 PID: 553 Comm: kworker/38:1 Kdump: loaded Tainted: G           OE     5.6.0-deli-v5.6-2848-g3f3109b0e75f #1
[  584.120782] Hardware name: Supermicro SYS-7049GP-TRT/X11DPG-QT, BIOS 3.1 05/23/2019
[  584.121223] Workqueue: events amdgpu_ras_do_recovery [amdgpu]
[  584.121638] Call Trace:
[  584.122050]  dump_stack+0x98/0xd5
[  584.122499]  __lock_acquire+0x1139/0x16e0
[  584.122931]  ? trace_hardirqs_on+0x3b/0xf0
[  584.123358]  ? cancel_delayed_work+0xa6/0xc0
[  584.123771]  lock_acquire+0xb8/0x1c0
[  584.124197]  ? amdgpu_device_gpu_recover+0x262/0x1030 [amdgpu]
[  584.124599]  down_write+0x49/0x120
[  584.125032]  ? amdgpu_device_gpu_recover+0x262/0x1030 [amdgpu]
[  584.125472]  amdgpu_device_gpu_recover+0x262/0x1030 [amdgpu]
[  584.125910]  ? amdgpu_ras_error_query+0x1b8/0x2a0 [amdgpu]
[  584.126367]  amdgpu_ras_do_recovery+0x159/0x190 [amdgpu]
[  584.126789]  process_one_work+0x29e/0x630
[  584.127208]  worker_thread+0x3c/0x3f0
[  584.127621]  ? __kthread_parkme+0x61/0x90
[  584.128014]  kthread+0x12f/0x150
[  584.128402]  ? process_one_work+0x630/0x630
[  584.128790]  ? kthread_park+0x90/0x90
[  584.129174]  ret_from_fork+0x3a/0x50

Each adev has owned lock_class_key to avoid false positive
recursive locking.

v2:
1. register adev->lock_key into lockdep, otherwise lockdep will
report the below warning

[ 1216.705820] BUG: key ffff890183b647d0 has not been registered!
[ 1216.705924] ------------[ cut here ]------------
[ 1216.705972] DEBUG_LOCKS_WARN_ON(1)
[ 1216.705997] WARNING: CPU: 20 PID: 541 at kernel/locking/lockdep.c:3743 lockdep_init_map+0x150/0x210

v3:
change to use down_write_nest_lock to annotate the false dead-lock
warning.

Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Signed-off-by: Dennis Li <Dennis.Li@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: add debugfs interface for RAP test

After amdgpu driver loading successfully, we can use
RAP debugfs interface <debugfs_dir>/dri/xxx/rap_test
to trigger RAP test.

Currently only L0 validate test is supported.

v2: refine amdgpu_rap.h

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

drm/amdgpu: enable RAP TA load

Enable the RAP TA loading path and add RAP test
trigger interface.

v2: fix potential mem leak issue

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

drm/amdgpu: add RAP TA header file

The RAP TA contains tests used to verify if
RAP(Register Access Policy), or otherwise known
as Security Policy is applied correctly
by PSP BL&TOS.

The RAP test is a measure to ensure that we reduce
the avenue of complexity and mistakes when dealing
with RAP in post-si execution, where debugging failures
related to RAP is quite difficult and expensive.

v2: add introduction for RAP TA

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

drm/amdgpu: reconfigure spm golden settings on Navi1x after GFXOFF exit(v3)

On Navi1x, the SPM golden settings are lost after GFXOFF
enter/exit, so reconfigure the golden settings after GFXOFF
exit.

Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Tianci.Yin <tianci.yin@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: add interface amdgpu_gfx_init_spm_golden for Navi1x

On Navi1x, the SPM golden settings are lost after GFXOFF
enter/exit, so reconfiguration is needed. Make the
configuration code as an interface for future use.

Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Reviewed-by: Luben Tuikov <luben.tuikov@amd.com>
Reviewed-by: Feifei Xu <Feifei.Xu@amd.com>
Signed-off-by: Tianci.Yin <tianci.yin@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: add debugfs node to toggle ras error cnt harvest

Before ras recovery is issued, user could operate this debugfs
node to enable/disable the harvest of all RAS IPs' ras error
count registers, which will help keep hardware's registers'
status instead of cleaning up them.

Signed-off-by: Guchun Chen <guchun.chen@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Reviewed-by: Dennis Li <Dennis.Li@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: bypass querying ras error count registers

Once ras recovery is issued by ras sync flood interrupt or
ras controller interrupt, add this guard to bypass or execute
ras error count register harvest of all IPs.

Signed-off-by: Guchun Chen <guchun.chen@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Reviewed-by: Dennis Li <Dennis.Li@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: Enable P2P dmabuf over XGMI

Access the exported P2P dmabuf over XGMI, if available.
Otherwise, fall back to the existing PCIe method.

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

drm/amd/display: convert to use le16_add_cpu()

Convert cpu_to_le16(le16_to_cpu(E1) + E2) to use le16_add_cpu().

Signed-off-by: Qinglang Miao <miaoqinglang@huawei.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu/display: drop unused function

This is not longer used as of the latest rework of this
code so drop it to avoid a unused function warning.

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

drm/amd/display: add DCN support for aarch64

This adds ARM64 support into the DCN. This mainly enables support
for Navi graphics cards. The dcn10 changes haven't been tested,
since I don't have the relevant hardware available, but there
is no way to conditionally disable them, so I've done them anyway.

Signed-off-by: Daniel Kolesa <daniel@octaforge.org>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu/display: use GFP_ATOMIC in dcn20_validate_bandwidth_internal

GFP_KERNEL may and will sleep, and this is being executed in
a non-preemptible context; this will mess things up since it's
called inbetween DC_FP_START/END, and rescheduling will result
in the DC_FP_END later being called in a different context (or
just crashing if any floating point/vector registers/instructions
are used after the call is resumed in a different context).

Signed-off-by: Daniel Kolesa <daniel@octaforge.org>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Blank stream before destroying HDCP session

[Why]
Stream disable sequence incorretly destroys HDCP session while stream is
not blanked and while audio is not muted. This sequence causes a flash
of corruption during mode change and an audio click.

[How]
Change sequence to blank stream before destroying HDCP session. Audio will
also be muted by blanking the stream.

Cc: stable@vger.kernel.org
Signed-off-by: Jaehyun Chung <jaehyun.chung@amd.com>
Reviewed-by: Alvin Lee <Alvin.Lee2@amd.com>
Acked-by: Qingqing Zhuo <qingqing.zhuo@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Fix EDID parsing after resume from suspend

[Why]
Resuming from suspend, CEA blocks from EDID are not parsed and no video
modes can support YUV420. When this happens, output bpc cannot go over
8-bit with 4K modes on HDMI.

[How]
In amdgpu_dm_update_connector_after_detect(), drm_add_edid_modes() is
called after drm_connector_update_edid_property() to fully parse EDID
and update display info.

Cc: stable@vger.kernel.org
Signed-off-by: Stylon Wang <stylon.wang@amd.com>
Reviewed-by: Nicholas Kazlauskas <Nicholas.Kazlauskas@amd.com>
Acked-by: Qingqing Zhuo <qingqing.zhuo@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Disconnect pipe separetely when disable pipe split

[Why]
When changing pixel formats for HDR (e.g. ARGB -> FP16)
there are configurations that change from 2 pipes to 1 pipe.
In these cases, it seems that disconnecting MPCC and doing
a surface update at the same time(after unlocking) causes
some registers to be updated slightly faster than others
after unlocking (e.g. if the pixel format is updated to FP16
before the new surface address is programmed, we get
corruption on the screen because the pixel formats aren't
matching). We separate disconnecting MPCC from the rest
of  the  pipe programming sequence to prevent this.

[How]
Move MPCC disconnect into separate operation than the
rest of the pipe programming.

Signed-off-by: Alvin Lee <alvin.lee2@amd.com>
Reviewed-by: Aric Cyr <Aric.Cyr@amd.com>
Acked-by: Qingqing Zhuo <qingqing.zhuo@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>