drm/amdgpu: Let KFD use more VMIDs on Arcturus

When there is no graphics support, KFD can use more of the VMIDs. Graphics
VMIDs are only used for video decoding/encoding and post processing. With
two VCE engines, there is no reason to reserve more than 2 VMIDs for that.

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

drm/amdgpu: Clean up KFD VMID assignment

The KFD VMID assignment was hard-coded in a few places. Consolidate that in
a single variable adev->vm_manager.first_kfd_vmid. The value is still
assigned in gmc-ip-version-specific code.

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

drm/amdgpu: SI support for UVD and VCE power managment

Port functionality from the Radeon driver to support
UVD and VCE power management.

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

drm/amdgpu: SI support for VCE clock control

Port functionality from the Radeon driver to support
VCE clock control.

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

drm/amdgpu: SI support for UVD clock control

Port functionality from the Radeon driver to support
UVD clock control.

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

Merge tag 'amd-drm-next-5.9-2020-07-01' of git://people.freedesktop.org/~agd5f/linux into drm-next

amd-drm-next-5.9-2020-07-01:

amdgpu:
- DC DMUB updates
- HDCP fixes
- Thermal interrupt fixes
- Add initial support for Sienna Cichlid GPU
- Add support for unique id on Arcturus
- Major swSMU code cleanup
- Skip BAR resizing if the bios already did id
- Fixes for DCN bandwidth calculations
- Runtime PM reference count fixes
- Add initial UVD support for SI
- Add support for ASSR on eDP links
- Lots of misc fixes and cleanups
- Enable runtime PM on vega10 boards that support BACO
- RAS fixes
- SR-IOV fixes
- Use IP discovery table on renoir
- DC stream synchronization fixes

amdkfd:
- Track SDMA usage per process
- Fix GCC10 compiler warnings
- Locking fix

radeon:
- Default to on chip GART for AGP boards on all arches
- Runtime PM reference count fixes

UAPI:
- Update comments to clarify MTYPE

From: Alex Deucher <alexdeucher@gmail.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20200701155041.1102829-1-alexander.deucher@amd.com
Signed-off-by: Dave Airlie <airlied@redhat.com>

drm/amd/powerplay: Fix NULL dereference in lock_bus() on Vega20 w/o RAS

I updated my system with Radeon VII from kernel 5.6 to kernel 5.7, and
following started to happen on each boot:

...
BUG: kernel NULL pointer dereference, address: 0000000000000128
...
CPU: 9 PID: 1940 Comm: modprobe Tainted: G            E     5.7.2-200.im0.fc32.x86_64 #1
Hardware name: System manufacturer System Product Name/PRIME X570-P, BIOS 1407 04/02/2020
RIP: 0010:lock_bus+0x42/0x60 [amdgpu]
...
Call Trace:
 i2c_smbus_xfer+0x3d/0xf0
 i2c_default_probe+0xf3/0x130
 i2c_detect.isra.0+0xfe/0x2b0
 ? kfree+0xa3/0x200
 ? kobject_uevent_env+0x11f/0x6a0
 ? i2c_detect.isra.0+0x2b0/0x2b0
 __process_new_driver+0x1b/0x20
 bus_for_each_dev+0x64/0x90
 ? 0xffffffffc0f34000
 i2c_register_driver+0x73/0xc0
 do_one_initcall+0x46/0x200
 ? _cond_resched+0x16/0x40
 ? kmem_cache_alloc_trace+0x167/0x220
 ? do_init_module+0x23/0x260
 do_init_module+0x5c/0x260
 __do_sys_init_module+0x14f/0x170
 do_syscall_64+0x5b/0xf0
 entry_SYSCALL_64_after_hwframe+0x44/0xa9
...

Error appears when some i2c device driver tries to probe for devices
using adapter registered by `smu_v11_0_i2c_eeprom_control_init()`.
Code supporting this adapter requires `adev->psp.ras.ras` to be not
NULL, which is true only when `amdgpu_ras_init()` detects HW support by
calling `amdgpu_ras_check_supported()`.

Before 9015d60c9ee1, adapter was registered by

-> amdgpu_device_ip_init()
  -> amdgpu_ras_recovery_init()
    -> amdgpu_ras_eeprom_init()
      -> smu_v11_0_i2c_eeprom_control_init()

after verifying that `adev->psp.ras.ras` is not NULL in
`amdgpu_ras_recovery_init()`. Currently it is registered
unconditionally by

-> amdgpu_device_ip_init()
  -> pp_sw_init()
    -> hwmgr_sw_init()
      -> vega20_smu_init()
        -> smu_v11_0_i2c_eeprom_control_init()

Fix simply adds HW support check (ras == NULL => no support) before
calling `smu_v11_0_i2c_eeprom_control_{init,fini}()`.

Please note that there is a chance that similar fix is also required for
CHIP_ARCTURUS. I do not know whether any actual Arcturus hardware without
RAS exist, and whether calling `smu_i2c_eeprom_init()` makes any sense
when there is no HW support.

Cc: stable@vger.kernel.org
Fixes: 9015d60c9ee1 ("drm/amdgpu: Move EEPROM I2C adapter to amdgpu_device")
Signed-off-by: Ivan Mironov <mironov.ivan@gmail.com>
Tested-by: Bjorn Nostvold <bjorn.nostvold@gmail.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: enable runtime pm on vega10 when noretry=0

The failures with ROCm only happen with noretry=1, so
enable runtime pm when noretry=0 (the current default).

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

drm/amdgpu: rework runtime pm enablement for BACO

Add a switch statement to simplify asic checks.  Note
that BACO is not supported on APUs, so there is no
need to check them.

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

drm/amdgpu: call release_firmware() without a NULL check

The release_firmware() function is NULL tolerant so we do not need
to check for NULL param before calling it.

Signed-off-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/amdkfd: Fix circular locking dependency warning

[  150.887733] ======================================================
[  150.893903] WARNING: possible circular locking dependency detected
[  150.905917] ------------------------------------------------------
[  150.912129] kfdtest/4081 is trying to acquire lock:
[  150.917002] ffff8f7f3762e118 (&mm->mmap_sem#2){++++}, at:
                                 __might_fault+0x3e/0x90
[  150.924490]
               but task is already holding lock:
[  150.930320] ffff8f7f49d229e8 (&dqm->lock_hidden){+.+.}, at:
                                destroy_queue_cpsch+0x29/0x210 [amdgpu]
[  150.939432]
               which lock already depends on the new lock.

[  150.947603]
               the existing dependency chain (in reverse order) is:
[  150.955074]
               -> #3 (&dqm->lock_hidden){+.+.}:
[  150.960822]        __mutex_lock+0xa1/0x9f0
[  150.964996]        evict_process_queues_cpsch+0x22/0x120 [amdgpu]
[  150.971155]        kfd_process_evict_queues+0x3b/0xc0 [amdgpu]
[  150.977054]        kgd2kfd_quiesce_mm+0x25/0x60 [amdgpu]
[  150.982442]        amdgpu_amdkfd_evict_userptr+0x35/0x70 [amdgpu]
[  150.988615]        amdgpu_mn_invalidate_hsa+0x41/0x60 [amdgpu]
[  150.994448]        __mmu_notifier_invalidate_range_start+0xa4/0x240
[  151.000714]        copy_page_range+0xd70/0xd80
[  151.005159]        dup_mm+0x3ca/0x550
[  151.008816]        copy_process+0x1bdc/0x1c70
[  151.013183]        _do_fork+0x76/0x6c0
[  151.016929]        __x64_sys_clone+0x8c/0xb0
[  151.021201]        do_syscall_64+0x4a/0x1d0
[  151.025404]        entry_SYSCALL_64_after_hwframe+0x49/0xbe
[  151.030977]
               -> #2 (&adev->notifier_lock){+.+.}:
[  151.036993]        __mutex_lock+0xa1/0x9f0
[  151.041168]        amdgpu_mn_invalidate_hsa+0x30/0x60 [amdgpu]
[  151.047019]        __mmu_notifier_invalidate_range_start+0xa4/0x240
[  151.053277]        copy_page_range+0xd70/0xd80
[  151.057722]        dup_mm+0x3ca/0x550
[  151.061388]        copy_process+0x1bdc/0x1c70
[  151.065748]        _do_fork+0x76/0x6c0
[  151.069499]        __x64_sys_clone+0x8c/0xb0
[  151.073765]        do_syscall_64+0x4a/0x1d0
[  151.077952]        entry_SYSCALL_64_after_hwframe+0x49/0xbe
[  151.083523]
               -> #1 (mmu_notifier_invalidate_range_start){+.+.}:
[  151.090833]        change_protection+0x802/0xab0
[  151.095448]        mprotect_fixup+0x187/0x2d0
[  151.099801]        setup_arg_pages+0x124/0x250
[  151.104251]        load_elf_binary+0x3a4/0x1464
[  151.108781]        search_binary_handler+0x6c/0x210
[  151.113656]        __do_execve_file.isra.40+0x7f7/0xa50
[  151.118875]        do_execve+0x21/0x30
[  151.122632]        call_usermodehelper_exec_async+0x17e/0x190
[  151.128393]        ret_from_fork+0x24/0x30
[  151.132489]
               -> #0 (&mm->mmap_sem#2){++++}:
[  151.138064]        __lock_acquire+0x11a1/0x1490
[  151.142597]        lock_acquire+0x90/0x180
[  151.146694]        __might_fault+0x68/0x90
[  151.150879]        read_sdma_queue_counter+0x5f/0xb0 [amdgpu]
[  151.156693]        update_sdma_queue_past_activity_stats+0x3b/0x90 [amdgpu]
[  151.163725]        destroy_queue_cpsch+0x1ae/0x210 [amdgpu]
[  151.169373]        pqm_destroy_queue+0xf0/0x250 [amdgpu]
[  151.174762]        kfd_ioctl_destroy_queue+0x32/0x70 [amdgpu]
[  151.180577]        kfd_ioctl+0x223/0x400 [amdgpu]
[  151.185284]        ksys_ioctl+0x8f/0xb0
[  151.189118]        __x64_sys_ioctl+0x16/0x20
[  151.193389]        do_syscall_64+0x4a/0x1d0
[  151.197569]        entry_SYSCALL_64_after_hwframe+0x49/0xbe
[  151.203141]
               other info that might help us debug this:

[  151.211140] Chain exists of:
                 &mm->mmap_sem#2 --> &adev->notifier_lock --> &dqm->lock_hidden

[  151.222535]  Possible unsafe locking scenario:

[  151.228447]        CPU0                    CPU1
[  151.232971]        ----                    ----
[  151.237502]   lock(&dqm->lock_hidden);
[  151.241254]                                lock(&adev->notifier_lock);
[  151.247774]                                lock(&dqm->lock_hidden);
[  151.254038]   lock(&mm->mmap_sem#2);

This commit fixes the warning by ensuring get_user() is not called
while reading SDMA stats with dqm_lock held as get_user() could cause a
page fault which leads to the circular locking scenario.

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

drm/radeon: fix array out-of-bounds read and write issues

There is an off-by-one bounds check on the index into arrays
table->mc_reg_address and table->mc_reg_table_entry[k].mc_data[j] that
can lead to reads and writes outside of arrays. Fix the bound checking
off-by-one error.

Addresses-Coverity: ("Out-of-bounds read/write")
Fixes: cc8dbbb4f62a ("drm/radeon: add dpm support for CI dGPUs (v2)")
Signed-off-by: Colin Ian King <colin.king@canonical.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: ensure 0 is returned for success in jpeg_v2_5_wait_for_idle

In the cases where adev->jpeg.num_jpeg_inst is zero or the condition
adev->jpeg.harvest_config & (1 << i) is always non-zero the variable
ret is never set to an error condition and the function returns
an uninitialized value in ret.  Since the only exit condition at
the end if the function is a success then explicitly return
0 rather than a potentially uninitialized value in ret.

Addresses-Coverity: ("Uninitialized scalar variable")
Fixes: 14f43e8f88c5 ("drm/amdgpu: move JPEG2.5 out from VCN2.5")
Signed-off-by: Colin Ian King <colin.king@canonical.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: make sure to reserve tmr region on all asics which support it

This includes older APUs like renoir.

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

drm/amdgpu/display: Unlock mutex on error

Make sure we pass through ret label to unlock the mutex.

Signed-off-by: John van der Kamp <sjonny@suffe.me.uk>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd: fix potential memleak in err branch

The function kobject_init_and_add alloc memory like:
kobject_init_and_add->kobject_add_varg->kobject_set_name_vargs
->kvasprintf_const->kstrdup_const->kstrdup->kmalloc_track_caller
->kmalloc_slab, in err branch this memory not free. If use
kmemleak, this path maybe catched.
These changes are to add kobject_put in kobject_init_and_add
failed branch, fix potential memleak.

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

drm/radeon: fix fb_div check in ni_init_smc_spll_table()

clk_s is checked twice in a row in ni_init_smc_spll_table().
fb_div should be checked instead.

Fixes: 69e0b57a91ad ("drm/radeon/kms: add dpm support for cayman (v5)")
Cc: stable@vger.kernel.org
Signed-off-by: Denis Efremov <efremov@linux.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/powerplay: return current DCEFCLK on sysfs read (v2)

This fixes incorrect output when reading pp_dpm_dcefclk sysfs.

Output before this patch:
cat /sys/devices/pci0000:<snip>/pp_dpm_dcefclk
0: 506Mhz
1: 42949671Mhz *
2: 1266Mhz

v2: squash in arcturus fix

Signed-off-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/powerplay: fix compilation warning

If (smu)->ppt_funcs->intf is not initialized then
smu_send_smc_msg(smu, SMU_MSG_ReadSerialNumBottom32, &bottom32) and
smu_send_smc_msg(smu, SMU_MSG_ReadSerialNumTop32, &top32) will not
touch bottom32 and top32.

Signed-off-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/display: 3.2.91

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

drm/amd/display: Fix ineffective setting of max bpc property

[Why]
Regression was introduced where setting max bpc property has no effect
on the atomic check and final commit. It has the same effect as max bpc
being stuck at 8.

[How]
Correctly propagate max bpc with the new connector state.

Signed-off-by: Stylon Wang <stylon.wang@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: [FW Promotion] Release 1.0.19

[Header Changes]
- Add debug flag for psr to use hw locking mgr state machine

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: DP link layer test 4.2.1.1 fix due to specs update

[why]
DP link layer CTS specs updated to change the test parameters in test
4.2.1.1.
Before it requires source to delay 400us on aux no reply.
With the specs updates Errata5, it requires source to delay 3.2ms
(based on LTTPR aux timeout)
This causes our test to fail after updating with the latest test
equipment firmware.

[how]
the change is to allow LTTPR 3.2ms aux timeout delay by default.
And set back to 400us if LTTPR feature is not enabled.
We will set 3.2ms and always enable LTTPR non transparent mode
if LTTPR feature is enabled and LTTPR is present.

Signed-off-by: Wenjing Liu <wenjing.liu@amd.com>
Reviewed-by: Jun Lei <Jun.Lei@amd.com>
Acked-by: Krunoslav Kovac <Krunoslav.Kovac@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: allow query ddc data over aux to be read only operation

[why]
Two issues:
1. Add read only operation support for query ddc data over aux.
2. Fix a bug where if read size is multiple of 16,
mot of the last read transaction will not be set to 0.

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

drm/amd/display: Compare v_front_porch when checking if streams are synchronizable

[Why]
If the front porch of the two timings differ, then there may not be
enough time while both streams are in vertical blank to perform a memory
clock change. This can hang the system.

[How]
Check the each streams timing.v_front_porch when determining if the two
streams are synchronizable.

Signed-off-by: David Galiffi <David.Galiffi@amd.com>
Reviewed-by: Jun Lei <Jun.Lei@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: enable seamless boot for dcn30

why:
seamless boots requires split of init_hw into hw and pipes to work. This
was implemented in dcn10_init_hw but did not apply yet to dcn30.

how:
Copy over dcn10_init_hw and adapt it to dcn30 using recent changes to
dcn3.  Behavior will be different in init sequence.

Signed-off-by: Martin Leung <martin.leung@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: Red screen observed on startup

[Why]
We try to to change new_clocks->dppclk_khz to 100000 when
new_clocks->dppclk_khz is 0

[How]
Don't change new_clocks->dppclk_khz value when new_clocks->dppclk_khz is
0

Signed-off-by: Peikang Zhang <peikang.zhang@amd.com>
Reviewed-by: Yongqiang Sun <yongqiang.sun@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Allow 4 split on 10K 420 modes

[Why]
10K YCbCr420 does not need ODM 4:1, but it requires MPC 4 split
indicated on the flags.

[How]
Make pixel encoding and resolution size specific workaround to enable
ODM combine on YCbCr420 high resolution modes.

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

drm/amd/display: VSC SDP supported for SST

[why]
If a typeC to HDMI dongle supports YCbCr420 pass through and VSC
colorimetry and pixel encoding formats in the Extended Receiver
Capability, we shall allow VSC SDP to be used.

[How]
The Extended Receiver Capability field shall check the
EXTENDED_RECEIVER_CAPABILITY_FIELD_PRESENT bit in the
TRAINING_AUX_RD_INTERVAL register.  Removed DPCD rev checking for VSC
SDP.

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

drm/amd/display: Fill in dmub_srv fw_version from firmware metadata

[Why]
DMCUB firmware version is now available from firmware metadata block.

We should be passing this into dmub_srv so we can know when to apply
firmware version specific functionality like using CW4 only instead
of the REGION4.

[How]
We don't have the helpers for DM to actually extract out firmware
metadata block themselves.

We could add that and add helpers in DM to grab this, but not every
creation sequence has firmware instruction before dmub_srv_create.

Easiest way to handle this is to fill this in automatically per DM in
the place we do have it - when calculating the region parameters. But
only fill it in if DM already hasn't in case we need to override with
a specific version.

We aren't do anything firmware version specific in dmub_srv_create
today that does require fw_version, so while it's a little unituitive
to do it when calculating region parameters it works for now.

Signed-off-by: Nicholas Kazlauskas <nicholas.kazlauskas@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: Enable output_bpc property on all outputs

[Why]
Connector property output_bpc is available on DP/eDP only. New IGT tests
would benifit if this property works on HDMI.

[How]
Enable this read-only property on all types of connectors.

Signed-off-by: Stylon Wang <stylon.wang@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: Force ODM combine on 5K+ 420 modes

[Why]
All YCbCr420 resolutions 5K and above have tiling and discoloration
issues.  The issue can be remedied by forcing ODM combine from 5K to 8K.
10K resolution requires ODM 4:1.  The mechanism of what the real problem
is, that is inherent in ODM combine programming, doesn't seem to be
pointed at singular register programming (CLK, MPC, DCSURF, etc.), and
needs more in-depth programming sequence review for these new use case
scenarios.  Until then, workaround to enable ODM combine is proposed.
While it is not our policy, HW spreadsheet also recommends turning on
ODM for these scenario to lower the voltage.

[How]
Make pixel encoding and resolution size specific workaround to enable
ODM combine on YCbCr420 high resolution modes.

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

drm/amd/display: fix 4to1 odm MPC_OUT_FLOW_CONTROL_COUNT

Signed-off-by: Dmytro Laktyushkin <Dmytro.Laktyushkin@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: use dispclk AVFS for dppclk

[Why]
There is using pixelclk AVFS for dppclk, that would cause issue.

[How]
To use dispclk AVFS for both dispclk and dppclk.  There would choose
dppclk for request voltage when dispclk wouldn't be updated case.  If
dispclk need to be updated, then it'll choose the bigger one from dppclk
and dispclk for request voltage.

Signed-off-by: Brandon Syu <Brandon.Syu@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: add mechanism to skip DCN init

[Why]
If optimized init is done in FW. DCN init be skipped in driver. This
need to be communicated between driver and fw and maintain backwards
compatibility.

[How]
Use DMUB scratch 0 bit 2 to indicate optimized init done in fw and
use DMUB scatch 4 bit 0 to indicate drive supports the optimized flow
so FW will perform it.

Signed-off-by: Eric Yang <Eric.Yang2@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: fine tune logic of edid max TMDS clock check

[WHY]
Check max_tmds_clk_mhz firstly will restrict pixel clock under HDMI
1.4, thus HDMI2.0 port can't correctly support 4K 60Hz.

[HOW]
Fine tune the logic to check max_forum_tmds_clk_mhz firstly.

Signed-off-by: Dale Zhao <dale.zhao@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: Added local_sink null check before access

[Why & How]
Need to check if local_sink is NULL before accessing.

Signed-off-by: Jake Wang <haonan.wang2@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: clip plane rects in DM before passing into DC

[Why]
DC global validation can fail when userspace requests to draw large
plane without performing the clipping themselves.

This is observed in the IGT kms_plane panning tests for 4K displays
where they draw an 8K plane without any clipping while expecting only
the top 4K to be drawn.

[How]
DRM already has helpers to take care of the clipping necessary and to
mark whether a plane is visible or not, so make use of these helpers
in DM before passing the plane to DC.

Signed-off-by: Aurabindo Pillai <aurabindo.pillai@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: 3.2.90

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

drm/amd/display: [FW Promotion] Release 1.0.18

[Header Changes]
- Update scratch information for boot status

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: Correctly respond in psr enablement interface

[Why]
dc_link_set_psr_allow_active() always returns true, even in the case
that PSR is not supported.

[How]
Hook up the return value of dc_link_set_psr_allow_active().

Signed-off-by: Camille Cho <Camille.Cho@amd.com>
Reviewed-by: Josip Pavic <Josip.Pavic@amd.com>
Acked-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 DML failures caused by doubled stereo viewport

[Why]
Side-by-side and Top-and-bottom stereo configurations fail DML mode
validation due to Viewport exceeded.

This is because we consider the planes as being pipe split in pipe
population so we end up doubling the viewport width, eg. from 4k to 8k.

[How]
These pipes technically aren't hsplit, so add a check for determining
whether is_hsplit should be set.

Signed-off-by: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com>
Reviewed-by: Dmytro Laktyushkin <Dmytro.Laktyushkin@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: enable assr

[Why]
assr is content protection for eDP, in order to use it we need to call
psp ta (dtm)

[How]
We have a enable_assr callback, hook into this and call the correct psp
cmd id to enable assr.

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

drm/amd/display: implement edid max TMDS clock check in DC

[WHY]
Currently DC doesn't check requested pixel clock against an EDID
specified TMDS max clock if it exists, passing modes that should fail

[HOW]
Add max TMDS clk to edid caps and perform check during validation

Signed-off-by: Michael Strauss <michael.strauss@amd.com>
Reviewed-by: Eric Yang <eric.yang2@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 1.0.17

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: Fixed using wrong eDP power sequence function pointer

[why]
dc->hwss->edp_backlight_control is null, it would casue it only be off
main-link of eDP.  It is not worng behavior for eDP power sequence off.

[how]
Must use hwseq->funcs.edp_backlight_control finction pointer for edp
backlight.

Signed-off-by: Yi-Ling Chen <Yi-Ling.Chen2@amd.com>
Reviewed-by: Sung Lee <Sung.Lee@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: Fix calculation of virtual channel payload

[why]
The calculation of virtual channel payload would not take link settings
in account. As we calculate VCPI slots needed both PBN for stream and
also PBN per time slot.  Before we would use generic PBN per time slot,
which would not change with link settings causing wrong Payload
allocation.

[how]
Provide PBN per time slot for each Virtual channel payload calculation.

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

drm/amd/display: [FW Promotion] Release 1.0.16

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: Use dmub fw to lock pipe, cursor, dig

[Why]
Hw lock manager adds the ability to lock pipe, cursor, and dig in fw.

[How]
Send hw lock command to fw to lock pipe, cursor, and dig.

Signed-off-by: Wyatt Wood <wyatt.wood@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 amdgpu: SI UVD enabled on Verde, Tahiti, Pitcairn

Enable asic Verde, Tahiti and Pitcairn UVD block.

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

drm amdgpu: SI UVD enable for Oland

Enable Oland UVD block.

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

drm amdgpu: SI UVD add uvd_v3_1 to makefile

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

drm amdgpu: SI UVD context rreg/wreg

Implement SI uvd_ctx_rreg/wreg functions.

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

drm amdgpu: SI UVD v3_1 (v2)

Implement SI UVD. The SI UVD firmware requires validation.

v2: squash in missing brace fix (Colin Ian King)

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

drm amdgpu: SI UVD registers

Add SI UVD registers files.

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

drm amdgpu: SI UVD PACKET_TYPE0

Fix packet_type0 definition in sid.

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

drm amdgpu: SI UVD add firmwares

For implement UVD on SI asic, add Oland, Pitcairn, Verde, Tahiti firmwares.

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

drm/amdgpu/uvd3.x: fix register definition warnings

drop the duplicate register macros from sid.h and use the
standard ones in the oss register headers.

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

drm/amd/display: fix compilation error on allmodconfig

when compiled with allmodconfig option, there are error
messages as below:

ERROR: modpost:
"mod_color_is_table_init"
[drivers/gpu/drm/amd/amdgpu/amdgpu.ko] undefined!
ERROR: modpost:
"mod_color_get_table"
[drivers/gpu/drm/amd/amdgpu/amdgpu.ko] undefined!
ERROR: modpost:
"mod_color_set_table_init_state"
[drivers/gpu/drm/amd/amdgpu/amdgpu.ko] undefined!

To fix the issue, this commits removes
CONFIG_DRM_AMD_DC_DCN guard in color/makefile.

Signed-off-by: Qingqing Zhuo <qingqing.zhuo@amd.com>
CC: Lewis Huang <Lewis.Huang@amd.com>
CC: Aric Cyr <Aric.Cyr@amd.com>
CC: Alexander Deucher <Alexander.Deucher@amd.com>
CC: Harry Wentland <hwentlan@amd.com>
CC: Nicholas Kazlauskas <Nicholas.Kazlauskas@amd.com>
CC: Bhawanpreet Lakha <Bhawanpreet.Lakha@amd.com>
CC: Stephen Rothwell <sfr@canb.auug.org.au>
Acked-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: fix unused variable

SOC15_WAIT_ON_RREG's return value needn't always been handled by
caller. new design is to fix this kind of unused variable.

Signed-off-by: James Zhu <James.Zhu@amd.com>
Reported-by: kernel test robot <lkp@intel.com>
Acked-by: Leo Liu <leo.liu@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: add fw release for sdma v5_0

sdma fw isn't released when module exit

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

drm/amdgpu: sdma v5_2 ring bo mem leak

invoke amdgpu_ring_fini for each sdma instance

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

drm/amdgpu: remove perf level dpm in one-VF

On Navi12 platform, node power_dpm_force_performance_level
doesn't work correctly in one-VF mode with at least three
smu messages not supported:
SMU_MSG_SetSoftMaxByFreq
SMU_MSG_SetSoftMinByFreq
SMU_MSG_TransferTableDram2Smu

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

drm/amdgpu: add XGMI support for sienna cichlid

support for setting up XGMI FB address regions

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

drm/amdgpu: restrict the hw sched jobs number to power of two

the module parameter sched_hw_submission is probably from user mode,
and the kernel need to check whether it is legal.

1. align hw sched jobs to power of 2 and set minimum number is 2.
2. use kernel api is_power_of_2() to simplify driver code.

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/amd/display: Update bounding box states (v2)

[Why]
We need to update each p-state in the bounding box

[How]
Update states when assigning values to clocks

v2: squash in patch to set min values (Alex)

Signed-off-by: Alvin Lee <alvin.lee2@amd.com>
Signed-off-by: Jerry (Fangzhi) Zuo <Jerry.Zuo@amd.com>
Reviewed-by: Hersen Wu <hersenxs.wu@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: remove redundant initialization of variable ret

The variable ret is being initialized with a value that is never read
and it is being updated later with a new value.  The initialization is
redundant and can be removed.

Addresses-Coverity: ("Unused value")
Signed-off-by: Colin Ian King <colin.king@canonical.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: remove unused functions

Remove unused amdgpu_xgmi_hive_try_lock() and smu7_reset_asic_tasks().

Signed-off-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/display: properly guard the calls to swSMU functions

It's only applicable on newer asics.  We could end up here when
using DC on older asics like SI or KV.

Bug: https://gitlab.freedesktop.org/drm/amd/-/issues/1170
Reviewed-by: Evan Quan <evan.quan@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu/display bail early in dm_pp_get_static_clocks

If there are no supported callbacks.  We'll fall back to the
nominal clocks.

Bug: https://gitlab.freedesktop.org/drm/amd/-/issues/1170
Reviewed-by: Evan Quan <evan.quan@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu/vcn2.0: fix no previous prototype for functions

Include header file vcn_v2_0.h

Signed-off-by: James Zhu <James.Zhu@amd.com>
Reported-by: kernel test robot <lkp@intel.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu/vcn1.0: fix no previous prototype for functions

Include header file vcn_v1_0.h

Signed-off-by: James Zhu <James.Zhu@amd.com>
Reported-by: kernel test robot <lkp@intel.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu/pm: fix ref count leak when pm_runtime_get_sync fails

The call to pm_runtime_get_sync increments the counter even in case of
failure, leading to incorrect ref count.
In case of failure, decrement the ref count before returning.

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

drm/amdgpu/debugfs: fix ref count leak when pm_runtime_get_sync fails

The call to pm_runtime_get_sync increments the counter even in case of
failure, leading to incorrect ref count.
In case of failure, decrement the ref count before returning.

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

drm/amd/display: label internally used symbols as static

Used sparse(make C=1) to find these loose ends.

Signed-off-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/powerplay: label internally used symbols as static

Used sparse(make C=1) to find these loose ends.

Signed-off-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/amdkfd: label internally used symbols as static

Used sparse(make C=1) to find these loose ends.

v2:
removed unwanted extra line

Signed-off-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: label internally used symbols as static

Used sparse(make C=1) to find these loose ends.

v2:
removed unwanted extra line

Signed-off-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/jpeg2.0: fix no previous prototype for functions

Include header file jpeg_v2_0.h.

Signed-off-by: James Zhu <James.Zhu@amd.com>
Reported-by: kernel test robot <lkp@intel.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu/jpeg1.0: fix no previous prototype for functions

Include header file jpeg_v1_0.h.

Signed-off-by: James Zhu <James.Zhu@amd.com>
Reported-by: kernel test robot <lkp@intel.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdkfd: fix ref count leak when pm_runtime_get_sync fails

The call to pm_runtime_get_sync increments the counter even in case of
failure, leading to incorrect ref count.
In case of failure, decrement the ref count before returning.

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

drm/amdgpu/fence: fix ref count leak when pm_runtime_get_sync fails

The call to pm_runtime_get_sync increments the counter even in case of
failure, leading to incorrect ref count.
In case of failure, decrement the ref count before returning.

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

drm/amdgpu/debugfs: fix memory leak when amdgpu_virt_enable_access_debugfs failed

Fix memory leak in amdgpu_debugfs_gpr_read not freeing data when
amdgpu_virt_enable_access_debugfs failed.

Fixes: 95a2f917387a2 ("drm/amdgpu: restrict debugfs register access under SR-IOV")
Signed-off-by: Chen Tao <chentao107@huawei.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu/debugfs: fix memory leak when pm_runtime_get_sync failed

Fix memory leak in amdgpu_debugfs_gpr_read not freeing data when
pm_runtime_get_sync failed.

Fixes: a9ffe2a983383 ("drm/amdgpu/debugfs: properly handle runtime pm")
Signed-off-by: Chen Tao <chentao107@huawei.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: vcn_v2_5_mc_resume_dpg_mode() can be static

Fixes: c422a628925d ("drm/amdgpu: rename macro for VCN2.0 2.5 and 3.0")
Signed-off-by: kernel test robot <lkp@intel.com>
Signed-off-by: James Zhu <James.Zhu@amd.com>
Reviewed-by: James Zhu <James.Zhu@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: fix compiler warning

Fixes below warning:
drivers/gpu/drm/amd/amdgpu/df_v3_6.c: In function ‘df_v3_6_reset_perfmon_cntr’:
drivers/gpu/drm/amd/amdgpu/df_v3_6.c:571:2: warning: ‘hi_base_addr’ may be used
uninitialized in this function [-Wmaybe-uninitialized]
  571 |  df_v3_6_perfmon_wreg(adev, lo_base_addr, 0, hi_base_addr, 0);
      |  ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

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

drm/amd/smu: unify smu ppt callback macros

v1: add the new macro "smu_ppt_xxx()" to unify smu callback interfaces
v2: rename the macro smu_ppt_xxx to smu_ppt_funcs.

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

drm/amd/smu: unify pptable_func{} callback interface

the pptable_func callback sets should be has unify interface,
so use "smu" as the pptable_func interface first parameter.

fix interfaces:
1. i2c_eeprom_init
2. i2c_eeprom_fini

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

drm/amd/display: Update DCN3 bounding box

[Why]
We want to update the bounding box to have more granular control of the
DCFCLK.

[How]
Setup DCFCLK to use STA values and also optimal values based on
UCLK.

Signed-off-by: Alvin Lee <alvin.lee2@amd.com>
Signed-off-by: Jerry (Fangzhi) Zuo <Jerry.Zuo@amd.com>
Reviewed-by: Hersen Wu <hersenxs.wu@amd.com>
Reviewed-by: Bhawanpreet Lakha <Bhawanpreet.Lakha@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: remove distinction between explicit and implicit sync (v2)

According to Marek a pipeline sync should be inserted for implicit syncs well.

v2: bump the driver version

Signed-off-by: Christian König <christian.koenig@amd.com>
Tested-by: Marek Olšák <marek.olsak@amd.com>
Signed-off-by: Marek Olšák <marek.olsak@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/amdgpu: Fix offset for SQ_DEBUG_STS_GLOBAL on gfx10 (v2)

Despite having different IP offsets the computed address of the register(s)
are the same between gfx7..gfx10.  This patch fixes the offset relative
to the GC block on gfx10.

(v2): SQ_DEBUG_STS_GLOBAL2 is 0x10 higher ...

Signed-off-by: Tom St Denis <tom.stdenis@amd.com>
Acked-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/radeon: Fix reference count leaks caused by pm_runtime_get_sync

On calling pm_runtime_get_sync() the reference count of the device
is incremented. In case of failure, decrement the
reference count before returning the error.

Acked-by: Evan Quan <evan.quan@amd.com>
Signed-off-by: Aditya Pakki <pakki001@umn.edu>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu/display: fix ref count leak when pm_runtime_get_sync fails

The call to pm_runtime_get_sync increments the counter even in case of
failure, leading to incorrect ref count.
In case of failure, decrement the ref count before returning.

Signed-off-by: Navid Emamdoost <navid.emamdoost@gmail.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: fix ref count leak in amdgpu_display_crtc_set_config

in amdgpu_display_crtc_set_config, the call to pm_runtime_get_sync
increments the counter even in case of failure, leading to incorrect
ref count. In case of failure, decrement the ref count before returning.

Signed-off-by: Navid Emamdoost <navid.emamdoost@gmail.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amd/display: fix ref count leak in amdgpu_drm_ioctl

in amdgpu_drm_ioctl the call to pm_runtime_get_sync increments the
counter even in case of failure, leading to incorrect
ref count. In case of failure, decrement the ref count before returning.

Signed-off-by: Navid Emamdoost <navid.emamdoost@gmail.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: fix ref count leak in amdgpu_driver_open_kms

in amdgpu_driver_open_kms the call to pm_runtime_get_sync increments the
counter even in case of failure, leading to incorrect
ref count. In case of failure, decrement the ref count before returning.

Signed-off-by: Navid Emamdoost <navid.emamdoost@gmail.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/radeon: fix multiple reference count leak

On calling pm_runtime_get_sync() the reference count of the device
is incremented. In case of failure, decrement the
reference count before returning the error.

Signed-off-by: Aditya Pakki <pakki001@umn.edu>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu/gfx9: Fix incorrect firmware size calculation

[WHY]
The memcpy() function copies n bytes from memory area src to memory area
dest. So specify the firmware size in bytes.

[How]
Correct the calculation.

Acked-by: Evan Quan <evan.quan@amd.com>
Signed-off-by: Lei Guo <raykwok1150@163.com>
Reviewed-by: Junwei Zhang <zjunweihit@163.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drm/amdgpu: fix documentation around busy_percentage

Add rename the gpu busy percentage for consistency and
add the mem busy percentage documentation.

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

drm/amdgpu/pm: update comment to clarify Overdrive interfaces

Vega10 and previous asics use one interface, vega20 and newer
use another.

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>