]> git.proxmox.com Git - mirror_ubuntu-focal-kernel.git/log
mirror_ubuntu-focal-kernel.git
5 years agodrm/powerplay: Fix Vega20 Average Power value v4
Kent Russell [Thu, 22 Aug 2019 12:17:40 +0000 (08:17 -0400)]
drm/powerplay: Fix Vega20 Average Power value v4

The SMU changed reading from CurrSocketPower to AverageSocketPower, so
reflect this accordingly. This fixes the issue where Average Power
Consumption was being reported as 0 from SMU 40.46-onward

v2: Fixed headline prefix
v3: Add check for SMU version for proper compatibility
v4: Style fix

Signed-off-by: Kent Russell <kent.russell@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agoamd/amdgpu: add Arcturus vf DID support
Frank.Min [Thu, 22 Aug 2019 22:16:36 +0000 (17:16 -0500)]
amd/amdgpu: add Arcturus vf DID support

Reviewed-by: Feifei Xu <Feifei.Xu@amd.com>
Signed-off-by: Frank.Min <Frank.Min@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amdgpu: unity mc base address for arcturus
Frank.Min [Wed, 21 Aug 2019 03:05:04 +0000 (11:05 +0800)]
drm/amdgpu: unity mc base address for arcturus

arcturus for sriov would use the unified mc base address

Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Frank.Min <Frank.Min@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amdgpu: disable agp for sriov
Frank.Min [Wed, 21 Aug 2019 09:20:34 +0000 (17:20 +0800)]
drm/amdgpu: disable agp for sriov

Since agp is not used for sriov, just disable it

Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Frank.Min <Frank.Min@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amdgpu/powerplay: silence a warning in smu_v11_0_setup_pptable
Alex Deucher [Thu, 22 Aug 2019 03:23:18 +0000 (22:23 -0500)]
drm/amdgpu/powerplay: silence a warning in smu_v11_0_setup_pptable

I think gcc is confused as I don't see how size could be used
uninitialized, but go ahead and silence the warning.

Reviewed-by: Evan Quan <evan.quan@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amdkfd: remove set but not used variable 'pdd'
YueHaibing [Fri, 26 Jul 2019 14:00:54 +0000 (22:00 +0800)]
drm/amdkfd: remove set but not used variable 'pdd'

Fixes gcc '-Wunused-but-set-variable' warning:

drivers/gpu/drm/amd/amdgpu/../amdkfd/kfd_process.c: In function restore_process_worker:
drivers/gpu/drm/amd/amdgpu/../amdkfd/kfd_process.c:949:29: warning:
 variable pdd set but not used [-Wunused-but-set-variable]

It is not used since
commit 5b87245faf57 ("drm/amdkfd: Simplify kfd2kgd interface")

Reported-by: Hulk Robot <hulkci@huawei.com>
Signed-off-by: YueHaibing <yuehaibing@huawei.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amd/display: remove duplicated include from dc_link.c
YueHaibing [Fri, 26 Jul 2019 07:51:31 +0000 (15:51 +0800)]
drm/amd/display: remove duplicated include from dc_link.c

Remove duplicated include.

Reported-by: Hulk Robot <hulkci@huawei.com>
Signed-off-by: YueHaibing <yuehaibing@huawei.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amdgpu: remove duplicated include from gfx_v9_0.c
YueHaibing [Wed, 10 Jul 2019 01:57:20 +0000 (01:57 +0000)]
drm/amdgpu: remove duplicated include from gfx_v9_0.c

Remove duplicated include.

Signed-off-by: YueHaibing <yuehaibing@huawei.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amdgpu: remove set but not used variable 'psp_enabled'
YueHaibing [Thu, 27 Jun 2019 07:53:50 +0000 (07:53 +0000)]
drm/amdgpu: remove set but not used variable 'psp_enabled'

Fixes gcc '-Wunused-but-set-variable' warning:

drivers/gpu/drm/amd/amdgpu/nv.c: In function 'nv_common_early_init':
drivers/gpu/drm/amd/amdgpu/nv.c:471:7: warning:
 variable 'psp_enabled' set but not used [-Wunused-but-set-variable]

It's not used since inroduction in
commit c6b6a42175f5 ("drm/amdgpu: add navi10 common ip block (v3)")

Signed-off-by: YueHaibing <yuehaibing@huawei.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amd/display: Fix 32-bit divide error in wait_for_alt_mode
Nathan Chancellor [Tue, 20 Aug 2019 23:57:13 +0000 (16:57 -0700)]
drm/amd/display: Fix 32-bit divide error in wait_for_alt_mode

When building arm32 allyesconfig:

ld.lld: error: undefined symbol: __aeabi_uldivmod
>>> referenced by dc_link.c
>>> gpu/drm/amd/display/dc/core/dc_link.o:(wait_for_alt_mode) in archive drivers/built-in.a
>>> referenced by dc_link.c
>>> gpu/drm/amd/display/dc/core/dc_link.o:(wait_for_alt_mode) in archive drivers/built-in.a

time_taken_in_ns is of type unsigned long long so we need to use div_u64
to avoid this error.

Fixes: b5b1f4554904 ("drm/amd/display: Enable type C hotplug")
Reported-by: Randy Dunlap <rdunlap@infradead.org>
Acked-by: Randy Dunlap <rdunlap@infradead.org> # build-tested
Reviewed-by: Harry Wentland <harry.wentland@amd.com>
Signed-off-by: Nathan Chancellor <natechancellor@gmail.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amdgpu: prevent memory leaks in AMDGPU_CS ioctl
Nicolai Hähnle [Tue, 20 Aug 2019 13:39:53 +0000 (15:39 +0200)]
drm/amdgpu: prevent memory leaks in AMDGPU_CS ioctl

Error out if the AMDGPU_CS ioctl is called with multiple SYNCOBJ_OUT and/or
TIMELINE_SIGNAL chunks, since otherwise the last chunk wins while the
allocated array as well as the reference counts of sync objects are leaked.

Signed-off-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amd/display: Lock the CRTC when setting CRC source
Nicholas Kazlauskas [Tue, 20 Aug 2019 17:15:25 +0000 (13:15 -0400)]
drm/amd/display: Lock the CRTC when setting CRC source

[Why]
We need to ensure that we're holding the lock on the CRTC when setting
the CRC source since we're modifying the CRTC state directly.

We also need to wait for any outstanding non-blocking commits to finish
so they aren't reading state that's potentially being modified -
non-blocking commits don't hold the CRTC lock while doing commit tail
work.

[How]
Lock the CRTC using its mutex. While holding the lock check if there's
any commit active on the CRTC - if there is, it's non-blocking and
we should wait until it's finished by waiting for hw_done to be
signaled since that's the last point where we touch CRTC state.

Signed-off-by: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com>
Reviewed-by: David Francis <David.Francis@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amd/display: Split out DC programming for CRC capture
Nicholas Kazlauskas [Tue, 20 Aug 2019 14:16:14 +0000 (10:16 -0400)]
drm/amd/display: Split out DC programming for CRC capture

[Why]
Calling amdgpu_dm_crtc_set_crc_source in amdgpu_dm directly has the
consequence of adding additional vblank references or starting DPRX
CRC capture more than once without calling stop first.

Vblank references for CRC capture should be managed entirely by opening
and closing the CRC file from userspace.

Stream state also shouldn't be required on the CRC so we can close the
file after the CRTC has been disabled.

[How]
Do DC programming required for configuring CRC capture separately from
setting the source. Whenever we re-enable or reset a CRC this
programming should be reapplied.

CRC vblank reference handling in amdgpu_dm can be entirely dropped after
this.

Stream state also no longer needs to be required since we can just defer
the programming to when the stream is actually enabled.

Signed-off-by: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com>
Reviewed-by: David Francis <David.Francis@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amd/display: Use connector list for finding DPRX CRC aux
Nicholas Kazlauskas [Tue, 20 Aug 2019 14:10:08 +0000 (10:10 -0400)]
drm/amd/display: Use connector list for finding DPRX CRC aux

[Why]
This change is a refactor in preparation for adding locking and removing
the requirement for a stream state on the CRTC for enabling CRC capture
to fix igt@kms_plane_multiple@* warnings.

[How]
We can get the aux by finding the matching connector for the CRTC
with the assumption that we're not doing cloning.

Signed-off-by: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com>
Reviewed-by: David Francis <David.Francis@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amd/display: Check return code for CRC drm_crtc_vblank_get
Nicholas Kazlauskas [Tue, 20 Aug 2019 16:58:37 +0000 (12:58 -0400)]
drm/amd/display: Check return code for CRC drm_crtc_vblank_get

[Why]
The call to drm_crtc_vblank_get can fail if vblank is disabled and
we try to increment the reference.

Since drm_crtc_vblank_get internally drops the reference when it fails
it means the subsequent drm_crtc_vblank_put(...) when closing the file
drops a zero reference.

This was found via igt@kms_plane@pixel-format-pipe-A-planes.

[How]
Check the return code and return it on failure.

We wouldn't have been able to enable CRC reading anyway since vblank
wasn't enabled.

Signed-off-by: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com>
Reviewed-by: David Francis <David.Francis@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amd/amdgpu: disable MMHUB PG for navi10
Kenneth Feng [Tue, 20 Aug 2019 07:11:37 +0000 (15:11 +0800)]
drm/amd/amdgpu: disable MMHUB PG for navi10

Disable MMHUB PG for navi10 according to the production requirement.

Signed-off-by: Kenneth Feng <kenneth.feng@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Reviewed-by: Kevin Wang <kevin1.wang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amd/powerplay: correct typo
Evan Quan [Tue, 20 Aug 2019 04:22:16 +0000 (12:22 +0800)]
drm/amd/powerplay: correct typo

"COMPUTE" was wrongly spelled as "CUSTOM".

Signed-off-by: Evan Quan <evan.quan@amd.com>
Reviewed-by: Kevin Wang <kevin1.wang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amd/powerplay: correct SW smu11 thermal range settings
Evan Quan [Fri, 16 Aug 2019 09:11:46 +0000 (17:11 +0800)]
drm/amd/powerplay: correct SW smu11 thermal range settings

Problems with current settings:
1. The min value was overrided to 0 on Vega20 & Navi10. While
   the expected should be -273.15 C.
2. The thermal min/max threshold was output in wrong unit on
   Navi10 & Arcturus. As TEMP_RANGE_MIN/MAX is already in
   millicelsius. And "*1000" in smu_v11_0_start_thermal_control
   makes the output wrongly.

Signed-off-by: Evan Quan <evan.quan@amd.com>
Reviewed-by: Kenneth Feng <kenneth.feng@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amd/powerplay: set Arcturus default fclk as bootup value on dpm disabled
Evan Quan [Fri, 16 Aug 2019 05:52:26 +0000 (13:52 +0800)]
drm/amd/powerplay: set Arcturus default fclk as bootup value on dpm disabled

On fclk dpm disabled, the default dpm table will be setup with only one
level and clock frequency as bootup value.

Signed-off-by: Evan Quan <evan.quan@amd.com>
Reviewed-by: Kevin Wang <kevin1.wang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amd/powerplay: get bootup fclk value
Evan Quan [Fri, 16 Aug 2019 05:47:01 +0000 (13:47 +0800)]
drm/amd/powerplay: get bootup fclk value

This is available with firmwareinfo table v3.2 or later.

Signed-off-by: Evan Quan <evan.quan@amd.com>
Reviewed-by: Kevin Wang <kevin1.wang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amd/powerplay: expose supported clock domains only through sysfs
Evan Quan [Fri, 16 Aug 2019 03:34:12 +0000 (11:34 +0800)]
drm/amd/powerplay: expose supported clock domains only through sysfs

Do not expose those unsupported clock domains through sysfs on
Arcturus.

Signed-off-by: Evan Quan <evan.quan@amd.com>
Reviewed-by: Kevin Wang <kevin1.wang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amd/powerplay: update Arcturus smc fw and driver interface header
Evan Quan [Fri, 16 Aug 2019 02:55:21 +0000 (10:55 +0800)]
drm/amd/powerplay: update Arcturus smc fw and driver interface header

Update smc fw and driver interface header.

Signed-off-by: Evan Quan <evan.quan@amd.com>
Reviewed-by: Kevin Wang <kevin1.wang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amd/powerplay: remove duplicate macro smu_get_uclk_dpm_states in amdgpu_smu.h
Kevin Wang [Tue, 20 Aug 2019 05:28:51 +0000 (13:28 +0800)]
drm/amd/powerplay: remove duplicate macro smu_get_uclk_dpm_states in amdgpu_smu.h

remove duplicate macro smu_get_uclk_dpm_states in amdgpu_smu.h

"
 #define smu_get_uclk_dpm_states(smu, clocks_in_khz, num_states) \
         ((smu)->ppt_funcs->get_uclk_dpm_states ? (smu)->ppt_funcs->get_uclk_dpm_states((smu), (clocks_in_khz), (num_states)) : 0)
 #define smu_get_max_sustainable_clocks_by_dc(smu, max_clocks) \
         ((smu)->funcs->get_max_sustainable_clocks_by_dc ? (smu)->funcs->get_max_sustainable_clocks_by_dc((smu), (max_clocks)) : 0)
 #define smu_get_uclk_dpm_states(smu, clocks_in_khz, num_states) \
         ((smu)->ppt_funcs->get_uclk_dpm_states ? (smu)->ppt_funcs->get_uclk_dpm_states((smu), (clocks_in_khz), (num_states)) : 0)
"

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>
5 years agodrm/amd/powerplay: fix variable type errors in smu_v11_0_setup_pptable
Kevin Wang [Mon, 19 Aug 2019 15:38:02 +0000 (23:38 +0800)]
drm/amd/powerplay: fix variable type errors in smu_v11_0_setup_pptable

fix size type errors, from uint32_t to uint16_t.
it will cause only initializes the highest 16 bits in
smu_get_atom_data_table function.

bug report:
This fixes the following static checker warning.
        drivers/gpu/drm/amd/amdgpu/../powerplay/smu_v11_0.c:390 smu_v11_0_setup_pptable()
        warn: passing casted pointer '&size' to 'smu_get_atom_data_table()' 32 vs 16.

Signed-off-by: Kevin Wang <kevin1.wang@amd.com>
Reported-by: Dan Carpenter <dan.carpenter@oracle.com>
Reviewed-by: Evan Quan <evan.quan@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amd/powerplay: add smu_smc_read_sensor support for arcturus
Kevin Wang [Fri, 16 Aug 2019 07:17:42 +0000 (15:17 +0800)]
drm/amd/powerplay: add smu_smc_read_sensor support for arcturus

the bellow patch refine the sensor read sequence,
but missed to add arcuturs support. (arcuturs_ppt.c)
it will cause some sensor is not supported in arcturus.

drm/amd/powerplay: change smu_read_sensor sequence in smu

Signed-off-by: Kevin Wang <kevin1.wang@amd.com>
Reviewed-by: Evan Quan <evan.quan@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amdgpu: Fix a typo in the include header guard of 'navi12_ip_offset.h'
Christophe JAILLET [Sun, 18 Aug 2019 15:59:57 +0000 (17:59 +0200)]
drm/amdgpu: Fix a typo in the include header guard of 'navi12_ip_offset.h'

'_navi10_ip_offset_HEADER' is already used in 'navi10_ip_offset.h', so use
'_navi12_ip_offset_HEADER' instead here.

Reviewed-by: Xiaojie Yuan <xiaojie.yuan@amd.com>
Signed-off-by: Christophe JAILLET <christophe.jaillet@wanadoo.fr>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amdgpu/psp: move TMR to cpu invisible vram region
Tianci.Yin [Mon, 19 Aug 2019 07:30:22 +0000 (15:30 +0800)]
drm/amdgpu/psp: move TMR to cpu invisible vram region

so that more visible vram can be available for umd.

Reviewed-by: Christian König <christian.koenig@amd.com>.
Signed-off-by: Tianci.Yin <tianci.yin@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amdgpu: remove redundant argument for psp_funcs::cmd_submit callback
Xiaojie Yuan [Fri, 16 Aug 2019 10:00:54 +0000 (18:00 +0800)]
drm/amdgpu: remove redundant argument for psp_funcs::cmd_submit callback

Signed-off-by: Xiaojie Yuan <xiaojie.yuan@amd.com>
Reviewed-by: Evan Quan <evan.quan@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amdkfd: Fill the name field in node topology with asic name v2
Yong Zhao [Fri, 2 Aug 2019 02:55:50 +0000 (22:55 -0400)]
drm/amdkfd: Fill the name field in node topology with asic name v2

The name field in node topology has not been used. We re-purpose it to
hold the asic name, which can be queried by user space applications
through sysfs.

Signed-off-by: Yong Zhao <Yong.Zhao@amd.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amdgpu: Set no-retry as default.
Feifei Xu [Fri, 16 Aug 2019 03:08:52 +0000 (11:08 +0800)]
drm/amdgpu: Set no-retry as default.

This is to improve performance.

Signed-off-by: Feifei Xu <Feifei.Xu@amd.com>
Tested-by: Candice Li <candice.li@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amdgpu: add firmware header printing for psp fw loading (v2)
Xiaojie Yuan [Thu, 15 Aug 2019 09:44:22 +0000 (17:44 +0800)]
drm/amdgpu: add firmware header printing for psp fw loading (v2)

firmware header information is printed for direct fw loading but not
added for psp fw loading yet

v2: squash in warning fix (Alex)

Signed-off-by: Xiaojie Yuan <xiaojie.yuan@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amdgpu: fix debug level for ppt offset/size
Xiaojie Yuan [Thu, 15 Aug 2019 09:43:36 +0000 (17:43 +0800)]
drm/amdgpu: fix debug level for ppt offset/size

Signed-off-by: Xiaojie Yuan <xiaojie.yuan@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amdgpu: remove special autoload handling for navi12
Xiaojie Yuan [Thu, 15 Aug 2019 07:08:02 +0000 (15:08 +0800)]
drm/amdgpu: remove special autoload handling for navi12

s/r list in rlc firmware is ready, so remove the special autoload handling

Signed-off-by: Xiaojie Yuan <xiaojie.yuan@amd.com>
Acked-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amd/display: Calculate bpc based on max_requested_bpc
Nicholas Kazlauskas [Wed, 21 Aug 2019 15:27:13 +0000 (11:27 -0400)]
drm/amd/display: Calculate bpc based on max_requested_bpc

[Why]
The only place where state->max_bpc is updated on the connector is
at the start of atomic check during drm_atomic_connector_check. It
isn't updated when adding the connectors to the atomic state after
the fact. It also doesn't necessarily reflect the right value when
called in amdgpu during mode validation outside of atomic check.

This can cause the wrong bpc to be used even if the max_requested_bpc
is the correct value.

[How]
Don't rely on state->max_bpc reflecting the real bpc value and just
do the min(...) based on display info bpc and max_requested_bpc.

Fixes: 01933ba42d3d ("drm/amd/display: Use current connector state if NULL when checking bpc")
Signed-off-by: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com>
Reviewed-by: Leo Li <sunpeng.li@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amdgpu/gfx9: update pg_flags after determining if gfx off is possible
Alex Deucher [Thu, 15 Aug 2019 13:27:09 +0000 (08:27 -0500)]
drm/amdgpu/gfx9: update pg_flags after determining if gfx off is possible

We need to set certain power gating flags after we determine
if the firmware version is sufficient to support gfxoff.
Previously we set the pg flags in early init, but we later
we might have disabled gfxoff if the firmware versions didn't
support it.  Move adding the additional pg flags after we
determine whether or not to support gfxoff.

Fixes: 005440066f92 ("drm/amdgpu: enable gfxoff again on raven series (v2)")
Tested-by: Kai-Heng Feng <kai.heng.feng@canonical.com>
Tested-by: Tom St Denis <tom.stdenis@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Cc: Kai-Heng Feng <kai.heng.feng@canonical.com>
5 years agodrm/amdgpu: Use new mode2 reset interface for RV.
Andrey Grodzovsky [Fri, 26 Jul 2019 18:07:42 +0000 (14:07 -0400)]
drm/amdgpu: Use new mode2 reset interface for RV.

Integrate the mode2 reset into rest sequence.

v2:
Check ppfuncs pointer for NULL

Signed-off-by: Andrey Grodzovsky <andrey.grodzovsky@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Evan Quan <evan.quan@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amd/poweplay: Add amd_pm_funcs callback for mode 2
Andrey Grodzovsky [Wed, 14 Aug 2019 20:26:35 +0000 (16:26 -0400)]
drm/amd/poweplay: Add amd_pm_funcs callback for mode 2

Add callback to call the new mode2 reset interface.

Signed-off-by: Andrey Grodzovsky <andrey.grodzovsky@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Evan Quan <evan.quan@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amd/powerpay: Implement mode2 reset callback for SMU10
Andrey Grodzovsky [Wed, 14 Aug 2019 20:21:44 +0000 (16:21 -0400)]
drm/amd/powerpay: Implement mode2 reset callback for SMU10

Add implmenetion.

Signed-off-by: Andrey Grodzovsky <andrey.grodzovsky@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Evan Quan <evan.quan@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amd/powerplay: add mode2 reset callback for pp_smu_mgr
Andrey Grodzovsky [Wed, 14 Aug 2019 20:16:52 +0000 (16:16 -0400)]
drm/amd/powerplay: add mode2 reset callback for pp_smu_mgr

Also define reset modes (0, 1 and 2)

Signed-off-by: Andrey Grodzovsky <andrey.grodzovsky@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Evan Quan <evan.quan@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amd/powerplay: Fix meaning of 0x1E PPSMC_MSG
Andrey Grodzovsky [Fri, 26 Jul 2019 18:03:56 +0000 (14:03 -0400)]
drm/amd/powerplay: Fix meaning of 0x1E PPSMC_MSG

By comparing to windows it means ASIC reset.

Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Andrey Grodzovsky <andrey.grodzovsky@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodmr/amdgpu: Fix compile error with CONFIG_DRM_AMDGPU_GART_DEBUGFS
Andrey Grodzovsky [Wed, 14 Aug 2019 19:44:14 +0000 (15:44 -0400)]
dmr/amdgpu: Fix compile error with CONFIG_DRM_AMDGPU_GART_DEBUGFS

Double defintion of 'i'

Signed-off-by: Andrey Grodzovsky <andrey.grodzovsky@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amd/powerplay/smu7: enforce minimal VBITimeout (v2)
Ahzo [Mon, 5 Aug 2019 19:14:18 +0000 (21:14 +0200)]
drm/amd/powerplay/smu7: enforce minimal VBITimeout (v2)

This fixes screen corruption/flickering on 75 Hz displays.

v2: make print statement debug only (Alex)

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=102646
Reviewed-by: Evan Quan <evan.quan@amd.com>
Signed-off-by: Ahzo <Ahzo@tutanota.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amd/amdgpu: Update VM function pointer
Gang Ba [Wed, 14 Aug 2019 15:00:19 +0000 (11:00 -0400)]
drm/amd/amdgpu: Update VM function pointer

When VM state changed and system in large bar mode,
make sure to use CPU update function, otherwise use
SDMA function.

Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Gang Ba <gaba@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amdkfd: Fill amdgpu_task_info for KFD VMs
Yong Zhao [Wed, 7 Aug 2019 21:45:37 +0000 (17:45 -0400)]
drm/amdkfd: Fill amdgpu_task_info for KFD VMs

The amdgpu_task_info will be used when printing VM page fault for KFD
processes.

Signed-off-by: Yong Zhao <Yong.Zhao@amd.com>
Acked-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Harish Kasiviswanathan <harish.kasiviswanatha@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amdgpu: Set VM_L2_CNTL.PDE_FAULT_CLASSIFICATION to 0 for GFX10
Yong Zhao [Tue, 13 Aug 2019 18:38:03 +0000 (14:38 -0400)]
drm/amdgpu: Set VM_L2_CNTL.PDE_FAULT_CLASSIFICATION to 0 for GFX10

We have done this for pre-GFX10 asics, but GFX10 did not pick up the
new change. The below is the commit message for that change.

This is recommended by HW designers. Previously when it was set to 1,
the PDE walk error in VM fault will be treated as
PERMISSION_OR_INVALID_PAGE_FAULT rather than usually expected OTHER_FAULT.
As a result, the retry control in VM_CONTEXT*_CNTL will change accordingly.

The above behavior is kind of abnormal. Furthermore, the
PDE_FAULT_CLASSIFICATION == 1 feature was targeted for very old ASICs
and it never made it way to production. Therefore, we should set it to 0.

Signed-off-by: Yong Zhao <Yong.Zhao@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amdgpu: Add more page fault info printing for GFX10
Yong Zhao [Wed, 7 Aug 2019 20:21:09 +0000 (16:21 -0400)]
drm/amdgpu: Add more page fault info printing for GFX10

The printing we did for GFX9 was not propogated to GFX10 somehow, so fix
it now.

Signed-off-by: Yong Zhao <Yong.Zhao@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amdgpu: Add printing for RW extracted from VM_L2_PROTECTION_FAULT_STATUS
Yong Zhao [Tue, 13 Aug 2019 18:04:26 +0000 (14:04 -0400)]
drm/amdgpu: Add printing for RW extracted from VM_L2_PROTECTION_FAULT_STATUS

RW is also useful in most cases.

Signed-off-by: Yong Zhao <Yong.Zhao@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amdkfd/gfx10: Calling amdgpu functions to invalidate TLB
Oak Zeng [Thu, 1 Aug 2019 19:24:42 +0000 (14:24 -0500)]
drm/amdkfd/gfx10: Calling amdgpu functions to invalidate TLB

Calling amdgpu function to invalidate TLB, instead of using a
kfd implementation. Delete the kfd local TLB invalidation
implementation.

Signed-off-by: Oak Zeng <Oak.Zeng@amd.com>
Reviewed-by: Christian Konig <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amdgpu: Export function to flush TLB of specific vm hub
Oak Zeng [Thu, 1 Aug 2019 19:55:45 +0000 (14:55 -0500)]
drm/amdgpu: Export function to flush TLB of specific vm hub

This is for kfd to reuse amdgpu TLB invalidation function.
On gfx10, kfd only needs to flush TLB on gfx hub but not
on mm hub. So export a function for KFD flush TLB only on
specific hub.

Signed-off-by: Oak Zeng <Oak.Zeng@amd.com>
Reviewed-by: Christian Konig <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amd/display: Load NV12 SOC BB from firmware
Leo Li [Tue, 6 Aug 2019 17:50:28 +0000 (13:50 -0400)]
drm/amd/display: Load NV12 SOC BB from firmware

[Why]

Previous SOC bounding box firmware loading logic was for NV10, when we
still had it in firmware. Now that it's brought into driver code, and
NV12 BB is in firmware, this logic needs to be repurposed for NV12.

[How]

Set SOC_BOUNDING_BOX_VALID to false, and add the dcn_2_0_nv12_soc BB
struct. In init_soc_bounding_box, load firmware to nv12 BB instead.

In addition, conditionally update and patch the BB, depending on ASIC
REV.

Signed-off-by: Leo Li <sunpeng.li@amd.com>
Reviewed-by: Hersen Wu <hersenxs.wu@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amd/display: Enable MPO with pre-blend color processing (RGB)
Michael Strauss [Fri, 26 Jul 2019 16:04:12 +0000 (12:04 -0400)]
drm/amd/display: Enable MPO with pre-blend color processing (RGB)

[Why]
DCN10 performs color processing before MPC combination, causes color
shift in RGB colorspaces when positive brightness offset is applied
However, YCbCr is still unfixed and remains disabled

[How]
Add layerIndex to dc_plane_state and dc_plane_info structs
Re-enable MPO when brightness is adjusted and colorspace is not YCbCr
Set rear plane's brightness offset to 0 when front plane visible

Signed-off-by: Michael Strauss <michael.strauss@amd.com>
Reviewed-by: Anthony Koo <Anthony.Koo@amd.com>
Acked-by: Leo Li <sunpeng.li@amd.com>
Acked-by: Nicholas Kazlauskas <Nicholas.Kazlauskas@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amd/display: audio cannot switch to internal when display turns off
yanyan kang [Mon, 29 Jul 2019 09:46:44 +0000 (17:46 +0800)]
drm/amd/display: audio cannot switch to internal when display turns off

[why]
disable_az_endpoint has been skipped because
dc->debug.az_endpoint_mute_only = true.

[how]
set dc->debug.az_endpoint_mute_only false when PPLIB’s PME notification function
 pointer is not NULL at the dcn10_resource construct function,because right now
 SMU/PPLIB and DAL all have the AZ D3 force PME notification implemented. AZ D3 should work.

Signed-off-by: yanyan kang <Yanyan.Kang@amd.com>
Reviewed-by: Charlene Liu <Charlene.Liu@amd.com>
Acked-by: Leo Li <sunpeng.li@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amd/display: check hpd before retry verify link cap
Wenjing Liu [Fri, 26 Jul 2019 18:53:20 +0000 (14:53 -0400)]
drm/amd/display: check hpd before retry verify link cap

[why]
During detection link training if a display is disconnected,
the current code will retry 3 times of link training
on disconnected link before giving up.

[how]
Before each retry check for HPD status, only retry
verify link cap when HPD is still high.
Also put a 10ms delay between each retry to improve
the chance of success.

Signed-off-by: Wenjing Liu <Wenjing.Liu@amd.com>
Reviewed-by: Jun Lei <Jun.Lei@amd.com>
Acked-by: Abdoulaye Berthe <Abdoulaye.Berthe@amd.com>
Acked-by: Leo Li <sunpeng.li@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amd/display: enable dcn_mem_pwr as golden setting updates
Charlene Liu [Mon, 29 Jul 2019 15:59:33 +0000 (11:59 -0400)]
drm/amd/display: enable dcn_mem_pwr as golden setting updates

Enable dcn_mem_pwr as golden setting updates

Signed-off-by: Charlene Liu <charlene.liu@amd.com>
Reviewed-by: Dmytro Laktyushkin <Dmytro.Laktyushkin@amd.com>
Acked-by: Leo Li <sunpeng.li@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amd/display: 3.2.46
Anthony Koo [Mon, 29 Jul 2019 16:17:30 +0000 (12:17 -0400)]
drm/amd/display: 3.2.46

Signed-off-by: Anthony Koo <Anthony.Koo@amd.com>
Acked-by: Leo Li <sunpeng.li@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amd/display: fix dcn20 global sync dml param extraction
Dmytro Laktyushkin [Fri, 26 Jul 2019 21:32:02 +0000 (17:32 -0400)]
drm/amd/display: fix dcn20 global sync dml param extraction

Currently the paremeters are extracted as if dml is calculating
using pipes as we pass them in. in reality, dml internally merges
pipes into planes if pipe split is detected.

This change adds reverse logic to dcn20_calculate_dlg_params so
that the global sync parameters can be correctly extracted for
all the pipes when pipe split is enabled.

Signed-off-by: Dmytro Laktyushkin <Dmytro.Laktyushkin@amd.com>
Reviewed-by: Charlene Liu <Charlene.Liu@amd.com>
Acked-by: Leo Li <sunpeng.li@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amd/display: 3.2.45
Anthony Koo [Sat, 27 Jul 2019 19:10:54 +0000 (15:10 -0400)]
drm/amd/display: 3.2.45

Signed-off-by: Anthony Koo <Anthony.Koo@amd.com>
Acked-by: Leo Li <sunpeng.li@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amd/display: Add and refine DSC logs in enable sequence
Nikola Cornij [Thu, 25 Jul 2019 22:46:54 +0000 (18:46 -0400)]
drm/amd/display: Add and refine DSC logs in enable sequence

[why]
Some logs messages were not precise and some new log messages
were needed after "get packed PPS" function was introduced

Signed-off-by: Nikola Cornij <nikola.cornij@amd.com>
Reviewed-by: Wenjing Liu <Wenjing.Liu@amd.com>
Acked-by: Leo Li <sunpeng.li@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amd/display: Add Logging for Gamma Related information (2/2)
Wyatt Wood [Wed, 24 Jul 2019 17:37:08 +0000 (13:37 -0400)]
drm/amd/display: Add Logging for Gamma Related information (2/2)

[Why]
A recent bug showed that logging would be useful in debugging
various gamma issues.

[How]
Add logging in dc.

Signed-off-by: Wyatt Wood <wyatt.wood@amd.com>
Reviewed-by: Krunoslav Kovac <Krunoslav.Kovac@amd.com>
Acked-by: Leo Li <sunpeng.li@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amd/display: Remove redundant definition of dwb_source enums
Julian Parkin [Thu, 25 Jul 2019 20:53:27 +0000 (16:53 -0400)]
drm/amd/display: Remove redundant definition of dwb_source enums

There are repeated (but guarded) definitions of dwb_src enums. There are
also unused entires. Clean them up.

Signed-off-by: Julian Parkin <julian.parkin@amd.com>
Reviewed-by: Charlene Liu <Charlene.Liu@amd.com>
Acked-by: Leo Li <sunpeng.li@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amd/display: Add enum for H-timing divider mode
Qingqing Zhuo [Tue, 23 Jul 2019 16:24:24 +0000 (12:24 -0400)]
drm/amd/display: Add enum for H-timing divider mode

Add h_timing_div_mode enum to better reflect possible register
values. Replace previously programmed values with enum

Signed-off-by: Qingqing Zhuo <qingqing.zhuo@amd.com>
Reviewed-by: Dmytro Laktyushkin <Dmytro.Laktyushkin@amd.com>
Acked-by: Leo Li <sunpeng.li@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amd/display: Register VUPDATE_NO_LOCK interrupts for DCN2
Nicholas Kazlauskas [Thu, 25 Jul 2019 15:53:16 +0000 (11:53 -0400)]
drm/amd/display: Register VUPDATE_NO_LOCK interrupts for DCN2

[Why]
These are needed to send back DRM vblank events in the case where VRR
is on. Without the interrupt enabled we're deferring the events into the
vblank queue and userspace is left waiting forever to get back the
events they need.

Found using igt@kms_vrr - the test fails immediately due to vblank
timeout.

[How]
Register them the same way we're handling it for DCN1.

This fixes igt@kms_vrr for DCN2.

Signed-off-by: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com>
Reviewed-by: David Francis <David.Francis@amd.com>
Acked-by: Leo Li <sunpeng.li@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amd/display: wake up ogam mem pwr before programming ocsc
Charlene Liu [Wed, 24 Jul 2019 22:14:46 +0000 (18:14 -0400)]
drm/amd/display: wake up ogam mem pwr before programming ocsc

[Description]
OGAM_MEM_PWR could stay in light up when driver woke up to update gamma.
either disable MEM_LOW power feature or set to OGAM_bypass could make artificial color distortion goes away.
Easy reproduce after LOW_MEM Power feature enables and resume from S3.

Signed-off-by: Charlene Liu <charlene.liu@amd.com>
Reviewed-by: Julian Parkin <jparkin@amd.com>
Acked-by: Leo Li <sunpeng.li@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amd/display: Correct DSC PPS log
Nikola Cornij [Thu, 25 Jul 2019 21:27:03 +0000 (17:27 -0400)]
drm/amd/display: Correct DSC PPS log

[why]
A misleading message "Programming PPS" appears before both programming
and "query PPS" functions

[how]
Move the message from "log PPS" function to "program PPS" function

Signed-off-by: Nikola Cornij <nikola.cornij@amd.com>
Reviewed-by: Wenjing Liu <Wenjing.Liu@amd.com>
Acked-by: Leo Li <sunpeng.li@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amd/display: Check if set_blank_data_double_buffer exists before call
Vitaly Prosyak [Fri, 19 Jul 2019 21:07:33 +0000 (16:07 -0500)]
drm/amd/display: Check if set_blank_data_double_buffer exists before call

Not all ASIC types have this function implemented - check before
calling.

Signed-off-by: Vitaly Prosyak <vitaly.prosyak@amd.com>
Reviewed-by: Julian Parkin <jparkin@amd.com>
Acked-by: Leo Li <sunpeng.li@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amd/display: Fix type of ODMCombineType field
Ilya Bakoulin [Wed, 24 Jul 2019 16:36:55 +0000 (12:36 -0400)]
drm/amd/display: Fix type of ODMCombineType field

The type was changed previously to better reflect possible register
values.

Signed-off-by: Ilya Bakoulin <Ilya.Bakoulin@amd.com>
Reviewed-by: Dmytro Laktyushkin <Dmytro.Laktyushkin@amd.com>
Acked-by: Leo Li <sunpeng.li@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amd/display: Add Logging for Gamma Related information (1/2)
Wyatt Wood [Wed, 24 Jul 2019 17:29:38 +0000 (13:29 -0400)]
drm/amd/display: Add Logging for Gamma Related information (1/2)

[Why]
A recent bug showed that logging would be useful in
debugging various gamma issues.

[How]
Add new log types and logging code to the color module.

Signed-off-by: Wyatt Wood <wyatt.wood@amd.com>
Reviewed-by: Aric Cyr <Aric.Cyr@amd.com>
Acked-by: Leo Li <sunpeng.li@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amd/display: Remove 4:2:2 DSC support
Nikola Cornij [Tue, 23 Jul 2019 20:38:02 +0000 (16:38 -0400)]
drm/amd/display: Remove 4:2:2 DSC support

[why]
If DSC is available, a higher picture quality is achieved by using
DSC with 4:4:4 format. Using 4:2:2 instead does not offer any benefit
and would only introduce loss of quality. Removing it reduces
maintenance and testing effort.

Signed-off-by: Nikola Cornij <nikola.cornij@amd.com>
Reviewed-by: Jun Lei <Jun.Lei@amd.com>
Acked-by: Leo Li <sunpeng.li@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amd/display: reprogram VM config when system resume
Lewis Huang [Mon, 22 Jul 2019 19:23:32 +0000 (15:23 -0400)]
drm/amd/display: reprogram VM config when system resume

[Why]
The vm config will be clear to 0 when system enter S4. It will
cause hubbub didn't know how to fetch data when system resume.
The flip always pending because earliest_inuse_address and
request_address are different.

[How]
Reprogram VM config when system resume

Signed-off-by: Lewis Huang <Lewis.Huang@amd.com>
Reviewed-by: Jun Lei <Jun.Lei@amd.com>
Acked-by: Eric Yang <eric.yang2@amd.com>
Acked-by: Leo Li <sunpeng.li@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amd/display: Enable type C hotplug
Eric Yang [Thu, 11 Jul 2019 02:41:51 +0000 (22:41 -0400)]
drm/amd/display: Enable type C hotplug

[Why and How]
We want to change where timing is done for alt mode.
Some of the commented out #ifs are needed for DCN20
so we enable them for that case.

Signed-off-by: Eric Yang <Eric.Yang2@amd.com>
Reviewed-by: Eric Yang <eric.yang2@amd.com>
Acked-by: Leo Li <sunpeng.li@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amd/display: HUBP/HUBBUB register programming fixes
Ilya Bakoulin [Mon, 22 Jul 2019 18:12:25 +0000 (14:12 -0400)]
drm/amd/display: HUBP/HUBBUB register programming fixes

[Why]
- Need to change interface function signature / add an enum
  to reflect the available register field values

[How]
- Add a new enum and modify existing functions to use it instead
  of bool

Signed-off-by: Ilya Bakoulin <Ilya.Bakoulin@amd.com>
Reviewed-by: Charlene Liu <Charlene.Liu@amd.com>
Acked-by: Leo Li <sunpeng.li@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amd/display: update optc odm interface for more than 2 opps
Dmytro Laktyushkin [Fri, 19 Jul 2019 15:43:39 +0000 (11:43 -0400)]
drm/amd/display: update optc odm interface for more than 2 opps

Current optc odm interface only accepts 2 opps, we need to
expand this to allow 4 to 1 odm combine.

Signed-off-by: Dmytro Laktyushkin <Dmytro.Laktyushkin@amd.com>
Reviewed-by: Charlene Liu <Charlene.Liu@amd.com>
Acked-by: Leo Li <sunpeng.li@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amd/display: Update DML parameters
Ilya Bakoulin [Wed, 17 Jul 2019 22:33:48 +0000 (18:33 -0400)]
drm/amd/display: Update DML parameters

[Why]
Need to add DML struct members that were omitted in previous
DML implemenations.

[How]
- Add missing enum values
- Add missing struct members
- Set new input values in the fetch functions

Signed-off-by: Ilya Bakoulin <Ilya.Bakoulin@amd.com>
Reviewed-by: Charlene Liu <Charlene.Liu@amd.com>
Acked-by: Leo Li <sunpeng.li@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amd/display: Remove duplicate interface for programming FB
Julian Parkin [Mon, 15 Jul 2019 16:16:01 +0000 (12:16 -0400)]
drm/amd/display: Remove duplicate interface for programming FB

[Why]
There are currently two interfaces for exactly the same thing:
hupb_update_dchub in hupb and update_dchub in hubbub. The hubbub
version is currently unused past dcn10, largely because the call
from the dcn10 hardware sequencer does not call through the
interface, so the hupb interface was used instead. This is
confusing because of the duplicate code, the unused functions,
and the fact that more that one block currently owns this set
of registers.

[How]
Remove the hubp interface entirely, as well as the register
declarations that are not longer needed because of this. Change
the call site to always call the hubbub version through the
interface. Fix the update_dchub function in dcn20_hubbub.c to
program the correct registers for dcn20.

Signed-off-by: Julian Parkin <julian.parkin@amd.com>
Reviewed-by: Aric Cyr <Aric.Cyr@amd.com>
Acked-by: Leo Li <sunpeng.li@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amd/display: fix pipe selection logic in validate
Jun Lei [Mon, 22 Jul 2019 13:45:20 +0000 (09:45 -0400)]
drm/amd/display: fix pipe selection logic in validate

[why]
Resource mapping done in dcn20_validate_bandwidth has a flaw:  When a full
update is performed, the HWSS will only update the MPCC tree for the stream
that is updated as opposed to all streams.  This means that when mapping pipes
in validation, care must be taken to not change any existing mapping, otherwise it
leads to partial hw programming

[how]
it's not strictly necessary to track which stream/mpcc tree is being updated, but
rather it's sufficient to compare current and new state and just keep pipes that were
previously already mapped unchanged.

Signed-off-by: Jun Lei <Jun.Lei@amd.com>
Reviewed-by: Eric Yang <eric.yang2@amd.com>
Acked-by: Leo Li <sunpeng.li@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amd/display: 3.2.44
Aric Cyr [Fri, 19 Jul 2019 21:41:56 +0000 (17:41 -0400)]
drm/amd/display: 3.2.44

Signed-off-by: Aric Cyr <aric.cyr@amd.com>
Acked-by: Leo Li <sunpeng.li@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amd/display: Set DSC before DIG front-end is connected to its back-end
Nikola Cornij [Wed, 17 Jul 2019 23:02:14 +0000 (19:02 -0400)]
drm/amd/display: Set DSC before DIG front-end is connected to its back-end

[why]
At the time DIG FE is connected to its BE, the clocks in OTG are enabled and
PHY will also be set up. When DSC has to be used to fit the stream into the
available bandwidth, without DSC being set DIG could get exposed to the
higer bandwidth it (or link) could handle. This causes the HW to "reject"
video enable setup (the register shows that video enable was attempted, but
the status bit shows it as disabled).

[how]
- Separate DSC setup into DSC register config and DSC PPS SDP setup

- Move most of the DSC setup (register config) to before
  dcn10_link_encoder_connect_dig_be_to_fe() is  called

- Set up DSC PPS SDP after DIG FE is connected to its BE. This is because
  setting DSC PPS SDP before that has no effect.

Signed-off-by: Nikola Cornij <nikola.cornij@amd.com>
Reviewed-by: Jun Lei <Jun.Lei@amd.com>
Acked-by: Leo Li <sunpeng.li@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amd/display: Change DSC policy from slices per column to minimum slice height
Nikola Cornij [Thu, 18 Jul 2019 23:13:26 +0000 (19:13 -0400)]
drm/amd/display: Change DSC policy from slices per column to minimum slice height

[why] Minimum slice height is recommended by VESA DSC Spreadsheet user guide

Signed-off-by: Nikola Cornij <nikola.cornij@amd.com>
Reviewed-by: Jun Lei <Jun.Lei@amd.com>
Acked-by: Leo Li <sunpeng.li@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amd/display: Add HLG support in color module
Vitaly Prosyak [Thu, 18 Jul 2019 18:07:03 +0000 (13:07 -0500)]
drm/amd/display: Add HLG support in color module

[Why & How]
Support hlg OETF and EOTF based on BT.2100-2
Follow up is required.

Signed-off-by: Vitaly Prosyak <vitaly.prosyak@amd.com>
Reviewed-by: Krunoslav Kovac <Krunoslav.Kovac@amd.com>
Acked-by: Leo Li <sunpeng.li@amd.com>
Acked-by: Vitaly Prosyak <Vitaly.Prosyak@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amd/display: clean up DML for DCN2x
Jun Lei [Thu, 18 Jul 2019 14:02:40 +0000 (10:02 -0400)]
drm/amd/display: clean up DML for DCN2x

[why]
Previous "less risky" implemenation of 3 tiered fallback is no longer necessary since
DMLv2 has gone through proper validation.  v2 can now be used as the default and 1
level of fallback can be removed

[how]
remove previous workaround implemenation

Signed-off-by: Jun Lei <Jun.Lei@amd.com>
Reviewed-by: Nicholas Kazlauskas <Nicholas.Kazlauskas@amd.com>
Acked-by: Leo Li <sunpeng.li@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amd/display: enabling seamless boot sequence for dcn2
Martin Leung [Wed, 17 Jul 2019 20:08:19 +0000 (16:08 -0400)]
drm/amd/display: enabling seamless boot sequence for dcn2

[Why]
Seamless boot (building SW state inheriting BIOS-initialized timing) was
enabled on DCN2, including fixes

[How]
Includes fixes for MPC, DPPCLK, and DIG FE mapping/OTG source select/
Pixel clock.

This is part 2 of 2 for seamless boot NV10

Signed-off-by: Martin Leung <martin.leung@amd.com>
Reviewed-by: Jun Lei <Jun.Lei@amd.com>
Acked-by: Leo Li <sunpeng.li@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amd/display: fix dcn-specific clk_mgr init_clocks
Martin Leung [Sun, 28 Jul 2019 15:09:11 +0000 (11:09 -0400)]
drm/amd/display: fix dcn-specific clk_mgr init_clocks

[Why]
underflow seen on certain monitor setups caused by making dcnxx_init_hw
generic

[How]
by moving dcn20_init_hw into dcn10, we added a dcn-specific clk_mgr
init (dc->clk_mgr->funcs->init_clocks()). Thus, put old clk_mgr
memset in an else statement so both memsets don't get set

Signed-off-by: Martin Leung <martin.leung@amd.com>
Reviewed-by: Anthony Koo <Anthony.Koo@amd.com>
Acked-by: Leo Li <sunpeng.li@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amd/display: Make init_hw and init_pipes generic for seamless boot
Martin Leung [Tue, 9 Jul 2019 19:15:17 +0000 (15:15 -0400)]
drm/amd/display: Make init_hw and init_pipes generic for seamless boot

[Why]
For seamless boot the init_hw sequence must be split into
actual hardware vs pipes, in order to defer pipe initialization to set mode
and skip of pipe-destructive sequences

[How]
made dcn10_init_hw and dcn10_init_pipes generic for future dcns to inherit
deleted dcn20 specific versions. This is part 1 of a 2 partimplementation
of seamless boot

Signed-off-by: Martin Leung <martin.leung@amd.com>
Reviewed-by: Jun Lei <Jun.Lei@amd.com>
Acked-by: Leo Li <sunpeng.li@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amd/display: 3.2.43
Aric Cyr [Mon, 15 Jul 2019 14:34:30 +0000 (10:34 -0400)]
drm/amd/display: 3.2.43

Signed-off-by: Aric Cyr <aric.cyr@amd.com>
Acked-by: Leo Li <sunpeng.li@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amd/display: wait for pending complete when enabling a plane
Jun Lei [Tue, 16 Jul 2019 17:02:41 +0000 (13:02 -0400)]
drm/amd/display: wait for pending complete when enabling a plane

[why]
When planes are enabled, they must be enabled using VSYNC update (not immediate).
However, before the VUPDATE occurs, DM may call with an "immediate" flip which is address
only.  This operation would normally be okay, but if the locking for immediate flip happens
to occur before the VUPDATE associated with the initial plane enablement, it will cause HW
to hang.

[how]
HWSS should enforce plane enable in HW to be synchronous with the call that enables the plane.

Signed-off-by: Jun Lei <Jun.Lei@amd.com>
Reviewed-by: Eric Yang <eric.yang2@amd.com>
Acked-by: Leo Li <sunpeng.li@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amd/display: refactor gpio to allocate hw_container in constructor
Su Sung Chung [Mon, 8 Jul 2019 15:31:39 +0000 (11:31 -0400)]
drm/amd/display: refactor gpio to allocate hw_container in constructor

[why]
if dynamic allocation fails during gpio_open, it will cause crash due to
page fault.

[how]
handle allocation when gpio object gets created and prevent from calling
gpio_open if allocation failed

Signed-off-by: Su Sung Chung <Su.Chung@amd.com>
Reviewed-by: Jun Lei <Jun.Lei@amd.com>
Acked-by: Leo Li <sunpeng.li@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amd/display: Add work-around option to skip DCN20 clock updates
Jaehyun Chung [Mon, 15 Jul 2019 21:35:56 +0000 (17:35 -0400)]
drm/amd/display: Add work-around option to skip DCN20 clock updates

[Why] Auto Overclock Memory fails for some systems that don't support
p-state.

[How] Implement the workaround, and it's corresponding enable flag.

Signed-off-by: Jaehyun Chung <jaehyun.chung@amd.com>
Reviewed-by: Alvin Lee <Alvin.Lee2@amd.com>
Acked-by: Leo Li <sunpeng.li@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amd/display: fixup DPP programming sequence
Jun Lei [Mon, 15 Jul 2019 14:41:47 +0000 (10:41 -0400)]
drm/amd/display: fixup DPP programming sequence

[why]
DC does not correct account for the fact that DPP DTO is double buffered while DPP ref is not.
This means that when DPP ref clock is lowered when it's "safe to lower", the DPP blocks that need
an increased divider will temporarily have actual DPP clock drop below minimum while DTO
double buffering takes effect.  This results in temporary underflow.

[how]
To fix this, DPP clock cannot be programmed atomically, but rather be broken up into the DTO and the
ref.  Each has a separate "safe to lower" logic.  When doing "prepare" the ref and dividers may only increase.
When doing "optimize", both may decrease.  It is guaranteed that we won't exceed max DPP clock because
we do not use dividers larger than 1.

Signed-off-by: Jun Lei <Jun.Lei@amd.com>
Reviewed-by: Eric Yang <eric.yang2@amd.com>
Acked-by: Leo Li <sunpeng.li@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amd/display: make firmware info only load once during dc_bios create
Dmytro Laktyushkin [Fri, 12 Jul 2019 19:06:06 +0000 (15:06 -0400)]
drm/amd/display: make firmware info only load once during dc_bios create

Currently every time DC wants to access firmware info we make a call
into VBIOS. This makes no sense as there is nothing that can change
runtime inside fw info and can cause issues when calling unstable
bios during bringup.

This change eliminate this behavior by only calling bios once for fw
info and keeping it stored as part of dc_bios.

Signed-off-by: Dmytro Laktyushkin <Dmytro.Laktyushkin@amd.com>
Reviewed-by: Chris Park <Chris.Park@amd.com>
Acked-by: Leo Li <sunpeng.li@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amd/display: Synchronous DisplayPort Link Training
David Galiffi [Mon, 24 Jun 2019 14:34:13 +0000 (10:34 -0400)]
drm/amd/display: Synchronous DisplayPort Link Training

[WHY]
We require a method to perform synchronous link training.

[HOW]
Sync LT is broken into 3 basic steps.
"Begin" starts the state machine, and resets "preferred" link settings.
"Attempt" will attempt to train the link with a given set of training
parameters.
"End" stops the state machine, and will optionally disable the link phy.
Between "Begin" and "End" DPCD:600h must not be set to "2"
(D3:Powered Down).
Between "Begin" and "End", there may be multiple "Attempts" with different
training parameters.

Signed-off-by: David Galiffi <david.galiffi@amd.com>
Reviewed-by: Jun Lei <Jun.Lei@amd.com>
Acked-by: Leo Li <sunpeng.li@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amd/display: Improve sharing of HUBBUB register lists
Julian Parkin [Fri, 12 Jul 2019 21:30:06 +0000 (17:30 -0400)]
drm/amd/display: Improve sharing of HUBBUB register lists

Add DCN20 common register list that contains registers shared
between DCN20 generations.

Signed-off-by: Julian Parkin <julian.parkin@amd.com>
Reviewed-by: Charlene Liu <Charlene.Liu@amd.com>
Acked-by: Aric Cyr <Aric.Cyr@amd.com>
Acked-by: Leo Li <sunpeng.li@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amd/display: Add 22, 24, and 26 degamma
Vitaly Prosyak [Wed, 10 Jul 2019 19:20:08 +0000 (14:20 -0500)]
drm/amd/display: Add 22, 24, and 26 degamma

[Why & How]
Support degamma ROM and RAM based on hardware capabilities.
Some refactoring into color module

Signed-off-by: Vitaly Prosyak <vitaly.prosyak@amd.com>
Reviewed-by: Gary Kattan <Gary.Kattan@amd.com>
Reviewed-by: Nevenko Stupar <Nevenko.Stupar@amd.com>
Acked-by: Leo Li <sunpeng.li@amd.com>
Acked-by: Vitaly Prosyak <Vitaly.Prosyak@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amd/display: Fix frames_to_insert math
Bayan Zabihiyan [Wed, 10 Jul 2019 20:00:53 +0000 (16:00 -0400)]
drm/amd/display: Fix frames_to_insert math

[Why]
The math on deciding on how many
"frames to insert" sometimes sent us over the max refresh rate.
Also integer overflow can occur if we have high refresh rates.

[How]
Instead of clipping the  frame duration such that it doesn’t go below the min,
just remove a frame from the number of frames to insert. +
Use unsigned long long for intermediate calculations to prevent
integer overflow.

Signed-off-by: Bayan Zabihiyan <bayan.zabihiyan@amd.com>
Reviewed-by: Aric Cyr <Aric.Cyr@amd.com>
Acked-by: Leo Li <sunpeng.li@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amd/display: fix issue where 252-255 values are clipped
Anthony Koo [Fri, 12 Jul 2019 14:52:54 +0000 (10:52 -0400)]
drm/amd/display: fix issue where 252-255 values are clipped

[Why]
When endpoint is at the boundary of a region, such as at 2^0=1
we find that the last segment has a sharp slope and some points
are clipped at the top.

[How]
If end point is 1, which is exactly at the 2^0 region boundary, we
need to program an additional region beyond this point.

Signed-off-by: Anthony Koo <Anthony.Koo@amd.com>
Reviewed-by: Aric Cyr <Aric.Cyr@amd.com>
Acked-by: Leo Li <sunpeng.li@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amd/display: reset hdmi tmds rate and data scramble on pipe reset
Wenjing Liu [Wed, 10 Jul 2019 22:35:18 +0000 (18:35 -0400)]
drm/amd/display: reset hdmi tmds rate and data scramble on pipe reset

[why]
hdmi data scramble and tmds rate is not reset during pipe reset.

[how]
reset hdmi tmds rate and data scramble on pipe reset

Signed-off-by: Wenjing Liu <Wenjing.Liu@amd.com>
Reviewed-by: Chris Park <Chris.Park@amd.com>
Acked-by: Leo Li <sunpeng.li@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amd/display: reset drr programming on pipe reset
Wenjing Liu [Wed, 10 Jul 2019 22:31:38 +0000 (18:31 -0400)]
drm/amd/display: reset drr programming on pipe reset

[why]
drr is still enabled after driver is unloaded causing black screen

[how]
disable drr during pipe reset.

Signed-off-by: Wenjing Liu <Wenjing.Liu@amd.com>
Reviewed-by: Jun Lei <Jun.Lei@amd.com>
Acked-by: Leo Li <sunpeng.li@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amd/display: Add DFS reference clock field
Yongqiang Sun [Tue, 25 Jun 2019 23:08:50 +0000 (19:08 -0400)]
drm/amd/display: Add DFS reference clock field

Add to clk_mgr_internal struct, for future use.

Signed-off-by: Yongqiang Sun <yongqiang.sun@amd.com>
Reviewed-by: Yongqiang Sun <yongqiang.sun@amd.com>
Acked-by: Leo Li <sunpeng.li@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amd/display: Add PIXEL_RATE control regs for more instances
Yongqiang Sun [Sat, 22 Jun 2019 22:52:41 +0000 (18:52 -0400)]
drm/amd/display: Add PIXEL_RATE control regs for more instances

For use by future ASICs

Signed-off-by: Sung Lee <sung.lee@amd.com>
Signed-off-by: Yongqiang Sun <yongqiang.sun@amd.com>
Reviewed-by: Yongqiang Sun <yongqiang.sun@amd.com>
Acked-by: Leo Li <sunpeng.li@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/scheduler: use job count instead of peek
Christian König [Fri, 9 Aug 2019 15:27:21 +0000 (17:27 +0200)]
drm/scheduler: use job count instead of peek

The spsc_queue_peek function is accessing queue->head which belongs to
the consumer thread and shouldn't be accessed by the producer

This is fixing a rare race condition when destroying entities.

Signed-off-by: Christian König <christian.koenig@amd.com>
Acked-by: Andrey Grodzovsky <andrey.grodzovsky@amd.com>
Reviewed-by: Monk.liu@amd.com
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amdgpu: MODULE_FIRMWARE requires linux/module.h
Stephen Rothwell [Tue, 13 Aug 2019 08:10:38 +0000 (18:10 +1000)]
drm/amdgpu: MODULE_FIRMWARE requires linux/module.h

Fixes: 6a7a0bdbfa0c ("drm/amdgpu: add psp_v12_0 for renoir (v2)")
Reviewed-by: Huang Rui <ray.huang@amd.com>
Signed-off-by: Stephen Rothwell <sfr@canb.auug.org.au>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>