]> git.proxmox.com Git - mirror_ubuntu-eoan-kernel.git/log
mirror_ubuntu-eoan-kernel.git
6 years agodrm/amdgpu: rename amdgpu_get_pcie_info
Alex Deucher [Fri, 15 Dec 2017 21:49:33 +0000 (16:49 -0500)]
drm/amdgpu: rename amdgpu_get_pcie_info

add device to the name for consistency.

Acked-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
6 years agodrm/amdgpu: move amdgpu_need_backup to amdgpu_object.c
Alex Deucher [Fri, 15 Dec 2017 21:45:02 +0000 (16:45 -0500)]
drm/amdgpu: move amdgpu_need_backup to amdgpu_object.c

It's the only place it's used.

Acked-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
6 years agodrm/amdgpu: rename amdgpu_gpu_recover
Alex Deucher [Fri, 15 Dec 2017 21:40:49 +0000 (16:40 -0500)]
drm/amdgpu: rename amdgpu_gpu_recover

add device to the name for consistency.

Acked-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
6 years agodrm/amdgpu: move dummy page functions to amdgpu_gart.c
Alex Deucher [Fri, 15 Dec 2017 21:28:33 +0000 (16:28 -0500)]
drm/amdgpu: move dummy page functions to amdgpu_gart.c

It's the only place they are used.

Acked-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
6 years agodrm/amdgpu: rename amdgpu_need_post
Alex Deucher [Fri, 15 Dec 2017 21:22:11 +0000 (16:22 -0500)]
drm/amdgpu: rename amdgpu_need_post

add device to the name for consistency.

Acked-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
6 years agodrm/amdgpu: rename ip block helper functions
Alex Deucher [Fri, 15 Dec 2017 21:18:00 +0000 (16:18 -0500)]
drm/amdgpu: rename ip block helper functions

add device to the name for consistency.

Acked-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
6 years agodrm/amdgpu: move fw_reserve functions to amdgpu_ttm.c
Alex Deucher [Thu, 14 Dec 2017 21:39:02 +0000 (16:39 -0500)]
drm/amdgpu: move fw_reserve functions to amdgpu_ttm.c

It's the only place they are used.

Acked-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
6 years agodrm/amdgpu: rename amdgpu_*_location functions
Alex Deucher [Thu, 14 Dec 2017 21:33:36 +0000 (16:33 -0500)]
drm/amdgpu: rename amdgpu_*_location functions

add device to the name for consistency.

Acked-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
6 years agodrm/amdgpu: move amdgpu_doorbell_get_kfd_info to amdgpu_amdkfd.c
Alex Deucher [Thu, 14 Dec 2017 21:27:11 +0000 (16:27 -0500)]
drm/amdgpu: move amdgpu_doorbell_get_kfd_info to amdgpu_amdkfd.c

It's the only place it's used.

Acked-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
6 years agodrm/amdgpu: rename amdgpu_pci_config_reset
Alex Deucher [Thu, 14 Dec 2017 21:22:53 +0000 (16:22 -0500)]
drm/amdgpu: rename amdgpu_pci_config_reset

add device for consistency with other functions in this file.

Acked-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
6 years agodrm/amdgpu: rename amdgpu_program_register_sequence
Alex Deucher [Thu, 14 Dec 2017 21:20:19 +0000 (16:20 -0500)]
drm/amdgpu: rename amdgpu_program_register_sequence

add device for consistency with other functions in this file.

Acked-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
6 years agodrm/amdgpu: rename amdgpu_wb_* functions
Alex Deucher [Thu, 14 Dec 2017 21:03:43 +0000 (16:03 -0500)]
drm/amdgpu: rename amdgpu_wb_* functions

add device for consistency.

Acked-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
6 years agodrm/amdgpu: move debugfs functions to their own file
Alex Deucher [Thu, 14 Dec 2017 20:23:14 +0000 (15:23 -0500)]
drm/amdgpu: move debugfs functions to their own file

amdgpu_device.c was getting pretty cluttered.

Acked-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
6 years agodrm/amdgpu: rename amdgpu_suspend to amdgpu_device_ip_suspend
Alex Deucher [Thu, 14 Dec 2017 21:47:40 +0000 (16:47 -0500)]
drm/amdgpu: rename amdgpu_suspend to amdgpu_device_ip_suspend

for consistency with the other functions in that file.

Acked-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
6 years agodrm/amdgpu: use consistent naming for static funcs in amdgpu_device.c
Alex Deucher [Thu, 14 Dec 2017 20:02:39 +0000 (15:02 -0500)]
drm/amdgpu: use consistent naming for static funcs in amdgpu_device.c

Prefix the functions with device or device_ip for functions which
deal with ip blocks for consistency.

Acked-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
6 years agodrm/amdgpu: move atom functions from amdgpu_device.c
Alex Deucher [Thu, 14 Dec 2017 19:32:53 +0000 (14:32 -0500)]
drm/amdgpu: move atom functions from amdgpu_device.c

and move them to amdgpu_atombios.c for consistency.

Acked-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
6 years agodrm/amd/display: remove redundant null check of array 'data'
Colin Ian King [Fri, 15 Dec 2017 10:53:43 +0000 (10:53 +0000)]
drm/amd/display: remove redundant null check of array 'data'

The null check on aconnector->base.edid_blob_ptr->data is redundant
since data is an array and can never be null.  Remove it.

Detected by CoverityScan, CID#1460369 ("Array compared against 0")

Acked-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Harry Wentland <harry.wentland@amd.com>
Signed-off-by: Colin Ian King <colin.king@canonical.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
6 years agodrm/amdgpu: setup the shared and private apertures on gfx9
Alex Deucher [Fri, 8 Dec 2017 20:09:20 +0000 (15:09 -0500)]
drm/amdgpu: setup the shared and private apertures on gfx9

Same as previous asics.  This was not yet set for gfx9.

Acked-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
6 years agodrm/amdgpu: Simplify amdgpu_lockup_timeout usage.
Andrey Grodzovsky [Wed, 13 Dec 2017 19:36:53 +0000 (14:36 -0500)]
drm/amdgpu: Simplify amdgpu_lockup_timeout usage.

With introduction of amdgpu_gpu_recovery we don't need any more
to rely on amdgpu_lockup_timeout == 0 for disabling GPU reset.

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>
6 years agodrm/amdgpu: Add gpu_recovery parameter
Andrey Grodzovsky [Tue, 12 Dec 2017 19:09:30 +0000 (14:09 -0500)]
drm/amdgpu: Add gpu_recovery parameter

Add new parameter to control GPU recovery procedure.

v2:
Add auto logic where reset is disabled for bare metal and enabled
for SR-IOV.
Allow forced reset from debugfs.

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>
6 years agodrm/ttm: use an ttm operation ctx for ttm_bo_move_xxx
Roger He [Fri, 8 Dec 2017 12:19:32 +0000 (20:19 +0800)]
drm/ttm: use an ttm operation ctx for ttm_bo_move_xxx

include ttm_bo_move_memcpy and ttm_bo_move_ttm

Signed-off-by: Roger He <Hongbo.He@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
6 years agodrm/ttm: enable eviction for Per-VM-BO
Roger He [Wed, 6 Dec 2017 06:16:09 +0000 (14:16 +0800)]
drm/ttm: enable eviction for Per-VM-BO

allow eviction of BOs reserved by the caller when they are
not part of the current working set.

Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Roger He <Hongbo.He@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
6 years agodrm/amdgpu: set allow_reserved_eviction and resv when bo allocation and cs
Roger He [Fri, 8 Dec 2017 05:31:52 +0000 (13:31 +0800)]
drm/amdgpu: set allow_reserved_eviction and resv when bo allocation and cs

enable eviction of other per VM BOs during allocation and allows
reaping of deleted BOs during CS.

Reviewed-by: Chunming Zhou <david1.zhou@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Roger He <Hongbo.He@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
6 years agodrm/ttm: add allow_reserved_eviction and resv into ttm_operation_ctx
Roger He [Fri, 8 Dec 2017 03:36:46 +0000 (11:36 +0800)]
drm/ttm: add allow_reserved_eviction and resv into ttm_operation_ctx

allow_reserved_eviction: Allow eviction of reserved BOs
resv: Reservation object to allow reserved evictions with

Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Roger He <Hongbo.He@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
6 years agodrm/amdgpu: add enumerate for PDB/PTB v3
Chunming Zhou [Wed, 13 Dec 2017 06:22:54 +0000 (14:22 +0800)]
drm/amdgpu: add enumerate for PDB/PTB v3

v2:
  remove SUBPTB member
v3:
  remove last_level, use AMDGPU_VM_PTB directly instead.

Signed-off-by: Chunming Zhou <david1.zhou@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
6 years agodrm/amd/display: dal 3.1.25
Tony Cheng [Tue, 28 Nov 2017 15:10:48 +0000 (10:10 -0500)]
drm/amd/display: dal 3.1.25

Signed-off-by: Tony Cheng <tony.cheng@amd.com>
Reviewed-by: Tony Cheng <Tony.Cheng@amd.com>
Acked-by: Harry Wentland <harry.wentland@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
6 years agodrm/amd/display: Add TODO item to remove vector.c
Harry Wentland [Thu, 23 Nov 2017 15:37:29 +0000 (10:37 -0500)]
drm/amd/display: Add TODO item to remove vector.c

Signed-off-by: Harry Wentland <harry.wentland@amd.com>
Reviewed-by: Jordan Lazare <Jordan.Lazare@amd.com>
Acked-by: Harry Wentland <harry.wentland@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
6 years agodrm/amd/display: Remove grph_object_id.c and move function to bios_parser
Harry Wentland [Thu, 23 Nov 2017 15:31:50 +0000 (10:31 -0500)]
drm/amd/display: Remove grph_object_id.c and move function to bios_parser

Signed-off-by: Harry Wentland <harry.wentland@amd.com>
Reviewed-by: Jordan Lazare <Jordan.Lazare@amd.com>
Acked-by: Harry Wentland <harry.wentland@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
6 years agodrm/amd/display: Remove redundant NULL check in DCE11 HWSS
Harry Wentland [Thu, 23 Nov 2017 02:05:55 +0000 (21:05 -0500)]
drm/amd/display: Remove redundant NULL check in DCE11 HWSS

We already check this a couple lines earlier.

Signed-off-by: Harry Wentland <harry.wentland@amd.com>
Reviewed-by: Jordan Lazare <Jordan.Lazare@amd.com>
Acked-by: Harry Wentland <harry.wentland@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
6 years agodrm/amd/display: use clamping rather than truncation for CM fp conversions
Dmytro Laktyushkin [Tue, 28 Nov 2017 21:20:45 +0000 (16:20 -0500)]
drm/amd/display: use clamping rather than truncation for CM fp conversions

Signed-off-by: Dmytro Laktyushkin <Dmytro.Laktyushkin@amd.com>
Reviewed-by: Tony Cheng <Tony.Cheng@amd.com>
Acked-by: Harry Wentland <harry.wentland@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
6 years agodrm/amd/display: remove format_control from set_cursor_attributes
Yue Hin Lau [Tue, 28 Nov 2017 17:05:08 +0000 (12:05 -0500)]
drm/amd/display: remove format_control from set_cursor_attributes

Signed-off-by: Yue Hin Lau <Yuehin.Lau@amd.com>
Reviewed-by: Eric Bernstein <Eric.Bernstein@amd.com>
Acked-by: Harry Wentland <harry.wentland@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
6 years agodrm/amd/display: Move wait for hpd ready out from edp power control.
Yongqiang Sun [Fri, 24 Nov 2017 21:31:03 +0000 (16:31 -0500)]
drm/amd/display: Move wait for hpd ready out from edp power control.

It may take over 200ms for wait hpd ready. To optimize the resume time,
we can power on eDP in init_hw, wait for hpd ready when doing link
training.

also create separate eDP enable function to make sure eDP is powered up
before doing and DPCD access, as HPD low will result in DPDC transaction
failure.

After optimization,
setpowerstate 145ms -> 9.8ms,
DPMS 387ms -> 18.9ms

Signed-off-by: Yongqiang Sun <yongqiang.sun@amd.com>
Signed-off-by: Tony Cheng <tony.cheng@amd.com>
Reviewed-by: Tony Cheng <Tony.Cheng@amd.com>
Acked-by: Harry Wentland <harry.wentland@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
6 years agodrm/amd/display: fix missing pixel clock adjustment for dongle
Eric Yang [Fri, 10 Nov 2017 15:44:24 +0000 (10:44 -0500)]
drm/amd/display: fix missing pixel clock adjustment for dongle

Signed-off-by: Eric Yang <Eric.Yang2@amd.com>
Reviewed-by: Tony Cheng <Tony.Cheng@amd.com>
Reviewed-by: Andrew Jiang <Andrew.Jiang@amd.com>
Acked-by: Harry Wentland <harry.wentland@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
6 years agodrm/amd/display: Move OPP mpc tree initialization to hw_init
Eric Bernstein [Fri, 24 Nov 2017 22:51:34 +0000 (17:51 -0500)]
drm/amd/display: Move OPP mpc tree initialization to hw_init

Move OPP initialization of mpc tree parameters to hw_init function.

Signed-off-by: Eric Bernstein <eric.bernstein@amd.com>
Reviewed-by: Tony Cheng <Tony.Cheng@amd.com>
Acked-by: Harry Wentland <harry.wentland@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
6 years agodrm/amd/display: use REG_UPDATE for MPC mux
Eric Bernstein [Mon, 27 Nov 2017 15:55:52 +0000 (10:55 -0500)]
drm/amd/display: use REG_UPDATE for MPC mux

Use REG_UPDATE instead of REG_SET for programming MPC out mux.

Signed-off-by: Eric Bernstein <eric.bernstein@amd.com>
Reviewed-by: Tony Cheng <Tony.Cheng@amd.com>
Acked-by: Harry Wentland <harry.wentland@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
6 years agodrm/amd/display: Correct fixed point calculation.
Vitaly Prosyak [Fri, 24 Nov 2017 21:52:33 +0000 (15:52 -0600)]
drm/amd/display: Correct fixed point calculation.

When convert from fixed31_32 to other fixed point
format use math operation round instead of floor.

Signed-off-by: Vitaly Prosyak <vitaly.prosyak@amd.com>
Reviewed-by: Roman Li <Roman.Li@amd.com>
Reviewed-by: Krunoslav Kovac <Krunoslav.Kovac@amd.com>
Acked-by: Harry Wentland <harry.wentland@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
6 years agodrm/amd/display: OPTC cleanup/implementation
Yue Hin Lau [Wed, 22 Nov 2017 21:48:35 +0000 (16:48 -0500)]
drm/amd/display: OPTC cleanup/implementation

Signed-off-by: Yue Hin Lau <Yuehin.Lau@amd.com>
Reviewed-by: Tony Cheng <Tony.Cheng@amd.com>
Acked-by: Harry Wentland <harry.wentland@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
6 years agodrm/amd/display: Move unity TF type to predefined types
Vitaly Prosyak [Thu, 23 Nov 2017 15:42:22 +0000 (09:42 -0600)]
drm/amd/display: Move unity TF type to predefined types

Also handle fixpoint y values for CM curves

Signed-off-by: Vitaly Prosyak <vitaly.prosyak@amd.com>
Reviewed-by: Tony Cheng <Tony.Cheng@amd.com>
Acked-by: Harry Wentland <harry.wentland@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
6 years agodrm/amd/display: dpp clean up
Yue Hin Lau [Thu, 23 Nov 2017 16:32:18 +0000 (11:32 -0500)]
drm/amd/display: dpp clean up

Signed-off-by: Yue Hin Lau <Yuehin.Lau@amd.com>
Reviewed-by: Tony Cheng <Tony.Cheng@amd.com>
Acked-by: Harry Wentland <harry.wentland@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
6 years agodrm/amd/display: add assert to verify dcn_calc input validity
Dmytro Laktyushkin [Thu, 23 Nov 2017 17:08:13 +0000 (12:08 -0500)]
drm/amd/display: add assert to verify dcn_calc input validity

This reverts commit 978482d0de86 Revert noisy assert messages

Signed-off-by: Dmytro Laktyushkin <Dmytro.Laktyushkin@amd.com>
Reviewed-by: Tony Cheng <Tony.Cheng@amd.com>
Acked-by: Harry Wentland <harry.wentland@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
6 years agodrm/amd/display: Add dppclk to dcn_bw_clocks
Dmytro Laktyushkin [Tue, 14 Nov 2017 16:52:11 +0000 (11:52 -0500)]
drm/amd/display: Add dppclk to dcn_bw_clocks

Signed-off-by: Dmytro Laktyushkin <Dmytro.Laktyushkin@amd.com>
Reviewed-by: Tony Cheng <Tony.Cheng@amd.com>
Acked-by: Harry Wentland <harry.wentland@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
6 years agodrm/amd/display: set chroma taps to 1 when not scaling
Dmytro Laktyushkin [Mon, 13 Nov 2017 22:03:53 +0000 (17:03 -0500)]
drm/amd/display: set chroma taps to 1 when not scaling

Signed-off-by: Dmytro Laktyushkin <Dmytro.Laktyushkin@amd.com>
Reviewed-by: Tony Cheng <Tony.Cheng@amd.com>
Acked-by: Harry Wentland <harry.wentland@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
6 years agodrm/amd/display: Print DC_VER at DC init
Harry Wentland [Tue, 21 Nov 2017 18:19:32 +0000 (13:19 -0500)]
drm/amd/display: Print DC_VER at DC init

This has proven helpful on other OSes to give a quick state of the
DC driver when a bug is reported.

Signed-off-by: Harry Wentland <harry.wentland@amd.com>
Reviewed-by: Tony Cheng <Tony.Cheng@amd.com>
Acked-by: Harry Wentland <harry.wentland@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
6 years agodrm/amd/display: really fix time out in init sequence
Tony Cheng [Wed, 22 Nov 2017 16:51:30 +0000 (11:51 -0500)]
drm/amd/display: really fix time out in init sequence

REG_UPDATE_2 return the reg value it write out through MMIO
we need to do a REG_READ to confirm the value is written out

Signed-off-by: Tony Cheng <tony.cheng@amd.com>
Reviewed-by: Tony Cheng <Tony.Cheng@amd.com>
Acked-by: Harry Wentland <harry.wentland@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
6 years agodrm/amd/display: Add disclaimer to BW and DML code provided by HW
Harry Wentland [Wed, 29 Nov 2017 15:28:29 +0000 (10:28 -0500)]
drm/amd/display: Add disclaimer to BW and DML code provided by HW

This code can sometimes look troubling but we trust it as it comes from
HW teams with a guarantee of correctness. Add a note to these files to
explain this.

v2: thing -> things

Signed-off-by: Harry Wentland <harry.wentland@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
6 years agodrm/amd/display: Use macro for isnan check
Harry Wentland [Wed, 29 Nov 2017 15:16:51 +0000 (10:16 -0500)]
drm/amd/display: Use macro for isnan check

In code provided by HW teams we do a NaN check on floats
by comparing the number against itself. This confuses most
people including myself. Macro it out to make it self-explanatory.

Don't do a NaN check for int.

v2: parantheses around 'number' expression

Signed-off-by: Harry Wentland <harry.wentland@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
6 years agodrm/amd/display: dal 3.1.24
Tony Cheng [Wed, 22 Nov 2017 02:40:34 +0000 (21:40 -0500)]
drm/amd/display: dal 3.1.24

Signed-off-by: Tony Cheng <tony.cheng@amd.com>
Reviewed-by: Tony Cheng <Tony.Cheng@amd.com>
Acked-by: Harry Wentland <harry.wentland@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
6 years agodrm/amd/display: Set mpcc_disconnect_pending during MPC reset
Tony Cheng [Tue, 21 Nov 2017 22:51:50 +0000 (17:51 -0500)]
drm/amd/display: Set mpcc_disconnect_pending during MPC reset

This prevents an issue where the MPCC will not go to idle due to us not
waiting for it to become idle during disable plane.

Signed-off-by: Andrew Jiang <Andrew.Jiang@amd.com>
Signed-off-by: Tony Cheng <tony.cheng@amd.com>
Reviewed-by: Tony Cheng <Tony.Cheng@amd.com>
Acked-by: Harry Wentland <harry.wentland@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
6 years agodrm/amd/display: Refine update flags usage in update_dchubp_dpp
Andrew Jiang [Tue, 21 Nov 2017 20:59:42 +0000 (15:59 -0500)]
drm/amd/display: Refine update flags usage in update_dchubp_dpp

- Only update DPP clock if it's a full update.
- Program viewport on position change. This caused SLS regressions.

Signed-off-by: Andrew Jiang <Andrew.Jiang@amd.com>
Reviewed-by: Tony Cheng <Tony.Cheng@amd.com>
Acked-by: Harry Wentland <harry.wentland@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
6 years agodrm/amd/display: Cache cursor position.
Yongqiang Sun [Tue, 21 Nov 2017 21:12:23 +0000 (16:12 -0500)]
drm/amd/display: Cache cursor position.

When programming cursor position after front end programmed,
if position is already set previously, it doesn't make sense
to program position with all 0.
Cache position and use the parameter after front end programming.

Signed-off-by: Yongqiang Sun <yongqiang.sun@amd.com>
Reviewed-by: Tony Cheng <Tony.Cheng@amd.com>
Acked-by: Harry Wentland <harry.wentland@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
6 years agodrm/amd/display: Program cursor regs after context swapped.
Yongqiang Sun [Tue, 21 Nov 2017 20:42:17 +0000 (15:42 -0500)]
drm/amd/display: Program cursor regs after context swapped.

Cursor is abnormal after pipe_ctx is changed in context.
Cause: cursor attributes is programmed right after front end
programming, but it use old pipe_ctx to program which is not
updated yet.
Solution:
Program cursor regs after context swapped.

Signed-off-by: Yongqiang Sun <yongqiang.sun@amd.com>
Reviewed-by: Tony Cheng <Tony.Cheng@amd.com>
Acked-by: Harry Wentland <harry.wentland@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
6 years agodrm/amd/display: fix recout_skip calculation when rotating 180 or 270
Yongqiang Sun [Tue, 21 Nov 2017 18:45:51 +0000 (13:45 -0500)]
drm/amd/display: fix recout_skip calculation when rotating 180 or 270

Fixed fliped landscape and fliped portrait hard hang.

Signed-off-by: Yongqiang Sun <yongqiang.sun@amd.com>
Reviewed-by: Tony Cheng <Tony.Cheng@amd.com>
Acked-by: Harry Wentland <harry.wentland@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
6 years agodrm/amd/display: Add is_tiling_rotated flag to plane_state
Eric Murphy-Zaremba [Fri, 17 Nov 2017 21:29:00 +0000 (16:29 -0500)]
drm/amd/display: Add is_tiling_rotated flag to plane_state

Signed-off-by: Eric Murphy-Zaremba <Eric.Murphy-zaremba@amd.com>
Reviewed-by: Aric Cyr <Aric.Cyr@amd.com>
Acked-by: Harry Wentland <harry.wentland@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
6 years agodrm/amd/display: OPP DPG test pattern
Eric Bernstein [Fri, 17 Nov 2017 22:21:26 +0000 (17:21 -0500)]
drm/amd/display: OPP DPG test pattern

Create opp_set_test_pattern function with similar interface
and implementation as timing generator test pattern.

Signed-off-by: Eric Bernstein <eric.bernstein@amd.com>
Reviewed-by: Tony Cheng <Tony.Cheng@amd.com>
Acked-by: Harry Wentland <harry.wentland@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
6 years agodrm/amd/display: Build unity lut for shaper
Vitaly Prosyak [Tue, 14 Nov 2017 23:12:52 +0000 (17:12 -0600)]
drm/amd/display: Build unity lut for shaper

Add color module to diagnostic compilation

Signed-off-by: Vitaly Prosyak <vitaly.prosyak@amd.com>
Reviewed-by: Charlene Liu <Charlene.Liu@amd.com>
Acked-by: Harry Wentland <harry.wentland@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
6 years agodrm/amd/display: Reset MPCC muxes during init
Eric Bernstein [Thu, 16 Nov 2017 20:34:50 +0000 (15:34 -0500)]
drm/amd/display: Reset MPCC muxes during init

During HW initialization, instead of assuming or
detecting the existing MPCC mux configuration and
then removing existing planes, reset all the MPCC
muxes.

Signed-off-by: Eric Bernstein <eric.bernstein@amd.com>
Reviewed-by: Tony Cheng <Tony.Cheng@amd.com>
Acked-by: Harry Wentland <harry.wentland@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
6 years agodrm/amd/display: CNVC pseudocode review follow up
Yue Hin Lau [Tue, 14 Nov 2017 17:13:27 +0000 (12:13 -0500)]
drm/amd/display: CNVC pseudocode review follow up

Signed-off-by: Yue Hin Lau <Yuehin.Lau@amd.com>
Reviewed-by: Eric Bernstein <Eric.Bernstein@amd.com>
Acked-by: Harry Wentland <harry.wentland@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
6 years agodrm/amd/display: Disable plane right after disconnected
Yongqiang Sun [Fri, 17 Nov 2017 15:44:15 +0000 (10:44 -0500)]
drm/amd/display: Disable plane right after disconnected

HDR display playing video underflow is observed when switching
to full screen due to program a lower watermark right after unlock otg.

Instead of disable plane in next flip coming, if there is a
plane disconnected, after otg unlock wait for mpcc idle and disable
the plane, then program watermark. So there is enough warter mark to make
sure current frame data pass through.

Signed-off-by: Yongqiang Sun <yongqiang.sun@amd.com>
Reviewed-by: Tony Cheng <Tony.Cheng@amd.com>
Acked-by: Harry Wentland <harry.wentland@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
6 years agodrm/amd/display: dal 3.1.23
Tony Cheng [Fri, 17 Nov 2017 04:37:10 +0000 (23:37 -0500)]
drm/amd/display: dal 3.1.23

Signed-off-by: Tony Cheng <tony.cheng@amd.com>
Reviewed-by: Tony Cheng <Tony.Cheng@amd.com>
Acked-by: Harry Wentland <harry.wentland@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
6 years agodrm/amd/display: Clean up os_types.h a bit
Harry Wentland [Thu, 16 Nov 2017 00:27:37 +0000 (19:27 -0500)]
drm/amd/display: Clean up os_types.h a bit

Signed-off-by: Harry Wentland <harry.wentland@amd.com>
Reviewed-by: Tony Cheng <Tony.Cheng@amd.com>
Acked-by: Harry Wentland <harry.wentland@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
6 years agodrm/amd/display: Fix access of wrong array element TF format conversion
Harry Wentland [Fri, 10 Nov 2017 17:12:40 +0000 (12:12 -0500)]
drm/amd/display: Fix access of wrong array element TF format conversion

Found by smatch:
drivers/gpu/drm/amd/amdgpu/../display/dc/dce110/dce110_hw_sequencer.c:357
convert_to_custom_float() error: buffer overflow 'arr_points' 2 <= 2
drivers/gpu/drm/amd/amdgpu/../display/dc/dce110/dce110_hw_sequencer.c:358
convert_to_custom_float() warn: buffer overflow 'arr_points' 2 <= 2

Regression:
drm/amd/display: Remove extra arr_points element

Signed-off-by: Harry Wentland <harry.wentland@amd.com>
Reviewed-by: Tony Cheng <Tony.Cheng@amd.com>
Acked-by: Harry Wentland <harry.wentland@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
6 years agodrm/amd/display: dal 3.1.22
Tony Cheng [Thu, 16 Nov 2017 23:03:52 +0000 (18:03 -0500)]
drm/amd/display: dal 3.1.22

Signed-off-by: Tony Cheng <tony.cheng@amd.com>
Reviewed-by: Tony Cheng <Tony.Cheng@amd.com>
Acked-by: Harry Wentland <harry.wentland@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
6 years agodrm/amd/display: Add dcc_change surface update flag
Andrew Jiang [Thu, 16 Nov 2017 22:08:44 +0000 (17:08 -0500)]
drm/amd/display: Add dcc_change surface update flag

Program the DCC registers when dcc_change is true.

Signed-off-by: Andrew Jiang <Andrew.Jiang@amd.com>
Reviewed-by: Tony Cheng <Tony.Cheng@amd.com>
Acked-by: Harry Wentland <harry.wentland@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
6 years agodrm/amd/display: Implement work around for optc underflow.
Yongqiang Sun [Wed, 15 Nov 2017 21:21:34 +0000 (16:21 -0500)]
drm/amd/display: Implement work around for optc underflow.

Work around for a hw bug causing optc underflow if blank data
double buffer disable and remove mpcc.
Checking optc status after otg unlock, after wait mpcc idle
check status again, if optc underflow just happens after wait
mpcc idle, clear underflow status and enable blank data double
buffer.

Signed-off-by: Yongqiang Sun <yongqiang.sun@amd.com>
Reviewed-by: Tony Cheng <Tony.Cheng@amd.com>
Acked-by: Harry Wentland <harry.wentland@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
6 years agodrm/amd/display: Change optimized_required logic
Andrew Jiang [Thu, 16 Nov 2017 20:51:21 +0000 (15:51 -0500)]
drm/amd/display: Change optimized_required logic

Rather than setting it every time there's a full update with surface
count > 0, set it when we need to do plane_atomic_disconnect. Also make
sure that we unset the flag in plane_atomic_disable, so that in the
event we run through a sequence where we do disconnect followed by an
immediate disable, we do not do unnecessarily request a passive flip to
do the optimization.

Signed-off-by: Andrew Jiang <Andrew.Jiang@amd.com>
Reviewed-by: Tony Cheng <Tony.Cheng@amd.com>
Acked-by: Harry Wentland <harry.wentland@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
6 years agodrm/amd/display: Add optimized_required flag
Harry Wentland [Wed, 22 Nov 2017 20:59:39 +0000 (15:59 -0500)]
drm/amd/display: Add optimized_required flag

Signed-off-by: Harry Wentland <harry.wentland@amd.com>
Reviewed-by: Leo (Sunpeng) Li <sunpeng.li@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
6 years agodrm/amd/display: Use same wait mpcc idle function.
Yongqiang Sun [Wed, 15 Nov 2017 21:12:19 +0000 (16:12 -0500)]
drm/amd/display: Use same wait mpcc idle function.

There is already wait mpcc idle function. It is better
to use the same function for all wait mpcc idle.

Signed-off-by: Yongqiang Sun <yongqiang.sun@amd.com>
Reviewed-by: Tony Cheng <Tony.Cheng@amd.com>
Acked-by: Harry Wentland <harry.wentland@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
6 years agodrm/amd/display: MPC updates
Eric Bernstein [Thu, 16 Nov 2017 19:29:10 +0000 (14:29 -0500)]
drm/amd/display: MPC updates

Fix update_mpcc logic to only call assert_mpcc_idle_before_connect
if mpcc is not already being used (and required removal).
Update set_out_rate_control to include optional flow control parameter.
In init_mpcc_from_hw check for case where bot_sel is same as mpcc_id.

Signed-off-by: Eric Bernstein <eric.bernstein@amd.com>
Reviewed-by: Tony Cheng <Tony.Cheng@amd.com>
Acked-by: Harry Wentland <harry.wentland@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
6 years agodrm/amd/display: Fixed read wrong reg to get bot_sel.
Yongqiang Sun [Thu, 16 Nov 2017 17:43:59 +0000 (12:43 -0500)]
drm/amd/display: Fixed read wrong reg to get bot_sel.

Signed-off-by: Yongqiang Sun <yongqiang.sun@amd.com>
Reviewed-by: Tony Cheng <Tony.Cheng@amd.com>
Acked-by: Harry Wentland <harry.wentland@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
6 years agodrm/amd/display: Fix potential mem leak in DC construct
Harry Wentland [Sat, 11 Nov 2017 01:17:03 +0000 (20:17 -0500)]
drm/amd/display: Fix potential mem leak in DC construct

Found by smatch:
drivers/gpu/drm/amd/amdgpu/../display/dc/core/dc.c:506 construct() warn:
possible memory leak of 'dc_ctx'
drivers/gpu/drm/amd/amdgpu/../display/dc/core/dc.c:506 construct() warn:
possible memory leak of 'dc_vbios'
drivers/gpu/drm/amd/amdgpu/../display/dc/core/dc.c:506 construct() warn:
possible memory leak of 'dcn_ip'
drivers/gpu/drm/amd/amdgpu/../display/dc/core/dc.c:506 construct() warn:
possible memory leak of 'dcn_soc'

Signed-off-by: Harry Wentland <harry.wentland@amd.com>
Reviewed-by: Tony Cheng <Tony.Cheng@amd.com>
Acked-by: Harry Wentland <harry.wentland@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
6 years agodrm/amd/display: Remove redundant checks in set_default_colors
Harry Wentland [Fri, 10 Nov 2017 17:08:13 +0000 (12:08 -0500)]
drm/amd/display: Remove redundant checks in set_default_colors

pipe_ctx->stream and pipe_ctx->plane_state are never NULL

Found by smatch:
drivers/gpu/drm/amd/amdgpu/../display/dc/dce110/dce110_hw_sequencer.c:2111
set_default_colors() error: we previously assumed 'pipe_ctx->stream'
could be null (see line 2101)

Signed-off-by: Harry Wentland <harry.wentland@amd.com>
Reviewed-by: Tony Cheng <Tony.Cheng@amd.com>
Acked-by: Harry Wentland <harry.wentland@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
6 years agodrm/amd/display: Remove plane_res.mi check in dce110_apply_ctx_for_surface
Harry Wentland [Fri, 10 Nov 2017 17:00:41 +0000 (12:00 -0500)]
drm/amd/display: Remove plane_res.mi check in dce110_apply_ctx_for_surface

plane_res.mi (memory interface) can never be NULL for DCE110

Found by smatch:
drivers/gpu/drm/amd/amdgpu/../display/dc/dce110/dce110_hw_sequencer.c:2881
dce110_apply_ctx_for_surface() error: we previously assumed
'pipe_ctx->plane_res.mi' could be null (see line 2873)

Signed-off-by: Harry Wentland <harry.wentland@amd.com>
Reviewed-by: Tony Cheng <Tony.Cheng@amd.com>
Acked-by: Harry Wentland <harry.wentland@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
6 years agodrm/amd/display: Remove PSR functions in Linux
Harry Wentland [Sat, 11 Nov 2017 01:01:38 +0000 (20:01 -0500)]
drm/amd/display: Remove PSR functions in Linux

NULL check issue found by smatch:
drivers/gpu/drm/amd/amdgpu/../display/dc/core/dc_link.c:1976
dc_link_setup_psr() warn: variable dereferenced before check 'link' (see
line 1970)

We don't use these functions so might as well remove them.

Signed-off-by: Harry Wentland <harry.wentland@amd.com>
Reviewed-by: Tony Cheng <Tony.Cheng@amd.com>
Acked-by: Harry Wentland <harry.wentland@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
6 years agodrm/amd/display: Integrating MPC pseudocode
Eric Bernstein [Mon, 6 Nov 2017 21:38:55 +0000 (16:38 -0500)]
drm/amd/display: Integrating MPC pseudocode

Integrating MPC pseudocode to support new blending cases
with secondary MPCC list.
This includes a design change to MPC data structures and
interfaces.

Signed-off-by: Eric Bernstein <eric.bernstein@amd.com>
Reviewed-by: Tony Cheng <Tony.Cheng@amd.com>
Acked-by: Harry Wentland <harry.wentland@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
6 years agodrm/amd/display: dal 3.1.21
Tony Cheng [Thu, 16 Nov 2017 04:58:20 +0000 (23:58 -0500)]
drm/amd/display: dal 3.1.21

Signed-off-by: Tony Cheng <tony.cheng@amd.com>
Reviewed-by: Tony Cheng <Tony.Cheng@amd.com>
Acked-by: Harry Wentland <harry.wentland@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
6 years agodrm/amd/display: try to find matching audio inst for enc inst first
Charlene Liu [Wed, 15 Nov 2017 23:55:57 +0000 (18:55 -0500)]
drm/amd/display: try to find matching audio inst for enc inst first

[Description]
in eDP+ HDMI/DP clone or extended configuration, audio inst changed from inst 1 to inst0.
No failure related this though, just playback device endpoint inst changed.
Also  remove one addition register read.

Signed-off-by: Charlene Liu <charlene.liu@amd.com>
Reviewed-by: Tony Cheng <Tony.Cheng@amd.com>
Acked-by: Harry Wentland <harry.wentland@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
6 years agodrm/amd/display: fix seq issue: turn on clock before programming afmt.
Charlene Liu [Wed, 15 Nov 2017 23:27:31 +0000 (18:27 -0500)]
drm/amd/display: fix seq issue: turn on clock before programming afmt.

Signed-off-by: Charlene Liu <charlene.liu@amd.com>
Reviewed-by: Krunoslav Kovac <Krunoslav.Kovac@amd.com>
Acked-by: Harry Wentland <harry.wentland@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
6 years agodrm/amd/display: Remove unnecessary wait mpcc idle.
Yongqiang Sun [Wed, 15 Nov 2017 21:06:15 +0000 (16:06 -0500)]
drm/amd/display: Remove unnecessary wait mpcc idle.

Before power gate plane, mpcc idle wait is processed,
no need to wait another time.

Signed-off-by: Yongqiang Sun <yongqiang.sun@amd.com>
Reviewed-by: Tony Cheng <Tony.Cheng@amd.com>
Acked-by: Harry Wentland <harry.wentland@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
6 years agodrm/amd/pp: reset dpm level when adjust power state
Rex Zhu [Wed, 13 Dec 2017 09:39:07 +0000 (17:39 +0800)]
drm/amd/pp: reset dpm level when adjust power state

Acked-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Rex Zhu <Rex.Zhu@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
6 years agodrm/ttm: init locked again to prevent incorrect unlock
Roger He [Thu, 7 Dec 2017 04:51:19 +0000 (12:51 +0800)]
drm/ttm: init locked again to prevent incorrect unlock

Signed-off-by: Roger He <Hongbo.He@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
6 years agodrm/amd/pp: implement dpm_get_sclk/mclk for RV
Rex Zhu [Tue, 12 Dec 2017 07:06:10 +0000 (15:06 +0800)]
drm/amd/pp: implement dpm_get_sclk/mclk for RV

RV implementation was missing these callbacks.  Used
to fetch the clock values for other components.

Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Evan Quan <evan.quan@amd.com>
Signed-off-by: Rex Zhu <Rex.Zhu@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
6 years agodrm/amdgpu: fix huge page setting for ATS case
Chunming Zhou [Mon, 11 Dec 2017 07:55:03 +0000 (15:55 +0800)]
drm/amdgpu: fix huge page setting for ATS case

Signed-off-by: Chunming Zhou <david1.zhou@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
6 years agodrm/amdgpu: drop amdgpu_atombios_scratch_regs_save/restore
Alex Deucher [Tue, 12 Dec 2017 20:26:10 +0000 (15:26 -0500)]
drm/amdgpu: drop amdgpu_atombios_scratch_regs_save/restore

No longer used.

Reviewed-by: Harry Wentland <harry.wentland@amd.com>
Acked-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
6 years agodrm/amdgpu: drop scratch regs save and restore from GPU reset handling
Alex Deucher [Tue, 12 Dec 2017 20:22:56 +0000 (15:22 -0500)]
drm/amdgpu: drop scratch regs save and restore from GPU reset handling

The expectation is that the base driver doesn't mess with these.
Some components interact with these directly so let the components
handle these directly.

Reviewed-by: Harry Wentland <harry.wentland@amd.com>
Acked-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
6 years agodrm/amdgpu: drop scratch regs save and restore from S3/S4 handling
Alex Deucher [Tue, 12 Dec 2017 20:20:22 +0000 (15:20 -0500)]
drm/amdgpu: drop scratch regs save and restore from S3/S4 handling

The expectation is that the base driver doesn't mess with these.
Some components interact with these directly so let the components
handle these directly.

Reviewed-by: Harry Wentland <harry.wentland@amd.com>
Acked-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
6 years agodrm/amdgpu: remove some old gc 9.x registers
Alex Deucher [Fri, 8 Dec 2017 22:02:24 +0000 (17:02 -0500)]
drm/amdgpu: remove some old gc 9.x registers

Leftover from bring up.

Acked-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
6 years agodrm/amdgpu: drop soc15_init_golden_registers
Alex Deucher [Fri, 8 Dec 2017 18:18:23 +0000 (13:18 -0500)]
drm/amdgpu: drop soc15_init_golden_registers

The golden register arrays were empty so the function was
effectively useless.

Acked-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
6 years agodrm/amdgpu: drop the bios scratch reg callbacks from nbio
Alex Deucher [Fri, 8 Dec 2017 18:11:33 +0000 (13:11 -0500)]
drm/amdgpu: drop the bios scratch reg callbacks from nbio

They are not used any longer.  We get the scratch register
locations from the vbios directly now.

Acked-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
6 years agodrm/amdgpu: convert nbio to use callbacks (v2)
Alex Deucher [Fri, 8 Dec 2017 18:07:58 +0000 (13:07 -0500)]
drm/amdgpu: convert nbio to use callbacks (v2)

Cleans up and consolidates all of the per-asic logic.

v2: squash in "drm/amdgpu: fix NULL err for sriov detect" (Chunming)

Acked-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
6 years agodrm/amdgpu: make function names consistent in nbio files
Alex Deucher [Fri, 8 Dec 2017 16:39:49 +0000 (11:39 -0500)]
drm/amdgpu: make function names consistent in nbio files

All functions should have nbio_v* prefix.

Acked-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
6 years agodrm/amdgpu: correct vce fw data and stack size
Frank Min [Mon, 6 Nov 2017 07:34:55 +0000 (15:34 +0800)]
drm/amdgpu: correct vce fw data and stack size

this fix the VCE world switch hang issue

Signed-off-by: Frank Min <Frank.Min@amd.com>
Acked-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
6 years agodrm/amdgpu: fix MAP_QUEUES paramter
Monk Liu [Thu, 23 Nov 2017 10:38:59 +0000 (18:38 +0800)]
drm/amdgpu: fix MAP_QUEUES paramter

Should be 0.

Signed-off-by: Monk Liu <Monk.Liu@amd.com>
Acked-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
6 years agodrm/amdgpu: no need with INT for fence polling
Monk Liu [Mon, 4 Dec 2017 12:46:17 +0000 (20:46 +0800)]
drm/amdgpu: no need with INT for fence polling

We are polling so no need for INT.

Signed-off-by: Monk Liu <Monk.Liu@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
6 years agodrm/amdgpu: no need to evict VRAM in device_fini
Monk Liu [Wed, 22 Nov 2017 11:21:43 +0000 (19:21 +0800)]
drm/amdgpu: no need to evict VRAM in device_fini

this VRAM evict is not needed and also cost 2seconds
to finish because the IRQ is software side disabled
before it.

Signed-off-by: Monk Liu <Monk.Liu@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
6 years agodrm/ttm: max_cpages is in unit of native page
Monk Liu [Fri, 1 Dec 2017 10:23:56 +0000 (18:23 +0800)]
drm/ttm: max_cpages is in unit of native page

fix calculation.

Signed-off-by: Monk Liu <Monk.Liu@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
6 years agodrm/ttm: fix incorrect calculate on shrink_pages
Monk Liu [Fri, 1 Dec 2017 10:21:34 +0000 (18:21 +0800)]
drm/ttm: fix incorrect calculate on shrink_pages

shrink_pages is in unit of Order after ttm_page_pool_free,
but it is used by nr_free in next round so need change
it into native page unit

Signed-off-by: Monk Liu <Monk.Liu@amd.com>
Reviewed-by: Roger He <Hongbo.He@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
6 years agodrm/amdgpu: allow get_vm_pde to change flags as well
Christian König [Wed, 29 Nov 2017 12:27:26 +0000 (13:27 +0100)]
drm/amdgpu: allow get_vm_pde to change flags as well

And also provide the level for which we need a PDE.

Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Chunming Zhou <david1.zhou@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
6 years agodrm/amdgpu: batch PDE updates again
Christian König [Thu, 30 Nov 2017 18:08:05 +0000 (19:08 +0100)]
drm/amdgpu: batch PDE updates again

Now instead of one submission for each PDE batch them together over all
PDs who need an update.

Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Chunming Zhou <david1.zhou@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
6 years agodrm/amdgpu: remove keeping the addr of the VM PDs
Christian König [Thu, 30 Nov 2017 14:41:28 +0000 (15:41 +0100)]
drm/amdgpu: remove keeping the addr of the VM PDs

No more double house keeping.

Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Chunming Zhou <david1.zhou@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
6 years agodrm/amdgpu: remove last_entry_used from the VM code
Christian König [Thu, 30 Nov 2017 14:28:03 +0000 (15:28 +0100)]
drm/amdgpu: remove last_entry_used from the VM code

Not needed any more.

Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Chunming Zhou <david1.zhou@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>