platform/kernel/linux-rpi.git
5 years agodrm/amdgpu: remove chash
Christian König [Wed, 7 Nov 2018 12:55:01 +0000 (13:55 +0100)]
drm/amdgpu: remove chash

Remove the chash implementation for now since it isn't used any more.

Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amdgpu: use ring/hash for fault handling on GMC9 v3
Christian König [Wed, 7 Nov 2018 12:55:01 +0000 (13:55 +0100)]
drm/amdgpu: use ring/hash for fault handling on GMC9 v3

Further testing showed that the idea with the chash doesn't work as expected.
Especially we can't predict when we can remove the entries from the hash again.

So replace the chash with a ring buffer/hash mix where entries in the container
age automatically based on their timestamp.

v2: use ring buffer / hash mix
v3: check the timeout to make sure all entries age

Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com> (v2)
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amdgpu: limit the number of IVs processed at once
Christian König [Tue, 5 Mar 2019 14:13:52 +0000 (15:13 +0100)]
drm/amdgpu: limit the number of IVs processed at once

Only process a maximum of 32 IVs before writing back the RPTR. This improves
hw handling when we get close to an overflow in the ring buffer.

Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amdgpu: enable IH ring 1&2 for Vega20 as well
Christian König [Mon, 4 Mar 2019 18:59:36 +0000 (19:59 +0100)]
drm/amdgpu: enable IH ring 1&2 for Vega20 as well

That doesn't seem to have any negative effects.

Signed-off-by: Christian König <christian.koenig@amd.com>
Acked-by: Chunming Zhou <david1.zhou@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amdgpu: enable IH doorbell for ring 1&2 on Vega
Christian König [Wed, 27 Feb 2019 14:05:20 +0000 (15:05 +0100)]
drm/amdgpu: enable IH doorbell for ring 1&2 on Vega

The doorbells should already be reserved, just enable them.

Signed-off-by: Christian König <christian.koenig@amd.com>
Acked-by: Chunming Zhou <david1.zhou@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amdgpu: change Vega IH ring 1 config
Christian König [Wed, 27 Feb 2019 13:54:23 +0000 (14:54 +0100)]
drm/amdgpu: change Vega IH ring 1 config

Disable overflow and enable full drain. This makes fault handling on ring 1
much more reliable since we don't generate back pressure any more.

Signed-off-by: Christian König <christian.koenig@amd.com>
Acked-by: Chunming Zhou <david1.zhou@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amdgpu: Only clear dumb buffers if ring is enabled
Nicholas Kazlauskas [Mon, 11 Mar 2019 18:16:29 +0000 (14:16 -0400)]
drm/amdgpu: Only clear dumb buffers if ring is enabled

The buffers should be cleared when possible but we also don't want
buffer creation to fail in the rare case where the ring isn't ready
during the call. This could happen during some suspend/resume sequences.

Cc: Christian König <ckoenig.leichtzumerken@gmail.com>
Signed-off-by: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amdgpu: Clear VRAM for DRM dumb_create buffers
Nicholas Kazlauskas [Fri, 8 Mar 2019 15:38:16 +0000 (10:38 -0500)]
drm/amdgpu: Clear VRAM for DRM dumb_create buffers

The dumb_create API isn't intended for high performance rendering
and it's more useful for userspace (ie. IGT) to have them precleared.

The bonus here is that we also won't needlessly leak whatever was
previously in VRAM, but it also probably wasn't sensitive if it was
going through this API.

Signed-off-by: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amdgpu: fix semicolon.cocci warnings
kbuild test robot [Wed, 6 Mar 2019 05:26:11 +0000 (13:26 +0800)]
drm/amdgpu: fix semicolon.cocci warnings

drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c:405:2-3: Unneeded semicolon
drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c:435:2-3: Unneeded semicolon

 Remove unneeded semicolon.

Generated by: scripts/coccinelle/misc/semicolon.cocci

CC: xinhui pan <xinhui.pan@amd.com>
Signed-off-by: kbuild test robot <fengguang.wu@intel.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amdgpu: add new ras workflow control flags
xinhui pan [Mon, 11 Mar 2019 07:23:00 +0000 (15:23 +0800)]
drm/amdgpu: add new ras workflow control flags

add ras post init function.
Do some initialization after all IP have finished their late init.

Add new member flags which will control the ras work flow.
For now, vbios enable ras for us on boot. That might change in the
future.
So there should be a flag from vbios to tell us if ras is enabled or not
on boot. Looks like there is no such info now.

Other bits of the flags are reserved to control other parts of ras.

Signed-off-by: xinhui pan <xinhui.pan@amd.com>
Reviewed-by: Evan Quan <evan.quan@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amdgpu: let ras initialization a little noticeable
xinhui pan [Tue, 12 Mar 2019 09:15:57 +0000 (17:15 +0800)]
drm/amdgpu: let ras initialization a little noticeable

add drm info output if ras initialized successfully.
add ras atomfirmware sanity check.

Signed-off-by: xinhui pan <xinhui.pan@amd.com>
Reviewed-by: Evan Quan <evan.quan@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amdgpu: Fix lockdep warning more gracely
xinhui pan [Mon, 11 Mar 2019 11:34:23 +0000 (19:34 +0800)]
drm/amdgpu: Fix lockdep warning more gracely

lockdep need a static key.
Previously we set ignore bit to avoid the warning.
Now call sysfs_attr_init to initialize the static key.

Signed-off-by: xinhui pan <xinhui.pan@amd.com>
Reviewed-and-Tested-by: Andrey Grodzovsky <andrey.grodzovsky@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amdgpu: Fix ras debugfs data parse
xinhui pan [Mon, 11 Mar 2019 10:10:57 +0000 (18:10 +0800)]
drm/amdgpu: Fix ras debugfs data parse

Unzero char is accepted by sscanf, so when data is structure but
unexpectedly return error invalid;

Signed-off-by: xinhui pan <xinhui.pan@amd.com>
Reviewed-by: Feifei Xu <Feifei.Xu@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amdgpu: add new member hw_supported
xinhui pan [Mon, 11 Mar 2019 06:12:40 +0000 (14:12 +0800)]
drm/amdgpu: add new member hw_supported

Currently, it is not clear how ras is supported. Both software and
hardware can set the supported. That is confusing.

Fix it by adding new member hw_supported.

Signed-off-by: xinhui pan <xinhui.pan@amd.com>
Reviewed-by: Evan Quan <evan.quan@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amdgpu: Fix warning when lockdep is enabled
xinhui pan [Mon, 11 Mar 2019 06:28:06 +0000 (14:28 +0800)]
drm/amdgpu: Fix warning when lockdep is enabled

Set ignore bit to satisfy locpdep.

Signed-off-by: xinhui pan <xinhui.pan@amd.com>
Reviewed-by: Evan Quan <evan.quan@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amdgpu: Fix NULL pointer when ta is missing
xinhui pan [Mon, 11 Mar 2019 04:17:14 +0000 (12:17 +0800)]
drm/amdgpu: Fix NULL pointer when ta is missing

Ta is optional, so check if ta firmware is loaded or not.

Signed-off-by: xinhui pan <xinhui.pan@amd.com>
Reviewed-by: Evan Quan <evan.quan@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amdgpu: fix ras parameter descriptions
Evan Quan [Thu, 7 Mar 2019 06:46:38 +0000 (14:46 +0800)]
drm/amdgpu: fix ras parameter descriptions

The descriptions of modinfo wrongly show two parameters
for each feature(see below). This patch can fix this
incorrect outputs.

parm:           amdgpu_ras_enable:Enable RAS features on the GPU (0 = disable, 1 = enable, -1 = auto (default))
parm:           ras_enable:int
parm:           amdgpu_ras_mask:Mask of RAS features to enable (default 0xffffffff), only valid when ras_enable == 1
parm:           ras_mask:uint

Signed-off-by: Evan Quan <evan.quan@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: xinhui pan <xinhui.pan@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amdgpu: export both supported and enabled ras features
xinhui pan [Thu, 7 Mar 2019 06:01:11 +0000 (14:01 +0800)]
drm/amdgpu: export both supported and enabled ras features

Signed-off-by: xinhui pan <xinhui.pan@amd.com>
Reviewed-by: Evan Quan <evan.quan@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amdgpu: lookup vbios table to check ecc capability
xinhui pan [Thu, 7 Mar 2019 03:49:26 +0000 (11:49 +0800)]
drm/amdgpu: lookup vbios table to check ecc capability

Signed-off-by: xinhui pan <xinhui.pan@amd.com>
Reviewed-by: Evan Quan <evan.quan@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amdgpu: query sram ecc/ecc availability from atombios
Hawking Zhang [Thu, 7 Mar 2019 03:00:20 +0000 (11:00 +0800)]
drm/amdgpu: query sram ecc/ecc availability from atombios

query sram ecc capability via amdgpu_atomfirmware_ecc_default_enabled
query ecc availability via amdgpu_atomfirmware_sram_ecc_supported

Signed-off-by: Hawking Zhang <Hawking.Zhang@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amdgpu: add atomfirmware helper function to query sram ecc caps
Hawking Zhang [Thu, 7 Mar 2019 02:52:43 +0000 (10:52 +0800)]
drm/amdgpu: add atomfirmware helper function to query sram ecc caps

sram ecc capability could be get from firmware_capability field in firmwareinfo table

Signed-off-by: Hawking Zhang <Hawking.Zhang@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amdgpu: add atomfirmware helper function to query ecc status
Hawking Zhang [Thu, 7 Mar 2019 02:14:13 +0000 (10:14 +0800)]
drm/amdgpu: add atomfirmware helper function to query ecc status

ecc default status (enabled or disabled) could be get from umc_config field in umc_info table

Signed-off-by: Hawking Zhang <Hawking.Zhang@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amdgpu: update atomfirmware header with ecc related members
Hawking Zhang [Thu, 7 Mar 2019 02:10:16 +0000 (10:10 +0800)]
drm/amdgpu: update atomfirmware header with ecc related members

add new umc_info structures and new firmware_capability defines

Signed-off-by: Hawking Zhang <Hawking.Zhang@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amdgpu: handle ras resume
xinhui pan [Thu, 7 Mar 2019 02:46:54 +0000 (10:46 +0800)]
drm/amdgpu: handle ras resume

Suspend will put irq, so resume need get irq back.
And in the same time, skip other ras initialization.

Signed-off-by: xinhui pan <xinhui.pan@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amdkfd: add RAS ECC event support (v3)
Eric Huang [Fri, 11 Jan 2019 19:38:51 +0000 (14:38 -0500)]
drm/amdkfd: add RAS ECC event support (v3)

RAS ECC event will combine with GPU reset event, due to
ECC interrupts are caused by uncorrectable error that triggers
GPU reset.

v2: Fix misleading-indentation warning
v3: fix build with CONFIG_HSA_AMD disabled

Signed-off-by: Eric Huang <JinhuiEric.Huang@amd.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amdkfd: add RAS capabilities in topology for Vega20 (v2)
Eric Huang [Fri, 11 Jan 2019 19:52:22 +0000 (14:52 -0500)]
drm/amdkfd: add RAS capabilities in topology for Vega20 (v2)

It is to collaborate with HSA_CAPABILITY in libhsakmt.

v2: squash in NULL pointer check

Signed-off-by: Eric Huang <JinhuiEric.Huang@amd.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amdgpu: add human readable debugfs control support (v2)
xinhui pan [Fri, 1 Mar 2019 08:32:11 +0000 (16:32 +0800)]
drm/amdgpu: add human readable debugfs control support (v2)

Currently, the debugfs control node can't parse bash-like commands.
Now add such support for any tester that uses scripts.

v2: squash in fixes for input validation

Signed-off-by: xinhui pan <xinhui.pan@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amdgpu: skip gpu reset when ras error occured
xinhui pan [Fri, 1 Mar 2019 05:29:13 +0000 (13:29 +0800)]
drm/amdgpu: skip gpu reset when ras error occured

gpu reset is not stable on vega20 A1.

Signed-off-by: xinhui pan <xinhui.pan@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amdgpu: add ioctl query for enabled ras features (v2)
xinhui pan [Mon, 17 Dec 2018 07:09:07 +0000 (15:09 +0800)]
drm/amdgpu: add ioctl query for enabled ras features (v2)

Add a query for userspace to check which RAS features
are enabled.

v2: squash in warning fix

Signed-off-by: xinhui pan <xinhui.pan@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amdgpu: Add a new flag to AMDGPU_CTX_OP_QUERY_STATE2
xinhui pan [Mon, 17 Dec 2018 06:31:12 +0000 (14:31 +0800)]
drm/amdgpu: Add a new flag to AMDGPU_CTX_OP_QUERY_STATE2

Add AMDGPU_CTX_QUERY2_FLAGS_RAS_CE/UE which indicate if any error happened
between previous query and this query.

Signed-off-by: xinhui pan <xinhui.pan@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amdgpu: enable ras on gmc9
xinhui pan [Wed, 23 Jan 2019 11:03:25 +0000 (19:03 +0800)]
drm/amdgpu: enable ras on gmc9

Signed-off-by: xinhui pan <xinhui.pan@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amdgpu: enable ras on gfx9 (v2)
Feifei Xu [Fri, 7 Dec 2018 09:52:20 +0000 (17:52 +0800)]
drm/amdgpu: enable ras on gfx9 (v2)

Register ecc interrupts and ecc interrupt handler on gfx9.
Add ras support on gfx9

v2: squash in warning fix

Signed-off-by: Feifei Xu <Feifei.Xu@amd.com>
Signed-off-by: xinhui pan <xinhui.pan@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amdgpu: enable ras on sdma4
xinhui pan [Wed, 28 Nov 2018 13:14:56 +0000 (21:14 +0800)]
drm/amdgpu: enable ras on sdma4

register IH, enable ras features on sdma.
create sysfs debugfs file for sdma.

Signed-off-by: xinhui pan <xinhui.pan@amd.com>
Signed-off-by: Feifei Xu <Feifei.Xu@amd.com>
Signed-off-by: Eric Huang <JinhuiEric.Huang@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amdgpu: reserve bad pages during recovery
xinhui pan [Mon, 21 Jan 2019 12:33:48 +0000 (20:33 +0800)]
drm/amdgpu: reserve bad pages during recovery

Mark vram pages with errors as bad and prevent the driver
from using them.

Signed-off-by: xinhui pan <xinhui.pan@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amdgpu: add debugfs ctrl node
xinhui pan [Thu, 31 Jan 2019 08:55:07 +0000 (16:55 +0800)]
drm/amdgpu: add debugfs ctrl node

allow userspace enable/disable ras

Signed-off-by: xinhui pan <xinhui.pan@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amdgpu: add amdgpu_ras.c to support ras (v2)
xinhui pan [Wed, 31 Oct 2018 06:38:28 +0000 (14:38 +0800)]
drm/amdgpu: add amdgpu_ras.c to support ras (v2)

add obj management.
add feature control.
add debugfs infrastructure.
add sysfs infrastructure.
add IH infrastructure.
add recovery infrastructure.

It is a framework. Other IPs need call amdgpu_ras_xxx function instead of
psp_ras_xxx functions.

v2: squash in warning fixes

Signed-off-by: xinhui pan <xinhui.pan@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amdgpu: add psp cmd submit timeout
xinhui pan [Wed, 23 Jan 2019 03:35:31 +0000 (11:35 +0800)]
drm/amdgpu: add psp cmd submit timeout

Signed-off-by: xinhui pan <xinhui.pan@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amdgpu: add psp v11 ras callback
xinhui pan [Wed, 14 Nov 2018 10:06:33 +0000 (18:06 +0800)]
drm/amdgpu: add psp v11 ras callback

Add trigger_error and cure_posion.

Acked-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: xinhui pan <xinhui.pan@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amdgpu: add psp ras subsystem infrastructure (v2)
xinhui pan [Wed, 21 Nov 2018 03:17:49 +0000 (11:17 +0800)]
drm/amdgpu: add psp ras subsystem infrastructure (v2)

Add ras fw loading, init, terminate.
Add ras cmd submit helper.
Add ras feature enable/disable common function.

v2: squash in unused variable warning fix

Signed-off-by: xinhui pan <xinhui.pan@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amdgpu: add psp ras callback func and macro
xinhui pan [Tue, 30 Oct 2018 05:44:46 +0000 (13:44 +0800)]
drm/amdgpu: add psp ras callback func and macro

Define the driver side interface for ras ta.

Acked-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: xinhui pan <xinhui.pan@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amdgpu: add ta_ras_if.h
xinhui pan [Tue, 30 Oct 2018 05:44:46 +0000 (13:44 +0800)]
drm/amdgpu: add ta_ras_if.h

Signed-off-by: xinhui pan <xinhui.pan@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amdgpu: add module parameters for ras
xinhui pan [Thu, 25 Oct 2018 06:35:13 +0000 (14:35 +0800)]
drm/amdgpu: add module parameters for ras

Allow RAS feature enable/disable via boot parameter.

Signed-off-by: xinhui pan <xinhui.pan@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amdgpu: export ta fw info
xinhui pan [Tue, 20 Nov 2018 02:22:36 +0000 (10:22 +0800)]
drm/amdgpu: export ta fw info

Output the ta fw, aka xgmi/ras, via debugfs.

Signed-off-by: xinhui pan <xinhui.pan@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amdgpu: add ta ras fw info (v2)
xinhui pan [Tue, 20 Nov 2018 02:32:00 +0000 (10:32 +0800)]
drm/amdgpu: add ta ras fw info (v2)

Add ras fw part, xgmi and ras fw are combined together in ta binary.
Reading the data from the info is not implemented yet.

v2: squash in "drm/amdgpu: fix NULL pointer when ta is missing"

Signed-off-by: xinhui pan <xinhui.pan@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amd/powerplay: move the smc_if_version to asic file
Kevin Wang [Mon, 11 Mar 2019 03:00:48 +0000 (11:00 +0800)]
drm/amd/powerplay: move the smc_if_version to asic file

each asic may be has different smc if version,
so move its to asic file to implement.

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: simplify sw-smu message map macro
Kevin Wang [Wed, 6 Mar 2019 11:45:16 +0000 (19:45 +0800)]
drm/amd/powerplay: simplify sw-smu message map macro

simplify macro of MSG_MAP for sw-smu

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/amdgpu: Cosmetic change for calling func amdgpu_gmc_vram_location
Oak Zeng [Thu, 7 Mar 2019 23:00:20 +0000 (17:00 -0600)]
drm/amdgpu: Cosmetic change for calling func amdgpu_gmc_vram_location

Use function parameter mc as the second parameter of amdgpu_gmc_vram_location,
so codes look more consistent.

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: Fix plane address updates for video surface formats
Nicholas Kazlauskas [Mon, 11 Mar 2019 13:38:35 +0000 (09:38 -0400)]
drm/amd/display: Fix plane address updates for video surface formats

[Why]
For new DC planes the correct plane address fields are filled based
on whether the plane had a graphics or video format.

However, when we perform stream and plane updates using DC we only ever
fill in the graphics format fields. This causing corruption and hangs
when using video surface formats like NV12 for planes.

[How]
Use the same logic everywhere we update dc_plane_address - always
fill in the correct fields based on the surface format type.

There are 3 places this is done:

- Atomic check, during DC plane creation
- Atomic commit, during plane prepare_fb
- Atomic commit tail, during amdgpu_dm_commit_planes

We use the fill_plane_tiling_attributes in all 3 locations and it
already needs the address to update DCC attributes, so the surface
address update logic can be moved into this helper.

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: Move IB pool init and fini v2
Andrey Grodzovsky [Wed, 6 Mar 2019 21:16:28 +0000 (16:16 -0500)]
drm/amdgpu: Move IB pool init and fini v2

Problem:
Using SDMA for TLB invalidation in certain ASICs exposed a problem
of IB pool not being ready while SDMA already up on Init and already
shutt down while SDMA still running on Fini. This caused
IB allocation failure. Temproary fix was commited into a
bringup branch but this is the generic fix.

Fix:
Init IB pool rigth after GMC is ready but before SDMA is ready.
Do th opposite for Fini.

v2: Remove restriction on SDMA early init and move amdgpu_ib_pool_fini

Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Andrey Grodzovsky <andrey.grodzovsky@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amd/display: On DCN1, Wait for vupdate on cursor updates
David Francis [Thu, 21 Feb 2019 20:04:05 +0000 (15:04 -0500)]
drm/amd/display: On DCN1, Wait for vupdate on cursor updates

[Why]
Cursor updates must acquire the pipe control lock to
prevent vupdate from triggering in the middle of cursor
programming. On DCN1 the pipe control lock prevents
pageflips from occurring. This means that a cursor update
right before vupdate can delay a pending pageflip

[How]
If the time until the next vupdate is less than a
conservative estimate of the cursor programming time,
wait until the vupdate has passed before locking.

Signed-off-by: David Francis <David.Francis@amd.com>
Reviewed-by: Harry Wentland <Harry.Wentland@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: add i2c over aux failure handling
Wenjing Liu [Sat, 23 Feb 2019 21:07:45 +0000 (16:07 -0500)]
drm/amd/display: add i2c over aux failure handling

[why]
We will not retry when EDID read failure using i2c over aux

[how]
treat i2c over aux failure the same as defer

Signed-off-by: Wenjing Liu <Wenjing.Liu@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: fix up reference clock abstractions
Jun Lei [Fri, 22 Feb 2019 21:50:00 +0000 (16:50 -0500)]
drm/amd/display: fix up reference clock abstractions

[why]
"reference clock" is a very overloaded variable in DC and causes confusion
as there are multiple sources of reference clock, which may be different values
incorrect input values to DML will cause DCHUB to be programmed improperly
and lead to hard to debug underflow issues

[how]
instead of using ref clock everywhere, specify WHICH ref clock:
- xtalin
- dccg refclk
- dchub refclk

these are all distinct values which may not be equal

Signed-off-by: Jun Lei <Jun.Lei@amd.com>
Reviewed-by: Yongqiang Sun <yongqiang.sun@amd.com>
Acked-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: Expose support for alpha blending on overlays
Nicholas Kazlauskas [Thu, 21 Feb 2019 19:57:48 +0000 (14:57 -0500)]
drm/amd/display: Expose support for alpha blending on overlays

[Why]
The DRM overlay planes DM exposes support RGBA formats but are currently
forced as fully opaque over whatever they overlay.

[How]
Expose DRM blending mode and alpha properties to userspace.

The overlays exposed support per-pixel pre-multiplied alpha along with
global plane opacity.

Signed-off-by: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com>
Reviewed-by: Harry Wentland <Harry.Wentland@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 HW i2c arbitration with dmcu
Charlene Liu [Fri, 22 Feb 2019 18:55:46 +0000 (13:55 -0500)]
drm/amd/display: add HW i2c arbitration with dmcu

Signed-off-by: Charlene Liu <charlene.liu@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: implement bounding box update based on uclk breakdown
Jun Lei [Tue, 19 Feb 2019 15:33:21 +0000 (10:33 -0500)]
drm/amd/display: implement bounding box update based on uclk breakdown

[why]
Actual breakdown of DPM level varies by SKU (for the same family)
DC needs some ability to ammend pre-silicon numbers

Signed-off-by: Jun Lei <Jun.Lei@amd.com>
Reviewed-by: Tony Cheng <Tony.Cheng@amd.com>
Acked-by: Leo Li <sunpeng.li@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amd/display: 3.2.21
Aric Cyr [Fri, 22 Feb 2019 22:02:14 +0000 (17:02 -0500)]
drm/amd/display: 3.2.21

Signed-off-by: Aric Cyr <aric.cyr@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: reduce abm min reduction, deviation gain and contrast factor
Josip Pavic [Fri, 22 Feb 2019 21:55:08 +0000 (16:55 -0500)]
drm/amd/display: reduce abm min reduction, deviation gain and contrast factor

[Why]
Increased power savings are desired for ABM 2.2.

[How]
Reduce the minimum reduction level, the deviation gain and the contrast factor
to allow for more aggressive operation of the algorithm.

Signed-off-by: Josip Pavic <Josip.Pavic@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: Rename is_hdmi to is_hdmi_tmds type
Eric Bernstein [Fri, 15 Feb 2019 14:27:22 +0000 (09:27 -0500)]
drm/amd/display: Rename is_hdmi to is_hdmi_tmds type

HDMI has TMDS and FRL signal types. Be specific about what is used.

Signed-off-by: Eric Bernstein <eric.bernstein@amd.com>
Reviewed-by: Nevenko Stupar <Nevenko.Stupar@amd.com>
Reviewed-by: Tony Cheng <Tony.Cheng@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 pipe lock during stream update
Wenjing Liu [Wed, 20 Feb 2019 19:00:55 +0000 (14:00 -0500)]
drm/amd/display: add pipe lock during stream update

[why]
Stream update will adjust both info packets and stream params,
need to make sure all things are applied togather.

[how]
add pipe lock during stream update

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: Increase DP blank timeout from 30 ms to 50 ms
Ken Chalmers [Fri, 15 Feb 2019 15:31:52 +0000 (09:31 -0600)]
drm/amd/display: Increase DP blank timeout from 30 ms to 50 ms

[Why]
At 24 Hz, a frame is 41.7 ms, so a 30 ms wait can (and does often)
timeout.

[How]
Bump timeout from 30 ms to 50 ms.

Signed-off-by: Ken Chalmers <ken.chalmers@amd.com>
Reviewed-by: Tony Cheng <Tony.Cheng@amd.com>
Acked-by: Leo Li <sunpeng.li@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amd/display: Combine field toggle macro and sequence write macro.
Yongqiang Sun [Fri, 15 Feb 2019 22:05:43 +0000 (17:05 -0500)]
drm/amd/display: Combine field toggle macro and sequence write macro.

[Why]
field toggle write is actual field sequence write with the same
field name.

[How]
Use REG_UPDATE_SEQ_2 for both sequence write and toggle.
Rename REG_UPDATE_1by1_3 to REG_UPDATE_SEQ_3.

Signed-off-by: Yongqiang Sun <yongqiang.sun@amd.com>
Reviewed-by: Tony Cheng <Tony.Cheng@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 PSR SMU Interrupt support
SivapiriyanKumarasamy [Mon, 10 Sep 2018 22:59:42 +0000 (18:59 -0400)]
drm/amd/display: Add PSR SMU Interrupt support

[WHY]
We have new bios capabilities enabling s0i2 entry on SMU interrupt. We want
this interrupt to be fired on PSR transitions such that we enter s0i2
when entering PSR active.

[HOW]
Add code to send the SMU interrupt with the appropriate
staticscreen flag when entering and exting PSR. Protect this
code with a config flag since it currently impacts BL PWM.

Signed-off-by: SivapiriyanKumarasamy <sivapiriyan.kumarasamy@amd.com>
Reviewed-by: Anthony Koo <Anthony.Koo@amd.com>
Acked-by: Leo Li <sunpeng.li@amd.com>
Acked-by: Tony Cheng <Tony.Cheng@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amd/display: Refactor reg_set and reg_update.
Yongqiang Sun [Wed, 13 Feb 2019 15:35:43 +0000 (10:35 -0500)]
drm/amd/display: Refactor reg_set and reg_update.

[Why]
Current reg update and reg set use same functions and
only delta is update reads reg value and call update function.

[How]
Refactor reg update and reg set functions.
1.Implement different functions for reg update and reg set.
2.Wrap same process to a help function, both reg update and
reg set will call it.

Signed-off-by: Yongqiang Sun <yongqiang.sun@amd.com>
Reviewed-by: Tony Cheng <Tony.Cheng@amd.com>
Acked-by: Leo Li <sunpeng.li@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amd/display: Free DCN version of stream encoder
Eric Bernstein [Wed, 13 Feb 2019 22:56:38 +0000 (17:56 -0500)]
drm/amd/display: Free DCN version of stream encoder

Cross a TODO item off the list. Cleanup SIGNAL_TYPE_HDMI_FRL, it's
not currently supported.

Signed-off-by: Eric Bernstein <eric.bernstein@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: clean up dml_init_instance
Dmytro Laktyushkin [Mon, 11 Feb 2019 22:18:41 +0000 (17:18 -0500)]
drm/amd/display: clean up dml_init_instance

Get rid of DV style dml init in favour of the cleaner DC
style.

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/powerplay: apply Vega20 BACO workaround
Evan Quan [Thu, 7 Mar 2019 02:20:12 +0000 (10:20 +0800)]
drm/amd/powerplay: apply Vega20 BACO workaround

Applied vdci flush workaround for Vega20 BACO.

Signed-off-by: Evan Quan <evan.quan@amd.com>
Acked-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amdgpu: XGMI pstate switch initial support
shaoyunl [Thu, 21 Feb 2019 21:30:25 +0000 (16:30 -0500)]
drm/amdgpu: XGMI pstate switch initial support

Driver vote low to high pstate switch whenever there is an outstanding
XGMI mapping request. Driver vote high to low pstate when all the
outstanding XGMI mapping is terminated.

Signed-off-by: shaoyunl <shaoyun.liu@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amdgpu: Enable XGMI mapping for peer device
shaoyunl [Fri, 22 Feb 2019 21:20:38 +0000 (16:20 -0500)]
drm/amdgpu: Enable XGMI mapping for peer device

Adjust vram base offset for XGMI mapping when update the PT entry so
the address will fall into correct XGMI aperture for peer device

Signed-off-by: shaoyunl <shaoyun.liu@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amdgpu: Add sysfs entries for xgmi hive v2.
Andrey Grodzovsky [Tue, 5 Mar 2019 15:39:08 +0000 (10:39 -0500)]
drm/amdgpu: Add sysfs entries  for xgmi hive v2.

For each device a file xgmi_device_id is created.
On the first device a subdirectory named xgmi_hive_info is created,
It contains  a file named hive_id and symlinks named node 1-4 linking
to each device in the hive.

v2: Return error codes instead of '-1' and few misspellings.

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/display: Don't ASSERT when total_planes == AMDGPU_MAX_PLANES
Nicholas Kazlauskas [Tue, 5 Mar 2019 17:24:08 +0000 (12:24 -0500)]
drm/amd/display: Don't ASSERT when total_planes == AMDGPU_MAX_PLANES

[Why]
Can happen on ASICs with 6 planes, but this isn't a bug since we haven't
written outside the array.

[How]
Use <= instead of <.

Cc: Leo Li <sunpeng.li@amd.com>
Signed-off-by: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com>
Reported-by: Michel Dänzer <michel.daenzer@amd.com>
Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
Tested-by: Michel Dänzer <michel.daenzer@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amd/display: Drop atomic_obj_lock for private obj
Nicholas Kazlauskas [Tue, 5 Mar 2019 14:14:17 +0000 (09:14 -0500)]
drm/amd/display: Drop atomic_obj_lock for private obj

[Why]
New DRM versions manage locking for private objects for us, so this
is no longer needed.

This also prevents a WARN_ON from occurring when the private object is
duplicated during the forced atomic commit that occurs from the HPD
handler.

The HPD handler calls drm_modeset_lock_all before the forced commit
and if the private object is duplicated then the
DEBUG_LOCKS_WARN_ON(ww_ctx->done_acquire) warning will be triggered
since we're trying to lock something when everything should have
already been locked.

[How]
Drop the lock and let DRM manage this.

Cc: Harry Wentland <harry.wentland@amd.com>
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/amd/amdgpu: Add ENGINE_CNTL register to vcn10 headers
Tom St Denis [Mon, 4 Mar 2019 15:58:44 +0000 (10:58 -0500)]
drm/amd/amdgpu: Add ENGINE_CNTL register to vcn10 headers

Signed-off-by: Tom St Denis <tom.stdenis@amd.com>
Acked-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amdgpu: allow huge invalid mappings on GMC8
Christian König [Mon, 4 Feb 2019 14:00:34 +0000 (15:00 +0100)]
drm/amdgpu: allow huge invalid mappings on GMC8

Only GMC9 supports true huge pages, but we can still free invalid mappings
on GMC8.

Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Acked-by: Huang Rui <ray.huang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amdgpu: drop the huge page flag
Christian König [Fri, 1 Feb 2019 16:11:29 +0000 (17:11 +0100)]
drm/amdgpu: drop the huge page flag

Not needed any more since we now free PDs/PTs on demand.

Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Acked-by: Huang Rui <ray.huang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amdgpu: free PDs/PTs on demand
Christian König [Fri, 1 Feb 2019 13:02:10 +0000 (14:02 +0100)]
drm/amdgpu: free PDs/PTs on demand

When something is unmapped we now free the affected PDs/PTs again.

Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Acked-by: Huang Rui <ray.huang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amdgpu: allocate VM PDs/PTs on demand
Christian König [Wed, 30 Jan 2019 15:07:29 +0000 (16:07 +0100)]
drm/amdgpu: allocate VM PDs/PTs on demand

Let's start to allocate VM PDs/PTs on demand instead of pre-allocating
them during mapping.

Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Acked-by: Huang Rui <ray.huang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amdgpu: let amdgpu_vm_clear_bo figure out ats status v2
Christian König [Thu, 16 Aug 2018 08:49:41 +0000 (10:49 +0200)]
drm/amdgpu: let amdgpu_vm_clear_bo figure out ats status v2

Instead of providing it from outside figure out the ats status in the
function itself from the data structures.

v2: simplify finding the right level
v3: partially revert changes from v2, more cleanup and split code
    into more functions.

Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Acked-by: Huang Rui <ray.huang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amdgpu: rework shadow handling during PD clear v3
Christian König [Thu, 16 Aug 2018 11:23:48 +0000 (13:23 +0200)]
drm/amdgpu: rework shadow handling during PD clear v3

This way we only deal with the real BO in here.

v2: use a do { ... } while loop instead
v3: fix NULL pointer in v2

Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Acked-by: Huang Rui <ray.huang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amd/powerplay: fix pcie sysfs interface when set wrong value
Likun Gao [Mon, 4 Mar 2019 02:46:27 +0000 (10:46 +0800)]
drm/amd/powerplay: fix pcie sysfs interface when set wrong value

The operation of mutex_unlock smu->mutex should be done when forced
level is larger than NUM_LINK_LEVELS in the function of force_clk_levels.

Reported-by: Julia Lawall <julia.lawall@lip6.fr>
Signed-off-by: Likun Gao <Likun.Gao@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 memdup.cocci warnings
kbuild test robot [Sat, 2 Mar 2019 06:50:36 +0000 (07:50 +0100)]
drm/amd/powerplay: fix memdup.cocci warnings

Simplify the code a bit by using kmemdup instead of kzalloc and memcpy.

Generated by: scripts/coccinelle/api/memdup.cocci

CC: Likun Gao <Likun.Gao@amd.com>
Reviewed-by: Likun Gao <Likun.Gao@amd.com>
Acked-by: Huang Rui <ray.huang@amd.com>
Signed-off-by: kbuild test robot <fengguang.wu@intel.com>
Signed-off-by: Julia Lawall <julia.lawall@lip6.fr>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amdgpu: fix missing assignment of error return code to variable ret
Colin Ian King [Sat, 2 Mar 2019 22:16:58 +0000 (22:16 +0000)]
drm/amdgpu: fix missing assignment of error return code to variable ret

An earlier commit replaced ttm_bo_wait with amdgpu_bo_sync_wait and
removed the error return assignment to variable ret. Fix this by adding
the assignment back. Also break line to clean up checkpatch overly
long line warning.

Detected by CoverityScan, CID#1477327 ("Logically dead code")

Fixes: c60cd590cb7d ("drm/amdgpu: Replace ttm_bo_wait with amdgpu_bo_sync_wait")
Signed-off-by: Colin Ian King <colin.king@canonical.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amdgpu: also reroute VMC and UMD to IH ring 1 on Vega 20
Christian König [Mon, 4 Mar 2019 18:34:34 +0000 (19:34 +0100)]
drm/amdgpu: also reroute VMC and UMD to IH ring 1 on Vega 20

Same patch we alredy did for Vega10. Just re-route page faults to a separate
ring to avoid drowning in interrupts.

Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amdgpu: reroute VMC and UMD to IH ring 1
Christian König [Fri, 2 Nov 2018 14:00:16 +0000 (15:00 +0100)]
drm/amdgpu: reroute VMC and UMD to IH ring 1

Page faults can easily overwhelm the interrupt handler.

So to make sure that we never lose valuable interrupts on the primary ring
we re-route page faults to IH ring 1.

Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amd/display: Fix Divide by 0 in memory calculations
Murton Liu [Fri, 15 Feb 2019 20:05:43 +0000 (15:05 -0500)]
drm/amd/display: Fix Divide by 0 in memory calculations

Check if we get any values equal to 0, and set to 1 if so.

Signed-off-by: Murton Liu <murton.liu@amd.com>
Reviewed-by: Aric Cyr <Aric.Cyr@amd.com>
Acked-by: Bhawanpreet Lakha <Bhawanpreet.Lakha@amd.com>
Acked-by: Sivapiriyan Kumarasamy <Sivapiriyan.Kumarasamy@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amdgpu: add thick tile mode settings for Oland of gfx6
Tao Zhou [Fri, 1 Mar 2019 06:01:04 +0000 (14:01 +0800)]
drm/amdgpu: add thick tile mode settings for Oland of gfx6

Adding thick tile mode for Oland to prevent UMD from getting mode value 0

Signed-off-by: Tao Zhou <tao.zhou1@amd.com>
Tested-by: Hui.Deng <hui.deng@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amdgpu/gfx_v8_0: Mark expected switch fall-through
Gustavo A. R. Silva [Fri, 1 Mar 2019 20:58:32 +0000 (14:58 -0600)]
drm/amdgpu/gfx_v8_0: Mark expected switch fall-through

In preparation to enabling -Wimplicit-fallthrough, mark switch
cases where we are expecting to fall through.

This patch fixes the following warning:

drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c: In function ‘gfx_v8_0_tiling_mode_table_init’:
./include/linux/device.h:1487:2: warning: this statement may fall through [-Wimplicit-fallthrough=]
  _dev_warn(dev, dev_fmt(fmt), ##__VA_ARGS__)
  ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c:3236:3: note: in expansion of macro ‘dev_warn’
   dev_warn(adev->dev,
   ^~~~~~~~
drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c:3240:2: note: here
  case CHIP_CARRIZO:
  ^~~~

Warning level 3 was used: -Wimplicit-fallthrough=3

This patch is part of the ongoing efforts to enable
-Wimplicit-fallthrough.

Signed-off-by: Gustavo A. R. Silva <gustavo@embeddedor.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/ttm: Fix spelling of "KiB"
Jakub Wilk [Thu, 28 Feb 2019 16:48:34 +0000 (17:48 +0100)]
drm/ttm: Fix spelling of "KiB"

The symbol for binary prefix kibi is "Ki", with uppercase K.
(In contrast, the symbol for decimal kilo is lowercase "k".)

Signed-off-by: Jakub Wilk <jwilk@jwilk.net>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amd/powerplay: fix semicolon code style issue
Yang Wei [Thu, 28 Feb 2019 14:48:04 +0000 (22:48 +0800)]
drm/amd/powerplay: fix semicolon code style issue

Delete superfluous semicolons.

Signed-off-by: Yang Wei <yang.wei9@zte.com.cn>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amd/powerplay: fix raven issue for sw smu
Huang Rui [Wed, 27 Feb 2019 10:51:01 +0000 (18:51 +0800)]
drm/amd/powerplay: fix raven issue for sw smu

Raven series doesn't support sw smu so far, so we have to add checking to
confirm it return false while the asic is raven.

Signed-off-by: Huang Rui <ray.huang@amd.com>
Reviewed-by: Likun Gao <Likun.Gao@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amdgpu: Bump amdgpu version for per-flip plane tiling updates
Nicholas Kazlauskas [Wed, 27 Feb 2019 17:56:36 +0000 (12:56 -0500)]
drm/amdgpu: Bump amdgpu version for per-flip plane tiling updates

To help xf86-video-amdgpu and mesa know DC supports updating the
tiling attributes for a framebuffer per-flip.

Cc: Michel Dänzer <michel@daenzer.net>
Signed-off-by: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com>
Acked-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Marek Olšák <marek.olsak@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amd/display: Add pp_smu null pointer check
Charlene Liu [Tue, 19 Feb 2019 17:14:19 +0000 (12:14 -0500)]
drm/amd/display: Add pp_smu null pointer check

res_pool->pp_smu may be NULL. Check before use

Signed-off-by: Charlene Liu <charlene.liu@amd.com>
Reviewed-by: Krunoslav Kovac <Krunoslav.Kovac@amd.com>
Acked-by: Bhawanpreet Lakha <Bhawanpreet.Lakha@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amd/display: 3.2.20
Aric Cyr [Fri, 15 Feb 2019 20:14:37 +0000 (15:14 -0500)]
drm/amd/display: 3.2.20

Signed-off-by: Aric Cyr <aric.cyr@amd.com>
Reviewed-by: Aric Cyr <Aric.Cyr@amd.com>
Acked-by: Bhawanpreet Lakha <Bhawanpreet.Lakha@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amd/display: Fix soft hang issue when some DPCD data invalid
Anthony Koo [Fri, 15 Feb 2019 19:19:30 +0000 (14:19 -0500)]
drm/amd/display: Fix soft hang issue when some DPCD data invalid

[Why]
AUX transaction returns success, but data has invalid lane count and rate
which when passed to VBIOS command table causes it to soft hang

[How]
Do some sanity checking and fail if the DPCD caps are invalid.

Signed-off-by: Anthony Koo <Anthony.Koo@amd.com>
Reviewed-by: Tony Cheng <Tony.Cheng@amd.com>
Acked-by: Bhawanpreet Lakha <Bhawanpreet.Lakha@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amd/display: Update plane tiling attributes for stream updates
Nicholas Kazlauskas [Mon, 28 Jan 2019 16:02:08 +0000 (11:02 -0500)]
drm/amd/display: Update plane tiling attributes for stream updates

[Why]
Tiling and DCC attributes can change when swapping framebuffers but
these will only ever get updated on full commits where
state->allow_modeset is true. But for the page-flip IOCTL
state->allow_modeset = false so these aren't updated and DCC changes
aren't being programmed.

[How]
Split out updating the tiling and DCC attributes into its own function
that's called when creating the plane for the first time but also
called when switching framebuffers before commit stream updates.

These functions also don't modify the plane state directly since the
plane state being used as reference is often the current one
being used by the hardware.

DC needs to be able to compare plane_state to the stream update
structures to determine if the hardware programming should happen. If
it's modified in place then nothing will be programmed and we can run
into concurrency issues.

Signed-off-by: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com>
Reviewed-by: Harry Wentland <Harry.Wentland@amd.com>
Acked-by: Bhawanpreet Lakha <Bhawanpreet.Lakha@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amd/display: Re-add custom degamma support
David Francis [Mon, 11 Feb 2019 19:15:19 +0000 (14:15 -0500)]
drm/amd/display: Re-add custom degamma support

[Why]
The dc_gamma_type CUSTOM_GAMMA is used to represent degamma
mappings passed in by drm. This type of gamma must be interpolated
into a transfer function by apply_1d_lut.  The line in
mod_color_calculate_degamma_params that handled this case
was erroneously removed.

[How]
For CUSTOM_GAMMA degamma, calculate the lut as before.

Signed-off-by: David Francis <David.Francis@amd.com>
Reviewed-by: Krunoslav Kovac <Krunoslav.Kovac@amd.com>
Acked-by: Bhawanpreet Lakha <Bhawanpreet.Lakha@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agoRevert "drm/amd/display: dcn add check surface in_use"
Charlene Liu [Thu, 14 Feb 2019 00:48:31 +0000 (19:48 -0500)]
Revert "drm/amd/display: dcn add check surface in_use"

This reverts commit 6bff1cc7780cca2fd2a775aa7b18b789e2a1b608.

[Description]
Revert since this will be checked at CP side.

Signed-off-by: Charlene Liu <Charlene.Liu@amd.com>
Reviewed-by: Tony Cheng <Tony.Cheng@amd.com>
Acked-by: Bhawanpreet Lakha <Bhawanpreet.Lakha@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amd/display: half bandwidth for YCbCr420 during validation
Martin Leung [Wed, 13 Feb 2019 22:06:31 +0000 (17:06 -0500)]
drm/amd/display: half bandwidth for YCbCr420 during validation

[Why]
used to be unable to run 4:2:0 if using a dongle because 4k60 bandwidth
exceeded dongle caps

[How]
half pixel clock during comparison to dongle cap. *Could get stuck on black
screen on monitor that don't support 420 but will be selecting 420 as
preferred mode*

Signed-off-by: Martin Leung <martin.leung@amd.com>
Reviewed-by: Wenjing Liu <Wenjing.Liu@amd.com>
Acked-by: Aidan Wood <Aidan.Wood@amd.com>
Acked-by: Bhawanpreet Lakha <Bhawanpreet.Lakha@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amd/display: Keep clocks high before seamless boot done
Anthony Koo [Sat, 9 Feb 2019 01:50:51 +0000 (20:50 -0500)]
drm/amd/display: Keep clocks high before seamless boot done

[Why]
UEFI boot usually uses a boot profile that uses higher clocks
and watermark settings.
UEFI boot surface is less optimal, for example it uses linear surface

[How]
Before we finish our seamless boot sequence, keep the clock and
watermark settings from boot.
Update to optimal settings only after first flip away from UEFI
frame buffer.

Signed-off-by: Anthony Koo <Anthony.Koo@amd.com>
Reviewed-by: Aric Cyr <Aric.Cyr@amd.com>
Acked-by: Bhawanpreet Lakha <Bhawanpreet.Lakha@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amd/display: Refactor pageflips plane commit
David Francis [Wed, 13 Feb 2019 18:20:11 +0000 (13:20 -0500)]
drm/amd/display: Refactor pageflips plane commit

[Why]
commit_planes is indented quite far

[How]
Move the pageflip code from an if statement to after a
continue

Signed-off-by: David Francis <David.Francis@amd.com>
Reviewed-by: Nicholas Kazlauskas <Nicholas.Kazlauskas@amd.com>
Acked-by: Bhawanpreet Lakha <Bhawanpreet.Lakha@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
5 years agodrm/amd/display: Allow pflips from a framebuffer to itself
David Francis [Wed, 6 Feb 2019 19:01:29 +0000 (14:01 -0500)]
drm/amd/display: Allow pflips from a framebuffer to itself

[Why]
IGT expects that pageflips can be triggered with the same
framebuffer before and after the commit

[How]
Expand the definition of pageflip to include any change
with an old framebuffer and a new framebuffer, even if they're
the same

Signed-off-by: David Francis <David.Francis@amd.com>
Reviewed-by: Nicholas Kazlauskas <Nicholas.Kazlauskas@amd.com>
Acked-by: Bhawanpreet Lakha <Bhawanpreet.Lakha@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>