platform/kernel/linux-rpi.git
2 years agodrm/amd/display: Fix HDCP 1.X 1A-04 failing
Bhawanpreet Lakha [Thu, 20 Oct 2022 15:46:50 +0000 (11:46 -0400)]
drm/amd/display: Fix HDCP 1.X 1A-04 failing

[Why]
On some linux based OS, the hotplug->HDCP start takes longer than
4seconds (by ~100ms) This is due to the HDCP delay (3 sec) so we only
have 1 second to hotplug->stream enablement, which is not enough for
certain OS configs.

[How]
Change the Delay to 0 seconds. From testing it seems like 0 Seconds can
pass 1.x and 2.x compliances

Tested-by: Mark Broadworth <mark.broadworth@amd.com>
Reviewed-by: Aurabindo Pillai <Aurabindo.Pillai@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Bhawanpreet Lakha <Bhawanpreet.Lakha@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2 years agodrm/amd/display: Remove optimization for VRR updates
Alvin Lee [Thu, 20 Oct 2022 15:46:49 +0000 (11:46 -0400)]
drm/amd/display: Remove optimization for VRR updates

Optimization caused unexpected regression, so remove for now.

Tested-by: Mark Broadworth <mark.broadworth@amd.com>
Reviewed-by: Aric Cyr <Aric.Cyr@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Alvin Lee <Alvin.Lee2@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2 years agodrm/amd/display: Set memclk levels to be at least 1 for dcn32
Dillon Varone [Thu, 20 Oct 2022 15:46:48 +0000 (11:46 -0400)]
drm/amd/display: Set memclk levels to be at least 1 for dcn32

[Why]
Cannot report 0 memclk levels even when SMU does not provide any.

[How]
When memclk levels reported by SMU is 0, set levels to 1.

Tested-by: Mark Broadworth <mark.broadworth@amd.com>
Reviewed-by: Martin Leung <Martin.Leung@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Dillon Varone <Dillon.Varone@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2 years agodrm/amd/display: Update latencies on DCN321
Dillon Varone [Thu, 20 Oct 2022 15:46:47 +0000 (11:46 -0400)]
drm/amd/display: Update latencies on DCN321

Update DF related latencies based on new measurements.

Tested-by: Mark Broadworth <mark.broadworth@amd.com>
Reviewed-by: Jun Lei <Jun.Lei@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Dillon Varone <Dillon.Varone@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2 years agodrm/amd/display: 3.2.208
Aric Cyr [Thu, 20 Oct 2022 15:46:46 +0000 (11:46 -0400)]
drm/amd/display: 3.2.208

DC version 3.2.208 brings along the following:

* Add more kernel doc
* Enable secure display on DCN21
* Limit dcn32 to 1950Mhz display clock
* PSR code refactor
* Rework audio stream sequence
* Generic bug fixes

Tested-by: Mark Broadworth <mark.broadworth@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Aric Cyr <Aric.Cyr@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2 years agodrm/amd/display: Document part of the DMUB cmd
Anthony Koo [Thu, 20 Oct 2022 15:46:45 +0000 (11:46 -0400)]
drm/amd/display: Document part of the DMUB cmd

Add a simple documentation in the dmub_cmd.

Tested-by: Mark Broadworth <mark.broadworth@amd.com>
Reviewed-by: Aric Cyr <Aric.Cyr@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Anthony Koo <Anthony.Koo@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2 years agodrm/amd/display: Limit dcn32 to 1950Mhz display clock
Jun Lei [Thu, 20 Oct 2022 15:46:44 +0000 (11:46 -0400)]
drm/amd/display: Limit dcn32 to 1950Mhz display clock

[why]
Hardware team recommends we limit dispclock to 1950Mhz for all DCN3.2.x

[how]
Limit to 1950 when initializing clocks.

Tested-by: Mark Broadworth <mark.broadworth@amd.com>
Reviewed-by: Alvin Lee <Alvin.Lee2@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Jun Lei <jun.lei@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2 years agodrm/amd/display: Drop struct crc_region and reuse struct rect
Alan Liu [Thu, 20 Oct 2022 15:46:43 +0000 (11:46 -0400)]
drm/amd/display: Drop struct crc_region and reuse struct rect

[Why]
reuse struct rect rather than adding a new struct.

[How]
- Userspace keeps inputting x_start, y_start, x_end, y_end
- We translate data to x, y, width, height in code flow to store
- translate back to x_start, y_start, x_end, y_end before programming HW

Tested-by: Mark Broadworth <mark.broadworth@amd.com>
Reviewed-by: Wayne Lin <Wayne.Lin@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Alan Liu <HaoPing.Liu@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2 years agodrm/amd/display: Implement secure display on DCN21
Alan Liu [Thu, 20 Oct 2022 15:46:42 +0000 (11:46 -0400)]
drm/amd/display: Implement secure display on DCN21

[Why]
Porting secure display feature from DCN10 to DCN21. Support single
display for now and will extend to multiple displays.

[How]
- use workqueue to offload works for dmub or dmcu firmware
- after receiving ROI update from userspace, set skip_frame_cnt to 1
- refactor amdgpu_dm_crtc_handle_crc_window_irq()
- disable PSR before activating secure_display on a crtc
- check if secure_display is activated before enabling psr
- only work for single display for now.

Tested-by: Mark Broadworth <mark.broadworth@amd.com>
Reviewed-by: Wayne Lin <Wayne.Lin@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Alan Liu <HaoPing.Liu@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2 years agodrm/amd/display: Ignore Cable ID Feature
Fangzhi Zuo [Thu, 20 Oct 2022 15:46:41 +0000 (11:46 -0400)]
drm/amd/display: Ignore Cable ID Feature

Ignore cable ID for DP2 receivers that does not support the feature.

Tested-by: Mark Broadworth <mark.broadworth@amd.com>
Reviewed-by: Roman Li <Roman.Li@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Fangzhi Zuo <Jerry.Zuo@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2 years agodrm/amd/display: Update DSC capabilitie for DCN314
Leo Chen [Thu, 20 Oct 2022 15:46:40 +0000 (11:46 -0400)]
drm/amd/display: Update DSC capabilitie for DCN314

dcn314 has 4 DSC - conflicted hardware document updated and confirmed.

Tested-by: Mark Broadworth <mark.broadworth@amd.com>
Reviewed-by: Charlene Liu <Charlene.Liu@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Leo Chen <sancchen@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2 years agodrm/amd/display: Revert logic for plane modifiers
Joaquín Ignacio Aramendía [Mon, 24 Oct 2022 11:33:59 +0000 (08:33 -0300)]
drm/amd/display: Revert logic for plane modifiers

This file was split in commit 5d945cbcd4b16a29d6470a80dfb19738f9a4319f
("drm/amd/display: Create a file dedicated to planes") and the logic in
dm_plane_format_mod_supported() function got changed by a switch logic.
That change broke drm_plane modifiers setting on series 5000 APUs
(tested on OXP mini AMD 5800U and HP Dev One 5850U PRO)
leading to Gamescope not working as reported on GitHub[1]

To reproduce the issue, enter a TTY and run:

$ gamescope -- vkcube

With said commit applied it will abort. This one restores the old logic,
fixing the issue that affects Gamescope.

[1](https://github.com/Plagman/gamescope/issues/624)

Cc: <stable@vger.kernel.org> # 6.0.x
Signed-off-by: Joaquín Ignacio Aramendía <samsagax@gmail.com>
Reviewed-by: Bas Nieuwenhuizen <bas@basnieuwenhuizen.nl>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2 years agodrm/radeon: fix repeated words in comments
wangjianli [Sat, 22 Oct 2022 06:04:17 +0000 (14:04 +0800)]
drm/radeon: fix repeated words in comments

Delete the redundant word 'the'.

Signed-off-by: wangjianli <wangjianli@cdjrlc.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2 years agoamd/amdgpu: fix repeated words in comments
wangjianli [Sat, 22 Oct 2022 06:01:55 +0000 (14:01 +0800)]
amd/amdgpu: fix repeated words in comments

Delete the redundant word 'the'.

Signed-off-by: wangjianli <wangjianli@cdjrlc.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2 years agodrm/amd/display: refine wake up aux in retrieve link caps
Lewis Huang [Thu, 20 Oct 2022 15:46:39 +0000 (11:46 -0400)]
drm/amd/display: refine wake up aux in retrieve link caps

[Why]
Read set_power_state dpcd after HPD cause USB4 CTS 4.2.1.1

[How]
Read LTTPR caps first. If aux channel not ready, wake up aux channel.
If wake up aux channel return pass, retrieve lttpr caps again. If wake
up aux channel return false, register a detection retry timer.

Tested-by: Mark Broadworth <mark.broadworth@amd.com>
Reviewed-by: Wenjing Liu <Wenjing.Liu@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Lewis Huang <Lewis.Huang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2 years agodrm/amd/display: Add events log to trace OPTC lock and unlock
Rodrigo Siqueira [Thu, 20 Oct 2022 15:46:38 +0000 (11:46 -0400)]
drm/amd/display: Add events log to trace OPTC lock and unlock

As an attempt to offer more DCN debug tools for cases where the OPTC can
hang, this commit introduces a trace event responsible for showing OPTC
status when it requests lock and unlock.

Tested-by: Mark Broadworth <mark.broadworth@amd.com>
Reviewed-by: Aurabindo Pillai <Aurabindo.Pillai@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2 years agodrm/amd/display: Convert documentation to a kernel-doc
Rodrigo Siqueira [Thu, 20 Oct 2022 15:46:37 +0000 (11:46 -0400)]
drm/amd/display: Convert documentation to a kernel-doc

The dc_dmub_srv file has a lot of documentation associated with SubVP
that could be converted to a kernel-doc. This commit just changes the
comment style to a kernel-doc.

Tested-by: Mark Broadworth <mark.broadworth@amd.com>
Reviewed-by: Aurabindo Pillai <Aurabindo.Pillai@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2 years agodrm/amd/display: Refactor eDP PSR codes
Ian Chen [Thu, 20 Oct 2022 15:46:36 +0000 (11:46 -0400)]
drm/amd/display: Refactor eDP PSR codes

We split out PSR config from "global" to "per-panel" config settings.

Tested-by: Mark Broadworth <mark.broadworth@amd.com>
Reviewed-by: Robin Chen <robin.chen@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Ian Chen <ian.chen@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2 years agodrm/amd/display: correctly populate dcn315 clock table
Dmytro Laktyushkin [Thu, 20 Oct 2022 15:46:35 +0000 (11:46 -0400)]
drm/amd/display: correctly populate dcn315 clock table

Fix incorrect pstate read order as well as min and max state logic.

Tested-by: Mark Broadworth <mark.broadworth@amd.com>
Reviewed-by: Charlene Liu <Charlene.Liu@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Dmytro Laktyushkin <Dmytro.Laktyushkin@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2 years agodrm/amd/display: remove audio mute control in hpo dp
Wenjing Liu [Thu, 20 Oct 2022 15:46:34 +0000 (11:46 -0400)]
drm/amd/display: remove audio mute control in hpo dp

VPG doesn't have the ability to mute audio output by sending all 0s in
audio SDP. The existing implemention is disabling audio SDP instead.
This is same as what dp_audio_enable does. Since it is no longer
referenced by any callers, we decided to remove this interface for
simplicity.

Tested-by: Mark Broadworth <mark.broadworth@amd.com>
Reviewed-by: Charlene Liu <Charlene.Liu@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Wenjing Liu <wenjing.liu@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2 years agodrm/amd/display: move stream encoder audio setup to link_hwss
Wenjing Liu [Thu, 20 Oct 2022 15:46:33 +0000 (11:46 -0400)]
drm/amd/display: move stream encoder audio setup to link_hwss

Unify stream encoder audio setup interface.

Tested-by: Mark Broadworth <mark.broadworth@amd.com>
Reviewed-by: Charlene Liu <Charlene.Liu@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Wenjing Liu <wenjing.liu@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2 years agodrm/amd/display: refactor enable/disable audio stream hw sequence
Wenjing Liu [Thu, 20 Oct 2022 15:46:32 +0000 (11:46 -0400)]
drm/amd/display: refactor enable/disable audio stream hw sequence

[why]
1. As recommended by hardware team, don't enable APG when stream is not
   enabled.
2. Move audio stream encoder programming into link_hwss.

[how]
1. Merge dp_audio_enable into enable audio stream hw sequence.
2. Move stream encoder programming into link hwss level to unify stream
   encoder programming interface.

Tested-by: Mark Broadworth <mark.broadworth@amd.com>
Reviewed-by: Charlene Liu <Charlene.Liu@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Wenjing Liu <wenjing.liu@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2 years agodrm/amd/display: Remove FPU guards from the DML folder
Rodrigo Siqueira [Thu, 20 Oct 2022 15:46:31 +0000 (11:46 -0400)]
drm/amd/display: Remove FPU guards from the DML folder

As part of the programming expectation for using DML functions, DC
requires that any DML function invoked outside DML uses:

 DC_FP_START();
 ... dml function ...
 DC_FP_END();

Additionally, all the DML functions that can be invoked outside the DML
folder call the function dc_assert_fp_enabled(), which is responsible
for triggering a warning in the case that the DML function was not
guarded by the DC_FP_START/END. For this reason, call DC_FP_START/END
inside DML is wrong, and this commit removes all of those references.

Tested-by: Mark Broadworth <mark.broadworth@amd.com>
Reviewed-by: Nevenko Stupar <Nevenko.Stupar@amd.com>
Reviewed-by: Jun Lei <Jun.Lei@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2 years agodrm/amd/display: Add UHBR135 and UHBR20 into debugfs
Fangzhi Zuo [Thu, 20 Oct 2022 15:46:30 +0000 (11:46 -0400)]
drm/amd/display: Add UHBR135 and UHBR20 into debugfs

Add support to manually force link rate to UHBR135 (0x546) and UHBR20
(0x7d0).

Tested-by: Mark Broadworth <mark.broadworth@amd.com>
Reviewed-by: Roman Li <Roman.Li@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Fangzhi Zuo <Jerry.Zuo@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2 years agodrm/amdgpu: disallow gfxoff until GC IP blocks complete s2idle resume
Prike Liang [Fri, 21 Oct 2022 02:04:40 +0000 (10:04 +0800)]
drm/amdgpu: disallow gfxoff until GC IP blocks complete s2idle resume

In the S2idle suspend/resume phase the gfxoff is keeping functional so
some IP blocks will be likely to reinitialize at gfxoff entry and that
will result in failing to program GC registers.Therefore, let disallow
gfxoff until AMDGPU IPs reinitialized completely.

Signed-off-by: Prike Liang <Prike.Liang@amd.com>
Acked-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2 years agodrm/amdgpu: skip mes self test for gc 11.0.3 in recover
YuBiao Wang [Wed, 19 Oct 2022 03:36:32 +0000 (11:36 +0800)]
drm/amdgpu: skip mes self test for gc 11.0.3 in recover

Temporary disable mes self teset for gc 11.0.3 during gpu_recovery.

Signed-off-by: YuBiao Wang <YuBiao.Wang@amd.com>
Acked-by: Luben Tuikov <luben.tuikov@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2 years agodrm/amdkfd: introduce dummy cache info for property asic
Prike Liang [Fri, 21 Oct 2022 20:38:48 +0000 (16:38 -0400)]
drm/amdkfd: introduce dummy cache info for property asic

This dummy cache info will enable kfd base function support.

Signed-off-by: Prike Liang <Prike.Liang@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2 years agodrm/amdkfd: correct the cache info for gfx1036
Jesse Zhang [Tue, 11 Oct 2022 05:23:10 +0000 (05:23 +0000)]
drm/amdkfd: correct the cache info for gfx1036

correct the cache information for gfx1036

Acked-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Yifan Zhang <yifan1.zhang@amd.com>
Signed-off-by: Yifan Zhang <yifan1.zhang@amd.com>
Signed-off-by: Jesse Zhang <jesse.zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2 years agodrm/amdkfd: update gfx1037 Lx cache setting
Prike Liang [Thu, 20 Oct 2022 06:44:26 +0000 (14:44 +0800)]
drm/amdkfd: update gfx1037 Lx cache setting

Update the gfx1037 L1/L2 cache setting.

Signed-off-by: Prike Liang <Prike.Liang@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2 years agodrm/amd: Add IMU fw version to fw version queries
David Francis [Fri, 16 Sep 2022 15:15:01 +0000 (11:15 -0400)]
drm/amd: Add IMU fw version to fw version queries

IMU is a new firmware for GFX11.

There are four means by which firmware version can be queried
from the driver: device attributes, vf2pf, debugfs,
and the AMDGPU_INFO_FW_VERSION option in the amdgpu info ioctl.

Add IMU as an option for those four methods.

V2: Added debugfs

Reviewed-by: Likun Gao <Likun.Gao@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: David Francis <David.Francis@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2 years agodrm/amdkfd: use vma_lookup() instead of find_vma()
Deming Wang [Fri, 7 Oct 2022 02:26:51 +0000 (22:26 -0400)]
drm/amdkfd: use vma_lookup() instead of find_vma()

Using vma_lookup() verifies the start address is contained in the found
vma.  This results in easier to read the code.

Signed-off-by: Deming Wang <wangdeming@inspur.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2 years agodrm/amd/display: make dcn32_mpc_funcs static
ruanjinjie [Wed, 19 Oct 2022 06:23:47 +0000 (14:23 +0800)]
drm/amd/display: make dcn32_mpc_funcs static

The symbol is not used outside of the file, so mark it static.

Fixes the following warning:

drivers/gpu/drm/amd/amdgpu/../display/dc/dcn32/dcn32_mpc.c:985:24: warning:
symbol 'dcn32_mpc_funcs' was not declared. Should it be static?

Reviewed-by: Harry Wentland <harry.wentland@amd.com>
Signed-off-by: ruanjinjie <ruanjinjie@huawei.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2 years agodrm/amd/display: make dcn32_mmhubbub_funcs static
ruanjinjie [Wed, 19 Oct 2022 06:31:15 +0000 (14:31 +0800)]
drm/amd/display: make dcn32_mmhubbub_funcs static

The symbol is not used outside of the file, so mark it static.

Fixes the following warning:

drivers/gpu/drm/amd/amdgpu/../display/dc/dcn32/dcn32_mmhubbub.c:214:28:
warning: symbol 'dcn32_mmhubbub_funcs' was
not declared. Should it be static?

Reviewed-by: Harry Wentland <harry.wentland@amd.com>
Signed-off-by: ruanjinjie <ruanjinjie@huawei.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2 years agodrm/amd/display: Make some symbols static
ruanjinjie [Wed, 19 Oct 2022 07:45:02 +0000 (15:45 +0800)]
drm/amd/display: Make some symbols static

These symbols qp_table_422_10bpc_min, qp_table_444_8bpc_max,
qp_table_420_12bpc_max, qp_table_444_10bpc_min, qp_table_420_8bpc_max,
qp_table_444_8bpc_min, qp_table_444_12bpc_min, qp_table_420_12bpc_min,
qp_table_422_12bpc_min, qp_table_422_12bpc_max, qp_table_444_12bpc_max,
qp_table_420_8bpc_min, qp_table_422_8bpc_min, qp_table_422_10bpc_max,
qp_table_420_10bpc_max, qp_table_420_10bpc_min, qp_table_444_10bpc_max,
qp_table_422_8bpc_max are not used outside of the file,
so mark them static.

./drivers/gpu/drm/amd/amdgpu/../display/dc/dml/dsc/qp_tables.h:28:18: warning:
symbol 'qp_table_422_10bpc_min' was not declared. Should it be static?

./drivers/gpu/drm/amd/amdgpu/../display/dc/dml/dsc/qp_tables.h:61:18: warning:
symbol 'qp_table_444_8bpc_max' was not declared. Should it be static?

./drivers/gpu/drm/amd/amdgpu/../display/dc/dml/dsc/qp_tables.h:102:18: warning:
symbol 'qp_table_420_12bpc_max' was not declared. Should it be static?

./drivers/gpu/drm/amd/amdgpu/../display/dc/dml/dsc/qp_tables.h:135:18: warning:
symbol 'qp_table_444_10bpc_min' was not declared. Should it be static?

./drivers/gpu/drm/amd/amdgpu/../display/dc/dml/dsc/qp_tables.h:188:18: warning:
symbol 'qp_table_420_8bpc_max' was not declared. Should it be static?

./drivers/gpu/drm/amd/amdgpu/../display/dc/dml/dsc/qp_tables.h:209:18: warning:
symbol 'qp_table_444_8bpc_min' was not declared. Should it be static?

./drivers/gpu/drm/amd/amdgpu/../display/dc/dml/dsc/qp_tables.h:250:18: warning:
symbol 'qp_table_444_12bpc_min' was not declared. Should it be static?

./drivers/gpu/drm/amd/amdgpu/../display/dc/dml/dsc/qp_tables.h:315:18: warning:
symbol 'qp_table_420_12bpc_min' was not declared. Should it be static?

./drivers/gpu/drm/amd/amdgpu/../display/dc/dml/dsc/qp_tables.h:348:18: warning:
symbol 'qp_table_422_12bpc_min' was not declared. Should it be static?

./drivers/gpu/drm/amd/amdgpu/../display/dc/dml/dsc/qp_tables.h:389:18: warning:
symbol 'qp_table_422_12bpc_max' was not declared. Should it be static?

./drivers/gpu/drm/amd/amdgpu/../display/dc/dml/dsc/qp_tables.h:430:18: warning:
symbol 'qp_table_444_12bpc_max' was not declared. Should it be static?

./drivers/gpu/drm/amd/amdgpu/../display/dc/dml/dsc/qp_tables.h:495:18: warning:
symbol 'qp_table_420_8bpc_min' was not declared. Should it be static?

./drivers/gpu/drm/amd/amdgpu/../display/dc/dml/dsc/qp_tables.h:516:18: warning:
symbol 'qp_table_422_8bpc_min' was not declared. Should it be static?

./drivers/gpu/drm/amd/amdgpu/../display/dc/dml/dsc/qp_tables.h:541:18: warning:
symbol 'qp_table_422_10bpc_max' was not declared. Should it be static?

./drivers/gpu/drm/amd/amdgpu/../display/dc/dml/dsc/qp_tables.h:574:16: warning:
symbol 'qp_table_420_10bpc_max' was not declared. Should it be static?

./drivers/gpu/drm/amd/amdgpu/../display/dc/dml/dsc/qp_tables.h:601:18: warning:
symbol 'qp_table_420_10bpc_min' was not declared. Should it be static?

./drivers/gpu/drm/amd/amdgpu/../display/dc/dml/dsc/qp_tables.h:628:18: warning:
symbol 'qp_table_444_10bpc_max' was not declared. Should it be static?

./drivers/gpu/drm/amd/amdgpu/../display/dc/dml/dsc/qp_tables.h:681:18: warning:
symbol 'qp_table_422_8bpc_max' was not declared. Should it be static?

Signed-off-by: ruanjinjie <ruanjinjie@huawei.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2 years agodrm/amdgpu: fix sdma doorbell init ordering on APUs
Alex Deucher [Wed, 19 Oct 2022 20:57:42 +0000 (16:57 -0400)]
drm/amdgpu: fix sdma doorbell init ordering on APUs

Commit 8795e182b02d ("PCI/portdrv: Don't disable AER reporting in get_port_device_capability()")
uncovered a bug in amdgpu that required a reordering of the driver
init sequence to avoid accessing a special register on the GPU
before it was properly set up leading to an PCI AER error.  This
reordering uncovered a different hw programming ordering dependency
in some APUs where the SDMA doorbells need to be programmed before
the GFX doorbells. To fix this, move the SDMA doorbell programming
back into the soc15 common code, but use the actual doorbell range
values directly rather than the values stored in the ring structure
since those will not be initialized at this point.

This is a partial revert, but with the doorbell assignment
fixed so the proper doorbell index is set before it's used.

Fixes: e3163bc8ffdfdb ("drm/amdgpu: move nbio sdma_doorbell_range() into sdma code for vega")
Acked-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Cc: skhan@linuxfoundation.org
2 years agodrm/amd/display: Don't return false if no stream
Alvin Lee [Thu, 6 Oct 2022 21:26:49 +0000 (17:26 -0400)]
drm/amd/display: Don't return false if no stream

pipe_ctx[i] exists even if the pipe is not
in use. If the pipe is not in use it will
always have a null stream, so don't return
false in this case.

Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Reviewed-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Acked-by: Qingqing Zhuo <qingqing.zhuo@amd.com>
Signed-off-by: Alvin Lee <Alvin.Lee2@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2 years agodrm/amd/display: Remove wrong pipe control lock
Rodrigo Siqueira [Thu, 6 Oct 2022 21:26:48 +0000 (17:26 -0400)]
drm/amd/display: Remove wrong pipe control lock

When using a device based on DCN32/321,
we have an issue where a second
4k@60Hz display does not light up,
and the system becomes unresponsive
for a few minutes. In the debug process,
it was possible to see a hang
in the function dcn20_post_unlock_program_front_end
in this part:

for (j = 0; j < TIMEOUT_FOR_PIPE_ENABLE_MS*1000
&& hubp->funcs->hubp_is_flip_pending(hubp); j++)
mdelay(1);
}

The hubp_is_flip_pending always returns positive
for waiting pending flips which is a symptom of
pipe hang. Additionally, the dmesg log shows
this message after a few minutes:

  BUG: soft lockup - CPU#4 stuck for 26s!
  ...
  [  +0.000003]  dcn20_post_unlock_program_front_end+0x112/0x340 [amdgpu]
  [  +0.000171]  dc_commit_state_no_check+0x63d/0xbf0 [amdgpu]
  [  +0.000155]  ? dc_validate_global_state+0x358/0x3d0 [amdgpu]
  [  +0.000154]  dc_commit_state+0xe2/0xf0 [amdgpu]

This confirmed the hypothesis that we had a pipe
hanging somewhere. Next, after checking the
ftrace entries, we have the below weird
sequence:

 [..]
  2)               |        dcn10_lock_all_pipes [amdgpu]() {
  2)   0.120 us    |          optc1_is_tg_enabled [amdgpu]();
  2)               |          dcn20_pipe_control_lock [amdgpu]() {
  2)               |            dc_dmub_srv_clear_inbox0_ack [amdgpu]() {
  2)   0.121 us    |              amdgpu_dm_dmub_reg_write [amdgpu]();
  2)   0.551 us    |            }
  2)               |            dc_dmub_srv_send_inbox0_cmd [amdgpu]() {
  2)   0.110 us    |              amdgpu_dm_dmub_reg_write [amdgpu]();
  2)   0.511 us    |            }
  2)               |            dc_dmub_srv_wait_for_inbox0_ack [amdgpu]() {
  2)   0.110 us    |              amdgpu_dm_dmub_reg_read [amdgpu]();
  2)   0.110 us    |              amdgpu_dm_dmub_reg_read [amdgpu]();
  2)   0.110 us    |              amdgpu_dm_dmub_reg_read [amdgpu]();
  2)   0.110 us    |              amdgpu_dm_dmub_reg_read [amdgpu]();
  2)   0.110 us    |              amdgpu_dm_dmub_reg_read [amdgpu]();
  2)   0.110 us    |              amdgpu_dm_dmub_reg_read [amdgpu]();
  2)   0.110 us    |              amdgpu_dm_dmub_reg_read [amdgpu]();
 [..]

We are not expected to read from dmub register
so many times and for so long. From the trace log,
it was possible to identify that the function
dcn20_pipe_control_lock was triggering the dmub
operation when it was unnecessary and causing
the hang issue. This commit drops the unnecessary
dmub code and, consequently, fixes the second display not
lighting up the issue.

Tested-by: Daniel Wheeler <daniel.wheeler@amd.com>
Acked-by: Qingqing Zhuo <qingqing.zhuo@amd.com>
Signed-off-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2 years agodrm/amd/pm: allow gfxoff on gc_11_0_3
Kenneth Feng [Thu, 20 Oct 2022 07:25:25 +0000 (15:25 +0800)]
drm/amd/pm: allow gfxoff on gc_11_0_3

allow gfxoff on gc_11_0_3

Signed-off-by: Kenneth Feng <kenneth.feng@amd.com>
Reviewed-by: Yang Wang <kevinyang.wang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2 years agodrm/amd/display: Use update plane and stream routine for DCN32x
Rodrigo Siqueira [Tue, 18 Oct 2022 12:28:25 +0000 (08:28 -0400)]
drm/amd/display: Use update plane and stream routine for DCN32x

Sub-viewport (Subvp) feature is used for changing MCLK without causing
any display artifact, requiring special treatment from the plane and
stream perspective since DC needs to read data from the cache when using
subvp. However, the function dc_commit_updates_for_stream does not
provide all the support needed by this feature which will make this
function legacy at some point. For this reason, this commit enables
dc_update_planes_and_stream for ASICs that support this feature but
preserves the old behavior for other ASICs. However,
dc_update_planes_and_stream should replace dc_commit_updates_for_stream
for all ASICs since it does most of the tasks executed by
dc_commit_updates_for_stream with other extra operations, but we need to
run tests before making this change.

Reviewed-by: Harry Wentland <harry.wentland@amd.com>
Signed-off-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Co-developed-by: Aurabindo Pillai <aurabindo.pillai@amd.com>
Signed-off-by: Aurabindo Pillai <aurabindo.pillai@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2 years agodrm/amd/display: Avoid ABM when ODM combine is enabled for eDP
Rodrigo Siqueira [Tue, 18 Oct 2022 12:28:24 +0000 (08:28 -0400)]
drm/amd/display: Avoid ABM when ODM combine is enabled for eDP

ODM to combine on the eDP panel with ABM causes the color difference to
the panel since the ABM module only sets one pipe. Hence, this commit
blocks ABM in case of ODM combined on eDP.

Reviewed-by: Harry Wentland <harry.wentland@amd.com>
Signed-off-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Co-developed-by: Aurabindo Pillai <aurabindo.pillai@amd.com>
Signed-off-by: Aurabindo Pillai <aurabindo.pillai@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2 years agodrm/amd/display: Update OTG instance in the commit stream
Rodrigo Siqueira [Tue, 18 Oct 2022 12:28:23 +0000 (08:28 -0400)]
drm/amd/display: Update OTG instance in the commit stream

OTG instance is not updated in dc_commit_state_no_check for newly
committed streams because mode_change is not set. Notice that OTG update
is part of the software state, and after hardware programming, it must
be updated; for this reason, this commit updates the OTG offset right
after hardware programming.

Reviewed-by: Harry Wentland <harry.wentland@amd.com>
Signed-off-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Co-developed-by: Aurabindo Pillai <aurabindo.pillai@amd.com>
Signed-off-by: Aurabindo Pillai <aurabindo.pillai@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2 years agodrm/amd/display: Handle seamless boot stream
Rodrigo Siqueira [Tue, 18 Oct 2022 12:28:22 +0000 (08:28 -0400)]
drm/amd/display: Handle seamless boot stream

A seamless boot stream has hardware resources assigned to it, and adding
a new stream means rebuilding the current assignment. It is desirable to
avoid this situation since it may cause light-up issues on the VGA
monitor on USB-C. This commit swaps the seamless boot stream to pipe 0
(if necessary) to ensure that the pipe context matches.

Reviewed-by: Harry Wentland <harry.wentland@amd.com>
Signed-off-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Co-developed-by: Aurabindo Pillai <aurabindo.pillai@amd.com>
Signed-off-by: Aurabindo Pillai <aurabindo.pillai@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2 years agodrm/amd/display: Add function for validate and update new stream
Rodrigo Siqueira [Tue, 18 Oct 2022 12:28:21 +0000 (08:28 -0400)]
drm/amd/display: Add function for validate and update new stream

DC stream can be seen as a representation of the DCN backend or the data
struct that represents the center of the display pipeline. The front end
(i.e., planes) is connected to the DC stream, and in its turn, streams
are connected to the DC link. Due to this dynamic, DC must handle the
following scenarios:

1. A stream is removed;
2. A new stream is created;
3. An unchanged stream had some updates on its planes.

These combinations require that the new stream data struct become
updated and has a valid global state. For handling multiple corner cases
associated with stream operations, this commit introduces a function
dedicated to manipulating stream changes and invokes the state
validation function after that.

Reviewed-by: Harry Wentland <harry.wentland@amd.com>
Signed-off-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Co-developed-by: Aurabindo Pillai <aurabindo.pillai@amd.com>
Signed-off-by: Aurabindo Pillai <aurabindo.pillai@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2 years agodrm/amd/display: Handle virtual hardware detect
Rodrigo Siqueira [Tue, 18 Oct 2022 12:28:20 +0000 (08:28 -0400)]
drm/amd/display: Handle virtual hardware detect

If virtual hardware is detected, there is no reason to run the full
dc_commit_streams process, and DC can return true immediately.

Reviewed-by: Harry Wentland <harry.wentland@amd.com>
Signed-off-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Co-developed-by: Aurabindo Pillai <aurabindo.pillai@amd.com>
Signed-off-by: Aurabindo Pillai <aurabindo.pillai@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2 years agodrm/amd/display: Include surface of unaffected streams
Rodrigo Siqueira [Tue, 18 Oct 2022 12:28:19 +0000 (08:28 -0400)]
drm/amd/display: Include surface of unaffected streams

The commit stream function does not include surfaces of unaffected
streams, which may lead to some blank screens during mode change in some
edge cases. This commit adds surfaces of unaffected streams followed by
kernel-doc for documenting some of the fields that participate in this
change.

v2: squash in kerneldoc warning fix (Alex)

Reviewed-by: Harry Wentland <harry.wentland@amd.com>
Signed-off-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Co-developed-by: Aurabindo Pillai <aurabindo.pillai@amd.com>
Signed-off-by: Aurabindo Pillai <aurabindo.pillai@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2 years agodrm/amd/display: Copy DC context in the commit streams
Rodrigo Siqueira [Tue, 18 Oct 2022 12:28:18 +0000 (08:28 -0400)]
drm/amd/display: Copy DC context in the commit streams

DC adds an instance of DML (which contains VBA) to each context, and
multiple threads might write back to the global VBA resulting in data
overwriting. To keep the consistency with other parts of the DC code,
this commit changes dc_commit_streams to copy the current DC state, and
as a result, it also changes the function signature to expect streams
instead of a context.

Reviewed-by: Harry Wentland <harry.wentland@amd.com>
Signed-off-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Co-developed-by: Aurabindo Pillai <aurabindo.pillai@amd.com>
Signed-off-by: Aurabindo Pillai <aurabindo.pillai@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2 years agodrm/amd/display: Enable new commit sequence only for DCN32x
Rodrigo Siqueira [Tue, 18 Oct 2022 12:28:17 +0000 (08:28 -0400)]
drm/amd/display: Enable new commit sequence only for DCN32x

Change commit sequence will impact all ASICs. It is prudent to run this
update in small steps to keep issues under control and avoid any
potential regression. With this idea in mind, this commit is preparation
work for the complete transition to the new commit sequence. To maintain
this change manageable across multiple ASICs, this commit adds a new
function named dc_commit_streams which is a copy of the dc_commit_state
with some minor changes. Finally, inside the dc_commit_state, we check
if we are using DCN32x or above and enable the new sequence only for
those devices.

Reviewed-by: Harry Wentland <harry.wentland@amd.com>
Signed-off-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Co-developed-by: Aurabindo Pillai <aurabindo.pillai@amd.com>
Signed-off-by: Aurabindo Pillai <aurabindo.pillai@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2 years agodrm/amd/display: Rework context change check
Rodrigo Siqueira [Tue, 18 Oct 2022 12:28:16 +0000 (08:28 -0400)]
drm/amd/display: Rework context change check

Context change is all about streams; for this reason, this commit
renames context_changed to streams_changed. Additionally, to make this
function more flexible, this commit changes the function signature to
receive the stream array and the stream count as a parameter.

Reviewed-by: Harry Wentland <harry.wentland@amd.com>
Signed-off-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Co-developed-by: Aurabindo Pillai <aurabindo.pillai@amd.com>
Signed-off-by: Aurabindo Pillai <aurabindo.pillai@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2 years agodrm/amd/display: Check if link state is valid
Rodrigo Siqueira [Tue, 18 Oct 2022 12:28:15 +0000 (08:28 -0400)]
drm/amd/display: Check if link state is valid

The link state is set to false if there is no link and local sink. Even
though the stream state may not change, it is desirable to commit the
new stream when HPD goes low to high.

Reviewed-by: Harry Wentland <harry.wentland@amd.com>
Signed-off-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Co-developed-by: Aurabindo Pillai <aurabindo.pillai@amd.com>
Signed-off-by: Aurabindo Pillai <aurabindo.pillai@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2 years agodrm/amdkfd: Fix memory leak in kfd_mem_dmamap_userptr()
Rafael Mendonca [Tue, 18 Oct 2022 01:27:50 +0000 (22:27 -0300)]
drm/amdkfd: Fix memory leak in kfd_mem_dmamap_userptr()

If the number of pages from the userptr BO differs from the SG BO then the
allocated memory for the SG table doesn't get freed before returning
-EINVAL, which may lead to a memory leak in some error paths. Fix this by
checking the number of pages before allocating memory for the SG table.

Fixes: 264fb4d332f5 ("drm/amdgpu: Add multi-GPU DMA mapping helpers")
Signed-off-by: Rafael Mendonca <rafaelmendsr@gmail.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2 years agodrm/amdgpu: Remove ATC L2 access for MMHUB 2.1.x
Lijo Lazar [Fri, 30 Sep 2022 05:13:08 +0000 (10:43 +0530)]
drm/amdgpu: Remove ATC L2 access for MMHUB 2.1.x

MMHUB 2.1.x versions don't have ATCL2. Remove accesses to ATCL2 registers.

Since they are non-existing registers, read access will cause a
'Completer Abort' and gets reported when AER is enabled with the below patch.
Tagging with the patch so that this is backported along with it.

v2: squash in uninitialized warning fix (Nathan Chancellor)

Fixes: 8795e182b02d ("PCI/portdrv: Don't disable AER reporting in get_port_device_capability()")

Signed-off-by: Lijo Lazar <lijo.lazar@amd.com>
Reviewed-by: Guchun Chen <guchun.chen@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Cc: stable@vger.kernel.org
2 years agodrm/amdgpu: Adjust MES polling timeout for sriov
Yiqing Yao [Tue, 18 Oct 2022 07:17:59 +0000 (15:17 +0800)]
drm/amdgpu: Adjust MES polling timeout for sriov

[why]
MES response time in sriov may be longer than default value
due to reset or init in other VF. A timeout value specific
to sriov is needed.

[how]
When in sriov, adjust the timeout value to calculated
worst case scenario.

Signed-off-by: Yiqing Yao <yiqing.yao@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2 years agodrm/amd/pm: update driver-if header for smu_v13_0_10
Kenneth Feng [Wed, 19 Oct 2022 03:24:05 +0000 (11:24 +0800)]
drm/amd/pm: update driver-if header for smu_v13_0_10

update driver-if header for smu_v13_0_10 and merge with smu_v13_0_0

Signed-off-by: Kenneth Feng <kenneth.feng@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2 years agodrm/amdgpu: fix pstate setting issue
Chengming Gui [Tue, 18 Oct 2022 09:31:38 +0000 (17:31 +0800)]
drm/amdgpu: fix pstate setting issue

[WHY]
0, original pstate X
1, ctx_A_create -> ctx_A->stable_pstate = X
2, ctx_A_set_pstate (Y) -> current pstate is Y (PEAK or STANDARD)
3, ctx_B_create -> ctx_B->stable_pstate =  Y
4, ctx_A_destroy -> restore pstate to X
5, ctx_B_destroy -> restore pstate to Y
Above sequence will cause final pstate is wrong (Y), should be original X.

[HOW]
When ctx_B create,
if  ctx_A touched pstate setting
(not auto, stable_pstate_ctx != NULL),
set ctx_B->stable_pstate the same value as ctx_A saved,
if stable_pstate_ctx == NULL,
fetch current pstate to fill
ctx_B->stable_pstate.

Signed-off-by: Chengming Gui <Jack.Gui@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2 years agodrm/amdgpu/powerplay/psm: Fix memory leak in power state init
Rafael Mendonca [Tue, 18 Oct 2022 00:53:57 +0000 (21:53 -0300)]
drm/amdgpu/powerplay/psm: Fix memory leak in power state init

Commit 902bc65de0b3 ("drm/amdgpu/powerplay/psm: return an error in power
state init") made the power state init function return early in case of
failure to get an entry from the powerplay table, but it missed to clean up
the allocated memory for the current power state before returning.

Fixes: 902bc65de0b3 ("drm/amdgpu/powerplay/psm: return an error in power state init")
Reviewed-by: Evan Quan <evan.quan@amd.com>
Signed-off-by: Rafael Mendonca <rafaelmendsr@gmail.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2 years agodrm/amdgpu: Fix for BO move issue
Arunpravin Paneer Selvam [Tue, 18 Oct 2022 14:08:38 +0000 (07:08 -0700)]
drm/amdgpu: Fix for BO move issue

A user reported a bug on CAPE VERDE system where uvd_v3_1
IP component failed to initialize as there is an issue with
BO move code from one memory to other.

In function amdgpu_mem_visible() called by amdgpu_bo_move(),
when there are no blocks to compare or if we have a single
block then break the loop.

Fixes: 312b4dc11d4f ("drm/amdgpu: Fix VRAM BO swap issue")
Signed-off-by: Arunpravin Paneer Selvam <Arunpravin.PaneerSelvam@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2 years agodrm/amdgpu: dequeue mes scheduler during fini
YuBiao Wang [Thu, 13 Oct 2022 03:31:55 +0000 (11:31 +0800)]
drm/amdgpu: dequeue mes scheduler during fini

[Why]
If mes is not dequeued during fini, mes will be in an uncleaned state
during reload, then mes couldn't receive some commands which leads to
reload failure.

[How]
Perform MES dequeue via MMIO after all the unmap jobs are done by mes
and before kiq fini.

v2: Move the dequeue operation inside kiq_hw_fini.

Signed-off-by: YuBiao Wang <YuBiao.Wang@amd.com>
Reviewed-by: Jack Xiao <Jack.Xiao@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2 years agodrm/amd/pm: enable thermal alert on smu_v13_0_10
Kenneth Feng [Fri, 14 Oct 2022 07:19:51 +0000 (15:19 +0800)]
drm/amd/pm: enable thermal alert on smu_v13_0_10

enable thermal alert on smu_v13_0_10

Signed-off-by: Kenneth Feng <kenneth.feng@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2 years agodrm/amdgpu: Program GC registers through RLCG interface in gfx_v11/gmc_v11
Yifan Zha [Wed, 7 Sep 2022 06:13:02 +0000 (14:13 +0800)]
drm/amdgpu: Program GC registers through RLCG interface in gfx_v11/gmc_v11

[Why]
L1 blocks most of GC registers accessing by MMIO.

[How]
Use RLCG interface to program GC registers under SRIOV VF in full access time.

Signed-off-by: Yifan Zha <Yifan.Zha@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Acked-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2 years agodrm/amd/display: add an ASSERT() to irq service functions
Hamza Mahfooz [Fri, 14 Oct 2022 14:29:46 +0000 (10:29 -0400)]
drm/amd/display: add an ASSERT() to irq service functions

Currently, if we encounter unimplemented functions, it is difficult to
tell what caused them just by looking at dmesg and that is compounded by
the fact that it is often hard to reproduce said issues, for instance we
have had reports of this condition being triggered when removing a
secondary display that is setup in mirror mode and is connected using
usb-c. So, to have access to more detailed debugging information, add an
ASSERT() to dal_irq_service_ack() and dal_irq_service_set() that only
triggers when we encounter an unimplemented function.

Reviewed-by: Harry Wentland <harry.wentland@amd.com>
Signed-off-by: Hamza Mahfooz <hamza.mahfooz@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2 years agodrm/amdkfd: Fix type of reset_type parameter in hqd_destroy() callback
Nathan Chancellor [Mon, 17 Oct 2022 16:28:38 +0000 (09:28 -0700)]
drm/amdkfd: Fix type of reset_type parameter in hqd_destroy() callback

When booting a kernel compiled with CONFIG_CFI_CLANG on a machine with
an RX 6700 XT, there is a CFI failure in kfd_destroy_mqd_cp():

  [   12.894543] CFI failure at kfd_destroy_mqd_cp+0x2a/0x40 [amdgpu] (target: hqd_destroy_v10_3+0x0/0x260 [amdgpu]; expected type: 0x8594d794)

Clang's kernel Control Flow Integrity (kCFI) makes sure that all
indirect call targets have a type that exactly matches the function
pointer prototype. In this case, hqd_destroy()'s third parameter,
reset_type, should have a type of 'uint32_t' but every implementation of
this callback has a third parameter type of 'enum kfd_preempt_type'.

Update the function pointer prototype to match reality so that there is
no more CFI violation.

Link: https://github.com/ClangBuiltLinux/linux/issues/1738
Signed-off-by: Nathan Chancellor <nathan@kernel.org>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2 years agodrm/amd/amdgpu: Replace kmap() with kmap_local_page()
Fabio M. De Francesco [Sun, 16 Oct 2022 17:41:58 +0000 (19:41 +0200)]
drm/amd/amdgpu: Replace kmap() with kmap_local_page()

kmap() is being deprecated in favor of kmap_local_page().

There are two main problems with kmap(): (1) It comes with an overhead as
mapping space is restricted and protected by a global lock for
synchronization and (2) it also requires global TLB invalidation when the
kmap’s pool wraps and it might block when the mapping space is fully
utilized until a slot becomes available.

With kmap_local_page() the mappings are per thread, CPU local, can take
page faults, and can be called from any context (including interrupts).
It is faster than kmap() in kernels with HIGHMEM enabled. Furthermore,
the tasks can be preempted and, when they are scheduled to run again, the
kernel virtual addresses are restored and are still valid.

Since its use in amdgpu/amdgpu_ttm.c is safe, it should be preferred.

Therefore, replace kmap() with kmap_local_page() in amdgpu/amdgpu_ttm.c.

Suggested-by: Ira Weiny <ira.weiny@intel.com>
Acked-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Fabio M. De Francesco <fmdefrancesco@gmail.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2 years agodrm/radeon: Replace kmap() with kmap_local_page()
Fabio M. De Francesco [Thu, 13 Oct 2022 21:07:14 +0000 (23:07 +0200)]
drm/radeon: Replace kmap() with kmap_local_page()

The use of kmap() is being deprecated in favor of kmap_local_page().

There are two main problems with kmap(): (1) It comes with an overhead as
the mapping space is restricted and protected by a global lock for
synchronization and (2) it also requires global TLB invalidation when the
kmap’s pool wraps and it might block when the mapping space is fully
utilized until a slot becomes available.

With kmap_local_page() the mappings are per thread, CPU local, can take
page faults, and can be called from any context (including interrupts).
It is faster than kmap() in kernels with HIGHMEM enabled. Furthermore,
the tasks can be preempted and, when they are scheduled to run again, the
kernel virtual addresses are restored and still valid.

Therefore, replace kmap() with kmap_local_page() in radeon_ttm_gtt_read().

Cc: "Venkataramanan, Anirudh" <anirudh.venkataramanan@intel.com>
Reviewed-by: Kees Cook <keescook@chromium.org>
Reviewed-by: Christian König <christian.koenig@amd.com>
Suggested-by: Ira Weiny <ira.weiny@intel.com>
Signed-off-by: Fabio M. De Francesco <fmdefrancesco@gmail.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2 years agodrm/amd/display: Increase frame size limit for display_mode_vba_util_32.o
Guenter Roeck [Thu, 13 Oct 2022 18:25:23 +0000 (11:25 -0700)]
drm/amd/display: Increase frame size limit for display_mode_vba_util_32.o

Building 32-bit images may fail with the following error.

drivers/gpu/drm/amd/amdgpu/../display/dc/dml/dcn32/display_mode_vba_util_32.c:
In function ‘dml32_UseMinimumDCFCLK’:
drivers/gpu/drm/amd/amdgpu/../display/dc/dml/dcn32/display_mode_vba_util_32.c:3142:1:
error: the frame size of 1096 bytes is larger than 1024 bytes

This is seen when building i386:allmodconfig with any of the following
compilers.

gcc (Debian 12.2.0-3) 12.2.0
gcc (Ubuntu 9.4.0-1ubuntu1~20.04.1) 9.4.0

The problem is not seen if the compiler supports GCC_PLUGIN_LATENT_ENTROPY
because in that case CONFIG_FRAME_WARN is already set to 2048 even for
32-bit builds.

dml32_UseMinimumDCFCLK() was introduced with commit dda4fb85e433
("drm/amd/display: DML changes for DCN32/321"). It declares a large
number of local variables. Increase the frame size for the affected
file to 2048, similar to other files in the same directory, to enable
32-bit build tests with affected compilers.

Fixes: dda4fb85e433 ("drm/amd/display: DML changes for DCN32/321")
Cc: Aurabindo Pillai <aurabindo.pillai@amd.com>
Reported-by: Łukasz Bartosik <ukaszb@google.com>
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2 years agodrm/amdgpu: move convert_error_address out of umc_ras
Hawking Zhang [Fri, 14 Oct 2022 07:17:43 +0000 (15:17 +0800)]
drm/amdgpu: move convert_error_address out of umc_ras

RAS error address translation algorithm is common
across dGPU and A + A platform as along as the SOC
integrates the same generation of UMC IP.

UMC RAS is managed by x86 MCA on A + A platform,
umc_ras in GPU driver is not initialized at all on
A + A platform. In such case, any umc_ras callback
implemented for dGPU config shouldn't be invoked
from A + A specific callback.

The change moves convert_error_address out of dGPU
umc_ras structure and makes it share between A + A
and dGPU config.

Signed-off-by: Hawking Zhang <Hawking.Zhang@amd.com>
Reviewed-by: Stanley Yang <Stanley.Yang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2 years agodrm/amd/pm: add SMU IP v13.0.4 IF version define to V7
Tim Huang [Thu, 29 Sep 2022 07:06:47 +0000 (15:06 +0800)]
drm/amd/pm: add SMU IP v13.0.4 IF version define to V7

The pmfw has changed the driver interface version, so keep same with the
fw.

Signed-off-by: Tim Huang <tim.huang@amd.com>
Reviewed-by: Mario Limonciello <mario.limonciello@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Cc: stable@vger.kernel.org # 6.0.x
2 years agodrm/amd/pm: update SMU IP v13.0.4 driver interface version
Tim Huang [Thu, 29 Sep 2022 06:39:21 +0000 (14:39 +0800)]
drm/amd/pm: update SMU IP v13.0.4 driver interface version

Update the SMU driver interface version to V7.

Signed-off-by: Tim Huang <tim.huang@amd.com>
Reviewed-by: Mario Limonciello <mario.limonciello@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Cc: stable@vger.kernel.org # 6.0.x
2 years agodrm/amd/pm: Init pm_attr_list when dpm is disabled
ZhenGuo Yin [Wed, 12 Oct 2022 08:54:38 +0000 (16:54 +0800)]
drm/amd/pm: Init pm_attr_list when dpm is disabled

[Why]
In SRIOV multi-vf, dpm is always disabled, and pm_attr_list won't
be initialized. There will be a NULL pointer call trace after
removing the dpm check condition in amdgpu_pm_sysfs_fini.
BUG: kernel NULL pointer dereference, address: 0000000000000000
RIP: 0010:amdgpu_device_attr_remove_groups+0x20/0x90 [amdgpu]
Call Trace:
  <TASK>
  amdgpu_pm_sysfs_fini+0x2f/0x40 [amdgpu]
  amdgpu_device_fini_hw+0xdf/0x290 [amdgpu]

[How]
List pm_attr_list should be initialized when dpm is disabled.

Fixes: a6ad27cec585fe ("drm/amd/pm: Remove redundant check condition")
Signed-off-by: ZhenGuo Yin <zhenguo.yin@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2 years agodrm/amd/pm: disable cstate feature for gpu reset scenario
Evan Quan [Thu, 29 Sep 2022 02:50:44 +0000 (10:50 +0800)]
drm/amd/pm: disable cstate feature for gpu reset scenario

Suggested by PMFW team and same as what did for gfxoff feature.
This can address some Mode1Reset failures observed on SMU13.0.0.

Signed-off-by: Evan Quan <evan.quan@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Reviewed-by: Lijo Lazar <lijo.lazar@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Cc: stable@vger.kernel.org # 6.0.x
2 years agodrm/amd/pm: fulfill SMU13.0.7 cstate control interface
Evan Quan [Thu, 29 Sep 2022 02:30:01 +0000 (10:30 +0800)]
drm/amd/pm: fulfill SMU13.0.7 cstate control interface

Fulfill the functionality for cstate control.

Signed-off-by: Evan Quan <evan.quan@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Reviewed-by: Lijo Lazar <lijo.lazar@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Cc: stable@vger.kernel.org # 6.0.x
2 years agodrm/amd/pm: fulfill SMU13.0.0 cstate control interface
Evan Quan [Thu, 29 Sep 2022 02:24:51 +0000 (10:24 +0800)]
drm/amd/pm: fulfill SMU13.0.0 cstate control interface

Fulfill the functionality for cstate control.

Signed-off-by: Evan Quan <evan.quan@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Reviewed-by: Lijo Lazar <lijo.lazar@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Cc: stable@vger.kernel.org # 6.0.x
2 years agodrm/amdgpu: Add sriov vf ras support in amdgpu_ras_asic_supported
YiPeng Chai [Wed, 28 Sep 2022 07:52:02 +0000 (15:52 +0800)]
drm/amdgpu: Add sriov vf ras support in amdgpu_ras_asic_supported

V2:
Add sriov vf ras support in amdgpu_ras_asic_supported.

Signed-off-by: YiPeng Chai <YiPeng.Chai@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2 years agodrm/amdgpu: Enable ras support for mp0 v13_0_0 and v13_0_10
YiPeng Chai [Tue, 27 Sep 2022 06:06:42 +0000 (14:06 +0800)]
drm/amdgpu: Enable ras support for mp0 v13_0_0 and v13_0_10

V1:
Enable ras support for CHIP_IP_DISCOVERY asic type.

V2:
1. Change commit comment.
2. Enable ras support for mp0 v13_0_0 and v13_0_10.

Signed-off-by: YiPeng Chai <YiPeng.Chai@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2 years agodrm/amdgpu: Enable gmc soft reset on gmc_v11_0_3
YiPeng Chai [Tue, 27 Sep 2022 05:16:27 +0000 (13:16 +0800)]
drm/amdgpu: Enable gmc soft reset on gmc_v11_0_3

Enable gmc soft reset on gmc_v11_0_3.

Signed-off-by: YiPeng Chai <YiPeng.Chai@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2 years agodrm/amdgpu: skip mes self test for gc 11.0.3
Likun Gao [Mon, 25 Jul 2022 12:02:40 +0000 (20:02 +0800)]
drm/amdgpu: skip mes self test for gc 11.0.3

Temporary disable mes self teset for gc 11.0.3.

Signed-off-by: Likun Gao <Likun.Gao@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2 years agodrm/amd/pm: skip loading pptable from driver on secure board for smu_v13_0_10
Kenneth Feng [Mon, 26 Sep 2022 09:15:04 +0000 (17:15 +0800)]
drm/amd/pm: skip loading pptable from driver on secure board for smu_v13_0_10

skip loading pptable from driver on secure board since it's loaded from psp.

Signed-off-by: Kenneth Feng <kenneth.feng@amd.com>
Reviewed-by: Guan Yu <Guan.Yu@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2 years agodrm/amd/amdgpu: enable gfx clock gating features on smu_v13_0_10
Kenneth Feng [Wed, 7 Sep 2022 09:05:34 +0000 (17:05 +0800)]
drm/amd/amdgpu: enable gfx clock gating features on smu_v13_0_10

enable gfx clock gating features on smu_v13_0_10

Signed-off-by: Kenneth Feng <kenneth.feng@amd.com>
Reviewed-by: Jack Gui <Jack.Gui@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2 years agodrm/amd/pm: remove the pptable id override on smu_v13_0_10
Kenneth Feng [Wed, 7 Sep 2022 07:40:34 +0000 (15:40 +0800)]
drm/amd/pm: remove the pptable id override on smu_v13_0_10

remove the pptable id override on smu_v13_0_10,
and the id is fetched from vbios now.

Signed-off-by: Kenneth Feng <kenneth.feng@amd.com>
Reviewed-by: Likun Gao <Likun.Gao@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2 years agodrm/amd/pm: temporarily disable thermal alert on smu_v13_0_10
Kenneth Feng [Sat, 30 Jul 2022 02:58:37 +0000 (10:58 +0800)]
drm/amd/pm: temporarily disable thermal alert on smu_v13_0_10

temporarily disable thermal alert on smu_v13_0_10 due to kfd test fail.
will enable it again after confirming the thermal hardware setting.

Signed-off-by: Kenneth Feng <kenneth.feng@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2 years agodrm/amdgpu: Revert "drm/amdgpu: getting fan speed pwm for vega10 properly"
Asher Song [Fri, 14 Oct 2022 03:36:33 +0000 (11:36 +0800)]
drm/amdgpu: Revert "drm/amdgpu: getting fan speed pwm for vega10 properly"

This reverts commit 16fb4dca95daa9d8e037201166a58de8284f4268.

Unfortunately, that commit causes fan monitors can't be read and written
properly.

Fixes: 16fb4dca95daa9 ("drm/amdgpu: getting fan speed pwm for vega10 properly")
Signed-off-by: Asher Song <Asher.Song@amd.com>
Reviewed-by: Guchun Chen <guchun.chen@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2 years agodrm/amdgpu: Refactor mode2 reset logic for v11.0.7
Victor Zhao [Thu, 13 Oct 2022 07:53:19 +0000 (15:53 +0800)]
drm/amdgpu: Refactor mode2 reset logic for v11.0.7

- refactor mode2 on v11.0.7 to align with aldebaran
- comment out using mode2 reset as default for now, will introduce
another controller to replace previous reset_level_mask

v2: squash in unused variable removal (Alex)

Signed-off-by: Victor Zhao <Victor.Zhao@amd.com>
Reviewed-by: Lijo Lazar <lijo.lazar@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2 years agoRevert "drm/amdgpu: let mode2 reset fallback to default when failure"
Victor Zhao [Thu, 13 Oct 2022 03:06:33 +0000 (11:06 +0800)]
Revert "drm/amdgpu: let mode2 reset fallback to default when failure"

This reverts commit dac6b80818ac2353631c5a33d140d8d5508e2957.

This commit reverted the AMDGPU_SKIP_MODE2_RESET as it conflicts with
the original design of reset handler. Will redesign it.

Fixes: dac6b80818ac23 ("drm/amdgpu: let mode2 reset fallback to default when failure")
Signed-off-by: Victor Zhao <Victor.Zhao@amd.com>
Reviewed-by: Lijo Lazar <lijo.lazar@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2 years agoRevert "drm/amdgpu: add debugfs amdgpu_reset_level"
Victor Zhao [Thu, 13 Oct 2022 02:42:04 +0000 (10:42 +0800)]
Revert "drm/amdgpu: add debugfs amdgpu_reset_level"

This reverts commit 5bd8d53f6fa53eab5433698d1362dae2aa53c1cc.

This commit breaks the reset logic for aldebaran, revert it for now.
Will move the mask inside the reset handler.

Fixes: 5bd8d53f6fa53e ("drm/amdgpu: add debugfs amdgpu_reset_level")
Signed-off-by: Victor Zhao <Victor.Zhao@amd.com>
Reviewed-by: Lijo Lazar <lijo.lazar@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2 years agodrm/amdgpu: set vm_update_mode=0 as default for Sienna Cichlid in SRIOV case
Danijel Slivka [Tue, 4 Oct 2022 13:39:44 +0000 (15:39 +0200)]
drm/amdgpu: set vm_update_mode=0 as default for Sienna Cichlid in SRIOV case

For asic with VF MMIO access protection avoid using CPU for VM table updates.
CPU pagetable updates have issues with HDP flush as VF MMIO access protection
blocks write to mmBIF_BX_DEV0_EPF0_VF0_HDP_MEM_COHERENCY_FLUSH_CNTL register
during sriov runtime.

v3: introduce virtualization capability flag AMDGPU_VF_MMIO_ACCESS_PROTECT
which indicates that VF MMIO write access is not allowed in sriov runtime

Signed-off-by: Danijel Slivka <danijel.slivka@amd.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2 years agodrm/amdgpu/si_dma: remove unused variable in si_dma_stop()
Yang Yingliang [Wed, 12 Oct 2022 07:37:07 +0000 (15:37 +0800)]
drm/amdgpu/si_dma: remove unused variable in si_dma_stop()

After commit 571c05365892 ("drm/amdgpu: switch sdma buffer function
tear down to a helper"), the variable 'ring' is not used anymore, it
can be removed.

Signed-off-by: Yang Yingliang <yangyingliang@huawei.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2 years agodrm/amdgpu: convert amdgpu_amdkfd_gpuvm.c to IP version checks
Alex Deucher [Mon, 3 Oct 2022 19:09:57 +0000 (15:09 -0400)]
drm/amdgpu: convert amdgpu_amdkfd_gpuvm.c to IP version checks

For consistency with the rest of the code.

Acked-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2 years agodrm/amdgpu: convert vega20_ih.c to IP version checks
Alex Deucher [Mon, 3 Oct 2022 19:00:47 +0000 (15:00 -0400)]
drm/amdgpu: convert vega20_ih.c to IP version checks

For consistency with newer asics.

Acked-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2 years agodrm/amdgpu: update psp_fw_type enum in amdgpu_ucode header
Hawking Zhang [Tue, 31 Aug 2021 07:30:08 +0000 (15:30 +0800)]
drm/amdgpu: update psp_fw_type enum in amdgpu_ucode header

To match with the definition in psp firmware

Acked-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Hawking Zhang <Hawking.Zhang@amd.com>
Reviewed-by: Le Ma <le.ma@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2 years agodrm/amdgpu: extend HWIP_MAX_INSTANCE to 28
Hawking Zhang [Tue, 17 May 2022 14:10:41 +0000 (22:10 +0800)]
drm/amdgpu: extend HWIP_MAX_INSTANCE to 28

more ip instances are available

Acked-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2 years agodrm/amdgpu: allow secure submission on gfx11 and sdma6
Yifan Zhang [Sun, 9 Oct 2022 06:37:06 +0000 (14:37 +0800)]
drm/amdgpu: allow secure submission on gfx11 and sdma6

This patch to allow secure submission on gfx11 and sdma6.

Signed-off-by: Yifan Zhang <yifan1.zhang@amd.com>
Reviewed-by: Tim Huang <Tim.Huang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2 years agodrm/amdgpu: add tmz support for GC 11.0.1
Yifan Zhang [Sun, 9 Oct 2022 06:35:20 +0000 (14:35 +0800)]
drm/amdgpu: add tmz support for GC 11.0.1

this patch to add tmz support for GC 11.0.1.

Signed-off-by: Yifan Zhang <yifan1.zhang@amd.com>
Reviewed-by: Tim Huang <Tim.Huang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2 years agoLinux 6.1-rc1
Linus Torvalds [Sun, 16 Oct 2022 22:36:24 +0000 (15:36 -0700)]
Linux 6.1-rc1

2 years agoMerge tag 'random-6.1-rc1-for-linus' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Sun, 16 Oct 2022 22:27:07 +0000 (15:27 -0700)]
Merge tag 'random-6.1-rc1-for-linus' of git://git./linux/kernel/git/crng/random

Pull more random number generator updates from Jason Donenfeld:
 "This time with some large scale treewide cleanups.

  The intent of this pull is to clean up the way callers fetch random
  integers. The current rules for doing this right are:

   - If you want a secure or an insecure random u64, use get_random_u64()

   - If you want a secure or an insecure random u32, use get_random_u32()

     The old function prandom_u32() has been deprecated for a while
     now and is just a wrapper around get_random_u32(). Same for
     get_random_int().

   - If you want a secure or an insecure random u16, use get_random_u16()

   - If you want a secure or an insecure random u8, use get_random_u8()

   - If you want secure or insecure random bytes, use get_random_bytes().

     The old function prandom_bytes() has been deprecated for a while
     now and has long been a wrapper around get_random_bytes()

   - If you want a non-uniform random u32, u16, or u8 bounded by a
     certain open interval maximum, use prandom_u32_max()

     I say "non-uniform", because it doesn't do any rejection sampling
     or divisions. Hence, it stays within the prandom_*() namespace, not
     the get_random_*() namespace.

     I'm currently investigating a "uniform" function for 6.2. We'll see
     what comes of that.

  By applying these rules uniformly, we get several benefits:

   - By using prandom_u32_max() with an upper-bound that the compiler
     can prove at compile-time is ≤65536 or ≤256, internally
     get_random_u16() or get_random_u8() is used, which wastes fewer
     batched random bytes, and hence has higher throughput.

   - By using prandom_u32_max() instead of %, when the upper-bound is
     not a constant, division is still avoided, because
     prandom_u32_max() uses a faster multiplication-based trick instead.

   - By using get_random_u16() or get_random_u8() in cases where the
     return value is intended to indeed be a u16 or a u8, we waste fewer
     batched random bytes, and hence have higher throughput.

  This series was originally done by hand while I was on an airplane
  without Internet. Later, Kees and I worked on retroactively figuring
  out what could be done with Coccinelle and what had to be done
  manually, and then we split things up based on that.

  So while this touches a lot of files, the actual amount of code that's
  hand fiddled is comfortably small"

* tag 'random-6.1-rc1-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/crng/random:
  prandom: remove unused functions
  treewide: use get_random_bytes() when possible
  treewide: use get_random_u32() when possible
  treewide: use get_random_{u8,u16}() when possible, part 2
  treewide: use get_random_{u8,u16}() when possible, part 1
  treewide: use prandom_u32_max() when possible, part 2
  treewide: use prandom_u32_max() when possible, part 1

2 years agoMerge tag 'perf-tools-for-v6.1-2-2022-10-16' of git://git.kernel.org/pub/scm/linux...
Linus Torvalds [Sun, 16 Oct 2022 22:14:29 +0000 (15:14 -0700)]
Merge tag 'perf-tools-for-v6.1-2-2022-10-16' of git://git./linux/kernel/git/acme/linux

Pull more perf tools updates from Arnaldo Carvalho de Melo:

 - Use BPF CO-RE (Compile Once, Run Everywhere) to support old kernels
   when using bperf (perf BPF based counters) with cgroups.

 - Support HiSilicon PCIe Performance Monitoring Unit (PMU), that
   monitors bandwidth, latency, bus utilization and buffer occupancy.

   Documented in Documentation/admin-guide/perf/hisi-pcie-pmu.rst.

 - User space tasks can migrate between CPUs, so when tracing selected
   CPUs, system-wide sideband is still needed, fix it in the setup of
   Intel PT on hybrid systems.

 - Fix metricgroups title message in 'perf list', it should state that
   the metrics groups are to be used with the '-M' option, not '-e'.

 - Sync the msr-index.h copy with the kernel sources, adding support for
   using "AMD64_TSC_RATIO" in filter expressions in 'perf trace' as well
   as decoding it when printing the MSR tracepoint arguments.

 - Fix program header size and alignment when generating a JIT ELF in
   'perf inject'.

 - Add multiple new Intel PT 'perf test' entries, including a jitdump
   one.

 - Fix the 'perf test' entries for 'perf stat' CSV and JSON output when
   running on PowerPC due to an invalid topology number in that arch.

 - Fix the 'perf test' for arm_coresight failures on the ARM Juno
   system.

 - Fix the 'perf test' attr entry for PERF_FORMAT_LOST, adding this
   option to the or expression expected in the intercepted
   perf_event_open() syscall.

 - Add missing condition flags ('hs', 'lo', 'vc', 'vs') for arm64 in the
   'perf annotate' asm parser.

 - Fix 'perf mem record -C' option processing, it was being chopped up
   when preparing the underlying 'perf record -e mem-events' and thus
   being ignored, requiring using '-- -C CPUs' as a workaround.

 - Improvements and tidy ups for 'perf test' shell infra.

 - Fix Intel PT information printing segfault in uClibc, where a NULL
   format was being passed to fprintf.

* tag 'perf-tools-for-v6.1-2-2022-10-16' of git://git.kernel.org/pub/scm/linux/kernel/git/acme/linux: (23 commits)
  tools arch x86: Sync the msr-index.h copy with the kernel sources
  perf auxtrace arm64: Add support for parsing HiSilicon PCIe Trace packet
  perf auxtrace arm64: Add support for HiSilicon PCIe Tune and Trace device driver
  perf auxtrace arm: Refactor event list iteration in auxtrace_record__init()
  perf tests stat+json_output: Include sanity check for topology
  perf tests stat+csv_output: Include sanity check for topology
  perf intel-pt: Fix system_wide dummy event for hybrid
  perf intel-pt: Fix segfault in intel_pt_print_info() with uClibc
  perf test: Fix attr tests for PERF_FORMAT_LOST
  perf test: test_intel_pt.sh: Add 9 tests
  perf inject: Fix GEN_ELF_TEXT_OFFSET for jit
  perf test: test_intel_pt.sh: Add jitdump test
  perf test: test_intel_pt.sh: Tidy some alignment
  perf test: test_intel_pt.sh: Print a message when skipping kernel tracing
  perf test: test_intel_pt.sh: Tidy some perf record options
  perf test: test_intel_pt.sh: Fix return checking again
  perf: Skip and warn on unknown format 'configN' attrs
  perf list: Fix metricgroups title message
  perf mem: Fix -C option behavior for perf mem record
  perf annotate: Add missing condition flags for arm64
  ...

2 years agoMerge tag 'kbuild-fixes-v6.1' of git://git.kernel.org/pub/scm/linux/kernel/git/masahi...
Linus Torvalds [Sun, 16 Oct 2022 18:12:22 +0000 (11:12 -0700)]
Merge tag 'kbuild-fixes-v6.1' of git://git./linux/kernel/git/masahiroy/linux-kbuild

Pull Kbuild fixes from Masahiro Yamada:

 - Fix CONFIG_DEBUG_INFO_DWARF_TOOLCHAIN_DEFAULT=y compile error for the
   combination of Clang >= 14 and GAS <= 2.35.

 - Drop vmlinux.bz2 from the rpm package as it just annoyingly increased
   the package size.

 - Fix modpost error under build environments using musl.

 - Make *.ll files keep value names for easier debugging

 - Fix single directory build

 - Prevent RISC-V from selecting the broken DWARF5 support when Clang
   and GAS are used together.

* tag 'kbuild-fixes-v6.1' of git://git.kernel.org/pub/scm/linux/kernel/git/masahiroy/linux-kbuild:
  lib/Kconfig.debug: Add check for non-constant .{s,u}leb128 support to DWARF5
  kbuild: fix single directory build
  kbuild: add -fno-discard-value-names to cmd_cc_ll_c
  scripts/clang-tools: Convert clang-tidy args to list
  modpost: put modpost options before argument
  kbuild: Stop including vmlinux.bz2 in the rpm's
  Kconfig.debug: add toolchain checks for DEBUG_INFO_DWARF_TOOLCHAIN_DEFAULT
  Kconfig.debug: simplify the dependency of DEBUG_INFO_DWARF4/5

2 years agoMerge tag 'clk-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/clk/linux
Linus Torvalds [Sun, 16 Oct 2022 18:08:19 +0000 (11:08 -0700)]
Merge tag 'clk-for-linus' of git://git./linux/kernel/git/clk/linux

Pull more clk updates from Stephen Boyd:
 "This is the final part of the clk patches for this merge window.

  The clk rate range series needed another week to fully bake. Maxime
  fixed the bug that broke clk notifiers and prevented this from being
  included in the first pull request. He also added a unit test on top
  to make sure it doesn't break so easily again. The majority of the
  series fixes up how the clk_set_rate_*() APIs work, particularly
  around when the rate constraints are dropped and how they move around
  when reparenting clks. Overall it's a much needed improvement to the
  clk rate range APIs that used to be pretty broken if you looked
  sideways.

  Beyond the core changes there are a few driver fixes for a compilation
  issue or improper data causing clks to fail to register or have the
  wrong parents. These are good to get in before the first -rc so that
  the system actually boots on the affected devices"

* tag 'clk-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/clk/linux: (31 commits)
  clk: tegra: Fix Tegra PWM parent clock
  clk: at91: fix the build with binutils 2.27
  clk: qcom: gcc-msm8660: Drop hardcoded fixed board clocks
  clk: mediatek: clk-mux: Add .determine_rate() callback
  clk: tests: Add tests for notifiers
  clk: Update req_rate on __clk_recalc_rates()
  clk: tests: Add missing test case for ranges
  clk: qcom: clk-rcg2: Take clock boundaries into consideration for gfx3d
  clk: Introduce the clk_hw_get_rate_range function
  clk: Zero the clk_rate_request structure
  clk: Stop forwarding clk_rate_requests to the parent
  clk: Constify clk_has_parent()
  clk: Introduce clk_core_has_parent()
  clk: Switch from __clk_determine_rate to clk_core_round_rate_nolock
  clk: Add our request boundaries in clk_core_init_rate_req
  clk: Introduce clk_hw_init_rate_request()
  clk: Move clk_core_init_rate_req() from clk_core_round_rate_nolock() to its caller
  clk: Change clk_core_init_rate_req prototype
  clk: Set req_rate on reparenting
  clk: Take into account uncached clocks in clk_set_rate_range()
  ...

2 years agoMerge tag '6.1-rc-smb3-client-fixes-part2' of git://git.samba.org/sfrench/cifs-2.6
Linus Torvalds [Sun, 16 Oct 2022 18:01:40 +0000 (11:01 -0700)]
Merge tag '6.1-rc-smb3-client-fixes-part2' of git://git.samba.org/sfrench/cifs-2.6

Pull more cifs updates from Steve French:

 - fix a regression in guest mounts to old servers

 - improvements to directory leasing (caching directory entries safely
   beyond the root directory)

 - symlink improvement (reducing roundtrips needed to process symlinks)

 - an lseek fix (to problem where some dir entries could be skipped)

 - improved ioctl for returning more detailed information on directory
   change notifications

 - clarify multichannel interface query warning

 - cleanup fix (for better aligning buffers using ALIGN and round_up)

 - a compounding fix

 - fix some uninitialized variable bugs found by Coverity and the kernel
   test robot

* tag '6.1-rc-smb3-client-fixes-part2' of git://git.samba.org/sfrench/cifs-2.6:
  smb3: improve SMB3 change notification support
  cifs: lease key is uninitialized in two additional functions when smb1
  cifs: lease key is uninitialized in smb1 paths
  smb3: must initialize two ACL struct fields to zero
  cifs: fix double-fault crash during ntlmssp
  cifs: fix static checker warning
  cifs: use ALIGN() and round_up() macros
  cifs: find and use the dentry for cached non-root directories also
  cifs: enable caching of directories for which a lease is held
  cifs: prevent copying past input buffer boundaries
  cifs: fix uninitialised var in smb2_compound_op()
  cifs: improve symlink handling for smb2+
  smb3: clarify multichannel warning
  cifs: fix regression in very old smb1 mounts
  cifs: fix skipping to incorrect offset in emit_cached_dirents

2 years agoRevert "cpumask: fix checking valid cpu range".
Tetsuo Handa [Sat, 15 Oct 2022 15:53:51 +0000 (00:53 +0900)]
Revert "cpumask: fix checking valid cpu range".

This reverts commit 78e5a3399421 ("cpumask: fix checking valid cpu range").

syzbot is hitting WARN_ON_ONCE(cpu >= nr_cpumask_bits) warning at
cpu_max_bits_warn() [1], for commit 78e5a3399421 ("cpumask: fix checking
valid cpu range") is broken.  Obviously that patch hits WARN_ON_ONCE()
when e.g.  reading /proc/cpuinfo because passing "cpu + 1" instead of
"cpu" will trivially hit cpu == nr_cpumask_bits condition.

Although syzbot found this problem in linux-next.git on 2022/09/27 [2],
this problem was not fixed immediately.  As a result, that patch was
sent to linux.git before the patch author recognizes this problem, and
syzbot started failing to test changes in linux.git since 2022/10/10
[3].

Andrew Jones proposed a fix for x86 and riscv architectures [4].  But
[2] and [5] indicate that affected locations are not limited to arch
code.  More delay before we find and fix affected locations, less tested
kernel (and more difficult to bisect and fix) before release.

We should have inspected and fixed basically all cpumask users before
applying that patch.  We should not crash kernels in order to ask
existing cpumask users to update their code, even if limited to
CONFIG_DEBUG_PER_CPU_MAPS=y case.

Link: https://syzkaller.appspot.com/bug?extid=d0fd2bf0dd6da72496dd
Link: https://syzkaller.appspot.com/bug?extid=21da700f3c9f0bc40150
Link: https://syzkaller.appspot.com/bug?extid=51a652e2d24d53e75734
Link: https://lkml.kernel.org/r/20221014155845.1986223-1-ajones@ventanamicro.com
Link: https://syzkaller.appspot.com/bug?extid=4d46c43d81c3bd155060
Reported-by: Andrew Jones <ajones@ventanamicro.com>
Reported-by: syzbot+d0fd2bf0dd6da72496dd@syzkaller.appspotmail.com
Signed-off-by: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>
Cc: Yury Norov <yury.norov@gmail.com>
Cc: Borislav Petkov <bp@alien8.de>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2 years agolib/Kconfig.debug: Add check for non-constant .{s,u}leb128 support to DWARF5
Nathan Chancellor [Fri, 14 Oct 2022 20:42:11 +0000 (13:42 -0700)]
lib/Kconfig.debug: Add check for non-constant .{s,u}leb128 support to DWARF5

When building with a RISC-V kernel with DWARF5 debug info using clang
and the GNU assembler, several instances of the following error appear:

  /tmp/vgettimeofday-48aa35.s:2963: Error: non-constant .uleb128 is not supported

Dumping the .s file reveals these .uleb128 directives come from
.debug_loc and .debug_ranges:

  .Ldebug_loc0:
          .byte   4                               # DW_LLE_offset_pair
          .uleb128 .Lfunc_begin0-.Lfunc_begin0    #   starting offset
          .uleb128 .Ltmp1-.Lfunc_begin0           #   ending offset
          .byte   1                               # Loc expr size
          .byte   90                              # DW_OP_reg10
          .byte   0                               # DW_LLE_end_of_list

  .Ldebug_ranges0:
          .byte   4                               # DW_RLE_offset_pair
          .uleb128 .Ltmp6-.Lfunc_begin0           #   starting offset
          .uleb128 .Ltmp27-.Lfunc_begin0          #   ending offset
          .byte   4                               # DW_RLE_offset_pair
          .uleb128 .Ltmp28-.Lfunc_begin0          #   starting offset
          .uleb128 .Ltmp30-.Lfunc_begin0          #   ending offset
          .byte   0                               # DW_RLE_end_of_list

There is an outstanding binutils issue to support a non-constant operand
to .sleb128 and .uleb128 in GAS for RISC-V but there does not appear to
be any movement on it, due to concerns over how it would work with
linker relaxation.

To avoid these build errors, prevent DWARF5 from being selected when
using clang and an assembler that does not have support for these symbol
deltas, which can be easily checked in Kconfig with as-instr plus the
small test program from the dwz test suite from the binutils issue.

Link: https://sourceware.org/bugzilla/show_bug.cgi?id=27215
Link: https://github.com/ClangBuiltLinux/linux/issues/1719
Signed-off-by: Nathan Chancellor <nathan@kernel.org>
Reviewed-by: Nick Desaulniers <ndesaulniers@google.com>
Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>
2 years agokbuild: fix single directory build
Masahiro Yamada [Fri, 14 Oct 2022 20:18:11 +0000 (05:18 +0900)]
kbuild: fix single directory build

Commit f110e5a250e3 ("kbuild: refactor single builds of *.ko") was wrong.

KBUILD_MODULES _is_ needed for single builds.

Otherwise, "make foo/bar/baz/" does not build module objects at all.

Fixes: f110e5a250e3 ("kbuild: refactor single builds of *.ko")
Reported-by: David Sterba <dsterba@suse.cz>
Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>
Tested-by: David Sterba <dsterba@suse.com>