Le Ma [Fri, 11 Oct 2019 10:21:16 +0000 (18:21 +0800)]
drm/amdgpu/soc15: disable doorbell interrupt as part of BACO entry sequence
Workaround to make RAS recovery work in BACO reset.
Signed-off-by: Le Ma <le.ma@amd.com> Reviewed-by: Alex Deucher <alexander.deucher@amd.com> Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Hans de Goede [Thu, 10 Oct 2019 16:28:17 +0000 (18:28 +0200)]
drm/amdgpu: Bail earlier when amdgpu.cik_/si_support is not set to 1
Bail from the pci_driver probe function instead of from the drm_driver
load function.
This avoid /dev/dri/card0 temporarily getting registered and then
unregistered again, sending unwanted add / remove udev events to
userspace.
Specifically this avoids triggering the (userspace) bug fixed by this
plymouth merge-request:
https://gitlab.freedesktop.org/plymouth/plymouth/merge_requests/59
Note that despite that being a userspace bug, not sending unnecessary
udev events is a good idea in general.
BugLink: https://bugzilla.redhat.com/show_bug.cgi?id=1490490 Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch> Signed-off-by: Hans de Goede <hdegoede@redhat.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Xiaojie Yuan [Wed, 9 Oct 2019 10:52:51 +0000 (18:52 +0800)]
drm/amd/powerplay: re-enable FW_DSTATE feature bit
SMU firmware has fix the bug, so remove this workaround.
Signed-off-by: Xiaojie Yuan <xiaojie.yuan@amd.com> Acked-by: Alex Deucher <alexander.deucher@amd.com> Reviewed-by: Kevin Wang <kevin1.wang@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Kevin Wang <kevin1.wang@amd.com> Reviewed-by: Kenneth Feng <kenneth.feng@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
@@
type T;
identifier swap.TEMP;
@@
(
- T TEMP;
|
- T TEMP = {...};
)
... when != TEMP
Cc: Harry Wentland <harry.wentland@amd.com> Cc: Leo Li <sunpeng.li@amd.com> Cc: Alex Deucher <alexander.deucher@amd.com> Cc: "Christian König" <christian.koenig@amd.com> Cc: "David (ChunMing) Zhou" <David1.Zhou@amd.com> Cc: amd-gfx@lists.freedesktop.org Reviewed-by: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com> Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
chen gong [Sun, 29 Sep 2019 02:58:43 +0000 (10:58 +0800)]
drm/amdgpu: Do not implement power-on for SDMA after do mode2 reset on Renoir
Find that ring sdma0 test failed if turn on SDMA powergating after do
mode2 reset.
Perhaps the mode2 reset does not reset the SDMA PG state, SDMA is
already powered up so there is no need to ask the SMU to power it up
again. So I skip this function for a moment.
Signed-off-by: chen gong <curry.gong@amd.com> Reviewed-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Xiaojie Yuan [Wed, 9 Oct 2019 08:40:59 +0000 (16:40 +0800)]
drm/amd/powerplay: add more feature bits
Additional features that can be enabled in the SMU.
Signed-off-by: Xiaojie Yuan <xiaojie.yuan@amd.com> Acked-by: Alex Deucher <alexander.deucher@amd.com> Reviewed-by: Kevin Wang <kevin1.wang@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Alex Deucher [Wed, 2 Oct 2019 21:10:24 +0000 (16:10 -0500)]
drm/amdgpu/powerplay: fix typo in mvdd table setup
Polaris and vegam use count for the value rather than
level. This looks like a copy paste typo from when
the code was adapted from previous asics.
I'm not sure that the SMU actually uses this value, so
I don't know that it actually is a bug per se.
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=108609 Reported-by: Robert Strube <rstrube@gmail.com> Reviewed-by: Evan Quan <evan.quan@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Nirmoy Das [Fri, 4 Oct 2019 09:53:37 +0000 (11:53 +0200)]
drm/amdgpu: fix memory leak
cleanup error handling code and make sure temporary info array
with the handles are freed by amdgpu_bo_list_put() on
idr_replace()'s failure.
Signed-off-by: Nirmoy Das <nirmoy.das@amd.com> Reviewed-by: Christian König <christian.koenig@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Luben Tuikov [Wed, 2 Oct 2019 04:02:18 +0000 (00:02 -0400)]
drm/amdgpu: Use the ALIGN() macro
Use the ALIGN() macro to set "num_dw" to a
multiple of 8, i.e. lower 3 bits cleared.
Signed-off-by: Luben Tuikov <luben.tuikov@amd.com> Reviewed-by: Christian König <christian.koenig@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Eric Yang [Thu, 22 Aug 2019 15:54:41 +0000 (11:54 -0400)]
drm/amd/display: hook up notify watermark ranges and get clock table
[Why]
Previously SMU was giving us 0s for the clock table. Now they have valid
clock table. We should use theirs. Also, need to send SMU watermark
ranges for selecting optimal watermarks.
Signed-off-by: Eric Yang <Eric.Yang2@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Sung Lee [Tue, 24 Sep 2019 17:20:33 +0000 (13:20 -0400)]
drm/amd/display: Skip DIG Check if Link is Virtual for Display Count
[WHY]
Without a check for virtual links, every link's DIG was getting
checked for enabled or disabled. If link was virtual, since it
did not have a DIG, this would cause issues.
[HOW]
Skip DIG Enable check if link is virtual and add virtual link to
to display count.
Signed-off-by: Sung Lee <sung.lee@amd.com> Reviewed-by: Yongqiang Sun <yongqiang.sun@amd.com> Acked-by: Bhawanpreet Lakha <Bhawanpreet.Lakha@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Wayne Lin [Thu, 19 Sep 2019 09:41:02 +0000 (17:41 +0800)]
drm/amd/display: correct stream LTE_340MCSC_SCRAMBLE value
[Why]
HDMI 2.0 requires scrambling under specific conditions. We refer to
stream property LTE_340MCSC_SCRAMBLE to determine whether en/dis
scrambling.
While creating stream for sink, we setup LTE_340MCSC_SCRAMBLE by
referring to edid_caps. However, dm_helpers_parse_edid_caps()
doesn't construct HDMI Forum block data for edid_caps.
Moreover, fill_stream_properties_from_drm_display_mode() aslo
unconsciously clear the LTE_340MCSC_SCRAMBLE flag.
[How]
Drm already provides drm_display_info to refer HDMI Forum vsdb info.
Set stream LTE_340MCSC_SCRAMBLE by drm_display_info and remove
memset in fill_stream_properties_from_drm_display_mode()
Signed-off-by: Wayne Lin <Wayne.Lin@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>
David Galiffi [Sat, 21 Sep 2019 00:20:23 +0000 (20:20 -0400)]
drm/amd/display: Fix dongle_caps containing stale information.
[WHY]
During detection:
function: get_active_converter_info populates link->dpcd_caps.dongle_caps
only when dpcd_rev >= DPCD_REV_11 and DWN_STRM_PORTX_TYPE is
DOWN_STREAM_DETAILED_HDMI or DOWN_STREAM_DETAILED_DP_PLUS_PLUS.
Otherwise, it is not cleared, and stale information remains.
During mode validation:
function: dp_active_dongle_validate_timing reads
link->dpcd_caps.dongle_caps->dongle_type to determine the maximum
pixel clock to support. This information is now stale and no longer
valid.
[HOW]
dp_active_dongle_validate_timing should be using
link->dpcd_caps->dongle_type instead.
Signed-off-by: David Galiffi <david.galiffi@amd.com> Reviewed-by: Jun Lei <Jun.Lei@amd.com> Acked-by: Bhawanpreet Lakha <Bhawanpreet.Lakha@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
drm/amd/display: Add capability check for static ramp calc
[Why]
Static ramp to max refresh rate does not have capability check on
calculated v_total. Programming a lower v_total_min and max than the
total causes continuous spurious HPDs.
[How]
Add a capability check after v_total calculation similar to calculate
v_total helper functions.
Signed-off-by: Jaehyun Chung <jaehyun.chung@amd.com> Reviewed-by: Anthony Koo <Anthony.Koo@amd.com> Acked-by: Bhawanpreet Lakha <Bhawanpreet.Lakha@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Xiaodong Yan [Fri, 30 Aug 2019 09:07:01 +0000 (17:07 +0800)]
drm/amd/display: make aux defer delay and aux sw start delay seperate
[why]
1. defer delay and sw start delay has been mixed up, defer delay was
programmed to AUX_SW_CONTROL:AUX_SW_START_DELAY.
2. There's no delay for defer
[how]
1. Set aux sw start to 0
2. Add delay for defer scenario
Signed-off-by: Xiaodong Yan <Xiaodong.Yan@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>
Leo Li [Fri, 20 Sep 2019 13:43:36 +0000 (09:43 -0400)]
drm/amd/display: Fix maybe-uninitialized warning
[Why]
Compiling with GCC 9.1.0 gives the following warning (I have
warnings-as-errors enabled):
drivers/gpu/drm/amd/amdgpu/../dal-dev/dc/core/dc.c: In function 'dc_validate_seamless_boot_timing':
drivers/gpu/drm/amd/amdgpu/../dal-dev/dc/core/dc.c:1180:8: error: 'se' may be used uninitialized in this function [-Werror=maybe-uninitialized]
1180 | if (!se->funcs->dp_get_pixel_format(
| ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1181 | se,
| ~~~
1182 | &hw_crtc_timing.pixel_encoding,
| ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1183 | &hw_crtc_timing.display_color_depth))
[How]
Initialize se to NULL.
Signed-off-by: Leo Li <sunpeng.li@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>
Martin Leung [Tue, 17 Sep 2019 18:50:22 +0000 (14:50 -0400)]
drm/amd/display: add more checks to validate seamless boot timing
[why]
we found using an active DP to HDMI panel that we weren't validating
dp_pixel_format and hardware timing v_front_porch, causing screen to
blank and/or corrupt while attempting a seamless boot.
[how]
added checks during dc_validate_seamless_boot_timing for these values
Signed-off-by: Martin Leung <martin.leung@amd.com> Reviewed-by: Anthony Koo <Anthony.Koo@amd.com> Acked-by: Bhawanpreet Lakha <Bhawanpreet.Lakha@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
drm/amd/display: add new active dongle to existent w/a
[Why & How]
Dongle 0x00E04C power down all internal circuits including
AUX communication preventing reading DPCD table.
Encoder will skip DP RX power down on disable output
to keep receiver powered all the time.
Signed-off-by: Vitaly Prosyak <vitaly.prosyak@amd.com> Reviewed-by: Charlene Liu <Charlene.Liu@amd.com> Acked-by: Bhawanpreet Lakha <Bhawanpreet.Lakha@amd.com> Acked-by: Vitaly Prosyak <Vitaly.Prosyak@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Joseph Gravenor [Mon, 16 Sep 2019 19:13:33 +0000 (15:13 -0400)]
drm/amd/display: add guard for SMU ver, for 48mhz clk
[why]
dp_48m_refclk_driver_pwdn is persistent through S3 and S5.
This was worked arround in SMU FW 55.21.0. Earlier FW don't have this fix
so we will hang on reboot
[how]
add a guard for smu versions before SMU FW 55.21.0
Signed-off-by: Joseph Gravenor <joseph.gravenor@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Eric Yang [Thu, 18 Jul 2019 17:56:59 +0000 (13:56 -0400)]
drm/amd/display: fix code to control 48mhz refclk
[Why]
The SMU message to enable this feature looks at argument. Previous code
didn't send right argument. This change will allow the feature to be
be enabled.
[How]
Fixed one issue where SMU message to enable the feature was sent without
setting the parameter.
Signed-off-by: Eric Yang <Eric.Yang2@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Eric Yang [Fri, 6 Sep 2019 22:26:23 +0000 (18:26 -0400)]
drm/amd/display: exit PSR during detection
[Why]
If 48mhz refclk is turned off during PSR, we will have issue doing
link training during detection.
[How]
Get out of PSR before detection
Signed-off-by: Eric Yang <Eric.Yang2@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>
Julian Parkin [Thu, 29 Aug 2019 21:06:05 +0000 (17:06 -0400)]
drm/amd/display: Program DWB watermarks from correct state
[Why]
When diags adds a DWB via a stream update, we calculate MMHUBBUB
paramaters, but dc->current_state has not yet been updated
when the DWB programming happens. This leads to overflow on
high bandwidth tests since the incorrect MMHUBBUB arbitration
parameters are programmed.
[How]
Pass the updated context down to the (enable|update)_writeback functions
so that they can use the correct watermarks when programming MMHUBBUB.
Josip Pavic [Thu, 12 Sep 2019 19:40:08 +0000 (15:40 -0400)]
drm/amd/display: wait for set pipe mcp command completion
[Why]
When the driver sends a pipe set command to the DMCU FW, it does not wait
for the command to complete. This can lead to unpredictable behavior if,
for example, the driver were to request a pipe disable to the FW via MCP,
then power down some hardware before the firmware has completed processing
the command.
[How]
Wait for the DMCU FW to finish processing set pipe commands
Signed-off-by: Josip Pavic <Josip.Pavic@amd.com> Reviewed-by: Anthony Koo <Anthony.Koo@amd.com> Acked-by: Bhawanpreet Lakha <Bhawanpreet.Lakha@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
drm/amd/display: Restore should_update_pstate_support after bad revert
[Why]
This function was mistakenly reverted as part of a legitimate revert. The
old version that was reverted to has bad logic, and is causing situations
where p-state change support is being toggled when it shouldn't be,
resulting in hangs.
Signed-off-by: Joshua Aberback <joshua.aberback@amd.com> Reviewed-by: Jun Lei <Jun.Lei@amd.com> Acked-by: Bhawanpreet Lakha <Bhawanpreet.Lakha@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Lewis Huang [Thu, 5 Sep 2019 07:33:58 +0000 (15:33 +0800)]
drm/amd/display: check phy dpalt lane count config
[Why]
Type-c PHY config is not align with dpcd lane count.
When those values didn't match, it cause driver do
link training with 4 lane but phy only can output 2 lane.
The link trainig always fail.
[How]
1. Modify get_max_link_cap function. According DPALT_DP4
to update max lane count.
2. Add dp_mst_verify_link_cap to handle MST case because
we didn't call dp_mst_verify_link_cap for MST case.
Signed-off-by: Lewis Huang <Lewis.Huang@amd.com> Reviewed-by: Wenjing Liu <Wenjing.Liu@amd.com> Acked-by: Bhawanpreet Lakha <Bhawanpreet.Lakha@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Wayne Lin [Tue, 3 Sep 2019 21:31:16 +0000 (05:31 +0800)]
drm/amd/display: build up VSIF infopacket
[Why]
Didn't send VSIF infopacket when it's 4k mode defined in HDMI 1.4b.
For HDMI 1.4b, While displaying 4k modes, it should send VSP.
[How]
Call mod_build_hf_vsif_infopacket() function to build info frame
and send it.
Signed-off-by: Wayne Lin <Wayne.Lin@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>
drm/amd/display: Add missing shifts and masks for dpp registers on dcn2
[Why]
The register CM_TEST_DEBUG_DATA is used in dpp1_program_input_csc, which is
called from dpp2_cnv_setup, but the shifts and masks for the fields of that
register are not initialized for dcn2. This causes all reads of that register
to return 0.
Jun Lei [Mon, 9 Sep 2019 15:05:33 +0000 (11:05 -0400)]
drm/amd/display: add explicit comparator as default optimization check
[why]
memcmp is vulnerable to regression due to dc_clocks structures not being
organized properly (not "current" clock related structures being at the beginning
of the structure) and causes unnecessary setting of the optimize bit
[how]
add a dcn sepcific comparator, implement for dcn2
Signed-off-by: Jun Lei <Jun.Lei@amd.com> Reviewed-by: Anthony Koo <Anthony.Koo@amd.com> Acked-by: Bhawanpreet Lakha <Bhawanpreet.Lakha@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Wayne Lin [Tue, 3 Sep 2019 21:18:31 +0000 (05:18 +0800)]
drm/amd/display: add support for VSIP info packet
[Why]
The vendor specific infoframe is needed for HDMI while displaying
specific modes.
DC supports sending it, but we aren't currently building it
[How]
Add mod_build_hf_vsif_infopacket() to build the vendor specific
info packet.
Signed-off-by: Wayne Lin <Wayne.Lin@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>
Joseph Gravenor [Wed, 4 Sep 2019 16:43:05 +0000 (12:43 -0400)]
drm/amd/display: fix hotplug during display off
[why]
HPD is not suppressed when we lower
clocks on renoir. B/c of this we do link
training when the 48mhz refclk is off, which
will cause ASIC hang.
[how]
Exit optimized power state for detection purpose.
Signed-off-by: Joseph Gravenor <joseph.gravenor@amd.com> Reviewed-by: Eric Yang <eric.yang2@amd.com> Acked-by: Bhawanpreet Lakha <Bhawanpreet.Lakha@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Julian Parkin [Thu, 22 Aug 2019 20:32:05 +0000 (16:32 -0400)]
drm/amd/display: Reprogram FMT on pipe change
[Why]
When planes are added or removed from a stream, the change
in pipe usage from dynamic MPC combine can cause a second
stream using ODM combine to pick a different pipe to combine with.
In this scenario, a different OPP is connected to the ODM without
programming its FMT.
[How]
Reprogram the FMT in dcn20_program_pipe whenever a pipe is
newly enabled, or when its opp changes.
Wenjing Liu [Tue, 3 Sep 2019 21:15:48 +0000 (17:15 -0400)]
drm/amd/display: skip enable stream on disconnected display
[why]
Virtual signal means there is no display attached.
In this case we will assign a virtual signal type to the stream.
We should only enable the front end of the stream but not the back end.
[how]
When stream is enabling with virtual signal type, skip backend programming.
Signed-off-by: Wenjing Liu <Wenjing.Liu@amd.com> Reviewed-by: Jun Lei <Jun.Lei@amd.com> Acked-by: Bhawanpreet Lakha <Bhawanpreet.Lakha@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Wayne Lin [Tue, 3 Sep 2019 22:12:22 +0000 (06:12 +0800)]
drm/amd/display: Correct values in AVI infoframe
[Why]
While displaying 4k modes defined in HDMI1.4b, should set VIC to 0
and use VSP HDMI_VIC to indicate the mode.
[How]
Use functions defined in drm to set up the VIC correspondingly.
Signed-off-by: Wayne Lin <Wayne.Lin@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>
[Why]
To compare the crc of the framebuffer data at input of
display pipeline with the crc of the otg, we need to
disable pixel formatter's dynamic expansion feature during
crc capture and keep it enable in the normal operation.
[HOW]
Expose a new interface in DM and dc for pixel formatter
(fmt dynamic bitdepth expansion control). Interface control
the FMT_DYNAMIC_EXP_EN bit, during crc capture keep
it disabled.
Signed-off-by: Robin Singh <robin.singh@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>
Oak Zeng [Mon, 7 Oct 2019 20:32:23 +0000 (15:32 -0500)]
drm/amdgpu: Enable gfx cache probing on HDP write for arcturus
This allows gfx cache to be probed and invalidated (for none-dirty cache lines)
on a HDP write (from either another GPU or CPU). This should work only for the
memory mapped as RW memory type newly added for arcturus, to achieve some cache
coherence b/t multiple memory clients.
Signed-off-by: Oak Zeng <Oak.Zeng@amd.com> Acked-by: Christian Konig <christian.koenig@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Oak Zeng [Mon, 7 Oct 2019 20:21:03 +0000 (15:21 -0500)]
drm/amdgpu: Clean up gmc_v9_0_gart_enable
Many logic in this function are HDP set up,
not gart set up. Moved those logic to gmc_v9_0_hw_init.
No functional change.
Signed-off-by: Oak Zeng <Oak.Zeng@amd.com> Acked-by: Christian konig <christian.koenig@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Marek Olšák <marek.olsak@amd.com> Reviewed-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Participate in device cgroup. All kfd devices are exposed via /dev/kfd.
So use /dev/dri/renderN node.
Before exposing the device to a task check if it has permission to
access it. If the task (based on its cgroup) can access /dev/dri/renderN
then expose the device via kfd node.
If the task cannot access /dev/dri/renderN then process device data
(pdd) is not created. This will ensure that task cannot use the device.
In sysfs topology, all device nodes are visible irrespective of the task
cgroup. The sysfs node directories are created at driver load time and
cannot be changed dynamically. However, access to information inside
nodes is controlled based on the task's cgroup permissions.
Signed-off-by: Harish Kasiviswanathan <Harish.Kasiviswanathan@amd.com> Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
All AMD compute devices are exported via single device node /dev/kfd. As
a result devices cannot be controlled individually using device cgroup.
AMD compute devices will rely on its graphics counterpart that exposes
/dev/dri/renderN node for each device. For each task (based on its
cgroup), KFD driver will check if /dev/dri/renderN node is accessible
before exposing it.
Signed-off-by: Harish Kasiviswanathan <Harish.Kasiviswanathan@amd.com> Acked-by: Tejun Heo <tj@kernel.org> Acked-by: Felix Kuehling <Felix.Kuehling@amd.com> Reviewed-by: Roman Gushchin <guro@fb.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Ori Messinger [Wed, 2 Oct 2019 14:02:07 +0000 (10:02 -0400)]
drm/amdgpu: Report vram vendor with sysfs (v3)
The vram vendor can be found as a separate sysfs file at:
/sys/class/drm/card[X]/device/mem_info_vram_vendor
The vram vendor is displayed as a string value.
v2: Use correct bit masking, and cache vram_vendor in gmc
v3: Drop unused functions for vram width, type, and vendor
Signed-off-by: Ori Messinger <ori.messinger@amd.com> Reviewed-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Dan Carpenter [Mon, 7 Oct 2019 08:52:10 +0000 (11:52 +0300)]
drm/amdkfd: Fix a && vs || typo
In the current code if "device_info" is ever NULL then the kernel will
Oops so probably || was intended instead of &&.
Fixes: e392c887df97 ("drm/amdkfd: Use array to probe kfd2kgd_calls") Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Dan Carpenter [Mon, 7 Oct 2019 09:02:06 +0000 (12:02 +0300)]
drm/amd/powerplay: Fix error handling in smu_init_fb_allocations()
The error handling is off by one. We should not free the first
"tables[i].bo" without decrementing "i" because that might result in a
double free. The second problem is that when an error occurs, then the
zeroth element "tables[0].bo" isn't freed.
I had make "i" signed int for the error handling to work, so I just
updated "ret" as well as a clean up.
Fixes: f96357a991b9 ("drm/amd/powerplay: implement smu_init(fini)_fb_allocations function") Reviewed-by: Kevin Wang <kevin1.wang@amd.com> Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Dan Carpenter [Mon, 7 Oct 2019 09:04:54 +0000 (12:04 +0300)]
drm/amd/powerplay: unlock on error in smu_resume()
This function needs to drop the mutex before returning.
Fixes: f7e3a5776fa6 ("drm/amd/powerplay: check SMU engine readiness before proceeding on S3 resume") Reviewed-by: Kevin Wang <kevin1.wang@amd.com> Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
zhengbin [Sat, 5 Oct 2019 02:44:36 +0000 (10:44 +0800)]
drm/amd/display: Remove set but not used variables 'pp_smu', 'old_pipe'
Fixes gcc '-Wunused-but-set-variable' warning:
drivers/gpu/drm/amd/display/dc/dce110/dce110_hw_sequencer.c: In function dce110_enable_audio_stream:
drivers/gpu/drm/amd/display/dc/dce110/dce110_hw_sequencer.c:949:23: warning: variable pp_smu set but not used [-Wunused-but-set-variable]
drivers/gpu/drm/amd/display/dc/dce110/dce110_hw_sequencer.c: In function dce110_disable_audio_stream:
drivers/gpu/drm/amd/display/dc/dce110/dce110_hw_sequencer.c:983:23: warning: variable pp_smu set but not used [-Wunused-but-set-variable]
drivers/gpu/drm/amd/display/dc/dce110/dce110_hw_sequencer.c: In function dce110_program_front_end_for_pipe:
drivers/gpu/drm/amd/display/dc/dce110/dce110_hw_sequencer.c:2429:19: warning: variable old_pipe set but not used [-Wunused-but-set-variable]
'pp_smu' is not used since commit 170a2398d2d8 ("drm/amd/display:
make clk_mgr call enable_pme_wa")
'old_pipe' is not used since commit 65d38262b3e8 ("drm/amd/display:
fbc state could not reach while enable fbc")
Reviewed-by: Harry Wentland <harry.wentland@amd.com> Reported-by: Hulk Robot <hulkci@huawei.com> Signed-off-by: zhengbin <zhengbin13@huawei.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
zhengbin [Sat, 5 Oct 2019 02:44:35 +0000 (10:44 +0800)]
drm/amd/display: Remove set but not used variable 'pixel_width'
Fixes gcc '-Wunused-but-set-variable' warning:
drivers/gpu/drm/amd/display/dc/dcn20/dcn20_dpp.c: In function dpp2_get_optimal_number_of_taps:
drivers/gpu/drm/amd/display/dc/dcn20/dcn20_dpp.c:359:11: warning: variable pixel_width set but not used [-Wunused-but-set-variable]
It is not used since commit f7de96ee8b5f ("drm/amd/display:
Add DCN2 DPP")
Reviewed-by: Harry Wentland <harry.wentland@amd.com> Reported-by: Hulk Robot <hulkci@huawei.com> Signed-off-by: zhengbin <zhengbin13@huawei.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
zhengbin [Sat, 5 Oct 2019 02:44:34 +0000 (10:44 +0800)]
drm/amd/display: Remove set but not used variables 'h_ratio_chroma', 'v_ratio_chroma'
Fixes gcc '-Wunused-but-set-variable' warning:
drivers/gpu/drm/amd/display/dc/dcn20/dcn20_dwb_scl.c: In function dwb_program_horz_scalar:
drivers/gpu/drm/amd/display/dc/dcn20/dcn20_dwb_scl.c:725:11: warning: variable h_ratio_chroma set but not used [-Wunused-but-set-variable]
drivers/gpu/drm/amd/display/dc/dcn20/dcn20_dwb_scl.c: In function dwb_program_vert_scalar:
drivers/gpu/drm/amd/display/dc/dcn20/dcn20_dwb_scl.c:806:11: warning: variable v_ratio_chroma set but not used [-Wunused-but-set-variable]
They are not used since commit 345429a67c48 ("drm/amd/display:
Add DCN2 DWB")
Reviewed-by: Harry Wentland <harry.wentland@amd.com> Reported-by: Hulk Robot <hulkci@huawei.com> Signed-off-by: zhengbin <zhengbin13@huawei.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
zhengbin [Sat, 5 Oct 2019 02:44:33 +0000 (10:44 +0800)]
drm/amd/display: Remove set but not used variable 'source_bpp'
Fixes gcc '-Wunused-but-set-variable' warning:
drivers/gpu/drm/amd/display/dc/dsc/rc_calc.c: In function calc_rc_params:
drivers/gpu/drm/amd/display/dc/dsc/rc_calc.c:180:6: warning: variable source_bpp set but not used [-Wunused-but-set-variable]
It is not used since commit 97bda0322b8a ("drm/amd/display:
Add DSC support for Navi (v2)")
Reviewed-by: Harry Wentland <harry.wentland@amd.com> Reported-by: Hulk Robot <hulkci@huawei.com> Signed-off-by: zhengbin <zhengbin13@huawei.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
zhengbin [Sat, 5 Oct 2019 02:44:32 +0000 (10:44 +0800)]
drm/amd/display: Make function wait_for_alt_mode static
Fix sparse warnings:
drivers/gpu/drm/amd/display/dc/core/dc_link.c:687:6: warning: symbol 'wait_for_alt_mode' was not declared. Should it be static?
Reviewed-by: Harry Wentland <harry.wentland@amd.com> Reported-by: Hulk Robot <hulkci@huawei.com> Signed-off-by: zhengbin <zhengbin13@huawei.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Colin Ian King [Sat, 5 Oct 2019 17:58:08 +0000 (18:58 +0100)]
drm/amdkfd: add missing void argument to function kgd2kfd_init
Function kgd2kfd_init is missing a void argument, add it
to clean up the non-ANSI function declaration.
Acked-by: Randy Dunlap <rdunlap@infradead.org> Signed-off-by: Colin Ian King <colin.king@canonical.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
YueHaibing [Sun, 6 Oct 2019 10:57:35 +0000 (18:57 +0800)]
drm/amd/display: remove set but not used variable 'core_freesync'
Fixes gcc '-Wunused-but-set-variable' warning:
rivers/gpu/drm/amd/amdgpu/../display/modules/freesync/freesync.c:
In function mod_freesync_get_settings:
drivers/gpu/drm/amd/amdgpu/../display/modules/freesync/freesync.c:984:24:
warning: variable core_freesync set but not used [-Wunused-but-set-variable]
It is not used since commit 98e6436d3af5 ("drm/amd/display: Refactor FreeSync module")
Reviewed-by: Harry Wentland <harry.wentland@amd.com> Reported-by: Hulk Robot <hulkci@huawei.com> Signed-off-by: YueHaibing <yuehaibing@huawei.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Oak Zeng [Fri, 4 Oct 2019 14:34:29 +0000 (09:34 -0500)]
drm/amdkfd: Print more sdma engine hqds in debug fs
Previously only PCIe-optimized SDMA engine hqds were
exposed in debug fs. Print all SDMA engine hqds.
Reported-by: Jonathan Kim <Jonathan.Kim@amd.com> Signed-off-by: Jonathan Kim <Jonathan.Kim@amd.com> Signed-off-by: Oak Zeng <Oak.Zeng@amd.com> Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Oak Zeng [Fri, 4 Oct 2019 14:28:21 +0000 (09:28 -0500)]
drm/amdkfd: Fix MQD size calculation
On device initialization, a chunk of GTT memory is pre-allocated for
HIQ and all SDMA queues mqd. The size of this allocation was wrong.
The correct sdma engine number should be PCIe-optimized SDMA engine
number plus xgmi SDMA engine number.
Reported-by: Jonathan Kim <Jonathan.Kim@amd.com> Signed-off-by: Jonathan Kim <Jonathan.Kim@amd.com> Signed-off-by: Oak Zeng <Oak.Zeng@amd.com> Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Felix Kuehling [Thu, 3 Oct 2019 21:54:57 +0000 (17:54 -0400)]
drm/amdgpu: Fix error handling in amdgpu_ras_recovery_init
Don't set a struct pointer to NULL before freeing its members. It's
hard to see what's happening due to a local pointer-to-pointer data
aliasing con->eh_data.
Signed-off-by: Felix Kuehling <Felix.Kuehling@amd.com> Tested-by: Philip Cox <Philip.Cox@amd.com> Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Colin Ian King [Thu, 3 Oct 2019 21:40:49 +0000 (22:40 +0100)]
drm/amdgpu: remove redundant variable r and redundant return statement
There is a return statement that is not reachable and a variable that
is not used. Remove them.
Addresses-Coverity: ("Structurally dead code") Fixes: de7b45babd9b ("drm/amdgpu: cleanup creating BOs at fixed location (v2)") Reviewed-by: Christian König <christian.koenig@amd.com> Signed-off-by: Colin Ian King <colin.king@canonical.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
The boolean variable pasid_mapping_needed is not initialized and
there are code paths that do not assign it any value before it is
is read later. Fix this by initializing pasid_mapping_needed to
false.
Addresses-Coverity: ("Uninitialized scalar variable") Fixes: 6817bf283b2b ("drm/amdgpu: grab the id mgr lock while accessing passid_mapping") Reviewed-by: Christian König <christian.koenig@amd.com> Signed-off-by: Colin Ian King <colin.king@canonical.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
zhengbin [Fri, 4 Oct 2019 09:37:40 +0000 (17:37 +0800)]
drm/amd/display: Make some functions static
Fix sparse warnings:
drivers/gpu/drm/amd/amdgpu/../display/amdgpu_dm/amdgpu_dm_hdcp.c:32:6: warning: symbol 'lp_write_i2c' was not declared. Should it be static?
drivers/gpu/drm/amd/amdgpu/../display/amdgpu_dm/amdgpu_dm_hdcp.c:42:6: warning: symbol 'lp_read_i2c' was not declared. Should it be static?
drivers/gpu/drm/amd/amdgpu/../display/amdgpu_dm/amdgpu_dm_hdcp.c:52:6: warning: symbol 'lp_write_dpcd' was not declared. Should it be static?
drivers/gpu/drm/amd/amdgpu/../display/amdgpu_dm/amdgpu_dm_hdcp.c:59:6: warning: symbol 'lp_read_dpcd' was not declared. Should it be static?
Reported-by: Hulk Robot <hulkci@huawei.com> Signed-off-by: zhengbin <zhengbin13@huawei.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Leo Liu [Tue, 1 Oct 2019 22:12:45 +0000 (18:12 -0400)]
drm/amdgpu: add code comment in vcn_v2_5_hw_init
Add a comment to VCN 2.5 encode ring
Signed-off-by: Leo Liu <leo.liu@amd.com> Reviewed-by: Christian König <christian.koenig@amd.com> Reviewed-by: James Zhu <James.Zhu@amd.com> Cc: Alex Deucher <alexander.deucher@amd.com> Cc: Christian König <christian.koenig@amd.com> Cc: David (ChunMing) Zhou <David1.Zhou@amd.com> Cc: David Airlie <airlied@linux.ie> Cc: Daniel Vetter <daniel@ffwll.ch> Cc: amd-gfx@lists.freedesktop.org Cc: dri-devel@lists.freedesktop.org Cc: linux-kernel@vger.kernel.org Cc: Gustavo A. R. Silva <gustavo@embeddedor.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Leo Liu [Tue, 1 Oct 2019 22:08:46 +0000 (18:08 -0400)]
drm/amdgpu/vcn: use amdgpu_ring_test_helper
Instead of amdgpu_ring_test_ring, so the helper function determines
whether the ring is ready
Signed-off-by: Leo Liu <leo.liu@amd.com> Reviewed-by: Christian König <christian.koenig@amd.com> Reviewed-by: James Zhu <James.Zhu@amd.com> Acked-by: Gustavo A. R. Silva <gustavo@embeddedor.com> Cc: Gustavo A. R. Silva <gustavo@embeddedor.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
There is a spelling mistake in the macros H1_A45_AUTHENICATED and
D1_A4_AUTHENICATED, fix these by adding the missing T.
Reviewed-by: Harry Wentland <harry.wentland@amd.com> Signed-off-by: Colin Ian King <colin.king@canonical.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
The type definition for 'uint' clashes with the generic kernel
headers:
drivers/gpu/drm/amd/amdgpu/../display/dc/dml/dcn21/display_mode_vba_21.c:43:22: error: redefinition of typedef 'uint' is a C11 feature [-Werror,-Wtypedef-redefinition]
include/linux/types.h:92:23: note: previous definition is here
Just remove this type and use plain 'unsigned int' consistently,
as it is already use almost everywhere in this file.
Arnd Bergmann [Wed, 2 Oct 2019 12:01:23 +0000 (14:01 +0200)]
drm/amdgpu: hide another #warning
An earlier patch of mine disabled some #warning statements
that get in the way of build testing, but then another
instance was added around the same time.
Remove that as well.
Fixes: b5203d16aef4 ("drm/amd/amdgpu: hide #warning for missing DC config") Fixes: e1c14c43395c ("drm/amdgpu: Enable DC on Renoir") Signed-off-by: Arnd Bergmann <arnd@arndb.de> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Arnd Bergmann [Wed, 2 Oct 2019 12:01:22 +0000 (14:01 +0200)]
drm/amdgpu: make pmu support optional, again
When CONFIG_PERF_EVENTS is disabled, we cannot compile the pmu
portion of the amdgpu driver:
drivers/gpu/drm/amd/amdgpu/amdgpu_pmu.c:48:38: error: no member named 'hw' in 'struct perf_event'
struct hw_perf_event *hwc = &event->hw;
~~~~~ ^
drivers/gpu/drm/amd/amdgpu/amdgpu_pmu.c:51:13: error: no member named 'attr' in 'struct perf_event'
if (event->attr.type != event->pmu->type)
~~~~~ ^
...
The same bug was already fixed by commit d155bef0636e ("amdgpu: make pmu
support optional") but broken again by what looks like an incorrectly
rebased patch.
drm/amdgpu: remove set but not used variable 'pipe'
Fixes gcc '-Wunused-but-set-variable' warning:
rivers/gpu/drm/amd/amdgpu/amdgpu_gfx.c: In function
‘amdgpu_gfx_graphics_queue_acquire’:
drivers/gpu/drm/amd/amdgpu/amdgpu_gfx.c:234:16: warning:
variable ‘pipe’ set but not used [-Wunused-but-set-variable]
It is never used, so can be removed.
Reported-by: Hulk Robot <hulkci@huawei.com> Signed-off-by: yu kuai <yukuai3@huawei.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Navid Emamdoost [Wed, 2 Oct 2019 03:46:07 +0000 (22:46 -0500)]
drm/amdgpu: fix multiple memory leaks in acp_hw_init
In acp_hw_init there are some allocations that needs to be released in
case of failure:
1- adev->acp.acp_genpd should be released if any allocation attemp for
adev->acp.acp_cell, adev->acp.acp_res or i2s_pdata fails.
2- all of those allocations should be released if
mfd_add_hotplug_devices or pm_genpd_add_device fail.
3- Release is needed in case of time out values expire.
Reviewed-by: Christian König <christian.koenig@amd.com> Signed-off-by: Navid Emamdoost <navid.emamdoost@gmail.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Alex Deucher [Tue, 1 Oct 2019 21:45:27 +0000 (16:45 -0500)]
drm/amdgpu: don't increment vram lost if we are in hibernation
We reset the GPU as part of our hibernation sequence so we need
to make sure we don't mark vram as lost in that case.
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=111879 Acked-by: Christian König <christian.koenig@amd.com> Reviewed-by: Huang Rui <ray.huang@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
shaoyunl [Tue, 1 Oct 2019 19:52:31 +0000 (15:52 -0400)]
drm/amdgpu : enable msix for amdgpu driver
We might used out of the msi resources in some cloud project
which have a lot gpu devices(including PF and VF), msix can
provide enough resources from system level view
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>