drm/amd/display: Fix vblank and pageflip event handling for FreeSync
[Why]
We're sending the drm vblank event a frame too early in the case where
the pageflip happens close to VUPDATE and ends up blocking the signal.
The implementation in DM was previously correct *before* we started
sending vblank events from VSTARTUP unconditionally to handle cases
where HUBP was off, OTG was ON and userspace was still requesting some
DRM planes enabled. As part of that patch series we dropped VUPDATE
since it was deemed close enough to VSTARTUP, but there's a key
difference betweeen VSTARTUP and VUPDATE - the VUPDATE signal can be
blocked if we're holding the pipe lock.
There was a fix recently to revert the unconditional behavior for the
DCN VSTARTUP vblank event since it was sending the pageflip event on
the wrong frame - once again, due to blocking VUPDATE and having the
address start scanning out two frames later.
The problem with this fix is it didn't update the logic that calls
drm_crtc_handle_vblank(), so the timestamps are totally bogus now.
[How]
Essentially reverts most of the original VSTARTUP series but retains
the behavior to send back events when active planes == 0.
Some refactoring/cleanup was done to not have duplicated code in both
the handlers.
Fixes: 16f17eda8bad ("drm/amd/display: Send vblank and user events at vsartup for DCN") Fixes: 3a2ce8d66a4b ("drm/amd/display: Disable VUpdate interrupt for DCN hardware") Fixes: 2b5aed9ac3f7 ("drm/amd/display: Fix pageflip event race condition for DCN.") Signed-off-by: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com> Reviewed-by: Leo Li <sunpeng.li@amd.com> Acked-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Felix Kuehling [Tue, 5 May 2020 18:02:43 +0000 (14:02 -0400)]
drm/amdgpu: Use GEM obj reference for KFD BOs
Releasing the AMDGPU BO ref directly leads to problems when BOs were
exported as DMA bufs. Releasing the GEM reference makes sure that the
AMDGPU/TTM BO is not freed too early.
Also take a GEM reference when importing BOs from DMABufs to keep
references to imported BOs balances properly.
Signed-off-by: Felix Kuehling <Felix.Kuehling@amd.com> Tested-by: Alex Sierra <alex.sierra@amd.com> Acked-by: Christian König <christian.koenig@amd.com> Reviewed-by: Alex Sierra <alex.sierra@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Evan Quan [Thu, 30 Apr 2020 06:38:39 +0000 (14:38 +0800)]
drm/amd/powerplay: perform PG ungate prior to CG ungate
Since gfxoff should be disabled first before trying to access those
GC registers.
Signed-off-by: Evan Quan <evan.quan@amd.com> Reviewed-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Evan Quan [Thu, 30 Apr 2020 03:24:02 +0000 (11:24 +0800)]
drm/amdgpu: drop unnecessary cancel_delayed_work_sync on PG ungate
As this is already properly handled in amdgpu_gfx_off_ctrl(). In fact,
this unnecessary cancel_delayed_work_sync may leave a small time window
for race condition and is dangerous.
Signed-off-by: Evan Quan <evan.quan@amd.com> Reviewed-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Evan Quan [Thu, 30 Apr 2020 03:15:19 +0000 (11:15 +0800)]
drm/amdgpu: disable MGCG/MGLS also on gfx CG ungate
Otherwise, MGCG/MGLS will be left enabled.
Signed-off-by: Evan Quan <evan.quan@amd.com> Reviewed-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Christian König [Fri, 31 Jan 2020 13:58:05 +0000 (14:58 +0100)]
drm/amdgpu: use the BAR if possible in amdgpu_device_vram_access v2
This should speed up debugging VRAM access a lot.
v2: add HDP flush/invalidate
Unrevert: RAS issue at root of the issue has been addressed
Signed-off-by: Christian König <christian.koenig@amd.com> Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com> Acked-by: Jonathan Kim <Jonathan.Kim@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Kent Russell <kent.russell@amd.com> Acked-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
drm/amd/display: Add dm support for DP 1.4 Compliance edid corruption test
It works together with drm framework
"drm: Add support for DP 1.4 Compliance edid corruption test"
Add the edid validity check scenario when edid base block is read back
with error. Send back real edid checksum and enable fail-safe mode in DC.
Signed-off-by: Jerry (Fangzhi) Zuo <Jerry.Zuo@amd.com> Reviewed-by: Harry Wentland <harry.wentland@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Arnd Bergmann [Tue, 5 May 2020 14:01:52 +0000 (16:01 +0200)]
drm/amdgpu: allocate large structures dynamically
After the structure was padded to 1024 bytes, it is no longer
suitable for being a local variable, as the function surpasses
the warning limit for 32-bit architectures:
drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c:587:5: error: stack frame size of 1072 bytes in function 'amdgpu_ras_feature_enable' [-Werror,-Wframe-larger-than=]
int amdgpu_ras_feature_enable(struct amdgpu_device *adev,
^
Use kzalloc() instead to get it from the heap.
Fixes: a0d254820f43 ("drm/amdgpu: update RAS TA to Host interface") Acked-by: Christian König <christian.koenig@amd.com> Signed-off-by: Arnd Bergmann <arnd@arndb.de> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
drm/amdgpu: Avoid integer overflow in amdgpu_device_suspend_display_audio
When building with Clang:
drivers/gpu/drm/amd/amdgpu/amdgpu_device.c:4160:53: warning: overflow in
expression; result is -294967296 with type 'long' [-Winteger-overflow]
expires = ktime_get_mono_fast_ns() + NSEC_PER_SEC * 4L;
^
1 warning generated.
Multiplication happens first due to order of operations and both
NSEC_PER_SEC and 4 are long literals so the expression overflows. To
avoid this, make 4 an unsigned long long literal, which matches the
type of expires (u64).
Fixes: 3f12acc8d6d4 ("drm/amdgpu: put the audio codec into suspend state before gpu reset V3") Link: https://github.com/ClangBuiltLinux/linux/issues/1017 Signed-off-by: Nathan Chancellor <natechancellor@gmail.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
drm/amd/display: Prevent dpcd reads with passive dongles
[why]
During hotplug, a DP port may be connected to the sink through
passive adapter which does not support DPCD reads. Issuing reads
without checking for this condition will result in errors
[how]
Ensure the link is in aux_mode before initiating operation that result
in a DPCD read.
Signed-off-by: Aurabindo Pillai <aurabindo.pillai@amd.com> Reviewed-by: Harry Wentland <Harry.Wentland@amd.com> Acked-by: Aurabindo Pillai <aurabindo.pillai@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
[Why]
Want to make use of detile buffer of all 4 pipes to maximize amount of
data stored to hide certain memory latency cases.
[How]
In case of 1 plane and 1 stream, program 4 pipes to each retrieve 1/4 of
plane later mixed together by the MPCs. Added support for transition
from 4 to 1 MPC to 2 to 1 MPC or no pipe split case and vice versa.
Currently, only enabled if debug flag is set.
yanyan kang [Mon, 20 Apr 2020 06:40:43 +0000 (14:40 +0800)]
drm/amd/display: The external monitor will show gray screen during SUT reboot
[Why]
same with CL#1711022(correcting yuv420 black color in function dcn10_blank_pixel_data,program_scaler),
yuv420 black color also needs to be correct when enabling HDMI stream at the resume procedure.
[How]
correcting the yuv420 black color according to the way how 420 is packed :2 channels carry Y component,
1 channel alternate between Cb and Cr.
Signed-off-by: yanyan kang <Yanyan.Kang@amd.com> Reviewed-by: Eric Yang <eric.yang2@amd.com> Acked-by: Aurabindo Pillai <aurabindo.pillai@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Anthony Koo [Wed, 22 Apr 2020 02:01:30 +0000 (22:01 -0400)]
drm/amd/display: add addition dc type to translate to dmub fw type
[Why]
For a type like PSR version, it makes sense for most of the code
to include a dc type, instead of having this a fw type define since
this is a capability and type exposed by dc.
Especially if it doesn't even need to communicate with the fw.
The code that is packing the firmware command message
should be the one who needs to translate the psr version
into a command that the firmware understands.
[How]
Add a dc_psr_version enum.
Signed-off-by: Anthony Koo <Anthony.Koo@amd.com> Reviewed-by: Aric Cyr <Aric.Cyr@amd.com> Acked-by: Aurabindo Pillai <aurabindo.pillai@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Alvin Lee [Tue, 21 Apr 2020 16:29:12 +0000 (12:29 -0400)]
drm/amd/display: Allow PState switch in VBLANK one display VACTIVE
[Why]
For certain display configurations we want to allow PSTATE
switch when one display can switch in VACTIVE and the
other display can switch in VBLANK
[How]
Add extra condition to dcn2 pstate support check
Signed-off-by: Alvin Lee <alvin.lee2@amd.com> Reviewed-by: Jun Lei <Jun.Lei@amd.com> Acked-by: Aurabindo Pillai <aurabindo.pillai@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Roman Li [Wed, 26 Feb 2020 22:30:29 +0000 (17:30 -0500)]
drm/amd/display: fix counter in wait_for_no_pipes_pending
[Why]
Wait counter is not being reset for each pipe.
[How]
Move counter reset into pipe loop scope.
Signed-off-by: Roman Li <roman.li@amd.com> Reviewed-by: Zhan Liu <Zhan.Liu@amd.com> Acked-by: Aurabindo Pillai <aurabindo.pillai@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
drm/amd/display: Mode change with same timing causing long display blank
[Why]
What a mode change is requested for the same timing a full stream reset
can occur in some cases which causes monitor to blank for a few seconds.
[How]
Do not consider infoframe updates as needing a full stream reset as they
will be handled on the first flip after a modeset when surface
information is available.
Signed-off-by: Aric Cyr <aric.cyr@amd.com> Reviewed-by: Anthony Koo <Anthony.Koo@amd.com> Acked-by: Aurabindo Pillai <aurabindo.pillai@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Sung Lee [Mon, 20 Apr 2020 15:38:30 +0000 (11:38 -0400)]
drm/amd/display: Update DCN2.1 DV Code Revision
[WHY & HOW]
There is a problem in hscale_pixel_rate, the bug
causes DCN to be more optimistic (more likely to underflow)
in upscale cases during prefetch.
This commit ports the fix from DV code to address these issues.
Signed-off-by: Sung Lee <sung.lee@amd.com> Reviewed-by: Dmytro Laktyushkin <Dmytro.Laktyushkin@amd.com> Acked-by: Aurabindo Pillai <aurabindo.pillai@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
amd/amdgpu: Limit rlcg write registers only for nv12
Create gfx_v10_0_rlc_funcs_sriov for nv12 with rlcg_write function pointers be initialized
so driver can use RLCG to write aceess CSIB and CP_ME_CNTL registers when nv12 in sriov mode
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>
Yong Zhao [Wed, 4 Mar 2020 21:52:41 +0000 (16:52 -0500)]
drm/amdkfd: Use a systematic method to calculate queue mask bit
The queue mask used for set_resources always assumes the queue number
per pipe is 8, so KFD needs to align with that by using function
amdgpu_queue_mask_bit_to_set_resource_bit().
Signed-off-by: Yong Zhao <Yong.Zhao@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Hawking Zhang [Thu, 16 Jan 2020 17:10:52 +0000 (01:10 +0800)]
drm/amdgpu: skip reservation of discovery tmr region in pre-Navi
IP discovery is only supported in Navi series and onwards.
There is no need to reserve a portion of vram as discovery
tmr region for pre-Navi adapters.
drm/amdgpu: change HWIP from UVD to VCN for VCN2.5
Rename hardware IP name from UVD to VCN to reduce confusion.
Hardware IP name UVD and VCN are equivalent for VCN2.5 asics.
Use name VCN for future VCN based asics.
V2: update description
V3: rebase
Signed-off-by: Boyuan Zhang <boyuan.zhang@amd.com> Reviewed-by: James Zhu <james.zhu@amd.com> Reviewed-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Felix Kuehling [Wed, 29 Apr 2020 23:38:19 +0000 (19:38 -0400)]
drm/amdgpu: Add missing parameter description in comments
Kerneldoc comments should describe all function parameters.
Signed-off-by: Felix Kuehling <Felix.Kuehling@amd.com> Reviewed-by: Kent Russell <kent.russell@amd.com> Reviewed-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Felix Kuehling [Wed, 29 Apr 2020 23:36:06 +0000 (19:36 -0400)]
drm/amdkfd: Fix comment formatting
Corrected two function names. Added a missing space.
Signed-off-by: Felix Kuehling <Felix.Kuehling@amd.com> Reviewed-by: Kent Russell <kent.russell@amd.com> Reviewed-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Oak Zeng [Fri, 1 Jun 2018 21:25:06 +0000 (17:25 -0400)]
drm/amdgpu: Changed CU reservation golden settings
With previous golden settings, compute task can't use
reserved LDS (32K) on CU0 and CU1. On 64K LDS system,
if compute work group allocate more than 32K LDS, then
it can't be dispatched to CU0 and CU1 because of the
reservation. This enables compute task to use reserved
LDS on CU0 and CU1.
Signed-off-by: Oak Zeng <Oak.Zeng@amd.com> Acked-by: Alex Deucher <alexander.deucher@amd.com> Reviewed-by: Felix Kuehling <felix.kuehling@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Felix Kuehling <Felix.Kuehling@amd.com>
Ori Messinger [Wed, 21 Aug 2019 14:48:43 +0000 (10:48 -0400)]
drm/amdkfd: Report domain with topology
PCI domain has moved to 32-bits to accommodate virtualization,
so a 32-bit integer is exposed for domain to reflect this change.
Domain can be found in here:
/sys/class/kfd/kfd/topology/nodes/X/properties
Where X is the card number
Signed-off-by: Ori Messinger <ori.messinger@amd.com> Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com> Signed-off-by: Felix Kuehling <Felix.Kuehling@amd.com> Acked-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Zheng Bin [Thu, 30 Apr 2020 02:26:53 +0000 (10:26 +0800)]
drm/amdgpu: remove set but not used variable 'direct_poll' in vcn_v2_5.c
Fixes gcc '-Wunused-but-set-variable' warning:
drivers/gpu/drm/amd/amdgpu/vcn_v2_5.c:1170:39: warning: variable ‘direct_poll’ set but not used [-Wunused-but-set-variable]
It is introduced by commit 7daaebfea5e5 ("drm/amdgpu:
add VCN2.5 sriov start for Arctrus"), but never used,
so remove it.
Reviewed-by: Christian König <christian.koenig@amd.com> Reported-by: Hulk Robot <hulkci@huawei.com> Signed-off-by: Zheng Bin <zhengbin13@huawei.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Zheng Bin [Thu, 30 Apr 2020 02:26:52 +0000 (10:26 +0800)]
drm/amdgpu: remove set but not used variable 'direct_poll' in vcn_v2_0.c
Fixes gcc '-Wunused-but-set-variable' warning:
drivers/gpu/drm/amd/amdgpu/vcn_v2_0.c:1848:39: warning: variable ‘direct_poll’ set but not used [-Wunused-but-set-variable]
It is introduced by commit dd26858a9cd8 ("drm/amdgpu:
implement initialization part on VCN2.0 for SRIOV"), but never used,
so remove it.
Reviewed-by: Christian König <christian.koenig@amd.com> Reported-by: Hulk Robot <hulkci@huawei.com> Signed-off-by: Zheng Bin <zhengbin13@huawei.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Zheng Bin [Thu, 30 Apr 2020 02:26:51 +0000 (10:26 +0800)]
drm/amdgpu: remove set but not used variable 'priority'
Fixes gcc '-Wunused-but-set-variable' warning:
drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c:1211:26: warning: variable ‘priority’ set but not used
It is not used since commit 33abcb1f5a17 ("drm/amdgpu:
set compute queue priority at mqd_init")
Reviewed-by: Christian König <christian.koenig@amd.com> Reported-by: Hulk Robot <hulkci@huawei.com> Signed-off-by: Zheng Bin <zhengbin13@huawei.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Daniel Kolesa [Wed, 29 Apr 2020 15:02:36 +0000 (17:02 +0200)]
drm/amd/display: work around fp code being emitted outside of DC_FP_START/END
The dcn20_validate_bandwidth function would have code touching the
incorrect registers emitted outside of the boundaries of the
DC_FP_START/END macros, at least on ppc64le. Work around the
problem by wrapping the whole function instead.
Signed-off-by: Daniel Kolesa <daniel@octaforge.org> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Bernard Zhao [Sun, 26 Apr 2020 13:12:44 +0000 (06:12 -0700)]
drm/radeon: cleanup coding style a bit
Maybe no need to check ws before kmalloc, kmalloc will check
itself, kmalloc`s logic is if ptr is NULL, kmalloc will just
return
Reviewed-by: Christian König <christian.koenig@amd.com> Signed-off-by: Bernard Zhao <bernard@vivo.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Evan Quan [Tue, 21 Apr 2020 02:47:35 +0000 (10:47 +0800)]
drm/amdgpu: put the audio codec into suspend state before gpu reset V3
At default, the autosuspend delay of audio controller is 3S. If the
gpu reset is triggered within 3S(after audio controller idle),
the audio controller may be unable into suspended state. Then
the sudden gpu reset will cause some audio errors. The change
here is targeted to resolve this.
However if the audio controller is in use when the gpu reset
triggered, this change may be still not enough to put the
audio controller into suspend state. Under this case, the
gpu reset will still proceed but there will be a warning
message printed("failed to suspend display audio").
V2: limit this for BACO and mode1 reset only
V3: try 1st to use pm_runtime_autosuspend_expiration() to
query how much time is left. Use default setting on
failure
Signed-off-by: Evan Quan <evan.quan@amd.com> Reviewed-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Jason Yan [Tue, 28 Apr 2020 06:31:01 +0000 (14:31 +0800)]
drm/amd/display: remove duplicate assignment of dcn21_funcs members
Fix the following coccicheck warning:
drivers/gpu/drm/amd/display/dc/dcn21/dcn21_init.c:31:51-52:
exit_optimized_pwr_state: first occurrence line 86, second occurrence
line 92
drivers/gpu/drm/amd/display/dc/dcn21/dcn21_init.c:31:51-52:
optimize_pwr_state: first occurrence line 85, second occurrence line 91
drivers/gpu/drm/amd/display/dc/dcn21/dcn21_init.c:31:51-52:
set_cursor_attribute: first occurrence line 71, second occurrence line
89
drivers/gpu/drm/amd/display/dc/dcn21/dcn21_init.c:31:51-52:
set_cursor_position: first occurrence line 70, second occurrence line 88
drivers/gpu/drm/amd/display/dc/dcn21/dcn21_init.c:31:51-52:
set_cursor_sdr_white_level: first occurrence line 72, second occurrence
line 90
Signed-off-by: Jason Yan <yanaijie@huawei.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
drm/amdkfd: Enable over-subscription with >1 GWS queue
The current GWS usage model will only allows a single GWS-enabled
process to be active on the GPU at once. This ensures that a
barrier-using kernel gets a known amount of GPU hardware, to
prevent deadlock due to inability to go beyond the GWS barrier.
The HWS watches how many GWS entries are assigned to each process,
and goes into over-subscription mode when two processes need more
than the 64 that are available. The current KFD method for working
with this is to allocate all 64 GWS entries to each GWS-capable
process.
When more than one GWS-enabled process is in the runlist, we must
make sure the runlist is in over-subscription mode, so that the
HWS gets a chained RUN_LIST packet and continues scheduling
kernels.
Signed-off-by: Joseph Greathouse <Joseph.Greathouse@amd.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>
Rather than only enabling GWS support based on the hws_gws_support
modparm, also check whether the GPU's HWS firmware supports GWS.
Leave the old modparm in place in case users want to test GWS
on GPUs not yet in the support list.
v2: fix broken syntax from the first patch.
Signed-off-by: Joseph Greathouse <Joseph.Greathouse@amd.com> Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Oak Zeng [Tue, 7 May 2019 03:11:14 +0000 (22:11 -0500)]
drm/amdkfd: New IOCTL to allocate queue GWS (v2)
Add a new kfd ioctl to allocate queue GWS. Queue
GWS is released on queue destroy.
v2: re-introduce this API with the following fixes squashed in:
- drm/amdkfd: fix null pointer dereference on dev
- drm/amdkfd: Return proper error code for gws alloc API
- drm/amdkfd: Remove GPU ID in GWS queue creation
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>
Alex Sierra [Fri, 10 Apr 2020 16:59:47 +0000 (11:59 -0500)]
drm/amdgpu: pass unlocked flag to params at amdgpu_vm_bo_update_mapping
Pass unlocked flag value to amdgpu_vm_update_params.unlocked
struct member at amdgpu_vm_bo_update_mapping.
Reviewed-by: Christian König <christian.koenig@amd.com> Signed-off-by: Alex Sierra <alex.sierra@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Christian König [Tue, 7 Apr 2020 12:54:23 +0000 (14:54 +0200)]
drm/amdgpu: add new unlocked flag for PTE updates
For HMM support we need the ability to invalidate PTEs from
a MM callback where we can't lock the root PD.
Add a new flag to better support this instead of assuming
that all invalidation updates are unlocked.
Signed-off-by: Christian König <christian.koenig@amd.com> Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Christian König [Wed, 1 Apr 2020 10:50:56 +0000 (12:50 +0200)]
drm/amdgpu: rename direct to immediate for VM updates
To avoid confusion with direct ring submissions rename bottom
of pipe VM table changes to immediate updates.
Signed-off-by: Christian König <christian.koenig@amd.com> Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Christian König [Wed, 1 Apr 2020 09:18:21 +0000 (11:18 +0200)]
drm/amdgpu: cleanup IB pool handling a bit
Fix the coding style, move and rename the definitions to
better match what they are supposed to be doing.
Signed-off-by: Christian König <christian.koenig@amd.com> Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Christian König [Mon, 16 Mar 2020 13:33:33 +0000 (14:33 +0100)]
drm/amdgpu: partial revert VM sync changes
We still need to add the VM update fences to the root PD.
So make sure to never sync to those implicitely.
Signed-off-by: Christian König <christian.koenig@amd.com> Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Alex Deucher [Thu, 23 Apr 2020 20:45:10 +0000 (16:45 -0400)]
drm/amdgpu: check ring type for secure IBs
We don't support secure operation on compute rings at the
moment so reject them.
Reviewed-by: Andrey Grodzovsky <andrey.grodzovsky@amd.com> Reviewed-by: Aaron Liu <aaron.liu@amd.com> Reviewed-by: Christian König <christian.koenig@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Christian König [Fri, 3 Apr 2020 13:56:12 +0000 (15:56 +0200)]
drm/amdgpu: fix size calculation in amdgpu_ttm_copy_mem_to_mem
When the node is larger than 4GB we overrun the size calculation.
Fix this by correctly limiting the size to the window as well.
Signed-off-by: Christian König <christian.koenig@amd.com> Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Christian König [Wed, 18 Mar 2020 09:28:20 +0000 (10:28 +0100)]
drm/amdgpu: cleanup amdgpu_ttm_copy_mem_to_mem and amdgpu_map_buffer v2
Cleanup amdgpu_ttm_copy_mem_to_mem by using fewer variables
for the same value.
Rename amdgpu_map_buffer to amdgpu_ttm_map_buffer, move it
to avoid the forward decleration, cleanup by moving the map
decission into the function and add some documentation.
No functional change.
v2: add some more cleanup suggested by Felix
Signed-off-by: Christian König <christian.koenig@amd.com> Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com> Reviewed-by: Huang Rui <ray.huang@amd.com> Tested-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Huang Rui [Mon, 9 Mar 2020 18:52:06 +0000 (14:52 -0400)]
drm/amdgpu: Fix per-IB secure flag GFX hang
Since commit "Move to a per-IB secure flag (TMZ)",
we've been seeing hangs in GFX. We need to send
FRAME CONTROL stop/start back-to-back, every time
we flip the TMZ flag. That is, when we transition
from TMZ to non-TMZ we have to send a stop with
TMZ followed by a start with non-TMZ, and
similarly for transitioning from non-TMZ into TMZ.
This patch implements this, thus fixing the GFX
hang.
v1 -> v2:
As suggested by Luben, and accept part of implemetation from this patch:
- Put "secure" closed to the loop and use optimization
- Change "secure" to bool again, and move "secure == -1" out of loop.
v3: Small fixes/optimizations.
Reported-and-Tested-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com> Signed-off-by: Huang Rui <ray.huang@amd.com> Reviewed-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Luben Tuikov <luben.tuikov@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Luben Tuikov [Wed, 26 Feb 2020 20:30:36 +0000 (15:30 -0500)]
drm/amdgpu: Fine-grained TMZ support
Add fine-grained per-ASIC TMZ support.
At the moment TMZ support is experimental for all
ASICs which support it.
Signed-off-by: Luben Tuikov <luben.tuikov@amd.com> Reviewed-by: Alex Deucher <alexander.deucher@amd.com> Reviewed-by: Christian König <christian.koenig@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Christian König [Mon, 2 Mar 2020 12:00:07 +0000 (13:00 +0100)]
drm/amdgpu: stop evicting encrypted BOs to swap
Swapping out encrypted BOs doesn't work because they can't change
their physical location without going through a bounce copy.
As a workaround disable evicting encrypted BOs to the system
domain for now.
Signed-off-by: Christian König <christian.koenig@amd.com> Reviewed-by: Alex Deucher <alexander.deucher@amd.com> Tested-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Christian König [Fri, 28 Feb 2020 13:48:06 +0000 (14:48 +0100)]
drm/amdgpu: add TMZ handling to amdgpu_move_blit
This way we should be at least able to move buffers from VRAM to GTT.
Signed-off-by: Christian König <christian.koenig@amd.com> Reviewed-by: Alex Deucher <alexander.deucher@amd.com> Tested-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Christian König [Fri, 6 Mar 2020 19:36:43 +0000 (14:36 -0500)]
drm/amdgpu: also add the TMZ flag to GART
This is necessary for TMZ handling during buffer moves and scanout.
Signed-off-by: Christian König <christian.koenig@amd.com> Reviewed-by: Alex Deucher <alexander.deucher@amd.com> Tested-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Huang Rui [Tue, 18 Feb 2020 05:07:42 +0000 (13:07 +0800)]
drm/amdgpu: fix the wrong logic checking when secure buffer is created (v3)
While the current amdgpu doesn't support TMZ, it will return the error if user
mode would like to allocate secure buffer.
v2: we didn't need this checking anymore.
v3: only print message once time.
Signed-off-by: Huang Rui <ray.huang@amd.com> Reviewed-by: Christian König <christian.koenig@amd.com> Acked-by: Nirmoy Das <Nirmoy.Das@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Luben Tuikov [Thu, 19 Mar 2020 20:47:51 +0000 (16:47 -0400)]
drm/amdgpu: implement TMZ accessor (v3)
Implement an accessor of adev->tmz.enabled. Let not
code around access it as "if (adev->tmz.enabled)"
as the organization may change. Instead...
Recruit "bool amdgpu_is_tmz(adev)" to return
exactly this Boolean value. That is, this function
is now an accessor of an already initialized and
set adev and adev->tmz.
Add "void amdgpu_gmc_tmz_set(adev)" to check and
set adev->gmc.tmz_enabled at initialization
time. After which one uses "bool
amdgpu_is_tmz(adev)" to query whether adev
supports TMZ.
Also, remove circular header file include.
v2: Remove amdgpu_tmz.[ch] as requested.
v3: Move TMZ into GMC.
Signed-off-by: Luben Tuikov <luben.tuikov@amd.com> Acked-by: Christian König <christian.koenig@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Huang Rui [Tue, 14 Jan 2020 10:55:22 +0000 (18:55 +0800)]
drm/amdgpu: remove the alignment placeholder for secure buffer
The alignment should match the page size for secure buffer, so we didn't
configure it anymore.
Signed-off-by: Huang Rui <ray.huang@amd.com> Reviewed-by: Aaron Liu <aaron.liu@amd.com> Reviewed-by: Christian König <christian.koenig@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Alex Deucher [Wed, 27 Nov 2019 20:55:35 +0000 (15:55 -0500)]
drm/amdgpu: move CS secure flag next the structs where it's used
So it's not mixed up with the CTX stuff.
Reviewed-by: Zhan Liu <zhan.liu@amd.com> Reviewed-by: Christian König <christian.koenig@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Aaron Liu [Tue, 15 Oct 2019 08:47:44 +0000 (16:47 +0800)]
drm/amdgpu: enable TMZ bit in sdma copy pkt for sdma v4
Enable sdma TMZ mode via setting TMZ bit in sdma copy pkt
for sdma v4
Signed-off-by: Aaron Liu <aaron.liu@amd.com> Reviewed-by: Christian König <christian.koenig@amd.com> Reviewed-by: Huang Rui <ray.huang@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Aaron Liu [Tue, 15 Oct 2019 07:45:23 +0000 (15:45 +0800)]
drm/amdgpu: expand amdgpu_copy_buffer interface with tmz parameter
This patch expands amdgpu_copy_buffer interface with tmz parameter.
Signed-off-by: Aaron Liu <aaron.liu@amd.com> Reviewed-by: Christian König <christian.koenig@amd.com> Reviewed-by: Huang Rui <ray.huang@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Aaron Liu [Tue, 15 Oct 2019 07:37:48 +0000 (15:37 +0800)]
drm/amdgpu: expand sdma copy_buffer interface with tmz parameter
This patch expands sdma copy_buffer interface with tmz parameter.
Signed-off-by: Aaron Liu <aaron.liu@amd.com> Reviewed-by: Christian König <christian.koenig@amd.com> Reviewed-by: Huang Rui <ray.huang@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Alex Deucher [Thu, 8 Aug 2019 03:32:46 +0000 (22:32 -0500)]
drm/amdgpu: set TMZ bits in PTEs for secure BO (v4)
If a buffer object is secure, i.e. created with
AMDGPU_GEM_CREATE_ENCRYPTED, then the TMZ bit of
the PTEs that belong the buffer object should be
set.
v1: design and draft the skeletion of TMZ bits setting on PTEs (Alex)
v2: return failure once create secure BO on non-TMZ platform (Ray)
v3: amdgpu_bo_encrypted() only checks the BO (Luben)
v4: move TMZ flag setting into amdgpu_vm_bo_update (Christian)
Signed-off-by: Alex Deucher <alexander.deucher@amd.com> Reviewed-by: Huang Rui <ray.huang@amd.com> Signed-off-by: Huang Rui <ray.huang@amd.com> Signed-off-by: Luben Tuikov <luben.tuikov@amd.com> Reviewed-by: Alex Deucher <alexander.deucher@amd.com> Reviewed-by: Christian König <christian.koenig@amd.com>
Huang Rui [Thu, 8 Aug 2019 12:05:15 +0000 (20:05 +0800)]
drm/amdgpu: job is secure iff CS is secure (v5)
Mark a job as secure, if and only if the command
submission flag has the secure flag set.
v2: fix the null job pointer while in vmid 0
submission.
v3: Context --> Command submission.
v4: filling cs parser with cs->in.flags
v5: move the job secure flag setting out of amdgpu_cs_submit()
Signed-off-by: Huang Rui <ray.huang@amd.com> Signed-off-by: Luben Tuikov <luben.tuikov@amd.com> Reviewed-by: Alex Deucher <alexander.deucher@amd.com> Reviewed-by: Christian König <christian.koenig@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
drm/amdgpu: expand the context control interface with trust flag
This patch expands the context control function to support trusted flag while we
want to set command buffer in trusted mode.
Signed-off-by: Huang Rui <ray.huang@amd.com> Reviewed-by: Alex Deucher <alexander.deucher@amd.com> Acked-by: Christian König <christian.koenig@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Huang Rui [Mon, 9 Mar 2020 17:53:37 +0000 (13:53 -0400)]
drm/amdgpu: expand the emit tmz interface with trusted flag
This patch expands the emit_tmz function to support trusted flag while we want
to set command buffer in trusted mode.
Signed-off-by: Huang Rui <ray.huang@amd.com> Reviewed-by: Alex Deucher <alexander.deucher@amd.com> Acked-by: Christian König <christian.koenig@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Huang Rui [Thu, 8 Aug 2019 09:00:16 +0000 (17:00 +0800)]
drm/amdgpu: add tmz bit in frame control packet
This patch adds tmz bit in frame control pm4 packet, and it will used in future.
Signed-off-by: Huang Rui <ray.huang@amd.com> Reviewed-by: Alex Deucher <alexander.deucher@amd.com> Reviewed-by: Christian König <christian.koenig@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Huang Rui [Thu, 19 Mar 2020 20:46:10 +0000 (16:46 -0400)]
drm/amdgpu: add function to check tmz capability (v4)
Add a function to check tmz capability with kernel parameter and ASIC type.
v2: use a per device tmz variable instead of global amdgpu_tmz.
v3: refine the comments for the function. (Luben)
v4: add amdgpu_tmz.c/h for future use.
Signed-off-by: Huang Rui <ray.huang@amd.com> Reviewed-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Huang Rui [Sun, 4 Aug 2019 08:33:16 +0000 (16:33 +0800)]
drm/amdgpu: add tmz feature parameter (v2)
This patch adds tmz parameter to enable/disable
the feature in the amdgpu kernel module. Nomally,
by default, it should be auto (rely on the
hardware capability).
But right now, it need to set "off" to avoid
breaking other developers' work because it's not
totally completed.
Will set "auto" till the feature is stable and
completely verified.
v2: add "auto" option for future use.
Signed-off-by: Huang Rui <ray.huang@amd.com> Reviewed-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
drm/amdgpu: add UAPI to create secure commands (v3)
Add a flag to the command submission IOCTL
structure which when present indicates that this
command submission should be treated as
secure. The kernel driver uses this flag to
determine whether the engine should be
transitioned to secure or unsecure, or the work
can be submitted to a secure queue depending on
the IP.
v3: the flag is now at command submission IOCTL
Signed-off-by: Luben Tuikov <luben.tuikov@amd.com> Reviewed-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
projects / mirror_ubuntu-kernels.git / log