Commit Graph

677571 Commits

Author SHA1 Message Date
Monk Liu
4fbf87e2fe drm/amdgpu:don't invoke srio-gpu-reset in gpu-reset (v2)
because we don't want to do sriov-gpu-reset under certain
cases, so just split those two funtion and don't invoke
sr-iov one from bare-metal one.

V2:
remove debugfs_gpu_reset routine on SRIOV case.

Signed-off-by: Monk Liu <Monk.Liu@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2017-05-24 17:40:37 -04:00
Chunming Zhou
bea396726d drm/amdgpu: id reset count only is updated when used end v2
before that, we have function to check if reset happens by using reset count.
v2: always update reset count after vm flush

Signed-off-by: Chunming Zhou <David1.Zhou@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2017-05-24 17:40:36 -04:00
Chunming Zhou
b9bf33d5ac drm/amdgpu: make pipeline sync be in same place v2
v2: directly return for 'if' case.

Signed-off-by: Chunming Zhou <David1.Zhou@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2017-05-24 17:40:35 -04:00
Chunming Zhou
df83d1ebc9 drm/amdgpu: add sched sync for amdgpu job v2
this is an improvement for previous patch, the sched_sync is to store fence
that could be skipped as scheduled, when job is executed, we didn't need
pipeline_sync if all fences in sched_sync are signalled, otherwise insert
pipeline_sync still.

v2: handle error when adding fence to sync failed.

Signed-off-by: Chunming Zhou <David1.Zhou@amd.com>
Reviewed-by: Junwei Zhang <Jerry.Zhang@amd.com> (v1)
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2017-05-24 17:40:35 -04:00
Christian König
a022c54e60 drm/amdgpu: remove unsed amdgpu_gem_handle_lockup (v2)
This kind of reset handling was removed a long time ago.

v2: fix warning (Alex)

Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2017-05-24 17:40:34 -04:00
Chunming Zhou
6643be65d9 drm/amdgpu: print when gpu reset successed
Signed-off-by: Chunming Zhou <David1.Zhou@amd.com>
Reviewed-by: Roger.He <Hongbo.He@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2017-05-24 17:40:33 -04:00
Chunming Zhou
fcf0649fcc drm/amdgpu: fix ring0 failed on pro card
the root cause is vram content is lost completely after pci reset.

Signed-off-by: Chunming Zhou <David1.Zhou@amd.com>
Reviewed-by: Roger.He <Hongbo.He@amd.com>
Acked-by: Christian König <christian.koenig@amd.com>
Acked-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2017-05-24 17:40:32 -04:00
Roger.He
738f64ccc2 drm/amdgpu: extend lock range for race condition when gpu reset
to cover below case:
1. A task gart bind/unbind but not add to adev->gtt_list yet
2. at this time gpu reset, gtt only recover those gtt in adev->gtt_list

Reviewed-by: Chunming Zhou <david1.zhou@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Roger.He <Hongbo.He@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2017-05-24 17:40:31 -04:00
Alex Xie
455a7bc27c drm/amdgpu: Fix comments in source code
Signed-off-by: Alex Xie <AlexBin.Xie@amd.com>
Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2017-05-24 17:40:31 -04:00
Alex Xie
ea81a173ff drm/amdgpu: fix errors in comments.
Signed-off-by: Alex Xie <AlexBin.Xie@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2017-05-24 17:40:30 -04:00
Alex Deucher
67fb56a6dd drm/amdgpu/gfx9: move define to header file
rather than defining it locally.

Reviewed-by: Junwei Zhang <Jerry.Zhang@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2017-05-24 17:40:29 -04:00
Nikola Pajkovsky
5b9c58f997 drm/amd/amdgpu: get rid of else branch
else branch is pointless if it's right at the end of function and use
unlikely() on err path.

Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Nikola Pajkovsky <npajkovsky@suse.cz>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2017-05-24 17:40:28 -04:00
Monk Liu
503bb31be4 drm/amdgpu:cleanup flag not used
Signed-off-by: Monk Liu <Monk.Liu@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2017-05-24 17:40:27 -04:00
Monk Liu
3b4d68e993 drm/amdgpu:use FRAME_CNTL for new GFX ucode (v2)
AI affected:

CP/HW team requires KMD insert FRAME_CONTROL(end) after
the last IB and before the fence of this DMAframe.

this is to make sure the cache are flushed, and it's a must
change no matter MCBP/SR-IOV or bare-metal case because new
CP hw won't do the cache flush for each IB anymore, it just
leaves it to KMD now.

with this patch, certain MCBP hang issue when rendering
vulkan/chained-ib are resolved.

v2: drop gfx8 changes.  gfx8 is not affected (Alex)

Signed-off-by: Monk Liu <Monk.Liu@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2017-05-24 17:40:26 -04:00
Monk Liu
d951eeddfa drm/amdgpu:new PM4 entry for VI/AI
TMZ package will be used for VULKAN/CHAINED-IB MCBP

Signed-off-by: Monk Liu <Monk.Liu@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2017-05-24 17:40:26 -04:00
Monk Liu
635e713298 drm/amdgpu:change SR-IOV DMAframe scheme
According to CP/hw team requirment, to support PAL/CHAINED-IB
MCBP, kernel driver must guarantee DE_META must be inserted
right prior to the work_load DE IB (with PREEMPT flag), there
cannot be any non-work_load DE IB between-in DE_META and
work_load DE IB.

Signed-off-by: Monk Liu <Monk.Liu@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2017-05-24 17:40:25 -04:00
Monk Liu
9524354358 drm/amdgpu:unify gfx8/9 ce/de meta_data
Signed-off-by: Monk Liu <Monk.Liu@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2017-05-24 17:40:24 -04:00
Monk Liu
eaa05d5288 drm/amdgpu:cleanup indent/format for gfx_v9_0.c
Signed-off-by: Monk Liu <Monk.Liu@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2017-05-24 17:40:23 -04:00
Frank Min
b48622b088 drm/amdgpu: clean doorbell after sending init table to mmsch
According to HW design, need to clean doorbell after setup MMSCH
table.

Signed-off-by: Frank Min <Frank.Min@amd.com>
Signed-off-by: Xiangliang Yu <Xiangliang.Yu@amd.com>
Reviewed-by: Monk Liu <Monk.Liu@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2017-05-24 17:40:22 -04:00
Xiangliang Yu
034b6867a4 drm/amdgpu/virt: change AI ack-irq message to debug level
Change message to debug level as VI does.

Signed-off-by: Xiangliang Yu <Xiangliang.Yu@amd.com>
Reviewed-by: Monk Liu <Monk.Liu@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2017-05-24 17:40:22 -04:00
Xiangliang Yu
943cafb825 drm/amdgpu/psp: Do not load asd for SRIOV
If psp version doesn't match asd version, asd loading will be
failed. Add workaround to bypass it for sriov.

Signed-off-by: Daniel Wang <Daniel.Wang2@amd.com>
Signed-off-by: Xiangliang Yu <Xiangliang.Yu@amd.com>
Reviewed-by: Monk Liu <Monk.Liu@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2017-05-24 17:40:21 -04:00
Trigger Huang
5dd696ae5d drm/amdgpu: Bypass GMC/UVD/VCE hw_fini in SR-IOV
On vega10, some hw finish operations should not be applied in SR-IOV
case. This works as workaround to fix multi-VFs reboot/shutdown
issues.

Signed-off-by: Trigger Huang <trigger.huang@amd.com>
Reviewed-by: Xiangliang Yu <Xiangliang.Yu@amd.com>
Reviewed-by: Monk Liu <monk.liu@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2017-05-24 17:40:20 -04:00
Monk Liu
2cb681b6e4 drm/amdgpu:re-write sriov_reinit_early/late (v2)
1,this way we make those routines compatible with the sequence
  requirment for both Tonga and Vega10
2,ignore PSP hw init when doing TDR, because for SR-IOV device
the ucode won't get lost after VF FLR, so no need to invoke PSP
doing the ucode reloading again.

v2: squash in ARRAY_SIZE fix

Signed-off-by: Monk Liu <Monk.Liu@amd.com>
Reviewed-by: Xiangliang Yu <Xiangliang.Yu@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2017-05-24 17:40:19 -04:00
Monk Liu
17b2e332a2 drm/amdgpu:need som change on vega10 mailbox
if sriov gpu reset is invoked by job timeout, it is run
in a global work-queue which is very slow and better not call
msleep ortherwise it takes long time to get back CPU.

so make below changes:

1: Change msleep 1 to mdelay 5
2: Ignore the ack fail from pf after time out,
   because VF FLR will clear ack, sometime VF FLR is done
   prior to the beginning of poll_ack so we can ignore this ack

TODO:
Put job_timedout (and the following gpu reset) in a driver thread,
instead of the global work_struct.

Signed-off-by: Monk Liu <Monk.Liu@amd.com>
Reviewed-by: Xiangliang Yu <Xiangliang.Yu@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2017-05-24 17:40:18 -04:00
Monk Liu
3af906f0cf drm/amdgpu:fix cannot receive rcv/ack irq bug
Signed-off-by: Monk Liu <Monk.Liu@amd.com>
Reviewed-by: Xiangliang Yu <Xiangliang.Yu@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2017-05-24 17:40:18 -04:00
Monk Liu
ff82577a10 drm/amdgpu:kiq reg access need timeout(v2)
this is to prevent fence forever waiting if FLR occured
during register accessing.

v2:
use define instead of hardcode for the timeout msec

Signed-off-by: Monk Liu <Monk.Liu@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2017-05-24 17:40:17 -04:00
Alex Deucher
2fdde9fa97 drm/amdgpu/gfx9: wait for completion in KIQ init
We need to make sure the various init sequences submitted
to KIQ complete before testing the rings.

Reviewed-by: monk liu <monk.liu@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2017-05-24 17:40:16 -04:00
Alex Deucher
f1f7b44378 drm/amdgpu/gfx9: use new KIQ packet defines
Rather than magic numbers.

Reviewed-by: monk liu <monk.liu@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2017-05-24 17:40:15 -04:00
Alex Deucher
495a746354 drm/amdgpu: add KIQ packet defines to soc15d.h
Will be used in subsequent commits rather rather than
magic numbers.

Reviewed-by: monk liu <monk.liu@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2017-05-24 17:40:14 -04:00
Alex Deucher
b98724db7f drm/amdgpu/gfx9: clear the compute ring on reset
To be consistent with gfx8.

Reviewed-by: monk liu <monk.liu@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2017-05-24 17:40:13 -04:00
Alex Deucher
0ef376cacb drm/amdgpu/gfx9: create mqd backups
And properly synchronize them with the master during
queue init.

Reviewed-by: monk liu <monk.liu@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2017-05-24 17:40:13 -04:00
Shaoyun Liu
cdf6adb28f drm/amdgpu: Move kiq ring lock out of virt structure
The usage of kiq should not depend on the virtualization.

Signed-off-by: Shaoyun Liu <Shaoyun.Liu@amd.com>
Reviewed-by:Andres Rodriquez <andresx7@gmail.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2017-05-24 17:40:12 -04:00
Chunming Zhou
b98b8dbc39 drm/amdgpu: bump module verion for reserved vmid
Interface to reserve a vmid for a specific process to
add in shader debugging that requries a fixed vmid.

Signed-off-by: Chunming Zhou <David1.Zhou@amd.com>
Reviewed-by: Junwei Zhang <Jerry.Zhang@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2017-05-24 17:40:11 -04:00
Chunming Zhou
7a63eb23d8 drm/amdgpu: implement grab reserved vmid V4
Implement the vmid reservation.

v2: move sync waiting only when flush needs
v3: fix racy
v4: peek fence instead of get fence, and fix potential context starved.

Signed-off-by: Chunming Zhou <David1.Zhou@amd.com>
Reviewed-by: Junwei Zhang <Jerry.Zhang@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2017-05-24 17:40:10 -04:00
Chunming Zhou
c350577073 drm/amdgpu: add limitation for dedicated vm number v4
Limit reserved vmids to 1 to avoid taking too many
out of commission and starving the system.

v2: move #define to amdgpu_vm.h
v3: move reserved vmid counter to id_manager,
and increase counter before allocating vmid
v4: rename to reserved_vmid_num

Signed-off-by: Chunming Zhou <David1.Zhou@amd.com>
Reviewed-by: Junwei Zhang <Jerry.Zhang@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2017-05-24 17:40:09 -04:00
Chunming Zhou
1e9ef26fb3 drm/amdgpu: reserve/unreserve vmid by vm ioctl v4
add reserve/unreserve vmid funtions. Used to reserve
vmids for certain shader debugging functionality that
required a fixed vmid for the life of the debug.

v3:
only reserve vmid from gfxhub
v4:
fix racy condition

Signed-off-by: Chunming Zhou <David1.Zhou@amd.com>
Reviewed-by: Junwei Zhang <Jerry.Zhang@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2017-05-24 17:40:09 -04:00
Chunming Zhou
36bbf3bf9b drm/amdgpu: add reserved vmid field in vm struct v2
v2: rename dedicated_vmid to reserved_vmid

Signed-off-by: Chunming Zhou <David1.Zhou@amd.com>
Reviewed-by: Junwei Zhang <Jerry.Zhang@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2017-05-24 17:40:08 -04:00
Chunming Zhou
cfbcacf428 drm/amdgpu: add vm ioctl
It will be used for reserving vmid for shader debugging
that requires a fixed vmid.

v2: fix warning (Alex)

Signed-off-by: Chunming Zhou <David1.Zhou@amd.com>
Reviewed-by: Junwei Zhang <Jerry.Zhang@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2017-05-24 17:40:07 -04:00
Trigger Huang
63a7c7487f drm/amdgpu: Enable chained IB MCBP support
Support for MCBP/Virtualization in combination with chained IBs is
formal released on firmware feature version #46. So enable it
according to firmware feature version, otherwise, world switch will
hang.

Signed-off-by: Trigger Huang <trigger.huang@amd.com>
Reviewed-by: Monk Liu <monk.liu@amd.com>
Reviewed-by: Xiangliang Yu <Xiangliang.Yu@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2017-05-24 17:40:06 -04:00
Rex Zhu
fe723cd3bf drm/amdgpu:fix get wrong gfx always on cu masks.
Bug: SWDEV-117987: Always on CU mask broken for gfx7+

Signed-off-by: Rex Zhu <Rex.Zhu@amd.com>
Acked-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2017-05-24 17:40:05 -04:00
Rex Zhu
94c9ceade9 drm/amdgpu: fix s3 ring test failed on Vi caused by KIQ enabled.
Signed-off-by: Rex Zhu <Rex.Zhu@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2017-05-24 17:40:05 -04:00
Xiangliang Yu
ab276632ec drm/amdgpu/virt: change the place of virt_init_setting
Change place of virt_init_setting function so that can cover the
cg and pg flags configuration.

Signed-off-by: Xiangliang Yu <Xiangliang.Yu@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2017-05-24 17:40:04 -04:00
Xiangliang Yu
213cacefcd drm/amdgpu/virt: bypass cg and pg setting for SRIOV
GPU hypervisor cover all settings of CG and PG, so guest doesn't
need to do anything. Bypass it.

Signed-off-by: Frank Min <Frank.Min@amd.com>
Signed-off-by: Xiangliang Yu <Xiangliang.Yu@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2017-05-24 17:40:03 -04:00
Christian König
3032f350ba drm/amdgpu: drop support for per ASIC read registers
Only per family registers are still used.

Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2017-05-24 17:40:02 -04:00
Christian König
97fcc76b67 drm/amdgpu: drop support for untouched registers
I couldn't figure out what this was original good for, but we
don't use it any more.

Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2017-05-24 17:40:01 -04:00
Rex Zhu
ca541f3316 drm/amdgpu: delete redundant kiq irq funcs type check in gfx8.
Signed-off-by: Rex Zhu <Rex.Zhu@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2017-05-24 17:40:00 -04:00
Rex Zhu
2d0806cabb drm/amdgpu: fix typo in dmesg in gfx_v8_0_kiq_kcq_disable.
Signed-off-by: Rex Zhu <Rex.Zhu@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2017-05-24 17:40:00 -04:00
Xiaojie Yuan
e6f7c765e7 drm/amdgpu: add HDMI audio support for si dce6
Signed-off-by: Xiaojie Yuan <Xiaojie.Yuan@amd.com>
Reviewed-by: Edward O'Callaghan <funfunctor@folklore1984.net>
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>
2017-05-24 17:39:59 -04:00
Xiaojie Yuan
4caca70668 drm/amdgpu: add DP audio support for si dce6 (v3)
v2: refine dce_v6_0_audio_endpt_wreg() and unify inconsistent method names
v3: fix num_pins for tahiti, pitcairn, verde and oland

Signed-off-by: Xiaojie Yuan <Xiaojie.Yuan@amd.com>
Reviewed-by: Edward O'Callaghan <funfunctor@folklore1984.net>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Acked-by: Junwei Zhang <Jerry.Zhang@amd.com>
Acked-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2017-05-24 17:39:58 -04:00
Alex Deucher
6a124e675a drm/amdgpu/gfx8: move CP_PQ_STATUS after doorbell range setting (v2)
I'm not sure if the order matters, but it seems like it makes
more sense to set this after the range is programmed.

v2: rebase (Alex)

Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2017-05-24 17:39:57 -04:00