because if the fence is really signaled, it could already
released so the fence pointer is a wild pointer, but if
we use job->base.node we are safe because job will not
be released untill amdgpu_job_timedout finished.
Signed-off-by: Monk Liu <Monk.Liu@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
1,TDR will kickout guilty job if it hang exceed the threshold
of the given one from kernel paramter "job_hang_limit", that
way a bad command stream will not infinitly cause GPU hang.
by default this threshold is 1 so a job will be kicked out
after it hang.
2,if a job timeout TDR routine will not reset all sched/ring,
instead if will only reset on the givn one which is indicated
by @job of amdgpu_sriov_gpu_reset, that way we don't need to
reset and recover each sched/ring if we already know which job
cause GPU hang.
3,unblock sriov_gpu_reset for AI family.
V2:
1:put kickout guilty job after sched parked.
2:since parking scheduler prior to kickout already occupies a
while, we can do last check on the in question job before
doing hw_reset.
TODO:
1:when a job is considered as guilty, we should mark some flag
in its fence status flag, and let UMD side aware that this
fence signaling is not due to job complete but job hang.
2:if gpu reset cause all video memory lost, we need introduce
a new policy to implement TDR, like drop all jobs not yet
signaled, and all IOCTL on this device will return ERROR
DEVICE_LOST.
this will be implemented later.
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>
We don't need a scheduler for KIQ.
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>
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>
that way we can know which job cause hang and
can do per sched reset/recovery instead of all
sched.
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>
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>
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>
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>
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>
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>
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>
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>
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>
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>
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>
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>
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>
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>
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>
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>
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>
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>
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>
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>
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>
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>
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>
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>
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>
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>
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>
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>
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>
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>
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>
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>
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>
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>
