Handle all DMA IOMMU group related dependencies before the
group is removed and we try to access it after free.

v2:
Move the actul handling function to TTM
Signed-off-by: NAndrey Grodzovsky <andrey.grodzovsky@amd.com>
Reviewed-by: NChristian König <christian.koenig@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

adev->rmmio is set to be NULL in amdgpu_device_unmap_mmio to prevent
access after pci_remove, however, in SRIOV case, amdgpu_virt_release_full_gpu
will still use adev->rmmio for access after amdgpu_device_unmap_mmio.
The patch is to move such SRIOV calling earlier to fini_early stage.

Fixes: 07775fc1 ("drm/amdgpu: Unmap all MMIO mappings")
Cc: Andrey Grodzovsky <andrey.grodzovsky@amd.com>
Signed-off-by: NLeslie Shi <Yuliang.Shi@amd.com>
Signed-off-by: NGuchun Chen <guchun.chen@amd.com>
Reviewed-by: NAndrey Grodzovsky <andrey.grodzovsky@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

Separate iommu_resume from kfd_resume, and move it before
other amdgpu ip init/resume.

Bug: https://bugzilla.kernel.org/show_bug.cgi?id=211277Signed-off-by: NJames Zhu <James.Zhu@amd.com>
Reviewed-by: NFelix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>
Cc: stable@vger.kernel.org

Separate iommu_resume from kfd_resume, and move it before
other amdgpu ip init/resume.

Bug: https://bugzilla.kernel.org/show_bug.cgi?id=211277Signed-off-by: NJames Zhu <James.Zhu@amd.com>
Reviewed-by: NFelix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

The AtomicOp Requester Enable bit is reserved in VFs and the PF value applies to all
associated VFs. so guest driver can not directly enable the atomicOps for VF, it
depends on PF to enable it. In current design, amdgpu driver  will get the enabled
atomicOps bits through private pf2vf data
Signed-off-by: Nshaoyunl <shaoyun.liu@amd.com>
Reviewed-by: NFelix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

schedule_delayed_work does not push back the work if it was already
scheduled before, so amdgpu_device_delay_enable_gfx_off ran ~100 ms
after the first time GFXOFF was disabled and re-enabled, even if GFXOFF
was disabled and re-enabled again during those 100 ms.

This resulted in frame drops / stutter with the upcoming mutter 41
release on Navi 14, due to constantly enabling GFXOFF in the HW and
disabling it again (for getting the GPU clock counter).

To fix this, call cancel_delayed_work_sync when the disable count
transitions from 0 to 1, and only schedule the delayed work on the
reverse transition, not if the disable count was already 0. This makes
sure the delayed work doesn't run at unexpected times, and allows it to
be lock-free.

v2:
* Use cancel_delayed_work_sync & mutex_trylock instead of
  mod_delayed_work.
v3:
* Make amdgpu_device_delay_enable_gfx_off lock-free (Christian König)
v4:
* Fix race condition between amdgpu_gfx_off_ctrl incrementing
  adev->gfx.gfx_off_req_count and amdgpu_device_delay_enable_gfx_off
  checking for it to be 0 (Evan Quan)

Cc: stable@vger.kernel.org
Reviewed-by: NEvan Quan <evan.quan@amd.com>
Reviewed-by: Lijo Lazar <lijo.lazar@amd.com> # v3
Acked-by: Christian König <christian.koenig@amd.com> # v3
Signed-off-by: NMichel Dänzer <mdaenzer@redhat.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

schedule_delayed_work does not push back the work if it was already
scheduled before, so amdgpu_device_delay_enable_gfx_off ran ~100 ms
after the first time GFXOFF was disabled and re-enabled, even if GFXOFF
was disabled and re-enabled again during those 100 ms.

This resulted in frame drops / stutter with the upcoming mutter 41
release on Navi 14, due to constantly enabling GFXOFF in the HW and
disabling it again (for getting the GPU clock counter).

To fix this, call cancel_delayed_work_sync when the disable count
transitions from 0 to 1, and only schedule the delayed work on the
reverse transition, not if the disable count was already 0. This makes
sure the delayed work doesn't run at unexpected times, and allows it to
be lock-free.

v2:
* Use cancel_delayed_work_sync & mutex_trylock instead of
  mod_delayed_work.
v3:
* Make amdgpu_device_delay_enable_gfx_off lock-free (Christian König)
v4:
* Fix race condition between amdgpu_gfx_off_ctrl incrementing
  adev->gfx.gfx_off_req_count and amdgpu_device_delay_enable_gfx_off
  checking for it to be 0 (Evan Quan)

Cc: stable@vger.kernel.org
Reviewed-by: NEvan Quan <evan.quan@amd.com>
Reviewed-by: Lijo Lazar <lijo.lazar@amd.com> # v3
Acked-by: Christian König <christian.koenig@amd.com> # v3
Signed-off-by: NMichel Dänzer <mdaenzer@redhat.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

[Why]
In some cases when we unload driver, warning call trace
will show up in vram_mgr_fini which claims that LRU is not empty, caused
by the ttm bo inside delay deleted queue.

[How]
We should flush delayed work to make sure the delay deleting is done.
Signed-off-by: NYuBiao Wang <YuBiao.Wang@amd.com>
Reviewed-by: NAndrey Grodzovsky <andrey.grodzovsky@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

Why: Previously hw fence is alloced separately with job.
It caused historical lifetime issues and corner cases.
The ideal situation is to take fence to manage both job
and fence's lifetime, and simplify the design of gpu-scheduler.

How:
We propose to embed hw_fence into amdgpu_job.
1. We cover the normal job submission by this method.
2. For ib_test, and submit without a parent job keep the
legacy way to create a hw fence separately.
v2:
use AMDGPU_FENCE_FLAG_EMBED_IN_JOB_BIT to show that the fence is
embedded in a job.
v3:
remove redundant variable ring in amdgpu_job
v4:
add tdr sequence support for this feature. Add a job_run_counter to
indicate whether this job is a resubmit job.
v5
add missing handling in amdgpu_fence_enable_signaling
Signed-off-by: NJingwen Chen <Jingwen.Chen2@amd.com>
Signed-off-by: NJack Zhang <Jack.Zhang7@hotmail.com>
Reviewed-by: NAndrey Grodzovsky <andrey.grodzovsky@amd.com>
Reviewed by: Monk Liu <monk.liu@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

When init failed in early init stage, amdgpu_object has
not been initialized, so hasn't the ttm delayed queue functions.
Signed-off-by: NYuBiao Wang <YuBiao.Wang@amd.com>
Reviewed-by: NEmily.Deng <Emily.Deng@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

In amdgpu_fence_driver_hw_fini, no need to call drm_sched_fini to stop
scheduler in s3 test, otherwise, fence related failure will arrive
after resume. To fix this and for a better clean up, move drm_sched_fini
from fence_hw_fini to fence_sw_fini, as it's part of driver shutdown, and
should never be called in hw_fini.

v2: rename amdgpu_fence_driver_init to amdgpu_fence_driver_sw_init,
to keep sw_init and sw_fini paired.

Bug: https://gitlab.freedesktop.org/drm/amd/-/issues/1668
Fixes: 8d35a259 ("drm/amdgpu: adjust fence driver enable sequence")
Suggested-by: NChristian König <christian.koenig@amd.com>
Tested-by: NMike Lothian <mike@fireburn.co.uk>
Signed-off-by: NGuchun Chen <guchun.chen@amd.com>
Reviewed-by: NChristian König <christian.koenig@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

Fence driver was enabled per ring when sw init on per IP block before.
Change to enable all the fence driver at the same time after
amdgpu_device_ip_init finished.
Rename some function related to fence to make it reasonable for read.
Signed-off-by: NLikun Gao <Likun.Gao@amd.com>
Reviewed-by: NHawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

This reverts commit 4192f7b5.

It is not true (as stated in the reverted commit changelog) that we never
unmap the BAR on failure; it actually does happen properly on
amdgpu_driver_load_kms() -> amdgpu_driver_unload_kms() ->
amdgpu_device_fini() error path.

What's worse, this commit actually completely breaks resource freeing on
probe failure (like e.g. failure to load microcode), as
amdgpu_driver_unload_kms() notices adev->rmmio being NULL and bails too
early, leaving all the resources that'd normally be freed in
amdgpu_acpi_fini() and amdgpu_device_fini() still hanging around, leading
to all sorts of oopses when someone tries to, for example, access the
sysfs and procfs resources which are still around while the driver is
gone.

Fixes: 4192f7b5 ("drm/amdgpu: unmap register bar on device init failure")
Reported-by: NVojtech Pavlik <vojtech@ucw.cz>
Signed-off-by: NJiri Kosina <jkosina@suse.cz>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>
Cc: stable@vger.kernel.org

On Sienna Cichlid, in pass-through mode, if we unload the driver in BACO
mode(RTPM), then the kernel would receive thousands of interrupts.
That's because there is doorbell monitor interrupt on BIF, so KVM keeps
injecting interrupts to the guest VM. So we should clear the doorbell
interrupt status after BACO exit.

v2: Modify coding style and commit message
Signed-off-by: NChengzhe Liu <ChengZhe.Liu@amd.com>
Reviewed-by: NLuben Tuikov <luben.tuikov@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

Use FAMILY_NV for cyan_skillfish.
Signed-off-by: NTao Zhou <tao.zhou1@amd.com>
Reviewed-by: NHawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

Add cyan_skillfish asic family.
Signed-off-by: NTao Zhou <tao.zhou1@amd.com>
Reviewed-by: NHawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

[Why]
Currently all timedout job will be considered to be guilty. In SRIOV
multi-vf use case, the vf flr happens first and then job time out is
found. There can be several jobs timeout during a very small time slice.
And if the innocent sdma job time out is found before the real bad
job, then the innocent sdma job will be set to guilty. This will lead
to a page fault after resubmitting job.

[How]
If the job is a kernel job, we will always consider it not guilty
Reviewed-by: NChristian König <christian.koenig@amd.com>
Signed-off-by: NJingwen Chen <Jingwen.Chen2@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

The callback functions are used for SRIOV read/write instead
of just for rlcg read/write
Signed-off-by: NRoy Sun <Roy.Sun@amd.com>
Reviewed-by: NZhou pengju <pengju.zhou@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

The VGA arbitration is entirely based on pci_dev structures, so just pass
that back to the set_vga_decode callback.
Signed-off-by: NChristoph Hellwig <hch@lst.de>
Link: https://patchwork.freedesktop.org/patch/msgid/20210716061634.2446357-8-hch@lst.deAcked-by: NChristian König <christian.koenig@amd.com>
Signed-off-by: NChristian König <christian.koenig@amd.com>

All callers pass NULL as the irq_set_state argument, so remove it and
the ->irq_set_state member in struct vga_device.
Signed-off-by: NChristoph Hellwig <hch@lst.de>
Link: https://patchwork.freedesktop.org/patch/msgid/20210716061634.2446357-7-hch@lst.deAcked-by: NChristian König <christian.koenig@amd.com>
Signed-off-by: NChristian König <christian.koenig@amd.com>

Add a trivial wrapper for the unregister case that sets all fields to
NULL.
Signed-off-by: NChristoph Hellwig <hch@lst.de>
Link: https://patchwork.freedesktop.org/patch/msgid/20210716061634.2446357-6-hch@lst.deAcked-by: NChristian König <christian.koenig@amd.com>
Signed-off-by: NChristian König <christian.koenig@amd.com>

split amdgpu_device_access_vram()
1. amdgpu_device_mm_access(): using MM_INDEX/MM_DATA to access vram
2. amdgpu_device_aper_access(): using vram aperature to access vram (option)
Signed-off-by: NKevin Wang <kevin1.wang@amd.com>
Reviewed-by: NChristian König <christian.koenig@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

This reverts commit 4192f7b5.

It is not true (as stated in the reverted commit changelog) that we never
unmap the BAR on failure; it actually does happen properly on
amdgpu_driver_load_kms() -> amdgpu_driver_unload_kms() ->
amdgpu_device_fini() error path.

What's worse, this commit actually completely breaks resource freeing on
probe failure (like e.g. failure to load microcode), as
amdgpu_driver_unload_kms() notices adev->rmmio being NULL and bails too
early, leaving all the resources that'd normally be freed in
amdgpu_acpi_fini() and amdgpu_device_fini() still hanging around, leading
to all sorts of oopses when someone tries to, for example, access the
sysfs and procfs resources which are still around while the driver is
gone.

Fixes: 4192f7b5 ("drm/amdgpu: unmap register bar on device init failure")
Reported-by: NVojtech Pavlik <vojtech@ucw.cz>
Signed-off-by: NJiri Kosina <jkosina@suse.cz>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

In some asics, we need to adjust the behavior according to the apu flags
at very early stage.
Signed-off-by: NHuang Rui <ray.huang@amd.com>
Reviewed-by: NAaron Liu <aaron.liu@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

Enable gpu recovery for beige_goby.
Signed-off-by: NChengming Gui <Jack.Gui@amd.com>
Reviewed-by: NHawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

Move shadow_list to struct amdgpu_bo_vm as shadow BOs
are part of PT/PD BOs.
Signed-off-by: NNirmoy Das <nirmoy.das@amd.com>
Reviewed-by: NChristian König <christian.koenig@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

To enable output on real display instead of virtual.
Acked-by: NHuang Rui <ray.huang@amd.com>
Signed-off-by: NNicholas Kazlauskas <nicholas.kazlauskas@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

This patch adds yellow carp support for gpu_info firmware and ip
block setting.
Signed-off-by: NAaron Liu <aaron.liu@amd.com>
Reviewed-by: NHuang Rui <ray.huang@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

This patch adds yellow carp to amd_asic_type enum and amdgpu_asic_name[].
Signed-off-by: NAaron Liu <aaron.liu@amd.com>
Reviewed-by: NHuang Rui <ray.huang@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

Integrate two generic functions to determine if HDP
flush is needed for all Asics.
Signed-off-by: NEric Huang <jinhuieric.huang@amd.com>
Reviewed-by: NChristian König <christian.koenig@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

Drop DC initialization when DCN is harvested in VBIOS. The way
doesn't affect virtual display ip initialization.
Signed-off-by: NLikun Gao  <Likun.Gao@amd.com>
Signed-off-by: NAsher Song <Asher.Song@amd.com>
Reviewed-by: NGuchun Chen <guchun.chen@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

When we want to decouble resource management from buffer management we need to
be able to handle resources separately.

Add a resource pointer and rename bo->mem so that all code needs to
change to access the pointer instead.

No functional change.
Signed-off-by: NChristian König <christian.koenig@amd.com>
Reviewed-by: NMatthew Auld <matthew.auld@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20210430092508.60710-4-christian.koenig@amd.com

enable smart shift on dGPU if it is part of HG system and
the platform supports ATCS method to handle power shift.

V2: avoid psc updates in baco enter and exit (Lijo)
    fix alignment (Shashank)
V3: rebased on unified ATCS handling. (Alex)
V4: check for return value and warn on failed update (Shashank)
    return 0 if device does not support smart shift.  (Lizo)
V5: rebased on ATPX/ATCS structures global (Alex)
Signed-off-by: NSathishkumar S <sathishkumar.sundararaju@amd.com>
Reviewed-by: NLijo Lazar <lijo.lazar@amd.com>
Reviewed-by: NShashank Sharma <shashank.sharma@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

Fixes the following W=1 kernel build warning(s):

 drivers/gpu/drm/amd/amdgpu/amdgpu_device.c:4624:6: warning: no previous prototype for ‘amdgpu_device_recheck_guilty_jobs’ [-Wmissing-prototypes]

Cc: Alex Deucher <alexander.deucher@amd.com>
Cc: "Christian König" <christian.koenig@amd.com>
Cc: David Airlie <airlied@linux.ie>
Cc: Daniel Vetter <daniel@ffwll.ch>
Cc: Sumit Semwal <sumit.semwal@linaro.org>
Cc: amd-gfx@lists.freedesktop.org
Cc: dri-devel@lists.freedesktop.org
Cc: linux-media@vger.kernel.org
Cc: linaro-mm-sig@lists.linaro.org
Signed-off-by: NLee Jones <lee.jones@linaro.org>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

Problem:
drivers/gpu/drm/amd/amdgpu/amdgpu_device.c:332:1: warning: unused label 'exit' [-Wunused-label]
exit:
^~~~~

Fix: Put #ifdef CONFIG_64BIT around exit
Reported-by: Nkernel test robot <lkp@intel.com>
Signed-off-by: NAndrey Grodzovsky <andrey.grodzovsky@amd.com>
Reviewed-by: NChristian König <christian.koenig@amd.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20210525184431.1170373-1-andrey.grodzovsky@amd.com

Drop DC initialization when DCN is harvested in VBIOS. The way
doesn't affect virtual display ip initialization.
Signed-off-by: NLikun Gao  <Likun.Gao@amd.com>
Signed-off-by: NAsher Song <Asher.Song@amd.com>
Reviewed-by: NGuchun Chen <guchun.chen@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

Make it's name not feature but function descriptive.
Signed-off-by: NAndrey Grodzovsky <andrey.grodzovsky@amd.com>
Reviewed-by: NAlex Deucher <alexander.deucher@amd.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20210521204122.762288-1-andrey.grodzovsky@amd.com

This patch series are used for GC/MMHUB(part)/IH_RB_CNTL
indirect access in the SRIOV environment.

There are 4 bits, controlled by host, to control
if GC/MMHUB(part)/IH_RB_CNTL indirect access enabled.
(one bit is master bit controls other 3 bits)

For GC registers, changing all the register access from MMIO to
RLC and use RLC as the default access method in the full access time.

For partial MMHUB registers, changing their access from MMIO to
RLC in the full access time, the remaining registers
keep the original access method.

For IH_RB_CNTL register, changing it's access from MMIO to PSP.
Signed-off-by: NPeng Ju Zhou <PengJu.Zhou@amd.com>
Reviewed-by: NAlex Deucher <alexander.deucher@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

Access to those must be prevented post pci_remove

v6: Drop BOs list, unampping VRAM BAR is enough.
v8:
Add condition of xgmi.connected_to_cpu to MTTR
handling and remove MTTR handling from the old place.
Signed-off-by: NAndrey Grodzovsky <andrey.grodzovsky@amd.com>
Reviewed-by: NAlex Deucher <alexander.deucher@amd.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20210517193105.491461-1-andrey.grodzovsky@amd.com

This should prevent writing to memory or IO ranges possibly
already allocated for other uses after our device is removed.

v5:
Protect more places wher memcopy_to/form_io takes place
Protect IB submissions

v6: Switch to !drm_dev_enter instead of scoping entire code
with brackets.

v7:
Drop guard of HW ring commands emission protection since they
are in GART and not in MMIO.
Signed-off-by: NAndrey Grodzovsky <andrey.grodzovsky@amd.com>
Reviewed-by: NAlex Deucher <alexander.deucher@amd.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20210512142648.666476-10-andrey.grodzovsky@amd.com