1. Move the variables of ras block instance members from
   specific xxx_ras_fini to general ras_fini call.
2. Function calls inside the modules only use parameters
   passed from xxx_ras_fini instead of ras block instance
   members.
Signed-off-by: Nyipechai <YiPeng.Chai@amd.com>
Reviewed-by: NTao Zhou <tao.zhou1@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

Modify .ras_fini function pointer parameter so that
we can remove redundant intermediate calls in some
ras blocks.
Signed-off-by: Nyipechai <YiPeng.Chai@amd.com>
Reviewed-by: NTao Zhou <tao.zhou1@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

[Why]
PSR Power on/off is done in PSR. Add a dc_debug option
and dmub setting to use PHY implementation of this instead.

[How]
Add a dc_debug option and dmub setting to use
PHY FSM Power up/down for PSR.
Co-authored-by: NShah Dharati <dharati.shah@amd.com>
Reviewed-by: NHansen Dsouza <hansen.dsouza@amd.com>
Acked-by: NAlan Liu <HaoPing.Liu@amd.com>
Signed-off-by: NShah Dharati <dharati.shah@amd.com>
Tested-by: NDaniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

[WHY?]
Some projectors support frame alternate 3D modes at 120Hz, but DAL3 does
not create timings. Most active DP to HDMI dongles do not translate
infoframes properly to use HW packing stereo mode.

[HOW?]
Create frame alternate 3D timings for displays that support it. Disable HW
packing 3D mode on DP active dongles.
Reviewed-by: NMartin Leung <Martin.Leung@amd.com>
Acked-by: NAlan Liu <HaoPing.Liu@amd.com>
Signed-off-by: NDillon Varone <dillon.varone@amd.com>
Tested-by: NDaniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

Clang static analysis reports this error
amdgpu_debugfs.c:1690:9: warning: 1st function call
  argument is an uninitialized value
  tmp = krealloc_array(tmp, i + 1,
        ^~~~~~~~~~~~~~~~~~~~~~~~~~~

realloc uses tmp, so tmp can not be garbage.
And the return needs to be checked.

Fixes: 5ce5a584 ("drm/amdgpu: add debugfs for reset registers list")
Reviewed-by: NChristian König <christian.koenig@amd.com>
Signed-off-by: NTom Rix <trix@redhat.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

[Why]
HDMI Compliance requires VIC to be set to 0
on 2D mode if HDMI_VIC is present.

[How]
When VIC and HDMI_VIC is both present,
reset VIC to 0.
Reviewed-by: NMartin Leung <Martin.Leung@amd.com>
Acked-by: NAlan Liu <HaoPing.Liu@amd.com>
Signed-off-by: NChris Park <Chris.Park@amd.com>
Tested-by: NDaniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

[Why & How]
To align coding style for how we use this across DCN. The resource
creation ones can remain static, however.
Reviewed-by: NEric Yang <Eric.Yang2@amd.com>
Acked-by: NAlan Liu <HaoPing.Liu@amd.com>
Signed-off-by: NNicholas Kazlauskas <nicholas.kazlauskas@amd.com>
Tested-by: NDaniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

RDPCS programming is done in DMUB remove legacy invalid code
Reviewed-by: NAlvin Lee <Alvin.Lee2@amd.com>
Acked-by: NAlan Liu <HaoPing.Liu@amd.com>
Signed-off-by: NHansen Dsouza <Hansen.Dsouza@amd.com>
Tested-by: NDaniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

To remove duplicate code, unify event message format and simplify new
event add in the following patches.

Use KFD_SMI_EVENT_MSG_SIZE to define msg size, the same size will be
used in user space to alloc the msg receive buffer.
Signed-off-by: NPhilip Yang <Philip.Yang@amd.com>
Reviewed-by: NFelix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

sizeof(buf) is 8 bytes because it is defined as unsigned char *buf,
each SMI event read only copy max 8 bytes to user buffer. Correct this
by using the buf allocate size.
Signed-off-by: NPhilip Yang <Philip.Yang@amd.com>
Reviewed-by: NFelix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

This reverts commit 3abfe30d.

To fix deadlock in kFDSVMEvictTest when xnack off.
Signed-off-by: NPhilip Yang <Philip.Yang@amd.com>
Reviewed-by: NFelix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

In file vega10_hwmgr.c, the names of struct vega10_power_state *
and struct pp_power_state * are confusingly used, which may lead
to some confusion.

Status quo is that variables of type struct vega10_power_state *
are named "vega10_ps", "ps", "vega10_power_state". A more
appropriate usage is that struct are named "ps" is used for
variabled of type struct pp_power_state *.

So rename struct vega10_power_state * which are named "ps" and
"vega10_power_state" to "vega10_ps", I also renamed "psa" to
"vega10_psa" and "psb" to "vega10_psb" to make it more clearly.

The rows longer than 100 columns are involved.
Reviewed-by: NEvan Quan <evan.quan@amd.com>
Signed-off-by: NMeng Tang <tangmeng@uniontech.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

Don't fill up the logs with:

[253557.859575] [drm:amdgpu_dm_atomic_check [amdgpu]] DSC precompute is not needed.
[253557.892966] [drm:amdgpu_dm_atomic_check [amdgpu]] DSC precompute is not needed.
[253557.926070] [drm:amdgpu_dm_atomic_check [amdgpu]] DSC precompute is not needed.
[253557.959344] [drm:amdgpu_dm_atomic_check [amdgpu]] DSC precompute is not needed.

which prints many times a second, when the kernel is run with
drm.debug=2.

Instead of DRM_DEBUG_DRIVER(), make it DRM_INFO_ONCE().

Cc: Alex Deucher <Alexander.Deucher@amd.com>
Cc: Roman Li <Roman.Li@amd.com>
Cc: Felix Kuehling <felix.kuehling@amd.com>
Cc: Hersen Wu <hersenwu@amd.com>
Cc: Daniel Wheeler <daniel.wheeler@amd.com>
Fixes: 17ce8a69 ("drm/amd/display: Add dsc pre-validation in atomic check")
Signed-off-by: NLuben Tuikov <luben.tuikov@amd.com>
Reviewed-by: NRoman Li <Roman.Li@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

mtk_drm_crtc_atomic_disable will send an async cmd to cmdq driver,
so it may not finish when cmdq_suspend is called sometimes.

Add wait_for_event after sending async disable plane cmd to make
sure the lastest cmd is done before cmdq_suspend.
Signed-off-by: Njason-jh.lin <jason-jh.lin@mediatek.com>
Signed-off-by: NChun-Kuang Hu <chunkuang.hu@kernel.org>
Link: https://patchwork.kernel.org/project/linux-mediatek/patch/20211202064039.20797-2-jason-jh.lin@mediatek.com/

this patch adds gfxoff support for smu 13.0.5
Signed-off-by: NYifan Zhang <yifan1.zhang@amd.com>
Reviewed-by: NHuang Rui <ray.huang@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

Based on smu 13.0.5 features, refine pp table code.
Signed-off-by: NYifan Zhang <yifan1.zhang@amd.com>
Reviewed-by: NHuang Rui <ray.huang@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

Silence [-Wmissing-prototypes] sparse warnings from the display folder
such as:

../drivers/gpu/drm/amd/amdgpu/../display/dc/clk_mgr/dcn315/dcn315_smu.c:126:5: warning: no previous prototype for ‘dcn315_smu_send_msg_with_param’ [-Wmissing-prototypes]
  126 | int dcn315_smu_send_msg_with_param(
      |     ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

v2: drop unused function dcn315_smu_set_voltage_via_phyclk() to avoid
    adding a warning (Alex)

Cc: Qingqing Zhuo (Lillian) <qingqing.zhuo@amd.com>
Reviewed-by: NHarry Wentland <harry.wentland@amd.com>
Reported-by: Nkernel test robot <lkp@intel.com>
Signed-off-by: NMagali Lemes <magalilemes00@gmail.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

Silence the following sparse warnings:

../drivers/gpu/drm/amd/amdgpu/../display/dc/dce112/dce112_resource.c:865:16:
sparse: warning: Using plain integer as NULL pointer

../drivers/gpu/drm/amd/amdgpu/../display/dc/dce110/dce110_hw_sequencer.c:1588:84:
sparse: warning: Using plain integer as NULL pointer

../drivers/gpu/drm/amd/amdgpu/../display/dc/dce110/dce110_hw_sequencer.c:2725:84:
sparse: warning: Using plain integer as NULL pointer

../drivers/gpu/drm/amd/amdgpu/../display/dc/core/dc_resource.c:1919:16:
sparse: warning: Using plain integer as NULL pointer

v2: drop removal of default case to avoid adding warnings (Alex)
Reviewed-by: NHarry Wentland <harry.wentland@amd.com>
Signed-off-by: NMagali Lemes <magalilemes00@gmail.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

When CONFIG_DRM_AMD_DC_DCN is not set, the function
'dm_helpers_enable_periodic_detection' doesn't have its prototype defined,
causing the following warning:

drivers/gpu/drm/amd/amdgpu/../display/amdgpu_dm/amdgpu_dm_helpers.c:805:6:
warning: no previous prototype for function 'dm_helpers_enable_periodic_detection' [-Wmissing-prototypes]
   void dm_helpers_enable_periodic_detection(struct dc_context *ctx, bool enable)
        ^
   drivers/gpu/drm/amd/amdgpu/../display/amdgpu_dm/amdgpu_dm_helpers.c:805:1:
note: declare 'static' if the function is not intended to be used outside
of this translation unit
   void dm_helpers_enable_periodic_detection(struct dc_context *ctx, bool enable)
   ^
   static
   1 warning generated.

This commit silences this warning by adding a conditional directive
around the mentioned function, and also corrects a small spelling error.
Reviewed-by: NHarry Wentland <harry.wentland@amd.com>
Reported-by: Nkernel test robot <lkp@intel.com>
Signed-off-by: NMagali Lemes <magalilemes00@gmail.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

Part of the documentation of the 'dc_process_dmub_aux_transfer_async'
function was misplaced, being put together with the
‘dc_enable_dmub_notifications’ documentation. This caused the following
warning:

drivers/gpu/drm/amd/amdgpu/../display/dc/core/dc.c:3757: warning:
expecting prototype for dc_process_dmub_aux_transfer_async(). Prototype
was for dc_enable_dmub_notifications() instead

This commit fixes the warning by placing the function documentations in
their right place.
Reviewed-by: NHarry Wentland <harry.wentland@amd.com>
Reported-by: Nkernel test robot <lkp@intel.com>
Signed-off-by: NMagali Lemes <magalilemes00@gmail.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

[Why]

During DC init, we read power management tables from PMFW. This info is
exchanged in the form of a binary blob inside gpu memory. In order to
parse the binary blob, the correct struct needs to be used.

[How]

Fix dcn316's definition of the DfPstateTable_t struct to align with PMFW
Signed-off-by: NLeo Li <sunpeng.li@amd.com>
Acked-by: NAlex Deucher <alexander.deucher@amd.com>
Acked-by: NHarry Wentland <harry.wentland@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

According to my investigation of the state of PCI
reset recently it's not working. The reason is
due to the fact the kernel PCI code rejects SBR
when there are more then one PF under same bridge
which we always have (at least AUDIO PF but usually
more) and that because SBR will reset all the PFS
and devices under the same bridge as you and you
cannot assume they support SBR.
Once we anble FLR support we can reenable this option as
FLR is doable on single PF and doens't have this
restriction.
Reviewed-by: NAlex Deucher <alexander.deucher@amd.com>
Signed-off-by: NAndrey Grodzovsky <andrey.grodzovsky@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

This parameter controls xGMI p2p communication, which is enabled by
default. However, it can be disabled by setting it to 0. In case xGMI
p2p is disabled in a dGPU, PCIe p2p interface will be used instead.
This parameter is ignored in GPUs that do not support xGMI
p2p configuration.
Signed-off-by: NAlex Sierra <alex.sierra@amd.com>
Acked-by: NLuben Tuikov <luben.tuikov@amd.com>
Acked-by: NHarish Kasiviswanathan <Harish.Kasiviswanathan@amd.com>
Reviewed-by: NFelix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

SMU MSG index should be used as parameter.
Signed-off-by: NYifan Zhang <yifan1.zhang@amd.com>
Acked-by: NAlex Deucher <alexander.deucher@amd.com>
Reviewed-by: NHuang Rui <ray.huang@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

To fix applications running across multiple GPU config hang.
Signed-off-by: NXiaogang Chen <xiaogang.chen@amd.com>
Reviewed-by: NAlex Deucher <alexander.deucher@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

This patch prohibits performing of set commands on all hwmon attributes
through sysfs in ONEVF mode.
Signed-off-by: NDanijel Slivka <danijel.slivka@amd.com>
Acked-by: NHarish Kasiviswanathan <Harish.Kasiviswanathan@amd.com>
Reviewed-by: NHarish Kasiviswanathan <Harish.Kasiviswanathan@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

Workstation application ANSA/META v21.1.4 get this error dmesg when
running CI test suite provided by ANSA/META:
[drm:amdgpu_gem_va_ioctl [amdgpu]] *ERROR* Couldn't update BO_VA (-16)

This is caused by:
1. create a 256MB buffer in invisible VRAM
2. CPU map the buffer and access it causes vm_fault and try to move
   it to visible VRAM
3. force visible VRAM space and traverse all VRAM bos to check if
   evicting this bo is valuable
4. when checking a VM bo (in invisible VRAM), amdgpu_vm_evictable()
   will set amdgpu_vm->evicting, but latter due to not in visible
   VRAM, won't really evict it so not add it to amdgpu_vm->evicted
5. before next CS to clear the amdgpu_vm->evicting, user VM ops
   ioctl will pass amdgpu_vm_ready() (check amdgpu_vm->evicted)
   but fail in amdgpu_vm_bo_update_mapping() (check
   amdgpu_vm->evicting) and get this error log

This error won't affect functionality as next CS will finish the
waiting VM ops. But we'd better clear the error log by checking
the amdgpu_vm->evicting flag in amdgpu_vm_ready() to stop calling
amdgpu_vm_bo_update_mapping() later.

Another reason is amdgpu_vm->evicted list holds all BOs (both
user buffer and page table), but only page table BOs' eviction
prevent VM ops. amdgpu_vm->evicting flag is set only for page
table BOs, so we should use evicting flag instead of evicted list
in amdgpu_vm_ready().

The side effect of this change is: previously blocked VM op (user
buffer in "evicted" list but no page table in it) gets done
immediately.

v2: update commit comments.
Acked-by: NPaul Menzel <pmenzel@molgen.mpg.de>
Reviewed-by: NChristian König <christian.koenig@amd.com>
Signed-off-by: NQiang Yu <qiang.yu@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>
Cc: stable@vger.kernel.org

vkms leverages common amdgpu framebuffer creation, and
also as it does not support FB modifier, there is no need
to check tiling flags when initing framebuffer when virtual
display is enabled.

This can fix below calltrace:

amdgpu 0000:00:08.0: GFX9+ requires FB check based on format modifier
WARNING: CPU: 0 PID: 1023 at drivers/gpu/drm/amd/amdgpu/amdgpu_display.c:1150 amdgpu_display_framebuffer_init+0x8e7/0xb40 [amdgpu]

v2: check adev->enable_virtual_display instead as vkms can be
	enabled in bare metal as well.
Signed-off-by: NLeslie Shi <Yuliang.Shi@amd.com>
Signed-off-by: NGuchun Chen <guchun.chen@amd.com>
Reviewed-by: NAlex Deucher <alexander.deucher@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

This reverts commit 4046afce.

No need to support modifier in virtual kms, otherwise, in SRIOV
mode, when lanuching X server, set crtc will fail due to mismatch
between primary plane modifier and framebuffer modifier.
Signed-off-by: NGuchun Chen <guchun.chen@amd.com>
Reviewed-by: NAlex Deucher <alexander.deucher@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

The GPU reset function of raven2 is not maintained or tested, so it should be
very unstable.

Now the amdgpu_asic_reset function is added to amdgpu_pmops_suspend, which
causes the S3 test of raven2 to fail, so the asic_reset of raven2 is ignored
here.

Fixes: daf8de08 ("drm/amdgpu: always reset the asic in suspend (v2)")
Signed-off-by: NChen Gong <curry.gong@amd.com>
Acked-by: NAlex Deucher <alexander.deucher@amd.com>
Reviewed-by: NMario Limonciello <mario.limonciello@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>
Cc: stable@vger.kernel.org

[Why]
Found when running igt@kms_atomic.

Userspace attempts to do a TEST_COMMIT when 0 streams which calls
dc_remove_stream_from_ctx. This in turn calls link_enc_unassign
which ends up modifying stream->link = NULL directly, causing the
global link_enc to be removed preventing further link activity
and future link validation from passing.

[How]
We take care of link_enc unassignment at the start of
link_enc_cfg_link_encs_assign so this call is no longer necessary.

Fixes global state from being modified while unlocked.
Reviewed-by: NJimmy Kizito <Jimmy.Kizito@amd.com>
Acked-by: NJasdeep Dhillon <jdhillon@amd.com>
Signed-off-by: NNicholas Kazlauskas <nicholas.kazlauskas@amd.com>
Tested-by: NDaniel Wheeler <daniel.wheeler@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>
Cc: stable@vger.kernel.org

commit 0064b0ce ("drm/amd/pm: enable ASPM by default") enabled ASPM
by default but a variety of hardware configurations it turns out that this
caused a regression.

* PPC64LE hardware does not support ASPM at a hardware level.
  CONFIG_PCIEASPM is often disabled on these architectures.
* Some dGPUs on ALD platforms don't work with ASPM enabled and PCIe subsystem
  disables it

Check with the PCIe subsystem to see that ASPM has been enabled
or not.

Fixes: 0064b0ce ("drm/amd/pm: enable ASPM by default")
Link: https://wiki.raptorcs.com/w/images/a/ad/P9_PHB_version1.0_27July2018_pub.pdf
Link: https://gitlab.freedesktop.org/drm/amd/-/issues/1723
Link: https://gitlab.freedesktop.org/drm/amd/-/issues/1739
Link: https://gitlab.freedesktop.org/drm/amd/-/issues/1885
Link: https://gitlab.freedesktop.org/drm/amd/-/issues/1907
Tested-by: koba.ko@canonical.com
Reviewed-by: NAlex Deucher <alexander.deucher@amd.com>
Signed-off-by: NMario Limonciello <mario.limonciello@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>
Cc: stable@vger.kernel.org

disocvery -> discovery
Reviewed-by: NGuchun Chen <guchun.chen@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

[why]
DML validation fails when we connect two or
more displays with HDR. Need to increase
DRAM BW to make the validation passing.
Following the value from DCN31.

[how]
Change the max DRAM BW DML field to 60%.
Reviewed-by: NCharlene Liu <Charlene.Liu@amd.com>
Acked-by: NQingqing Zhuo <qingqing.zhuo@amd.com>
Signed-off-by: NSung Joon Kim <sungkim@amd.com>
Reviewed-by: NHarry Wentland <harry.wentland@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

[Why]
Detbuffer size is dynamically set for dcn31x. At certain moment,
compbuf+(def size * num pipes) > config return buffer size causing
flickering. This is easily reproducible when MPO is
enabled with two displays.

[How]
At prepare BW, use the min comp buffer size. When it is to
optimize BW, set compbuf size back to maximum possible size.
Reviewed-by: NCharlene Liu <Charlene.Liu@amd.com>
Acked-by: NQingqing Zhuo <qingqing.zhuo@amd.com>
Signed-off-by: NDuncan Ma <duncanma@amd.com>
Reviewed-by: NHarry Wentland <harry.wentland@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

[Why & How]
Due to how pmfw fills out the table when dcfclk states are disabled,
using dcfclk based clk table would cause a no read situation.
Revert the change to prevent underflow until a better solution is coded.
Reviewed-by: NCharlene Liu <Charlene.Liu@amd.com>
Acked-by: NQingqing Zhuo <qingqing.zhuo@amd.com>
Signed-off-by: NDmytro Laktyushkin <Dmytro.Laktyushkin@amd.com>
Reviewed-by: NHarry Wentland <harry.wentland@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

[WHy & How]
Unbounded requesting is unsupported on pipe split modes
and this change prevents us running into such a situation
with wide modes.
Reviewed-by: NCharlene Liu <Charlene.Liu@amd.com>
Acked-by: NQingqing Zhuo <qingqing.zhuo@amd.com>
Signed-off-by: NDmytro Laktyushkin <Dmytro.Laktyushkin@amd.com>
Reviewed-by: NHarry Wentland <harry.wentland@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

Dump the list of register values to trace event on GPU reset.
Signed-off-by: NSomalapuram Amaranath <Amaranath.Somalapuram@amd.com>
Reviewed-by: NChristian König <christian.koenig@amd.com>
Reviewed-by: NAndrey Grodzovsky <andrey.grodzovsky@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

List of register populated for dump collection during the GPU reset.
Signed-off-by: NSomalapuram Amaranath <Amaranath.Somalapuram@amd.com>
Reviewed-by: NChristian König <christian.koenig@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

Workstation application ANSA/META v21.1.4 get this error dmesg when
running CI test suite provided by ANSA/META:
[drm:amdgpu_gem_va_ioctl [amdgpu]] *ERROR* Couldn't update BO_VA (-16)

This is caused by:
1. create a 256MB buffer in invisible VRAM
2. CPU map the buffer and access it causes vm_fault and try to move
   it to visible VRAM
3. force visible VRAM space and traverse all VRAM bos to check if
   evicting this bo is valuable
4. when checking a VM bo (in invisible VRAM), amdgpu_vm_evictable()
   will set amdgpu_vm->evicting, but latter due to not in visible
   VRAM, won't really evict it so not add it to amdgpu_vm->evicted
5. before next CS to clear the amdgpu_vm->evicting, user VM ops
   ioctl will pass amdgpu_vm_ready() (check amdgpu_vm->evicted)
   but fail in amdgpu_vm_bo_update_mapping() (check
   amdgpu_vm->evicting) and get this error log

This error won't affect functionality as next CS will finish the
waiting VM ops. But we'd better clear the error log by checking
the amdgpu_vm->evicting flag in amdgpu_vm_ready() to stop calling
amdgpu_vm_bo_update_mapping() later.

Another reason is amdgpu_vm->evicted list holds all BOs (both
user buffer and page table), but only page table BOs' eviction
prevent VM ops. amdgpu_vm->evicting flag is set only for page
table BOs, so we should use evicting flag instead of evicted list
in amdgpu_vm_ready().

The side effect of this change is: previously blocked VM op (user
buffer in "evicted" list but no page table in it) gets done
immediately.

v2: update commit comments.
Acked-by: NPaul Menzel <pmenzel@molgen.mpg.de>
Reviewed-by: NChristian König <christian.koenig@amd.com>
Signed-off-by: NQiang Yu <qiang.yu@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>