use bo_create/free_kernel instead of manually doing it
Signed-off-by: NMonk Liu <Monk.Liu@amd.com>
Reviewed-by: NChristian König <christian.koenig@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

instead of doing it in each GFX ip's sw_fini
Signed-off-by: NMonk Liu <Monk.Liu@amd.com>
Reviewed-by: NChristian König <christian.koenig@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

Signed-off-by: NMonk Liu <Monk.Liu@amd.com>
Reviewed-by: NChristian König <christian.koenig@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

NULL pointer is because original logic will step into
set_pde_pte() even after the gart.ptr is freed due to
there are twice gart_unbind() on all gart area.

also, there are other minor fixes:
1,since gart_init only create dummy page, the corresponding
gart_fini shouldn't do more like unbinding all GART, this is
unnecessary because in driver fini stage all GART unbinding
had already been done during each IP's SW_FINI (GMC's
SW_FINI is the last one called), so remove the step
for the GART unbinding in gart_fini().

2,gart_fini() is already invoked during each GMC IP's gart_fini
routine,e.g. gmc_vx_0_gart_fini(), so no need to manually
call it during ttm_fini().

3,amdgpu_gem_force_release() should be put ahead of
amdgpu_vm_manager_fini()
Signed-off-by: NMonk Liu <Monk.Liu@amd.com>
Reviewed-by: NChristian König <christian.koenig@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

for gart_ram_alloc/free, they are never used in driver thus
ripe them out totally.

for gart_vram_pin/unpin, they are not needed becuase we can
use bo_creat_kernel/free to replace the original manual way
in the gart_vram_alloc/free, thus gart_vram_pin/unpin can
also be riped out.
Signed-off-by: NMonk Liu <Monk.Liu@amd.com>
Reviewed-by: NChristian König <christian.koenig@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

Signed-off-by: NMonk Liu <Monk.Liu@amd.com>
Reviewed-by: NChristian König <christian.koenig@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

This can be triggered by userspace, e.g. trying to allocate too large a
BO, so it shouldn't log anything by default.

Callers need to handle failure anyway.
Reviewed-by: NAlex Deucher <alexander.deucher@amd.com>
Reviewed-by: NChristian König <christian.koenig@amd.com>
Signed-off-by: NMichel Dänzer <michel.daenzer@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

keep consistency with threshold of swapout
Reviewed-by: NChristian König <christian.koenig@amd.com>
Signed-off-by: NRoger He <Hongbo.He@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

Make initialization code check the ECC related registers, which are initialized
by the VBIOS, to see if ECC is present and initialized and DRM_INFO() the
result.
Signed-off-by: NDavid Panariti <David.Panariti@amd.com>
Reviewed-by: NAlex Deucher <alexander.deucher@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

Use the cached values rather than hardcoding it.
Acked-by: NChristian König <christian.koenig@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

Use the cached values rather than hardcoding it.
Acked-by: NChristian König <christian.koenig@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

Use the cached values rather than hardcoding it.
Acked-by: NChristian König <christian.koenig@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

Variable pf2vf_ver is assigned but never read, it is redundant and
hence can be removed.

Cleans up clang warning:
drivers/gpu/drm/amd/amdgpu/amdgpu_virt.c:310:3: warning: Value stored
to 'pf2vf_ver' is never read
Reivewed-by: NHorace Chen <horace.chen@amd.com>
Signed-off-by: NColin Ian King <colin.king@canonical.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

When UVD bo is created, its size is based on the information from firmware
header (ucode_size_bytes). The same value should be be used when programming
UVD mc controller offsets, otherwise it can happen that
(mmUVD_VCPU_CACHE_OFFSET2 + mmUVD_VCPU_CACHE_SIZE2) will point
AMDGPU_GPU_PAGE_SIZE bytes after the UVD bo end.

Second issue is that when programming the mmUVD_VCPU_CACHE_SIZE0 register,
AMDGPU_UVD_FIRMWARE_OFFSET should be taken into account. If it isn't,
(mmUVD_VCPU_CACHE_OFFSET2 + mmUVD_VCPU_CACHE_SIZE2) will always point
AMDGPU_UVD_FIRMWARE_OFFSET bytes after the UVD bo end.

v2: move firmware size calculation into macro definition
v3: align firmware size to the gpu page size
Reviewed-by: NChristian König <christian.koenig@amd.com>
Signed-off-by: NPiotr Redlewski <predlewski@gmail.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

Validates BO size against each requested domain's total memory.

v2:
Make GTT size check a MUST to allow fall back to GTT.
Rmove redundant NULL check.
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>

It's pointless to have the same value twice, just always use max_pfn.
Signed-off-by: NChristian König <christian.koenig@amd.com>
Reviewed-by: NAlex Deucher <alexander.deucher@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

Not sure what that should originally been good for, but it doesn't seem
to make any sense any more.
Signed-off-by: NChristian König <christian.koenig@amd.com>
Reviewed-by: NAlex Deucher <alexander.deucher@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

Otherwise, they could be optimized by scheduled fence.
Signed-off-by: NChunming Zhou <david1.zhou@amd.com>
Reviewed-by: NChristian König <christian.koenig@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

Do ring clear before ring test, otherwise compute ring test will
fail after gpu resetting. Still can't find the root cause, just
workaround it.
Signed-off-by: NXiangliang.Yu <Xiangliang.Yu@amd.com>
Acked-by: NMonk Liu <Monk.Liu@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

Retry at drm_dev_register instead of amdgpu_device_init.
Reviewed-by: NChristian König <christian.koenig@amd.com>
Signed-off-by: NPixel Ding <Pixel.Ding@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

Signed-off-by: NMonk Liu <Monk.Liu@amd.com>
Acked-by: NAlex Deucher <alexander.deucher@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

It introduces 900ms latency in exclusive mode which causes failure
of driver loading. Host can resize the BAR before guest staring,
so the resizing is not necessary here.
Signed-off-by: NPixel Ding <Pixel.Ding@amd.com>
Reviewed-by: NChristian König <christian.koenig@amd.com>
Reviewed-by: NAlex Deucher <alexander.deucher@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

Signed-off-by: Npding <Pixel.Ding@amd.com>
Reviewed-by: NFelix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

Also finalize kfd inside amdgpu_device_fini. kfd device_init needs
SRIOV exclusive accessing. Try to gather exclusive accessing to
reduce time consuming.
Signed-off-by: Npding <Pixel.Ding@amd.com>
Reviewed-by: NFelix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

Just allocate the GART space and fill it.

This prevents forcing the BO to be idle.

v2: don't unbind/bind at all, just fill the allocated GART space
Signed-off-by: NChristian König <christian.koenig@amd.com>
Acked-by: NAlex Deucher <alexander.deucher@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

We actually don't bind here, but rather allocate GART space if necessary.
Signed-off-by: NChristian König <christian.koenig@amd.com>
Reviewed-by: NAlex Deucher <alexander.deucher@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

Signed-off-by: NHawking Zhang <Hawking.Zhang@amd.com>
Acked-by: NChristian König <christian.koenig@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

Try to resize BAR0 to let CPU access all of VRAM.

v2: rebased, style cleanups, disable mem decode before resize,
    handle gmc_v9 as well, round size up to power of two.
v3: handle gmc_v6 as well, release and reassign all BARs in the driver.
v4: rename new function to amdgpu_device_resize_fb_bar,
    reenable mem decoding only if all resources are assigned.
v5: reorder resource release, return -ENODEV instead of BUG_ON().
v6: squash in rebase fix
Signed-off-by: NChristian König <christian.koenig@amd.com>
Reviewed-by: NAlex Deucher <alexander.deucher@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

The previous solution will create a zero buffer on the system
domain and then move the zeroes to the VRAM. This will break the
original data on the VRAM.

Refine the code to create bo on VRAM domain directly and then remove
and re-create mem node to the exact position before bo_pin. This can
avoid breaking the data and will not cause eviction.
Signed-off-by: NHorace Chen <horace.chen@amd.com>
Reviewed-by: Nmonk liu <monk.liu@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

This is caused of that hypervisor fails to handle request, one known
issue is MMIO unblocking timeout. In theory we can retry init here.
Signed-off-by: Npding <Pixel.Ding@amd.com>
Reviewed-by: NXiangliang Yu <Xiangliang.Yu@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

Reported-by: NSun Gary <Gary.Sun@amd.com>
Signed-off-by: Npding <Pixel.Ding@amd.com>
Reviewed-by: NXiangliang Yu <Xiangliang.Yu@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

Not used anymore.
Signed-off-by: NMichel Dänzer <michel.daenzer@amd.com>
Reviewed-by: NAlex Deucher <alexander.deucher@amd.com>
Acked-by: NChristian König <christian.koenig@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

Hardcoding the maximum numbers could result in spurious error messages
from the IRQ state callbacks, e.g. on Polaris 11/12:

[drm:dce_v11_0_set_pageflip_irq_state [amdgpu]] *ERROR* invalid pageflip crtc 5
[drm:amdgpu_irq_disable_all [amdgpu]] *ERROR* error disabling interrupt (-22)
Signed-off-by: NMichel Dänzer <michel.daenzer@amd.com>
Reviewed-by: NAlex Deucher <alexander.deucher@amd.com>
Acked-by: NChristian König <christian.koenig@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

The GTT manager handles the GART address space anyway, so it is
completely pointless to keep the same information around twice.

v2: rebased
Signed-off-by: NChristian König <christian.koenig@amd.com>
Reviewed-by: NChunming Zhou <david1.zhou@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

Rename amdgpu_gtt_mgr_is_allocated() to amdgpu_gtt_mgr_has_gart_addr() and use
that instead.

v2: rename the function as well.
Signed-off-by: NChristian König <christian.koenig@amd.com>
Reviewed-by: NChunming Zhou <david1.zhou@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

Signed-off-by: NMonk Liu <Monk.Liu@amd.com>
Acked-by: NChristian König <christian.koenig@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

for SR-IOV when doing gpu reset this routine shouldn't do
resource allocating otherwise memory leak
Signed-off-by: NMonk Liu <Monk.Liu@amd.com>
Acked-by: NChristian König <christian.koenig@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

1,no sriov check since gpu recover is unified
2,need CPU_ACCESS_REQUIRED flag for VRAM if SRIOV
because otherwise after following PIN the first allocated
VRAM bo is wasted due to some TTM mgr reason.
Signed-off-by: NMonk Liu <Monk.Liu@amd.com>
Reviewed-by: NChristian König <christian.koenig@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

since now gpu reset is unified with gpu_recover
for both bare-metal and SR-IOV:

1)rename in_sriov_reset to in_gpu_reset
2)move lock_reset from adev->virt to adev
Signed-off-by: NMonk Liu <Monk.Liu@amd.com>
Reviewed-by: NChristian König <christian.koenig@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>

1,new imple names amdgpu_gpu_recover which gives more hint
on what it does compared with gpu_reset

2,gpu_recover unify bare-metal and SR-IOV, only the asic reset
part is implemented differently

3,gpu_recover will increase hang job karma and mark its entity/context
as guilty if exceeds limit

V2:

4,in scheduler main routine the job from guilty context  will be immedialy
fake signaled after it poped from queue and its fence be set with
"-ECANCELED" error

5,in scheduler recovery routine all jobs from the guilty entity would be
dropped

6,in run_job() routine the real IB submission would be skipped if @skip parameter
equales true or there was VRAM lost occured.

V3:

7,replace deprecated gpu reset, use new gpu recover
Signed-off-by: NMonk Liu <Monk.Liu@amd.com>
Reviewed-by: NChristian König <christian.koenig@amd.com>
Signed-off-by: NAlex Deucher <alexander.deucher@amd.com>