On Wed, Jun 1, 2022 at 10:03 AM Christian König christian.koenig@amd.com wrote:
Am 01.06.22 um 02:40 schrieb Bas Nieuwenhuizen:
This should be okay because moves themselves use KERNEL usage and hence still sync with BOOKKEEP usage. Then any later submits still wait on any pending VM operations.
(i.e. we only made VM ops not wait on BOOKKEEP submits, not the other way around)
Well NAK again. This allows access to freed up memory and is a complete no-go.
How does this allow access to freed memory? Worst I can see is that the unmap happens earlier if the app/drivers gets the waits wrong, which wouldn't give access after the underlying BO is freed?
Regards, Christian.
Signed-off-by: Bas Nieuwenhuizen bas@basnieuwenhuizen.nl
drivers/gpu/drm/amd/amdgpu/amdgpu_vm_cpu.c | 2 +- drivers/gpu/drm/amd/amdgpu/amdgpu_vm_sdma.c | 2 +- 2 files changed, 2 insertions(+), 2 deletions(-)
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm_cpu.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm_cpu.c index f10332e1c6c0..31bc73fd1fae 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm_cpu.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm_cpu.c @@ -51,7 +51,7 @@ static int amdgpu_vm_cpu_prepare(struct amdgpu_vm_update_params *p, if (!resv) return 0;
return amdgpu_bo_sync_wait_resv(p->adev, resv, sync_mode, sync_mode, p->vm, true);
return amdgpu_bo_sync_wait_resv(p->adev, resv, sync_mode, AMDGPU_SYNC_EXPLICIT, p->vm, true);}
/**
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm_sdma.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm_sdma.c index 63b484dc76c5..c8d5898bea11 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm_sdma.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm_sdma.c @@ -75,7 +75,7 @@ static int amdgpu_vm_sdma_prepare(struct amdgpu_vm_update_params *p, if (!resv) return 0;
return amdgpu_sync_resv(p->adev, &p->job->sync, resv, sync_mode, sync_mode, p->vm);
return amdgpu_sync_resv(p->adev, &p->job->sync, resv, sync_mode, AMDGPU_SYNC_EXPLICIT, p->vm);}
/**