Please also include a patch description so we understand why this change is needed.
Alex
From: amd-gfx <amd-gfx-bounces@xxxxxxxxxxxxxxxxxxxxx> on behalf of Christian König <ckoenig.leichtzumerken@xxxxxxxxx>
Sent: Wednesday, August 14, 2019 11:06 AM To: Ba, Gang <Gang.Ba@xxxxxxx>; amd-gfx@xxxxxxxxxxxxxxxxxxxxx <amd-gfx@xxxxxxxxxxxxxxxxxxxxx> Cc: Ba, Gang <Gang.Ba@xxxxxxx> Subject: Re: [PATCH] Update VM function point when VM state changed. Am 14.08.19 um 17:01 schrieb Gang Ba:
> Change-Id: Ibcfada560a00c9aeebfd922ae48de920e44a5866 > Signed-off-by: Gang Ba <gaba@xxxxxxx> Good catch, but there is a problem below. > --- > drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c | 6 ++++++ > 1 file changed, 6 insertions(+) > > diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c > index c1baf3d..72dd434 100644 > --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c > +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c > @@ -2862,6 +2862,12 @@ int amdgpu_vm_make_compute(struct amdgpu_device *adev, struct amdgpu_vm *vm, uns > WARN_ONCE((vm->use_cpu_for_update && !amdgpu_gmc_vram_full_visible(&adev->gmc)), > "CPU update of VM recommended only for large BAR system\n"); > > + if (vm->use_cpu_for_update) > + vm->update_funcs = &amdgpu_vm_cpu_funcs; > + else > + vm->update_funcs = &amdgpu_vm_sdma_funcs; > + vm->last_update = NULL; You can't just set last_update to NULL or your would leak the memory the fence occupies. So this needs to be: dma_fence_put(vm->last_update); vm->last_update = NULL; Regards, Christian. > + > if (vm->pasid) { > unsigned long flags; > _______________________________________________ amd-gfx mailing list amd-gfx@xxxxxxxxxxxxxxxxxxxxx https://lists.freedesktop.org/mailman/listinfo/amd-gfx |
_______________________________________________ amd-gfx mailing list amd-gfx@xxxxxxxxxxxxxxxxxxxxx https://lists.freedesktop.org/mailman/listinfo/amd-gfx