From: Bas Nieuwenhuizen Date: Wed, 31 Jan 2018 12:58:55 +0000 (+0100) Subject: drm/amdgpu: Fix always_valid bos multiple LRU insertions. X-Git-Url: https://git.stricted.de/?a=commitdiff_plain;h=c73d9e350085a479d787d6d6dff3599fa6952491;p=GitHub%2FLineageOS%2Fandroid_kernel_motorola_exynos9610.git drm/amdgpu: Fix always_valid bos multiple LRU insertions. commit a20ee0b1f8b42e2568f3a4408003d22b2dfcc706 upstream. If these bos are evicted and are in the validated list things blow up, so do not put them in there. Notably, that tries to add the bo to the LRU twice, which results in a BUG_ON in ttm_bo.c. While for the bo_list an alternative would be to not allow always valid bos in there, that does not work for the user fence. v2: Fixed whitespace issue pointed out by checkpatch.pl Signed-off-by: Bas Nieuwenhuizen Reviewed-by: Christian König Signed-off-by: Alex Deucher Cc: stable@vger.kernel.org Signed-off-by: Greg Kroah-Hartman --- diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_bo_list.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_bo_list.c index 59089e027f4d..92be7f6de197 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_bo_list.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_bo_list.c @@ -233,8 +233,10 @@ void amdgpu_bo_list_get_list(struct amdgpu_bo_list *list, for (i = 0; i < list->num_entries; i++) { unsigned priority = list->array[i].priority; - list_add_tail(&list->array[i].tv.head, - &bucket[priority]); + if (!list->array[i].robj->parent) + list_add_tail(&list->array[i].tv.head, + &bucket[priority]); + list->array[i].user_pages = NULL; } diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c index b5aa8e6f8e0b..5f892ad6476e 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c @@ -522,7 +522,7 @@ static int amdgpu_cs_parser_bos(struct amdgpu_cs_parser *p, INIT_LIST_HEAD(&duplicates); amdgpu_vm_get_pd_bo(&fpriv->vm, &p->validated, &p->vm_pd); - if (p->uf_entry.robj) + if (p->uf_entry.robj && !p->uf_entry.robj->parent) list_add(&p->uf_entry.tv.head, &p->validated); if (need_mmap_lock)