From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on aws-us-west-2-korg-lkml-1.web.codeaurora.org Received: from gabe.freedesktop.org (gabe.freedesktop.org [131.252.210.177]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by smtp.lore.kernel.org (Postfix) with ESMTPS id B1F9BCCD185 for ; Thu, 16 Oct 2025 02:40:15 +0000 (UTC) Received: from gabe.freedesktop.org (localhost [127.0.0.1]) by gabe.freedesktop.org (Postfix) with ESMTP id 2574910E92A; Thu, 16 Oct 2025 02:40:15 +0000 (UTC) Received: from us-smtp-delivery-44.mimecast.com (us-smtp-delivery-44.mimecast.com [205.139.111.44]) by gabe.freedesktop.org (Postfix) with ESMTPS id 394F310E92A for ; Thu, 16 Oct 2025 02:40:14 +0000 (UTC) Received: from mx-prod-mc-08.mail-002.prod.us-west-2.aws.redhat.com (ec2-35-165-154-97.us-west-2.compute.amazonaws.com [35.165.154.97]) by relay.mimecast.com with ESMTP with STARTTLS (version=TLSv1.3, cipher=TLS_AES_256_GCM_SHA384) id us-mta-688-HCEj1lEbO1q2cjFbFSrZ1w-1; Wed, 15 Oct 2025 22:34:01 -0400 X-MC-Unique: HCEj1lEbO1q2cjFbFSrZ1w-1 X-Mimecast-MFC-AGG-ID: HCEj1lEbO1q2cjFbFSrZ1w_1760582039 Received: from mx-prod-int-06.mail-002.prod.us-west-2.aws.redhat.com (mx-prod-int-06.mail-002.prod.us-west-2.aws.redhat.com [10.30.177.93]) (using TLSv1.3 with cipher TLS_AES_256_GCM_SHA384 (256/256 bits) key-exchange X25519 server-signature RSA-PSS (2048 bits) server-digest SHA256) (No client certificate requested) by mx-prod-mc-08.mail-002.prod.us-west-2.aws.redhat.com (Postfix) with ESMTPS id CB9A4180034C; Thu, 16 Oct 2025 02:33:59 +0000 (UTC) Received: from dreadlord.taild9177d.ts.net (unknown [10.67.32.64]) by mx-prod-int-06.mail-002.prod.us-west-2.aws.redhat.com (Postfix) with ESMTP id 8C036180057A; Thu, 16 Oct 2025 02:33:53 +0000 (UTC) From: Dave Airlie To: dri-devel@lists.freedesktop.org, tj@kernel.org, christian.koenig@amd.com, Johannes Weiner , Michal Hocko , Roman Gushchin , Shakeel Butt , Muchun Song Cc: cgroups@vger.kernel.org, Dave Chinner , Waiman Long , simona@ffwll.ch Subject: [PATCH 14/16] amdgpu: add support for memory cgroups Date: Thu, 16 Oct 2025 12:31:42 +1000 Message-ID: <20251016023205.2303108-15-airlied@gmail.com> In-Reply-To: <20251016023205.2303108-1-airlied@gmail.com> References: <20251016023205.2303108-1-airlied@gmail.com> MIME-Version: 1.0 X-Scanned-By: MIMEDefang 3.4.1 on 10.30.177.93 X-Mimecast-Spam-Score: 0 X-Mimecast-MFC-PROC-ID: XBbUIyMiaPyQ63ThkzaAtE1MgDeEqNSm-y_MyKMC2us_1760582039 X-Mimecast-Originator: gmail.com Content-Transfer-Encoding: quoted-printable content-type: text/plain; charset=WINDOWS-1252; x-default=true X-BeenThere: dri-devel@lists.freedesktop.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: Direct Rendering Infrastructure - Development List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: dri-devel-bounces@lists.freedesktop.org Sender: "dri-devel" From: Dave Airlie This adds support for adding a obj cgroup to a buffer object, and passing in the placement flags to make sure it's accounted properly. Signed-off-by: Dave Airlie --- drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c | 1 + drivers/gpu/drm/amd/amdgpu/amdgpu_object.c | 13 +++++++++---- drivers/gpu/drm/amd/amdgpu/amdgpu_object.h | 1 + drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c | 2 ++ mm/memcontrol.c | 1 + 5 files changed, 14 insertions(+), 4 deletions(-) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c b/drivers/gpu/drm/amd/= amdgpu/amdgpu_gem.c index b7ebae289bea..85ff70a399bc 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c @@ -225,6 +225,7 @@ int amdgpu_gem_object_create(struct amdgpu_device *adev= , unsigned long size, =09bp.domain =3D initial_domain; =09bp.bo_ptr_size =3D sizeof(struct amdgpu_bo); =09bp.xcp_id_plus1 =3D xcp_id_plus1; +=09bp.objcg =3D get_obj_cgroup_from_current(); =20 =09r =3D amdgpu_bo_create_user(adev, &bp, &ubo); =09if (r) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c b/drivers/gpu/drm/a= md/amdgpu/amdgpu_object.c index e08f58de4b17..d6a6f7e17a2a 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c @@ -167,7 +167,7 @@ void amdgpu_bo_placement_from_domain(struct amdgpu_bo *= abo, u32 domain) =09=09places[c].mem_type =3D =09=09=09abo->flags & AMDGPU_GEM_CREATE_PREEMPTIBLE ? =09=09=09AMDGPU_PL_PREEMPT : TTM_PL_TT; -=09=09places[c].flags =3D 0; +=09=09places[c].flags =3D TTM_PL_FLAG_MEMCG; =09=09/* =09=09 * When GTT is just an alternative to VRAM make sure that we =09=09 * only use it as fallback and still try to fill up VRAM first. @@ -182,7 +182,7 @@ void amdgpu_bo_placement_from_domain(struct amdgpu_bo *= abo, u32 domain) =09=09places[c].fpfn =3D 0; =09=09places[c].lpfn =3D 0; =09=09places[c].mem_type =3D TTM_PL_SYSTEM; -=09=09places[c].flags =3D 0; +=09=09places[c].flags =3D TTM_PL_FLAG_MEMCG; =09=09c++; =09} =20 @@ -662,16 +662,21 @@ int amdgpu_bo_create(struct amdgpu_device *adev, =09=09size =3D ALIGN(size, PAGE_SIZE); =09} =20 -=09if (!amdgpu_bo_validate_size(adev, size, bp->domain)) +=09if (!amdgpu_bo_validate_size(adev, size, bp->domain)) { +=09=09obj_cgroup_put(bp->objcg); =09=09return -ENOMEM; +=09} =20 =09BUG_ON(bp->bo_ptr_size < sizeof(struct amdgpu_bo)); =20 =09*bo_ptr =3D NULL; =09bo =3D kvzalloc(bp->bo_ptr_size, GFP_KERNEL); -=09if (bo =3D=3D NULL) +=09if (bo =3D=3D NULL) { +=09=09obj_cgroup_put(bp->objcg); =09=09return -ENOMEM; +=09} =09drm_gem_private_object_init(adev_to_drm(adev), &bo->tbo.base, size); +=09ttm_bo_set_cgroup(&bo->tbo, bp->objcg); /* hand the reference to the tt= m bo */ =09bo->tbo.base.funcs =3D &amdgpu_gem_object_funcs; =09bo->vm_bo =3D NULL; =09bo->preferred_domains =3D bp->preferred_domain ? bp->preferred_domain : diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.h b/drivers/gpu/drm/a= md/amdgpu/amdgpu_object.h index 656b8a931dae..b07a168a6665 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.h @@ -55,6 +55,7 @@ struct amdgpu_bo_param { =09enum ttm_bo_type=09=09type; =09bool=09=09=09=09no_wait_gpu; =09struct dma_resv=09=09=09*resv; +=09struct obj_cgroup *objcg; =09void=09=09=09=09(*destroy)(struct ttm_buffer_object *bo); =09/* xcp partition number plus 1, 0 means any partition */ =09int8_t=09=09=09=09xcp_id_plus1; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c b/drivers/gpu/drm/amd/= amdgpu/amdgpu_ttm.c index bcab4a83137b..7642c17ebda7 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c @@ -152,11 +152,13 @@ static void amdgpu_evict_flags(struct ttm_buffer_obje= ct *bo, =09=09=09amdgpu_bo_placement_from_domain(abo, AMDGPU_GEM_DOMAIN_GTT | =09=09=09=09=09=09=09AMDGPU_GEM_DOMAIN_CPU); =09=09} +=09=09abo->placements[0].flags &=3D ~TTM_PL_FLAG_MEMCG; =09=09break; =09case TTM_PL_TT: =09case AMDGPU_PL_PREEMPT: =09default: =09=09amdgpu_bo_placement_from_domain(abo, AMDGPU_GEM_DOMAIN_CPU); +=09=09abo->placements[0].flags &=3D ~TTM_PL_FLAG_MEMCG; =09=09break; =09} =09*placement =3D abo->placement; diff --git a/mm/memcontrol.c b/mm/memcontrol.c index ece340f3e391..b5782f6d21c2 100644 --- a/mm/memcontrol.c +++ b/mm/memcontrol.c @@ -2721,6 +2721,7 @@ __always_inline struct obj_cgroup *current_obj_cgroup= (void) =20 =09return objcg; } +EXPORT_SYMBOL_GPL(current_obj_cgroup); =20 struct obj_cgroup *get_obj_cgroup_from_folio(struct folio *folio) { --=20 2.51.0