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 486B8CAC599 for ; Tue, 16 Sep 2025 13:03:35 +0000 (UTC) Received: from gabe.freedesktop.org (localhost [127.0.0.1]) by gabe.freedesktop.org (Postfix) with ESMTP id 0C70810E7D3; Tue, 16 Sep 2025 13:03:35 +0000 (UTC) Authentication-Results: gabe.freedesktop.org; dkim=pass (2048-bit key; unprotected) header.d=intel.com header.i=@intel.com header.b="mKMxzv2Z"; dkim-atps=neutral Received: from mgamail.intel.com (mgamail.intel.com [192.198.163.11]) by gabe.freedesktop.org (Postfix) with ESMTPS id 1ACA210E7D3 for ; Tue, 16 Sep 2025 13:03:33 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=intel.com; i=@intel.com; q=dns/txt; s=Intel; t=1758027813; x=1789563813; h=message-id:date:mime-version:subject:to:cc:references: from:in-reply-to:content-transfer-encoding; bh=PCrpYGBp7y1JMaIPklpa0DytX0azC/2Q32vX4RzRb5k=; b=mKMxzv2ZC9ZG643JI9w8nj5kAWDm0627YLFpUBVtpi+jTCodzsHtzZLZ uiKqC+ngFOLsa9yJIylVwG3dqZQl1G/fjj7I+AtCi6MGeNqXjZfNXjPit dBstkx9sD2+o0j7yLHnVzprSRaTxfNnGb1AD7r2RBCcpq0L28PvdfjcLl QRc5aALY6/zHuZVyD1x3q47ryTnCZrv+V+HqBT5Ug7LIQ5Z9hFwEfU9JU UMpnBoBB+IXXybQNUG5sz5tCO39yH88LHaXDOWg8tK2tpopd9zhFkJlxN ex47Wx1Qjen1Ga2UZwCocfP/ASxX1u9J9IEkAC4QIkYzV5rnpK0teGNuE A==; X-CSE-ConnectionGUID: 8EP2sD1ISBGyiYYnxW8yMw== X-CSE-MsgGUID: mDs5ZpV0QniRThhxa7/Lkw== X-IronPort-AV: E=McAfee;i="6800,10657,11555"; a="70928412" X-IronPort-AV: E=Sophos;i="6.18,269,1751266800"; d="scan'208";a="70928412" Received: from orviesa010.jf.intel.com ([10.64.159.150]) by fmvoesa105.fm.intel.com with ESMTP/TLS/ECDHE-RSA-AES256-GCM-SHA384; 16 Sep 2025 06:03:33 -0700 X-CSE-ConnectionGUID: BnwxnEbjRreiI9HhBSnCcQ== X-CSE-MsgGUID: LS6yIRNKQVGqHcbbxc7WMA== X-ExtLoop1: 1 X-IronPort-AV: E=Sophos;i="6.18,269,1751266800"; d="scan'208";a="174225419" Received: from abityuts-desk.ger.corp.intel.com (HELO [10.245.245.199]) ([10.245.245.199]) by orviesa010-auth.jf.intel.com with ESMTP/TLS/ECDHE-RSA-AES256-GCM-SHA384; 16 Sep 2025 06:03:30 -0700 Message-ID: <53d50dff-89eb-4de0-befc-4bb2552c5e21@intel.com> Date: Tue, 16 Sep 2025 14:03:28 +0100 MIME-Version: 1.0 User-Agent: Mozilla Thunderbird Subject: Re: [RFC PATCH] drm/xe/dma-buf: Allow pinning of p2p dma-buf To: =?UTF-8?Q?Thomas_Hellstr=C3=B6m?= , intel-xe@lists.freedesktop.org Cc: Dave Airlie , Simona Vetter , Joonas Lahtinen , Maarten Lankhorst , Matthew Brost , Rodrigo Vivi , Lucas De Marchi References: <20250916115322.23293-1-thomas.hellstrom@linux.intel.com> Content-Language: en-GB From: Matthew Auld In-Reply-To: <20250916115322.23293-1-thomas.hellstrom@linux.intel.com> Content-Type: text/plain; charset=UTF-8; format=flowed Content-Transfer-Encoding: 8bit X-BeenThere: intel-xe@lists.freedesktop.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: Intel Xe graphics driver List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: intel-xe-bounces@lists.freedesktop.org Sender: "Intel-xe" On 16/09/2025 12:53, Thomas Hellström wrote: > RDMA NICs typically requires the VRAM dma-bufs to be pinned in > VRAM for pcie-p2p communication, since they don't fully support > the move_notify() scheme. We would like to support that. > > However allowing unaccounted pinning of VRAM creates a DOS vector > so up until now we haven't allowed it. > > However with cgroups support in TTM, the amount of VRAM allocated > to a cgroup can be limited, and since also the pinned memory is > accounted as allocated VRAM we should be safe. > > An analogy with system memory can be made if we observe the > similarity with kernel system memory that is allocated as the > result of user-space action and that is accounted using __GFP_ACCOUNT. > > Ideally, to be more flexible, we would add a "pinned_memory", > or possibly "kernel_memory" limit to the dmem cgroups controller, > that would additionally limit the memory that is pinned in this way. > If we let that limit default to the dmem::max limit we can > introduce that without needing to care about regressions. > > Considering that we already pin VRAM in this way for at least > page-table memory and LRC memory, and the above path to greater > flexibility, allow this also for dma-bufs. > > Cc: Dave Airlie > Cc: Simona Vetter > Cc: Joonas Lahtinen > Cc: Maarten Lankhorst > Cc: Matthew Brost > Cc: Rodrigo Vivi > Cc: Lucas De Marchi > Signed-off-by: Thomas Hellström > --- > drivers/gpu/drm/xe/tests/xe_dma_buf.c | 13 +++++++++ > drivers/gpu/drm/xe/xe_dma_buf.c | 41 +++++++++++++++++---------- > 2 files changed, 39 insertions(+), 15 deletions(-) > > diff --git a/drivers/gpu/drm/xe/tests/xe_dma_buf.c b/drivers/gpu/drm/xe/tests/xe_dma_buf.c > index a7e548a2bdfb..1f88ca71820c 100644 > --- a/drivers/gpu/drm/xe/tests/xe_dma_buf.c > +++ b/drivers/gpu/drm/xe/tests/xe_dma_buf.c > @@ -31,6 +31,7 @@ static void check_residency(struct kunit *test, struct xe_bo *exported, > struct drm_exec *exec) > { > struct dma_buf_test_params *params = to_dma_buf_test_params(test->priv); > + struct dma_buf_attachment *attach; > u32 mem_type; > int ret; > > @@ -88,6 +89,18 @@ static void check_residency(struct kunit *test, struct xe_bo *exported, > > KUNIT_EXPECT_TRUE(test, xe_bo_is_mem_type(exported, mem_type)); > > + /* Check that we can pin without migrating. */ > + attach = list_first_entry_or_null(&dmabuf->attachments, typeof(*attach), node); > + if (attach) { > + int err = dma_buf_pin(attach); > + > + if (!err) { > + KUNIT_EXPECT_TRUE(test, xe_bo_is_mem_type(exported, mem_type)); > + dma_buf_unpin(attach); > + } > + KUNIT_EXPECT_EQ(test, err, 0); > + } > + > if (params->force_different_devices) > KUNIT_EXPECT_TRUE(test, xe_bo_is_mem_type(imported, XE_PL_TT)); > else > diff --git a/drivers/gpu/drm/xe/xe_dma_buf.c b/drivers/gpu/drm/xe/xe_dma_buf.c > index a7d67725c3ee..54e42960daad 100644 > --- a/drivers/gpu/drm/xe/xe_dma_buf.c > +++ b/drivers/gpu/drm/xe/xe_dma_buf.c > @@ -48,32 +48,43 @@ static void xe_dma_buf_detach(struct dma_buf *dmabuf, > > static int xe_dma_buf_pin(struct dma_buf_attachment *attach) > { > - struct drm_gem_object *obj = attach->dmabuf->priv; > + struct dma_buf *dmabuf = attach->dmabuf; > + struct drm_gem_object *obj = dmabuf->priv; > struct xe_bo *bo = gem_to_xe_bo(obj); > struct xe_device *xe = xe_bo_device(bo); > struct drm_exec *exec = XE_VALIDATION_UNSUPPORTED; > + bool allow_vram = true; > int ret; > > - /* > - * For now only support pinning in TT memory, for two reasons: > - * 1) Avoid pinning in a placement not accessible to some importers. > - * 2) Pinning in VRAM requires PIN accounting which is a to-do. > - */ > - if (xe_bo_is_pinned(bo) && !xe_bo_is_mem_type(bo, XE_PL_TT)) { > + if (!IS_ENABLED(CONFIG_DMABUF_MOVE_NOTIFY)) { > + allow_vram = false; > + } else { > + list_for_each_entry(attach, &dmabuf->attachments, node) { > + if (!attach->peer2peer) { > + allow_vram = false; > + break; > + } > + } > + } > + > + if (xe_bo_is_pinned(bo) && !xe_bo_is_mem_type(bo, XE_PL_TT) && > + !(xe_bo_is_vram(bo) && allow_vram)) { > drm_dbg(&xe->drm, "Can't migrate pinned bo for dma-buf pin.\n"); > return -EINVAL; > } > > - ret = xe_bo_migrate(bo, XE_PL_TT, NULL, exec); > - if (ret) { > - if (ret != -EINTR && ret != -ERESTARTSYS) > - drm_dbg(&xe->drm, > - "Failed migrating dma-buf to TT memory: %pe\n", > - ERR_PTR(ret)); > - return ret; > + if (!allow_vram) { > + ret = xe_bo_migrate(bo, XE_PL_TT, NULL, exec); > + if (ret) { > + if (ret != -EINTR && ret != -ERESTARTSYS) > + drm_dbg(&xe->drm, > + "Failed migrating dma-buf to TT memory: %pe\n", > + ERR_PTR(ret)); > + return ret; > + } > } > > - ret = xe_bo_pin_external(bo, true, exec); > + ret = xe_bo_pin_external(bo, !allow_vram, exec); Are we also missing save/restore support for such objects? Or at least I can't see where the save flow is happening for externally pinned VRAM? > xe_assert(xe, !ret); > > return 0;