From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from mgamail.intel.com (mgamail.intel.com [192.198.163.7]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by smtp.subspace.kernel.org (Postfix) with ESMTPS id A819E43784E for ; Mon, 6 Jul 2026 18:03:57 +0000 (UTC) Authentication-Results: smtp.subspace.kernel.org; arc=none smtp.client-ip=192.198.163.7 ARC-Seal:i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1783361041; cv=none; b=o9spZg5uhhyA/ihQ5F8qLf05K2DNwglM5EYOqZWP2OQMVzRBHYEVXjbE+fd4WkWuDudxsuDxJCvovTtgfkWpcLFERK+6bMy5KjdMFQOsIUoXhGrbvRacnYqYa4o3OvlecJ5ewAWJRZzu4sVd1pXCFTk08tzpPchCKi7mxbpXtLI= ARC-Message-Signature:i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1783361041; c=relaxed/simple; bh=2kcQBkGpNPCI1AvVOTzKbPcT9nWlCRpBwE3xZaqBWA0=; h=Message-ID:Subject:From:To:Cc:Date:In-Reply-To:References: Content-Type:MIME-Version; b=IRSYH2O02/Su8COwKr5VpSftSMlZweNPhdE7yuKR7iIyjjFwSrpvJxs40XbZFOdpA74upvwbnr4mW8TgpIITKO2rL/I90PPwC5e22FMiXY9Qwa6ue2n2l567Kj+IpBQJU7xh7jsNUyjezNymu9R1fqJIhNXZ/pf7AUoYEd2ywFs= ARC-Authentication-Results:i=1; smtp.subspace.kernel.org; dmarc=pass (p=none dis=none) header.from=linux.intel.com; spf=pass smtp.mailfrom=linux.intel.com; dkim=pass (2048-bit key) header.d=intel.com header.i=@intel.com header.b=UnYAJCg1; arc=none smtp.client-ip=192.198.163.7 Authentication-Results: smtp.subspace.kernel.org; dmarc=pass (p=none dis=none) header.from=linux.intel.com Authentication-Results: smtp.subspace.kernel.org; spf=pass smtp.mailfrom=linux.intel.com Authentication-Results: smtp.subspace.kernel.org; dkim=pass (2048-bit key) header.d=intel.com header.i=@intel.com header.b="UnYAJCg1" DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=intel.com; i=@intel.com; q=dns/txt; s=Intel; t=1783361038; x=1814897038; h=message-id:subject:from:to:cc:date:in-reply-to: references:content-transfer-encoding:mime-version; bh=2kcQBkGpNPCI1AvVOTzKbPcT9nWlCRpBwE3xZaqBWA0=; b=UnYAJCg1NtxBK2AtJDtjUtBnk65QmdNcBqzujLvRjXLqynn24jbhR9+q wNyf9rimiObZKSXWV4WzBLAK30zAHeV4d/kYJbO9shaFRIm5H876qfWON BiWmnuGsWqW6QpLjJrCkPGUZIWmUJPnnp2k36+Vl7F270DL0fttqh5hKT HvyDhsq0OaBAATqaJv+qnQaFz1eHuLS9mh2c7vjRjShZc/Toz6/LtiAda Zr+kRWm1/t5Fb9jjKXWDIiod4qKphVoqrNXDY5uiAGi8rO6/5JCFJvBTu 0QSz5Y7+X5UDo/Sa52aPRaj0zQkjKrEl4UoNmjSZNQJBQG+HrvGeN38VT A==; X-CSE-ConnectionGUID: OHrmQD3TQmeO37N9ZkZS7w== X-CSE-MsgGUID: NDcL9NXJRMuS6lyA+2ueiQ== X-IronPort-AV: E=McAfee;i="6800,10657,11839"; a="109547925" X-IronPort-AV: E=Sophos;i="6.25,151,1779174000"; d="scan'208";a="109547925" Received: from orviesa001.jf.intel.com ([10.64.159.141]) by fmvoesa101.fm.intel.com with ESMTP/TLS/ECDHE-RSA-AES256-GCM-SHA384; 06 Jul 2026 11:03:56 -0700 X-CSE-ConnectionGUID: 59Kn16hXStuL4+5PuDS4bQ== X-CSE-MsgGUID: XAixNBXJS1ivs9FoPT5rTg== X-ExtLoop1: 1 X-IronPort-AV: E=Sophos;i="6.25,151,1779174000"; d="scan'208";a="291944761" Received: from conormcd-mobl2.ger.corp.intel.com (HELO [10.245.244.132]) ([10.245.244.132]) by smtpauth.intel.com with ESMTP/TLS/ECDHE-RSA-AES256-GCM-SHA384; 06 Jul 2026 11:03:52 -0700 Message-ID: <257957966546bbccdf18f883cab690646c18ab80.camel@linux.intel.com> Subject: Re: [PATCH 02/10] drm/ttm: replace TTMs refcount with the DRM refcount v4 From: Thomas =?ISO-8859-1?Q?Hellstr=F6m?= To: Matthew Brost Cc: Christian =?ISO-8859-1?Q?K=F6nig?= , Natalie Vock , Maarten Lankhorst , Maxime Ripard , Thomas Zimmermann , David Airlie , Simona Vetter , Jani Nikula , Joonas Lahtinen , Rodrigo Vivi , Tvrtko Ursulin , Huang Rui , Matthew Auld , Alex Deucher , dri-devel@lists.freedesktop.org, linux-kernel@vger.kernel.org, intel-gfx@lists.freedesktop.org, intel-xe@lists.freedesktop.org, amd-gfx@lists.freedesktop.org Date: Mon, 06 Jul 2026 20:03:49 +0200 In-Reply-To: References: <20260703-ttm_2_drm_exec-v1-0-43685ac1286b@gmx.de> <20260703-ttm_2_drm_exec-v1-2-43685ac1286b@gmx.de> <88614266-6ee3-4488-9e2f-030ae0266e0e@amd.com> <7031c630c987940d814eaef9d1696969be12d81b.camel@linux.intel.com> Organization: Intel Sweden AB, Registration Number: 556189-6027 Content-Type: text/plain; charset="UTF-8" Content-Transfer-Encoding: quoted-printable User-Agent: Evolution 3.58.3 (3.58.3-1.fc43) Precedence: bulk X-Mailing-List: linux-kernel@vger.kernel.org List-Id: List-Subscribe: List-Unsubscribe: MIME-Version: 1.0 On Mon, 2026-07-06 at 10:51 -0700, Matthew Brost wrote: > On Mon, Jul 06, 2026 at 07:01:27PM +0200, Thomas Hellstr=C3=B6m wrote: > > On Mon, 2026-07-06 at 16:49 +0200, Christian K=C3=B6nig wrote: > > > On 7/6/26 15:14, Thomas Hellstr=C3=B6m wrote: > > > > On Fri, 2026-07-03 at 18:31 +0200, Natalie Vock wrote: > > > > > From: Christian K=C3=B6nig > > > > >=20 > > > > > Instead of keeping a separate reference count for the TTM > > > > > object > > > > > also > > > > > use > > > > > the reference count for DRM GEM objects inside TTM. > > > > >=20 > > > > > Apart from avoiding two reference counts for one object this > > > > > approach > > > > > has > > > > > the clear advantage of being able to use drm_exec inside TTM. > > > > >=20 > > > > > v2: adjust XE assert as well and re-enable disabled test > > > > > v3: handle another case in i915 > > > > > v4: set GEM driver funcs of transfer BOs to point to the TTM > > > > > free > > > > > callback (Natalie) > > > >=20 > > > > I think the main review issue from the last time this was on > > > > the > > > > table > > > > was that we shouldn't resurrect the gem refcount. Apart from > > > > the > > > > risc > > > > of getting barriers wrong, both xe and IIRC i915 partly rely on > > > > the > > > > gem > > > > refcount never being resurrected and that callbacks for bos > > > > with > > > > zero > > > > gem refcount means that the gem part of the object is unusable. > > >=20 > > > I've spend quite some time thinking about that and came to the > > > conclusion that this is actually harmless. > > >=20 > > > The drivers shouldn't be able to see the resurected BO, except if > > > they go over the LRU list manually (which they shouldn't). > >=20 > > The shrinker uses the TTM helpers for this. Basically the check > > needs > > to be ported to use the zombie interface but the present change > > also > > widens the window where we can't evict / shrink at all due to zero > > refcounts. > >=20 > > While it might be made harmless, resurrecting a refcount like this > > is > > IMO not something that should leak into the gem refcount. Nobody > > else > > does this in the kernel tree. The bo in reality becomes a zombie > > once > > the gem refcount reaches zero. >=20 > When you say resurrecting a refcount - what exactly do you mean by > this? >=20 > ref -> 0 -> init ref count 1 -> 0 again? Resending with full CC list. Yes. The destructor re-initializes the refcount to 1. Previously this was confined to the TTM refcount with limited visibility. /Thomas >=20 > Matt >=20 > >=20 > > >=20 > > > >=20 > > > > For example xe_bo.c: > > > >=20 > > > > if (!xe_bo_is_xe_bo(bo) || > > > > !xe_bo_get_unless_zero(xe_bo)) > > > > return xe_bo_shrink_purge(ctx, bo, scanned); > > > >=20 > > > > So IIRC the conclusion was when removing the ttm refcount we > > > > shouldn't > > > > attempt to resurrect the gem one. If the get_unless_zero() > > > > fails > > > > during > > > > evict walk, we simply find something to wait for. See previous > > > > discussion there. > > >=20 > > > Yeah, I considered that as well but the problem is we often > > > doesn't > > > have anything to wait on. > >=20 > > That's not the conclusion of the previous discussion? > >=20 > > https://lore.kernel.org/dri-devel/20250716160555.20217-2-christian.koen= ig@amd.com/#r > >=20 > > >=20 > > > >=20 > > > > I fully support removing the ttm refcount, but not if it means > > > > resurrecting the gem refcount. > > > >=20 > > > > If we want to sidestep that problem, in favour of getting the > > > > proposed > > > > locking functionality in and future proof it, I suggest using > > > >=20 > > > > https://lore.kernel.org/all/20260605112700.181040-1-thomas.hellstro= m@linux.intel.com/ > > > >=20 > > > > And rebase this series on that. This means we can use the ttm > > > > refcount > > > > for the transaction refcounting, and also that if we add a dma- > > > > buf > > > > map > > > > interface with a dma_resv_txn_obj, we could use that to also > > > > have > > > > exhaustive eviction that originates from a dma_buf map. > > >=20 > > > I don't think that this is a good idea. It just adds another > > > layer of > > > abstraction and doesn't solve the problem in any way possible. > >=20 > > This comment confuses me. Exactly what problem isn't solved by > > this, > > and which of the stated benefits/use-cases in the cover-letter do > > you > > think aren't worthwhile? > >=20 > > Also for reference: (Section at the end and follow-up messages) > > https://lore.kernel.org/all/3716d43462188590743060755b37e3d060f7600f.ca= mel@linux.intel.com/ > >=20 > > Thanks, > > Thomas > >=20 > >=20 > > > Regards, > > > Christian. > > >=20 > > > >=20 > > > > /Thomas > > > >=20 > > > > >=20 > > > > > Signed-off-by: tChristian K=C3=B6nig > > > > > Signed-off-by: Natalie Vock > > > > > --- > > > > > =C2=A0drivers/gpu/drm/i915/gem/i915_gem_ttm.c=C2=A0=C2=A0=C2=A0= =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 |=C2=A0 35 +++-- > > > > > - > > > > > =C2=A0drivers/gpu/drm/ttm/tests/ttm_bo_validate_test.c |=C2=A0=C2= =A0 8 +- > > > > > =C2=A0drivers/gpu/drm/ttm/tests/ttm_kunit_helpers.c=C2=A0=C2=A0= =C2=A0 |=C2=A0=C2=A0 2 - > > > > > =C2=A0drivers/gpu/drm/ttm/ttm_bo.c=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0= =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2= =A0=C2=A0=C2=A0 | 135 > > > > > +++++++++++-- > > > > > ---------- > > > > > =C2=A0drivers/gpu/drm/ttm/ttm_bo_internal.h=C2=A0=C2=A0=C2=A0=C2= =A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 |=C2=A0 17 ++- > > > > > =C2=A0drivers/gpu/drm/ttm/ttm_bo_util.c=C2=A0=C2=A0=C2=A0=C2=A0= =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 |=C2=A0 = 15 ++- > > > > > =C2=A0drivers/gpu/drm/xe/xe_bo.c=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2= =A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0= =C2=A0=C2=A0=C2=A0=C2=A0 |=C2=A0=C2=A0 2 +- > > > > > =C2=A0include/drm/ttm/ttm_bo.h=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2= =A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0= =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 |=C2=A0=C2=A0 9 -- > > > > > =C2=A08 files changed, 111 insertions(+), 112 deletions(-) > > > > >=20 > > > > > diff --git a/drivers/gpu/drm/i915/gem/i915_gem_ttm.c > > > > > b/drivers/gpu/drm/i915/gem/i915_gem_ttm.c > > > > > index df3fcc2b1248e..642296602de69 100644 > > > > > --- a/drivers/gpu/drm/i915/gem/i915_gem_ttm.c > > > > > +++ b/drivers/gpu/drm/i915/gem/i915_gem_ttm.c > > > > > @@ -939,7 +939,7 @@ void i915_ttm_adjust_lru(struct > > > > > drm_i915_gem_object *obj) > > > > > =C2=A0 * Don't manipulate the TTM LRUs while in TTM bo > > > > > destruction. > > > > > =C2=A0 * We're called through > > > > > i915_ttm_delete_mem_notify(). > > > > > =C2=A0 */ > > > > > - if (!kref_read(&bo->kref)) > > > > > + if (!kref_read(&bo->base.refcount)) > > > > > =C2=A0 return; > > > > > =C2=A0 > > > > > =C2=A0 /* > > > > > @@ -957,30 +957,21 @@ void i915_ttm_adjust_lru(struct > > > > > drm_i915_gem_object *obj) > > > > > =C2=A0 * > > > > > =C2=A0 * TODO: consider maybe also bumping the shrinker > > > > > list > > > > > here > > > > > when we have > > > > > =C2=A0 * already unpinned it, which should give us > > > > > something > > > > > more > > > > > like an LRU. > > > > > - * > > > > > - * TODO: There is a small window of opportunity for > > > > > this > > > > > function to > > > > > - * get called from eviction after we've dropped the > > > > > last > > > > > GEM > > > > > refcount, > > > > > - * but before the TTM deleted flag is set on the > > > > > object. > > > > > Avoid > > > > > - * adjusting the shrinker list in such cases, since > > > > > the > > > > > object is > > > > > - * not available to the shrinker anyway due to its > > > > > zero > > > > > refcount. > > > > > - * To fix this properly we should move to a TTM > > > > > shrinker > > > > > LRU > > > > > list for > > > > > - * these objects. > > > > > =C2=A0 */ > > > > > - if (kref_get_unless_zero(&obj->base.refcount)) { > > > > > - if (shrinkable !=3D obj->mm.ttm_shrinkable) { > > > > > - if (shrinkable) { > > > > > - if (obj->mm.madv =3D=3D > > > > > I915_MADV_WILLNEED) > > > > > - > > > > > __i915_gem_object_ma > > > > > ke_s > > > > > hrinkable(obj); > > > > > - else > > > > > - > > > > > __i915_gem_object_ma > > > > > ke_p > > > > > urgeable(obj); > > > > > - } else { > > > > > - > > > > > i915_gem_object_make_unshrin > > > > > kabl > > > > > e(obj); > > > > > - } > > > > > - > > > > > - obj->mm.ttm_shrinkable =3D shrinkable; > > > > > + i915_gem_object_get(obj); > > > > > + if (shrinkable !=3D obj->mm.ttm_shrinkable) { > > > > > + if (shrinkable) { > > > > > + if (obj->mm.madv =3D=3D > > > > > I915_MADV_WILLNEED) > > > > > + __i915_gem_object_make_shrin > > > > > kabl > > > > > e(ob > > > > > j); > > > > > + else > > > > > + __i915_gem_object_make_purge > > > > > able > > > > > (obj > > > > > ); > > > > > + } else { > > > > > + i915_gem_object_make_unshrinkable(ob > > > > > j); > > > > > =C2=A0 } > > > > > - i915_gem_object_put(obj); > > > > > + > > > > > + obj->mm.ttm_shrinkable =3D shrinkable; > > > > > =C2=A0 } > > > > > + i915_gem_object_put(obj); > > > > > =C2=A0 > > > > > =C2=A0 /* > > > > > =C2=A0 * Put on the correct LRU list depending on the MADV > > > > > status > > > > > diff --git a/drivers/gpu/drm/ttm/tests/ttm_bo_validate_test.c > > > > > b/drivers/gpu/drm/ttm/tests/ttm_bo_validate_test.c > > > > > index 56ad8ef325840..904cb4da6c9b3 100644 > > > > > --- a/drivers/gpu/drm/ttm/tests/ttm_bo_validate_test.c > > > > > +++ b/drivers/gpu/drm/ttm/tests/ttm_bo_validate_test.c > > > > > @@ -127,7 +127,7 @@ static void > > > > > ttm_bo_init_reserved_sys_man(struct > > > > > kunit *test) > > > > > =C2=A0 dma_resv_unlock(bo->base.resv); > > > > > =C2=A0 > > > > > =C2=A0 KUNIT_EXPECT_EQ(test, err, 0); > > > > > - KUNIT_EXPECT_EQ(test, kref_read(&bo->kref), 1); > > > > > + KUNIT_EXPECT_EQ(test, kref_read(&bo->base.refcount), > > > > > 1); > > > > > =C2=A0 KUNIT_EXPECT_PTR_EQ(test, bo->bdev, priv->ttm_dev); > > > > > =C2=A0 KUNIT_EXPECT_EQ(test, bo->type, bo_type); > > > > > =C2=A0 KUNIT_EXPECT_EQ(test, bo->page_alignment, > > > > > PAGE_SIZE); > > > > > @@ -176,7 +176,7 @@ static void > > > > > ttm_bo_init_reserved_mock_man(struct > > > > > kunit *test) > > > > > =C2=A0 dma_resv_unlock(bo->base.resv); > > > > > =C2=A0 > > > > > =C2=A0 KUNIT_EXPECT_EQ(test, err, 0); > > > > > - KUNIT_EXPECT_EQ(test, kref_read(&bo->kref), 1); > > > > > + KUNIT_EXPECT_EQ(test, kref_read(&bo->base.refcount), > > > > > 1); > > > > > =C2=A0 KUNIT_EXPECT_PTR_EQ(test, bo->bdev, priv->ttm_dev); > > > > > =C2=A0 KUNIT_EXPECT_EQ(test, bo->type, bo_type); > > > > > =C2=A0 KUNIT_EXPECT_EQ(test, ctx.bytes_moved, size); > > > > > @@ -969,6 +969,8 @@ static void > > > > > ttm_bo_validate_allowed_only_evict(struct kunit *test) > > > > > =C2=A0 ttm_mock_manager_fini(priv->ttm_dev, mem_multihop); > > > > > =C2=A0} > > > > > =C2=A0 > > > > > +extern const struct drm_gem_object_funcs > > > > > ttm_deleted_object_funcs; > > > > > + > > > > > =C2=A0static void ttm_bo_validate_deleted_evict(struct kunit > > > > > *test) > > > > > =C2=A0{ > > > > > =C2=A0 struct ttm_operation_ctx ctx_init =3D { }, ctx_val=C2=A0 = =3D > > > > > { }; > > > > > @@ -999,7 +1001,7 @@ static void > > > > > ttm_bo_validate_deleted_evict(struct > > > > > kunit *test) > > > > > =C2=A0 KUNIT_EXPECT_EQ(test, > > > > > ttm_resource_manager_usage(man), > > > > > big); > > > > > =C2=A0 > > > > > =C2=A0 dma_resv_unlock(bo_big->base.resv); > > > > > - bo_big->deleted =3D true; > > > > > + bo_big->base.funcs =3D &ttm_deleted_object_funcs; > > > > > =C2=A0 > > > > > =C2=A0 bo_small =3D ttm_bo_kunit_init(test, test->priv, > > > > > small, > > > > > NULL); > > > > > =C2=A0 bo_small->type =3D bo_type; > > > > > diff --git a/drivers/gpu/drm/ttm/tests/ttm_kunit_helpers.c > > > > > b/drivers/gpu/drm/ttm/tests/ttm_kunit_helpers.c > > > > > index 5cfe8f3f80d75..b7ab19e0e4b2b 100644 > > > > > --- a/drivers/gpu/drm/ttm/tests/ttm_kunit_helpers.c > > > > > +++ b/drivers/gpu/drm/ttm/tests/ttm_kunit_helpers.c > > > > > @@ -189,8 +189,6 @@ struct ttm_buffer_object > > > > > *ttm_bo_kunit_init(struct kunit *test, > > > > > =C2=A0 bo->bdev =3D devs->ttm_dev; > > > > > =C2=A0 bo->destroy =3D dummy_ttm_bo_destroy; > > > > > =C2=A0 > > > > > - kref_init(&bo->kref); > > > > > - > > > > > =C2=A0 return bo; > > > > > =C2=A0} > > > > > =C2=A0EXPORT_SYMBOL_GPL(ttm_bo_kunit_init); > > > > > diff --git a/drivers/gpu/drm/ttm/ttm_bo.c > > > > > b/drivers/gpu/drm/ttm/ttm_bo.c > > > > > index 3980f376e3ba4..2b470c1746f60 100644 > > > > > --- a/drivers/gpu/drm/ttm/ttm_bo.c > > > > > +++ b/drivers/gpu/drm/ttm/ttm_bo.c > > > > > @@ -246,88 +246,84 @@ static void > > > > > ttm_bo_delayed_delete(struct > > > > > work_struct *work) > > > > > =C2=A0 ttm_bo_put(bo); > > > > > =C2=A0} > > > > > =C2=A0 > > > > > -static void ttm_bo_release(struct kref *kref) > > > > > +/* > > > > > + * All other callbacks should never ever be called on a > > > > > deleted > > > > > TTM > > > > > object. > > > > > + */ > > > > > +const struct drm_gem_object_funcs ttm_deleted_object_funcs =3D > > > > > { > > > > > + .free =3D ttm_bo_free > > > > > +}; > > > > > +EXPORT_SYMBOL_FOR_TESTS_ONLY(ttm_deleted_object_funcs); > > > > > + > > > > > +/* Returns true if the BO is about to get deleted */ > > > > > +static bool ttm_bo_is_zombie(struct ttm_buffer_object *bo) > > > > > +{ > > > > > + return bo->base.funcs =3D=3D &ttm_deleted_object_funcs; > > > > > +} > > > > > + > > > > > +void ttm_bo_fini(struct ttm_buffer_object *bo) > > > > > =C2=A0{ > > > > > - struct ttm_buffer_object *bo =3D > > > > > - =C2=A0=C2=A0=C2=A0 container_of(kref, struct ttm_buffer_object, > > > > > kref); > > > > > =C2=A0 struct ttm_device *bdev =3D bo->bdev; > > > > > =C2=A0 int ret; > > > > > =C2=A0 > > > > > =C2=A0 WARN_ON_ONCE(bo->pin_count); > > > > > =C2=A0 WARN_ON_ONCE(bo->bulk_move); > > > > > =C2=A0 > > > > > - if (!bo->deleted) { > > > > > - ret =3D ttm_bo_individualize_resv(bo); > > > > > - if (ret) { > > > > > - /* Last resort, if we fail to > > > > > allocate > > > > > memory for the > > > > > - * fences block for the BO to become > > > > > idle > > > > > - */ > > > > > - dma_resv_wait_timeout(bo->base.resv, > > > > > - =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 > > > > > DMA_RESV_USAGE_BOOKKEEP, false, > > > > > - =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 30 * HZ); > > > > > - } > > > > > + ret =3D ttm_bo_individualize_resv(bo); > > > > > + if (ret) { > > > > > + /* Last resort, if we fail to allocate > > > > > memory > > > > > for > > > > > the > > > > > + * fences block for the BO to become idle > > > > > + */ > > > > > + dma_resv_wait_timeout(bo->base.resv, > > > > > DMA_RESV_USAGE_BOOKKEEP, > > > > > + =C2=A0=C2=A0=C2=A0=C2=A0=C2=A0 false, 30 * HZ); > > > > > + } > > > > > =C2=A0 > > > > > - if (bdev->funcs->release_notify) > > > > > - bdev->funcs->release_notify(bo); > > > > > + if (bo->bdev->funcs->release_notify) > > > > > + bo->bdev->funcs->release_notify(bo); > > > > > =C2=A0 > > > > > - drm_vma_offset_remove(bdev->vma_manager, > > > > > &bo- > > > > > > base.vma_node); > > > > > - ttm_mem_io_free(bdev, bo->resource); > > > > > + drm_vma_offset_remove(bdev->vma_manager, &bo- > > > > > > base.vma_node); > > > > > + ttm_mem_io_free(bdev, bo->resource); > > > > > =C2=A0 > > > > > - if (!dma_resv_test_signaled(&bo->base._resv, > > > > > - =C2=A0=C2=A0=C2=A0 > > > > > DMA_RESV_USAGE_BOOKKEEP) > > > > > > >=20 > > > > > - =C2=A0=C2=A0=C2=A0 (want_init_on_free() && (bo->ttm !=3D > > > > > NULL)) > > > > > > >=20 > > > > > - =C2=A0=C2=A0=C2=A0 bo->type =3D=3D ttm_bo_type_sg || > > > > > - =C2=A0=C2=A0=C2=A0 !dma_resv_trylock(bo->base.resv)) { > > > > > - /* The BO is not idle, resurrect it > > > > > for > > > > > delayed destroy */ > > > > > - ttm_bo_flush_all_fences(bo); > > > > > - bo->deleted =3D true; > > > > > + if (!dma_resv_test_signaled(&bo->base._resv, > > > > > DMA_RESV_USAGE_BOOKKEEP) || > > > > > + =C2=A0=C2=A0=C2=A0 (want_init_on_free() && (bo->ttm !=3D NULL))= || > > > > > + =C2=A0=C2=A0=C2=A0 bo->type =3D=3D ttm_bo_type_sg || > > > > > + =C2=A0=C2=A0=C2=A0 !dma_resv_trylock(bo->base.resv)) { > > > > > + /* The BO is not idle, resurrect it for > > > > > delayed > > > > > destroy */ > > > > > + ttm_bo_flush_all_fences(bo); > > > > > =C2=A0 > > > > > - spin_lock(&bdev->lru_lock); > > > > > - > > > > > - /* > > > > > - * Make pinned bos immediately > > > > > available > > > > > to > > > > > - * shrinkers, now that they are > > > > > queued > > > > > for > > > > > - * destruction. > > > > > - * > > > > > - * FIXME: QXL is triggering this. > > > > > Can be > > > > > removed when the > > > > > - * driver is fixed. > > > > > - */ > > > > > - if (bo->pin_count) { > > > > > - bo->pin_count =3D 0; > > > > > - > > > > > ttm_resource_move_to_lru_tai > > > > > l(bo- > > > > > > resource); > > > > > - } > > > > > + spin_lock(&bo->bdev->lru_lock); > > > > > =C2=A0 > > > > > - kref_init(&bo->kref); > > > > > - spin_unlock(&bdev->lru_lock); > > > > > + /* > > > > > + * Make pinned bos immediately available to > > > > > + * shrinkers, now that they are queued for > > > > > + * destruction. > > > > > + * > > > > > + * FIXME: QXL is triggering this. Can be > > > > > removed > > > > > when the > > > > > + * driver is fixed. > > > > > + */ > > > > > + if (bo->pin_count) { > > > > > + bo->pin_count =3D 0; > > > > > + ttm_resource_move_to_lru_tail(bo- > > > > > > resource); > > > > > + } > > > > > =C2=A0 > > > > > - INIT_WORK(&bo->delayed_delete, > > > > > ttm_bo_delayed_delete); > > > > > + kref_init(&bo->base.refcount); > > > > > + bo->base.funcs =3D &ttm_deleted_object_funcs; > > > > > + spin_unlock(&bo->bdev->lru_lock); > > > > > =C2=A0 > > > > > - /* Schedule the worker on the > > > > > closest > > > > > NUMA > > > > > node. This > > > > > - * improves performance since system > > > > > memory > > > > > might be > > > > > - * cleared on free and that is best > > > > > done > > > > > on > > > > > a CPU core > > > > > - * close to it. > > > > > - */ > > > > > - queue_work_node(bdev->pool.nid, > > > > > bdev- > > > > > > wq, > > > > > &bo->delayed_delete); > > > > > - return; > > > > > - } > > > > > + INIT_WORK(&bo->delayed_delete, > > > > > ttm_bo_delayed_delete); > > > > > =C2=A0 > > > > > + /* Schedule the worker on the closest NUMA > > > > > node. > > > > > This > > > > > + * improves performance since system memory > > > > > might be > > > > > + * cleared on free and that is best done on > > > > > a > > > > > CPU > > > > > core > > > > > + * close to it. > > > > > + */ > > > > > + queue_work_node(bdev->pool.nid, bdev->wq, > > > > > &bo- > > > > > > delayed_delete); > > > > > + } else { > > > > > =C2=A0 ttm_bo_cleanup_memtype_use(bo); > > > > > =C2=A0 dma_resv_unlock(bo->base.resv); > > > > > - } > > > > > =C2=A0 > > > > > - atomic_dec(&ttm_glob.bo_count); > > > > > - bo->destroy(bo); > > > > > -} > > > > > - > > > > > -/* TODO: remove! */ > > > > > -void ttm_bo_put(struct ttm_buffer_object *bo) > > > > > -{ > > > > > - kref_put(&bo->kref, ttm_bo_release); > > > > > -} > > > > > - > > > > > -void ttm_bo_fini(struct ttm_buffer_object *bo) > > > > > -{ > > > > > - ttm_bo_put(bo); > > > > > + atomic_dec(&ttm_glob.bo_count); > > > > > + bo->destroy(bo); > > > > > + } > > > > > =C2=A0} > > > > > =C2=A0EXPORT_SYMBOL(ttm_bo_fini); > > > > > =C2=A0 > > > > > @@ -470,7 +466,7 @@ int ttm_bo_evict_first(struct ttm_device > > > > > *bdev, > > > > > struct ttm_resource_manager *man > > > > > =C2=A0 if (!bo->resource || bo->resource->mem_type !=3D > > > > > mem_type) > > > > > =C2=A0 goto out_bo_moved; > > > > > =C2=A0 > > > > > - if (bo->deleted) { > > > > > + if (ttm_bo_is_zombie(bo)) { > > > > > =C2=A0 ret =3D ttm_bo_wait_ctx(bo, ctx); > > > > > =C2=A0 if (!ret) > > > > > =C2=A0 ttm_bo_cleanup_memtype_use(bo); > > > > > @@ -524,7 +520,7 @@ static s64 ttm_bo_evict_cb(struct > > > > > ttm_lru_walk > > > > > *walk, struct ttm_buffer_object * > > > > > =C2=A0 if (bo->pin_count || !bo->bdev->funcs- > > > > > > eviction_valuable(bo, > > > > > evict_walk->place)) > > > > > =C2=A0 return 0; > > > > > =C2=A0 > > > > > - if (bo->deleted) { > > > > > + if (ttm_bo_is_zombie(bo)) { > > > > > =C2=A0 lret =3D ttm_bo_wait_ctx(bo, walk->arg.ctx); > > > > > =C2=A0 if (!lret) > > > > > =C2=A0 ttm_bo_cleanup_memtype_use(bo); > > > > > @@ -624,7 +620,6 @@ static int ttm_bo_evict_alloc(struct > > > > > ttm_device > > > > > *bdev, > > > > > =C2=A0void ttm_bo_pin(struct ttm_buffer_object *bo) > > > > > =C2=A0{ > > > > > =C2=A0 dma_resv_assert_held(bo->base.resv); > > > > > - WARN_ON_ONCE(!kref_read(&bo->kref)); > > > > > =C2=A0 spin_lock(&bo->bdev->lru_lock); > > > > > =C2=A0 if (bo->resource) > > > > > =C2=A0 ttm_resource_del_bulk_move(bo->resource, > > > > > bo); > > > > > @@ -643,7 +638,6 @@ EXPORT_SYMBOL(ttm_bo_pin); > > > > > =C2=A0void ttm_bo_unpin(struct ttm_buffer_object *bo) > > > > > =C2=A0{ > > > > > =C2=A0 dma_resv_assert_held(bo->base.resv); > > > > > - WARN_ON_ONCE(!kref_read(&bo->kref)); > > > > > =C2=A0 if (WARN_ON_ONCE(!bo->pin_count)) > > > > > =C2=A0 return; > > > > > =C2=A0 > > > > > @@ -934,7 +928,6 @@ int ttm_bo_init_reserved(struct > > > > > ttm_device > > > > > *bdev, > > > > > struct ttm_buffer_object *bo, > > > > > =C2=A0{ > > > > > =C2=A0 int ret; > > > > > =C2=A0 > > > > > - kref_init(&bo->kref); > > > > > =C2=A0 bo->bdev =3D bdev; > > > > > =C2=A0 bo->type =3D type; > > > > > =C2=A0 bo->page_alignment =3D alignment; > > > > > @@ -1131,8 +1124,8 @@ ttm_bo_swapout_cb(struct ttm_lru_walk > > > > > *walk, > > > > > struct ttm_buffer_object *bo) > > > > > =C2=A0 goto out; > > > > > =C2=A0 } > > > > > =C2=A0 > > > > > - if (bo->deleted) { > > > > > - pgoff_t num_pages =3D tt->num_pages; > > > > > + if (ttm_bo_is_zombie(bo)) { > > > > > + pgoff_t num_pages =3D bo->ttm->num_pages; > > > > > =C2=A0 > > > > > =C2=A0 ret =3D ttm_bo_wait_ctx(bo, ctx); > > > > > =C2=A0 if (ret) > > > > > diff --git a/drivers/gpu/drm/ttm/ttm_bo_internal.h > > > > > b/drivers/gpu/drm/ttm/ttm_bo_internal.h > > > > > index e0d48eac74b03..ded2a47be0bcb 100644 > > > > > --- a/drivers/gpu/drm/ttm/ttm_bo_internal.h > > > > > +++ b/drivers/gpu/drm/ttm/ttm_bo_internal.h > > > > > @@ -27,6 +27,14 @@ > > > > > =C2=A0 > > > > > =C2=A0#include > > > > > =C2=A0 > > > > > +static inline void ttm_bo_free(struct drm_gem_object *gobj) > > > > > +{ > > > > > + struct ttm_buffer_object *bo =3D container_of(gobj, > > > > > typeof(*bo), base); > > > > > + > > > > > + atomic_dec(&ttm_glob.bo_count); > > > > > + bo->destroy(bo); > > > > > +} > > > > > + > > > > > =C2=A0/** > > > > > =C2=A0 * ttm_bo_get - reference a struct ttm_buffer_object > > > > > =C2=A0 * > > > > > @@ -34,7 +42,7 @@ > > > > > =C2=A0 */ > > > > > =C2=A0static inline void ttm_bo_get(struct ttm_buffer_object *bo) > > > > > =C2=A0{ > > > > > - kref_get(&bo->kref); > > > > > + drm_gem_object_get(&bo->base); > > > > > =C2=A0} > > > > > =C2=A0 > > > > > =C2=A0/** > > > > > @@ -50,11 +58,14 @@ static inline void ttm_bo_get(struct > > > > > ttm_buffer_object *bo) > > > > > =C2=A0static inline __must_check struct ttm_buffer_object * > > > > > =C2=A0ttm_bo_get_unless_zero(struct ttm_buffer_object *bo) > > > > > =C2=A0{ > > > > > - if (!kref_get_unless_zero(&bo->kref)) > > > > > + if (!kref_get_unless_zero(&bo->base.refcount)) > > > > > =C2=A0 return NULL; > > > > > =C2=A0 return bo; > > > > > =C2=A0} > > > > > =C2=A0 > > > > > -void ttm_bo_put(struct ttm_buffer_object *bo); > > > > > +static inline void ttm_bo_put(struct ttm_buffer_object *bo) > > > > > +{ > > > > > + drm_gem_object_put(&bo->base); > > > > > +} > > > > > =C2=A0 > > > > > =C2=A0#endif > > > > > diff --git a/drivers/gpu/drm/ttm/ttm_bo_util.c > > > > > b/drivers/gpu/drm/ttm/ttm_bo_util.c > > > > > index 3e3c201a02226..7ed085adf1c9b 100644 > > > > > --- a/drivers/gpu/drm/ttm/ttm_bo_util.c > > > > > +++ b/drivers/gpu/drm/ttm/ttm_bo_util.c > > > > > @@ -41,6 +41,18 @@ > > > > > =C2=A0 > > > > > =C2=A0#include "ttm_bo_internal.h" > > > > > =C2=A0 > > > > > +static void ttm_transfer_object_free(struct drm_gem_object > > > > > *obj) > > > > > +{ > > > > > + struct ttm_buffer_object *bo =3D > > > > > + container_of(obj, struct ttm_buffer_object, > > > > > base); > > > > > + > > > > > + ttm_bo_fini(bo); > > > > > +} > > > > > + > > > > > +const struct drm_gem_object_funcs ttm_transfer_object_funcs > > > > > =3D { > > > > > + .free =3D ttm_transfer_object_free, > > > > > +}; > > > > > + > > > > > =C2=A0struct ttm_transfer_obj { > > > > > =C2=A0 struct ttm_buffer_object base; > > > > > =C2=A0 struct ttm_buffer_object *bo; > > > > > @@ -247,7 +259,8 @@ static int > > > > > ttm_buffer_object_transfer(struct > > > > > ttm_buffer_object *bo, > > > > > =C2=A0 atomic_inc(&ttm_glob.bo_count); > > > > > =C2=A0 drm_vma_node_reset(&fbo->base.base.vma_node); > > > > > =C2=A0 > > > > > - kref_init(&fbo->base.kref); > > > > > + kref_init(&fbo->base.base.refcount); > > > > > + fbo->base.base.funcs =3D &ttm_transfer_object_funcs; > > > > > =C2=A0 fbo->base.destroy =3D &ttm_transfered_destroy; > > > > > =C2=A0 fbo->base.pin_count =3D 0; > > > > > =C2=A0 if (bo->type !=3D ttm_bo_type_sg) > > > > > diff --git a/drivers/gpu/drm/xe/xe_bo.c > > > > > b/drivers/gpu/drm/xe/xe_bo.c > > > > > index 85e6d9a0f575b..5843f850339c7 100644 > > > > > --- a/drivers/gpu/drm/xe/xe_bo.c > > > > > +++ b/drivers/gpu/drm/xe/xe_bo.c > > > > > @@ -1651,7 +1651,7 @@ static bool > > > > > xe_ttm_bo_lock_in_destructor(struct > > > > > ttm_buffer_object *ttm_bo) > > > > > =C2=A0 struct xe_device *xe =3D ttm_to_xe_device(ttm_bo- > > > > > >bdev); > > > > > =C2=A0 bool locked; > > > > > =C2=A0 > > > > > - xe_assert(xe, !kref_read(&ttm_bo->kref)); > > > > > + xe_assert(xe, !kref_read(&ttm_bo->base.refcount)); > > > > > =C2=A0 > > > > > =C2=A0 /* > > > > > =C2=A0 * We can typically only race with TTM trylocking > > > > > under > > > > > the > > > > > diff --git a/include/drm/ttm/ttm_bo.h > > > > > b/include/drm/ttm/ttm_bo.h > > > > > index 8310bc3d55f90..1eae9eea5ff32 100644 > > > > > --- a/include/drm/ttm/ttm_bo.h > > > > > +++ b/include/drm/ttm/ttm_bo.h > > > > > @@ -78,11 +78,8 @@ enum ttm_bo_type { > > > > > =C2=A0 * @type: The bo type. > > > > > =C2=A0 * @page_alignment: Page alignment. > > > > > =C2=A0 * @destroy: Destruction function. If NULL, kfree is used. > > > > > - * @kref: Reference count of this buffer object. When this > > > > > refcount > > > > > reaches > > > > > - * zero, the object is destroyed or put on the delayed > > > > > delete > > > > > list. > > > > > =C2=A0 * @resource: structure describing current placement. > > > > > =C2=A0 * @ttm: TTM structure holding system pages. > > > > > - * @deleted: True if the object is only a zombie and already > > > > > deleted. > > > > > =C2=A0 * @bulk_move: The bulk move object. > > > > > =C2=A0 * @priority: Priority for LRU, BOs with lower priority are > > > > > evicted > > > > > first. > > > > > =C2=A0 * @pin_count: Pin count. > > > > > @@ -109,17 +106,11 @@ struct ttm_buffer_object { > > > > > =C2=A0 uint32_t page_alignment; > > > > > =C2=A0 void (*destroy) (struct ttm_buffer_object *); > > > > > =C2=A0 > > > > > - /* > > > > > - * Members not needing protection. > > > > > - */ > > > > > - struct kref kref; > > > > > - > > > > > =C2=A0 /* > > > > > =C2=A0 * Members protected by the bo::resv::reserved lock. > > > > > =C2=A0 */ > > > > > =C2=A0 struct ttm_resource *resource; > > > > > =C2=A0 struct ttm_tt *ttm; > > > > > - bool deleted; > > > > > =C2=A0 struct ttm_lru_bulk_move *bulk_move; > > > > > =C2=A0 unsigned priority; > > > > > =C2=A0 unsigned pin_count;