From nobody Wed Oct 8 02:04:52 2025 Received: from bali.collaboradmins.com (bali.collaboradmins.com [148.251.105.195]) (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 099EA2BE7B3 for ; Thu, 3 Jul 2025 20:53:53 +0000 (UTC) Authentication-Results: smtp.subspace.kernel.org; arc=none smtp.client-ip=148.251.105.195 ARC-Seal: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1751576035; cv=none; b=XKo6Bw0uOxryAuKz/ySfY1FGpiXuj+f7jv8zrKV+cBqZNEeye+9PS3Sbda0KUAYZz6NhSq3yyo9vutjaf/aneOCapABU88SippRHQUWp/Te4V4GKKZLekaHj84j4vIL5CdSyLlT7zlsvwMSmwN5Tx8PfwScA8RiO6sVCJqSNzpg= ARC-Message-Signature: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1751576035; c=relaxed/simple; bh=pnptQuQ9ul8nylrVc2pKAZ2SD1Z/hkCiZqbmBBAxVdU=; h=From:To:Cc:Subject:Date:Message-ID:In-Reply-To:References: MIME-Version; b=lSgZRSnu2Wt274y+gTpSqEK0K2NRKyetIIrcRlWIX0OJ9WOOhF54Udf2knYLj1Ix6t4PasIIZv7nNrF1SRdBd5IhQFkyVatieCYKChteBdFQkBRG4buOD0kkkwY5DmzsBeL8Ip3N3TyQy/w634lSU/+4xXIfSvMA1ji4Qxl9Ack= ARC-Authentication-Results: i=1; smtp.subspace.kernel.org; dmarc=pass (p=none dis=none) header.from=collabora.com; spf=pass smtp.mailfrom=collabora.com; dkim=pass (2048-bit key) header.d=collabora.com header.i=@collabora.com header.b=PYdJcb0F; arc=none smtp.client-ip=148.251.105.195 Authentication-Results: smtp.subspace.kernel.org; dmarc=pass (p=none dis=none) header.from=collabora.com Authentication-Results: smtp.subspace.kernel.org; spf=pass smtp.mailfrom=collabora.com Authentication-Results: smtp.subspace.kernel.org; dkim=pass (2048-bit key) header.d=collabora.com header.i=@collabora.com header.b="PYdJcb0F" DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=collabora.com; s=mail; t=1751576031; bh=pnptQuQ9ul8nylrVc2pKAZ2SD1Z/hkCiZqbmBBAxVdU=; h=From:To:Cc:Subject:Date:In-Reply-To:References:From; b=PYdJcb0FkW1uBBauZkQEQ3IOCEqR5xAjkY9WwM9shmkmK9kSaJM94O5Ca9pxcAscx ljGBrYJysbFFryqoapW8CPuxfasZ9fXCSqbU/n7BT73nM6ecxWE/7JgDexq2gLoUfm o8N6vdOOwLqGLMyuEpnbESLCXl6GUINhem3qULEX+auyrfO3GZZi+NRqNVA0ONw/Dr rwRd6S9pT9XXdMwuqY3U57ew7xAkIEtEOYjmqob9dILtk0Wcs8q1RIdjbu7Uwh3wNt bTyV5iAR3yEYczXSGw2gMl9BPdZ2XtkDzXwKY4JM5ZVs85kNjlPr/sr26oVYYv4Usb Lm3cB2qQFi9BA== Received: from debian-rockchip-rock5b-rk3588.. (unknown [90.168.160.154]) (using TLSv1.3 with cipher TLS_AES_256_GCM_SHA384 (256/256 bits) key-exchange X25519 server-signature RSA-PSS (4096 bits) server-digest SHA256) (No client certificate requested) (Authenticated sender: nanokatze) by bali.collaboradmins.com (Postfix) with ESMTPSA id 18D3017E10C7; Thu, 3 Jul 2025 22:53:50 +0200 (CEST) From: Caterina Shablia To: "Maarten Lankhorst" , "Maxime Ripard" , "Thomas Zimmermann" , "David Airlie" , "Simona Vetter" , "Frank Binns" , "Matt Coster" , "Karol Herbst" , "Lyude Paul" , "Danilo Krummrich" , "Boris Brezillon" , "Steven Price" , "Liviu Dudau" , "Lucas De Marchi" , =?UTF-8?q?Thomas=20Hellstr=C3=B6m?= , "Rodrigo Vivi" Cc: dri-devel@lists.freedesktop.org, linux-kernel@vger.kernel.org, nouveau@lists.freedesktop.org, intel-xe@lists.freedesktop.org, Asahi Lina , Caterina Shablia Subject: [PATCH v3 6/7] drm/gpuvm: Add DRM_GPUVA_REPEAT flag and logic Date: Thu, 3 Jul 2025 20:52:58 +0000 Message-ID: <20250703205308.19419-7-caterina.shablia@collabora.com> X-Mailer: git-send-email 2.47.2 In-Reply-To: <20250703205308.19419-1-caterina.shablia@collabora.com> References: <20250703205308.19419-1-caterina.shablia@collabora.com> Precedence: bulk X-Mailing-List: linux-kernel@vger.kernel.org List-Id: List-Subscribe: List-Unsubscribe: MIME-Version: 1.0 Content-Transfer-Encoding: quoted-printable Content-Type: text/plain; charset="utf-8" From: Asahi Lina To be able to support "fake sparse" mappings without relying on GPU page fault handling, drivers may need to create large (e.g. 4GiB) mappings of the same page repeatedly (or same range of pages). Doing this through individual mappings would be very wasteful. This can be handled better by using a flag on map creation, but to do it safely, drm_gpuvm needs to be aware of this special case. Add a flag that signals that a given mapping is a page mapping, which is repeated all over the entire requested VA range. This tweaks the sm_map() logic to treat the GEM offsets differently when mappings are a repeated ones so they are not incremented as they would be with regular mappings. The size of the GEM portion to repeat is passed through drm_gpuva::gem::range. Most of the time it will be a page size, but it can be bigger as long as it's less that drm_gpuva::va::range, and drm_gpuva::gem::range is a multiple of drm_gpuva::va::range. Signed-off-by: Asahi Lina Signed-off-by: Caterina Shablia --- drivers/gpu/drm/drm_gpuvm.c | 71 +++++++++++++++++++++++++++++++++---- include/drm/drm_gpuvm.h | 43 +++++++++++++++++++++- 2 files changed, 107 insertions(+), 7 deletions(-) diff --git a/drivers/gpu/drm/drm_gpuvm.c b/drivers/gpu/drm/drm_gpuvm.c index a24b6159a0d4..7b0c90119d32 100644 --- a/drivers/gpu/drm/drm_gpuvm.c +++ b/drivers/gpu/drm/drm_gpuvm.c @@ -2063,6 +2063,7 @@ op_map_cb(const struct drm_gpuvm_ops *fn, void *priv, op.map.va.range =3D req->va.range; op.map.gem.obj =3D req->gem.obj; op.map.gem.offset =3D req->gem.offset; + op.map.gem.range =3D req->gem.range; op.map.flags =3D req->flags; =20 return fn->sm_step_map(&op, priv); @@ -2122,12 +2123,53 @@ static bool can_merge(struct drm_gpuvm *gpuvm, cons= t struct drm_gpuva *a, if (drm_WARN_ON(gpuvm->drm, b->va.addr > a->va.addr + a->va.range)) return false; =20 + if (a->flags & DRM_GPUVA_REPEAT) { + u64 va_diff =3D b->va.addr - a->va.addr; + + /* If this is a repeated mapping, both the GEM range + * and offset must match. + */ + if (a->gem.range !=3D b->gem.range || + a->gem.offset !=3D b->gem.offset) + return false; + + /* The difference between the VA addresses must be a + * multiple of the repeated range, otherwise there's + * a shift. + */ + if (do_div(va_diff, a->gem.range)) + return false; + + return true; + } + /* We intentionally ignore u64 underflows because all we care about * here is whether the VA diff matches the GEM offset diff. */ return b->va.addr - a->va.addr =3D=3D b->gem.offset - a->gem.offset; } =20 +static int check_map_req(struct drm_gpuvm *gpuvm, + const struct drm_gpuvm_map_req *req) +{ + if (unlikely(!drm_gpuvm_range_valid(gpuvm, req->va.addr, req->va.range))) + return -EINVAL; + + if (req->flags & DRM_GPUVA_REPEAT) { + u64 va_range =3D req->va.range; + + /* For a repeated mapping, GEM range must be > 0 + * and a multiple of the VA range. + */ + if (unlikely(!req->gem.range || + (va_range < req->gem.range) || + do_div(va_range, req->gem.range))) + return -EINVAL; + } + + return 0; +} + static int __drm_gpuvm_sm_map(struct drm_gpuvm *gpuvm, const struct drm_gpuvm_ops *ops, void *priv, @@ -2137,6 +2179,7 @@ __drm_gpuvm_sm_map(struct drm_gpuvm *gpuvm, struct drm_gpuva reqva =3D { .va.addr =3D req->va.addr, .va.range =3D req->va.range, + .gem.range =3D req->gem.range, .gem.offset =3D req->gem.offset, .gem.obj =3D req->gem.obj, .flags =3D req->flags, @@ -2144,7 +2187,8 @@ __drm_gpuvm_sm_map(struct drm_gpuvm *gpuvm, u64 req_end =3D req->va.addr + req->va.range; int ret; =20 - if (unlikely(!drm_gpuvm_range_valid(gpuvm, req->va.addr, req->va.range))) + ret =3D check_map_req(gpuvm, req); + if (unlikely(ret)) return -EINVAL; =20 drm_gpuvm_for_each_va_range_safe(va, next, gpuvm, req->va.addr, req_end) { @@ -2175,7 +2219,8 @@ __drm_gpuvm_sm_map(struct drm_gpuvm *gpuvm, .va.addr =3D req_end, .va.range =3D range - req->va.range, .gem.obj =3D obj, - .gem.offset =3D offset + req->va.range, + .gem.range =3D va->gem.range, + .gem.offset =3D offset, .flags =3D va->flags, }; struct drm_gpuva_op_unmap u =3D { @@ -2183,6 +2228,9 @@ __drm_gpuvm_sm_map(struct drm_gpuvm *gpuvm, .keep =3D merge, }; =20 + if (!(va->flags & DRM_GPUVA_REPEAT)) + n.gem.offset +=3D req->va.range; + ret =3D op_remap_cb(ops, priv, NULL, &n, &u); if (ret) return ret; @@ -2194,6 +2242,7 @@ __drm_gpuvm_sm_map(struct drm_gpuvm *gpuvm, .va.addr =3D addr, .va.range =3D ls_range, .gem.obj =3D obj, + .gem.range =3D va->gem.range, .gem.offset =3D offset, .flags =3D va->flags, }; @@ -2220,11 +2269,14 @@ __drm_gpuvm_sm_map(struct drm_gpuvm *gpuvm, .va.addr =3D req_end, .va.range =3D end - req_end, .gem.obj =3D obj, - .gem.offset =3D offset + ls_range + - req->va.range, + .gem.range =3D va->gem.range, + .gem.offset =3D offset, .flags =3D va->flags, }; =20 + if (!(va->flags & DRM_GPUVA_REPEAT)) + n.gem.offset +=3D ls_range + req->va.range; + ret =3D op_remap_cb(ops, priv, &p, &n, &u); if (ret) return ret; @@ -2250,7 +2302,8 @@ __drm_gpuvm_sm_map(struct drm_gpuvm *gpuvm, .va.addr =3D req_end, .va.range =3D end - req_end, .gem.obj =3D obj, - .gem.offset =3D offset + req_end - addr, + .gem.range =3D va->gem.range, + .gem.offset =3D offset, .flags =3D va->flags, }; struct drm_gpuva_op_unmap u =3D { @@ -2258,6 +2311,8 @@ __drm_gpuvm_sm_map(struct drm_gpuvm *gpuvm, .keep =3D merge, }; =20 + if (!(va->flags & DRM_GPUVA_REPEAT)) + n.gem.offset +=3D req_end - addr; =20 ret =3D op_remap_cb(ops, priv, NULL, &n, &u); if (ret) @@ -2295,6 +2350,7 @@ __drm_gpuvm_sm_unmap(struct drm_gpuvm *gpuvm, prev.va.addr =3D addr; prev.va.range =3D req_addr - addr; prev.gem.obj =3D obj; + prev.gem.range =3D va->gem.range; prev.gem.offset =3D offset; prev.flags =3D va->flags; =20 @@ -2305,7 +2361,10 @@ __drm_gpuvm_sm_unmap(struct drm_gpuvm *gpuvm, next.va.addr =3D req_end; next.va.range =3D end - req_end; next.gem.obj =3D obj; - next.gem.offset =3D offset + (req_end - addr); + prev.gem.range =3D va->gem.range; + next.gem.offset =3D offset; + if (!(va->flags & DRM_GPUVA_REPEAT)) + next.gem.offset +=3D req_end - addr; next.flags =3D va->flags; =20 next_split =3D true; diff --git a/include/drm/drm_gpuvm.h b/include/drm/drm_gpuvm.h index f77a89e791f1..629e8508f99f 100644 --- a/include/drm/drm_gpuvm.h +++ b/include/drm/drm_gpuvm.h @@ -56,10 +56,19 @@ enum drm_gpuva_flags { */ DRM_GPUVA_SPARSE =3D (1 << 1), =20 + /** + * @DRM_GPUVA_REPEAT: + * + * Flag indicating that the &drm_gpuva is a mapping of a GEM + * portion repeated multiple times to fill the virtual address + * range. + */ + DRM_GPUVA_REPEAT =3D (1 << 2), + /** * @DRM_GPUVA_USERBITS: user defined bits */ - DRM_GPUVA_USERBITS =3D (1 << 2), + DRM_GPUVA_USERBITS =3D (1 << 3), }; =20 /** @@ -111,6 +120,18 @@ struct drm_gpuva { */ u64 offset; =20 + /* + * @gem.range: the range of the GEM that is mapped + * + * When dealing with normal mappings, this must be zero. + * When flags has DRM_GPUVA_REPEAT set, this field must be + * smaller than va.range and va.range must be a multiple of + * gem.range. + * This is a u32 not a u64 because we expect repeated mappings + * to be pointing to relatively small portions of a GEM object. + */ + u32 range; + /** * @gem.obj: the mapped &drm_gem_object */ @@ -842,6 +863,17 @@ struct drm_gpuva_op_map { */ u64 offset; =20 + /* + * @gem.range: the range of the GEM that is mapped + * + * When dealing with normal mappings, this must be zero. + * When flags has DRM_GPUVA_REPEAT set, it must be smaller + * and be a multiple of va.range. This is a u32 not a u64 + * because we expect repeated mappings to be pointing to + * a relatively small portion of a GEM object. + */ + u32 range; + /** * @gem.obj: the &drm_gem_object to map */ @@ -1078,6 +1110,15 @@ struct drm_gpuvm_map_req { =20 /** @offset: offset in the GEM */ u64 offset; + + /** + * @range: size of the range of the GEM object to map + * + * Must be zero unless flags has DRM_GPUVA_REPEAT set. + * If DRM_GPUVA_REPEAT is set, this field must be less than va.range, + * and va.range must be a multiple of gem.range. + */ + u32 range; } gem; =20 /** @flags: combination of DRM_GPUVA_ flags describing the mapping proper= ties. */ --=20 2.47.2