{"id":819137,"url":"http://patchwork.ozlabs.org/api/patches/819137/?format=json","web_url":"http://patchwork.ozlabs.org/project/qemu-devel/patch/20170927125340.12360-2-berrange@redhat.com/","project":{"id":14,"url":"http://patchwork.ozlabs.org/api/projects/14/?format=json","name":"QEMU Development","link_name":"qemu-devel","list_id":"qemu-devel.nongnu.org","list_email":"qemu-devel@nongnu.org","web_url":"","scm_url":"","webscm_url":"","list_archive_url":"","list_archive_url_format":"","commit_url_format":""},"msgid":"<20170927125340.12360-2-berrange@redhat.com>","list_archive_url":null,"date":"2017-09-27T12:53:35","name":"[v4,1/6] block: use 1 MB bounce buffers for crypto instead of 16KB","commit_ref":null,"pull_url":null,"state":"new","archived":false,"hash":"cc3a90de4dbb2eb7e599690d4ba865ce2565f282","submitter":{"id":2694,"url":"http://patchwork.ozlabs.org/api/people/2694/?format=json","name":"Daniel P. Berrangé","email":"berrange@redhat.com"},"delegate":null,"mbox":"http://patchwork.ozlabs.org/project/qemu-devel/patch/20170927125340.12360-2-berrange@redhat.com/mbox/","series":[{"id":5359,"url":"http://patchwork.ozlabs.org/api/series/5359/?format=json","web_url":"http://patchwork.ozlabs.org/project/qemu-devel/list/?series=5359","date":"2017-09-27T12:53:35","name":"Misc improvements to crypto block driver","version":4,"mbox":"http://patchwork.ozlabs.org/series/5359/mbox/"}],"comments":"http://patchwork.ozlabs.org/api/patches/819137/comments/","check":"pending","checks":"http://patchwork.ozlabs.org/api/patches/819137/checks/","tags":{},"related":[],"headers":{"Return-Path":"","X-Original-To":"incoming@patchwork.ozlabs.org","Delivered-To":"patchwork-incoming@bilbo.ozlabs.org","Authentication-Results":["ozlabs.org;\n\tspf=permerror (mailfrom) smtp.mailfrom=nongnu.org\n\t(client-ip=2001:4830:134:3::11; helo=lists.gnu.org;\n\tenvelope-from=qemu-devel-bounces+incoming=patchwork.ozlabs.org@nongnu.org;\n\treceiver=)","ext-mx07.extmail.prod.ext.phx2.redhat.com;\n\tdmarc=none (p=none dis=none) header.from=redhat.com","ext-mx07.extmail.prod.ext.phx2.redhat.com;\n\tspf=fail smtp.mailfrom=berrange@redhat.com"],"Received":["from lists.gnu.org (lists.gnu.org [IPv6:2001:4830:134:3::11])\n\t(using TLSv1 with cipher AES256-SHA (256/256 bits))\n\t(No client certificate requested)\n\tby ozlabs.org (Postfix) with ESMTPS id 3y2Hpp6Tv3z9tXQ\n\tfor ;\n\tWed, 27 Sep 2017 22:54:45 +1000 (AEST)","from localhost ([::1]:54691 helo=lists.gnu.org)\n\tby lists.gnu.org with esmtp (Exim 4.71) (envelope-from\n\t)\n\tid 1dxBrO-0004Ky-1R\n\tfor incoming@patchwork.ozlabs.org; Wed, 27 Sep 2017 08:54:42 -0400","from eggs.gnu.org ([2001:4830:134:3::10]:49930)\n\tby lists.gnu.org with esmtp (Exim 4.71)\n\t(envelope-from ) id 1dxBqi-0004I2-Ac\n\tfor qemu-devel@nongnu.org; Wed, 27 Sep 2017 08:54:01 -0400","from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71)\n\t(envelope-from ) id 1dxBqe-0001Qr-Cf\n\tfor qemu-devel@nongnu.org; Wed, 27 Sep 2017 08:54:00 -0400","from mx1.redhat.com ([209.132.183.28]:52442)\n\tby eggs.gnu.org with esmtps (TLS1.0:DHE_RSA_AES_256_CBC_SHA1:32)\n\t(Exim 4.71) (envelope-from )\n\tid 1dxBqa-0001MO-Qo; Wed, 27 Sep 2017 08:53:53 -0400","from smtp.corp.redhat.com\n\t(int-mx01.intmail.prod.int.phx2.redhat.com [10.5.11.11])\n\t(using TLSv1.2 with cipher AECDH-AES256-SHA (256/256 bits))\n\t(No client certificate requested)\n\tby mx1.redhat.com (Postfix) with ESMTPS id BC13FC04D2E4;\n\tWed, 27 Sep 2017 12:53:51 +0000 (UTC)","from localhost.localdomain.com (unknown [10.42.22.189])\n\tby smtp.corp.redhat.com (Postfix) with ESMTP id 245927F760;\n\tWed, 27 Sep 2017 12:53:49 +0000 (UTC)"],"DMARC-Filter":"OpenDMARC Filter v1.3.2 mx1.redhat.com BC13FC04D2E4","From":"\"Daniel P. Berrange\" ","To":"qemu-devel@nongnu.org","Date":"Wed, 27 Sep 2017 13:53:35 +0100","Message-Id":"<20170927125340.12360-2-berrange@redhat.com>","In-Reply-To":"<20170927125340.12360-1-berrange@redhat.com>","References":"<20170927125340.12360-1-berrange@redhat.com>","X-Scanned-By":"MIMEDefang 2.79 on 10.5.11.11","X-Greylist":"Sender IP whitelisted, not delayed by milter-greylist-4.5.16\n\t(mx1.redhat.com [10.5.110.31]);\n\tWed, 27 Sep 2017 12:53:51 +0000 (UTC)","X-detected-operating-system":"by eggs.gnu.org: GNU/Linux 2.2.x-3.x [generic]\n\t[fuzzy]","X-Received-From":"209.132.183.28","Subject":"[Qemu-devel] [PATCH v4 1/6] block: use 1 MB bounce buffers for\n\tcrypto instead of 16KB","X-BeenThere":"qemu-devel@nongnu.org","X-Mailman-Version":"2.1.21","Precedence":"list","List-Id":"","List-Unsubscribe":",\n\t","List-Archive":"","List-Post":"","List-Help":"","List-Subscribe":",\n\t","Cc":"Kevin Wolf , qemu-block@nongnu.org,\n\tStefan Hajnoczi , Max Reitz ","Errors-To":"qemu-devel-bounces+incoming=patchwork.ozlabs.org@nongnu.org","Sender":"\"Qemu-devel\"\n\t"},"content":"Using 16KB bounce buffers creates a significant performance\npenalty for I/O to encrypted volumes on storage which high\nI/O latency (rotating rust & network drives), because it\ntriggers lots of fairly small I/O operations.\n\nOn tests with rotating rust, and cache=none|directsync,\nwrite speed increased from 2MiB/s to 32MiB/s, on a par\nwith that achieved by the in-kernel luks driver. With\nother cache modes the in-kernel driver is still notably\nfaster because it is able to report completion of the\nI/O request before any encryption is done, while the\nin-QEMU driver must encrypt the data before completion.\n\nSigned-off-by: Daniel P. Berrange \n---\n block/crypto.c | 28 +++++++++++++++-------------\n 1 file changed, 15 insertions(+), 13 deletions(-)","diff":"diff --git a/block/crypto.c b/block/crypto.c\nindex 58ef6f2f52..684cabeaf8 100644\n--- a/block/crypto.c\n+++ b/block/crypto.c\n@@ -379,7 +379,11 @@ static void block_crypto_close(BlockDriverState *bs)\n }\n \n \n-#define BLOCK_CRYPTO_MAX_SECTORS 32\n+/*\n+ * 1 MB bounce buffer gives good performance / memory tradeoff\n+ * when using cache=none|directsync.\n+ */\n+#define BLOCK_CRYPTO_MAX_IO_SIZE (1024 * 1024)\n \n static coroutine_fn int\n block_crypto_co_readv(BlockDriverState *bs, int64_t sector_num,\n@@ -396,12 +400,11 @@ block_crypto_co_readv(BlockDriverState *bs, int64_t sector_num,\n \n qemu_iovec_init(&hd_qiov, qiov->niov);\n \n- /* Bounce buffer so we have a linear mem region for\n- * entire sector. XXX optimize so we avoid bounce\n- * buffer in case that qiov->niov == 1\n+ /* Bounce buffer because we don't wish to expose cipher text\n+ * in qiov which points to guest memory.\n */\n cipher_data =\n- qemu_try_blockalign(bs->file->bs, MIN(BLOCK_CRYPTO_MAX_SECTORS * 512,\n+ qemu_try_blockalign(bs->file->bs, MIN(BLOCK_CRYPTO_MAX_IO_SIZE,\n qiov->size));\n if (cipher_data == NULL) {\n ret = -ENOMEM;\n@@ -411,8 +414,8 @@ block_crypto_co_readv(BlockDriverState *bs, int64_t sector_num,\n while (remaining_sectors) {\n cur_nr_sectors = remaining_sectors;\n \n- if (cur_nr_sectors > BLOCK_CRYPTO_MAX_SECTORS) {\n- cur_nr_sectors = BLOCK_CRYPTO_MAX_SECTORS;\n+ if (cur_nr_sectors > (BLOCK_CRYPTO_MAX_IO_SIZE / 512)) {\n+ cur_nr_sectors = (BLOCK_CRYPTO_MAX_IO_SIZE / 512);\n }\n \n qemu_iovec_reset(&hd_qiov);\n@@ -464,12 +467,11 @@ block_crypto_co_writev(BlockDriverState *bs, int64_t sector_num,\n \n qemu_iovec_init(&hd_qiov, qiov->niov);\n \n- /* Bounce buffer so we have a linear mem region for\n- * entire sector. XXX optimize so we avoid bounce\n- * buffer in case that qiov->niov == 1\n+ /* Bounce buffer because we're not permitted to touch\n+ * contents of qiov - it points to guest memory.\n */\n cipher_data =\n- qemu_try_blockalign(bs->file->bs, MIN(BLOCK_CRYPTO_MAX_SECTORS * 512,\n+ qemu_try_blockalign(bs->file->bs, MIN(BLOCK_CRYPTO_MAX_IO_SIZE,\n qiov->size));\n if (cipher_data == NULL) {\n ret = -ENOMEM;\n@@ -479,8 +481,8 @@ block_crypto_co_writev(BlockDriverState *bs, int64_t sector_num,\n while (remaining_sectors) {\n cur_nr_sectors = remaining_sectors;\n \n- if (cur_nr_sectors > BLOCK_CRYPTO_MAX_SECTORS) {\n- cur_nr_sectors = BLOCK_CRYPTO_MAX_SECTORS;\n+ if (cur_nr_sectors > (BLOCK_CRYPTO_MAX_IO_SIZE / 512)) {\n+ cur_nr_sectors = (BLOCK_CRYPTO_MAX_IO_SIZE / 512);\n }\n \n qemu_iovec_to_buf(qiov, bytes_done,\n","prefixes":["v4","1/6"]}