From nobody Sun May 5 17:42:20 2024 Delivered-To: importer@patchew.org Received-SPF: pass (zoho.com: domain of gnu.org designates 208.118.235.17 as permitted sender) client-ip=208.118.235.17; envelope-from=qemu-devel-bounces+importer=patchew.org@nongnu.org; helo=lists.gnu.org; Authentication-Results: mx.zohomail.com; spf=pass (zoho.com: domain of gnu.org designates 208.118.235.17 as permitted sender) smtp.mailfrom=qemu-devel-bounces+importer=patchew.org@nongnu.org; dmarc=fail(p=none dis=none) header.from=virtuozzo.com Return-Path: Received: from lists.gnu.org (lists.gnu.org [208.118.235.17]) by mx.zohomail.com with SMTPS id 1528485751459446.89255412312787; Fri, 8 Jun 2018 12:22:31 -0700 (PDT) Received: from localhost ([::1]:37778 helo=lists.gnu.org) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1fRMxy-0007oe-N4 for importer@patchew.org; Fri, 08 Jun 2018 15:22:30 -0400 Received: from eggs.gnu.org ([2001:4830:134:3::10]:38858) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1fRMwA-0006fM-39 for qemu-devel@nongnu.org; Fri, 08 Jun 2018 15:20:39 -0400 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1fRMw5-0004mI-6b for qemu-devel@nongnu.org; Fri, 08 Jun 2018 15:20:38 -0400 Received: from relay.sw.ru ([195.214.232.25]:46428) by eggs.gnu.org with esmtps (TLS1.0:DHE_RSA_AES_256_CBC_SHA1:32) (Exim 4.71) (envelope-from ) id 1fRMw4-0004kA-TE; Fri, 08 Jun 2018 15:20:33 -0400 Received: from msk-vpn.virtuozzo.com ([195.214.232.6] helo=kvm.sw.ru) by relay.sw.ru with esmtp (Exim 4.90_1) (envelope-from ) id 1fRMw0-0000OY-9V; Fri, 08 Jun 2018 22:20:28 +0300 From: Vladimir Sementsov-Ogievskiy To: qemu-devel@nongnu.org, qemu-block@nongnu.org Date: Fri, 8 Jun 2018 22:20:25 +0300 Message-Id: <20180608192027.284601-2-vsementsov@virtuozzo.com> X-Mailer: git-send-email 2.11.1 In-Reply-To: <20180608192027.284601-1-vsementsov@virtuozzo.com> References: <20180608192027.284601-1-vsementsov@virtuozzo.com> X-detected-operating-system: by eggs.gnu.org: GNU/Linux 3.x [fuzzy] X-Received-From: 195.214.232.25 Subject: [Qemu-devel] [PATCH 1/3] qemu-img: allow compressed not-in-order writes X-BeenThere: qemu-devel@nongnu.org X-Mailman-Version: 2.1.21 Precedence: list List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Cc: kwolf@redhat.com, vsementsov@virtuozzo.com, stefanha@gmail.com, pl@kamp.de, mreitz@redhat.com, den@openvz.org Errors-To: qemu-devel-bounces+importer=patchew.org@nongnu.org Sender: "Qemu-devel" X-ZohoMail: RSF_0 Z_629925259 SPT_0 Content-Transfer-Encoding: quoted-printable MIME-Version: 1.0 Content-Type: text/plain; charset="utf-8" No reason to forbid them, and they are needed to improve performance with compress-threads in further patches. Signed-off-by: Vladimir Sementsov-Ogievskiy --- qemu-img.c | 5 ----- 1 file changed, 5 deletions(-) diff --git a/qemu-img.c b/qemu-img.c index 75f1610aa0..df2657b9cb 100644 --- a/qemu-img.c +++ b/qemu-img.c @@ -2122,11 +2122,6 @@ static int img_convert(int argc, char **argv) goto fail_getopt; } =20 - if (!s.wr_in_order && s.compressed) { - error_report("Out of order write and compress are mutually exclusi= ve"); - goto fail_getopt; - } - if (tgt_image_opts && !skip_create) { error_report("--target-image-opts requires use of -n flag"); goto fail_getopt; --=20 2.11.1 From nobody Sun May 5 17:42:20 2024 Delivered-To: importer@patchew.org Received-SPF: pass (zoho.com: domain of gnu.org designates 208.118.235.17 as permitted sender) client-ip=208.118.235.17; envelope-from=qemu-devel-bounces+importer=patchew.org@nongnu.org; helo=lists.gnu.org; Authentication-Results: mx.zohomail.com; spf=pass (zoho.com: domain of gnu.org designates 208.118.235.17 as permitted sender) smtp.mailfrom=qemu-devel-bounces+importer=patchew.org@nongnu.org; dmarc=fail(p=none dis=none) header.from=virtuozzo.com Return-Path: Received: from lists.gnu.org (lists.gnu.org [208.118.235.17]) by mx.zohomail.com with SMTPS id 1528485872393396.22887367798955; Fri, 8 Jun 2018 12:24:32 -0700 (PDT) Received: from localhost ([::1]:37792 helo=lists.gnu.org) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1fRMzv-0000ws-Nm for importer@patchew.org; Fri, 08 Jun 2018 15:24:31 -0400 Received: from eggs.gnu.org ([2001:4830:134:3::10]:38861) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1fRMwA-0006fO-41 for qemu-devel@nongnu.org; Fri, 08 Jun 2018 15:20:39 -0400 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1fRMw5-0004m8-5R for qemu-devel@nongnu.org; Fri, 08 Jun 2018 15:20:38 -0400 Received: from relay.sw.ru ([195.214.232.25]:46436) by eggs.gnu.org with esmtps (TLS1.0:DHE_RSA_AES_256_CBC_SHA1:32) (Exim 4.71) (envelope-from ) id 1fRMw4-0004k8-TI; Fri, 08 Jun 2018 15:20:33 -0400 Received: from msk-vpn.virtuozzo.com ([195.214.232.6] helo=kvm.sw.ru) by relay.sw.ru with esmtp (Exim 4.90_1) (envelope-from ) id 1fRMw0-0000OY-Fk; Fri, 08 Jun 2018 22:20:28 +0300 From: Vladimir Sementsov-Ogievskiy To: qemu-devel@nongnu.org, qemu-block@nongnu.org Date: Fri, 8 Jun 2018 22:20:26 +0300 Message-Id: <20180608192027.284601-3-vsementsov@virtuozzo.com> X-Mailer: git-send-email 2.11.1 In-Reply-To: <20180608192027.284601-1-vsementsov@virtuozzo.com> References: <20180608192027.284601-1-vsementsov@virtuozzo.com> X-detected-operating-system: by eggs.gnu.org: GNU/Linux 3.x [fuzzy] X-Received-From: 195.214.232.25 Subject: [Qemu-devel] [PATCH 2/3] qcow2: refactor data compression X-BeenThere: qemu-devel@nongnu.org X-Mailman-Version: 2.1.21 Precedence: list List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Cc: kwolf@redhat.com, vsementsov@virtuozzo.com, stefanha@gmail.com, pl@kamp.de, mreitz@redhat.com, den@openvz.org Errors-To: qemu-devel-bounces+importer=patchew.org@nongnu.org Sender: "Qemu-devel" X-ZohoMail: RSF_0 Z_629925259 SPT_0 Content-Transfer-Encoding: quoted-printable MIME-Version: 1.0 Content-Type: text/plain; charset="utf-8" Make a separate function for compression to be parallelized later. - use .avail_aut field instead of .next_out to calculate size of compressed data. It looks more natural and it allows to keep dest to be void pointer - set avail_out to be at least one byte less than input, to be sure avoid inefficient compression earlier Signed-off-by: Vladimir Sementsov-Ogievskiy --- block/qcow2.c | 74 +++++++++++++++++++++++++++++++++++++------------------= ---- 1 file changed, 47 insertions(+), 27 deletions(-) diff --git a/block/qcow2.c b/block/qcow2.c index 549fee9b69..d4dbe329ab 100644 --- a/block/qcow2.c +++ b/block/qcow2.c @@ -22,11 +22,13 @@ * THE SOFTWARE. */ =20 +#define ZLIB_CONST +#include + #include "qemu/osdep.h" #include "block/block_int.h" #include "sysemu/block-backend.h" #include "qemu/module.h" -#include #include "qcow2.h" #include "qemu/error-report.h" #include "qapi/error.h" @@ -3674,6 +3676,45 @@ static int qcow2_truncate(BlockDriverState *bs, int6= 4_t offset, return 0; } =20 +/* qcow2_compress() + * + * @dest - destination buffer, at least of @size-1 bytes + * @src - source buffer, @size bytes + * + * Returns: compressed size on success + * -1 if compression is inefficient + * -2 on any other error + */ +static ssize_t qcow2_compress(void *dest, const void *src, size_t size) +{ + ssize_t ret; + z_stream strm; + + /* best compression, small window, no zlib header */ + memset(&strm, 0, sizeof(strm)); + ret =3D deflateInit2(&strm, Z_DEFAULT_COMPRESSION, Z_DEFLATED, + -12, 9, Z_DEFAULT_STRATEGY); + if (ret !=3D 0) { + return -2; + } + + strm.avail_in =3D size; + strm.next_in =3D src; + strm.avail_out =3D size - 1; + strm.next_out =3D dest; + + ret =3D deflate(&strm, Z_FINISH); + if (ret =3D=3D Z_STREAM_END) { + ret =3D size - 1 - strm.avail_out; + } else { + ret =3D (ret =3D=3D Z_OK ? -1 : -2); + } + + deflateEnd(&strm); + + return ret; +} + /* XXX: put compressed sectors first, then all the cluster aligned tables to avoid losing bytes in alignment */ static coroutine_fn int @@ -3683,8 +3724,8 @@ qcow2_co_pwritev_compressed(BlockDriverState *bs, uin= t64_t offset, BDRVQcow2State *s =3D bs->opaque; QEMUIOVector hd_qiov; struct iovec iov; - z_stream strm; - int ret, out_len; + int ret; + size_t out_len; uint8_t *buf, *out_buf; int64_t cluster_offset; =20 @@ -3717,32 +3758,11 @@ qcow2_co_pwritev_compressed(BlockDriverState *bs, u= int64_t offset, =20 out_buf =3D g_malloc(s->cluster_size); =20 - /* best compression, small window, no zlib header */ - memset(&strm, 0, sizeof(strm)); - ret =3D deflateInit2(&strm, Z_DEFAULT_COMPRESSION, - Z_DEFLATED, -12, - 9, Z_DEFAULT_STRATEGY); - if (ret !=3D 0) { - ret =3D -EINVAL; - goto fail; - } - - strm.avail_in =3D s->cluster_size; - strm.next_in =3D (uint8_t *)buf; - strm.avail_out =3D s->cluster_size; - strm.next_out =3D out_buf; - - ret =3D deflate(&strm, Z_FINISH); - if (ret !=3D Z_STREAM_END && ret !=3D Z_OK) { - deflateEnd(&strm); + out_len =3D qcow2_compress(out_buf, buf, s->cluster_size); + if (out_len =3D=3D -2) { ret =3D -EINVAL; goto fail; - } - out_len =3D strm.next_out - out_buf; - - deflateEnd(&strm); - - if (ret !=3D Z_STREAM_END || out_len >=3D s->cluster_size) { + } else if (out_len =3D=3D -1) { /* could not compress: write normal cluster */ ret =3D qcow2_co_pwritev(bs, offset, bytes, qiov, 0); if (ret < 0) { --=20 2.11.1 From nobody Sun May 5 17:42:20 2024 Delivered-To: importer@patchew.org Received-SPF: pass (zoho.com: domain of gnu.org designates 208.118.235.17 as permitted sender) client-ip=208.118.235.17; envelope-from=qemu-devel-bounces+importer=patchew.org@nongnu.org; helo=lists.gnu.org; Authentication-Results: mx.zohomail.com; spf=pass (zoho.com: domain of gnu.org designates 208.118.235.17 as permitted sender) smtp.mailfrom=qemu-devel-bounces+importer=patchew.org@nongnu.org; dmarc=fail(p=none dis=none) header.from=virtuozzo.com Return-Path: Received: from lists.gnu.org (lists.gnu.org [208.118.235.17]) by mx.zohomail.com with SMTPS id 1528485756112865.7593427109848; Fri, 8 Jun 2018 12:22:36 -0700 (PDT) Received: from localhost ([::1]:37779 helo=lists.gnu.org) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1fRMxz-0007pG-Ot for importer@patchew.org; Fri, 08 Jun 2018 15:22:31 -0400 Received: from eggs.gnu.org ([2001:4830:134:3::10]:38864) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1fRMwA-0006fP-57 for qemu-devel@nongnu.org; Fri, 08 Jun 2018 15:20:39 -0400 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1fRMw5-0004mU-BI for qemu-devel@nongnu.org; Fri, 08 Jun 2018 15:20:38 -0400 Received: from relay.sw.ru ([195.214.232.25]:46430) by eggs.gnu.org with esmtps (TLS1.0:DHE_RSA_AES_256_CBC_SHA1:32) (Exim 4.71) (envelope-from ) id 1fRMw5-0004k7-2s; Fri, 08 Jun 2018 15:20:33 -0400 Received: from msk-vpn.virtuozzo.com ([195.214.232.6] helo=kvm.sw.ru) by relay.sw.ru with esmtp (Exim 4.90_1) (envelope-from ) id 1fRMw0-0000OY-Js; Fri, 08 Jun 2018 22:20:28 +0300 From: Vladimir Sementsov-Ogievskiy To: qemu-devel@nongnu.org, qemu-block@nongnu.org Date: Fri, 8 Jun 2018 22:20:27 +0300 Message-Id: <20180608192027.284601-4-vsementsov@virtuozzo.com> X-Mailer: git-send-email 2.11.1 In-Reply-To: <20180608192027.284601-1-vsementsov@virtuozzo.com> References: <20180608192027.284601-1-vsementsov@virtuozzo.com> X-detected-operating-system: by eggs.gnu.org: GNU/Linux 3.x [fuzzy] X-Received-From: 195.214.232.25 Subject: [Qemu-devel] [PATCH 3/3] qcow2: add compress threads X-BeenThere: qemu-devel@nongnu.org X-Mailman-Version: 2.1.21 Precedence: list List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Cc: kwolf@redhat.com, vsementsov@virtuozzo.com, stefanha@gmail.com, pl@kamp.de, mreitz@redhat.com, den@openvz.org Errors-To: qemu-devel-bounces+importer=patchew.org@nongnu.org Sender: "Qemu-devel" X-ZohoMail: RSF_0 Z_629925259 SPT_0 Content-Transfer-Encoding: quoted-printable MIME-Version: 1.0 Content-Type: text/plain; charset="utf-8" Do data compression in separate threads. This significantly improve performance for qemu-img convert with -W (allow async writes) and -c (compressed) options. Signed-off-by: Vladimir Sementsov-Ogievskiy --- block/qcow2.h | 3 +++ block/qcow2.c | 62 +++++++++++++++++++++++++++++++++++++++++++++++++++++++= +++- 2 files changed, 64 insertions(+), 1 deletion(-) diff --git a/block/qcow2.h b/block/qcow2.h index 01b5250415..0bd21623c2 100644 --- a/block/qcow2.h +++ b/block/qcow2.h @@ -326,6 +326,9 @@ typedef struct BDRVQcow2State { * override) */ char *image_backing_file; char *image_backing_format; + + CoQueue compress_wait_queue; + int nb_compress_threads; } BDRVQcow2State; =20 typedef struct Qcow2COWRegion { diff --git a/block/qcow2.c b/block/qcow2.c index d4dbe329ab..91465893e2 100644 --- a/block/qcow2.c +++ b/block/qcow2.c @@ -42,6 +42,7 @@ #include "qapi/qobject-input-visitor.h" #include "qapi/qapi-visit-block-core.h" #include "crypto.h" +#include "block/thread-pool.h" =20 /* Differences with QCOW: @@ -1544,6 +1545,9 @@ static int coroutine_fn qcow2_do_open(BlockDriverStat= e *bs, QDict *options, qcow2_check_refcounts(bs, &result, 0); } #endif + + qemu_co_queue_init(&s->compress_wait_queue); + return ret; =20 fail: @@ -3715,6 +3719,62 @@ static ssize_t qcow2_compress(void *dest, const void= *src, size_t size) return ret; } =20 +#define MAX_COMPRESS_THREADS 4 + +typedef struct Qcow2CompressData { + void *dest; + const void *src; + size_t size; + ssize_t ret; +} Qcow2CompressData; + +static int qcow2_compress_pool_func(void *opaque) +{ + Qcow2CompressData *data =3D opaque; + + data->ret =3D qcow2_compress(data->dest, data->src, data->size); + + return 0; +} + +static void qcow2_compress_complete(void *opaque, int ret) +{ + qemu_coroutine_enter(opaque); +} + +/* See qcow2_compress definition for parameters description */ +static ssize_t qcow2_co_compress(BlockDriverState *bs, + void *dest, const void *src, size_t size) +{ + BDRVQcow2State *s =3D bs->opaque; + BlockAIOCB *acb; + ThreadPool *pool =3D aio_get_thread_pool(bdrv_get_aio_context(bs)); + Qcow2CompressData arg =3D { + .dest =3D dest, + .src =3D src, + .size =3D size, + }; + + while (s->nb_compress_threads >=3D MAX_COMPRESS_THREADS) { + qemu_co_queue_wait(&s->compress_wait_queue, NULL); + } + + s->nb_compress_threads++; + acb =3D thread_pool_submit_aio(pool, qcow2_compress_pool_func, &arg, + qcow2_compress_complete, + qemu_coroutine_self()); + + if (!acb) { + s->nb_compress_threads--; + return -EINVAL; + } + qemu_coroutine_yield(); + s->nb_compress_threads--; + qemu_co_queue_next(&s->compress_wait_queue); + + return arg.ret; +} + /* XXX: put compressed sectors first, then all the cluster aligned tables to avoid losing bytes in alignment */ static coroutine_fn int @@ -3758,7 +3818,7 @@ qcow2_co_pwritev_compressed(BlockDriverState *bs, uin= t64_t offset, =20 out_buf =3D g_malloc(s->cluster_size); =20 - out_len =3D qcow2_compress(out_buf, buf, s->cluster_size); + out_len =3D qcow2_co_compress(bs, out_buf, buf, s->cluster_size); if (out_len =3D=3D -2) { ret =3D -EINVAL; goto fail; --=20 2.11.1