From patchwork Thu Jan 31 15:19:12 2019 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Stefano Garzarella X-Patchwork-Id: 10790805 Return-Path: Received: from mail.wl.linuxfoundation.org (pdx-wl-mail.web.codeaurora.org [172.30.200.125]) by pdx-korg-patchwork-2.web.codeaurora.org (Postfix) with ESMTP id E28BF139A for ; Thu, 31 Jan 2019 15:23:10 +0000 (UTC) Received: from mail.wl.linuxfoundation.org (localhost [127.0.0.1]) by mail.wl.linuxfoundation.org (Postfix) with ESMTP id D2FA431404 for ; Thu, 31 Jan 2019 15:23:10 +0000 (UTC) Received: by mail.wl.linuxfoundation.org (Postfix, from userid 486) id C5A6731407; Thu, 31 Jan 2019 15:23:10 +0000 (UTC) X-Spam-Checker-Version: SpamAssassin 3.3.1 (2010-03-16) on pdx-wl-mail.web.codeaurora.org X-Spam-Level: X-Spam-Status: No, score=-2.9 required=2.0 tests=BAYES_00,MAILING_LIST_MULTI autolearn=ham version=3.3.1 Received: from lists.gnu.org (lists.gnu.org [209.51.188.17]) (using TLSv1 with cipher AES256-SHA (256/256 bits)) (No client certificate requested) by mail.wl.linuxfoundation.org (Postfix) with ESMTPS id 1922B31404 for ; Thu, 31 Jan 2019 15:23:10 +0000 (UTC) Received: from localhost ([127.0.0.1]:56354 helo=lists.gnu.org) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1gpEBJ-0006rw-G2 for patchwork-qemu-devel@patchwork.kernel.org; Thu, 31 Jan 2019 10:23:09 -0500 Received: from eggs.gnu.org ([209.51.188.92]:48073) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1gpE7z-0004KM-Kv for qemu-devel@nongnu.org; Thu, 31 Jan 2019 10:19:46 -0500 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1gpE7x-0007hL-KP for qemu-devel@nongnu.org; Thu, 31 Jan 2019 10:19:43 -0500 Received: from mx1.redhat.com ([209.132.183.28]:18354) by eggs.gnu.org with esmtps (TLS1.0:DHE_RSA_AES_256_CBC_SHA1:32) (Exim 4.71) (envelope-from ) id 1gpE7r-0007Yh-Lz; Thu, 31 Jan 2019 10:19:37 -0500 Received: from smtp.corp.redhat.com (int-mx08.intmail.prod.int.phx2.redhat.com [10.5.11.23]) (using TLSv1.2 with cipher AECDH-AES256-SHA (256/256 bits)) (No client certificate requested) by mx1.redhat.com (Postfix) with ESMTPS id 6AFEEA0915; Thu, 31 Jan 2019 15:19:33 +0000 (UTC) Received: from steredhat.redhat.com (ovpn-117-151.ams2.redhat.com [10.36.117.151]) by smtp.corp.redhat.com (Postfix) with ESMTP id 5C5A919C7B; Thu, 31 Jan 2019 15:19:30 +0000 (UTC) From: Stefano Garzarella To: qemu-devel@nongnu.org Date: Thu, 31 Jan 2019 16:19:12 +0100 Message-Id: <20190131151914.164903-4-sgarzare@redhat.com> In-Reply-To: <20190131151914.164903-1-sgarzare@redhat.com> References: <20190131151914.164903-1-sgarzare@redhat.com> MIME-Version: 1.0 X-Scanned-By: MIMEDefang 2.84 on 10.5.11.23 X-Greylist: Sender IP whitelisted, not delayed by milter-greylist-4.5.16 (mx1.redhat.com [10.5.110.39]); Thu, 31 Jan 2019 15:19:33 +0000 (UTC) X-detected-operating-system: by eggs.gnu.org: GNU/Linux 2.2.x-3.x [generic] X-Received-From: 209.132.183.28 Subject: [Qemu-devel] [PATCH v2 3/5] virtio-blk: add DISCARD and WRITE ZEROES features 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: Kevin Wolf , Laurent Vivier , Thomas Huth , Eduardo Habkost , qemu-block@nongnu.org, "Michael S. Tsirkin" , "Dr . David Alan Gilbert" , Max Reitz , Stefan Hajnoczi , Paolo Bonzini Errors-To: qemu-devel-bounces+patchwork-qemu-devel=patchwork.kernel.org@nongnu.org Sender: "Qemu-devel" X-Virus-Scanned: ClamAV using ClamSMTP This patch adds the support of DISCARD and WRITE ZEROES commands, that have been introduced in the virtio-blk protocol to have better performance when using SSD backend. We support only one segment per request since multiple segments are not widely used and there are no userspace APIs that allow applications to submit multiple segments in a single call. Signed-off-by: Stefano Garzarella Reviewed-by: Michael S. Tsirkin --- hw/block/virtio-blk.c | 173 +++++++++++++++++++++++++++++++++ include/hw/virtio/virtio-blk.h | 2 + 2 files changed, 175 insertions(+) diff --git a/hw/block/virtio-blk.c b/hw/block/virtio-blk.c index 542ec52536..34ee676895 100644 --- a/hw/block/virtio-blk.c +++ b/hw/block/virtio-blk.c @@ -147,6 +147,30 @@ out: aio_context_release(blk_get_aio_context(s->conf.conf.blk)); } +static void virtio_blk_discard_wzeroes_complete(void *opaque, int ret) +{ + VirtIOBlockReq *req = opaque; + VirtIOBlock *s = req->dev; + bool is_wzeroes = (virtio_ldl_p(VIRTIO_DEVICE(req->dev), &req->out.type) & + ~VIRTIO_BLK_T_BARRIER) == VIRTIO_BLK_T_WRITE_ZEROES; + + aio_context_acquire(blk_get_aio_context(s->conf.conf.blk)); + if (ret) { + if (virtio_blk_handle_rw_error(req, -ret, 0, is_wzeroes)) { + goto out; + } + } + + virtio_blk_req_complete(req, VIRTIO_BLK_S_OK); + if (is_wzeroes) { + block_acct_done(blk_get_stats(req->dev->blk), &req->acct); + } + virtio_blk_free_request(req); + +out: + aio_context_release(blk_get_aio_context(s->conf.conf.blk)); +} + #ifdef __linux__ typedef struct { @@ -480,6 +504,82 @@ static bool virtio_blk_sect_range_ok(VirtIOBlock *dev, return true; } +static uint8_t virtio_blk_handle_dwz(VirtIOBlockReq *req, bool is_wzeroes, + struct virtio_blk_discard_write_zeroes *dwz_hdr) +{ + VirtIOBlock *s = req->dev; + uint64_t sector; + uint32_t num_sectors, flags; + uint8_t err_status; + int bytes; + + sector = virtio_ldq_p(VIRTIO_DEVICE(req->dev), &dwz_hdr->sector); + num_sectors = virtio_ldl_p(VIRTIO_DEVICE(req->dev), &dwz_hdr->num_sectors); + flags = virtio_ldl_p(VIRTIO_DEVICE(req->dev), &dwz_hdr->flags); + + /* + * dwz_max_sectors is at most BDRV_REQUEST_MAX_SECTORS, this check + * make us sure that "num_sectors << BDRV_SECTOR_BITS" can fit in + * the integer variable. + */ + if (unlikely(num_sectors > s->conf.dwz_max_sectors)) { + err_status = VIRTIO_BLK_S_IOERR; + goto err; + } + + bytes = num_sectors << BDRV_SECTOR_BITS; + + if (unlikely(!virtio_blk_sect_range_ok(req->dev, sector, bytes))) { + err_status = VIRTIO_BLK_S_IOERR; + goto err; + } + + /* + * The device MUST set the status byte to VIRTIO_BLK_S_UNSUPP for discard + * and write zeroes commands if any unknown flag is set. + */ + if (unlikely(flags & ~VIRTIO_BLK_WRITE_ZEROES_FLAG_UNMAP)) { + err_status = VIRTIO_BLK_S_UNSUPP; + goto err; + } + + if (is_wzeroes) { /* VIRTIO_BLK_T_WRITE_ZEROES */ + int blk_aio_flags = 0; + + if (s->conf.wz_may_unmap && + flags & VIRTIO_BLK_WRITE_ZEROES_FLAG_UNMAP) { + blk_aio_flags |= BDRV_REQ_MAY_UNMAP; + } + + block_acct_start(blk_get_stats(req->dev->blk), &req->acct, bytes, + BLOCK_ACCT_WRITE); + + blk_aio_pwrite_zeroes(req->dev->blk, sector << BDRV_SECTOR_BITS, + bytes, blk_aio_flags, + virtio_blk_discard_wzeroes_complete, req); + } else { /* VIRTIO_BLK_T_DISCARD */ + /* + * The device MUST set the status byte to VIRTIO_BLK_S_UNSUPP for + * discard commands if the unmap flag is set. + */ + if (unlikely(flags & VIRTIO_BLK_WRITE_ZEROES_FLAG_UNMAP)) { + err_status = VIRTIO_BLK_S_UNSUPP; + goto err; + } + + blk_aio_pdiscard(req->dev->blk, sector << BDRV_SECTOR_BITS, bytes, + virtio_blk_discard_wzeroes_complete, req); + } + + return VIRTIO_BLK_S_OK; + +err: + if (is_wzeroes) { + block_acct_invalid(blk_get_stats(req->dev->blk), BLOCK_ACCT_WRITE); + } + return err_status; +} + static int virtio_blk_handle_request(VirtIOBlockReq *req, MultiReqBuffer *mrb) { uint32_t type; @@ -586,6 +686,45 @@ static int virtio_blk_handle_request(VirtIOBlockReq *req, MultiReqBuffer *mrb) virtio_blk_free_request(req); break; } + /* + * VIRTIO_BLK_T_DISCARD and VIRTIO_BLK_T_WRITE_ZEROES are defined with + * VIRTIO_BLK_T_OUT flag set. We masked this flag in the switch statement, + * so we must mask it for these requests, then we will check if it is set. + */ + case VIRTIO_BLK_T_DISCARD & ~VIRTIO_BLK_T_OUT: + case VIRTIO_BLK_T_WRITE_ZEROES & ~VIRTIO_BLK_T_OUT: + { + struct virtio_blk_discard_write_zeroes dwz_hdr; + size_t out_len = iov_size(out_iov, out_num); + bool is_wzeroes = (type & ~VIRTIO_BLK_T_BARRIER) == + VIRTIO_BLK_T_WRITE_ZEROES; + uint8_t err_status; + + /* + * Unsupported if VIRTIO_BLK_T_OUT is not set or the request contains + * more than one segment. + */ + if (unlikely(!(type & VIRTIO_BLK_T_OUT) || + out_len > sizeof(dwz_hdr))) { + virtio_blk_req_complete(req, VIRTIO_BLK_S_UNSUPP); + virtio_blk_free_request(req); + return 0; + } + + if (unlikely(iov_to_buf(out_iov, out_num, 0, &dwz_hdr, + sizeof(dwz_hdr)) != sizeof(dwz_hdr))) { + virtio_error(vdev, "virtio-blk discard/wzeroes header too short"); + return -1; + } + + err_status = virtio_blk_handle_dwz(req, is_wzeroes, &dwz_hdr); + if (err_status != VIRTIO_BLK_S_OK) { + virtio_blk_req_complete(req, err_status); + virtio_blk_free_request(req); + } + + break; + } default: virtio_blk_req_complete(req, VIRTIO_BLK_S_UNSUPP); virtio_blk_free_request(req); @@ -765,6 +904,22 @@ static void virtio_blk_update_config(VirtIODevice *vdev, uint8_t *config) blkcfg.alignment_offset = 0; blkcfg.wce = blk_enable_write_cache(s->blk); virtio_stw_p(vdev, &blkcfg.num_queues, s->conf.num_queues); + if (s->conf.discard_wzeroes) { + virtio_stl_p(vdev, &blkcfg.max_discard_sectors, + s->conf.dwz_max_sectors); + virtio_stl_p(vdev, &blkcfg.discard_sector_alignment, + blk_size >> BDRV_SECTOR_BITS); + virtio_stl_p(vdev, &blkcfg.max_write_zeroes_sectors, + s->conf.dwz_max_sectors); + blkcfg.write_zeroes_may_unmap = s->conf.wz_may_unmap; + /* + * We support only one segment per request since multiple segments + * are not widely used and there are no userspace APIs that allow + * applications to submit multiple segments in a single call. + */ + virtio_stl_p(vdev, &blkcfg.max_discard_seg, 1); + virtio_stl_p(vdev, &blkcfg.max_write_zeroes_seg, 1); + } memcpy(config, &blkcfg, sizeof(struct virtio_blk_config)); } @@ -811,6 +966,10 @@ static uint64_t virtio_blk_get_features(VirtIODevice *vdev, uint64_t features, if (s->conf.num_queues > 1) { virtio_add_feature(&features, VIRTIO_BLK_F_MQ); } + if (s->conf.discard_wzeroes) { + virtio_add_feature(&features, VIRTIO_BLK_F_DISCARD); + virtio_add_feature(&features, VIRTIO_BLK_F_WRITE_ZEROES); + } return features; } @@ -956,6 +1115,16 @@ static void virtio_blk_device_realize(DeviceState *dev, Error **errp) return; } + if (conf->discard_wzeroes) { + if (!conf->dwz_max_sectors || + conf->dwz_max_sectors > BDRV_REQUEST_MAX_SECTORS) { + error_setg(errp, "invalid dwz-max-sectors property (%" PRIu32 "), " + "must be between 1 and %lu", + conf->dwz_max_sectors, BDRV_REQUEST_MAX_SECTORS); + return; + } + } + virtio_init(vdev, "virtio-blk", VIRTIO_ID_BLOCK, sizeof(struct virtio_blk_config)); @@ -1028,6 +1197,10 @@ static Property virtio_blk_properties[] = { IOThread *), DEFINE_PROP_BIT("discard-wzeroes", VirtIOBlock, conf.discard_wzeroes, 0, true), + DEFINE_PROP_UINT32("discard-wzeroes-max-sectors", VirtIOBlock, + conf.dwz_max_sectors, BDRV_REQUEST_MAX_SECTORS), + DEFINE_PROP_BIT("wzeroes-may-unmap", VirtIOBlock, conf.wz_may_unmap, 0, + true), DEFINE_PROP_END_OF_LIST(), }; diff --git a/include/hw/virtio/virtio-blk.h b/include/hw/virtio/virtio-blk.h index c336afb4cd..4e9d4434ff 100644 --- a/include/hw/virtio/virtio-blk.h +++ b/include/hw/virtio/virtio-blk.h @@ -41,6 +41,8 @@ struct VirtIOBlkConf uint16_t num_queues; uint16_t queue_size; uint32_t discard_wzeroes; + uint32_t dwz_max_sectors; + uint32_t wz_may_unmap; }; struct VirtIOBlockDataPlane;