From patchwork Fri Feb 6 17:37:52 2015 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: "Denis V. Lunev" X-Patchwork-Id: 437430 Return-Path: X-Original-To: incoming@patchwork.ozlabs.org Delivered-To: patchwork-incoming@bilbo.ozlabs.org Received: from lists.gnu.org (lists.gnu.org [IPv6:2001:4830:134:3::11]) (using TLSv1 with cipher AES256-SHA (256/256 bits)) (No client certificate requested) by ozlabs.org (Postfix) with ESMTPS id A47A514012E for ; Sat, 7 Feb 2015 04:41:33 +1100 (AEDT) Received: from localhost ([::1]:49915 helo=lists.gnu.org) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1YJmuJ-0008O0-UI for incoming@patchwork.ozlabs.org; Fri, 06 Feb 2015 12:41:31 -0500 Received: from eggs.gnu.org ([2001:4830:134:3::10]:32978) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1YJmqL-0001og-Dj for qemu-devel@nongnu.org; Fri, 06 Feb 2015 12:37:26 -0500 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1YJmqG-0006IA-Fe for qemu-devel@nongnu.org; Fri, 06 Feb 2015 12:37:25 -0500 Received: from mailhub.sw.ru ([195.214.232.25]:37864 helo=relay.sw.ru) by eggs.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1YJmqG-0006I3-34 for qemu-devel@nongnu.org; Fri, 06 Feb 2015 12:37:20 -0500 Received: from hades.sw.ru ([10.30.8.132]) by relay.sw.ru (8.13.4/8.13.4) with ESMTP id t16HbD3m004098; Fri, 6 Feb 2015 20:37:18 +0300 (MSK) From: "Denis V. Lunev" To: Date: Fri, 6 Feb 2015 20:37:52 +0300 Message-Id: <1423244272-24887-3-git-send-email-den@openvz.org> X-Mailer: git-send-email 1.9.1 In-Reply-To: <1423244272-24887-1-git-send-email-den@openvz.org> References: <1423244272-24887-1-git-send-email-den@openvz.org> X-detected-operating-system: by eggs.gnu.org: OpenBSD 3.x X-Received-From: 195.214.232.25 Cc: Kevin Wolf , "Denis V. Lunev" , qemu-devel@nongnu.org, Paolo Bonzini Subject: [Qemu-devel] [PATCH 2/2] block: align bounce buffers to page X-BeenThere: qemu-devel@nongnu.org X-Mailman-Version: 2.1.14 Precedence: list List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: qemu-devel-bounces+incoming=patchwork.ozlabs.org@nongnu.org Sender: qemu-devel-bounces+incoming=patchwork.ozlabs.org@nongnu.org The following sequence int fd = open(argv[1], O_RDWR | O_CREAT | O_DIRECT, 0644); for (i = 0; i < 100000; i++) write(fd, buf, 4096); performs 5% better if buf is aligned to 4096 bytes. The difference is quite reliable. On the other hand we do not want at the moment to enforce bounce buffering if guest request is aligned to 512 bytes. The patch introduces new concept: minimal memory alignment for bounce buffers. Original so called "optimal" value is actually minimal required value for aligment. Optimal should be set to page size by default. There is no driver which should change this default at the moment. Signed-off-by: Denis V. Lunev CC: Paolo Bonzini CC: Kevin Wolf --- block.c | 20 ++++++++++++++++++-- block/raw-posix.c | 2 +- include/block/block.h | 2 ++ include/block/block_int.h | 3 +++ 4 files changed, 24 insertions(+), 3 deletions(-) diff --git a/block.c b/block.c index e98d651..569a2d8 100644 --- a/block.c +++ b/block.c @@ -232,6 +232,16 @@ size_t bdrv_opt_mem_align(BlockDriverState *bs) return bs->bl.opt_mem_alignment; } +size_t bdrv_min_mem_align(BlockDriverState *bs) +{ + if (!bs || !bs->drv) { + /* page size should be on the safe side */ + return getpagesize(); + } + + return bs->bl.min_mem_alignment; +} + /* check if the path starts with ":" */ int path_has_protocol(const char *path) { @@ -541,9 +551,11 @@ void bdrv_refresh_limits(BlockDriverState *bs, Error **errp) } bs->bl.opt_transfer_length = bs->file->bl.opt_transfer_length; bs->bl.max_transfer_length = bs->file->bl.max_transfer_length; + bs->bl.min_mem_alignment = bs->file->bl.min_mem_alignment; bs->bl.opt_mem_alignment = bs->file->bl.opt_mem_alignment; } else { - bs->bl.opt_mem_alignment = BDRV_SECTOR_SIZE; + bs->bl.min_mem_alignment = BDRV_SECTOR_SIZE; + bs->bl.opt_mem_alignment = getpagesize(); } if (bs->backing_hd) { @@ -558,6 +570,9 @@ void bdrv_refresh_limits(BlockDriverState *bs, Error **errp) bs->bl.max_transfer_length = MIN_NON_ZERO(bs->bl.max_transfer_length, bs->backing_hd->bl.max_transfer_length); + bs->bl.min_mem_alignment = + MAX(bs->bl.min_mem_alignment, + bs->backing_hd->bl.min_mem_alignment); bs->bl.opt_mem_alignment = MAX(bs->bl.opt_mem_alignment, bs->backing_hd->bl.opt_mem_alignment); @@ -1044,6 +1059,7 @@ static int bdrv_open_common(BlockDriverState *bs, BlockDriverState *file, } assert(bdrv_opt_mem_align(bs) != 0); + assert(bdrv_min_mem_align(bs) != 0); assert((bs->request_alignment != 0) || bs->sg); return 0; @@ -5331,7 +5347,7 @@ void *qemu_try_blockalign0(BlockDriverState *bs, size_t size) bool bdrv_qiov_is_aligned(BlockDriverState *bs, QEMUIOVector *qiov) { int i; - size_t alignment = bdrv_opt_mem_align(bs); + size_t alignment = bdrv_min_mem_align(bs); for (i = 0; i < qiov->niov; i++) { if ((uintptr_t) qiov->iov[i].iov_base % alignment) { diff --git a/block/raw-posix.c b/block/raw-posix.c index 9848f83..1205e9d 100644 --- a/block/raw-posix.c +++ b/block/raw-posix.c @@ -650,7 +650,7 @@ static void raw_refresh_limits(BlockDriverState *bs, Error **errp) BDRVRawState *s = bs->opaque; raw_probe_alignment(bs, s->fd, errp); - bs->bl.opt_mem_alignment = s->buf_align; + bs->bl.min_mem_alignment = s->buf_align; } static ssize_t handle_aiocb_ioctl(RawPosixAIOData *aiocb) diff --git a/include/block/block.h b/include/block/block.h index 3082d2b..b9b24b5 100644 --- a/include/block/block.h +++ b/include/block/block.h @@ -424,6 +424,8 @@ void bdrv_img_create(const char *filename, const char *fmt, /* Returns the alignment in bytes that is required so that no bounce buffer * is required throughout the stack */ +size_t bdrv_min_mem_align(BlockDriverState *bs); +/* Returns optimal alignment in bytes for bounce buffer */ size_t bdrv_opt_mem_align(BlockDriverState *bs); void bdrv_set_guest_block_size(BlockDriverState *bs, int align); void *qemu_blockalign(BlockDriverState *bs, size_t size); diff --git a/include/block/block_int.h b/include/block/block_int.h index e264be9..98b183c 100644 --- a/include/block/block_int.h +++ b/include/block/block_int.h @@ -296,6 +296,9 @@ typedef struct BlockLimits { int max_transfer_length; /* memory alignment so that no bounce buffer is needed */ + size_t min_mem_alignment; + + /* memory alignment for bounce buffer */ size_t opt_mem_alignment; } BlockLimits;