From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([2001:4830:134:3::10]:47061) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1XW21P-0004na-P9 for qemu-devel@nongnu.org; Mon, 22 Sep 2014 07:43:18 -0400 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1XW21J-0007Yy-KK for qemu-devel@nongnu.org; Mon, 22 Sep 2014 07:43:11 -0400 Received: from mx1.redhat.com ([209.132.183.28]:56964) by eggs.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1XW21J-0007WZ-Cg for qemu-devel@nongnu.org; Mon, 22 Sep 2014 07:43:05 -0400 From: Stefan Hajnoczi Date: Mon, 22 Sep 2014 12:41:41 +0100 Message-Id: <1411386150-24003-11-git-send-email-stefanha@redhat.com> In-Reply-To: <1411386150-24003-1-git-send-email-stefanha@redhat.com> References: <1411386150-24003-1-git-send-email-stefanha@redhat.com> Subject: [Qemu-devel] [PULL v2 10/59] linux-aio: Convert laio_aiocb_info.cancel to .cancel_async List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: qemu-devel@nongnu.org Cc: Peter Maydell , Fam Zheng , Stefan Hajnoczi From: Fam Zheng Just call io_cancel (2), if it fails, it means the request is not canceled, so the event loop will eventually call qemu_laio_process_completion. In qemu_laio_process_completion, change to call the cb unconditionally. It is required by bdrv_aio_cancel_async. Signed-off-by: Fam Zheng Signed-off-by: Stefan Hajnoczi --- block/linux-aio.c | 30 ++++++++---------------------- 1 file changed, 8 insertions(+), 22 deletions(-) diff --git a/block/linux-aio.c b/block/linux-aio.c index 9aca758..b67f56c 100644 --- a/block/linux-aio.c +++ b/block/linux-aio.c @@ -85,9 +85,8 @@ static void qemu_laio_process_completion(struct qemu_laio_state *s, ret = -EINVAL; } } - - laiocb->common.cb(laiocb->common.opaque, ret); } + laiocb->common.cb(laiocb->common.opaque, ret); qemu_aio_release(laiocb); } @@ -153,35 +152,22 @@ static void laio_cancel(BlockDriverAIOCB *blockacb) struct io_event event; int ret; - if (laiocb->ret != -EINPROGRESS) + if (laiocb->ret != -EINPROGRESS) { return; - - /* - * Note that as of Linux 2.6.31 neither the block device code nor any - * filesystem implements cancellation of AIO request. - * Thus the polling loop below is the normal code path. - */ + } ret = io_cancel(laiocb->ctx->ctx, &laiocb->iocb, &event); - if (ret == 0) { - laiocb->ret = -ECANCELED; + laiocb->ret = -ECANCELED; + if (ret != 0) { + /* iocb is not cancelled, cb will be called by the event loop later */ return; } - /* - * We have to wait for the iocb to finish. - * - * The only way to get the iocb status update is by polling the io context. - * We might be able to do this slightly more optimal by removing the - * O_NONBLOCK flag. - */ - while (laiocb->ret == -EINPROGRESS) { - qemu_laio_completion_cb(&laiocb->ctx->e); - } + laiocb->common.cb(laiocb->common.opaque, laiocb->ret); } static const AIOCBInfo laio_aiocb_info = { .aiocb_size = sizeof(struct qemu_laiocb), - .cancel = laio_cancel, + .cancel_async = laio_cancel, }; static void ioq_init(LaioQueue *io_q) -- 1.9.3