From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([2001:4830:134:3::10]:47305) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1VYSc4-0005mD-GR for qemu-devel@nongnu.org; Mon, 21 Oct 2013 23:26:41 -0400 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1VYSbv-0004Zw-IP for qemu-devel@nongnu.org; Mon, 21 Oct 2013 23:26:32 -0400 Received: from e28smtp07.in.ibm.com ([122.248.162.7]:51661) by eggs.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1VYSbu-0004Zk-Ug for qemu-devel@nongnu.org; Mon, 21 Oct 2013 23:26:23 -0400 Received: from /spool/local by e28smtp07.in.ibm.com with IBM ESMTP SMTP Gateway: Authorized Use Only! Violators will be prosecuted for from ; Tue, 22 Oct 2013 08:56:18 +0530 Received: from d28relay03.in.ibm.com (d28relay03.in.ibm.com [9.184.220.60]) by d28dlp02.in.ibm.com (Postfix) with ESMTP id BA064394004D for ; Tue, 22 Oct 2013 08:55:56 +0530 (IST) Received: from d28av04.in.ibm.com (d28av04.in.ibm.com [9.184.220.66]) by d28relay03.in.ibm.com (8.13.8/8.13.8/NCO v10.0) with ESMTP id r9M3T6c343974682 for ; Tue, 22 Oct 2013 08:59:06 +0530 Received: from d28av04.in.ibm.com (localhost [127.0.0.1]) by d28av04.in.ibm.com (8.14.4/8.14.4/NCO v10.0 AVout) with ESMTP id r9M3QGWF023725 for ; Tue, 22 Oct 2013 08:56:16 +0530 From: Lei Li Date: Tue, 22 Oct 2013 11:25:36 +0800 Message-Id: <1382412341-1173-13-git-send-email-lilei@linux.vnet.ibm.com> In-Reply-To: <1382412341-1173-1-git-send-email-lilei@linux.vnet.ibm.com> References: <1382412341-1173-1-git-send-email-lilei@linux.vnet.ibm.com> Subject: [Qemu-devel] [PATCH 12/17] migration-local: override hook_ram_load List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: qemu-devel@nongnu.org Cc: aarcange@redhat.com, aliguori@us.ibm.com, Lei Li , quintela@redhat.com, mdroth@linux.vnet.ibm.com, mrhines@linux.vnet.ibm.com, lagarcia@br.ibm.com, pbonzini@redhat.com, rcj@linux.vnet.ibm.com Override hook_ram_load to receive the pipe file descriptor passed by source process and page address which will be extracted to vmsplice the page data from pipe. Signed-off-by: Lei Li --- migration-local.c | 85 +++++++++++++++++++++++++++++++++++++++++++++++++++++ 1 files changed, 85 insertions(+), 0 deletions(-) diff --git a/migration-local.c b/migration-local.c index 7ccec30..083a24f 100644 --- a/migration-local.c +++ b/migration-local.c @@ -32,6 +32,8 @@ #include "trace.h" #include "qemu/osdep.h" +#define BEFORE_PIPE_FD 199 + //#define DEBUG_MIGRATION_LOCAL #ifdef DEBUG_MIGRATION_LOCAL @@ -50,6 +52,8 @@ typedef struct QEMUFileLocal { bool unix_page_flipping; } QEMUFileLocal; +static bool pipefd_passed; + static int qemu_local_get_sockfd(void *opaque) { QEMUFileLocal *s = opaque; @@ -64,6 +68,22 @@ static int qemu_local_get_buffer(void *opaque, uint8_t *buf, ssize_t len; for (;;) { + /* + * FIX ME: BEFORE_PIPE_FD is hard-coded and checked temporarily here + * because that the control message of passed pipe file descriptor + * might be 'eaten' to stream file by qemu_recv(), which would lead + * to the failure of recv_pipefd(), as it should stay in the socket + * and received by the real receiver recvmsg(). + * + * Although this message is followed by the first load_hook flags + * RAM_SAVE_FLAG_HOOK, the incoming side is hardly to avoid this + * as it would fill it into the stream file before any check action + * taken. Need to find a way out to fix this. + */ + if (size > BEFORE_PIPE_FD && !pipefd_passed) { + size = BEFORE_PIPE_FD; + } + len = qemu_recv(s->sockfd, buf, size, 0); if (len != -1) { break; @@ -115,6 +135,7 @@ static int qemu_local_close(void *opaque) } static int send_pipefd(int sockfd, int pipefd); +static int recv_pipefd(int sockfd); static int qemu_local_send_pipefd(QEMUFile *f, void *opaque, uint64_t flags) @@ -192,10 +213,74 @@ static size_t qemu_local_save_ram(QEMUFile *f, void *opaque, return RAM_SAVE_CONTROL_NOT_SUPP; } +static int qemu_local_ram_load(QEMUFile *f, void *opaque, + uint64_t flags) +{ + QEMUFileLocal *s = opaque; + ram_addr_t addr; + struct iovec iov; + ssize_t ret; + + /* Receive the pipe file descripter passed from source process */ + if (!pipefd_passed) { + s->pipefd[0] = recv_pipefd(s->sockfd); + if (s->pipefd[0] <= 0) { + fprintf(stderr, "failed to receive pipe fd: %d\n", s->pipefd[0]); + } else { + pipefd_passed = 1; + DPRINTF(stderr, "succeed\n"); + } + + return s->pipefd[0]; + } + + if (pipefd_passed) { + void *host; + + /* + * Extract the page address from the 8-byte record and + * read the page data from the pipe. + */ + addr = qemu_get_be64(s->file); + host = qemu_get_ram_ptr(addr); + + iov.iov_base = host; + iov.iov_len = TARGET_PAGE_SIZE; + + /* The flag SPLICE_F_MOVE is introduced in kernel for the page + * flipping feature in QEMU, which will movie pages rather than + * copying, previously unused. + * + * If a move is not possible the kernel will transparently falls + * back to copying data. + * + * For older kernels the SPLICE_F_MOVE would be ignored and a copy + * would occur. + */ + ret = vmsplice(s->pipefd[0], &iov, 1, SPLICE_F_MOVE); + if (ret == -1) { + if (errno != EAGAIN && errno != EINTR) { + fprintf(stderr, "vmsplice() load error: %s", strerror(errno)); + return ret; + } + DPRINTF("vmsplice load error\n"); + } else if (ret == 0) { + DPRINTF(stderr, "load_page: zero read\n"); + } + + DPRINTF("vmsplice (read): %zu\n", ret); + return ret; + } + + return 0; +} + + static const QEMUFileOps pipe_read_ops = { .get_fd = qemu_local_get_sockfd, .get_buffer = qemu_local_get_buffer, .close = qemu_local_close, + .hook_ram_load = qemu_local_ram_load }; static const QEMUFileOps pipe_write_ops = { -- 1.7.7.6