From 2302c1cafb13df23938b098d9c6595de52ec2577 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Marc-Andr=C3=A9=20Lureau?= Date: Sun, 1 Dec 2013 22:23:41 +0100 Subject: Split nbd block client code MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Signed-off-by: Marc-André Lureau Acked-by: Paolo Bonzini Signed-off-by: Gerd Hoffmann --- block/Makefile.objs | 2 +- block/nbd-client.c | 372 +++++++++++++++++++++++++++++++++++++++++++++++++++ block/nbd-client.h | 51 +++++++ block/nbd.c | 373 ++++------------------------------------------------ 4 files changed, 452 insertions(+), 346 deletions(-) create mode 100644 block/nbd-client.c create mode 100644 block/nbd-client.h (limited to 'block') diff --git a/block/Makefile.objs b/block/Makefile.objs index f43ecbc044..4e8c91ec34 100644 --- a/block/Makefile.objs +++ b/block/Makefile.objs @@ -10,7 +10,7 @@ block-obj-$(CONFIG_POSIX) += raw-posix.o block-obj-$(CONFIG_LINUX_AIO) += linux-aio.o ifeq ($(CONFIG_POSIX),y) -block-obj-y += nbd.o sheepdog.o +block-obj-y += nbd.o nbd-client.o sheepdog.o block-obj-$(CONFIG_LIBISCSI) += iscsi.o block-obj-$(CONFIG_CURL) += curl.o block-obj-$(CONFIG_RBD) += rbd.o diff --git a/block/nbd-client.c b/block/nbd-client.c new file mode 100644 index 0000000000..1abfc6a7f5 --- /dev/null +++ b/block/nbd-client.c @@ -0,0 +1,372 @@ +/* + * QEMU Block driver for NBD + * + * Copyright (C) 2008 Bull S.A.S. + * Author: Laurent Vivier + * + * Some parts: + * Copyright (C) 2007 Anthony Liguori + * + * Permission is hereby granted, free of charge, to any person obtaining a copy + * of this software and associated documentation files (the "Software"), to deal + * in the Software without restriction, including without limitation the rights + * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell + * copies of the Software, and to permit persons to whom the Software is + * furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER + * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, + * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN + * THE SOFTWARE. + */ + +#include "nbd-client.h" +#include "qemu/sockets.h" + +#define HANDLE_TO_INDEX(bs, handle) ((handle) ^ ((uint64_t)(intptr_t)bs)) +#define INDEX_TO_HANDLE(bs, index) ((index) ^ ((uint64_t)(intptr_t)bs)) + +static void nbd_reply_ready(void *opaque) +{ + NbdClientSession *s = opaque; + uint64_t i; + int ret; + + if (s->reply.handle == 0) { + /* No reply already in flight. Fetch a header. It is possible + * that another thread has done the same thing in parallel, so + * the socket is not readable anymore. + */ + ret = nbd_receive_reply(s->sock, &s->reply); + if (ret == -EAGAIN) { + return; + } + if (ret < 0) { + s->reply.handle = 0; + goto fail; + } + } + + /* There's no need for a mutex on the receive side, because the + * handler acts as a synchronization point and ensures that only + * one coroutine is called until the reply finishes. */ + i = HANDLE_TO_INDEX(s, s->reply.handle); + if (i >= MAX_NBD_REQUESTS) { + goto fail; + } + + if (s->recv_coroutine[i]) { + qemu_coroutine_enter(s->recv_coroutine[i], NULL); + return; + } + +fail: + for (i = 0; i < MAX_NBD_REQUESTS; i++) { + if (s->recv_coroutine[i]) { + qemu_coroutine_enter(s->recv_coroutine[i], NULL); + } + } +} + +static void nbd_restart_write(void *opaque) +{ + NbdClientSession *s = opaque; + + qemu_coroutine_enter(s->send_coroutine, NULL); +} + +static int nbd_co_send_request(NbdClientSession *s, + struct nbd_request *request, + QEMUIOVector *qiov, int offset) +{ + int rc, ret; + + qemu_co_mutex_lock(&s->send_mutex); + s->send_coroutine = qemu_coroutine_self(); + qemu_aio_set_fd_handler(s->sock, nbd_reply_ready, nbd_restart_write, s); + if (qiov) { + if (!s->is_unix) { + socket_set_cork(s->sock, 1); + } + rc = nbd_send_request(s->sock, request); + if (rc >= 0) { + ret = qemu_co_sendv(s->sock, qiov->iov, qiov->niov, + offset, request->len); + if (ret != request->len) { + rc = -EIO; + } + } + if (!s->is_unix) { + socket_set_cork(s->sock, 0); + } + } else { + rc = nbd_send_request(s->sock, request); + } + qemu_aio_set_fd_handler(s->sock, nbd_reply_ready, NULL, s); + s->send_coroutine = NULL; + qemu_co_mutex_unlock(&s->send_mutex); + return rc; +} + +static void nbd_co_receive_reply(NbdClientSession *s, + struct nbd_request *request, struct nbd_reply *reply, + QEMUIOVector *qiov, int offset) +{ + int ret; + + /* Wait until we're woken up by the read handler. TODO: perhaps + * peek at the next reply and avoid yielding if it's ours? */ + qemu_coroutine_yield(); + *reply = s->reply; + if (reply->handle != request->handle) { + reply->error = EIO; + } else { + if (qiov && reply->error == 0) { + ret = qemu_co_recvv(s->sock, qiov->iov, qiov->niov, + offset, request->len); + if (ret != request->len) { + reply->error = EIO; + } + } + + /* Tell the read handler to read another header. */ + s->reply.handle = 0; + } +} + +static void nbd_coroutine_start(NbdClientSession *s, + struct nbd_request *request) +{ + int i; + + /* Poor man semaphore. The free_sema is locked when no other request + * can be accepted, and unlocked after receiving one reply. */ + if (s->in_flight >= MAX_NBD_REQUESTS - 1) { + qemu_co_mutex_lock(&s->free_sema); + assert(s->in_flight < MAX_NBD_REQUESTS); + } + s->in_flight++; + + for (i = 0; i < MAX_NBD_REQUESTS; i++) { + if (s->recv_coroutine[i] == NULL) { + s->recv_coroutine[i] = qemu_coroutine_self(); + break; + } + } + + assert(i < MAX_NBD_REQUESTS); + request->handle = INDEX_TO_HANDLE(s, i); +} + +static void nbd_coroutine_end(NbdClientSession *s, + struct nbd_request *request) +{ + int i = HANDLE_TO_INDEX(s, request->handle); + s->recv_coroutine[i] = NULL; + if (s->in_flight-- == MAX_NBD_REQUESTS) { + qemu_co_mutex_unlock(&s->free_sema); + } +} + +static int nbd_co_readv_1(NbdClientSession *client, int64_t sector_num, + int nb_sectors, QEMUIOVector *qiov, + int offset) +{ + struct nbd_request request; + struct nbd_reply reply; + ssize_t ret; + + request.type = NBD_CMD_READ; + request.from = sector_num * 512; + request.len = nb_sectors * 512; + + nbd_coroutine_start(client, &request); + ret = nbd_co_send_request(client, &request, NULL, 0); + if (ret < 0) { + reply.error = -ret; + } else { + nbd_co_receive_reply(client, &request, &reply, qiov, offset); + } + nbd_coroutine_end(client, &request); + return -reply.error; + +} + +static int nbd_co_writev_1(NbdClientSession *client, int64_t sector_num, + int nb_sectors, QEMUIOVector *qiov, + int offset) +{ + struct nbd_request request; + struct nbd_reply reply; + ssize_t ret; + + request.type = NBD_CMD_WRITE; + if (!bdrv_enable_write_cache(client->bs) && + (client->nbdflags & NBD_FLAG_SEND_FUA)) { + request.type |= NBD_CMD_FLAG_FUA; + } + + request.from = sector_num * 512; + request.len = nb_sectors * 512; + + nbd_coroutine_start(client, &request); + ret = nbd_co_send_request(client, &request, qiov, offset); + if (ret < 0) { + reply.error = -ret; + } else { + nbd_co_receive_reply(client, &request, &reply, NULL, 0); + } + nbd_coroutine_end(client, &request); + return -reply.error; +} + +/* qemu-nbd has a limit of slightly less than 1M per request. Try to + * remain aligned to 4K. */ +#define NBD_MAX_SECTORS 2040 + +int nbd_client_session_co_readv(NbdClientSession *client, int64_t sector_num, + int nb_sectors, QEMUIOVector *qiov) +{ + int offset = 0; + int ret; + while (nb_sectors > NBD_MAX_SECTORS) { + ret = nbd_co_readv_1(client, sector_num, + NBD_MAX_SECTORS, qiov, offset); + if (ret < 0) { + return ret; + } + offset += NBD_MAX_SECTORS * 512; + sector_num += NBD_MAX_SECTORS; + nb_sectors -= NBD_MAX_SECTORS; + } + return nbd_co_readv_1(client, sector_num, nb_sectors, qiov, offset); +} + +int nbd_client_session_co_writev(NbdClientSession *client, int64_t sector_num, + int nb_sectors, QEMUIOVector *qiov) +{ + int offset = 0; + int ret; + while (nb_sectors > NBD_MAX_SECTORS) { + ret = nbd_co_writev_1(client, sector_num, + NBD_MAX_SECTORS, qiov, offset); + if (ret < 0) { + return ret; + } + offset += NBD_MAX_SECTORS * 512; + sector_num += NBD_MAX_SECTORS; + nb_sectors -= NBD_MAX_SECTORS; + } + return nbd_co_writev_1(client, sector_num, nb_sectors, qiov, offset); +} + +int nbd_client_session_co_flush(NbdClientSession *client) +{ + struct nbd_request request; + struct nbd_reply reply; + ssize_t ret; + + if (!(client->nbdflags & NBD_FLAG_SEND_FLUSH)) { + return 0; + } + + request.type = NBD_CMD_FLUSH; + if (client->nbdflags & NBD_FLAG_SEND_FUA) { + request.type |= NBD_CMD_FLAG_FUA; + } + + request.from = 0; + request.len = 0; + + nbd_coroutine_start(client, &request); + ret = nbd_co_send_request(client, &request, NULL, 0); + if (ret < 0) { + reply.error = -ret; + } else { + nbd_co_receive_reply(client, &request, &reply, NULL, 0); + } + nbd_coroutine_end(client, &request); + return -reply.error; +} + +int nbd_client_session_co_discard(NbdClientSession *client, int64_t sector_num, + int nb_sectors) +{ + struct nbd_request request; + struct nbd_reply reply; + ssize_t ret; + + if (!(client->nbdflags & NBD_FLAG_SEND_TRIM)) { + return 0; + } + request.type = NBD_CMD_TRIM; + request.from = sector_num * 512; + request.len = nb_sectors * 512; + + nbd_coroutine_start(client, &request); + ret = nbd_co_send_request(client, &request, NULL, 0); + if (ret < 0) { + reply.error = -ret; + } else { + nbd_co_receive_reply(client, &request, &reply, NULL, 0); + } + nbd_coroutine_end(client, &request); + return -reply.error; + +} + +static void nbd_teardown_connection(NbdClientSession *client) +{ + struct nbd_request request; + + request.type = NBD_CMD_DISC; + request.from = 0; + request.len = 0; + nbd_send_request(client->sock, &request); + + qemu_aio_set_fd_handler(client->sock, NULL, NULL, NULL); + closesocket(client->sock); + client->sock = -1; +} + +void nbd_client_session_close(NbdClientSession *client) +{ + nbd_teardown_connection(client); + g_free(client->export_name); + client->export_name = NULL; +} + +int nbd_client_session_init(NbdClientSession *client, + BlockDriverState *bs, int sock) +{ + int ret; + + /* NBD handshake */ + ret = nbd_receive_negotiate(sock, client->export_name, + &client->nbdflags, &client->size, + &client->blocksize); + if (ret < 0) { + logout("Failed to negotiate with the NBD server\n"); + closesocket(sock); + return ret; + } + + qemu_co_mutex_init(&client->send_mutex); + qemu_co_mutex_init(&client->free_sema); + client->bs = bs; + client->sock = sock; + + /* Now that we're connected, set the socket to be non-blocking and + * kick the reply mechanism. */ + qemu_set_nonblock(sock); + qemu_aio_set_fd_handler(sock, nbd_reply_ready, NULL, client); + + logout("Established connection with NBD server\n"); + return 0; +} diff --git a/block/nbd-client.h b/block/nbd-client.h new file mode 100644 index 0000000000..c271236911 --- /dev/null +++ b/block/nbd-client.h @@ -0,0 +1,51 @@ +#ifndef NBD_CLIENT_H +#define NBD_CLIENT_H + +#include "qemu-common.h" +#include "block/nbd.h" +#include "block/block_int.h" + +/* #define DEBUG_NBD */ + +#if defined(DEBUG_NBD) +#define logout(fmt, ...) \ + fprintf(stderr, "nbd\t%-24s" fmt, __func__, ##__VA_ARGS__) +#else +#define logout(fmt, ...) ((void)0) +#endif + +#define MAX_NBD_REQUESTS 16 + +typedef struct NbdClientSession { + int sock; + uint32_t nbdflags; + off_t size; + size_t blocksize; + + CoMutex send_mutex; + CoMutex free_sema; + Coroutine *send_coroutine; + int in_flight; + + Coroutine *recv_coroutine[MAX_NBD_REQUESTS]; + struct nbd_reply reply; + + char *export_name; /* An NBD server may export several devices */ + bool is_unix; + + BlockDriverState *bs; +} NbdClientSession; + +int nbd_client_session_init(NbdClientSession *client, + BlockDriverState *bs, int sock); +void nbd_client_session_close(NbdClientSession *client); + +int nbd_client_session_co_discard(NbdClientSession *client, int64_t sector_num, + int nb_sectors); +int nbd_client_session_co_flush(NbdClientSession *client); +int nbd_client_session_co_writev(NbdClientSession *client, int64_t sector_num, + int nb_sectors, QEMUIOVector *qiov); +int nbd_client_session_co_readv(NbdClientSession *client, int64_t sector_num, + int nb_sectors, QEMUIOVector *qiov); + +#endif /* NBD_CLIENT_H */ diff --git a/block/nbd.c b/block/nbd.c index c8deeee67f..be75ba074c 100644 --- a/block/nbd.c +++ b/block/nbd.c @@ -26,8 +26,7 @@ * THE SOFTWARE. */ -#include "qemu-common.h" -#include "block/nbd.h" +#include "block/nbd-client.h" #include "qemu/uri.h" #include "block/block_int.h" #include "qemu/module.h" @@ -40,37 +39,9 @@ #define EN_OPTSTR ":exportname=" -/* #define DEBUG_NBD */ - -#if defined(DEBUG_NBD) -#define logout(fmt, ...) \ - fprintf(stderr, "nbd\t%-24s" fmt, __func__, ##__VA_ARGS__) -#else -#define logout(fmt, ...) ((void)0) -#endif - -#define MAX_NBD_REQUESTS 16 -#define HANDLE_TO_INDEX(bs, handle) ((handle) ^ ((uint64_t)(intptr_t)bs)) -#define INDEX_TO_HANDLE(bs, index) ((index) ^ ((uint64_t)(intptr_t)bs)) - typedef struct BDRVNBDState { - int sock; - uint32_t nbdflags; - off_t size; - size_t blocksize; - - CoMutex send_mutex; - CoMutex free_sema; - Coroutine *send_coroutine; - int in_flight; - - Coroutine *recv_coroutine[MAX_NBD_REQUESTS]; - struct nbd_reply reply; - - bool is_unix; + NbdClientSession client; QemuOpts *socket_opts; - - char *export_name; /* An NBD server may export several devices */ } BDRVNBDState; static int nbd_parse_uri(const char *filename, QDict *options) @@ -227,9 +198,9 @@ static int nbd_config(BDRVNBDState *s, QDict *options) "be used at the same time."); return -EINVAL; } - s->is_unix = true; + s->client.is_unix = true; } else if (qdict_haskey(options, "host")) { - s->is_unix = false; + s->client.is_unix = false; } else { return -EINVAL; } @@ -247,162 +218,20 @@ static int nbd_config(BDRVNBDState *s, QDict *options) qemu_opt_set_number(s->socket_opts, "port", NBD_DEFAULT_PORT); } - s->export_name = g_strdup(qdict_get_try_str(options, "export")); - if (s->export_name) { + s->client.export_name = g_strdup(qdict_get_try_str(options, "export")); + if (s->client.export_name) { qdict_del(options, "export"); } return 0; } - -static void nbd_coroutine_start(BDRVNBDState *s, struct nbd_request *request) -{ - int i; - - /* Poor man semaphore. The free_sema is locked when no other request - * can be accepted, and unlocked after receiving one reply. */ - if (s->in_flight >= MAX_NBD_REQUESTS - 1) { - qemu_co_mutex_lock(&s->free_sema); - assert(s->in_flight < MAX_NBD_REQUESTS); - } - s->in_flight++; - - for (i = 0; i < MAX_NBD_REQUESTS; i++) { - if (s->recv_coroutine[i] == NULL) { - s->recv_coroutine[i] = qemu_coroutine_self(); - break; - } - } - - assert(i < MAX_NBD_REQUESTS); - request->handle = INDEX_TO_HANDLE(s, i); -} - -static void nbd_reply_ready(void *opaque) -{ - BDRVNBDState *s = opaque; - uint64_t i; - int ret; - - if (s->reply.handle == 0) { - /* No reply already in flight. Fetch a header. It is possible - * that another thread has done the same thing in parallel, so - * the socket is not readable anymore. - */ - ret = nbd_receive_reply(s->sock, &s->reply); - if (ret == -EAGAIN) { - return; - } - if (ret < 0) { - s->reply.handle = 0; - goto fail; - } - } - - /* There's no need for a mutex on the receive side, because the - * handler acts as a synchronization point and ensures that only - * one coroutine is called until the reply finishes. */ - i = HANDLE_TO_INDEX(s, s->reply.handle); - if (i >= MAX_NBD_REQUESTS) { - goto fail; - } - - if (s->recv_coroutine[i]) { - qemu_coroutine_enter(s->recv_coroutine[i], NULL); - return; - } - -fail: - for (i = 0; i < MAX_NBD_REQUESTS; i++) { - if (s->recv_coroutine[i]) { - qemu_coroutine_enter(s->recv_coroutine[i], NULL); - } - } -} - -static void nbd_restart_write(void *opaque) -{ - BDRVNBDState *s = opaque; - qemu_coroutine_enter(s->send_coroutine, NULL); -} - -static int nbd_co_send_request(BDRVNBDState *s, struct nbd_request *request, - QEMUIOVector *qiov, int offset) -{ - int rc, ret; - - qemu_co_mutex_lock(&s->send_mutex); - s->send_coroutine = qemu_coroutine_self(); - qemu_aio_set_fd_handler(s->sock, nbd_reply_ready, nbd_restart_write, s); - if (qiov) { - if (!s->is_unix) { - socket_set_cork(s->sock, 1); - } - rc = nbd_send_request(s->sock, request); - if (rc >= 0) { - ret = qemu_co_sendv(s->sock, qiov->iov, qiov->niov, - offset, request->len); - if (ret != request->len) { - rc = -EIO; - } - } - if (!s->is_unix) { - socket_set_cork(s->sock, 0); - } - } else { - rc = nbd_send_request(s->sock, request); - } - qemu_aio_set_fd_handler(s->sock, nbd_reply_ready, NULL, s); - s->send_coroutine = NULL; - qemu_co_mutex_unlock(&s->send_mutex); - return rc; -} - -static void nbd_co_receive_reply(BDRVNBDState *s, struct nbd_request *request, - struct nbd_reply *reply, - QEMUIOVector *qiov, int offset) -{ - int ret; - - /* Wait until we're woken up by the read handler. TODO: perhaps - * peek at the next reply and avoid yielding if it's ours? */ - qemu_coroutine_yield(); - *reply = s->reply; - if (reply->handle != request->handle) { - reply->error = EIO; - } else { - if (qiov && reply->error == 0) { - ret = qemu_co_recvv(s->sock, qiov->iov, qiov->niov, - offset, request->len); - if (ret != request->len) { - reply->error = EIO; - } - } - - /* Tell the read handler to read another header. */ - s->reply.handle = 0; - } -} - -static void nbd_coroutine_end(BDRVNBDState *s, struct nbd_request *request) -{ - int i = HANDLE_TO_INDEX(s, request->handle); - s->recv_coroutine[i] = NULL; - if (s->in_flight-- == MAX_NBD_REQUESTS) { - qemu_co_mutex_unlock(&s->free_sema); - } -} - static int nbd_establish_connection(BlockDriverState *bs) { BDRVNBDState *s = bs->opaque; int sock; - int ret; - off_t size; - size_t blocksize; - if (s->is_unix) { + if (s->client.is_unix) { sock = unix_socket_outgoing(qemu_opt_get(s->socket_opts, "path")); } else { sock = tcp_socket_outgoing_opts(s->socket_opts); @@ -417,50 +246,14 @@ static int nbd_establish_connection(BlockDriverState *bs) return -errno; } - /* NBD handshake */ - ret = nbd_receive_negotiate(sock, s->export_name, &s->nbdflags, &size, - &blocksize); - if (ret < 0) { - logout("Failed to negotiate with the NBD server\n"); - closesocket(sock); - return ret; - } - - /* Now that we're connected, set the socket to be non-blocking and - * kick the reply mechanism. */ - qemu_set_nonblock(sock); - qemu_aio_set_fd_handler(sock, nbd_reply_ready, NULL, s); - - s->sock = sock; - s->size = size; - s->blocksize = blocksize; - - logout("Established connection with NBD server\n"); - return 0; -} - -static void nbd_teardown_connection(BlockDriverState *bs) -{ - BDRVNBDState *s = bs->opaque; - struct nbd_request request; - - request.type = NBD_CMD_DISC; - request.from = 0; - request.len = 0; - nbd_send_request(s->sock, &request); - - qemu_aio_set_fd_handler(s->sock, NULL, NULL, NULL); - closesocket(s->sock); + return sock; } static int nbd_open(BlockDriverState *bs, QDict *options, int flags, Error **errp) { BDRVNBDState *s = bs->opaque; - int result; - - qemu_co_mutex_init(&s->send_mutex); - qemu_co_mutex_init(&s->free_sema); + int result, sock; /* Pop the config into our state object. Exit if invalid. */ result = nbd_config(s, options); @@ -471,172 +264,62 @@ static int nbd_open(BlockDriverState *bs, QDict *options, int flags, /* establish TCP connection, return error if it fails * TODO: Configurable retry-until-timeout behaviour. */ - result = nbd_establish_connection(bs); - - return result; -} - -static int nbd_co_readv_1(BlockDriverState *bs, int64_t sector_num, - int nb_sectors, QEMUIOVector *qiov, - int offset) -{ - BDRVNBDState *s = bs->opaque; - struct nbd_request request; - struct nbd_reply reply; - ssize_t ret; - - request.type = NBD_CMD_READ; - request.from = sector_num * 512; - request.len = nb_sectors * 512; - - nbd_coroutine_start(s, &request); - ret = nbd_co_send_request(s, &request, NULL, 0); - if (ret < 0) { - reply.error = -ret; - } else { - nbd_co_receive_reply(s, &request, &reply, qiov, offset); - } - nbd_coroutine_end(s, &request); - return -reply.error; - -} - -static int nbd_co_writev_1(BlockDriverState *bs, int64_t sector_num, - int nb_sectors, QEMUIOVector *qiov, - int offset) -{ - BDRVNBDState *s = bs->opaque; - struct nbd_request request; - struct nbd_reply reply; - ssize_t ret; - - request.type = NBD_CMD_WRITE; - if (!bdrv_enable_write_cache(bs) && (s->nbdflags & NBD_FLAG_SEND_FUA)) { - request.type |= NBD_CMD_FLAG_FUA; + sock = nbd_establish_connection(bs); + if (sock < 0) { + return sock; } - request.from = sector_num * 512; - request.len = nb_sectors * 512; - - nbd_coroutine_start(s, &request); - ret = nbd_co_send_request(s, &request, qiov, offset); - if (ret < 0) { - reply.error = -ret; - } else { - nbd_co_receive_reply(s, &request, &reply, NULL, 0); - } - nbd_coroutine_end(s, &request); - return -reply.error; + /* NBD handshake */ + return nbd_client_session_init(&s->client, bs, sock); } -/* qemu-nbd has a limit of slightly less than 1M per request. Try to - * remain aligned to 4K. */ -#define NBD_MAX_SECTORS 2040 - static int nbd_co_readv(BlockDriverState *bs, int64_t sector_num, int nb_sectors, QEMUIOVector *qiov) { - int offset = 0; - int ret; - while (nb_sectors > NBD_MAX_SECTORS) { - ret = nbd_co_readv_1(bs, sector_num, NBD_MAX_SECTORS, qiov, offset); - if (ret < 0) { - return ret; - } - offset += NBD_MAX_SECTORS * 512; - sector_num += NBD_MAX_SECTORS; - nb_sectors -= NBD_MAX_SECTORS; - } - return nbd_co_readv_1(bs, sector_num, nb_sectors, qiov, offset); + BDRVNBDState *s = bs->opaque; + + return nbd_client_session_co_readv(&s->client, sector_num, + nb_sectors, qiov); } static int nbd_co_writev(BlockDriverState *bs, int64_t sector_num, int nb_sectors, QEMUIOVector *qiov) { - int offset = 0; - int ret; - while (nb_sectors > NBD_MAX_SECTORS) { - ret = nbd_co_writev_1(bs, sector_num, NBD_MAX_SECTORS, qiov, offset); - if (ret < 0) { - return ret; - } - offset += NBD_MAX_SECTORS * 512; - sector_num += NBD_MAX_SECTORS; - nb_sectors -= NBD_MAX_SECTORS; - } - return nbd_co_writev_1(bs, sector_num, nb_sectors, qiov, offset); + BDRVNBDState *s = bs->opaque; + + return nbd_client_session_co_writev(&s->client, sector_num, + nb_sectors, qiov); } static int nbd_co_flush(BlockDriverState *bs) { BDRVNBDState *s = bs->opaque; - struct nbd_request request; - struct nbd_reply reply; - ssize_t ret; - - if (!(s->nbdflags & NBD_FLAG_SEND_FLUSH)) { - return 0; - } - request.type = NBD_CMD_FLUSH; - if (s->nbdflags & NBD_FLAG_SEND_FUA) { - request.type |= NBD_CMD_FLAG_FUA; - } - - request.from = 0; - request.len = 0; - - nbd_coroutine_start(s, &request); - ret = nbd_co_send_request(s, &request, NULL, 0); - if (ret < 0) { - reply.error = -ret; - } else { - nbd_co_receive_reply(s, &request, &reply, NULL, 0); - } - nbd_coroutine_end(s, &request); - return -reply.error; + return nbd_client_session_co_flush(&s->client); } static int nbd_co_discard(BlockDriverState *bs, int64_t sector_num, int nb_sectors) { BDRVNBDState *s = bs->opaque; - struct nbd_request request; - struct nbd_reply reply; - ssize_t ret; - if (!(s->nbdflags & NBD_FLAG_SEND_TRIM)) { - return 0; - } - request.type = NBD_CMD_TRIM; - request.from = sector_num * 512; - request.len = nb_sectors * 512; - - nbd_coroutine_start(s, &request); - ret = nbd_co_send_request(s, &request, NULL, 0); - if (ret < 0) { - reply.error = -ret; - } else { - nbd_co_receive_reply(s, &request, &reply, NULL, 0); - } - nbd_coroutine_end(s, &request); - return -reply.error; + return nbd_client_session_co_discard(&s->client, sector_num, + nb_sectors); } static void nbd_close(BlockDriverState *bs) { BDRVNBDState *s = bs->opaque; - g_free(s->export_name); - qemu_opts_del(s->socket_opts); - nbd_teardown_connection(bs); + qemu_opts_del(s->socket_opts); + nbd_client_session_close(&s->client); } static int64_t nbd_getlength(BlockDriverState *bs) { BDRVNBDState *s = bs->opaque; - return s->size; + return s->client.size; } static BlockDriver bdrv_nbd = { -- cgit v1.2.1 From e53a18e488c657bbc6f218ae60de8e813a912667 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Marc-Andr=C3=A9=20Lureau?= Date: Sun, 1 Dec 2013 22:23:42 +0100 Subject: nbd: don't change socket block during negotiate MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit The caller might handle non-blocking using coroutine. Leave the choice to the caller to use a blocking or non-blocking negotiate. Signed-off-by: Marc-André Lureau Acked-by: Paolo Bonzini Signed-off-by: Gerd Hoffmann --- block/nbd-client.c | 1 + 1 file changed, 1 insertion(+) (limited to 'block') diff --git a/block/nbd-client.c b/block/nbd-client.c index 1abfc6a7f5..693110d26d 100644 --- a/block/nbd-client.c +++ b/block/nbd-client.c @@ -348,6 +348,7 @@ int nbd_client_session_init(NbdClientSession *client, int ret; /* NBD handshake */ + qemu_set_block(sock); ret = nbd_receive_negotiate(sock, client->export_name, &client->nbdflags, &client->size, &client->blocksize); -- cgit v1.2.1 From e2bc625f9bbb3a5d3ef2cb0f14dd52b517b92ffd Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Marc-Andr=C3=A9=20Lureau?= Date: Sun, 1 Dec 2013 22:23:43 +0100 Subject: nbd: pass export name as init argument MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit There is no need to keep the export name around, and it seems a better fit as an argument in the init() call. Signed-off-by: Marc-André Lureau Acked-by: Paolo Bonzini Signed-off-by: Gerd Hoffmann --- block/nbd-client.c | 9 ++++----- block/nbd-client.h | 5 ++--- block/nbd.c | 13 ++++++++----- 3 files changed, 14 insertions(+), 13 deletions(-) (limited to 'block') diff --git a/block/nbd-client.c b/block/nbd-client.c index 693110d26d..3dfae1da09 100644 --- a/block/nbd-client.c +++ b/block/nbd-client.c @@ -338,18 +338,17 @@ static void nbd_teardown_connection(NbdClientSession *client) void nbd_client_session_close(NbdClientSession *client) { nbd_teardown_connection(client); - g_free(client->export_name); - client->export_name = NULL; } -int nbd_client_session_init(NbdClientSession *client, - BlockDriverState *bs, int sock) +int nbd_client_session_init(NbdClientSession *client, BlockDriverState *bs, + int sock, const char *export) { int ret; /* NBD handshake */ + logout("session init %s\n", export); qemu_set_block(sock); - ret = nbd_receive_negotiate(sock, client->export_name, + ret = nbd_receive_negotiate(sock, export, &client->nbdflags, &client->size, &client->blocksize); if (ret < 0) { diff --git a/block/nbd-client.h b/block/nbd-client.h index c271236911..f2a63378bb 100644 --- a/block/nbd-client.h +++ b/block/nbd-client.h @@ -30,14 +30,13 @@ typedef struct NbdClientSession { Coroutine *recv_coroutine[MAX_NBD_REQUESTS]; struct nbd_reply reply; - char *export_name; /* An NBD server may export several devices */ bool is_unix; BlockDriverState *bs; } NbdClientSession; -int nbd_client_session_init(NbdClientSession *client, - BlockDriverState *bs, int sock); +int nbd_client_session_init(NbdClientSession *client, BlockDriverState *bs, + int sock, const char *export_name); void nbd_client_session_close(NbdClientSession *client); int nbd_client_session_co_discard(NbdClientSession *client, int64_t sector_num, diff --git a/block/nbd.c b/block/nbd.c index be75ba074c..4455a134db 100644 --- a/block/nbd.c +++ b/block/nbd.c @@ -188,7 +188,7 @@ out: g_free(file); } -static int nbd_config(BDRVNBDState *s, QDict *options) +static int nbd_config(BDRVNBDState *s, QDict *options, char **export) { Error *local_err = NULL; @@ -218,8 +218,8 @@ static int nbd_config(BDRVNBDState *s, QDict *options) qemu_opt_set_number(s->socket_opts, "port", NBD_DEFAULT_PORT); } - s->client.export_name = g_strdup(qdict_get_try_str(options, "export")); - if (s->client.export_name) { + *export = g_strdup(qdict_get_try_str(options, "export")); + if (*export) { qdict_del(options, "export"); } @@ -253,10 +253,11 @@ static int nbd_open(BlockDriverState *bs, QDict *options, int flags, Error **errp) { BDRVNBDState *s = bs->opaque; + char *export = NULL; int result, sock; /* Pop the config into our state object. Exit if invalid. */ - result = nbd_config(s, options); + result = nbd_config(s, options, &export); if (result != 0) { return result; } @@ -270,7 +271,9 @@ static int nbd_open(BlockDriverState *bs, QDict *options, int flags, } /* NBD handshake */ - return nbd_client_session_init(&s->client, bs, sock); + result = nbd_client_session_init(&s->client, bs, sock, export); + g_free(export); + return result; } static int nbd_co_readv(BlockDriverState *bs, int64_t sector_num, -- cgit v1.2.1 From 5ad283ebb82c007f49535907874213cdb05c307a Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Marc-Andr=C3=A9=20Lureau?= Date: Sun, 1 Dec 2013 22:23:44 +0100 Subject: nbd: make nbd_client_session_close() idempotent MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Signed-off-by: Marc-André Lureau Acked-by: Paolo Bonzini Signed-off-by: Gerd Hoffmann --- block/nbd-client.c | 5 +++++ 1 file changed, 5 insertions(+) (limited to 'block') diff --git a/block/nbd-client.c b/block/nbd-client.c index 3dfae1da09..9156093dc0 100644 --- a/block/nbd-client.c +++ b/block/nbd-client.c @@ -337,7 +337,12 @@ static void nbd_teardown_connection(NbdClientSession *client) void nbd_client_session_close(NbdClientSession *client) { + if (!client->bs) { + return; + } + nbd_teardown_connection(client); + client->bs = NULL; } int nbd_client_session_init(NbdClientSession *client, BlockDriverState *bs, -- cgit v1.2.1 From 69152c09d337a8e1368e65bcd5a598def6c3079f Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Marc-Andr=C3=A9=20Lureau?= Date: Sun, 1 Dec 2013 22:23:45 +0100 Subject: nbd: finish any pending coroutine MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Make sure all pending coroutines are finished when closing the session. Signed-off-by: Marc-André Lureau Acked-by: Paolo Bonzini Signed-off-by: Gerd Hoffmann --- block/nbd-client.c | 21 ++++++++++++++++----- 1 file changed, 16 insertions(+), 5 deletions(-) (limited to 'block') diff --git a/block/nbd-client.c b/block/nbd-client.c index 9156093dc0..1973cf0dfe 100644 --- a/block/nbd-client.c +++ b/block/nbd-client.c @@ -32,6 +32,17 @@ #define HANDLE_TO_INDEX(bs, handle) ((handle) ^ ((uint64_t)(intptr_t)bs)) #define INDEX_TO_HANDLE(bs, index) ((index) ^ ((uint64_t)(intptr_t)bs)) +static void nbd_recv_coroutines_enter_all(NbdClientSession *s) +{ + int i; + + for (i = 0; i < MAX_NBD_REQUESTS; i++) { + if (s->recv_coroutine[i]) { + qemu_coroutine_enter(s->recv_coroutine[i], NULL); + } + } +} + static void nbd_reply_ready(void *opaque) { NbdClientSession *s = opaque; @@ -67,11 +78,7 @@ static void nbd_reply_ready(void *opaque) } fail: - for (i = 0; i < MAX_NBD_REQUESTS; i++) { - if (s->recv_coroutine[i]) { - qemu_coroutine_enter(s->recv_coroutine[i], NULL); - } - } + nbd_recv_coroutines_enter_all(s); } static void nbd_restart_write(void *opaque) @@ -330,6 +337,10 @@ static void nbd_teardown_connection(NbdClientSession *client) request.len = 0; nbd_send_request(client->sock, &request); + /* finish any pending coroutines */ + shutdown(client->sock, 2); + nbd_recv_coroutines_enter_all(client); + qemu_aio_set_fd_handler(client->sock, NULL, NULL, NULL); closesocket(client->sock); client->sock = -1; -- cgit v1.2.1 From b1b27b64262fdace45e5ab134c4438338076cb98 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Marc-Andr=C3=A9=20Lureau?= Date: Sun, 1 Dec 2013 22:23:46 +0100 Subject: nbd: avoid uninitialized warnings MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit ==15815== Thread 1: ==15815== Syscall param socketcall.sendto(msg) points to uninitialised byte(s) ==15815== at 0x65AD5CB: send (send.c:31) ==15815== by 0x37F84B: nbd_wr_sync (nbd.c:145) ==15815== by 0x37F94B: write_sync (nbd.c:186) ==15815== by 0x380FA9: nbd_send_request (nbd.c:681) ==15815== by 0x1C4A2D: nbd_teardown_connection (nbd-client.c:337) ==15815== by 0x1C4AD8: nbd_client_session_close (nbd-client.c:354) ==15815== by 0x1ED2D8: close_socketpair (spicebd.c:132) ==15815== by 0x1EE265: spice_close (spicebd.c:457) ==15815== by 0x1ACBF6: bdrv_close (block.c:1519) ==15815== by 0x1AD804: bdrv_delete (block.c:1772) ==15815== by 0x1B4136: bdrv_unref (block.c:4476) ==15815== by 0x1ACCE0: bdrv_close (block.c:1541) ==15815== Address 0x7feffef98 is on thread 1's stack Signed-off-by: Marc-André Lureau Acked-by: Paolo Bonzini Signed-off-by: Gerd Hoffmann --- block/nbd-client.c | 21 +++++++++------------ 1 file changed, 9 insertions(+), 12 deletions(-) (limited to 'block') diff --git a/block/nbd-client.c b/block/nbd-client.c index 1973cf0dfe..0922b78292 100644 --- a/block/nbd-client.c +++ b/block/nbd-client.c @@ -185,11 +185,10 @@ static int nbd_co_readv_1(NbdClientSession *client, int64_t sector_num, int nb_sectors, QEMUIOVector *qiov, int offset) { - struct nbd_request request; + struct nbd_request request = { .type = NBD_CMD_READ }; struct nbd_reply reply; ssize_t ret; - request.type = NBD_CMD_READ; request.from = sector_num * 512; request.len = nb_sectors * 512; @@ -209,11 +208,10 @@ static int nbd_co_writev_1(NbdClientSession *client, int64_t sector_num, int nb_sectors, QEMUIOVector *qiov, int offset) { - struct nbd_request request; + struct nbd_request request = { .type = NBD_CMD_WRITE }; struct nbd_reply reply; ssize_t ret; - request.type = NBD_CMD_WRITE; if (!bdrv_enable_write_cache(client->bs) && (client->nbdflags & NBD_FLAG_SEND_FUA)) { request.type |= NBD_CMD_FLAG_FUA; @@ -275,7 +273,7 @@ int nbd_client_session_co_writev(NbdClientSession *client, int64_t sector_num, int nbd_client_session_co_flush(NbdClientSession *client) { - struct nbd_request request; + struct nbd_request request = { .type = NBD_CMD_FLUSH }; struct nbd_reply reply; ssize_t ret; @@ -283,7 +281,6 @@ int nbd_client_session_co_flush(NbdClientSession *client) return 0; } - request.type = NBD_CMD_FLUSH; if (client->nbdflags & NBD_FLAG_SEND_FUA) { request.type |= NBD_CMD_FLAG_FUA; } @@ -305,14 +302,13 @@ int nbd_client_session_co_flush(NbdClientSession *client) int nbd_client_session_co_discard(NbdClientSession *client, int64_t sector_num, int nb_sectors) { - struct nbd_request request; + struct nbd_request request = { .type = NBD_CMD_TRIM }; struct nbd_reply reply; ssize_t ret; if (!(client->nbdflags & NBD_FLAG_SEND_TRIM)) { return 0; } - request.type = NBD_CMD_TRIM; request.from = sector_num * 512; request.len = nb_sectors * 512; @@ -330,11 +326,12 @@ int nbd_client_session_co_discard(NbdClientSession *client, int64_t sector_num, static void nbd_teardown_connection(NbdClientSession *client) { - struct nbd_request request; + struct nbd_request request = { + .type = NBD_CMD_DISC, + .from = 0, + .len = 0 + }; - request.type = NBD_CMD_DISC; - request.from = 0; - request.len = 0; nbd_send_request(client->sock, &request); /* finish any pending coroutines */ -- cgit v1.2.1