qemu-e2k/io/channel-socket.c

Ignoring revisions in .git-blame-ignore-revs. Click here to bypass and see the normal blame view.

953 lines
27 KiB
C
Raw Normal View History

/*
* QEMU I/O channels sockets driver
*
* Copyright (c) 2015 Red Hat, Inc.
*
* This library is free software; you can redistribute it and/or
* modify it under the terms of the GNU Lesser General Public
* License as published by the Free Software Foundation; either
* version 2.1 of the License, or (at your option) any later version.
*
* This library is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
* Lesser General Public License for more details.
*
* You should have received a copy of the GNU Lesser General Public
* License along with this library; if not, see <http://www.gnu.org/licenses/>.
*/
#include "qemu/osdep.h"
2016-03-14 09:01:28 +01:00
#include "qapi/error.h"
#include "qapi/qapi-visit-sockets.h"
#include "qemu/module.h"
#include "io/channel-socket.h"
io: follow coroutine AioContext in qio_channel_yield() The ongoing QEMU multi-queue block layer effort makes it possible for multiple threads to process I/O in parallel. The nbd block driver is not compatible with the multi-queue block layer yet because QIOChannel cannot be used easily from coroutines running in multiple threads. This series changes the QIOChannel API to make that possible. In the current API, calling qio_channel_attach_aio_context() sets the AioContext where qio_channel_yield() installs an fd handler prior to yielding: qio_channel_attach_aio_context(ioc, my_ctx); ... qio_channel_yield(ioc); // my_ctx is used here ... qio_channel_detach_aio_context(ioc); This API design has limitations: reading and writing must be done in the same AioContext and moving between AioContexts involves a cumbersome sequence of API calls that is not suitable for doing on a per-request basis. There is no fundamental reason why a QIOChannel needs to run within the same AioContext every time qio_channel_yield() is called. QIOChannel only uses the AioContext while inside qio_channel_yield(). The rest of the time, QIOChannel is independent of any AioContext. In the new API, qio_channel_yield() queries the AioContext from the current coroutine using qemu_coroutine_get_aio_context(). There is no need to explicitly attach/detach AioContexts anymore and qio_channel_attach_aio_context() and qio_channel_detach_aio_context() are gone. One coroutine can read from the QIOChannel while another coroutine writes from a different AioContext. This API change allows the nbd block driver to use QIOChannel from any thread. It's important to keep in mind that the block driver already synchronizes QIOChannel access and ensures that two coroutines never read simultaneously or write simultaneously. This patch updates all users of qio_channel_attach_aio_context() to the new API. Most conversions are simple, but vhost-user-server requires a new qemu_coroutine_yield() call to quiesce the vu_client_trip() coroutine when not attached to any AioContext. While the API is has become simpler, there is one wart: QIOChannel has a special case for the iohandler AioContext (used for handlers that must not run in nested event loops). I didn't find an elegant way preserve that behavior, so I added a new API called qio_channel_set_follow_coroutine_ctx(ioc, true|false) for opting in to the new AioContext model. By default QIOChannel uses the iohandler AioHandler. Code that formerly called qio_channel_attach_aio_context() now calls qio_channel_set_follow_coroutine_ctx(ioc, true) once after the QIOChannel is created. Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> Acked-by: Daniel P. Berrangé <berrange@redhat.com> Message-ID: <20230830224802.493686-5-stefanha@redhat.com> [eblake: also fix migration/rdma.c] Signed-off-by: Eric Blake <eblake@redhat.com>
2023-08-31 00:48:02 +02:00
#include "io/channel-util.h"
#include "io/channel-watch.h"
#include "trace.h"
#include "qapi/clone-visitor.h"
QIOChannelSocket: Implement io_writev zero copy flag & io_flush for CONFIG_LINUX For CONFIG_LINUX, implement the new zero copy flag and the optional callback io_flush on QIOChannelSocket, but enables it only when MSG_ZEROCOPY feature is available in the host kernel, which is checked on qio_channel_socket_connect_sync() qio_channel_socket_flush() was implemented by counting how many times sendmsg(...,MSG_ZEROCOPY) was successfully called, and then reading the socket's error queue, in order to find how many of them finished sending. Flush will loop until those counters are the same, or until some error occurs. Notes on using writev() with QIO_CHANNEL_WRITE_FLAG_ZERO_COPY: 1: Buffer - As MSG_ZEROCOPY tells the kernel to use the same user buffer to avoid copying, some caution is necessary to avoid overwriting any buffer before it's sent. If something like this happen, a newer version of the buffer may be sent instead. - If this is a problem, it's recommended to call qio_channel_flush() before freeing or re-using the buffer. 2: Locked memory - When using MSG_ZERCOCOPY, the buffer memory will be locked after queued, and unlocked after it's sent. - Depending on the size of each buffer, and how often it's sent, it may require a larger amount of locked memory than usually available to non-root user. - If the required amount of locked memory is not available, writev_zero_copy will return an error, which can abort an operation like migration, - Because of this, when an user code wants to add zero copy as a feature, it requires a mechanism to disable it, so it can still be accessible to less privileged users. Signed-off-by: Leonardo Bras <leobras@redhat.com> Reviewed-by: Peter Xu <peterx@redhat.com> Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> Reviewed-by: Juan Quintela <quintela@redhat.com> Message-Id: <20220513062836.965425-4-leobras@redhat.com> Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
2022-05-13 08:28:32 +02:00
#ifdef CONFIG_LINUX
#include <linux/errqueue.h>
#include <sys/socket.h>
#if (defined(MSG_ZEROCOPY) && defined(SO_ZEROCOPY))
#define QEMU_MSG_ZEROCOPY
#endif
#endif
#define SOCKET_MAX_FDS 16
SocketAddress *
qio_channel_socket_get_local_address(QIOChannelSocket *ioc,
Error **errp)
{
return socket_sockaddr_to_address(&ioc->localAddr,
ioc->localAddrLen,
errp);
}
SocketAddress *
qio_channel_socket_get_remote_address(QIOChannelSocket *ioc,
Error **errp)
{
return socket_sockaddr_to_address(&ioc->remoteAddr,
ioc->remoteAddrLen,
errp);
}
QIOChannelSocket *
qio_channel_socket_new(void)
{
QIOChannelSocket *sioc;
QIOChannel *ioc;
sioc = QIO_CHANNEL_SOCKET(object_new(TYPE_QIO_CHANNEL_SOCKET));
sioc->fd = -1;
QIOChannelSocket: Implement io_writev zero copy flag & io_flush for CONFIG_LINUX For CONFIG_LINUX, implement the new zero copy flag and the optional callback io_flush on QIOChannelSocket, but enables it only when MSG_ZEROCOPY feature is available in the host kernel, which is checked on qio_channel_socket_connect_sync() qio_channel_socket_flush() was implemented by counting how many times sendmsg(...,MSG_ZEROCOPY) was successfully called, and then reading the socket's error queue, in order to find how many of them finished sending. Flush will loop until those counters are the same, or until some error occurs. Notes on using writev() with QIO_CHANNEL_WRITE_FLAG_ZERO_COPY: 1: Buffer - As MSG_ZEROCOPY tells the kernel to use the same user buffer to avoid copying, some caution is necessary to avoid overwriting any buffer before it's sent. If something like this happen, a newer version of the buffer may be sent instead. - If this is a problem, it's recommended to call qio_channel_flush() before freeing or re-using the buffer. 2: Locked memory - When using MSG_ZERCOCOPY, the buffer memory will be locked after queued, and unlocked after it's sent. - Depending on the size of each buffer, and how often it's sent, it may require a larger amount of locked memory than usually available to non-root user. - If the required amount of locked memory is not available, writev_zero_copy will return an error, which can abort an operation like migration, - Because of this, when an user code wants to add zero copy as a feature, it requires a mechanism to disable it, so it can still be accessible to less privileged users. Signed-off-by: Leonardo Bras <leobras@redhat.com> Reviewed-by: Peter Xu <peterx@redhat.com> Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> Reviewed-by: Juan Quintela <quintela@redhat.com> Message-Id: <20220513062836.965425-4-leobras@redhat.com> Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
2022-05-13 08:28:32 +02:00
sioc->zero_copy_queued = 0;
sioc->zero_copy_sent = 0;
ioc = QIO_CHANNEL(sioc);
qio_channel_set_feature(ioc, QIO_CHANNEL_FEATURE_SHUTDOWN);
#ifdef WIN32
ioc->event = CreateEvent(NULL, FALSE, FALSE, NULL);
#endif
trace_qio_channel_socket_new(sioc);
return sioc;
}
static int
qio_channel_socket_set_fd(QIOChannelSocket *sioc,
int fd,
Error **errp)
{
if (sioc->fd != -1) {
error_setg(errp, "Socket is already open");
return -1;
}
sioc->fd = fd;
sioc->remoteAddrLen = sizeof(sioc->remoteAddr);
sioc->localAddrLen = sizeof(sioc->localAddr);
if (getpeername(fd, (struct sockaddr *)&sioc->remoteAddr,
&sioc->remoteAddrLen) < 0) {
if (errno == ENOTCONN) {
memset(&sioc->remoteAddr, 0, sizeof(sioc->remoteAddr));
sioc->remoteAddrLen = sizeof(sioc->remoteAddr);
} else {
error_setg_errno(errp, errno,
"Unable to query remote socket address");
goto error;
}
}
if (getsockname(fd, (struct sockaddr *)&sioc->localAddr,
&sioc->localAddrLen) < 0) {
error_setg_errno(errp, errno,
"Unable to query local socket address");
goto error;
}
#ifndef WIN32
if (sioc->localAddr.ss_family == AF_UNIX) {
QIOChannel *ioc = QIO_CHANNEL(sioc);
qio_channel_set_feature(ioc, QIO_CHANNEL_FEATURE_FD_PASS);
}
#endif /* WIN32 */
return 0;
error:
sioc->fd = -1; /* Let the caller close FD on failure */
return -1;
}
QIOChannelSocket *
qio_channel_socket_new_fd(int fd,
Error **errp)
{
QIOChannelSocket *ioc;
ioc = qio_channel_socket_new();
if (qio_channel_socket_set_fd(ioc, fd, errp) < 0) {
object_unref(OBJECT(ioc));
return NULL;
}
trace_qio_channel_socket_new_fd(ioc, fd);
return ioc;
}
int qio_channel_socket_connect_sync(QIOChannelSocket *ioc,
SocketAddress *addr,
Error **errp)
{
int fd;
trace_qio_channel_socket_connect_sync(ioc, addr);
fd = socket_connect(addr, errp);
if (fd < 0) {
trace_qio_channel_socket_connect_fail(ioc);
return -1;
}
trace_qio_channel_socket_connect_complete(ioc, fd);
if (qio_channel_socket_set_fd(ioc, fd, errp) < 0) {
close(fd);
return -1;
}
QIOChannelSocket: Implement io_writev zero copy flag & io_flush for CONFIG_LINUX For CONFIG_LINUX, implement the new zero copy flag and the optional callback io_flush on QIOChannelSocket, but enables it only when MSG_ZEROCOPY feature is available in the host kernel, which is checked on qio_channel_socket_connect_sync() qio_channel_socket_flush() was implemented by counting how many times sendmsg(...,MSG_ZEROCOPY) was successfully called, and then reading the socket's error queue, in order to find how many of them finished sending. Flush will loop until those counters are the same, or until some error occurs. Notes on using writev() with QIO_CHANNEL_WRITE_FLAG_ZERO_COPY: 1: Buffer - As MSG_ZEROCOPY tells the kernel to use the same user buffer to avoid copying, some caution is necessary to avoid overwriting any buffer before it's sent. If something like this happen, a newer version of the buffer may be sent instead. - If this is a problem, it's recommended to call qio_channel_flush() before freeing or re-using the buffer. 2: Locked memory - When using MSG_ZERCOCOPY, the buffer memory will be locked after queued, and unlocked after it's sent. - Depending on the size of each buffer, and how often it's sent, it may require a larger amount of locked memory than usually available to non-root user. - If the required amount of locked memory is not available, writev_zero_copy will return an error, which can abort an operation like migration, - Because of this, when an user code wants to add zero copy as a feature, it requires a mechanism to disable it, so it can still be accessible to less privileged users. Signed-off-by: Leonardo Bras <leobras@redhat.com> Reviewed-by: Peter Xu <peterx@redhat.com> Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> Reviewed-by: Juan Quintela <quintela@redhat.com> Message-Id: <20220513062836.965425-4-leobras@redhat.com> Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
2022-05-13 08:28:32 +02:00
#ifdef QEMU_MSG_ZEROCOPY
int ret, v = 1;
ret = setsockopt(fd, SOL_SOCKET, SO_ZEROCOPY, &v, sizeof(v));
if (ret == 0) {
/* Zero copy available on host */
qio_channel_set_feature(QIO_CHANNEL(ioc),
QIO_CHANNEL_FEATURE_WRITE_ZERO_COPY);
}
#endif
qio_channel_set_feature(QIO_CHANNEL(ioc),
QIO_CHANNEL_FEATURE_READ_MSG_PEEK);
return 0;
}
static void qio_channel_socket_connect_worker(QIOTask *task,
gpointer opaque)
{
QIOChannelSocket *ioc = QIO_CHANNEL_SOCKET(qio_task_get_source(task));
SocketAddress *addr = opaque;
Error *err = NULL;
qio_channel_socket_connect_sync(ioc, addr, &err);
qio_task_set_error(task, err);
}
void qio_channel_socket_connect_async(QIOChannelSocket *ioc,
SocketAddress *addr,
QIOTaskFunc callback,
gpointer opaque,
GDestroyNotify destroy,
GMainContext *context)
{
QIOTask *task = qio_task_new(
OBJECT(ioc), callback, opaque, destroy);
SocketAddress *addrCopy;
addrCopy = QAPI_CLONE(SocketAddress, addr);
/* socket_connect() does a non-blocking connect(), but it
* still blocks in DNS lookups, so we must use a thread */
trace_qio_channel_socket_connect_async(ioc, addr);
qio_task_run_in_thread(task,
qio_channel_socket_connect_worker,
addrCopy,
(GDestroyNotify)qapi_free_SocketAddress,
context);
}
int qio_channel_socket_listen_sync(QIOChannelSocket *ioc,
SocketAddress *addr,
int num,
Error **errp)
{
int fd;
trace_qio_channel_socket_listen_sync(ioc, addr, num);
fd = socket_listen(addr, num, errp);
if (fd < 0) {
trace_qio_channel_socket_listen_fail(ioc);
return -1;
}
trace_qio_channel_socket_listen_complete(ioc, fd);
if (qio_channel_socket_set_fd(ioc, fd, errp) < 0) {
close(fd);
return -1;
}
qio_channel_set_feature(QIO_CHANNEL(ioc), QIO_CHANNEL_FEATURE_LISTEN);
return 0;
}
struct QIOChannelListenWorkerData {
SocketAddress *addr;
int num; /* amount of expected connections */
};
static void qio_channel_listen_worker_free(gpointer opaque)
{
struct QIOChannelListenWorkerData *data = opaque;
qapi_free_SocketAddress(data->addr);
g_free(data);
}
static void qio_channel_socket_listen_worker(QIOTask *task,
gpointer opaque)
{
QIOChannelSocket *ioc = QIO_CHANNEL_SOCKET(qio_task_get_source(task));
struct QIOChannelListenWorkerData *data = opaque;
Error *err = NULL;
qio_channel_socket_listen_sync(ioc, data->addr, data->num, &err);
qio_task_set_error(task, err);
}
void qio_channel_socket_listen_async(QIOChannelSocket *ioc,
SocketAddress *addr,
int num,
QIOTaskFunc callback,
gpointer opaque,
GDestroyNotify destroy,
GMainContext *context)
{
QIOTask *task = qio_task_new(
OBJECT(ioc), callback, opaque, destroy);
struct QIOChannelListenWorkerData *data;
data = g_new0(struct QIOChannelListenWorkerData, 1);
data->addr = QAPI_CLONE(SocketAddress, addr);
data->num = num;
/* socket_listen() blocks in DNS lookups, so we must use a thread */
trace_qio_channel_socket_listen_async(ioc, addr, num);
qio_task_run_in_thread(task,
qio_channel_socket_listen_worker,
data,
qio_channel_listen_worker_free,
context);
}
int qio_channel_socket_dgram_sync(QIOChannelSocket *ioc,
SocketAddress *localAddr,
SocketAddress *remoteAddr,
Error **errp)
{
int fd;
trace_qio_channel_socket_dgram_sync(ioc, localAddr, remoteAddr);
fd = socket_dgram(remoteAddr, localAddr, errp);
if (fd < 0) {
trace_qio_channel_socket_dgram_fail(ioc);
return -1;
}
trace_qio_channel_socket_dgram_complete(ioc, fd);
if (qio_channel_socket_set_fd(ioc, fd, errp) < 0) {
close(fd);
return -1;
}
return 0;
}
struct QIOChannelSocketDGramWorkerData {
SocketAddress *localAddr;
SocketAddress *remoteAddr;
};
static void qio_channel_socket_dgram_worker_free(gpointer opaque)
{
struct QIOChannelSocketDGramWorkerData *data = opaque;
qapi_free_SocketAddress(data->localAddr);
qapi_free_SocketAddress(data->remoteAddr);
g_free(data);
}
static void qio_channel_socket_dgram_worker(QIOTask *task,
gpointer opaque)
{
QIOChannelSocket *ioc = QIO_CHANNEL_SOCKET(qio_task_get_source(task));
struct QIOChannelSocketDGramWorkerData *data = opaque;
Error *err = NULL;
/* socket_dgram() blocks in DNS lookups, so we must use a thread */
qio_channel_socket_dgram_sync(ioc, data->localAddr,
data->remoteAddr, &err);
qio_task_set_error(task, err);
}
void qio_channel_socket_dgram_async(QIOChannelSocket *ioc,
SocketAddress *localAddr,
SocketAddress *remoteAddr,
QIOTaskFunc callback,
gpointer opaque,
GDestroyNotify destroy,
GMainContext *context)
{
QIOTask *task = qio_task_new(
OBJECT(ioc), callback, opaque, destroy);
struct QIOChannelSocketDGramWorkerData *data = g_new0(
struct QIOChannelSocketDGramWorkerData, 1);
data->localAddr = QAPI_CLONE(SocketAddress, localAddr);
data->remoteAddr = QAPI_CLONE(SocketAddress, remoteAddr);
trace_qio_channel_socket_dgram_async(ioc, localAddr, remoteAddr);
qio_task_run_in_thread(task,
qio_channel_socket_dgram_worker,
data,
qio_channel_socket_dgram_worker_free,
context);
}
QIOChannelSocket *
qio_channel_socket_accept(QIOChannelSocket *ioc,
Error **errp)
{
QIOChannelSocket *cioc;
cioc = qio_channel_socket_new();
cioc->remoteAddrLen = sizeof(ioc->remoteAddr);
cioc->localAddrLen = sizeof(ioc->localAddr);
retry:
trace_qio_channel_socket_accept(ioc);
cioc->fd = qemu_accept(ioc->fd, (struct sockaddr *)&cioc->remoteAddr,
&cioc->remoteAddrLen);
if (cioc->fd < 0) {
if (errno == EINTR) {
goto retry;
}
error_setg_errno(errp, errno, "Unable to accept connection");
trace_qio_channel_socket_accept_fail(ioc);
goto error;
}
if (getsockname(cioc->fd, (struct sockaddr *)&cioc->localAddr,
&cioc->localAddrLen) < 0) {
error_setg_errno(errp, errno,
"Unable to query local socket address");
goto error;
}
#ifndef WIN32
if (cioc->localAddr.ss_family == AF_UNIX) {
QIOChannel *ioc_local = QIO_CHANNEL(cioc);
qio_channel_set_feature(ioc_local, QIO_CHANNEL_FEATURE_FD_PASS);
}
#endif /* WIN32 */
qio_channel_set_feature(QIO_CHANNEL(cioc),
QIO_CHANNEL_FEATURE_READ_MSG_PEEK);
trace_qio_channel_socket_accept_complete(ioc, cioc, cioc->fd);
return cioc;
error:
object_unref(OBJECT(cioc));
return NULL;
}
static void qio_channel_socket_init(Object *obj)
{
QIOChannelSocket *ioc = QIO_CHANNEL_SOCKET(obj);
ioc->fd = -1;
}
static void qio_channel_socket_finalize(Object *obj)
{
QIOChannelSocket *ioc = QIO_CHANNEL_SOCKET(obj);
if (ioc->fd != -1) {
QIOChannel *ioc_local = QIO_CHANNEL(ioc);
if (qio_channel_has_feature(ioc_local, QIO_CHANNEL_FEATURE_LISTEN)) {
Error *err = NULL;
socket_listen_cleanup(ioc->fd, &err);
if (err) {
error_report_err(err);
err = NULL;
}
}
#ifdef WIN32
qemu_socket_unselect(ioc->fd, NULL);
#endif
close(ioc->fd);
ioc->fd = -1;
}
}
#ifndef WIN32
static void qio_channel_socket_copy_fds(struct msghdr *msg,
int **fds, size_t *nfds)
{
struct cmsghdr *cmsg;
*nfds = 0;
*fds = NULL;
for (cmsg = CMSG_FIRSTHDR(msg); cmsg; cmsg = CMSG_NXTHDR(msg, cmsg)) {
int fd_size, i;
int gotfds;
if (cmsg->cmsg_len < CMSG_LEN(sizeof(int)) ||
cmsg->cmsg_level != SOL_SOCKET ||
cmsg->cmsg_type != SCM_RIGHTS) {
continue;
}
fd_size = cmsg->cmsg_len - CMSG_LEN(0);
if (!fd_size) {
continue;
}
gotfds = fd_size / sizeof(int);
*fds = g_renew(int, *fds, *nfds + gotfds);
memcpy(*fds + *nfds, CMSG_DATA(cmsg), fd_size);
for (i = 0; i < gotfds; i++) {
int fd = (*fds)[*nfds + i];
if (fd < 0) {
continue;
}
/* O_NONBLOCK is preserved across SCM_RIGHTS so reset it */
qemu_socket_set_block(fd);
#ifndef MSG_CMSG_CLOEXEC
qemu_set_cloexec(fd);
#endif
}
*nfds += gotfds;
}
}
static ssize_t qio_channel_socket_readv(QIOChannel *ioc,
const struct iovec *iov,
size_t niov,
int **fds,
size_t *nfds,
int flags,
Error **errp)
{
QIOChannelSocket *sioc = QIO_CHANNEL_SOCKET(ioc);
ssize_t ret;
struct msghdr msg = { NULL, };
char control[CMSG_SPACE(sizeof(int) * SOCKET_MAX_FDS)];
int sflags = 0;
memset(control, 0, CMSG_SPACE(sizeof(int) * SOCKET_MAX_FDS));
msg.msg_iov = (struct iovec *)iov;
msg.msg_iovlen = niov;
if (fds && nfds) {
msg.msg_control = control;
msg.msg_controllen = sizeof(control);
#ifdef MSG_CMSG_CLOEXEC
sflags |= MSG_CMSG_CLOEXEC;
#endif
}
if (flags & QIO_CHANNEL_READ_FLAG_MSG_PEEK) {
sflags |= MSG_PEEK;
}
retry:
ret = recvmsg(sioc->fd, &msg, sflags);
if (ret < 0) {
if (errno == EAGAIN) {
return QIO_CHANNEL_ERR_BLOCK;
}
if (errno == EINTR) {
goto retry;
}
error_setg_errno(errp, errno,
"Unable to read from socket");
return -1;
}
if (fds && nfds) {
qio_channel_socket_copy_fds(&msg, fds, nfds);
}
return ret;
}
static ssize_t qio_channel_socket_writev(QIOChannel *ioc,
const struct iovec *iov,
size_t niov,
int *fds,
size_t nfds,
int flags,
Error **errp)
{
QIOChannelSocket *sioc = QIO_CHANNEL_SOCKET(ioc);
ssize_t ret;
struct msghdr msg = { NULL, };
char control[CMSG_SPACE(sizeof(int) * SOCKET_MAX_FDS)];
size_t fdsize = sizeof(int) * nfds;
struct cmsghdr *cmsg;
QIOChannelSocket: Implement io_writev zero copy flag & io_flush for CONFIG_LINUX For CONFIG_LINUX, implement the new zero copy flag and the optional callback io_flush on QIOChannelSocket, but enables it only when MSG_ZEROCOPY feature is available in the host kernel, which is checked on qio_channel_socket_connect_sync() qio_channel_socket_flush() was implemented by counting how many times sendmsg(...,MSG_ZEROCOPY) was successfully called, and then reading the socket's error queue, in order to find how many of them finished sending. Flush will loop until those counters are the same, or until some error occurs. Notes on using writev() with QIO_CHANNEL_WRITE_FLAG_ZERO_COPY: 1: Buffer - As MSG_ZEROCOPY tells the kernel to use the same user buffer to avoid copying, some caution is necessary to avoid overwriting any buffer before it's sent. If something like this happen, a newer version of the buffer may be sent instead. - If this is a problem, it's recommended to call qio_channel_flush() before freeing or re-using the buffer. 2: Locked memory - When using MSG_ZERCOCOPY, the buffer memory will be locked after queued, and unlocked after it's sent. - Depending on the size of each buffer, and how often it's sent, it may require a larger amount of locked memory than usually available to non-root user. - If the required amount of locked memory is not available, writev_zero_copy will return an error, which can abort an operation like migration, - Because of this, when an user code wants to add zero copy as a feature, it requires a mechanism to disable it, so it can still be accessible to less privileged users. Signed-off-by: Leonardo Bras <leobras@redhat.com> Reviewed-by: Peter Xu <peterx@redhat.com> Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> Reviewed-by: Juan Quintela <quintela@redhat.com> Message-Id: <20220513062836.965425-4-leobras@redhat.com> Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
2022-05-13 08:28:32 +02:00
int sflags = 0;
memset(control, 0, CMSG_SPACE(sizeof(int) * SOCKET_MAX_FDS));
msg.msg_iov = (struct iovec *)iov;
msg.msg_iovlen = niov;
if (nfds) {
if (nfds > SOCKET_MAX_FDS) {
error_setg_errno(errp, EINVAL,
"Only %d FDs can be sent, got %zu",
SOCKET_MAX_FDS, nfds);
return -1;
}
msg.msg_control = control;
msg.msg_controllen = CMSG_SPACE(sizeof(int) * nfds);
cmsg = CMSG_FIRSTHDR(&msg);
cmsg->cmsg_len = CMSG_LEN(fdsize);
cmsg->cmsg_level = SOL_SOCKET;
cmsg->cmsg_type = SCM_RIGHTS;
memcpy(CMSG_DATA(cmsg), fds, fdsize);
}
QIOChannelSocket: Implement io_writev zero copy flag & io_flush for CONFIG_LINUX For CONFIG_LINUX, implement the new zero copy flag and the optional callback io_flush on QIOChannelSocket, but enables it only when MSG_ZEROCOPY feature is available in the host kernel, which is checked on qio_channel_socket_connect_sync() qio_channel_socket_flush() was implemented by counting how many times sendmsg(...,MSG_ZEROCOPY) was successfully called, and then reading the socket's error queue, in order to find how many of them finished sending. Flush will loop until those counters are the same, or until some error occurs. Notes on using writev() with QIO_CHANNEL_WRITE_FLAG_ZERO_COPY: 1: Buffer - As MSG_ZEROCOPY tells the kernel to use the same user buffer to avoid copying, some caution is necessary to avoid overwriting any buffer before it's sent. If something like this happen, a newer version of the buffer may be sent instead. - If this is a problem, it's recommended to call qio_channel_flush() before freeing or re-using the buffer. 2: Locked memory - When using MSG_ZERCOCOPY, the buffer memory will be locked after queued, and unlocked after it's sent. - Depending on the size of each buffer, and how often it's sent, it may require a larger amount of locked memory than usually available to non-root user. - If the required amount of locked memory is not available, writev_zero_copy will return an error, which can abort an operation like migration, - Because of this, when an user code wants to add zero copy as a feature, it requires a mechanism to disable it, so it can still be accessible to less privileged users. Signed-off-by: Leonardo Bras <leobras@redhat.com> Reviewed-by: Peter Xu <peterx@redhat.com> Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> Reviewed-by: Juan Quintela <quintela@redhat.com> Message-Id: <20220513062836.965425-4-leobras@redhat.com> Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
2022-05-13 08:28:32 +02:00
if (flags & QIO_CHANNEL_WRITE_FLAG_ZERO_COPY) {
#ifdef QEMU_MSG_ZEROCOPY
QIOChannelSocket: Implement io_writev zero copy flag & io_flush for CONFIG_LINUX For CONFIG_LINUX, implement the new zero copy flag and the optional callback io_flush on QIOChannelSocket, but enables it only when MSG_ZEROCOPY feature is available in the host kernel, which is checked on qio_channel_socket_connect_sync() qio_channel_socket_flush() was implemented by counting how many times sendmsg(...,MSG_ZEROCOPY) was successfully called, and then reading the socket's error queue, in order to find how many of them finished sending. Flush will loop until those counters are the same, or until some error occurs. Notes on using writev() with QIO_CHANNEL_WRITE_FLAG_ZERO_COPY: 1: Buffer - As MSG_ZEROCOPY tells the kernel to use the same user buffer to avoid copying, some caution is necessary to avoid overwriting any buffer before it's sent. If something like this happen, a newer version of the buffer may be sent instead. - If this is a problem, it's recommended to call qio_channel_flush() before freeing or re-using the buffer. 2: Locked memory - When using MSG_ZERCOCOPY, the buffer memory will be locked after queued, and unlocked after it's sent. - Depending on the size of each buffer, and how often it's sent, it may require a larger amount of locked memory than usually available to non-root user. - If the required amount of locked memory is not available, writev_zero_copy will return an error, which can abort an operation like migration, - Because of this, when an user code wants to add zero copy as a feature, it requires a mechanism to disable it, so it can still be accessible to less privileged users. Signed-off-by: Leonardo Bras <leobras@redhat.com> Reviewed-by: Peter Xu <peterx@redhat.com> Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> Reviewed-by: Juan Quintela <quintela@redhat.com> Message-Id: <20220513062836.965425-4-leobras@redhat.com> Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
2022-05-13 08:28:32 +02:00
sflags = MSG_ZEROCOPY;
#else
/*
* We expect QIOChannel class entry point to have
* blocked this code path already
*/
g_assert_not_reached();
QIOChannelSocket: Implement io_writev zero copy flag & io_flush for CONFIG_LINUX For CONFIG_LINUX, implement the new zero copy flag and the optional callback io_flush on QIOChannelSocket, but enables it only when MSG_ZEROCOPY feature is available in the host kernel, which is checked on qio_channel_socket_connect_sync() qio_channel_socket_flush() was implemented by counting how many times sendmsg(...,MSG_ZEROCOPY) was successfully called, and then reading the socket's error queue, in order to find how many of them finished sending. Flush will loop until those counters are the same, or until some error occurs. Notes on using writev() with QIO_CHANNEL_WRITE_FLAG_ZERO_COPY: 1: Buffer - As MSG_ZEROCOPY tells the kernel to use the same user buffer to avoid copying, some caution is necessary to avoid overwriting any buffer before it's sent. If something like this happen, a newer version of the buffer may be sent instead. - If this is a problem, it's recommended to call qio_channel_flush() before freeing or re-using the buffer. 2: Locked memory - When using MSG_ZERCOCOPY, the buffer memory will be locked after queued, and unlocked after it's sent. - Depending on the size of each buffer, and how often it's sent, it may require a larger amount of locked memory than usually available to non-root user. - If the required amount of locked memory is not available, writev_zero_copy will return an error, which can abort an operation like migration, - Because of this, when an user code wants to add zero copy as a feature, it requires a mechanism to disable it, so it can still be accessible to less privileged users. Signed-off-by: Leonardo Bras <leobras@redhat.com> Reviewed-by: Peter Xu <peterx@redhat.com> Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> Reviewed-by: Juan Quintela <quintela@redhat.com> Message-Id: <20220513062836.965425-4-leobras@redhat.com> Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
2022-05-13 08:28:32 +02:00
#endif
}
QIOChannelSocket: Implement io_writev zero copy flag & io_flush for CONFIG_LINUX For CONFIG_LINUX, implement the new zero copy flag and the optional callback io_flush on QIOChannelSocket, but enables it only when MSG_ZEROCOPY feature is available in the host kernel, which is checked on qio_channel_socket_connect_sync() qio_channel_socket_flush() was implemented by counting how many times sendmsg(...,MSG_ZEROCOPY) was successfully called, and then reading the socket's error queue, in order to find how many of them finished sending. Flush will loop until those counters are the same, or until some error occurs. Notes on using writev() with QIO_CHANNEL_WRITE_FLAG_ZERO_COPY: 1: Buffer - As MSG_ZEROCOPY tells the kernel to use the same user buffer to avoid copying, some caution is necessary to avoid overwriting any buffer before it's sent. If something like this happen, a newer version of the buffer may be sent instead. - If this is a problem, it's recommended to call qio_channel_flush() before freeing or re-using the buffer. 2: Locked memory - When using MSG_ZERCOCOPY, the buffer memory will be locked after queued, and unlocked after it's sent. - Depending on the size of each buffer, and how often it's sent, it may require a larger amount of locked memory than usually available to non-root user. - If the required amount of locked memory is not available, writev_zero_copy will return an error, which can abort an operation like migration, - Because of this, when an user code wants to add zero copy as a feature, it requires a mechanism to disable it, so it can still be accessible to less privileged users. Signed-off-by: Leonardo Bras <leobras@redhat.com> Reviewed-by: Peter Xu <peterx@redhat.com> Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> Reviewed-by: Juan Quintela <quintela@redhat.com> Message-Id: <20220513062836.965425-4-leobras@redhat.com> Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
2022-05-13 08:28:32 +02:00
retry:
QIOChannelSocket: Implement io_writev zero copy flag & io_flush for CONFIG_LINUX For CONFIG_LINUX, implement the new zero copy flag and the optional callback io_flush on QIOChannelSocket, but enables it only when MSG_ZEROCOPY feature is available in the host kernel, which is checked on qio_channel_socket_connect_sync() qio_channel_socket_flush() was implemented by counting how many times sendmsg(...,MSG_ZEROCOPY) was successfully called, and then reading the socket's error queue, in order to find how many of them finished sending. Flush will loop until those counters are the same, or until some error occurs. Notes on using writev() with QIO_CHANNEL_WRITE_FLAG_ZERO_COPY: 1: Buffer - As MSG_ZEROCOPY tells the kernel to use the same user buffer to avoid copying, some caution is necessary to avoid overwriting any buffer before it's sent. If something like this happen, a newer version of the buffer may be sent instead. - If this is a problem, it's recommended to call qio_channel_flush() before freeing or re-using the buffer. 2: Locked memory - When using MSG_ZERCOCOPY, the buffer memory will be locked after queued, and unlocked after it's sent. - Depending on the size of each buffer, and how often it's sent, it may require a larger amount of locked memory than usually available to non-root user. - If the required amount of locked memory is not available, writev_zero_copy will return an error, which can abort an operation like migration, - Because of this, when an user code wants to add zero copy as a feature, it requires a mechanism to disable it, so it can still be accessible to less privileged users. Signed-off-by: Leonardo Bras <leobras@redhat.com> Reviewed-by: Peter Xu <peterx@redhat.com> Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> Reviewed-by: Juan Quintela <quintela@redhat.com> Message-Id: <20220513062836.965425-4-leobras@redhat.com> Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
2022-05-13 08:28:32 +02:00
ret = sendmsg(sioc->fd, &msg, sflags);
if (ret <= 0) {
QIOChannelSocket: Implement io_writev zero copy flag & io_flush for CONFIG_LINUX For CONFIG_LINUX, implement the new zero copy flag and the optional callback io_flush on QIOChannelSocket, but enables it only when MSG_ZEROCOPY feature is available in the host kernel, which is checked on qio_channel_socket_connect_sync() qio_channel_socket_flush() was implemented by counting how many times sendmsg(...,MSG_ZEROCOPY) was successfully called, and then reading the socket's error queue, in order to find how many of them finished sending. Flush will loop until those counters are the same, or until some error occurs. Notes on using writev() with QIO_CHANNEL_WRITE_FLAG_ZERO_COPY: 1: Buffer - As MSG_ZEROCOPY tells the kernel to use the same user buffer to avoid copying, some caution is necessary to avoid overwriting any buffer before it's sent. If something like this happen, a newer version of the buffer may be sent instead. - If this is a problem, it's recommended to call qio_channel_flush() before freeing or re-using the buffer. 2: Locked memory - When using MSG_ZERCOCOPY, the buffer memory will be locked after queued, and unlocked after it's sent. - Depending on the size of each buffer, and how often it's sent, it may require a larger amount of locked memory than usually available to non-root user. - If the required amount of locked memory is not available, writev_zero_copy will return an error, which can abort an operation like migration, - Because of this, when an user code wants to add zero copy as a feature, it requires a mechanism to disable it, so it can still be accessible to less privileged users. Signed-off-by: Leonardo Bras <leobras@redhat.com> Reviewed-by: Peter Xu <peterx@redhat.com> Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> Reviewed-by: Juan Quintela <quintela@redhat.com> Message-Id: <20220513062836.965425-4-leobras@redhat.com> Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
2022-05-13 08:28:32 +02:00
switch (errno) {
case EAGAIN:
return QIO_CHANNEL_ERR_BLOCK;
QIOChannelSocket: Implement io_writev zero copy flag & io_flush for CONFIG_LINUX For CONFIG_LINUX, implement the new zero copy flag and the optional callback io_flush on QIOChannelSocket, but enables it only when MSG_ZEROCOPY feature is available in the host kernel, which is checked on qio_channel_socket_connect_sync() qio_channel_socket_flush() was implemented by counting how many times sendmsg(...,MSG_ZEROCOPY) was successfully called, and then reading the socket's error queue, in order to find how many of them finished sending. Flush will loop until those counters are the same, or until some error occurs. Notes on using writev() with QIO_CHANNEL_WRITE_FLAG_ZERO_COPY: 1: Buffer - As MSG_ZEROCOPY tells the kernel to use the same user buffer to avoid copying, some caution is necessary to avoid overwriting any buffer before it's sent. If something like this happen, a newer version of the buffer may be sent instead. - If this is a problem, it's recommended to call qio_channel_flush() before freeing or re-using the buffer. 2: Locked memory - When using MSG_ZERCOCOPY, the buffer memory will be locked after queued, and unlocked after it's sent. - Depending on the size of each buffer, and how often it's sent, it may require a larger amount of locked memory than usually available to non-root user. - If the required amount of locked memory is not available, writev_zero_copy will return an error, which can abort an operation like migration, - Because of this, when an user code wants to add zero copy as a feature, it requires a mechanism to disable it, so it can still be accessible to less privileged users. Signed-off-by: Leonardo Bras <leobras@redhat.com> Reviewed-by: Peter Xu <peterx@redhat.com> Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> Reviewed-by: Juan Quintela <quintela@redhat.com> Message-Id: <20220513062836.965425-4-leobras@redhat.com> Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
2022-05-13 08:28:32 +02:00
case EINTR:
goto retry;
QIOChannelSocket: Implement io_writev zero copy flag & io_flush for CONFIG_LINUX For CONFIG_LINUX, implement the new zero copy flag and the optional callback io_flush on QIOChannelSocket, but enables it only when MSG_ZEROCOPY feature is available in the host kernel, which is checked on qio_channel_socket_connect_sync() qio_channel_socket_flush() was implemented by counting how many times sendmsg(...,MSG_ZEROCOPY) was successfully called, and then reading the socket's error queue, in order to find how many of them finished sending. Flush will loop until those counters are the same, or until some error occurs. Notes on using writev() with QIO_CHANNEL_WRITE_FLAG_ZERO_COPY: 1: Buffer - As MSG_ZEROCOPY tells the kernel to use the same user buffer to avoid copying, some caution is necessary to avoid overwriting any buffer before it's sent. If something like this happen, a newer version of the buffer may be sent instead. - If this is a problem, it's recommended to call qio_channel_flush() before freeing or re-using the buffer. 2: Locked memory - When using MSG_ZERCOCOPY, the buffer memory will be locked after queued, and unlocked after it's sent. - Depending on the size of each buffer, and how often it's sent, it may require a larger amount of locked memory than usually available to non-root user. - If the required amount of locked memory is not available, writev_zero_copy will return an error, which can abort an operation like migration, - Because of this, when an user code wants to add zero copy as a feature, it requires a mechanism to disable it, so it can still be accessible to less privileged users. Signed-off-by: Leonardo Bras <leobras@redhat.com> Reviewed-by: Peter Xu <peterx@redhat.com> Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> Reviewed-by: Juan Quintela <quintela@redhat.com> Message-Id: <20220513062836.965425-4-leobras@redhat.com> Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
2022-05-13 08:28:32 +02:00
case ENOBUFS:
if (flags & QIO_CHANNEL_WRITE_FLAG_ZERO_COPY) {
QIOChannelSocket: Implement io_writev zero copy flag & io_flush for CONFIG_LINUX For CONFIG_LINUX, implement the new zero copy flag and the optional callback io_flush on QIOChannelSocket, but enables it only when MSG_ZEROCOPY feature is available in the host kernel, which is checked on qio_channel_socket_connect_sync() qio_channel_socket_flush() was implemented by counting how many times sendmsg(...,MSG_ZEROCOPY) was successfully called, and then reading the socket's error queue, in order to find how many of them finished sending. Flush will loop until those counters are the same, or until some error occurs. Notes on using writev() with QIO_CHANNEL_WRITE_FLAG_ZERO_COPY: 1: Buffer - As MSG_ZEROCOPY tells the kernel to use the same user buffer to avoid copying, some caution is necessary to avoid overwriting any buffer before it's sent. If something like this happen, a newer version of the buffer may be sent instead. - If this is a problem, it's recommended to call qio_channel_flush() before freeing or re-using the buffer. 2: Locked memory - When using MSG_ZERCOCOPY, the buffer memory will be locked after queued, and unlocked after it's sent. - Depending on the size of each buffer, and how often it's sent, it may require a larger amount of locked memory than usually available to non-root user. - If the required amount of locked memory is not available, writev_zero_copy will return an error, which can abort an operation like migration, - Because of this, when an user code wants to add zero copy as a feature, it requires a mechanism to disable it, so it can still be accessible to less privileged users. Signed-off-by: Leonardo Bras <leobras@redhat.com> Reviewed-by: Peter Xu <peterx@redhat.com> Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> Reviewed-by: Juan Quintela <quintela@redhat.com> Message-Id: <20220513062836.965425-4-leobras@redhat.com> Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
2022-05-13 08:28:32 +02:00
error_setg_errno(errp, errno,
"Process can't lock enough memory for using MSG_ZEROCOPY");
return -1;
}
break;
}
QIOChannelSocket: Implement io_writev zero copy flag & io_flush for CONFIG_LINUX For CONFIG_LINUX, implement the new zero copy flag and the optional callback io_flush on QIOChannelSocket, but enables it only when MSG_ZEROCOPY feature is available in the host kernel, which is checked on qio_channel_socket_connect_sync() qio_channel_socket_flush() was implemented by counting how many times sendmsg(...,MSG_ZEROCOPY) was successfully called, and then reading the socket's error queue, in order to find how many of them finished sending. Flush will loop until those counters are the same, or until some error occurs. Notes on using writev() with QIO_CHANNEL_WRITE_FLAG_ZERO_COPY: 1: Buffer - As MSG_ZEROCOPY tells the kernel to use the same user buffer to avoid copying, some caution is necessary to avoid overwriting any buffer before it's sent. If something like this happen, a newer version of the buffer may be sent instead. - If this is a problem, it's recommended to call qio_channel_flush() before freeing or re-using the buffer. 2: Locked memory - When using MSG_ZERCOCOPY, the buffer memory will be locked after queued, and unlocked after it's sent. - Depending on the size of each buffer, and how often it's sent, it may require a larger amount of locked memory than usually available to non-root user. - If the required amount of locked memory is not available, writev_zero_copy will return an error, which can abort an operation like migration, - Because of this, when an user code wants to add zero copy as a feature, it requires a mechanism to disable it, so it can still be accessible to less privileged users. Signed-off-by: Leonardo Bras <leobras@redhat.com> Reviewed-by: Peter Xu <peterx@redhat.com> Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> Reviewed-by: Juan Quintela <quintela@redhat.com> Message-Id: <20220513062836.965425-4-leobras@redhat.com> Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
2022-05-13 08:28:32 +02:00
error_setg_errno(errp, errno,
"Unable to write to socket");
return -1;
}
if (flags & QIO_CHANNEL_WRITE_FLAG_ZERO_COPY) {
sioc->zero_copy_queued++;
}
return ret;
}
#else /* WIN32 */
static ssize_t qio_channel_socket_readv(QIOChannel *ioc,
const struct iovec *iov,
size_t niov,
int **fds,
size_t *nfds,
int flags,
Error **errp)
{
QIOChannelSocket *sioc = QIO_CHANNEL_SOCKET(ioc);
ssize_t done = 0;
ssize_t i;
int sflags = 0;
if (flags & QIO_CHANNEL_READ_FLAG_MSG_PEEK) {
sflags |= MSG_PEEK;
}
for (i = 0; i < niov; i++) {
ssize_t ret;
retry:
ret = recv(sioc->fd,
iov[i].iov_base,
iov[i].iov_len,
sflags);
if (ret < 0) {
if (errno == EAGAIN) {
if (done) {
return done;
} else {
return QIO_CHANNEL_ERR_BLOCK;
}
} else if (errno == EINTR) {
goto retry;
} else {
error_setg_errno(errp, errno,
"Unable to read from socket");
return -1;
}
}
done += ret;
if (ret < iov[i].iov_len) {
return done;
}
}
return done;
}
static ssize_t qio_channel_socket_writev(QIOChannel *ioc,
const struct iovec *iov,
size_t niov,
int *fds,
size_t nfds,
int flags,
Error **errp)
{
QIOChannelSocket *sioc = QIO_CHANNEL_SOCKET(ioc);
ssize_t done = 0;
ssize_t i;
for (i = 0; i < niov; i++) {
ssize_t ret;
retry:
ret = send(sioc->fd,
iov[i].iov_base,
iov[i].iov_len,
0);
if (ret < 0) {
if (errno == EAGAIN) {
if (done) {
return done;
} else {
return QIO_CHANNEL_ERR_BLOCK;
}
} else if (errno == EINTR) {
goto retry;
} else {
error_setg_errno(errp, errno,
"Unable to write to socket");
return -1;
}
}
done += ret;
if (ret < iov[i].iov_len) {
return done;
}
}
return done;
}
#endif /* WIN32 */
QIOChannelSocket: Implement io_writev zero copy flag & io_flush for CONFIG_LINUX For CONFIG_LINUX, implement the new zero copy flag and the optional callback io_flush on QIOChannelSocket, but enables it only when MSG_ZEROCOPY feature is available in the host kernel, which is checked on qio_channel_socket_connect_sync() qio_channel_socket_flush() was implemented by counting how many times sendmsg(...,MSG_ZEROCOPY) was successfully called, and then reading the socket's error queue, in order to find how many of them finished sending. Flush will loop until those counters are the same, or until some error occurs. Notes on using writev() with QIO_CHANNEL_WRITE_FLAG_ZERO_COPY: 1: Buffer - As MSG_ZEROCOPY tells the kernel to use the same user buffer to avoid copying, some caution is necessary to avoid overwriting any buffer before it's sent. If something like this happen, a newer version of the buffer may be sent instead. - If this is a problem, it's recommended to call qio_channel_flush() before freeing or re-using the buffer. 2: Locked memory - When using MSG_ZERCOCOPY, the buffer memory will be locked after queued, and unlocked after it's sent. - Depending on the size of each buffer, and how often it's sent, it may require a larger amount of locked memory than usually available to non-root user. - If the required amount of locked memory is not available, writev_zero_copy will return an error, which can abort an operation like migration, - Because of this, when an user code wants to add zero copy as a feature, it requires a mechanism to disable it, so it can still be accessible to less privileged users. Signed-off-by: Leonardo Bras <leobras@redhat.com> Reviewed-by: Peter Xu <peterx@redhat.com> Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> Reviewed-by: Juan Quintela <quintela@redhat.com> Message-Id: <20220513062836.965425-4-leobras@redhat.com> Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
2022-05-13 08:28:32 +02:00
#ifdef QEMU_MSG_ZEROCOPY
static int qio_channel_socket_flush(QIOChannel *ioc,
Error **errp)
{
QIOChannelSocket *sioc = QIO_CHANNEL_SOCKET(ioc);
struct msghdr msg = {};
struct sock_extended_err *serr;
struct cmsghdr *cm;
char control[CMSG_SPACE(sizeof(*serr))];
int received;
int ret;
if (sioc->zero_copy_queued == sioc->zero_copy_sent) {
return 0;
}
QIOChannelSocket: Implement io_writev zero copy flag & io_flush for CONFIG_LINUX For CONFIG_LINUX, implement the new zero copy flag and the optional callback io_flush on QIOChannelSocket, but enables it only when MSG_ZEROCOPY feature is available in the host kernel, which is checked on qio_channel_socket_connect_sync() qio_channel_socket_flush() was implemented by counting how many times sendmsg(...,MSG_ZEROCOPY) was successfully called, and then reading the socket's error queue, in order to find how many of them finished sending. Flush will loop until those counters are the same, or until some error occurs. Notes on using writev() with QIO_CHANNEL_WRITE_FLAG_ZERO_COPY: 1: Buffer - As MSG_ZEROCOPY tells the kernel to use the same user buffer to avoid copying, some caution is necessary to avoid overwriting any buffer before it's sent. If something like this happen, a newer version of the buffer may be sent instead. - If this is a problem, it's recommended to call qio_channel_flush() before freeing or re-using the buffer. 2: Locked memory - When using MSG_ZERCOCOPY, the buffer memory will be locked after queued, and unlocked after it's sent. - Depending on the size of each buffer, and how often it's sent, it may require a larger amount of locked memory than usually available to non-root user. - If the required amount of locked memory is not available, writev_zero_copy will return an error, which can abort an operation like migration, - Because of this, when an user code wants to add zero copy as a feature, it requires a mechanism to disable it, so it can still be accessible to less privileged users. Signed-off-by: Leonardo Bras <leobras@redhat.com> Reviewed-by: Peter Xu <peterx@redhat.com> Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> Reviewed-by: Juan Quintela <quintela@redhat.com> Message-Id: <20220513062836.965425-4-leobras@redhat.com> Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
2022-05-13 08:28:32 +02:00
msg.msg_control = control;
msg.msg_controllen = sizeof(control);
memset(control, 0, sizeof(control));
ret = 1;
QIOChannelSocket: Implement io_writev zero copy flag & io_flush for CONFIG_LINUX For CONFIG_LINUX, implement the new zero copy flag and the optional callback io_flush on QIOChannelSocket, but enables it only when MSG_ZEROCOPY feature is available in the host kernel, which is checked on qio_channel_socket_connect_sync() qio_channel_socket_flush() was implemented by counting how many times sendmsg(...,MSG_ZEROCOPY) was successfully called, and then reading the socket's error queue, in order to find how many of them finished sending. Flush will loop until those counters are the same, or until some error occurs. Notes on using writev() with QIO_CHANNEL_WRITE_FLAG_ZERO_COPY: 1: Buffer - As MSG_ZEROCOPY tells the kernel to use the same user buffer to avoid copying, some caution is necessary to avoid overwriting any buffer before it's sent. If something like this happen, a newer version of the buffer may be sent instead. - If this is a problem, it's recommended to call qio_channel_flush() before freeing or re-using the buffer. 2: Locked memory - When using MSG_ZERCOCOPY, the buffer memory will be locked after queued, and unlocked after it's sent. - Depending on the size of each buffer, and how often it's sent, it may require a larger amount of locked memory than usually available to non-root user. - If the required amount of locked memory is not available, writev_zero_copy will return an error, which can abort an operation like migration, - Because of this, when an user code wants to add zero copy as a feature, it requires a mechanism to disable it, so it can still be accessible to less privileged users. Signed-off-by: Leonardo Bras <leobras@redhat.com> Reviewed-by: Peter Xu <peterx@redhat.com> Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> Reviewed-by: Juan Quintela <quintela@redhat.com> Message-Id: <20220513062836.965425-4-leobras@redhat.com> Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
2022-05-13 08:28:32 +02:00
while (sioc->zero_copy_sent < sioc->zero_copy_queued) {
received = recvmsg(sioc->fd, &msg, MSG_ERRQUEUE);
if (received < 0) {
switch (errno) {
case EAGAIN:
/* Nothing on errqueue, wait until something is available */
qio_channel_wait(ioc, G_IO_ERR);
continue;
case EINTR:
continue;
default:
error_setg_errno(errp, errno,
"Unable to read errqueue");
return -1;
}
}
cm = CMSG_FIRSTHDR(&msg);
if (cm->cmsg_level != SOL_IP && cm->cmsg_type != IP_RECVERR &&
cm->cmsg_level != SOL_IPV6 && cm->cmsg_type != IPV6_RECVERR) {
QIOChannelSocket: Implement io_writev zero copy flag & io_flush for CONFIG_LINUX For CONFIG_LINUX, implement the new zero copy flag and the optional callback io_flush on QIOChannelSocket, but enables it only when MSG_ZEROCOPY feature is available in the host kernel, which is checked on qio_channel_socket_connect_sync() qio_channel_socket_flush() was implemented by counting how many times sendmsg(...,MSG_ZEROCOPY) was successfully called, and then reading the socket's error queue, in order to find how many of them finished sending. Flush will loop until those counters are the same, or until some error occurs. Notes on using writev() with QIO_CHANNEL_WRITE_FLAG_ZERO_COPY: 1: Buffer - As MSG_ZEROCOPY tells the kernel to use the same user buffer to avoid copying, some caution is necessary to avoid overwriting any buffer before it's sent. If something like this happen, a newer version of the buffer may be sent instead. - If this is a problem, it's recommended to call qio_channel_flush() before freeing or re-using the buffer. 2: Locked memory - When using MSG_ZERCOCOPY, the buffer memory will be locked after queued, and unlocked after it's sent. - Depending on the size of each buffer, and how often it's sent, it may require a larger amount of locked memory than usually available to non-root user. - If the required amount of locked memory is not available, writev_zero_copy will return an error, which can abort an operation like migration, - Because of this, when an user code wants to add zero copy as a feature, it requires a mechanism to disable it, so it can still be accessible to less privileged users. Signed-off-by: Leonardo Bras <leobras@redhat.com> Reviewed-by: Peter Xu <peterx@redhat.com> Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> Reviewed-by: Juan Quintela <quintela@redhat.com> Message-Id: <20220513062836.965425-4-leobras@redhat.com> Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
2022-05-13 08:28:32 +02:00
error_setg_errno(errp, EPROTOTYPE,
"Wrong cmsg in errqueue");
return -1;
}
serr = (void *) CMSG_DATA(cm);
if (serr->ee_errno != SO_EE_ORIGIN_NONE) {
error_setg_errno(errp, serr->ee_errno,
"Error on socket");
return -1;
}
if (serr->ee_origin != SO_EE_ORIGIN_ZEROCOPY) {
error_setg_errno(errp, serr->ee_origin,
"Error not from zero copy");
return -1;
}
/* No errors, count successfully finished sendmsg()*/
sioc->zero_copy_sent += serr->ee_data - serr->ee_info + 1;
/* If any sendmsg() succeeded using zero copy, return 0 at the end */
if (serr->ee_code != SO_EE_CODE_ZEROCOPY_COPIED) {
ret = 0;
}
}
return ret;
}
#endif /* QEMU_MSG_ZEROCOPY */
static int
qio_channel_socket_set_blocking(QIOChannel *ioc,
bool enabled,
Error **errp)
{
QIOChannelSocket *sioc = QIO_CHANNEL_SOCKET(ioc);
if (enabled) {
qemu_socket_set_block(sioc->fd);
} else {
qemu_socket_set_nonblock(sioc->fd);
}
return 0;
}
static void
qio_channel_socket_set_delay(QIOChannel *ioc,
bool enabled)
{
QIOChannelSocket *sioc = QIO_CHANNEL_SOCKET(ioc);
int v = enabled ? 0 : 1;
setsockopt(sioc->fd,
IPPROTO_TCP, TCP_NODELAY,
&v, sizeof(v));
}
static void
qio_channel_socket_set_cork(QIOChannel *ioc,
bool enabled)
{
QIOChannelSocket *sioc = QIO_CHANNEL_SOCKET(ioc);
int v = enabled ? 1 : 0;
socket_set_cork(sioc->fd, v);
}
static int
qio_channel_socket_close(QIOChannel *ioc,
Error **errp)
{
QIOChannelSocket *sioc = QIO_CHANNEL_SOCKET(ioc);
int rc = 0;
Error *err = NULL;
if (sioc->fd != -1) {
#ifdef WIN32
qemu_socket_unselect(sioc->fd, NULL);
#endif
if (qio_channel_has_feature(ioc, QIO_CHANNEL_FEATURE_LISTEN)) {
socket_listen_cleanup(sioc->fd, errp);
}
if (close(sioc->fd) < 0) {
sioc->fd = -1;
error_setg_errno(&err, errno, "Unable to close socket");
error_propagate(errp, err);
return -1;
}
sioc->fd = -1;
}
return rc;
}
static int
qio_channel_socket_shutdown(QIOChannel *ioc,
QIOChannelShutdown how,
Error **errp)
{
QIOChannelSocket *sioc = QIO_CHANNEL_SOCKET(ioc);
int sockhow;
switch (how) {
case QIO_CHANNEL_SHUTDOWN_READ:
sockhow = SHUT_RD;
break;
case QIO_CHANNEL_SHUTDOWN_WRITE:
sockhow = SHUT_WR;
break;
case QIO_CHANNEL_SHUTDOWN_BOTH:
default:
sockhow = SHUT_RDWR;
break;
}
if (shutdown(sioc->fd, sockhow) < 0) {
error_setg_errno(errp, errno,
"Unable to shutdown socket");
return -1;
}
return 0;
}
static void qio_channel_socket_set_aio_fd_handler(QIOChannel *ioc,
io: follow coroutine AioContext in qio_channel_yield() The ongoing QEMU multi-queue block layer effort makes it possible for multiple threads to process I/O in parallel. The nbd block driver is not compatible with the multi-queue block layer yet because QIOChannel cannot be used easily from coroutines running in multiple threads. This series changes the QIOChannel API to make that possible. In the current API, calling qio_channel_attach_aio_context() sets the AioContext where qio_channel_yield() installs an fd handler prior to yielding: qio_channel_attach_aio_context(ioc, my_ctx); ... qio_channel_yield(ioc); // my_ctx is used here ... qio_channel_detach_aio_context(ioc); This API design has limitations: reading and writing must be done in the same AioContext and moving between AioContexts involves a cumbersome sequence of API calls that is not suitable for doing on a per-request basis. There is no fundamental reason why a QIOChannel needs to run within the same AioContext every time qio_channel_yield() is called. QIOChannel only uses the AioContext while inside qio_channel_yield(). The rest of the time, QIOChannel is independent of any AioContext. In the new API, qio_channel_yield() queries the AioContext from the current coroutine using qemu_coroutine_get_aio_context(). There is no need to explicitly attach/detach AioContexts anymore and qio_channel_attach_aio_context() and qio_channel_detach_aio_context() are gone. One coroutine can read from the QIOChannel while another coroutine writes from a different AioContext. This API change allows the nbd block driver to use QIOChannel from any thread. It's important to keep in mind that the block driver already synchronizes QIOChannel access and ensures that two coroutines never read simultaneously or write simultaneously. This patch updates all users of qio_channel_attach_aio_context() to the new API. Most conversions are simple, but vhost-user-server requires a new qemu_coroutine_yield() call to quiesce the vu_client_trip() coroutine when not attached to any AioContext. While the API is has become simpler, there is one wart: QIOChannel has a special case for the iohandler AioContext (used for handlers that must not run in nested event loops). I didn't find an elegant way preserve that behavior, so I added a new API called qio_channel_set_follow_coroutine_ctx(ioc, true|false) for opting in to the new AioContext model. By default QIOChannel uses the iohandler AioHandler. Code that formerly called qio_channel_attach_aio_context() now calls qio_channel_set_follow_coroutine_ctx(ioc, true) once after the QIOChannel is created. Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> Acked-by: Daniel P. Berrangé <berrange@redhat.com> Message-ID: <20230830224802.493686-5-stefanha@redhat.com> [eblake: also fix migration/rdma.c] Signed-off-by: Eric Blake <eblake@redhat.com>
2023-08-31 00:48:02 +02:00
AioContext *read_ctx,
IOHandler *io_read,
io: follow coroutine AioContext in qio_channel_yield() The ongoing QEMU multi-queue block layer effort makes it possible for multiple threads to process I/O in parallel. The nbd block driver is not compatible with the multi-queue block layer yet because QIOChannel cannot be used easily from coroutines running in multiple threads. This series changes the QIOChannel API to make that possible. In the current API, calling qio_channel_attach_aio_context() sets the AioContext where qio_channel_yield() installs an fd handler prior to yielding: qio_channel_attach_aio_context(ioc, my_ctx); ... qio_channel_yield(ioc); // my_ctx is used here ... qio_channel_detach_aio_context(ioc); This API design has limitations: reading and writing must be done in the same AioContext and moving between AioContexts involves a cumbersome sequence of API calls that is not suitable for doing on a per-request basis. There is no fundamental reason why a QIOChannel needs to run within the same AioContext every time qio_channel_yield() is called. QIOChannel only uses the AioContext while inside qio_channel_yield(). The rest of the time, QIOChannel is independent of any AioContext. In the new API, qio_channel_yield() queries the AioContext from the current coroutine using qemu_coroutine_get_aio_context(). There is no need to explicitly attach/detach AioContexts anymore and qio_channel_attach_aio_context() and qio_channel_detach_aio_context() are gone. One coroutine can read from the QIOChannel while another coroutine writes from a different AioContext. This API change allows the nbd block driver to use QIOChannel from any thread. It's important to keep in mind that the block driver already synchronizes QIOChannel access and ensures that two coroutines never read simultaneously or write simultaneously. This patch updates all users of qio_channel_attach_aio_context() to the new API. Most conversions are simple, but vhost-user-server requires a new qemu_coroutine_yield() call to quiesce the vu_client_trip() coroutine when not attached to any AioContext. While the API is has become simpler, there is one wart: QIOChannel has a special case for the iohandler AioContext (used for handlers that must not run in nested event loops). I didn't find an elegant way preserve that behavior, so I added a new API called qio_channel_set_follow_coroutine_ctx(ioc, true|false) for opting in to the new AioContext model. By default QIOChannel uses the iohandler AioHandler. Code that formerly called qio_channel_attach_aio_context() now calls qio_channel_set_follow_coroutine_ctx(ioc, true) once after the QIOChannel is created. Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> Acked-by: Daniel P. Berrangé <berrange@redhat.com> Message-ID: <20230830224802.493686-5-stefanha@redhat.com> [eblake: also fix migration/rdma.c] Signed-off-by: Eric Blake <eblake@redhat.com>
2023-08-31 00:48:02 +02:00
AioContext *write_ctx,
IOHandler *io_write,
void *opaque)
{
QIOChannelSocket *sioc = QIO_CHANNEL_SOCKET(ioc);
io: follow coroutine AioContext in qio_channel_yield() The ongoing QEMU multi-queue block layer effort makes it possible for multiple threads to process I/O in parallel. The nbd block driver is not compatible with the multi-queue block layer yet because QIOChannel cannot be used easily from coroutines running in multiple threads. This series changes the QIOChannel API to make that possible. In the current API, calling qio_channel_attach_aio_context() sets the AioContext where qio_channel_yield() installs an fd handler prior to yielding: qio_channel_attach_aio_context(ioc, my_ctx); ... qio_channel_yield(ioc); // my_ctx is used here ... qio_channel_detach_aio_context(ioc); This API design has limitations: reading and writing must be done in the same AioContext and moving between AioContexts involves a cumbersome sequence of API calls that is not suitable for doing on a per-request basis. There is no fundamental reason why a QIOChannel needs to run within the same AioContext every time qio_channel_yield() is called. QIOChannel only uses the AioContext while inside qio_channel_yield(). The rest of the time, QIOChannel is independent of any AioContext. In the new API, qio_channel_yield() queries the AioContext from the current coroutine using qemu_coroutine_get_aio_context(). There is no need to explicitly attach/detach AioContexts anymore and qio_channel_attach_aio_context() and qio_channel_detach_aio_context() are gone. One coroutine can read from the QIOChannel while another coroutine writes from a different AioContext. This API change allows the nbd block driver to use QIOChannel from any thread. It's important to keep in mind that the block driver already synchronizes QIOChannel access and ensures that two coroutines never read simultaneously or write simultaneously. This patch updates all users of qio_channel_attach_aio_context() to the new API. Most conversions are simple, but vhost-user-server requires a new qemu_coroutine_yield() call to quiesce the vu_client_trip() coroutine when not attached to any AioContext. While the API is has become simpler, there is one wart: QIOChannel has a special case for the iohandler AioContext (used for handlers that must not run in nested event loops). I didn't find an elegant way preserve that behavior, so I added a new API called qio_channel_set_follow_coroutine_ctx(ioc, true|false) for opting in to the new AioContext model. By default QIOChannel uses the iohandler AioHandler. Code that formerly called qio_channel_attach_aio_context() now calls qio_channel_set_follow_coroutine_ctx(ioc, true) once after the QIOChannel is created. Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com> Reviewed-by: Eric Blake <eblake@redhat.com> Acked-by: Daniel P. Berrangé <berrange@redhat.com> Message-ID: <20230830224802.493686-5-stefanha@redhat.com> [eblake: also fix migration/rdma.c] Signed-off-by: Eric Blake <eblake@redhat.com>
2023-08-31 00:48:02 +02:00
qio_channel_util_set_aio_fd_handler(sioc->fd, read_ctx, io_read,
sioc->fd, write_ctx, io_write,
opaque);
}
static GSource *qio_channel_socket_create_watch(QIOChannel *ioc,
GIOCondition condition)
{
QIOChannelSocket *sioc = QIO_CHANNEL_SOCKET(ioc);
return qio_channel_create_socket_watch(ioc,
sioc->fd,
condition);
}
static void qio_channel_socket_class_init(ObjectClass *klass,
void *class_data G_GNUC_UNUSED)
{
QIOChannelClass *ioc_klass = QIO_CHANNEL_CLASS(klass);
ioc_klass->io_writev = qio_channel_socket_writev;
ioc_klass->io_readv = qio_channel_socket_readv;
ioc_klass->io_set_blocking = qio_channel_socket_set_blocking;
ioc_klass->io_close = qio_channel_socket_close;
ioc_klass->io_shutdown = qio_channel_socket_shutdown;
ioc_klass->io_set_cork = qio_channel_socket_set_cork;
ioc_klass->io_set_delay = qio_channel_socket_set_delay;
ioc_klass->io_create_watch = qio_channel_socket_create_watch;
ioc_klass->io_set_aio_fd_handler = qio_channel_socket_set_aio_fd_handler;
QIOChannelSocket: Implement io_writev zero copy flag & io_flush for CONFIG_LINUX For CONFIG_LINUX, implement the new zero copy flag and the optional callback io_flush on QIOChannelSocket, but enables it only when MSG_ZEROCOPY feature is available in the host kernel, which is checked on qio_channel_socket_connect_sync() qio_channel_socket_flush() was implemented by counting how many times sendmsg(...,MSG_ZEROCOPY) was successfully called, and then reading the socket's error queue, in order to find how many of them finished sending. Flush will loop until those counters are the same, or until some error occurs. Notes on using writev() with QIO_CHANNEL_WRITE_FLAG_ZERO_COPY: 1: Buffer - As MSG_ZEROCOPY tells the kernel to use the same user buffer to avoid copying, some caution is necessary to avoid overwriting any buffer before it's sent. If something like this happen, a newer version of the buffer may be sent instead. - If this is a problem, it's recommended to call qio_channel_flush() before freeing or re-using the buffer. 2: Locked memory - When using MSG_ZERCOCOPY, the buffer memory will be locked after queued, and unlocked after it's sent. - Depending on the size of each buffer, and how often it's sent, it may require a larger amount of locked memory than usually available to non-root user. - If the required amount of locked memory is not available, writev_zero_copy will return an error, which can abort an operation like migration, - Because of this, when an user code wants to add zero copy as a feature, it requires a mechanism to disable it, so it can still be accessible to less privileged users. Signed-off-by: Leonardo Bras <leobras@redhat.com> Reviewed-by: Peter Xu <peterx@redhat.com> Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> Reviewed-by: Juan Quintela <quintela@redhat.com> Message-Id: <20220513062836.965425-4-leobras@redhat.com> Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
2022-05-13 08:28:32 +02:00
#ifdef QEMU_MSG_ZEROCOPY
ioc_klass->io_flush = qio_channel_socket_flush;
#endif
}
static const TypeInfo qio_channel_socket_info = {
.parent = TYPE_QIO_CHANNEL,
.name = TYPE_QIO_CHANNEL_SOCKET,
.instance_size = sizeof(QIOChannelSocket),
.instance_init = qio_channel_socket_init,
.instance_finalize = qio_channel_socket_finalize,
.class_init = qio_channel_socket_class_init,
};
static void qio_channel_socket_register_types(void)
{
type_register_static(&qio_channel_socket_info);
}
type_init(qio_channel_socket_register_types);