qemu-e2k/io/channel-command.c

416 lines
12 KiB
C
Raw Normal View History

/*
* QEMU I/O channels external command 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"
#include "io/channel-command.h"
#include "io/channel-watch.h"
2016-03-14 09:01:28 +01:00
#include "qapi/error.h"
#include "qemu/module.h"
#include "qemu/sockets.h"
#include "trace.h"
#ifndef WIN32
/**
* qio_channel_command_new_pid:
* @writefd: the FD connected to the command's stdin
* @readfd: the FD connected to the command's stdout
* @pid: the PID of the running child command
* @errp: pointer to a NULL-initialized error object
*
* Create a channel for performing I/O with the
* previously spawned command identified by @pid.
* The two file descriptors provide the connection
* to command's stdio streams, either one or which
* may be -1 to indicate that stream is not open.
*
* The channel will take ownership of the process
* @pid and will kill it when closing the channel.
* Similarly it will take responsibility for
* closing the file descriptors @writefd and @readfd.
*
* Returns: the command channel object, or NULL on error
*/
static QIOChannelCommand *
qio_channel_command_new_pid(int writefd,
int readfd,
pid_t pid)
{
QIOChannelCommand *ioc;
ioc = QIO_CHANNEL_COMMAND(object_new(TYPE_QIO_CHANNEL_COMMAND));
ioc->readfd = readfd;
ioc->writefd = writefd;
ioc->pid = pid;
trace_qio_channel_command_new_pid(ioc, writefd, readfd, pid);
return ioc;
}
QIOChannelCommand *
qio_channel_command_new_spawn(const char *const argv[],
int flags,
Error **errp)
{
pid_t pid = -1;
int stdinfd[2] = { -1, -1 };
int stdoutfd[2] = { -1, -1 };
int devnull = -1;
bool stdinnull = false, stdoutnull = false;
QIOChannelCommand *ioc;
flags = flags & O_ACCMODE;
if (flags == O_RDONLY) {
stdinnull = true;
}
if (flags == O_WRONLY) {
stdoutnull = true;
}
if (stdinnull || stdoutnull) {
devnull = open("/dev/null", O_RDWR);
if (devnull < 0) {
error_setg_errno(errp, errno,
"Unable to open /dev/null");
goto error;
}
}
if ((!stdinnull && !g_unix_open_pipe(stdinfd, FD_CLOEXEC, NULL)) ||
(!stdoutnull && !g_unix_open_pipe(stdoutfd, FD_CLOEXEC, NULL))) {
error_setg_errno(errp, errno,
"Unable to open pipe");
goto error;
}
pid = qemu_fork(errp);
if (pid < 0) {
goto error;
}
if (pid == 0) { /* child */
dup2(stdinnull ? devnull : stdinfd[0], STDIN_FILENO);
dup2(stdoutnull ? devnull : stdoutfd[1], STDOUT_FILENO);
/* Leave stderr connected to qemu's stderr */
if (!stdinnull) {
close(stdinfd[0]);
close(stdinfd[1]);
}
if (!stdoutnull) {
close(stdoutfd[0]);
close(stdoutfd[1]);
}
if (devnull != -1) {
close(devnull);
}
execv(argv[0], (char * const *)argv);
_exit(1);
}
if (!stdinnull) {
close(stdinfd[0]);
}
if (!stdoutnull) {
close(stdoutfd[1]);
}
ioc = qio_channel_command_new_pid(stdinnull ? devnull : stdinfd[1],
stdoutnull ? devnull : stdoutfd[0],
pid);
trace_qio_channel_command_new_spawn(ioc, argv[0], flags);
return ioc;
error:
if (devnull != -1) {
close(devnull);
}
if (stdinfd[0] != -1) {
close(stdinfd[0]);
}
if (stdinfd[1] != -1) {
close(stdinfd[1]);
}
if (stdoutfd[0] != -1) {
close(stdoutfd[0]);
}
if (stdoutfd[1] != -1) {
close(stdoutfd[1]);
}
return NULL;
}
#else /* WIN32 */
QIOChannelCommand *
qio_channel_command_new_spawn(const char *const argv[],
int flags,
Error **errp)
{
error_setg_errno(errp, ENOSYS,
"Command spawn not supported on this platform");
return NULL;
}
#endif /* WIN32 */
#ifndef WIN32
static int qio_channel_command_abort(QIOChannelCommand *ioc,
Error **errp)
{
pid_t ret;
int status;
int step = 0;
/* See if intermediate process has exited; if not, try a nice
* SIGTERM followed by a more severe SIGKILL.
*/
rewait:
trace_qio_channel_command_abort(ioc, ioc->pid);
ret = waitpid(ioc->pid, &status, WNOHANG);
trace_qio_channel_command_wait(ioc, ioc->pid, ret, status);
if (ret == (pid_t)-1) {
if (errno == EINTR) {
goto rewait;
} else {
error_setg_errno(errp, errno,
"Cannot wait on pid %llu",
(unsigned long long)ioc->pid);
return -1;
}
} else if (ret == 0) {
if (step == 0) {
kill(ioc->pid, SIGTERM);
} else if (step == 1) {
kill(ioc->pid, SIGKILL);
} else {
error_setg(errp,
"Process %llu refused to die",
(unsigned long long)ioc->pid);
return -1;
}
step++;
usleep(10 * 1000);
goto rewait;
}
return 0;
}
#endif /* ! WIN32 */
static void qio_channel_command_init(Object *obj)
{
QIOChannelCommand *ioc = QIO_CHANNEL_COMMAND(obj);
ioc->readfd = -1;
ioc->writefd = -1;
ioc->pid = -1;
}
static void qio_channel_command_finalize(Object *obj)
{
QIOChannelCommand *ioc = QIO_CHANNEL_COMMAND(obj);
if (ioc->readfd != -1) {
close(ioc->readfd);
}
if (ioc->writefd != -1 &&
ioc->writefd != ioc->readfd) {
close(ioc->writefd);
}
ioc->writefd = ioc->readfd = -1;
if (ioc->pid > 0) {
#ifndef WIN32
qio_channel_command_abort(ioc, NULL);
#endif
}
}
static ssize_t qio_channel_command_readv(QIOChannel *ioc,
const struct iovec *iov,
size_t niov,
int **fds,
size_t *nfds,
Error **errp)
{
QIOChannelCommand *cioc = QIO_CHANNEL_COMMAND(ioc);
ssize_t ret;
retry:
ret = readv(cioc->readfd, iov, niov);
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 command");
return -1;
}
return ret;
}
static ssize_t qio_channel_command_writev(QIOChannel *ioc,
const struct iovec *iov,
size_t niov,
int *fds,
size_t nfds,
int flags,
Error **errp)
{
QIOChannelCommand *cioc = QIO_CHANNEL_COMMAND(ioc);
ssize_t ret;
retry:
ret = writev(cioc->writefd, iov, niov);
if (ret <= 0) {
if (errno == EAGAIN) {
return QIO_CHANNEL_ERR_BLOCK;
}
if (errno == EINTR) {
goto retry;
}
error_setg_errno(errp, errno, "%s",
"Unable to write to command");
return -1;
}
return ret;
}
static int qio_channel_command_set_blocking(QIOChannel *ioc,
bool enabled,
Error **errp)
{
#ifdef WIN32
/* command spawn is not supported on win32 */
g_assert_not_reached();
#else
QIOChannelCommand *cioc = QIO_CHANNEL_COMMAND(ioc);
if (!g_unix_set_fd_nonblocking(cioc->writefd, !enabled, NULL) ||
!g_unix_set_fd_nonblocking(cioc->readfd, !enabled, NULL)) {
error_setg_errno(errp, errno, "Failed to set FD nonblocking");
return -1;
}
#endif
return 0;
}
static int qio_channel_command_close(QIOChannel *ioc,
Error **errp)
{
QIOChannelCommand *cioc = QIO_CHANNEL_COMMAND(ioc);
int rv = 0;
#ifndef WIN32
pid_t wp;
#endif
/* We close FDs before killing, because that
* gives a better chance of clean shutdown
*/
if (cioc->readfd != -1 &&
close(cioc->readfd) < 0) {
rv = -1;
}
if (cioc->writefd != -1 &&
cioc->writefd != cioc->readfd &&
close(cioc->writefd) < 0) {
rv = -1;
}
cioc->writefd = cioc->readfd = -1;
#ifndef WIN32
do {
wp = waitpid(cioc->pid, NULL, 0);
} while (wp == (pid_t)-1 && errno == EINTR);
if (wp == (pid_t)-1) {
error_setg_errno(errp, errno, "Failed to wait for pid %llu",
(unsigned long long)cioc->pid);
return -1;
}
#endif
if (rv < 0) {
error_setg_errno(errp, errno, "%s",
"Unable to close command");
}
return rv;
}
static void qio_channel_command_set_aio_fd_handler(QIOChannel *ioc,
AioContext *ctx,
IOHandler *io_read,
IOHandler *io_write,
void *opaque)
{
QIOChannelCommand *cioc = QIO_CHANNEL_COMMAND(ioc);
aio-posix: split poll check from ready handler Adaptive polling measures the execution time of the polling check plus handlers called when a polled event becomes ready. Handlers can take a significant amount of time, making it look like polling was running for a long time when in fact the event handler was running for a long time. For example, on Linux the io_submit(2) syscall invoked when a virtio-blk device's virtqueue becomes ready can take 10s of microseconds. This can exceed the default polling interval (32 microseconds) and cause adaptive polling to stop polling. By excluding the handler's execution time from the polling check we make the adaptive polling calculation more accurate. As a result, the event loop now stays in polling mode where previously it would have fallen back to file descriptor monitoring. The following data was collected with virtio-blk num-queues=2 event_idx=off using an IOThread. Before: 168k IOPS, IOThread syscalls: 9837.115 ( 0.020 ms): IO iothread1/620155 io_submit(ctx_id: 140512552468480, nr: 16, iocbpp: 0x7fcb9f937db0) = 16 9837.158 ( 0.002 ms): IO iothread1/620155 write(fd: 103, buf: 0x556a2ef71b88, count: 8) = 8 9837.161 ( 0.001 ms): IO iothread1/620155 write(fd: 104, buf: 0x556a2ef71b88, count: 8) = 8 9837.163 ( 0.001 ms): IO iothread1/620155 ppoll(ufds: 0x7fcb90002800, nfds: 4, tsp: 0x7fcb9f1342d0, sigsetsize: 8) = 3 9837.164 ( 0.001 ms): IO iothread1/620155 read(fd: 107, buf: 0x7fcb9f939cc0, count: 512) = 8 9837.174 ( 0.001 ms): IO iothread1/620155 read(fd: 105, buf: 0x7fcb9f939cc0, count: 512) = 8 9837.176 ( 0.001 ms): IO iothread1/620155 read(fd: 106, buf: 0x7fcb9f939cc0, count: 512) = 8 9837.209 ( 0.035 ms): IO iothread1/620155 io_submit(ctx_id: 140512552468480, nr: 32, iocbpp: 0x7fca7d0cebe0) = 32 174k IOPS (+3.6%), IOThread syscalls: 9809.566 ( 0.036 ms): IO iothread1/623061 io_submit(ctx_id: 140539805028352, nr: 32, iocbpp: 0x7fd0cdd62be0) = 32 9809.625 ( 0.001 ms): IO iothread1/623061 write(fd: 103, buf: 0x5647cfba5f58, count: 8) = 8 9809.627 ( 0.002 ms): IO iothread1/623061 write(fd: 104, buf: 0x5647cfba5f58, count: 8) = 8 9809.663 ( 0.036 ms): IO iothread1/623061 io_submit(ctx_id: 140539805028352, nr: 32, iocbpp: 0x7fd0d0388b50) = 32 Notice that ppoll(2) and eventfd read(2) syscalls are eliminated because the IOThread stays in polling mode instead of falling back to file descriptor monitoring. As usual, polling is not implemented on Windows so this patch ignores the new io_poll_read() callback in aio-win32.c. Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com> Reviewed-by: Stefano Garzarella <sgarzare@redhat.com> Message-id: 20211207132336.36627-2-stefanha@redhat.com [Fixed up aio_set_event_notifier() calls in tests/unit/test-fdmon-epoll.c added after this series was queued. --Stefan] Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
2021-12-07 14:23:31 +01:00
aio_set_fd_handler(ctx, cioc->readfd, false,
io_read, NULL, NULL, NULL, opaque);
aio_set_fd_handler(ctx, cioc->writefd, false,
NULL, io_write, NULL, NULL, opaque);
}
static GSource *qio_channel_command_create_watch(QIOChannel *ioc,
GIOCondition condition)
{
QIOChannelCommand *cioc = QIO_CHANNEL_COMMAND(ioc);
return qio_channel_create_fd_pair_watch(ioc,
cioc->readfd,
cioc->writefd,
condition);
}
static void qio_channel_command_class_init(ObjectClass *klass,
void *class_data G_GNUC_UNUSED)
{
QIOChannelClass *ioc_klass = QIO_CHANNEL_CLASS(klass);
ioc_klass->io_writev = qio_channel_command_writev;
ioc_klass->io_readv = qio_channel_command_readv;
ioc_klass->io_set_blocking = qio_channel_command_set_blocking;
ioc_klass->io_close = qio_channel_command_close;
ioc_klass->io_create_watch = qio_channel_command_create_watch;
ioc_klass->io_set_aio_fd_handler = qio_channel_command_set_aio_fd_handler;
}
static const TypeInfo qio_channel_command_info = {
.parent = TYPE_QIO_CHANNEL,
.name = TYPE_QIO_CHANNEL_COMMAND,
.instance_size = sizeof(QIOChannelCommand),
.instance_init = qio_channel_command_init,
.instance_finalize = qio_channel_command_finalize,
.class_init = qio_channel_command_class_init,
};
static void qio_channel_command_register_types(void)
{
type_register_static(&qio_channel_command_info);
}
type_init(qio_channel_command_register_types);