Commit 5a34dbb5 authored by Anthony Liguori's avatar Anthony Liguori
Browse files

Merge remote-tracking branch 'bonzini/threadpool' into staging 



* bonzini/threadpool: (39 commits)
  raw-win32: implement native asynchronous I/O
  raw-posix: move linux-aio.c to block/
  raw-win32: add emulated AIO support
  raw-posix: rename raw-posix-aio.h, hide unavailable prototypes
  raw: merge posix-aio-compat.c into block/raw-posix.c
  block: switch posix-aio-compat to threadpool
  threadpool: do not take lock in event_notifier_ready
  aio: add generic thread-pool facility
  qemu-thread: add QemuSemaphore
  linux-aio: use event notifiers
  aio: clean up now-unused functions
  main-loop: use aio_notify for qemu_notify_event
  main-loop: use GSource to poll AIO file descriptors
  aio: call aio_notify after setting I/O handlers
  aio: add aio_notify
  aio: make AioContexts GSources
  aio: add Win32 implementation
  aio: prepare for introducing GSource-based dispatch
  aio: add non-blocking variant of aio_wait
  aio: test node->deleted before calling io_flush
  ...

Signed-off-by: default avatarAnthony Liguori <aliguori@us.ibm.com>
parents 43552994 f563a5d7

Makefile

+3 −4
Original line number Diff line number Diff line
@@ -171,8 +171,7 @@ endif 
qemu-img.o: qemu-img-cmds.h



tools-obj-y = $(oslib-obj-y) $(trace-obj-y) qemu-tool.o qemu-timer.o \

	qemu-timer-common.o main-loop.o notify.o \

	iohandler.o cutils.o iov.o async.o error.o

	main-loop.o iohandler.o error.o

tools-obj-$(CONFIG_POSIX) += compatfd.o



qemu-img$(EXESUF): qemu-img.o $(tools-obj-y) $(block-obj-y)
@@ -181,7 +180,7 @@ qemu-io$(EXESUF): qemu-io.o cmd.o $(tools-obj-y) $(block-obj-y) 


qemu-bridge-helper$(EXESUF): qemu-bridge-helper.o



vscclient$(EXESUF): $(libcacard-y) $(oslib-obj-y) $(trace-obj-y) $(tools-obj-y) qemu-timer-common.o libcacard/vscclient.o

vscclient$(EXESUF): $(libcacard-y) $(oslib-obj-y) $(trace-obj-y) libcacard/vscclient.o

	$(call quiet-command,$(CC) $(LDFLAGS) -o $@ $^ $(libcacard_libs) $(LIBS),"  LINK  $@")



fsdev/virtfs-proxy-helper$(EXESUF): fsdev/virtfs-proxy-helper.o fsdev/virtio-9p-marshal.o oslib-posix.o $(trace-obj-y)
@@ -224,7 +223,7 @@ $(SRC_PATH)/qapi-schema.json $(SRC_PATH)/scripts/qapi-commands.py $(qapi-py) 
QGALIB_GEN=$(addprefix qga/qapi-generated/, qga-qapi-types.h qga-qapi-visit.h qga-qmp-commands.h)

$(qga-obj-y) qemu-ga.o: $(QGALIB_GEN)



qemu-ga$(EXESUF): qemu-ga.o $(qga-obj-y) $(tools-obj-y) $(qapi-obj-y) $(qobject-obj-y) $(version-obj-y)

qemu-ga$(EXESUF): qemu-ga.o $(qga-obj-y) $(oslib-obj-y) $(trace-obj-y) $(qapi-obj-y) $(qobject-obj-y) $(version-obj-y)



QEMULIBS=libuser libdis libdis-user

Makefile.objs

+10 −12
Original line number Diff line number Diff line
@@ -19,7 +19,7 @@ universal-obj-y += $(qom-obj-y) 


#######################################################################

# oslib-obj-y is code depending on the OS (win32 vs posix)

oslib-obj-y = osdep.o

oslib-obj-y = osdep.o cutils.o qemu-timer-common.o

oslib-obj-$(CONFIG_WIN32) += oslib-win32.o qemu-thread-win32.o

oslib-obj-$(CONFIG_POSIX) += oslib-posix.o qemu-thread-posix.o
@@ -41,12 +41,12 @@ coroutine-obj-$(CONFIG_WIN32) += coroutine-win32.o 
#######################################################################

# block-obj-y is code used by both qemu system emulation and qemu-img



block-obj-y = cutils.o iov.o cache-utils.o qemu-option.o module.o async.o

block-obj-y += nbd.o block.o blockjob.o aio.o aes.o qemu-config.o

block-obj-y += qemu-progress.o qemu-sockets.o uri.o notify.o

block-obj-y = iov.o cache-utils.o qemu-option.o module.o async.o

block-obj-y += nbd.o block.o blockjob.o aes.o qemu-config.o

block-obj-y += thread-pool.o qemu-progress.o qemu-sockets.o uri.o notify.o

block-obj-y += $(coroutine-obj-y) $(qobject-obj-y) $(version-obj-y)

block-obj-$(CONFIG_POSIX) += posix-aio-compat.o

block-obj-$(CONFIG_LINUX_AIO) += linux-aio.o

block-obj-$(CONFIG_POSIX) += event_notifier-posix.o aio-posix.o

block-obj-$(CONFIG_WIN32) += event_notifier-win32.o aio-win32.o

block-obj-y += block/

block-obj-y += $(qapi-obj-y) qapi-types.o qapi-visit.o
@@ -92,9 +92,8 @@ common-obj-y += ui/ 
common-obj-y += bt-host.o bt-vhci.o



common-obj-y += dma-helpers.o

common-obj-y += iov.o acl.o

common-obj-y += acl.o

common-obj-$(CONFIG_POSIX) += compatfd.o

common-obj-y += event_notifier.o

common-obj-y += qemu-timer.o qemu-timer-common.o

common-obj-y += qtest.o

common-obj-y += vl.o
@@ -113,7 +112,7 @@ endif 
user-obj-y =

user-obj-y += envlist.o path.o

user-obj-y += tcg-runtime.o host-utils.o

user-obj-y += cutils.o iov.o cache-utils.o

user-obj-y += cache-utils.o

user-obj-y += module.o

user-obj-y += qemu-user.o

user-obj-y += $(trace-obj-y)
@@ -228,9 +227,8 @@ universal-obj-y += $(qapi-obj-y) 
######################################################################

# guest agent



qga-obj-y = qga/ qemu-ga.o module.o

qga-obj-$(CONFIG_WIN32) += oslib-win32.o

qga-obj-$(CONFIG_POSIX) += oslib-posix.o qemu-sockets.o qemu-option.o

qga-obj-y = qga/ qemu-ga.o module.o qemu-tool.o

qga-obj-$(CONFIG_POSIX) += qemu-sockets.o qemu-option.o



vl.o: QEMU_CFLAGS+=$(GPROF_CFLAGS)

aio.c→aio-posix.c

+268 −0
Original line number Diff line number Diff line
@@ -18,20 +18,9 @@ 
#include "qemu-queue.h"

#include "qemu_socket.h"



typedef struct AioHandler AioHandler;



/* The list of registered AIO handlers */

static QLIST_HEAD(, AioHandler) aio_handlers;



/* This is a simple lock used to protect the aio_handlers list.  Specifically,

 * it's used to ensure that no callbacks are removed while we're walking and

 * dispatching callbacks.

 */

static int walking_handlers;



struct AioHandler

{

    int fd;

    GPollFD pfd;

    IOHandler *io_read;

    IOHandler *io_write;

    AioFlushHandler *io_flush;
@@ -40,12 +29,12 @@ struct AioHandler 
    QLIST_ENTRY(AioHandler) node;

};



static AioHandler *find_aio_handler(int fd)

static AioHandler *find_aio_handler(AioContext *ctx, int fd)

{

    AioHandler *node;



    QLIST_FOREACH(node, &aio_handlers, node) {

        if (node->fd == fd)

    QLIST_FOREACH(node, &ctx->aio_handlers, node) {

        if (node->pfd.fd == fd)

            if (!node->deleted)

                return node;

    }
@@ -53,7 +42,8 @@ static AioHandler *find_aio_handler(int fd) 
    return NULL;

}



int qemu_aio_set_fd_handler(int fd,

void aio_set_fd_handler(AioContext *ctx,

                        int fd,

                        IOHandler *io_read,

                        IOHandler *io_write,

                        AioFlushHandler *io_flush,
@@ -61,15 +51,18 @@ int qemu_aio_set_fd_handler(int fd, 
{

    AioHandler *node;



    node = find_aio_handler(fd);

    node = find_aio_handler(ctx, fd);



    /* Are we deleting the fd handler? */

    if (!io_read && !io_write) {

        if (node) {

            g_source_remove_poll(&ctx->source, &node->pfd);



            /* If the lock is held, just mark the node as deleted */

            if (walking_handlers)

            if (ctx->walking_handlers) {

                node->deleted = 1;

            else {

                node->pfd.revents = 0;

            } else {

                /* Otherwise, delete it for real.  We can't just mark it as

                 * deleted because deleted nodes are only cleaned up after

                 * releasing the walking_handlers lock.
@@ -82,113 +75,194 @@ int qemu_aio_set_fd_handler(int fd, 
        if (node == NULL) {

            /* Alloc and insert if it's not already there */

            node = g_malloc0(sizeof(AioHandler));

            node->fd = fd;

            QLIST_INSERT_HEAD(&aio_handlers, node, node);

            node->pfd.fd = fd;

            QLIST_INSERT_HEAD(&ctx->aio_handlers, node, node);



            g_source_add_poll(&ctx->source, &node->pfd);

        }

        /* Update handler with latest information */

        node->io_read = io_read;

        node->io_write = io_write;

        node->io_flush = io_flush;

        node->opaque = opaque;



        node->pfd.events = (io_read ? G_IO_IN | G_IO_HUP : 0);

        node->pfd.events |= (io_write ? G_IO_OUT : 0);

    }



    qemu_set_fd_handler2(fd, NULL, io_read, io_write, opaque);

    aio_notify(ctx);

}



    return 0;

void aio_set_event_notifier(AioContext *ctx,

                            EventNotifier *notifier,

                            EventNotifierHandler *io_read,

                            AioFlushEventNotifierHandler *io_flush)

{

    aio_set_fd_handler(ctx, event_notifier_get_fd(notifier),

                       (IOHandler *)io_read, NULL,

                       (AioFlushHandler *)io_flush, notifier);

}



void qemu_aio_flush(void)

bool aio_pending(AioContext *ctx)

{

    while (qemu_aio_wait());

    AioHandler *node;



    QLIST_FOREACH(node, &ctx->aio_handlers, node) {

        int revents;



        /*

         * FIXME: right now we cannot get G_IO_HUP and G_IO_ERR because

         * main-loop.c is still select based (due to the slirp legacy).

         * If main-loop.c ever switches to poll, G_IO_ERR should be

         * tested too.  Dispatching G_IO_ERR to both handlers should be

         * okay, since handlers need to be ready for spurious wakeups.

         */

        revents = node->pfd.revents & node->pfd.events;

        if (revents & (G_IO_IN | G_IO_HUP | G_IO_ERR) && node->io_read) {

            return true;

        }

        if (revents & (G_IO_OUT | G_IO_ERR) && node->io_write) {

            return true;

        }

    }



bool qemu_aio_wait(void)

    return false;

}



bool aio_poll(AioContext *ctx, bool blocking)

{

    static struct timeval tv0;

    AioHandler *node;

    fd_set rdfds, wrfds;

    int max_fd = -1;

    int ret;

    bool busy;

    bool busy, progress;



    progress = false;



    /*

     * If there are callbacks left that have been queued, we need to call then.

     * Do not call select in this case, because it is possible that the caller

     * does not need a complete flush (as is the case for qemu_aio_wait loops).

     */

    if (qemu_bh_poll()) {

    if (aio_bh_poll(ctx)) {

        blocking = false;

        progress = true;

    }



    /*

     * Then dispatch any pending callbacks from the GSource.

     *

     * We have to walk very carefully in case qemu_aio_set_fd_handler is

     * called while we're walking.

     */

    node = QLIST_FIRST(&ctx->aio_handlers);

    while (node) {

        AioHandler *tmp;

        int revents;



        ctx->walking_handlers++;



        revents = node->pfd.revents & node->pfd.events;

        node->pfd.revents = 0;



        /* See comment in aio_pending.  */

        if (revents & (G_IO_IN | G_IO_HUP | G_IO_ERR) && node->io_read) {

            node->io_read(node->opaque);

            progress = true;

        }

        if (revents & (G_IO_OUT | G_IO_ERR) && node->io_write) {

            node->io_write(node->opaque);

            progress = true;

        }



        tmp = node;

        node = QLIST_NEXT(node, node);



        ctx->walking_handlers--;



        if (!ctx->walking_handlers && tmp->deleted) {

            QLIST_REMOVE(tmp, node);

            g_free(tmp);

        }

    }



    if (progress && !blocking) {

        return true;

    }



    walking_handlers++;

    ctx->walking_handlers++;



    FD_ZERO(&rdfds);

    FD_ZERO(&wrfds);



    /* fill fd sets */

    busy = false;

    QLIST_FOREACH(node, &aio_handlers, node) {

    QLIST_FOREACH(node, &ctx->aio_handlers, node) {

        /* If there aren't pending AIO operations, don't invoke callbacks.

         * Otherwise, if there are no AIO requests, qemu_aio_wait() would

         * wait indefinitely.

         */

        if (node->io_flush) {

        if (!node->deleted && node->io_flush) {

            if (node->io_flush(node->opaque) == 0) {

                continue;

            }

            busy = true;

        }

        if (!node->deleted && node->io_read) {

            FD_SET(node->fd, &rdfds);

            max_fd = MAX(max_fd, node->fd + 1);

            FD_SET(node->pfd.fd, &rdfds);

            max_fd = MAX(max_fd, node->pfd.fd + 1);

        }

        if (!node->deleted && node->io_write) {

            FD_SET(node->fd, &wrfds);

            max_fd = MAX(max_fd, node->fd + 1);

            FD_SET(node->pfd.fd, &wrfds);

            max_fd = MAX(max_fd, node->pfd.fd + 1);

        }

    }



    walking_handlers--;

    ctx->walking_handlers--;



    /* No AIO operations?  Get us out of here */

    if (!busy) {

        return false;

        return progress;

    }



    /* wait until next event */

    ret = select(max_fd, &rdfds, &wrfds, NULL, NULL);

    ret = select(max_fd, &rdfds, &wrfds, NULL, blocking ? NULL : &tv0);



    /* if we have any readable fds, dispatch event */

    if (ret > 0) {

        /* we have to walk very carefully in case

         * qemu_aio_set_fd_handler is called while we're walking */

        node = QLIST_FIRST(&aio_handlers);

        node = QLIST_FIRST(&ctx->aio_handlers);

        while (node) {

            AioHandler *tmp;



            walking_handlers++;

            ctx->walking_handlers++;



            if (!node->deleted &&

                FD_ISSET(node->fd, &rdfds) &&

                FD_ISSET(node->pfd.fd, &rdfds) &&

                node->io_read) {

                node->io_read(node->opaque);

                progress = true;

            }

            if (!node->deleted &&

                FD_ISSET(node->fd, &wrfds) &&

                FD_ISSET(node->pfd.fd, &wrfds) &&

                node->io_write) {

                node->io_write(node->opaque);

                progress = true;

            }



            tmp = node;

            node = QLIST_NEXT(node, node);



            walking_handlers--;

            ctx->walking_handlers--;



            if (!walking_handlers && tmp->deleted) {

            if (!ctx->walking_handlers && tmp->deleted) {

                QLIST_REMOVE(tmp, node);

                g_free(tmp);

            }

        }

    }



    return true;

    return progress;

}

aio-win32.c

0 → 100644
+215 −0
Original line number Diff line number Diff line 
/*

 * QEMU aio implementation

 *

 * Copyright IBM Corp., 2008

 * Copyright Red Hat Inc., 2012

 *

 * Authors:

 *  Anthony Liguori   <aliguori@us.ibm.com>

 *  Paolo Bonzini     <pbonzini@redhat.com>

 *

 * This work is licensed under the terms of the GNU GPL, version 2.  See

 * the COPYING file in the top-level directory.

 *

 * Contributions after 2012-01-13 are licensed under the terms of the

 * GNU GPL, version 2 or (at your option) any later version.

 */



#include "qemu-common.h"

#include "block.h"

#include "qemu-queue.h"

#include "qemu_socket.h"



struct AioHandler {

    EventNotifier *e;

    EventNotifierHandler *io_notify;

    AioFlushEventNotifierHandler *io_flush;

    GPollFD pfd;

    int deleted;

    QLIST_ENTRY(AioHandler) node;

};



void aio_set_event_notifier(AioContext *ctx,

                            EventNotifier *e,

                            EventNotifierHandler *io_notify,

                            AioFlushEventNotifierHandler *io_flush)

{

    AioHandler *node;



    QLIST_FOREACH(node, &ctx->aio_handlers, node) {

        if (node->e == e && !node->deleted) {

            break;

        }

    }



    /* Are we deleting the fd handler? */

    if (!io_notify) {

        if (node) {

            g_source_remove_poll(&ctx->source, &node->pfd);



            /* If the lock is held, just mark the node as deleted */

            if (ctx->walking_handlers) {

                node->deleted = 1;

                node->pfd.revents = 0;

            } else {

                /* Otherwise, delete it for real.  We can't just mark it as

                 * deleted because deleted nodes are only cleaned up after

                 * releasing the walking_handlers lock.

                 */

                QLIST_REMOVE(node, node);

                g_free(node);

            }

        }

    } else {

        if (node == NULL) {

            /* Alloc and insert if it's not already there */

            node = g_malloc0(sizeof(AioHandler));

            node->e = e;

            node->pfd.fd = (uintptr_t)event_notifier_get_handle(e);

            node->pfd.events = G_IO_IN;

            QLIST_INSERT_HEAD(&ctx->aio_handlers, node, node);



            g_source_add_poll(&ctx->source, &node->pfd);

        }

        /* Update handler with latest information */

        node->io_notify = io_notify;

        node->io_flush = io_flush;

    }



    aio_notify(ctx);

}



bool aio_pending(AioContext *ctx)

{

    AioHandler *node;



    QLIST_FOREACH(node, &ctx->aio_handlers, node) {

        if (node->pfd.revents && node->io_notify) {

            return true;

        }

    }



    return false;

}



bool aio_poll(AioContext *ctx, bool blocking)

{

    AioHandler *node;

    HANDLE events[MAXIMUM_WAIT_OBJECTS + 1];

    bool busy, progress;

    int count;



    progress = false;



    /*

     * If there are callbacks left that have been queued, we need to call then.

     * Do not call select in this case, because it is possible that the caller

     * does not need a complete flush (as is the case for qemu_aio_wait loops).

     */

    if (aio_bh_poll(ctx)) {

        blocking = false;

        progress = true;

    }



    /*

     * Then dispatch any pending callbacks from the GSource.

     *

     * We have to walk very carefully in case qemu_aio_set_fd_handler is

     * called while we're walking.

     */

    node = QLIST_FIRST(&ctx->aio_handlers);

    while (node) {

        AioHandler *tmp;



        ctx->walking_handlers++;



        if (node->pfd.revents && node->io_notify) {

            node->pfd.revents = 0;

            node->io_notify(node->e);

            progress = true;

        }



        tmp = node;

        node = QLIST_NEXT(node, node);



        ctx->walking_handlers--;



        if (!ctx->walking_handlers && tmp->deleted) {

            QLIST_REMOVE(tmp, node);

            g_free(tmp);

        }

    }



    if (progress && !blocking) {

        return true;

    }



    ctx->walking_handlers++;



    /* fill fd sets */

    busy = false;

    count = 0;

    QLIST_FOREACH(node, &ctx->aio_handlers, node) {

        /* If there aren't pending AIO operations, don't invoke callbacks.

         * Otherwise, if there are no AIO requests, qemu_aio_wait() would

         * wait indefinitely.

         */

        if (!node->deleted && node->io_flush) {

            if (node->io_flush(node->e) == 0) {

                continue;

            }

            busy = true;

        }

        if (!node->deleted && node->io_notify) {

            events[count++] = event_notifier_get_handle(node->e);

        }

    }



    ctx->walking_handlers--;



    /* No AIO operations?  Get us out of here */

    if (!busy) {

        return progress;

    }



    /* wait until next event */

    for (;;) {

        int timeout = blocking ? INFINITE : 0;

        int ret = WaitForMultipleObjects(count, events, FALSE, timeout);



        /* if we have any signaled events, dispatch event */

        if ((DWORD) (ret - WAIT_OBJECT_0) >= count) {

            break;

        }



        blocking = false;



        /* we have to walk very carefully in case

         * qemu_aio_set_fd_handler is called while we're walking */

        node = QLIST_FIRST(&ctx->aio_handlers);

        while (node) {

            AioHandler *tmp;



            ctx->walking_handlers++;



            if (!node->deleted &&

                event_notifier_get_handle(node->e) == events[ret - WAIT_OBJECT_0] &&

                node->io_notify) {

                node->io_notify(node->e);

                progress = true;

            }



            tmp = node;

            node = QLIST_NEXT(node, node);



            ctx->walking_handlers--;



            if (!ctx->walking_handlers && tmp->deleted) {

                QLIST_REMOVE(tmp, node);

                g_free(tmp);

            }

        }

    }



    return progress;

}

arch_init.h

+1 −1
Original line number Diff line number Diff line
@@ -34,6 +34,6 @@ int tcg_available(void); 
int kvm_available(void);

int xen_available(void);



CpuDefinitionInfoList GCC_WEAK_DECL *arch_query_cpu_definitions(Error **errp);

CpuDefinitionInfoList *arch_query_cpu_definitions(Error **errp);



#endif