[Qemu-devel] [PATCH 2/5] Move AIO

[Laurent Vivier <Laurent.Vivier@bull.net>]

This patche moves raw AIO part from block-raw-posix.c to qemu-aio-raw.c to
be able to use raw AIO from other files.

Laurent
---
 Makefile          |    2 
 Makefile.target   |    2 
 block-raw-posix.c |  205 ++-------------------------------------------
 qemu-aio-posix.c  |  241 ++++++++++++++++++++++++++++++++++++++++++++++++++++++
 qemu-aio-posix.h  |   39 ++++++++
 5 files changed, 293 insertions(+), 196 deletions(-)

Index: qemu/Makefile.target
===================================================================
--- qemu.orig/Makefile.target	2008-01-23 09:18:17.000000000 +0100
+++ qemu/Makefile.target	2008-01-23 09:19:30.000000000 +0100
@@ -398,7 +398,7 @@ VL_OBJS=vl.o osdep.o monitor.o pci.o loa
 ifdef CONFIG_WIN32
 VL_OBJS+=block-raw-win32.o
 else
-VL_OBJS+=block-raw-posix.o
+VL_OBJS+=block-raw-posix.o qemu-aio-posix.o
 endif
 
 ifdef CONFIG_ALSA
Index: qemu/block-raw-posix.c
===================================================================
--- qemu.orig/block-raw-posix.c	2008-01-23 09:19:16.000000000 +0100
+++ qemu/block-raw-posix.c	2008-01-23 09:19:30.000000000 +0100
@@ -28,7 +28,7 @@
 #endif
 #include "block_int.h"
 #include <assert.h>
-#include <aio.h>
+#include "qemu-aio-posix.h"
 
 #ifdef CONFIG_COCOA
 #include <paths.h>
@@ -75,7 +75,7 @@
 #define FD_OPEN_TIMEOUT 1000
 
 typedef struct BDRVRawState {
-    int fd;
+    int fd;	/* must be the first field for qemu-aio-posix.c */
     int type;
     unsigned int lseek_err_cnt;
 #if defined(__linux__)
@@ -233,180 +233,18 @@ label__raw_write__success:
 /***********************************************************/
 /* Unix AIO using POSIX AIO */
 
-typedef struct RawAIOCB {
-    BlockDriverAIOCB common;
-    struct aiocb aiocb;
-    struct RawAIOCB *next;
-} RawAIOCB;
-
-static int aio_sig_num = SIGUSR2;
-static RawAIOCB *first_aio; /* AIO issued */
-static int aio_initialized = 0;
-
-static void aio_signal_handler(int signum)
-{
-#ifndef QEMU_IMG
-    CPUState *env = cpu_single_env;
-    if (env) {
-        /* stop the currently executing cpu because a timer occured */
-        cpu_interrupt(env, CPU_INTERRUPT_EXIT);
-#ifdef USE_KQEMU
-        if (env->kqemu_enabled) {
-            kqemu_cpu_interrupt(env);
-        }
-#endif
-    }
-#endif
-}
-
-void qemu_aio_init(void)
-{
-    struct sigaction act;
-
-    aio_initialized = 1;
-
-    sigfillset(&act.sa_mask);
-    act.sa_flags = 0; /* do not restart syscalls to interrupt select() */
-    act.sa_handler = aio_signal_handler;
-    sigaction(aio_sig_num, &act, NULL);
-
-#if defined(__GLIBC__) && defined(__linux__)
-    {
-        /* XXX: aio thread exit seems to hang on RedHat 9 and this init
-           seems to fix the problem. */
-        struct aioinit ai;
-        memset(&ai, 0, sizeof(ai));
-        ai.aio_threads = 1;
-        ai.aio_num = 1;
-        ai.aio_idle_time = 365 * 100000;
-        aio_init(&ai);
-    }
-#endif
-}
-
-void qemu_aio_poll(void)
-{
-    RawAIOCB *acb, **pacb;
-    int ret;
-
-    for(;;) {
-        pacb = &first_aio;
-        for(;;) {
-            acb = *pacb;
-            if (!acb)
-                goto the_end;
-            ret = aio_error(&acb->aiocb);
-            if (ret == ECANCELED) {
-                /* remove the request */
-                *pacb = acb->next;
-                qemu_aio_release(acb);
-            } else if (ret != EINPROGRESS) {
-                /* end of aio */
-                if (ret == 0) {
-                    ret = aio_return(&acb->aiocb);
-                    if (ret == acb->aiocb.aio_nbytes)
-                        ret = 0;
-                    else
-                        ret = -EINVAL;
-                } else {
-                    ret = -ret;
-                }
-                /* remove the request */
-                *pacb = acb->next;
-                /* call the callback */
-                acb->common.cb(acb->common.opaque, ret);
-                qemu_aio_release(acb);
-                break;
-            } else {
-                pacb = &acb->next;
-            }
-        }
-    }
- the_end: ;
-}
-
-/* Wait for all IO requests to complete.  */
-void qemu_aio_flush(void)
-{
-    qemu_aio_wait_start();
-    qemu_aio_poll();
-    while (first_aio) {
-        qemu_aio_wait();
-    }
-    qemu_aio_wait_end();
-}
-
-/* wait until at least one AIO was handled */
-static sigset_t wait_oset;
-
-void qemu_aio_wait_start(void)
-{
-    sigset_t set;
-
-    if (!aio_initialized)
-        qemu_aio_init();
-    sigemptyset(&set);
-    sigaddset(&set, aio_sig_num);
-    sigprocmask(SIG_BLOCK, &set, &wait_oset);
-}
-
-void qemu_aio_wait(void)
-{
-    sigset_t set;
-    int nb_sigs;
-
-#ifndef QEMU_IMG
-    if (qemu_bh_poll())
-        return;
-#endif
-    sigemptyset(&set);
-    sigaddset(&set, aio_sig_num);
-    sigwait(&set, &nb_sigs);
-    qemu_aio_poll();
-}
-
-void qemu_aio_wait_end(void)
-{
-    sigprocmask(SIG_SETMASK, &wait_oset, NULL);
-}
-
-static RawAIOCB *raw_aio_setup(BlockDriverState *bs,
+static BlockDriverAIOCB *raw_aio_read(BlockDriverState *bs,
         int64_t sector_num, uint8_t *buf, int nb_sectors,
         BlockDriverCompletionFunc *cb, void *opaque)
 {
-    BDRVRawState *s = bs->opaque;
     RawAIOCB *acb;
 
     if (fd_open(bs) < 0)
         return NULL;
 
-    acb = qemu_aio_get(bs, cb, opaque);
-    if (!acb)
-        return NULL;
-    acb->aiocb.aio_fildes = s->fd;
-    acb->aiocb.aio_sigevent.sigev_signo = aio_sig_num;
-    acb->aiocb.aio_sigevent.sigev_notify = SIGEV_SIGNAL;
-    acb->aiocb.aio_buf = buf;
-    acb->aiocb.aio_nbytes = nb_sectors * 512;
-    acb->aiocb.aio_offset = sector_num * 512;
-    acb->next = first_aio;
-    first_aio = acb;
-    return acb;
-}
+    acb = qemu_aio_read(bs, sector_num * 512, buf, nb_sectors * 512,
+                        cb, opaque);
 
-static BlockDriverAIOCB *raw_aio_read(BlockDriverState *bs,
-        int64_t sector_num, uint8_t *buf, int nb_sectors,
-        BlockDriverCompletionFunc *cb, void *opaque)
-{
-    RawAIOCB *acb;
-
-    acb = raw_aio_setup(bs, sector_num, buf, nb_sectors, cb, opaque);
-    if (!acb)
-        return NULL;
-    if (aio_read(&acb->aiocb) < 0) {
-        qemu_aio_release(acb);
-        return NULL;
-    }
     return &acb->common;
 }
 
@@ -416,41 +254,20 @@ static BlockDriverAIOCB *raw_aio_write(B
 {
     RawAIOCB *acb;
 
-    acb = raw_aio_setup(bs, sector_num, (uint8_t*)buf, nb_sectors, cb, opaque);
-    if (!acb)
-        return NULL;
-    if (aio_write(&acb->aiocb) < 0) {
-        qemu_aio_release(acb);
+    if (fd_open(bs) < 0)
         return NULL;
-    }
+
+    acb = qemu_aio_write(bs, sector_num * 512, buf, nb_sectors * 512,
+                         cb, opaque);
+
     return &acb->common;
 }
 
 static void raw_aio_cancel(BlockDriverAIOCB *blockacb)
 {
-    int ret;
     RawAIOCB *acb = (RawAIOCB *)blockacb;
-    RawAIOCB **pacb;
 
-    ret = aio_cancel(acb->aiocb.aio_fildes, &acb->aiocb);
-    if (ret == AIO_NOTCANCELED) {
-        /* fail safe: if the aio could not be canceled, we wait for
-           it */
-        while (aio_error(&acb->aiocb) == EINPROGRESS);
-    }
-
-    /* remove the callback from the queue */
-    pacb = &first_aio;
-    for(;;) {
-        if (*pacb == NULL) {
-            break;
-        } else if (*pacb == acb) {
-            *pacb = acb->next;
-            qemu_aio_release(acb);
-            break;
-        }
-        pacb = &acb->next;
-    }
+    qemu_aio_cancel(acb);
 }
 
 static void raw_close(BlockDriverState *bs)
Index: qemu/Makefile
===================================================================
--- qemu.orig/Makefile	2008-01-23 09:18:17.000000000 +0100
+++ qemu/Makefile	2008-01-23 09:19:30.000000000 +0100
@@ -136,7 +136,7 @@ QEMU_IMG_BLOCK_OBJS = $(BLOCK_OBJS)
 ifdef CONFIG_WIN32
 QEMU_IMG_BLOCK_OBJS += qemu-img-block-raw-win32.o
 else
-QEMU_IMG_BLOCK_OBJS += qemu-img-block-raw-posix.o
+QEMU_IMG_BLOCK_OBJS += qemu-img-block-raw-posix.o qemu-img-qemu-aio-posix.o
 endif
 
 ######################################################################
Index: qemu/qemu-aio-posix.c
===================================================================
--- /dev/null	1970-01-01 00:00:00.000000000 +0000
+++ qemu/qemu-aio-posix.c	2008-01-23 09:19:30.000000000 +0100
@@ -0,0 +1,241 @@
+/*
+ * QEMU Asynchronous I/O
+ *
+ * Copyright (c) 2006 Fabrice Bellard
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ */
+#include "qemu-common.h"
+#include <aio.h>
+#ifndef QEMU_IMG
+#include "qemu-timer.h"
+#include "exec-all.h"
+#endif
+#include "block_int.h"
+#include "qemu-aio-posix.h"
+
+static void aio_signal_handler(int signum)
+{
+#ifndef QEMU_IMG
+    CPUState *env = cpu_single_env;
+    if (env) {
+        /* stop the currently executing cpu because a timer occured */
+        cpu_interrupt(env, CPU_INTERRUPT_EXIT);
+#ifdef USE_KQEMU
+        if (env->kqemu_enabled) {
+            kqemu_cpu_interrupt(env);
+        }
+#endif
+    }
+#endif
+}
+
+static int aio_initialized = 0;
+static int aio_sig_num = SIGUSR2;
+static RawAIOCB *first_aio; /* AIO issued */
+
+void qemu_aio_init(void)
+{
+    struct sigaction act;
+
+    aio_initialized = 1;
+
+    sigfillset(&act.sa_mask);
+    act.sa_flags = 0; /* do not restart syscalls to interrupt select() */
+    act.sa_handler = aio_signal_handler;
+    sigaction(aio_sig_num, &act, NULL);
+
+#if defined(__GLIBC__) && defined(__linux__)
+    {
+        /* XXX: aio thread exit seems to hang on RedHat 9 and this init
+           seems to fix the problem. */
+        struct aioinit ai;
+        memset(&ai, 0, sizeof(ai));
+        ai.aio_threads = 1;
+        ai.aio_num = 1;
+        ai.aio_idle_time = 365 * 100000;
+        aio_init(&ai);
+    }
+#endif
+}
+
+void qemu_aio_poll(void)
+{
+    RawAIOCB *acb, **pacb;
+    int ret;
+
+    for(;;) {
+        pacb = &first_aio;
+        for(;;) {
+            acb = *pacb;
+            if (!acb)
+                goto the_end;
+            ret = aio_error(&acb->aiocb);
+            if (ret == ECANCELED) {
+                /* remove the request */
+                *pacb = acb->next;
+                qemu_aio_release(acb);
+            } else if (ret != EINPROGRESS) {
+                /* end of aio */
+                if (ret == 0) {
+                    ret = aio_return(&acb->aiocb);
+                    if (ret == acb->aiocb.aio_nbytes)
+                        ret = 0;
+                    else
+                        ret = -EINVAL;
+                } else {
+                    ret = -ret;
+                }
+                /* remove the request */
+                *pacb = acb->next;
+                /* call the callback */
+                acb->common.cb(acb->common.opaque, ret);
+                qemu_aio_release(acb);
+                break;
+            } else {
+                pacb = &acb->next;
+            }
+        }
+    }
+ the_end: ;
+}
+
+/* Wait for all IO requests to complete.  */
+void qemu_aio_flush(void)
+{
+    qemu_aio_wait_start();
+    qemu_aio_poll();
+    while (first_aio) {
+        qemu_aio_wait();
+    }
+    qemu_aio_wait_end();
+}
+
+/* wait until at least one AIO was handled */
+static sigset_t wait_oset;
+
+void qemu_aio_wait_start(void)
+{
+    sigset_t set;
+
+    if (!aio_initialized)
+        qemu_aio_init();
+    sigemptyset(&set);
+    sigaddset(&set, aio_sig_num);
+    sigprocmask(SIG_BLOCK, &set, &wait_oset);
+}
+
+void qemu_aio_wait(void)
+{
+    sigset_t set;
+    int nb_sigs;
+
+#ifndef QEMU_IMG
+    if (qemu_bh_poll())
+        return;
+#endif
+    sigemptyset(&set);
+    sigaddset(&set, aio_sig_num);
+    sigwait(&set, &nb_sigs);
+    qemu_aio_poll();
+}
+
+void qemu_aio_wait_end(void)
+{
+    sigprocmask(SIG_SETMASK, &wait_oset, NULL);
+}
+
+static RawAIOCB *qemu_aio_setup(BlockDriverState *bs,
+        int64_t offset, uint8_t *buf, int nb_bytes,
+        BlockDriverCompletionFunc *cb, void *opaque)
+{
+    int fd = *(int*)bs->opaque;
+    RawAIOCB *acb;
+
+    acb = qemu_aio_get(bs, cb, opaque);
+    if (!acb)
+        return NULL;
+    acb->aiocb.aio_fildes = fd;
+    acb->aiocb.aio_sigevent.sigev_signo = aio_sig_num;
+    acb->aiocb.aio_sigevent.sigev_notify = SIGEV_SIGNAL;
+    acb->aiocb.aio_buf = buf;
+    acb->aiocb.aio_nbytes = nb_bytes;
+    acb->aiocb.aio_offset = offset;
+    acb->next = first_aio;
+    first_aio = acb;
+    return acb;
+}
+
+RawAIOCB *qemu_aio_read(BlockDriverState *bs,
+        int64_t offset, uint8_t *buf, int nb_bytes,
+        BlockDriverCompletionFunc *cb, void *opaque)
+{
+    RawAIOCB *acb;
+
+    acb = qemu_aio_setup(bs, offset, buf, nb_bytes, cb, opaque);
+    if (!acb)
+        return NULL;
+    if (aio_read(&acb->aiocb) < 0) {
+        qemu_aio_release(acb);
+        return NULL;
+    }
+    return acb;
+}
+
+RawAIOCB *qemu_aio_write(BlockDriverState *bs,
+        int64_t offset, const uint8_t *buf, int nb_bytes,
+        BlockDriverCompletionFunc *cb, void *opaque)
+{
+    RawAIOCB *acb;
+
+    acb = qemu_aio_setup(bs, offset, (uint8_t*)buf, nb_bytes, cb, opaque);
+    if (!acb)
+        return NULL;
+    if (aio_write(&acb->aiocb) < 0) {
+        qemu_aio_release(acb);
+        return NULL;
+    }
+    return acb;
+}
+
+void qemu_aio_cancel(RawAIOCB *acb)
+{
+    int ret;
+    RawAIOCB **pacb;
+
+    ret = aio_cancel(acb->aiocb.aio_fildes, &acb->aiocb);
+    if (ret == AIO_NOTCANCELED) {
+        /* fail safe: if the aio could not be canceled, we wait for
+           it */
+        while (aio_error(&acb->aiocb) == EINPROGRESS);
+    }
+
+    /* remove the callback from the queue */
+    pacb = &first_aio;
+    for(;;) {
+        if (*pacb == NULL) {
+            break;
+        } else if (*pacb == acb) {
+            *pacb = acb->next;
+            qemu_aio_release(acb);
+            break;
+        }
+        pacb = &acb->next;
+    }
+}
Index: qemu/qemu-aio-posix.h
===================================================================
--- /dev/null	1970-01-01 00:00:00.000000000 +0000
+++ qemu/qemu-aio-posix.h	2008-01-23 09:19:30.000000000 +0100
@@ -0,0 +1,39 @@
+/*
+ * QEMU Asynchronous I/O
+ *
+ * Copyright (c) 2006 Fabrice Bellard
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
+ * THE SOFTWARE.
+ */
+
+#include <aio.h>
+
+typedef struct RawAIOCB {
+    BlockDriverAIOCB common;
+    struct aiocb aiocb;
+    struct RawAIOCB *next;
+} RawAIOCB;
+
+RawAIOCB *qemu_aio_read(BlockDriverState *bs,
+        int64_t offset, uint8_t *buf, int nb_bytes,
+        BlockDriverCompletionFunc *cb, void *opaque);
+RawAIOCB *qemu_aio_write(BlockDriverState *bs,
+        int64_t offset, const uint8_t *buf, int nb_bytes,
+        BlockDriverCompletionFunc *cb, void *opaque);
+void qemu_aio_cancel(RawAIOCB *acb);