[Qemu-devel] [RFC PATCH v0 1/3] gluster: Convert aio routines into coroutines

[Bharata B Rao <bharata@linux.vnet.ibm.com>]

Convert the read, write, flush and discard implementations from aio-based
ones to coroutine based ones.

Signed-off-by: Bharata B Rao <bharata@linux.vnet.ibm.com>
---
 block/gluster.c | 168 +++++++++++++++++++++-----------------------------------
 1 file changed, 63 insertions(+), 105 deletions(-)

diff --git a/block/gluster.c b/block/gluster.c
index 877686a..9f85228 100644
--- a/block/gluster.c
+++ b/block/gluster.c
@@ -26,11 +26,11 @@ typedef struct GlusterAIOCB {
     int ret;
     bool *finished;
     QEMUBH *bh;
+    Coroutine *coroutine;
 } GlusterAIOCB;
 
 typedef struct BDRVGlusterState {
     struct glfs *glfs;
-    int fds[2];
     struct glfs_fd *fd;
     int event_reader_pos;
     GlusterAIOCB *event_acb;
@@ -231,46 +231,23 @@ out:
     return NULL;
 }
 
-static void qemu_gluster_complete_aio(GlusterAIOCB *acb, BDRVGlusterState *s)
+static void qemu_gluster_complete_aio(void *opaque)
 {
-    int ret;
-    bool *finished = acb->finished;
-    BlockDriverCompletionFunc *cb = acb->common.cb;
-    void *opaque = acb->common.opaque;
-
-    if (!acb->ret || acb->ret == acb->size) {
-        ret = 0; /* Success */
-    } else if (acb->ret < 0) {
-        ret = acb->ret; /* Read/Write failed */
-    } else {
-        ret = -EIO; /* Partial read/write - fail it */
-    }
+    GlusterAIOCB *acb = (GlusterAIOCB *)opaque;
 
-    qemu_aio_release(acb);
-    cb(opaque, ret);
-    if (finished) {
-        *finished = true;
+    if (acb->ret == acb->size) {
+        acb->ret = 0;
+    } else if (acb->ret > 0) {
+        acb->ret = -EIO; /* Partial read/write - fail it */
     }
-}
 
-static void qemu_gluster_aio_event_reader(void *opaque)
-{
-    BDRVGlusterState *s = opaque;
-    ssize_t ret;
-
-    do {
-        char *p = (char *)&s->event_acb;
-
-        ret = read(s->fds[GLUSTER_FD_READ], p + s->event_reader_pos,
-                   sizeof(s->event_acb) - s->event_reader_pos);
-        if (ret > 0) {
-            s->event_reader_pos += ret;
-            if (s->event_reader_pos == sizeof(s->event_acb)) {
-                s->event_reader_pos = 0;
-                qemu_gluster_complete_aio(s->event_acb, s);
-            }
-        }
-    } while (ret < 0 && errno == EINTR);
+    qemu_bh_delete(acb->bh);
+    acb->bh = NULL;
+    qemu_coroutine_enter(acb->coroutine, NULL);
+    if (acb->finished) {
+        *acb->finished = true;
+    }
+    qemu_aio_release(acb);
 }
 
 /* TODO Convert to fine grained options */
@@ -309,7 +286,6 @@ static int qemu_gluster_open(BlockDriverState *bs,  QDict *options,
 
     filename = qemu_opt_get(opts, "filename");
 
-
     s->glfs = qemu_gluster_init(gconf, filename);
     if (!s->glfs) {
         ret = -errno;
@@ -329,18 +305,8 @@ static int qemu_gluster_open(BlockDriverState *bs,  QDict *options,
     s->fd = glfs_open(s->glfs, gconf->image, open_flags);
     if (!s->fd) {
         ret = -errno;
-        goto out;
     }
 
-    ret = qemu_pipe(s->fds);
-    if (ret < 0) {
-        ret = -errno;
-        goto out;
-    }
-    fcntl(s->fds[GLUSTER_FD_READ], F_SETFL, O_NONBLOCK);
-    qemu_aio_set_fd_handler(s->fds[GLUSTER_FD_READ],
-        qemu_gluster_aio_event_reader, NULL, s);
-
 out:
     qemu_opts_del(opts);
     qemu_gluster_gconf_free(gconf);
@@ -414,28 +380,20 @@ static const AIOCBInfo gluster_aiocb_info = {
     .cancel = qemu_gluster_aio_cancel,
 };
 
+/*
+ * AIO callback routine called from GlusterFS thread.
+ */
 static void gluster_finish_aiocb(struct glfs_fd *fd, ssize_t ret, void *arg)
 {
     GlusterAIOCB *acb = (GlusterAIOCB *)arg;
-    BlockDriverState *bs = acb->common.bs;
-    BDRVGlusterState *s = bs->opaque;
-    int retval;
 
     acb->ret = ret;
-    retval = qemu_write_full(s->fds[GLUSTER_FD_WRITE], &acb, sizeof(acb));
-    if (retval != sizeof(acb)) {
-        /*
-         * Gluster AIO callback thread failed to notify the waiting
-         * QEMU thread about IO completion.
-         */
-        error_report("Gluster AIO completion failed: %s", strerror(errno));
-        abort();
-    }
+    acb->bh = qemu_bh_new(qemu_gluster_complete_aio, acb);
+    qemu_bh_schedule(acb->bh);
 }
 
-static BlockDriverAIOCB *qemu_gluster_aio_rw(BlockDriverState *bs,
-        int64_t sector_num, QEMUIOVector *qiov, int nb_sectors,
-        BlockDriverCompletionFunc *cb, void *opaque, int write)
+static coroutine_fn int qemu_gluster_aio_rw(BlockDriverState *bs,
+        int64_t sector_num, int nb_sectors, QEMUIOVector *qiov, int write)
 {
     int ret;
     GlusterAIOCB *acb;
@@ -446,10 +404,11 @@ static BlockDriverAIOCB *qemu_gluster_aio_rw(BlockDriverState *bs,
     offset = sector_num * BDRV_SECTOR_SIZE;
     size = nb_sectors * BDRV_SECTOR_SIZE;
 
-    acb = qemu_aio_get(&gluster_aiocb_info, bs, cb, opaque);
+    acb = qemu_aio_get(&gluster_aiocb_info, bs, NULL, NULL);
     acb->size = size;
     acb->ret = 0;
     acb->finished = NULL;
+    acb->coroutine = qemu_coroutine_self();
 
     if (write) {
         ret = glfs_pwritev_async(s->fd, qiov->iov, qiov->niov, offset, 0,
@@ -462,11 +421,13 @@ static BlockDriverAIOCB *qemu_gluster_aio_rw(BlockDriverState *bs,
     if (ret < 0) {
         goto out;
     }
-    return &acb->common;
+
+    qemu_coroutine_yield();
+    return acb->ret;
 
 out:
     qemu_aio_release(acb);
-    return NULL;
+    return ret;
 }
 
 static int qemu_gluster_truncate(BlockDriverState *bs, int64_t offset)
@@ -482,47 +443,46 @@ static int qemu_gluster_truncate(BlockDriverState *bs, int64_t offset)
     return 0;
 }
 
-static BlockDriverAIOCB *qemu_gluster_aio_readv(BlockDriverState *bs,
-        int64_t sector_num, QEMUIOVector *qiov, int nb_sectors,
-        BlockDriverCompletionFunc *cb, void *opaque)
+static coroutine_fn int qemu_gluster_co_readv(BlockDriverState *bs,
+        int64_t sector_num, int nb_sectors, QEMUIOVector *qiov)
 {
-    return qemu_gluster_aio_rw(bs, sector_num, qiov, nb_sectors, cb, opaque, 0);
+    return qemu_gluster_aio_rw(bs, sector_num, nb_sectors, qiov, 0);
 }
 
-static BlockDriverAIOCB *qemu_gluster_aio_writev(BlockDriverState *bs,
-        int64_t sector_num, QEMUIOVector *qiov, int nb_sectors,
-        BlockDriverCompletionFunc *cb, void *opaque)
+static coroutine_fn int qemu_gluster_co_writev(BlockDriverState *bs,
+        int64_t sector_num, int nb_sectors, QEMUIOVector *qiov)
 {
-    return qemu_gluster_aio_rw(bs, sector_num, qiov, nb_sectors, cb, opaque, 1);
+    return qemu_gluster_aio_rw(bs, sector_num, nb_sectors, qiov, 1);
 }
 
-static BlockDriverAIOCB *qemu_gluster_aio_flush(BlockDriverState *bs,
-        BlockDriverCompletionFunc *cb, void *opaque)
+static coroutine_fn int qemu_gluster_co_flush_to_disk(BlockDriverState *bs)
 {
     int ret;
     GlusterAIOCB *acb;
     BDRVGlusterState *s = bs->opaque;
 
-    acb = qemu_aio_get(&gluster_aiocb_info, bs, cb, opaque);
+    acb = qemu_aio_get(&gluster_aiocb_info, bs, NULL, NULL);
     acb->size = 0;
     acb->ret = 0;
     acb->finished = NULL;
+    acb->coroutine = qemu_coroutine_self();
 
     ret = glfs_fsync_async(s->fd, &gluster_finish_aiocb, acb);
     if (ret < 0) {
         goto out;
     }
-    return &acb->common;
+
+    qemu_coroutine_yield();
+    return acb->ret;
 
 out:
     qemu_aio_release(acb);
-    return NULL;
+    return ret;
 }
 
 #ifdef CONFIG_GLUSTERFS_DISCARD
-static BlockDriverAIOCB *qemu_gluster_aio_discard(BlockDriverState *bs,
-        int64_t sector_num, int nb_sectors, BlockDriverCompletionFunc *cb,
-        void *opaque)
+static coroutine_fn int qemu_gluster_co_discard(BlockDriverState *bs,
+        int64_t sector_num, int nb_sectors)
 {
     int ret;
     GlusterAIOCB *acb;
@@ -533,20 +493,22 @@ static BlockDriverAIOCB *qemu_gluster_aio_discard(BlockDriverState *bs,
     offset = sector_num * BDRV_SECTOR_SIZE;
     size = nb_sectors * BDRV_SECTOR_SIZE;
 
-    acb = qemu_aio_get(&gluster_aiocb_info, bs, cb, opaque);
+    acb = qemu_aio_get(&gluster_aiocb_info, bs, NULL, NULL);
     acb->size = 0;
     acb->ret = 0;
     acb->finished = NULL;
+    acb->coroutine = qemu_coroutine_self();
 
     ret = glfs_discard_async(s->fd, offset, size, &gluster_finish_aiocb, acb);
     if (ret < 0) {
         goto out;
     }
-    return &acb->common;
 
+    qemu_coroutine_yield();
+    return acb->ret;
 out:
     qemu_aio_release(acb);
-    return NULL;
+    return ret;
 }
 #endif
 
@@ -581,10 +543,6 @@ static void qemu_gluster_close(BlockDriverState *bs)
 {
     BDRVGlusterState *s = bs->opaque;
 
-    close(s->fds[GLUSTER_FD_READ]);
-    close(s->fds[GLUSTER_FD_WRITE]);
-    qemu_aio_set_fd_handler(s->fds[GLUSTER_FD_READ], NULL, NULL, NULL);
-
     if (s->fd) {
         glfs_close(s->fd);
         s->fd = NULL;
@@ -618,12 +576,12 @@ static BlockDriver bdrv_gluster = {
     .bdrv_getlength               = qemu_gluster_getlength,
     .bdrv_get_allocated_file_size = qemu_gluster_allocated_file_size,
     .bdrv_truncate                = qemu_gluster_truncate,
-    .bdrv_aio_readv               = qemu_gluster_aio_readv,
-    .bdrv_aio_writev              = qemu_gluster_aio_writev,
-    .bdrv_aio_flush               = qemu_gluster_aio_flush,
+    .bdrv_co_readv                = qemu_gluster_co_readv,
+    .bdrv_co_writev               = qemu_gluster_co_writev,
+    .bdrv_co_flush_to_disk        = qemu_gluster_co_flush_to_disk,
     .bdrv_has_zero_init           = qemu_gluster_has_zero_init,
 #ifdef CONFIG_GLUSTERFS_DISCARD
-    .bdrv_aio_discard             = qemu_gluster_aio_discard,
+    .bdrv_co_discard              = qemu_gluster_co_discard,
 #endif
     .create_options               = qemu_gluster_create_options,
 };
@@ -639,12 +597,12 @@ static BlockDriver bdrv_gluster_tcp = {
     .bdrv_getlength               = qemu_gluster_getlength,
     .bdrv_get_allocated_file_size = qemu_gluster_allocated_file_size,
     .bdrv_truncate                = qemu_gluster_truncate,
-    .bdrv_aio_readv               = qemu_gluster_aio_readv,
-    .bdrv_aio_writev              = qemu_gluster_aio_writev,
-    .bdrv_aio_flush               = qemu_gluster_aio_flush,
+    .bdrv_co_readv                = qemu_gluster_co_readv,
+    .bdrv_co_writev               = qemu_gluster_co_writev,
+    .bdrv_co_flush_to_disk        = qemu_gluster_co_flush_to_disk,
     .bdrv_has_zero_init           = qemu_gluster_has_zero_init,
 #ifdef CONFIG_GLUSTERFS_DISCARD
-    .bdrv_aio_discard             = qemu_gluster_aio_discard,
+    .bdrv_co_discard              = qemu_gluster_co_discard,
 #endif
     .create_options               = qemu_gluster_create_options,
 };
@@ -660,12 +618,12 @@ static BlockDriver bdrv_gluster_unix = {
     .bdrv_getlength               = qemu_gluster_getlength,
     .bdrv_get_allocated_file_size = qemu_gluster_allocated_file_size,
     .bdrv_truncate                = qemu_gluster_truncate,
-    .bdrv_aio_readv               = qemu_gluster_aio_readv,
-    .bdrv_aio_writev              = qemu_gluster_aio_writev,
-    .bdrv_aio_flush               = qemu_gluster_aio_flush,
+    .bdrv_co_readv                = qemu_gluster_co_readv,
+    .bdrv_co_writev               = qemu_gluster_co_writev,
+    .bdrv_co_flush_to_disk        = qemu_gluster_co_flush_to_disk,
     .bdrv_has_zero_init           = qemu_gluster_has_zero_init,
 #ifdef CONFIG_GLUSTERFS_DISCARD
-    .bdrv_aio_discard             = qemu_gluster_aio_discard,
+    .bdrv_co_discard              = qemu_gluster_co_discard,
 #endif
     .create_options               = qemu_gluster_create_options,
 };
@@ -681,12 +639,12 @@ static BlockDriver bdrv_gluster_rdma = {
     .bdrv_getlength               = qemu_gluster_getlength,
     .bdrv_get_allocated_file_size = qemu_gluster_allocated_file_size,
     .bdrv_truncate                = qemu_gluster_truncate,
-    .bdrv_aio_readv               = qemu_gluster_aio_readv,
-    .bdrv_aio_writev              = qemu_gluster_aio_writev,
-    .bdrv_aio_flush               = qemu_gluster_aio_flush,
+    .bdrv_co_readv                = qemu_gluster_co_readv,
+    .bdrv_co_writev               = qemu_gluster_co_writev,
+    .bdrv_co_flush_to_disk        = qemu_gluster_co_flush_to_disk,
     .bdrv_has_zero_init           = qemu_gluster_has_zero_init,
 #ifdef CONFIG_GLUSTERFS_DISCARD
-    .bdrv_aio_discard             = qemu_gluster_aio_discard,
+    .bdrv_co_discard              = qemu_gluster_co_discard,
 #endif
     .create_options               = qemu_gluster_create_options,
 };
-- 
1.7.11.7