block: bdrv_reopen() patches

index 470bdcc..c3eb585 100644
@@ -664,7 +664,7 @@ static int bdrv_open_common(BlockDriverState *bs, const char *filename,
         open_flags |= BDRV_O_RDWR;
     }
 
-    bs->keep_read_only = bs->read_only = !(open_flags & BDRV_O_RDWR);
+    bs->read_only = !(open_flags & BDRV_O_RDWR);
 
     /* Open the image, either directly or using a protocol */
     if (drv->bdrv_file_open) {

@@ -804,6 +804,12 @@ int bdrv_open(BlockDriverState *bs, const char *filename, int flags,
         goto unlink_and_fail;
     }
 
+    if (flags & BDRV_O_RDWR) {
+        flags |= BDRV_O_ALLOW_RDWR;
+    }
+
+    bs->keep_read_only = !(flags & BDRV_O_ALLOW_RDWR);
+
     /* Open the image */
     ret = bdrv_open_common(bs, filename, flags, drv);
     if (ret < 0) {

@@ -833,12 +839,6 @@ int bdrv_open(BlockDriverState *bs, const char *filename, int flags,
             bdrv_close(bs);
             return ret;
         }
-        if (bs->is_temporary) {
-            bs->backing_hd->keep_read_only = !(flags & BDRV_O_RDWR);
-        } else {
-            /* base image inherits from "parent" */
-            bs->backing_hd->keep_read_only = bs->keep_read_only;
-        }
     }
 
     if (!bdrv_key_required(bs)) {

index 2e2be11..4d919c2 100644
@@ -80,6 +80,7 @@ typedef struct BlockDevOps {
 #define BDRV_O_COPY_ON_READ 0x0400 /* copy read backing sectors into image */
 #define BDRV_O_INCOMING    0x0800  /* consistency hint for incoming migration */
 #define BDRV_O_CHECK       0x1000  /* open solely for consistency check */
+#define BDRV_O_ALLOW_RDWR  0x2000  /* allow reopen to change from r/o to r/w */
 
 #define BDRV_O_CACHE_MASK  (BDRV_O_NOCACHE | BDRV_O_CACHE_WB | BDRV_O_NO_FLUSH)
 

index c3eb585..2b96c8e 100644
@@ -2164,6 +2164,13 @@ int bdrv_enable_write_cache(BlockDriverState *bs)
 void bdrv_set_enable_write_cache(BlockDriverState *bs, bool wce)
 {
     bs->enable_write_cache = wce;
+
+    /* so a reopen() will preserve wce */
+    if (wce) {
+        bs->open_flags |= BDRV_O_CACHE_WB;
+    } else {
+        bs->open_flags &= ~BDRV_O_CACHE_WB;
+    }
 }
 
 int bdrv_is_encrypted(BlockDriverState *bs)

index 2b96c8e..d625703 100644
@@ -859,6 +859,238 @@ unlink_and_fail:
     return ret;
 }
 
+typedef struct BlockReopenQueueEntry {
+     bool prepared;
+     BDRVReopenState state;
+     QSIMPLEQ_ENTRY(BlockReopenQueueEntry) entry;
+} BlockReopenQueueEntry;
+
+/*
+ * Adds a BlockDriverState to a simple queue for an atomic, transactional
+ * reopen of multiple devices.
+ *
+ * bs_queue can either be an existing BlockReopenQueue that has had QSIMPLE_INIT
+ * already performed, or alternatively may be NULL a new BlockReopenQueue will
+ * be created and initialized. This newly created BlockReopenQueue should be
+ * passed back in for subsequent calls that are intended to be of the same
+ * atomic 'set'.
+ *
+ * bs is the BlockDriverState to add to the reopen queue.
+ *
+ * flags contains the open flags for the associated bs
+ *
+ * returns a pointer to bs_queue, which is either the newly allocated
+ * bs_queue, or the existing bs_queue being used.
+ *
+ */
+BlockReopenQueue *bdrv_reopen_queue(BlockReopenQueue *bs_queue,
+                                    BlockDriverState *bs, int flags)
+{
+    assert(bs != NULL);
+
+    BlockReopenQueueEntry *bs_entry;
+    if (bs_queue == NULL) {
+        bs_queue = g_new0(BlockReopenQueue, 1);
+        QSIMPLEQ_INIT(bs_queue);
+    }
+
+    if (bs->file) {
+        bdrv_reopen_queue(bs_queue, bs->file, flags);
+    }
+
+    bs_entry = g_new0(BlockReopenQueueEntry, 1);
+    QSIMPLEQ_INSERT_TAIL(bs_queue, bs_entry, entry);
+
+    bs_entry->state.bs = bs;
+    bs_entry->state.flags = flags;
+
+    return bs_queue;
+}
+
+/*
+ * Reopen multiple BlockDriverStates atomically & transactionally.
+ *
+ * The queue passed in (bs_queue) must have been built up previous
+ * via bdrv_reopen_queue().
+ *
+ * Reopens all BDS specified in the queue, with the appropriate
+ * flags.  All devices are prepared for reopen, and failure of any
+ * device will cause all device changes to be abandonded, and intermediate
+ * data cleaned up.
+ *
+ * If all devices prepare successfully, then the changes are committed
+ * to all devices.
+ *
+ */
+int bdrv_reopen_multiple(BlockReopenQueue *bs_queue, Error **errp)
+{
+    int ret = -1;
+    BlockReopenQueueEntry *bs_entry, *next;
+    Error *local_err = NULL;
+
+    assert(bs_queue != NULL);
+
+    bdrv_drain_all();
+
+    QSIMPLEQ_FOREACH(bs_entry, bs_queue, entry) {
+        if (bdrv_reopen_prepare(&bs_entry->state, bs_queue, &local_err)) {
+            error_propagate(errp, local_err);
+            goto cleanup;
+        }
+        bs_entry->prepared = true;
+    }
+
+    /* If we reach this point, we have success and just need to apply the
+     * changes
+     */
+    QSIMPLEQ_FOREACH(bs_entry, bs_queue, entry) {
+        bdrv_reopen_commit(&bs_entry->state);
+    }
+
+    ret = 0;
+
+cleanup:
+    QSIMPLEQ_FOREACH_SAFE(bs_entry, bs_queue, entry, next) {
+        if (ret && bs_entry->prepared) {
+            bdrv_reopen_abort(&bs_entry->state);
+        }
+        g_free(bs_entry);
+    }
+    g_free(bs_queue);
+    return ret;
+}
+
+
+/* Reopen a single BlockDriverState with the specified flags. */
+int bdrv_reopen(BlockDriverState *bs, int bdrv_flags, Error **errp)
+{
+    int ret = -1;
+    Error *local_err = NULL;
+    BlockReopenQueue *queue = bdrv_reopen_queue(NULL, bs, bdrv_flags);
+
+    ret = bdrv_reopen_multiple(queue, &local_err);
+    if (local_err != NULL) {
+        error_propagate(errp, local_err);
+    }
+    return ret;
+}
+
+
+/*
+ * Prepares a BlockDriverState for reopen. All changes are staged in the
+ * 'opaque' field of the BDRVReopenState, which is used and allocated by
+ * the block driver layer .bdrv_reopen_prepare()
+ *
+ * bs is the BlockDriverState to reopen
+ * flags are the new open flags
+ * queue is the reopen queue
+ *
+ * Returns 0 on success, non-zero on error.  On error errp will be set
+ * as well.
+ *
+ * On failure, bdrv_reopen_abort() will be called to clean up any data.
+ * It is the responsibility of the caller to then call the abort() or
+ * commit() for any other BDS that have been left in a prepare() state
+ *
+ */
+int bdrv_reopen_prepare(BDRVReopenState *reopen_state, BlockReopenQueue *queue,
+                        Error **errp)
+{
+    int ret = -1;
+    Error *local_err = NULL;
+    BlockDriver *drv;
+
+    assert(reopen_state != NULL);
+    assert(reopen_state->bs->drv != NULL);
+    drv = reopen_state->bs->drv;
+
+    /* if we are to stay read-only, do not allow permission change
+     * to r/w */
+    if (!(reopen_state->bs->open_flags & BDRV_O_ALLOW_RDWR) &&
+        reopen_state->flags & BDRV_O_RDWR) {
+        error_set(errp, QERR_DEVICE_IS_READ_ONLY,
+                  reopen_state->bs->device_name);
+        goto error;
+    }
+
+
+    ret = bdrv_flush(reopen_state->bs);
+    if (ret) {
+        error_set(errp, ERROR_CLASS_GENERIC_ERROR, "Error (%s) flushing drive",
+                  strerror(-ret));
+        goto error;
+    }
+
+    if (drv->bdrv_reopen_prepare) {
+        ret = drv->bdrv_reopen_prepare(reopen_state, queue, &local_err);
+        if (ret) {
+            if (local_err != NULL) {
+                error_propagate(errp, local_err);
+            } else {
+                error_set(errp, QERR_OPEN_FILE_FAILED,
+                          reopen_state->bs->filename);
+            }
+            goto error;
+        }
+    } else {
+        /* It is currently mandatory to have a bdrv_reopen_prepare()
+         * handler for each supported drv. */
+        error_set(errp, QERR_BLOCK_FORMAT_FEATURE_NOT_SUPPORTED,
+                  drv->format_name, reopen_state->bs->device_name,
+                 "reopening of file");
+        ret = -1;
+        goto error;
+    }
+
+    ret = 0;
+
+error:
+    return ret;
+}
+
+/*
+ * Takes the staged changes for the reopen from bdrv_reopen_prepare(), and
+ * makes them final by swapping the staging BlockDriverState contents into
+ * the active BlockDriverState contents.
+ */
+void bdrv_reopen_commit(BDRVReopenState *reopen_state)
+{
+    BlockDriver *drv;
+
+    assert(reopen_state != NULL);
+    drv = reopen_state->bs->drv;
+    assert(drv != NULL);
+
+    /* If there are any driver level actions to take */
+    if (drv->bdrv_reopen_commit) {
+        drv->bdrv_reopen_commit(reopen_state);
+    }
+
+    /* set BDS specific flags now */
+    reopen_state->bs->open_flags         = reopen_state->flags;
+    reopen_state->bs->enable_write_cache = !!(reopen_state->flags &
+                                              BDRV_O_CACHE_WB);
+    reopen_state->bs->read_only = !(reopen_state->flags & BDRV_O_RDWR);
+}
+
+/*
+ * Abort the reopen, and delete and free the staged changes in
+ * reopen_state
+ */
+void bdrv_reopen_abort(BDRVReopenState *reopen_state)
+{
+    BlockDriver *drv;
+
+    assert(reopen_state != NULL);
+    drv = reopen_state->bs->drv;
+    assert(drv != NULL);
+
+    if (drv->bdrv_reopen_abort) {
+        drv->bdrv_reopen_abort(reopen_state);
+    }
+}
+
+
 void bdrv_close(BlockDriverState *bs)
 {
     bdrv_flush(bs);

index 4d919c2..b1095d8 100644
@@ -97,6 +97,15 @@ typedef enum {
     BDRV_ACTION_REPORT, BDRV_ACTION_IGNORE, BDRV_ACTION_STOP
 } BlockQMPEventAction;
 
+typedef QSIMPLEQ_HEAD(BlockReopenQueue, BlockReopenQueueEntry) BlockReopenQueue;
+
+typedef struct BDRVReopenState {
+    BlockDriverState *bs;
+    int flags;
+    void *opaque;
+} BDRVReopenState;
+
+
 void bdrv_iostatus_enable(BlockDriverState *bs);
 void bdrv_iostatus_reset(BlockDriverState *bs);
 void bdrv_iostatus_disable(BlockDriverState *bs);

@@ -131,6 +140,14 @@ int bdrv_parse_cache_flags(const char *mode, int *flags);
 int bdrv_file_open(BlockDriverState **pbs, const char *filename, int flags);
 int bdrv_open(BlockDriverState *bs, const char *filename, int flags,
               BlockDriver *drv);
+BlockReopenQueue *bdrv_reopen_queue(BlockReopenQueue *bs_queue,
+                                    BlockDriverState *bs, int flags);
+int bdrv_reopen_multiple(BlockReopenQueue *bs_queue, Error **errp);
+int bdrv_reopen(BlockDriverState *bs, int bdrv_flags, Error **errp);
+int bdrv_reopen_prepare(BDRVReopenState *reopen_state,
+                        BlockReopenQueue *queue, Error **errp);
+void bdrv_reopen_commit(BDRVReopenState *reopen_state);
+void bdrv_reopen_abort(BDRVReopenState *reopen_state);
 void bdrv_close(BlockDriverState *bs);
 int bdrv_attach_dev(BlockDriverState *bs, void *dev);
 void bdrv_attach_dev_nofail(BlockDriverState *bs, void *dev);

index 4452f6f..22b3d93 100644
@@ -139,6 +139,13 @@ struct BlockDriver {
     int instance_size;
     int (*bdrv_probe)(const uint8_t *buf, int buf_size, const char *filename);
     int (*bdrv_probe_device)(const char *filename);
+
+    /* For handling image reopen for split or non-split files */
+    int (*bdrv_reopen_prepare)(BDRVReopenState *reopen_state,
+                               BlockReopenQueue *queue, Error **errp);
+    void (*bdrv_reopen_commit)(BDRVReopenState *reopen_state);
+    void (*bdrv_reopen_abort)(BDRVReopenState *reopen_state);
+
     int (*bdrv_open)(BlockDriverState *bs, int flags);
     int (*bdrv_file_open)(BlockDriverState *bs, const char *filename, int flags);
     int (*bdrv_read)(BlockDriverState *bs, int64_t sector_num,

@@ -336,6 +343,7 @@ struct BlockDriverState {
 
     /* long-running background operation */
     BlockJob *job;
+
 };
 
 int get_tmp_filename(char *filename, int size);

index e5c2bcd..6a6181c 100644
@@ -245,6 +245,7 @@ typedef struct NICInfo NICInfo;
 typedef struct HCIInfo HCIInfo;
 typedef struct AudioState AudioState;
 typedef struct BlockDriverState BlockDriverState;
+typedef struct BDRVReopenState BDRVReopenState;
 typedef struct DriveInfo DriveInfo;
 typedef struct DisplayState DisplayState;
 typedef struct DisplayChangeListener DisplayChangeListener;

index 6be20b1..ee55f79 100644
@@ -185,6 +185,40 @@ static int raw_normalize_devicepath(const char **filename)
 }
 #endif
 
+static int raw_set_aio(void **aio_ctx, int *use_aio, int bdrv_flags)
+{
+#ifdef CONFIG_LINUX_AIO
+    int ret = -1;
+    assert(aio_ctx != NULL);
+    assert(use_aio != NULL);
+    /*
+     * Currently Linux do AIO only for files opened with O_DIRECT
+     * specified so check NOCACHE flag too
+     */
+    if ((bdrv_flags & (BDRV_O_NOCACHE|BDRV_O_NATIVE_AIO)) ==
+                      (BDRV_O_NOCACHE|BDRV_O_NATIVE_AIO)) {
+
+        /* if non-NULL, laio_init() has already been run */
+        if (*aio_ctx == NULL) {
+            *aio_ctx = laio_init();
+            if (!*aio_ctx) {
+                goto error;
+            }
+        }
+        *use_aio = 1;
+    } else {
+        *use_aio = 0;
+    }
+
+    ret = 0;
+
+error:
+    return ret;
+#else
+    return 0;
+#endif
+}
+
 static int raw_open_common(BlockDriverState *bs, const char *filename,
                            int bdrv_flags, int open_flags)
 {

@@ -239,26 +273,7 @@ static int raw_open_common(BlockDriverState *bs, const char *filename,
         goto out_free_buf;
     }
 
-#ifdef CONFIG_LINUX_AIO
-    /*
-     * Currently Linux do AIO only for files opened with O_DIRECT
-     * specified so check NOCACHE flag too
-     */
-    if ((bdrv_flags & (BDRV_O_NOCACHE|BDRV_O_NATIVE_AIO)) ==
-                      (BDRV_O_NOCACHE|BDRV_O_NATIVE_AIO)) {
-
-        s->aio_ctx = laio_init();
-        if (!s->aio_ctx) {
-            goto out_free_buf;
-        }
-        s->use_aio = 1;
-    } else
-#endif
-    {
-#ifdef CONFIG_LINUX_AIO
-        s->use_aio = 0;
-#endif
-    }
+    raw_set_aio(&s->aio_ctx, &s->use_aio, bdrv_flags);
 
 #ifdef CONFIG_XFS
     if (platform_test_xfs_fd(s->fd)) {

index ee55f79..7d3ac9d 100644
@@ -185,6 +185,28 @@ static int raw_normalize_devicepath(const char **filename)
 }
 #endif
 
+static void raw_parse_flags(int bdrv_flags, int *open_flags)
+{
+    assert(open_flags != NULL);
+
+    *open_flags |= O_BINARY;
+    *open_flags &= ~O_ACCMODE;
+    if (bdrv_flags & BDRV_O_RDWR) {
+        *open_flags |= O_RDWR;
+    } else {
+        *open_flags |= O_RDONLY;
+    }
+
+    /* Use O_DSYNC for write-through caching, no flags for write-back caching,
+     * and O_DIRECT for no caching. */
+    if ((bdrv_flags & BDRV_O_NOCACHE)) {
+        *open_flags |= O_DIRECT;
+    }
+    if (!(bdrv_flags & BDRV_O_CACHE_WB)) {
+        *open_flags |= O_DSYNC;
+    }
+}
+
 static int raw_set_aio(void **aio_ctx, int *use_aio, int bdrv_flags)
 {
 #ifdef CONFIG_LINUX_AIO

@@ -230,20 +252,8 @@ static int raw_open_common(BlockDriverState *bs, const char *filename,
         return ret;
     }
 
-    s->open_flags = open_flags | O_BINARY;
-    s->open_flags &= ~O_ACCMODE;
-    if (bdrv_flags & BDRV_O_RDWR) {
-        s->open_flags |= O_RDWR;
-    } else {
-        s->open_flags |= O_RDONLY;
-    }
-
-    /* Use O_DSYNC for write-through caching, no flags for write-back caching,
-     * and O_DIRECT for no caching. */
-    if ((bdrv_flags & BDRV_O_NOCACHE))
-        s->open_flags |= O_DIRECT;
-    if (!(bdrv_flags & BDRV_O_CACHE_WB))
-        s->open_flags |= O_DSYNC;
+    s->open_flags = open_flags;
+    raw_parse_flags(bdrv_flags, &s->open_flags);
 
     s->fd = -1;
     fd = qemu_open(filename, s->open_flags, 0644);

64 lines not on this page · raw message

index 0b96165..70cf700 100644
@@ -268,10 +268,6 @@ iscsi_aio_writev(BlockDriverState *bs, int64_t sector_num,
     acb->task->xfer_dir = SCSI_XFER_WRITE;
     acb->task->cdb_size = 16;
     acb->task->cdb[0] = 0x8a;
-    if (!(bs->open_flags & BDRV_O_CACHE_WB)) {
-        /* set FUA on writes when cache mode is write through */
-        acb->task->cdb[1] |= 0x04;
-    }
     lba = sector_qemu2lun(sector_num, iscsilun);
     *(uint32_t *)&acb->task->cdb[2]  = htonl(lba >> 32);
     *(uint32_t *)&acb->task->cdb[6]  = htonl(lba & 0xffffffff);

index 7d3ac9d..4a1047c 100644
@@ -202,9 +202,6 @@ static void raw_parse_flags(int bdrv_flags, int *open_flags)
     if ((bdrv_flags & BDRV_O_NOCACHE)) {
         *open_flags |= O_DIRECT;
     }
-    if (!(bdrv_flags & BDRV_O_CACHE_WB)) {
-        *open_flags |= O_DSYNC;
-    }
 }
 
 static int raw_set_aio(void **aio_ctx, int *use_aio, int bdrv_flags)

index 335c06a..78c8306 100644
@@ -92,9 +92,6 @@ static void raw_parse_flags(int flags, int *access_flags, DWORD *overlapped)
     if (flags & BDRV_O_NOCACHE) {
         *overlapped |= FILE_FLAG_NO_BUFFERING;
     }
-    if (!(flags & BDRV_O_CACHE_WB)) {
-        *overlapped |= FILE_FLAG_WRITE_THROUGH;
-    }
 }
 
 static int raw_open(BlockDriverState *bs, const char *filename, int flags)

index 5a0f79f..015a9db 100644
@@ -487,12 +487,6 @@ static int qemu_rbd_open(BlockDriverState *bs, const char *filename, int flags)
         rados_conf_set(s->cluster, "rbd_cache", "false");
     } else {
         rados_conf_set(s->cluster, "rbd_cache", "true");
-        if (!(flags & BDRV_O_CACHE_WB)) {
-            r = rados_conf_set(s->cluster, "rbd_cache_max_dirty", "0");
-            if (r < 0) {
-                rados_conf_set(s->cluster, "rbd_cache", "false");
-            }
-        }
     }
 
     if (strstr(conf, "conf=") == NULL) {

index df4f441..5b553db 100644
@@ -1114,14 +1114,12 @@ static int sd_open(BlockDriverState *bs, const char *filename, int flags)
         goto out;
     }
 
-    if (flags & BDRV_O_CACHE_WB) {
-        s->cache_enabled = 1;
-        s->flush_fd = connect_to_sdog(s->addr, s->port);
-        if (s->flush_fd < 0) {
-            error_report("failed to connect");
-            ret = s->flush_fd;
-            goto out;
-        }
+    s->cache_enabled = 1;
+    s->flush_fd = connect_to_sdog(s->addr, s->port);
+    if (s->flush_fd < 0) {
+        error_report("failed to connect");
+        ret = s->flush_fd;
+        goto out;
     }
 
     if (snapid || tag[0] != '\0') {

index 4a1047c..1ea09c0 100644
@@ -352,7 +352,7 @@ static BlockDriverAIOCB *raw_aio_submit(BlockDriverState *bs,
      * boundary.  Check if this is the case or tell the low-level
      * driver that it needs to copy the buffer.
      */
-    if (s->aligned_buf) {
+    if ((bs->open_flags & BDRV_O_NOCACHE)) {
         if (!qiov_is_aligned(bs, qiov)) {
             type |= QEMU_AIO_MISALIGNED;
 #ifdef CONFIG_LINUX_AIO

index 1ea09c0..6bf5480 100644
@@ -133,8 +133,6 @@ typedef struct BDRVRawState {
     int use_aio;
     void *aio_ctx;
 #endif
-    uint8_t *aligned_buf;
-    unsigned aligned_buf_size;
 #ifdef CONFIG_XFS
     bool is_xfs : 1;
 #endif

@@ -261,23 +259,10 @@ static int raw_open_common(BlockDriverState *bs, const char *filename,
         return ret;
     }
     s->fd = fd;
-    s->aligned_buf = NULL;
-
-    if ((bdrv_flags & BDRV_O_NOCACHE)) {
-        /*
-         * Allocate a buffer for read/modify/write cycles.  Chose the size
-         * pessimistically as we don't know the block size yet.
-         */
-        s->aligned_buf_size = 32 * MAX_BLOCKSIZE;
-        s->aligned_buf = qemu_memalign(MAX_BLOCKSIZE, s->aligned_buf_size);
-        if (s->aligned_buf == NULL) {
-            goto out_close;
-        }
-    }
 
     /* We're falling back to POSIX AIO in some cases so init always */
     if (paio_init() < 0) {
-        goto out_free_buf;
+        goto out_close;
     }
 
     raw_set_aio(&s->aio_ctx, &s->use_aio, bdrv_flags);

@@ -290,8 +275,6 @@ static int raw_open_common(BlockDriverState *bs, const char *filename,
 
     return 0;
 
-out_free_buf:
-    qemu_vfree(s->aligned_buf);
 out_close:
     qemu_close(fd);
     return -errno;

@@ -400,8 +383,6 @@ static void raw_close(BlockDriverState *bs)
     if (s->fd >= 0) {
         qemu_close(s->fd);
         s->fd = -1;
-        if (s->aligned_buf != NULL)
-            qemu_vfree(s->aligned_buf);
     }
 }
 

index 6bf5480..edd1eca 100644
@@ -138,6 +138,15 @@ typedef struct BDRVRawState {
 #endif
 } BDRVRawState;
 
+typedef struct BDRVRawReopenState {
+    int fd;
+    int open_flags;
+#ifdef CONFIG_LINUX_AIO
+    int use_aio;
+    void *aio_ctx;
+#endif
+} BDRVRawReopenState;
+
 static int fd_open(BlockDriverState *bs);
 static int64_t raw_getlength(BlockDriverState *bs);
 

@@ -288,6 +297,93 @@ static int raw_open(BlockDriverState *bs, const char *filename, int flags)
     return raw_open_common(bs, filename, flags, 0);
 }
 
+static int raw_reopen_prepare(BDRVReopenState *state,
+                              BlockReopenQueue *queue, Error **errp)
+{
+    BDRVRawState *s;
+    BDRVRawReopenState *raw_s;
+    int ret = 0;
+
+    assert(state != NULL);
+    assert(state->bs != NULL);
+
+    s = state->bs->opaque;
+
+    state->opaque = g_malloc0(sizeof(BDRVRawReopenState));
+    raw_s = state->opaque;
+    raw_s->use_aio = s->use_aio;
+    raw_s->aio_ctx = s->aio_ctx;
+
+    raw_parse_flags(state->flags, &raw_s->open_flags);
+    raw_set_aio(&raw_s->aio_ctx, &raw_s->use_aio, state->flags);
+
+    raw_s->fd = -1;
+
+    int fcntl_flags = O_APPEND | O_ASYNC | O_NONBLOCK;
+#ifdef O_NOATIME
+    fcntl_flags |= O_NOATIME;
+#endif
+    if ((raw_s->open_flags & ~fcntl_flags) == (s->open_flags & ~fcntl_flags)) {
+        /* dup the original fd */
+        /* TODO: use qemu fcntl wrapper */
+        raw_s->fd = fcntl(s->fd, F_DUPFD_CLOEXEC, 0);
+        if (raw_s->fd >= 0) {
+            ret = fcntl_setfl(raw_s->fd, raw_s->open_flags);
+            if (ret) {
+                qemu_close(raw_s->fd);
+                raw_s->fd = -1;
+            }
+        }
+    }
+
+    /* If we cannot use fctnl, or fcntl failed, fall back to qemu_open() */
+    if (raw_s->fd == -1) {
+        assert(!(raw_s->open_flags & O_CREAT));
+        raw_s->fd = qemu_open(state->bs->filename, raw_s->open_flags);
+        if (raw_s->fd == -1) {
+            ret = -1;
+        }
+    }
+
+    return ret;
+}
+
+
+static void raw_reopen_commit(BDRVReopenState *state)
+{
+    BDRVRawReopenState *raw_s = state->opaque;
+    BDRVRawState *s = state->bs->opaque;
+
+    s->open_flags = raw_s->open_flags;
+
+    qemu_close(s->fd);
+    s->fd = raw_s->fd;
+    s->use_aio = raw_s->use_aio;
+    s->aio_ctx = raw_s->aio_ctx;
+
+    g_free(state->opaque);
+    state->opaque = NULL;
+}
+
+
+static void raw_reopen_abort(BDRVReopenState *state)
+{
+    BDRVRawReopenState *raw_s = state->opaque;
+
+     /* nothing to do if NULL, we didn't get far enough */
+    if (raw_s == NULL) {
+        return;
+    }
+
+    if (raw_s->fd >= 0) {
+        qemu_close(raw_s->fd);
+        raw_s->fd = -1;
+    }
+    g_free(state->opaque);
+    state->opaque = NULL;
+}
+
+
 /* XXX: use host sector size if necessary with:
 #ifdef DIOCGSECTORSIZE
         {

@@ -738,6 +834,9 @@ static BlockDriver bdrv_file = {
     .instance_size = sizeof(BDRVRawState),
     .bdrv_probe = NULL, /* no probe for protocols */
     .bdrv_file_open = raw_open,
+    .bdrv_reopen_prepare = raw_reopen_prepare,
+    .bdrv_reopen_commit = raw_reopen_commit,
+    .bdrv_reopen_abort = raw_reopen_abort,
     .bdrv_close = raw_close,
     .bdrv_create = raw_create,
     .bdrv_co_discard = raw_co_discard,

index ff34ea4..253e949 100644
@@ -9,6 +9,14 @@ static int raw_open(BlockDriverState *bs, int flags)
     return 0;
 }
 
+/* We have nothing to do for raw reopen, stubs just return
+ * success */
+static int raw_reopen_prepare(BDRVReopenState *state,
+                              BlockReopenQueue *queue,  Error **errp)
+{
+    return 0;
+}
+
 static int coroutine_fn raw_co_readv(BlockDriverState *bs, int64_t sector_num,
                                      int nb_sectors, QEMUIOVector *qiov)
 {

@@ -115,6 +123,8 @@ static BlockDriver bdrv_raw = {
     .bdrv_open          = raw_open,
     .bdrv_close         = raw_close,
 
+    .bdrv_reopen_prepare  = raw_reopen_prepare,
+
     .bdrv_co_readv          = raw_co_readv,
     .bdrv_co_writev         = raw_co_writev,
     .bdrv_co_is_allocated   = raw_co_is_allocated,

index 21cb239..6c182ca 100644
@@ -505,6 +505,14 @@ out:
     return ret;
 }
 
+/* We have nothing to do for QED reopen, stubs just return
+ * success */
+static int bdrv_qed_reopen_prepare(BDRVReopenState *state,
+                                   BlockReopenQueue *queue, Error **errp)
+{
+    return 0;
+}
+
 static void bdrv_qed_close(BlockDriverState *bs)
 {
     BDRVQEDState *s = bs->opaque;

@@ -1564,6 +1572,7 @@ static BlockDriver bdrv_qed = {
     .bdrv_rebind              = bdrv_qed_rebind,
     .bdrv_open                = bdrv_qed_open,
     .bdrv_close               = bdrv_qed_close,
+    .bdrv_reopen_prepare      = bdrv_qed_reopen_prepare,
     .bdrv_create              = bdrv_qed_create,
     .bdrv_co_is_allocated     = bdrv_qed_co_is_allocated,
     .bdrv_make_empty          = bdrv_qed_make_empty,

index 8f183f1..aa5e603 100644
@@ -52,6 +52,7 @@ typedef struct {
     uint32_t magic;
     uint32_t len;
 } QCowExtension;
+
 #define  QCOW2_EXT_MAGIC_END 0
 #define  QCOW2_EXT_MAGIC_BACKING_FORMAT 0xE2792ACA
 #define  QCOW2_EXT_MAGIC_FEATURE_TABLE 0x6803f857

@@ -558,6 +559,14 @@ static int qcow2_set_key(BlockDriverState *bs, const char *key)
     return 0;
 }
 
+/* We have nothing to do for QCOW2 reopen, stubs just return
+ * success */
+static int qcow2_reopen_prepare(BDRVReopenState *state,
+                                BlockReopenQueue *queue, Error **errp)
+{
+    return 0;
+}
+
 static int coroutine_fn qcow2_co_is_allocated(BlockDriverState *bs,
         int64_t sector_num, int nb_sectors, int *pnum)
 {

@@ -1679,6 +1688,7 @@ static BlockDriver bdrv_qcow2 = {
     .bdrv_probe         = qcow2_probe,
     .bdrv_open          = qcow2_open,
     .bdrv_close         = qcow2_close,
+    .bdrv_reopen_prepare  = qcow2_reopen_prepare,
     .bdrv_create        = qcow2_create,
     .bdrv_co_is_allocated = qcow2_co_is_allocated,
     .bdrv_set_key       = qcow2_set_key,

index 7b5ab87..b239c82 100644
@@ -197,6 +197,15 @@ static int qcow_open(BlockDriverState *bs, int flags)
     return ret;
 }
 
+
+/* We have nothing to do for QCOW reopen, stubs just return
+ * success */
+static int qcow_reopen_prepare(BDRVReopenState *state,
+                               BlockReopenQueue *queue, Error **errp)
+{
+    return 0;
+}
+
 static int qcow_set_key(BlockDriverState *bs, const char *key)
 {
     BDRVQcowState *s = bs->opaque;

@@ -868,6 +877,7 @@ static BlockDriver bdrv_qcow = {
     .bdrv_probe		= qcow_probe,
     .bdrv_open		= qcow_open,
     .bdrv_close		= qcow_close,
+    .bdrv_reopen_prepare = qcow_reopen_prepare,
     .bdrv_create	= qcow_create,
 
     .bdrv_co_readv          = qcow_co_readv,

index bba4c61..f2e861b 100644
@@ -300,6 +300,40 @@ static int vmdk_is_cid_valid(BlockDriverState *bs)
     return 1;
 }
 
+/* Queue extents, if any, for reopen() */
+static int vmdk_reopen_prepare(BDRVReopenState *state,
+                               BlockReopenQueue *queue, Error **errp)
+{
+    BDRVVmdkState *s;
+    int ret = -1;
+    int i;
+    VmdkExtent *e;
+
+    assert(state != NULL);
+    assert(state->bs != NULL);
+
+    if (queue == NULL) {
+        error_set(errp, ERROR_CLASS_GENERIC_ERROR,
+                 "No reopen queue for VMDK extents");
+        goto exit;
+    }
+
+    s = state->bs->opaque;
+
+    assert(s != NULL);
+
+    for (i = 0; i < s->num_extents; i++) {
+        e = &s->extents[i];
+        if (e->file != state->bs->file) {
+            bdrv_reopen_queue(queue, e->file, state->flags);
+        }
+    }
+    ret = 0;
+
+exit:
+    return ret;
+}
+
 static int vmdk_parent_open(BlockDriverState *bs)
 {
     char *p_name;

@@ -1646,6 +1680,7 @@ static BlockDriver bdrv_vmdk = {
     .instance_size  = sizeof(BDRVVmdkState),
     .bdrv_probe     = vmdk_probe,
     .bdrv_open      = vmdk_open,
+    .bdrv_reopen_prepare = vmdk_reopen_prepare,
     .bdrv_read      = vmdk_co_read,
     .bdrv_write     = vmdk_co_write,
     .bdrv_close     = vmdk_close,

index 78c8306..8a698d3 100644
@@ -31,6 +31,7 @@
 #define FTYPE_FILE 0
 #define FTYPE_CD     1
 #define FTYPE_HARDDISK 2
+#define WINDOWS_VISTA 6
 
 typedef struct BDRVRawState {
     HANDLE hfile;

@@ -38,6 +39,10 @@ typedef struct BDRVRawState {
     char drive_path[16]; /* format: "d:\" */
 } BDRVRawState;
 
+typedef struct BDRVRawReopenState {
+    HANDLE hfile;
+} BDRVRawReopenState;
+
 int qemu_ftruncate64(int fd, int64_t length)
 {
     LARGE_INTEGER li;

@@ -117,6 +122,103 @@ static int raw_open(BlockDriverState *bs, const char *filename, int flags)
     return 0;
 }
 
+static int raw_reopen_prepare(BDRVReopenState *state,
+                              BlockReopenQueue *queue, Error **errp)
+{
+    BDRVRawState *s;
+    BDRVRawReopenState *raw_s;
+    int ret = 0;
+    int access_flags;
+    DWORD overlapped;
+    OSVERSIONINFO osvi;
+
+    assert(state != NULL);
+    assert(state->bs != NULL);
+
+    s = state->bs->opaque;
+
+    state->opaque = g_malloc0(sizeof(BDRVRawReopenState));
+    raw_s = state->opaque;
+
+    raw_parse_flags(state->flags, &access_flags, &overlapped);
+
+    ZeroMemory(&osvi, sizeof(OSVERSIONINFO));
+    osvi.dwOSVersionInfoSize = sizeof(OSVERSIONINFO);
+
+    GetVersionEx(&osvi);
+    raw_s->hfile = INVALID_HANDLE_VALUE;
+
+    if (osvi.dwMajorVersion >= WINDOWS_VISTA) {
+        raw_s->hfile = ReOpenFile(s->hfile, access_flags, FILE_SHARE_READ,
+                                  overlapped);
+    }
+
+    /* could not reopen the file handle, so fall back to opening
+     * new file (CreateFile) */
+    if (raw_s->hfile == INVALID_HANDLE_VALUE) {
+        raw_s->hfile = CreateFile(state->bs->filename, access_flags,
+                                  FILE_SHARE_READ, NULL, OPEN_EXISTING,
+                                  overlapped, NULL);
+        if (raw_s->hfile == INVALID_HANDLE_VALUE) {
+            /* this could happen because the access_flags requested are
+             * incompatible with the existing share mode of s->hfile,
+             * so our only option now is to close s->hfile, and try again.
+             * This could end badly */
+            CloseHandle(s->hfile);
+            s->hfile = INVALID_HANDLE_VALUE;
+            raw_s->hfile = CreateFile(state->bs->filename, access_flags,
+                                      FILE_SHARE_READ, NULL, OPEN_EXISTING,
+                                      overlapped, NULL);
+            if (raw_s->hfile == INVALID_HANDLE_VALUE) {
+                /* Unrecoverable error */
+                error_set(errp, ERROR_CLASS_GENERIC_ERROR,
+                          "Fatal error reopening %s file; file closed and cannot be opened\n",
+                          state->bs->filename);
+                ret = -1;
+            } else{
+                /* since we had to close the original, go ahead and
+                 * re-assign here */
+                s->hfile = raw_s->hfile;
+                raw_s->hfile = INVALID_HANDLE_VALUE;
+            }
+        }
+    }
+
+    return ret;
+}
+
+
+static void raw_reopen_commit(BDRVReopenState *state)
+{
+    BDRVRawReopenState *raw_s = state->opaque;
+    BDRVRawState *s = state->bs->opaque;
+
+    if (raw_s->hfile != INVALID_HANDLE_VALUE) {
+        CloseHandle(s->hfile);
+        s->hfile = raw_s->hfile;
+    }
+
+    g_free(state->opaque);
+    state->opaque = NULL;
+}
+
+
+static void raw_reopen_abort(BDRVReopenState *state)
+{
+    BDRVRawReopenState *raw_s = state->opaque;
+
+     /* nothing to do if NULL, we didn't get far enough */
+    if (raw_s == NULL) {
+        return;
+    }
+
+    if (raw_s->hfile != INVALID_HANDLE_VALUE) {
+        CloseHandle(raw_s->hfile);
+    }
+    g_free(state->opaque);
+    state->opaque = NULL;
+}
+
 static int raw_read(BlockDriverState *bs, int64_t sector_num,
                     uint8_t *buf, int nb_sectors)
 {

@@ -287,6 +389,9 @@ static BlockDriver bdrv_file = {
     .protocol_name	= "file",
     .instance_size	= sizeof(BDRVRawState),
     .bdrv_file_open	= raw_open,
+    .bdrv_reopen_prepare = raw_reopen_prepare,
+    .bdrv_reopen_commit = raw_reopen_commit,
+    .bdrv_reopen_abort = raw_reopen_abort,
     .bdrv_close		= raw_close,
     .bdrv_create	= raw_create,
 

index c4f1529..ce38395 100644
@@ -454,6 +454,12 @@ static int vdi_open(BlockDriverState *bs, int flags)
     return -1;
 }
 
+static int vdi_reopen_prepare(BDRVReopenState *state,
+                              BlockReopenQueue *queue, Error **errp)
+{
+    return 0;
+}
+
 static int coroutine_fn vdi_co_is_allocated(BlockDriverState *bs,
         int64_t sector_num, int nb_sectors, int *pnum)
 {

@@ -762,6 +768,7 @@ static BlockDriver bdrv_vdi = {
     .bdrv_probe = vdi_probe,
     .bdrv_open = vdi_open,
     .bdrv_close = vdi_close,
+    .bdrv_reopen_prepare = vdi_reopen_prepare,
     .bdrv_create = vdi_create,
     .bdrv_co_is_allocated = vdi_co_is_allocated,
     .bdrv_make_empty = vdi_make_empty,

index c0b82c4..b6bf52f 100644
@@ -265,6 +265,12 @@ static int vpc_open(BlockDriverState *bs, int flags)
     return err;
 }
 
+static int vpc_reopen_prepare(BDRVReopenState *state,
+                              BlockReopenQueue *queue, Error **errp)
+{
+    return 0;
+}
+
 /*
  * Returns the absolute byte offset of the given sector in the image file.
  * If the sector is not allocated, -1 is returned instead.

@@ -783,6 +789,7 @@ static BlockDriver bdrv_vpc = {
     .bdrv_probe     = vpc_probe,
     .bdrv_open      = vpc_open,
     .bdrv_close     = vpc_close,
+    .bdrv_reopen_prepare = vpc_reopen_prepare,
     .bdrv_create    = vpc_create,
 
     .bdrv_read              = vpc_co_read,

index d625703..ebdbf61 100644
@@ -1497,13 +1497,11 @@ int bdrv_check(BlockDriverState *bs, BdrvCheckResult *res, BdrvCheckMode fix)
 int bdrv_commit(BlockDriverState *bs)
 {
     BlockDriver *drv = bs->drv;
-    BlockDriver *backing_drv;
     int64_t sector, total_sectors;
     int n, ro, open_flags;
-    int ret = 0, rw_ret = 0;
+    int ret = 0;
     uint8_t *buf;
     char filename[1024];
-    BlockDriverState *bs_rw, *bs_ro;
 
     if (!drv)
         return -ENOMEDIUM;

@@ -1512,42 +1510,18 @@ int bdrv_commit(BlockDriverState *bs)
         return -ENOTSUP;
     }
 
-    if (bs->backing_hd->keep_read_only) {
-        return -EACCES;
-    }
-
     if (bdrv_in_use(bs) || bdrv_in_use(bs->backing_hd)) {
         return -EBUSY;
     }
 
-    backing_drv = bs->backing_hd->drv;
     ro = bs->backing_hd->read_only;
     strncpy(filename, bs->backing_hd->filename, sizeof(filename));
     open_flags =  bs->backing_hd->open_flags;
 
     if (ro) {
-        /* re-open as RW */
-        bdrv_delete(bs->backing_hd);
-        bs->backing_hd = NULL;
-        bs_rw = bdrv_new("");
-        rw_ret = bdrv_open(bs_rw, filename, open_flags | BDRV_O_RDWR,
-            backing_drv);
-        if (rw_ret < 0) {
-            bdrv_delete(bs_rw);
-            /* try to re-open read-only */
-            bs_ro = bdrv_new("");
-            ret = bdrv_open(bs_ro, filename, open_flags & ~BDRV_O_RDWR,
-                backing_drv);
-            if (ret < 0) {
-                bdrv_delete(bs_ro);
-                /* drive not functional anymore */
-                bs->drv = NULL;
-                return ret;
-            }
-            bs->backing_hd = bs_ro;
-            return rw_ret;
+        if (bdrv_reopen(bs->backing_hd, open_flags | BDRV_O_RDWR, NULL)) {
+            return -EACCES;
         }
-        bs->backing_hd = bs_rw;
     }
 
     total_sectors = bdrv_getlength(bs) >> BDRV_SECTOR_BITS;

@@ -1584,20 +1558,8 @@ ro_cleanup:
     g_free(buf);
 
     if (ro) {
-        /* re-open as RO */
-        bdrv_delete(bs->backing_hd);
-        bs->backing_hd = NULL;
-        bs_ro = bdrv_new("");
-        ret = bdrv_open(bs_ro, filename, open_flags & ~BDRV_O_RDWR,
-            backing_drv);
-        if (ret < 0) {
-            bdrv_delete(bs_ro);
-            /* drive not functional anymore */
-            bs->drv = NULL;
-            return ret;
-        }
-        bs->backing_hd = bs_ro;
-        bs->backing_hd->keep_read_only = 0;
+        /* ignoring error return here */
+        bdrv_reopen(bs->backing_hd, open_flags & ~BDRV_O_RDWR, NULL);
     }
 
     return ret;

index ebdbf61..5db1df1 100644
@@ -808,8 +808,6 @@ int bdrv_open(BlockDriverState *bs, const char *filename, int flags,
         flags |= BDRV_O_ALLOW_RDWR;
     }
 
-    bs->keep_read_only = !(flags & BDRV_O_ALLOW_RDWR);
-
     /* Open the image */
     ret = bdrv_open_common(bs, filename, flags, drv);
     if (ret < 0) {

index 22b3d93..ac4245c 100644
@@ -275,7 +275,6 @@ struct BlockDriverState {
     int64_t total_sectors; /* if we are reading a disk image, give its
                               size in sectors */
     int read_only; /* if true, the media is read only */
-    int keep_read_only; /* if true, the media was requested to stay read only */
     int open_flags; /* flags used to open the file, re-used for re-open */
     int encrypted; /* if true, the media is encrypted */
     int valid_key; /* if true, a valid encryption key has been set */