char *name;
int refcnt;
BdrvChild *root;
+ AioContext *ctx;
DriveInfo *legacy_dinfo; /* null unless created by drive_new() */
QTAILQ_ENTRY(BlockBackend) link; /* for block_backends */
QTAILQ_ENTRY(BlockBackend) monitor_link; /* for monitor_block_backends */
BlockBackendPublic public;
- void *dev; /* attached device model, if any */
- bool legacy_dev; /* true if dev is not a DeviceState */
- /* TODO change to DeviceState when all users are qdevified */
+ DeviceState *dev; /* attached device model, if any */
const BlockDevOps *dev_ops;
void *dev_opaque;
uint64_t shared_perm;
bool disable_perm;
+ bool allow_aio_context_change;
bool allow_write_beyond_eof;
NotifierList remove_bs_notifiers, insert_bs_notifiers;
* Accessed with atomic ops.
*/
unsigned int in_flight;
- AioWait wait;
};
typedef struct BlockBackendAIOCB {
abort();
}
static void blk_root_drained_begin(BdrvChild *child);
+static bool blk_root_drained_poll(BdrvChild *child);
static void blk_root_drained_end(BdrvChild *child);
static void blk_root_change_media(BdrvChild *child, bool load);
static void blk_root_resize(BdrvChild *child);
+static bool blk_root_can_set_aio_ctx(BdrvChild *child, AioContext *ctx,
+ GSList **ignore, Error **errp);
+static void blk_root_set_aio_ctx(BdrvChild *child, AioContext *ctx,
+ GSList **ignore);
+
static char *blk_root_get_parent_desc(BdrvChild *child)
{
BlockBackend *blk = child->opaque;
.get_parent_desc = blk_root_get_parent_desc,
.drained_begin = blk_root_drained_begin,
+ .drained_poll = blk_root_drained_poll,
.drained_end = blk_root_drained_end,
.activate = blk_root_activate,
.attach = blk_root_attach,
.detach = blk_root_detach,
+
+ .can_set_aio_ctx = blk_root_can_set_aio_ctx,
+ .set_aio_ctx = blk_root_set_aio_ctx,
};
/*
*
* Return the new BlockBackend on success, null on failure.
*/
-BlockBackend *blk_new(uint64_t perm, uint64_t shared_perm)
+BlockBackend *blk_new(AioContext *ctx, uint64_t perm, uint64_t shared_perm)
{
BlockBackend *blk;
blk = g_new0(BlockBackend, 1);
blk->refcnt = 1;
+ blk->ctx = ctx;
blk->perm = perm;
blk->shared_perm = shared_perm;
blk_set_enable_write_cache(blk, true);
+ blk->on_read_error = BLOCKDEV_ON_ERROR_REPORT;
+ blk->on_write_error = BLOCKDEV_ON_ERROR_ENOSPC;
+
block_acct_init(&blk->stats);
notifier_list_init(&blk->remove_bs_notifiers);
/*
* Creates a new BlockBackend, opens a new BlockDriverState, and connects both.
+ * The new BlockBackend is in the main AioContext.
*
* Just as with bdrv_open(), after having called this function the reference to
* @options belongs to the block layer (even on failure).
perm |= BLK_PERM_RESIZE;
}
- blk = blk_new(perm, BLK_PERM_ALL);
+ blk = blk_new(qemu_get_aio_context(), perm, BLK_PERM_ALL);
bs = bdrv_open(filename, reference, options, flags, errp);
if (!bs) {
blk_unref(blk);
return NULL;
}
- blk->root = bdrv_root_attach_child(bs, "root", &child_root,
+ blk->root = bdrv_root_attach_child(bs, "root", &child_root, blk->ctx,
perm, BLK_PERM_ALL, blk, errp);
if (!blk->root) {
- bdrv_unref(bs);
blk_unref(blk);
return NULL;
}
*/
void blk_ref(BlockBackend *blk)
{
+ assert(blk->refcnt > 0);
blk->refcnt++;
}
{
if (blk) {
assert(blk->refcnt > 0);
- if (!--blk->refcnt) {
+ if (blk->refcnt > 1) {
+ blk->refcnt--;
+ } else {
+ blk_drain(blk);
+ /* blk_drain() cannot resurrect blk, nobody held a reference */
+ assert(blk->refcnt == 1);
+ blk->refcnt = 0;
blk_delete(blk);
}
}
blk_update_root_state(blk);
+ /* bdrv_root_unref_child() will cause blk->root to become stale and may
+ * switch to a completion coroutine later on. Let's drain all I/O here
+ * to avoid that and a potential QEMU crash.
+ */
+ blk_drain(blk);
bdrv_root_unref_child(blk->root);
blk->root = NULL;
}
int blk_insert_bs(BlockBackend *blk, BlockDriverState *bs, Error **errp)
{
ThrottleGroupMember *tgm = &blk->public.throttle_group_member;
- blk->root = bdrv_root_attach_child(bs, "root", &child_root,
+ bdrv_ref(bs);
+ blk->root = bdrv_root_attach_child(bs, "root", &child_root, blk->ctx,
blk->perm, blk->shared_perm, blk, errp);
if (blk->root == NULL) {
return -EPERM;
}
- bdrv_ref(bs);
notifier_list_notify(&blk->insert_bs_notifiers, blk);
if (tgm->throttle_state) {
*shared_perm = blk->shared_perm;
}
-static int blk_do_attach_dev(BlockBackend *blk, void *dev)
+/*
+ * Attach device model @dev to @blk.
+ * Return 0 on success, -EBUSY when a device model is attached already.
+ */
+int blk_attach_dev(BlockBackend *blk, DeviceState *dev)
{
if (blk->dev) {
return -EBUSY;
blk_ref(blk);
blk->dev = dev;
- blk->legacy_dev = false;
blk_iostatus_reset(blk);
return 0;
}
-/*
- * Attach device model @dev to @blk.
- * Return 0 on success, -EBUSY when a device model is attached already.
- */
-int blk_attach_dev(BlockBackend *blk, DeviceState *dev)
-{
- return blk_do_attach_dev(blk, dev);
-}
-
-/*
- * Attach device model @dev to @blk.
- * @blk must not have a device model attached already.
- * TODO qdevified devices don't use this, remove when devices are qdevified
- */
-void blk_attach_dev_legacy(BlockBackend *blk, void *dev)
-{
- if (blk_do_attach_dev(blk, dev) < 0) {
- abort();
- }
- blk->legacy_dev = true;
-}
-
/*
* Detach device model @dev from @blk.
* @dev must be currently attached to @blk.
*/
-void blk_detach_dev(BlockBackend *blk, void *dev)
-/* TODO change to DeviceState *dev when all users are qdevified */
+void blk_detach_dev(BlockBackend *blk, DeviceState *dev)
{
assert(blk->dev == dev);
blk->dev = NULL;
/*
* Return the device model attached to @blk if any, else null.
*/
-void *blk_get_attached_dev(BlockBackend *blk)
-/* TODO change to return DeviceState * when all users are qdevified */
+DeviceState *blk_get_attached_dev(BlockBackend *blk)
{
return blk->dev;
}
* device attached to the BlockBackend. */
char *blk_get_attached_dev_id(BlockBackend *blk)
{
- DeviceState *dev;
-
- assert(!blk->legacy_dev);
- dev = blk->dev;
+ DeviceState *dev = blk->dev;
if (!dev) {
return g_strdup("");
} else if (dev->id) {
return g_strdup(dev->id);
}
- return object_get_canonical_path(OBJECT(dev));
+
+ return object_get_canonical_path(OBJECT(dev)) ?: g_strdup("");
}
/*
void blk_set_dev_ops(BlockBackend *blk, const BlockDevOps *ops,
void *opaque)
{
- /* All drivers that use blk_set_dev_ops() are qdevified and we want to keep
- * it that way, so we can assume blk->dev, if present, is a DeviceState if
- * blk->dev_ops is set. Non-device users may use dev_ops without device. */
- assert(!blk->legacy_dev);
-
blk->dev_ops = ops;
blk->dev_opaque = opaque;
bool tray_was_open, tray_is_open;
Error *local_err = NULL;
- assert(!blk->legacy_dev);
-
tray_was_open = blk_dev_is_tray_open(blk);
blk->dev_ops->change_media_cb(blk->dev_opaque, load, &local_err);
if (local_err) {
if (tray_was_open != tray_is_open) {
char *id = blk_get_attached_dev_id(blk);
- qapi_event_send_device_tray_moved(blk_name(blk), id, tray_is_open,
- &error_abort);
+ qapi_event_send_device_tray_moved(blk_name(blk), id, tray_is_open);
g_free(id);
}
}
void blk_iostatus_reset(BlockBackend *blk)
{
if (blk_iostatus_is_enabled(blk)) {
- BlockDriverState *bs = blk_bs(blk);
blk->iostatus = BLOCK_DEVICE_IO_STATUS_OK;
- if (bs && bs->job) {
- block_job_iostatus_reset(bs->job);
- }
}
}
blk->allow_write_beyond_eof = allow;
}
+void blk_set_allow_aio_context_change(BlockBackend *blk, bool allow)
+{
+ blk->allow_aio_context_change = allow;
+}
+
static int blk_check_byte_request(BlockBackend *blk, int64_t offset,
size_t size)
{
rwco->ret = blk_co_preadv(rwco->blk, rwco->offset, qiov->size,
qiov, rwco->flags);
+ aio_wait_kick();
}
static void blk_write_entry(void *opaque)
rwco->ret = blk_co_pwritev(rwco->blk, rwco->offset, qiov->size,
qiov, rwco->flags);
+ aio_wait_kick();
}
static int blk_prw(BlockBackend *blk, int64_t offset, uint8_t *buf,
int64_t bytes, CoroutineEntry co_entry,
BdrvRequestFlags flags)
{
- QEMUIOVector qiov;
- struct iovec iov;
- BlkRwCo rwco;
-
- iov = (struct iovec) {
- .iov_base = buf,
- .iov_len = bytes,
- };
- qemu_iovec_init_external(&qiov, &iov, 1);
-
- rwco = (BlkRwCo) {
+ QEMUIOVector qiov = QEMU_IOVEC_INIT_BUF(qiov, buf, bytes);
+ BlkRwCo rwco = {
.blk = blk,
.offset = offset,
.iobuf = &qiov,
return bdrv_make_zero(blk->root, flags);
}
-static void blk_inc_in_flight(BlockBackend *blk)
+void blk_inc_in_flight(BlockBackend *blk)
{
atomic_inc(&blk->in_flight);
}
-static void blk_dec_in_flight(BlockBackend *blk)
+void blk_dec_in_flight(BlockBackend *blk)
{
atomic_dec(&blk->in_flight);
- aio_wait_kick(&blk->wait);
+ aio_wait_kick();
}
static void error_callback_bh(void *opaque)
static void blk_aio_complete(BlkAioEmAIOCB *acb)
{
if (acb->has_returned) {
- blk_dec_in_flight(acb->rwco.blk);
acb->common.cb(acb->common.opaque, acb->rwco.ret);
+ blk_dec_in_flight(acb->rwco.blk);
qemu_aio_unref(acb);
}
}
rwco->ret = blk_co_ioctl(rwco->blk, rwco->offset,
qiov->iov[0].iov_base);
+ aio_wait_kick();
}
int blk_ioctl(BlockBackend *blk, unsigned long int req, void *buf)
return ret;
}
- return bdrv_co_pdiscard(blk_bs(blk), offset, bytes);
+ return bdrv_co_pdiscard(blk->root, offset, bytes);
}
int blk_co_flush(BlockBackend *blk)
{
BlkRwCo *rwco = opaque;
rwco->ret = blk_co_flush(rwco->blk);
+ aio_wait_kick();
}
int blk_flush(BlockBackend *blk)
}
/* We may have -ENOMEDIUM completions in flight */
- AIO_WAIT_WHILE(&blk->wait,
- blk_get_aio_context(blk),
- atomic_mb_read(&blk->in_flight) > 0);
+ AIO_WAIT_WHILE(blk_get_aio_context(blk),
+ atomic_mb_read(&blk->in_flight) > 0);
if (bs) {
bdrv_drained_end(bs);
aio_context_acquire(ctx);
/* We may have -ENOMEDIUM completions in flight */
- AIO_WAIT_WHILE(&blk->wait, ctx,
- atomic_mb_read(&blk->in_flight) > 0);
+ AIO_WAIT_WHILE(ctx, atomic_mb_read(&blk->in_flight) > 0);
aio_context_release(ctx);
}
qapi_event_send_block_io_error(blk_name(blk), !!bs,
bs ? bdrv_get_node_name(bs) : NULL, optype,
action, blk_iostatus_is_enabled(blk),
- error == ENOSPC, strerror(error),
- &error_abort);
+ error == ENOSPC, strerror(error));
}
/* This is done by device models because, while the block layer knows
}
}
-int blk_is_read_only(BlockBackend *blk)
+bool blk_is_read_only(BlockBackend *blk)
{
BlockDriverState *bs = blk_bs(blk);
}
}
-int blk_is_sg(BlockBackend *blk)
+bool blk_is_sg(BlockBackend *blk)
{
BlockDriverState *bs = blk_bs(blk);
if (!bs) {
- return 0;
+ return false;
}
return bdrv_is_sg(bs);
}
-int blk_enable_write_cache(BlockBackend *blk)
+bool blk_enable_write_cache(BlockBackend *blk)
{
return blk->enable_write_cache;
}
BlockDriverState *bs = blk_bs(blk);
char *id;
- /* blk_eject is only called by qdevified devices */
- assert(!blk->legacy_dev);
-
if (bs) {
bdrv_eject(bs, eject_flag);
}
* the frontend experienced a tray event. */
id = blk_get_attached_dev_id(blk);
qapi_event_send_device_tray_moved(blk_name(blk), id,
- eject_flag, &error_abort);
+ eject_flag);
g_free(id);
}
}
}
+/* Returns the minimum request alignment, in bytes; guaranteed nonzero */
+uint32_t blk_get_request_alignment(BlockBackend *blk)
+{
+ BlockDriverState *bs = blk_bs(blk);
+ return bs ? bs->bl.request_alignment : BDRV_SECTOR_SIZE;
+}
+
/* Returns the maximum transfer length, in bytes; guaranteed nonzero */
uint32_t blk_get_max_transfer(BlockBackend *blk)
{
AioContext *blk_get_aio_context(BlockBackend *blk)
{
- return bdrv_get_aio_context(blk_bs(blk));
+ BlockDriverState *bs = blk_bs(blk);
+
+ if (bs) {
+ AioContext *ctx = bdrv_get_aio_context(blk_bs(blk));
+ assert(ctx == blk->ctx);
+ }
+
+ return blk->ctx;
}
static AioContext *blk_aiocb_get_aio_context(BlockAIOCB *acb)
return blk_get_aio_context(blk_acb->blk);
}
-void blk_set_aio_context(BlockBackend *blk, AioContext *new_context)
+static int blk_do_set_aio_context(BlockBackend *blk, AioContext *new_context,
+ bool update_root_node, Error **errp)
{
BlockDriverState *bs = blk_bs(blk);
ThrottleGroupMember *tgm = &blk->public.throttle_group_member;
+ int ret;
if (bs) {
+ if (update_root_node) {
+ ret = bdrv_child_try_set_aio_context(bs, new_context, blk->root,
+ errp);
+ if (ret < 0) {
+ return ret;
+ }
+ }
if (tgm->throttle_state) {
bdrv_drained_begin(bs);
throttle_group_detach_aio_context(tgm);
throttle_group_attach_aio_context(tgm, new_context);
bdrv_drained_end(bs);
}
- bdrv_set_aio_context(bs, new_context);
}
+
+ blk->ctx = new_context;
+ return 0;
+}
+
+int blk_set_aio_context(BlockBackend *blk, AioContext *new_context,
+ Error **errp)
+{
+ return blk_do_set_aio_context(blk, new_context, true, errp);
+}
+
+static bool blk_root_can_set_aio_ctx(BdrvChild *child, AioContext *ctx,
+ GSList **ignore, Error **errp)
+{
+ BlockBackend *blk = child->opaque;
+
+ if (blk->allow_aio_context_change) {
+ return true;
+ }
+
+ /* Only manually created BlockBackends that are not attached to anything
+ * can change their AioContext without updating their user. */
+ if (!blk->name || blk->dev) {
+ /* TODO Add BB name/QOM path */
+ error_setg(errp, "Cannot change iothread of active block backend");
+ return false;
+ }
+
+ return true;
+}
+
+static void blk_root_set_aio_ctx(BdrvChild *child, AioContext *ctx,
+ GSList **ignore)
+{
+ BlockBackend *blk = child->opaque;
+ blk_do_set_aio_context(blk, ctx, false, &error_abort);
}
void blk_add_aio_context_notifier(BlockBackend *blk,
QEMUIOVector *qiov = rwco->iobuf;
rwco->ret = blk_co_pdiscard(rwco->blk, rwco->offset, qiov->size);
+ aio_wait_kick();
}
int blk_pdiscard(BlockBackend *blk, int64_t offset, int bytes)
}
}
+static bool blk_root_drained_poll(BdrvChild *child)
+{
+ BlockBackend *blk = child->opaque;
+ assert(blk->quiesce_counter);
+ return !!blk->in_flight;
+}
+
static void blk_root_drained_end(BdrvChild *child)
{
BlockBackend *blk = child->opaque;
int coroutine_fn blk_co_copy_range(BlockBackend *blk_in, int64_t off_in,
BlockBackend *blk_out, int64_t off_out,
- int bytes, BdrvRequestFlags flags)
+ int bytes, BdrvRequestFlags read_flags,
+ BdrvRequestFlags write_flags)
{
int r;
r = blk_check_byte_request(blk_in, off_in, bytes);
}
return bdrv_co_copy_range(blk_in->root, off_in,
blk_out->root, off_out,
- bytes, flags);
+ bytes, read_flags, write_flags);
+}
+
+const BdrvChild *blk_root(BlockBackend *blk)
+{
+ return blk->root;
}
projects / mirror_qemu.git / blobdiff