#include <xen/grant_table.h>
#include <xen/events.h>
#include <xen/page.h>
+#include <xen/platform_pci.h>
#include <xen/interface/grant_table.h>
#include <xen/interface/io/blkif.h>
*/
struct blkfront_info
{
+ struct mutex mutex;
struct xenbus_device *xbdev;
struct gendisk *gd;
int vdevice;
unsigned long shadow_free;
int feature_barrier;
int is_ready;
-
- /**
- * The number of people holding this device open. We won't allow a
- * hot-unplug unless this is 0.
- */
- int users;
};
static DEFINE_SPINLOCK(blkif_io_lock);
static int xlvbd_barrier(struct blkfront_info *info)
{
int err;
+ const char *barrier;
- err = blk_queue_ordered(info->rq,
- info->feature_barrier ? QUEUE_ORDERED_DRAIN : QUEUE_ORDERED_NONE);
+ switch (info->feature_barrier) {
+ case QUEUE_ORDERED_DRAIN: barrier = "enabled (drain)"; break;
+ case QUEUE_ORDERED_TAG: barrier = "enabled (tag)"; break;
+ case QUEUE_ORDERED_NONE: barrier = "disabled"; break;
+ default: return -EINVAL;
+ }
+
+ err = blk_queue_ordered(info->rq, info->feature_barrier);
if (err)
return err;
printk(KERN_INFO "blkfront: %s: barriers %s\n",
- info->gd->disk_name,
- info->feature_barrier ? "enabled" : "disabled");
+ info->gd->disk_name, barrier);
return 0;
}
info->rq = gd->queue;
info->gd = gd;
- if (info->feature_barrier)
- xlvbd_barrier(info);
+ xlvbd_barrier(info);
if (vdisk_info & VDISK_READONLY)
set_disk_ro(gd, 1);
printk(KERN_WARNING "blkfront: %s: write barrier op failed\n",
info->gd->disk_name);
error = -EOPNOTSUPP;
- info->feature_barrier = 0;
+ info->feature_barrier = QUEUE_ORDERED_NONE;
xlvbd_barrier(info);
}
/* fall through */
return err;
}
-
/**
* Entry point to this code when a new device is created. Allocate the basic
* structures and the ring buffer for communication with the backend, and
}
}
+ if (xen_hvm_domain()) {
+ char *type;
+ int len;
+ /* no unplug has been done: do not hook devices != xen vbds */
+ if (xen_platform_pci_unplug & XEN_UNPLUG_IGNORE) {
+ int major;
+
+ if (!VDEV_IS_EXTENDED(vdevice))
+ major = BLKIF_MAJOR(vdevice);
+ else
+ major = XENVBD_MAJOR;
+
+ if (major != XENVBD_MAJOR) {
+ printk(KERN_INFO
+ "%s: HVM does not support vbd %d as xen block device\n",
+ __FUNCTION__, vdevice);
+ return -ENODEV;
+ }
+ }
+ /* do not create a PV cdrom device if we are an HVM guest */
+ type = xenbus_read(XBT_NIL, dev->nodename, "device-type", &len);
+ if (IS_ERR(type))
+ return -ENODEV;
+ if (strncmp(type, "cdrom", 5) == 0) {
+ kfree(type);
+ return -ENODEV;
+ }
+ kfree(type);
+ }
info = kzalloc(sizeof(*info), GFP_KERNEL);
if (!info) {
xenbus_dev_fatal(dev, -ENOMEM, "allocating info structure");
return -ENOMEM;
}
+ mutex_init(&info->mutex);
info->xbdev = dev;
info->vdevice = vdevice;
info->connected = BLKIF_STATE_DISCONNECTED;
return err;
}
+static void
+blkfront_closing(struct blkfront_info *info)
+{
+ struct xenbus_device *xbdev = info->xbdev;
+ struct block_device *bdev = NULL;
+
+ mutex_lock(&info->mutex);
+
+ if (xbdev->state == XenbusStateClosing) {
+ mutex_unlock(&info->mutex);
+ return;
+ }
+
+ if (info->gd)
+ bdev = bdget_disk(info->gd, 0);
+
+ mutex_unlock(&info->mutex);
+
+ if (!bdev) {
+ xenbus_frontend_closed(xbdev);
+ return;
+ }
+
+ mutex_lock(&bdev->bd_mutex);
+
+ if (bdev->bd_openers) {
+ xenbus_dev_error(xbdev, -EBUSY,
+ "Device in use; refusing to close");
+ xenbus_switch_state(xbdev, XenbusStateClosing);
+ } else {
+ xlvbd_release_gendisk(info);
+ xenbus_frontend_closed(xbdev);
+ }
+
+ mutex_unlock(&bdev->bd_mutex);
+ bdput(bdev);
+}
/*
* Invoked when the backend is finally 'ready' (and has told produced
unsigned long sector_size;
unsigned int binfo;
int err;
+ int barrier;
switch (info->connected) {
case BLKIF_STATE_CONNECTED:
}
err = xenbus_gather(XBT_NIL, info->xbdev->otherend,
- "feature-barrier", "%lu", &info->feature_barrier,
+ "feature-barrier", "%lu", &barrier,
NULL);
+
+ /*
+ * If there's no "feature-barrier" defined, then it means
+ * we're dealing with a very old backend which writes
+ * synchronously; draining will do what needs to get done.
+ *
+ * If there are barriers, then we can do full queued writes
+ * with tagged barriers.
+ *
+ * If barriers are not supported, then there's no much we can
+ * do, so just set ordering to NONE.
+ */
if (err)
- info->feature_barrier = 0;
+ info->feature_barrier = QUEUE_ORDERED_DRAIN;
+ else if (barrier)
+ info->feature_barrier = QUEUE_ORDERED_TAG;
+ else
+ info->feature_barrier = QUEUE_ORDERED_NONE;
err = xlvbd_alloc_gendisk(sectors, info, binfo, sector_size);
if (err) {
enum xenbus_state backend_state)
{
struct blkfront_info *info = dev_get_drvdata(&dev->dev);
- struct block_device *bd;
dev_dbg(&dev->dev, "blkfront:blkback_changed to state %d.\n", backend_state);
break;
case XenbusStateClosing:
- if (info->gd == NULL) {
- xenbus_frontend_closed(dev);
- break;
- }
- bd = bdget_disk(info->gd, 0);
- if (bd == NULL)
- xenbus_dev_fatal(dev, -ENODEV, "bdget failed");
-
- mutex_lock(&bd->bd_mutex);
- if (info->users > 0)
- xenbus_dev_error(dev, -EBUSY,
- "Device in use; refusing to close");
- else {
- xlvbd_release_gendisk(info);
- xenbus_frontend_closed(info->xbdev);
- }
-
- mutex_unlock(&bd->bd_mutex);
- bdput(bd);
+ blkfront_closing(info);
break;
}
}
-static int blkfront_remove(struct xenbus_device *dev)
+static int blkfront_remove(struct xenbus_device *xbdev)
{
- struct blkfront_info *info = dev_get_drvdata(&dev->dev);
+ struct blkfront_info *info = dev_get_drvdata(&xbdev->dev);
+ struct block_device *bdev = NULL;
+ struct gendisk *disk;
- dev_dbg(&dev->dev, "blkfront_remove: %s removed\n", dev->nodename);
+ dev_dbg(&xbdev->dev, "%s removed", xbdev->nodename);
blkif_free(info, 0);
- if(info->users == 0)
+ mutex_lock(&info->mutex);
+
+ disk = info->gd;
+ if (disk)
+ bdev = bdget_disk(disk, 0);
+
+ info->xbdev = NULL;
+ mutex_unlock(&info->mutex);
+
+ if (!bdev) {
kfree(info);
- else
- info->xbdev = NULL;
+ return 0;
+ }
+
+ /*
+ * The xbdev was removed before we reached the Closed
+ * state. See if it's safe to remove the disk. If the bdev
+ * isn't closed yet, we let release take care of it.
+ */
+
+ mutex_lock(&bdev->bd_mutex);
+ info = disk->private_data;
+
+ dev_warn(disk_to_dev(disk),
+ "%s was hot-unplugged, %d stale handles\n",
+ xbdev->nodename, bdev->bd_openers);
+
+ if (info && !bdev->bd_openers) {
+ xlvbd_release_gendisk(info);
+ disk->private_data = NULL;
+ kfree(info);
+ }
+
+ mutex_unlock(&bdev->bd_mutex);
+ bdput(bdev);
return 0;
}
static int blkif_open(struct block_device *bdev, fmode_t mode)
{
- struct blkfront_info *info = bdev->bd_disk->private_data;
-
- if (!info->xbdev)
- return -ENODEV;
+ struct gendisk *disk = bdev->bd_disk;
+ struct blkfront_info *info;
+ int err = 0;
lock_kernel();
- info->users++;
- unlock_kernel();
- return 0;
+ info = disk->private_data;
+ if (!info) {
+ /* xbdev gone */
+ err = -ERESTARTSYS;
+ goto out;
+ }
+
+ mutex_lock(&info->mutex);
+
+ if (!info->gd)
+ /* xbdev is closed */
+ err = -ERESTARTSYS;
+
+ mutex_unlock(&info->mutex);
+
+out:
+ unlock_kernel();
+ return err;
}
static int blkif_release(struct gendisk *disk, fmode_t mode)
{
struct blkfront_info *info = disk->private_data;
+ struct block_device *bdev;
+ struct xenbus_device *xbdev;
+
lock_kernel();
- info->users--;
- if (info->users == 0) {
- /* Check whether we have been instructed to close. We will
- have ignored this request initially, as the device was
- still mounted. */
- struct xenbus_device *dev = info->xbdev;
-
- if (!dev) {
- xlvbd_release_gendisk(info);
- kfree(info);
- } else if (xenbus_read_driver_state(dev->otherend)
- == XenbusStateClosing && info->is_ready) {
- xlvbd_release_gendisk(info);
- xenbus_frontend_closed(dev);
- }
+
+ bdev = bdget_disk(disk, 0);
+ bdput(bdev);
+
+ if (bdev->bd_openers)
+ goto out;
+
+ /*
+ * Check if we have been instructed to close. We will have
+ * deferred this request, because the bdev was still open.
+ */
+
+ mutex_lock(&info->mutex);
+ xbdev = info->xbdev;
+
+ if (xbdev && xbdev->state == XenbusStateClosing) {
+ /* pending switch to state closed */
+ dev_info(disk_to_dev(bdev->bd_disk), "releasing disk\n");
+ xlvbd_release_gendisk(info);
+ xenbus_frontend_closed(info->xbdev);
+ }
+
+ mutex_unlock(&info->mutex);
+
+ if (!xbdev) {
+ /* sudden device removal */
+ dev_info(disk_to_dev(bdev->bd_disk), "releasing disk\n");
+ xlvbd_release_gendisk(info);
+ disk->private_data = NULL;
+ kfree(info);
}
+
+out:
unlock_kernel();
return 0;
}