[Xen-devel] [PATCH RFC 3/3] virtio infrastructure: example block driver

[Rusty Russell <rusty@xxxxxxxxxxxxxxx>]

Example block driver using virtio.

The block driver uses outbufs with sg[0] being the request information
(struct virtio_blk_outhdr) with the type, sector and inbuf id.  For a
write, the rest of the sg will contain the data to be written.

The first segment of the inbuf is a result code (struct
virtio_blk_inhdr).  For a read, the rest of the sg points to the input
buffer.

TODO:
        1) Ordered tag support.

Signed-off-by: Rusty Russell <rusty@xxxxxxxxxxxxxxx>
---
 drivers/block/Makefile     |    1
 drivers/block/virtio_blk.c |  269 ++++++++++++++++++++++++++++++++++++++++++++
 include/linux/virtio_blk.h |   29 ++++
 3 files changed, 299 insertions(+)

diff -r 8f6c1b0efb6a drivers/block/Makefile
--- a/drivers/block/Makefile    Thu May 31 17:54:08 2007 +1000
+++ b/drivers/block/Makefile    Thu May 31 17:54:13 2007 +1000
@@ -20,6 +20,7 @@ obj-$(CONFIG_BLK_CPQ_CISS_DA)  += cciss.
 obj-$(CONFIG_BLK_CPQ_CISS_DA)  += cciss.o
 obj-$(CONFIG_BLK_DEV_DAC960)   += DAC960.o
 obj-$(CONFIG_CDROM_PKTCDVD)    += pktcdvd.o
+obj-y                          += virtio_blk.o
 
 obj-$(CONFIG_BLK_DEV_UMEM)     += umem.o
 obj-$(CONFIG_BLK_DEV_NBD)      += nbd.o
diff -r 8f6c1b0efb6a drivers/block/virtio_blk.c
--- /dev/null   Thu Jan 01 00:00:00 1970 +0000
+++ b/drivers/block/virtio_blk.c        Thu May 31 17:55:12 2007 +1000
@@ -0,0 +1,270 @@
+#include <linux/spinlock.h>
+#include <linux/blkdev.h>
+#include <linux/hdreg.h>
+#include <linux/virtio.h>
+#include <linux/virtio_blk.h>
+
+static unsigned char virtblk_index = 'a';
+struct virtio_blk
+{
+       struct virtio_device *vdev;
+
+       spinlock_t lock;
+
+       /* The disk structure for the kernel. */
+       struct gendisk *disk;
+
+       /* Request tracking (by inbuf id). */
+       struct list_head reqs;
+
+       mempool_t *pool;
+
+       /* Scatterlist: can be too big for stack. */
+       struct scatterlist sg[1+MAX_PHYS_SEGMENTS];
+};
+
+struct virtblk_req
+{
+       struct list_head list;
+       struct request *req;
+       unsigned long out_id;
+       struct virtio_blk_outhdr out_hdr;
+       struct virtio_blk_inhdr in_hdr;
+       unsigned long in_used;
+};
+
+/* Jens gave me this nice helper to end all chunks of a request. */
+static void end_entire_request(struct request *req, int uptodate)
+{
+       if (end_that_request_first(req, uptodate, req->hard_nr_sectors))
+               BUG();
+       add_disk_randomness(req->rq_disk);
+       end_that_request_last(req, uptodate);
+}
+
+static bool do_write(request_queue_t *q, struct virtio_blk *vblk,
+                    struct virtblk_req *vbr)
+{
+       unsigned long num;
+
+       /* Set up for reply. */
+       vblk->sg[0].page = virt_to_page(&vbr->in_hdr);
+       vblk->sg[0].offset = offset_in_page(&vbr->in_hdr);
+       vblk->sg[0].length = sizeof(vbr->in_hdr);
+       vbr->out_hdr.id = vblk->vdev->ops->add_inbuf(vblk->vdev, vblk->sg, 1,
+                                                    &vbr->in_used);
+       if (IS_ERR_VALUE(vbr->out_hdr.id))
+               goto full;
+
+       /* First sg element points to output header. */
+       vblk->sg[0].page = virt_to_page(&vbr->out_hdr);
+       vblk->sg[0].offset = offset_in_page(&vbr->out_hdr);
+       vblk->sg[0].length = sizeof(vbr->out_hdr);
+
+       num = blk_rq_map_sg(q, vbr->req, vblk->sg+1);
+       vbr->out_id = vblk->vdev->ops->add_outbuf(vblk->vdev, vblk->sg, 1+num,
+                                                 NULL);
+       if (IS_ERR_VALUE(vbr->out_id))
+               goto detach_inbuf_full;
+
+       list_add_tail(&vbr->list, &vblk->reqs);
+       return true;
+
+detach_inbuf_full:
+       vblk->vdev->ops->detach_inbuf(vblk->vdev, vbr->out_hdr.id);
+full:
+       return false;
+}
+
+static bool do_read(request_queue_t *q, struct virtio_blk *vblk,
+                   struct virtblk_req *vbr)
+{
+       unsigned long num;
+
+       /* Set up for reply. */
+       vblk->sg[0].page = virt_to_page(&vbr->in_hdr);
+       vblk->sg[0].offset = offset_in_page(&vbr->in_hdr);
+       vblk->sg[0].length = sizeof(vbr->in_hdr);
+       num = blk_rq_map_sg(q, vbr->req, vblk->sg+1);
+       vbr->out_hdr.id = vblk->vdev->ops->add_inbuf(vblk->vdev, vblk->sg,
+                                                    1+num, &vbr->in_used);
+       if (IS_ERR_VALUE(vbr->out_hdr.id))
+               goto full;
+
+       vblk->sg[0].page = virt_to_page(&vbr->out_hdr);
+       vblk->sg[0].offset = offset_in_page(&vbr->out_hdr);
+       vblk->sg[0].length = sizeof(vbr->out_hdr);
+
+       vbr->out_id = vblk->vdev->ops->add_outbuf(vblk->vdev, vblk->sg, 1,
+                                                 NULL);
+       if (IS_ERR_VALUE(vbr->out_id))
+               goto detach_inbuf_full;
+
+       pr_debug("%s: read at offset %lu %lu->%lu\n",
+                vblk->disk->disk_name,
+                vbr->out_hdr.sector*512, vbr->out_id, vbr->out_hdr.id);
+       list_add_tail(&vbr->list, &vblk->reqs);
+       return true;
+
+detach_inbuf_full:
+       vblk->vdev->ops->detach_inbuf(vblk->vdev, vbr->out_hdr.id);
+full:
+       return false;
+}
+
+static void do_virtblk_request(request_queue_t *q)
+{
+       struct virtio_blk *vblk = NULL;
+       struct request *req;
+       struct virtblk_req *vbr;
+
+       while ((req = elv_next_request(q)) != NULL) {
+               vblk = req->rq_disk->private_data;
+
+               /* FIXME: handle these iff capable. */
+               if (!blk_fs_request(req)) {
+                       pr_debug("Got non-command 0x%08x\n", req->cmd_type);
+                       req->errors++;
+                       end_entire_request(req, 0);
+                       continue;
+               }
+
+               vbr = mempool_alloc(vblk->pool, GFP_ATOMIC);
+               if (!vbr)
+                       goto stop;
+
+               BUG_ON(req->nr_phys_segments > ARRAY_SIZE(vblk->sg));
+               vbr->req = req;
+               vbr->out_hdr.type = rq_data_dir(req);
+               vbr->out_hdr.sector = req->sector;
+               vbr->in_used = 0;
+
+               if (rq_data_dir(req) == WRITE) {
+                       if (!do_write(q, vblk, vbr))
+                               goto stop;
+               } else {
+                       if (!do_read(q, vblk, vbr))
+                               goto stop;
+               }
+               blkdev_dequeue_request(req);
+       }
+
+sync:
+       if (vblk)
+               virtio_sync(vblk->vdev);
+       return;
+
+stop:
+       /* Queue full?  Wait. */
+       blk_stop_queue(q);
+       mempool_free(vbr, vblk->pool);
+       goto sync;
+}
+
+int virtblk_interrupt(struct gendisk *disk)
+{
+       struct virtio_blk *vblk = disk->private_data;
+       struct virtblk_req *i, *next;
+       unsigned long flags;
+
+       spin_lock_irqsave(&vblk->lock, flags);
+       list_for_each_entry_safe(i, next, &vblk->reqs, list) {
+               /* Reply hasn't come back? */
+               if (!i->in_used)
+                       continue;
+               /* Make sure other side can no longer read/write */
+               vblk->vdev->ops->detach_outbuf(vblk->vdev, i->out_id);
+               vblk->vdev->ops->detach_inbuf(vblk->vdev, i->out_hdr.id);
+               pr_debug("%s: finished request %lu %s\n", disk->disk_name,
+                         i->out_id, i->in_hdr.status == 1 ? "OK" : "FAILED");
+               /* Finish and free request */
+               end_entire_request(i->req, i->in_hdr.status == 1);
+               list_del(&i->list);
+               mempool_free(i, vblk->pool);
+       }
+       blk_start_queue(disk->queue);
+       spin_unlock_irqrestore(&vblk->lock, flags);
+       return IRQ_HANDLED;
+}
+EXPORT_SYMBOL_GPL(virtblk_interrupt);
+
+static struct block_device_operations virtblk_fops = {
+       .owner = THIS_MODULE,
+};
+
+struct gendisk *virtblk_probe(struct virtio_device *vdev)
+{
+       struct virtio_blk *vblk;
+       int err, major;
+
+       vblk = kmalloc(sizeof(*vblk), GFP_KERNEL);
+       if (!vblk) {
+               err = -ENOMEM;
+               goto out;
+       }
+
+       spin_lock_init(&vblk->lock);
+       INIT_LIST_HEAD(&vblk->reqs);
+       vblk->vdev = vdev;
+
+       vblk->pool = mempool_create_kmalloc_pool(1,sizeof(struct virtblk_req));
+       if (!vblk->pool) {
+               err = -ENOMEM;
+               goto out_free_vblk;
+       }
+
+       major = register_blkdev(0, "virtblk");
+       if (major < 0) {
+               err = major;
+               goto out_mempool;
+       }
+
+       /* FIXME: How many partitions?  How long is a piece of string? */
+       vblk->disk = alloc_disk(1 << 3);
+       if (!vblk->disk) {
+               err = -ENOMEM;
+               goto out_unregister_blkdev;
+       }
+
+       vblk->disk->queue = blk_init_queue(do_virtblk_request, &vblk->lock);
+       if (!vblk->disk->queue) {
+               err = -ENOMEM;
+               goto out_put_disk;
+       }
+
+       sprintf(vblk->disk->disk_name, "vb%c", virtblk_index++);
+       vblk->disk->major = major;
+       vblk->disk->first_minor = 0;
+       vblk->disk->private_data = vblk;
+       vblk->disk->fops = &virtblk_fops;
+
+       /* Caller can do blk_queue_max_hw_segments(), set_capacity()
+        * etc then add_disk(). */
+       return vblk->disk;
+
+out_put_disk:
+       put_disk(vblk->disk);
+out_unregister_blkdev:
+       unregister_blkdev(major, "virtblk");
+out_mempool:
+       mempool_destroy(vblk->pool);
+out_free_vblk:
+       kfree(vblk);
+out:
+       return ERR_PTR(err);
+}
+EXPORT_SYMBOL_GPL(virtblk_probe);
+
+void virtblk_remove(struct gendisk *disk)
+{
+       struct virtio_blk *vblk = disk->private_data;
+       int major = vblk->disk->major;
+
+       BUG_ON(!list_empty(&vblk->reqs));
+       blk_cleanup_queue(vblk->disk->queue);
+       put_disk(vblk->disk);
+       unregister_blkdev(major, "virtblk");
+       mempool_destroy(vblk->pool);
+       kfree(vblk);
+}
+EXPORT_SYMBOL_GPL(virtblk_remove);
diff -r 8f6c1b0efb6a include/linux/virtio_blk.h
--- /dev/null   Thu Jan 01 00:00:00 1970 +0000
+++ b/include/linux/virtio_blk.h        Thu May 31 17:54:13 2007 +1000
@@ -0,0 +1,29 @@
+#ifndef _LINUX_VIRTIO_BLK_H
+#define _LINUX_VIRTIO_BLK_H
+#include <linux/types.h>
+struct gendisk;
+struct virtio_device;
+struct hd_geometry;
+
+/* This is the first element of the scatter-gather list. */
+struct virtio_blk_outhdr
+{
+       /* 0 == read, 1 == write */
+       u32 type;
+       /* Sector (ie. 512 byte offset) */
+       unsigned long sector;
+       /* Where to put reply. */
+       unsigned long id;
+};
+
+struct virtio_blk_inhdr
+{
+       /* 1 = OK, 0 = not ok. */
+       unsigned long status;
+};
+
+struct gendisk *virtblk_probe(struct virtio_device *vdev);
+int virtblk_interrupt(struct gendisk *disk);
+void virtblk_remove(struct gendisk *disk);
+
+#endif /* _LINUX_VIRTIO_BLK_H */



_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-devel