[PATCH] virtio_blk: add block topology support

[Christoph Hellwig <hch@lst.de>]

Allow reading various alignment values from the config page.  This
allows the guest to much better align I/O requests depending on the
storage topology.

Note that the formats for the config values appear a bit messed up,
but we follow the formats used by ATA and SCSI so they are expected in
the storage world.

Signed-off-by: Christoph Hellwig <hch@lst.de>

Index: linux-2.6/drivers/block/virtio_blk.c
===================================================================
--- linux-2.6.orig/drivers/block/virtio_blk.c	2010-01-29 11:13:38.509004113 +0100
+++ linux-2.6/drivers/block/virtio_blk.c	2010-01-29 19:36:44.183006458 +0100
@@ -243,10 +243,12 @@ static int index_to_minor(int index)
 static int __devinit virtblk_probe(struct virtio_device *vdev)
 {
 	struct virtio_blk *vblk;
+	struct request_queue *q;
 	int err;
 	u64 cap;
 	u32 v, blk_size, sg_elems, opt_io_size;
 	u16 min_io_size;
+	u8 physical_block_exp, alignment_offset;
 
 	if (index_to_minor(index) >= 1 << MINORBITS)
 		return -ENOSPC;
@@ -293,13 +295,13 @@ static int __devinit virtblk_probe(struc
 		goto out_mempool;
 	}
 
-	vblk->disk->queue = blk_init_queue(do_virtblk_request, &vblk->lock);
-	if (!vblk->disk->queue) {
+	q = vblk->disk->queue = blk_init_queue(do_virtblk_request, &vblk->lock);
+	if (!q) {
 		err = -ENOMEM;
 		goto out_put_disk;
 	}
 
-	vblk->disk->queue->queuedata = vblk;
+	q->queuedata = vblk;
 
 	if (index < 26) {
 		sprintf(vblk->disk->disk_name, "vd%c", 'a' + index % 26);
@@ -323,10 +325,10 @@ static int __devinit virtblk_probe(struc
 
 	/* If barriers are supported, tell block layer that queue is ordered */
 	if (virtio_has_feature(vdev, VIRTIO_BLK_F_FLUSH))
-		blk_queue_ordered(vblk->disk->queue, QUEUE_ORDERED_DRAIN_FLUSH,
+		blk_queue_ordered(q, QUEUE_ORDERED_DRAIN_FLUSH,
 				  virtblk_prepare_flush);
 	else if (virtio_has_feature(vdev, VIRTIO_BLK_F_BARRIER))
-		blk_queue_ordered(vblk->disk->queue, QUEUE_ORDERED_TAG, NULL);
+		blk_queue_ordered(q, QUEUE_ORDERED_TAG, NULL);
 
 	/* If disk is read-only in the host, the guest should obey */
 	if (virtio_has_feature(vdev, VIRTIO_BLK_F_RO))
@@ -345,14 +347,14 @@ static int __devinit virtblk_probe(struc
 	set_capacity(vblk->disk, cap);
 
 	/* We can handle whatever the host told us to handle. */
-	blk_queue_max_phys_segments(vblk->disk->queue, vblk->sg_elems-2);
-	blk_queue_max_hw_segments(vblk->disk->queue, vblk->sg_elems-2);
+	blk_queue_max_phys_segments(q, vblk->sg_elems-2);
+	blk_queue_max_hw_segments(q, vblk->sg_elems-2);
 
 	/* No need to bounce any requests */
-	blk_queue_bounce_limit(vblk->disk->queue, BLK_BOUNCE_ANY);
+	blk_queue_bounce_limit(q, BLK_BOUNCE_ANY);
 
 	/* No real sector limit. */
-	blk_queue_max_sectors(vblk->disk->queue, -1U);
+	blk_queue_max_sectors(q, -1U);
 
 	/* Host can optionally specify maximum segment size and number of
 	 * segments. */
@@ -360,29 +362,46 @@ static int __devinit virtblk_probe(struc
 				offsetof(struct virtio_blk_config, size_max),
 				&v);
 	if (!err)
-		blk_queue_max_segment_size(vblk->disk->queue, v);
+		blk_queue_max_segment_size(q, v);
 	else
-		blk_queue_max_segment_size(vblk->disk->queue, -1U);
+		blk_queue_max_segment_size(q, -1U);
 
 	/* Host can optionally specify the block size of the device */
 	err = virtio_config_val(vdev, VIRTIO_BLK_F_BLK_SIZE,
 				offsetof(struct virtio_blk_config, blk_size),
 				&blk_size);
 	if (!err)
-		blk_queue_logical_block_size(vblk->disk->queue, blk_size);
+		blk_queue_logical_block_size(q, blk_size);
 
 	/* Use topology information if available */
-	err = virtio_config_val(vdev, VIRTIO_BLK_F_BLK_TOPOLOGY,
+	err = virtio_config_val(vdev, VIRTIO_BLK_F_TOPOLOGY,
+			offsetof(struct virtio_blk_config, physical_block_exp),
+			&physical_block_exp);
+	if (!err && physical_block_exp) {
+		u16 physical_block_size =
+			queue_logical_block_size(q) * (1 << physical_block_exp);
+		blk_queue_physical_block_size(q, physical_block_size);
+	}
+
+	err = virtio_config_val(vdev, VIRTIO_BLK_F_TOPOLOGY,
+			offsetof(struct virtio_blk_config, alignment_offset),
+			&alignment_offset);
+	if (!err && alignment_offset) {
+		u16 alignment = queue_logical_block_size(q) * alignment_offset;
+		blk_queue_alignment_offset(q, alignment);
+	}
+
+	err = virtio_config_val(vdev, VIRTIO_BLK_F_TOPOLOGY,
 			offsetof(struct virtio_blk_config, min_io_size),
 			&min_io_size);
-	if (!err)
-		blk_queue_io_min(vblk->disk->queue, min_io_size);
+	if (!err && min_io_size)
+		blk_queue_io_min(q, min_io_size);
 
-	err = virtio_config_val(vdev, VIRTIO_BLK_F_BLK_TOPOLOGY,
+	err = virtio_config_val(vdev, VIRTIO_BLK_F_TOPOLOGY,
 			offsetof(struct virtio_blk_config, opt_io_size),
 			&opt_io_size);
-	if (!err)
-		blk_queue_io_opt(vblk->disk->queue, opt_io_size);
+	if (!err && opt_io_size)
+		blk_queue_io_opt(q, opt_io_size);
 
 
 	add_disk(vblk->disk);
Index: linux-2.6/include/linux/virtio_blk.h
===================================================================
--- linux-2.6.orig/include/linux/virtio_blk.h	2010-01-29 11:11:20.436254184 +0100
+++ linux-2.6/include/linux/virtio_blk.h	2010-01-29 19:35:44.431016843 +0100
@@ -15,6 +15,7 @@
 #define VIRTIO_BLK_F_BLK_SIZE	6	/* Block size of disk is available*/
 #define VIRTIO_BLK_F_SCSI	7	/* Supports scsi command passthru */
 #define VIRTIO_BLK_F_FLUSH	9	/* Cache flush command support */
+#define VIRTIO_BLK_F_TOPOLOGY	10	/* Topology information is available */
 
 struct virtio_blk_config {
 	/* The capacity (in 512-byte sectors). */
@@ -29,8 +30,20 @@ struct virtio_blk_config {
 		__u8 heads;
 		__u8 sectors;
 	} geometry;
+
 	/* block size of device (if VIRTIO_BLK_F_BLK_SIZE) */
 	__u32 blk_size;
+
+	/* the next 4 entries are covered by VIRTIO_BLK_F_TOPOLOGY  */
+	/* exponent for physical block per logical block. */
+	__u8 physical_block_exp;
+	/* alignment offset in logical blocks. */
+	__u8 alignment_offset;
+	/* minimum I/O size without performance penalty in bytes. */
+	__u16 min_io_size;
+	/* optimal sustained I/O size. */
+	__u32 opt_io_size;
+
 } __attribute__((packed));
 
 /*