[SPARC64]: Add Sun LDOM virtual disk driver.
authorDavid S. Miller <davem@sunset.davemloft.net>
Mon, 16 Jul 2007 11:03:56 +0000 (04:03 -0700)
committerDavid S. Miller <davem@sunset.davemloft.net>
Mon, 16 Jul 2007 11:03:56 +0000 (04:03 -0700)
Signed-off-by: David S. Miller <davem@davemloft.net>
drivers/block/Kconfig
drivers/block/Makefile
drivers/block/sunvdc.c [new file with mode: 0644]

index 6e23af1..d9867fd 100644 (file)
@@ -423,6 +423,13 @@ config ATA_OVER_ETH
        This driver provides Support for ATA over Ethernet block
        devices like the Coraid EtherDrive (R) Storage Blade.
 
+config SUNVDC
+       tristate "Sun Virtual Disk Client support"
+       depends on SUN_LDOMS
+       help
+         Support for virtual disk devices as a client under Sun
+         Logical Domains.
+
 source "drivers/s390/block/Kconfig"
 
 endif # BLK_DEV
index e5f98ac..43371c5 100644 (file)
@@ -19,6 +19,7 @@ obj-$(CONFIG_BLK_CPQ_DA)      += cpqarray.o
 obj-$(CONFIG_BLK_CPQ_CISS_DA)  += cciss.o
 obj-$(CONFIG_BLK_DEV_DAC960)   += DAC960.o
 obj-$(CONFIG_CDROM_PKTCDVD)    += pktcdvd.o
+obj-$(CONFIG_SUNVDC)           += sunvdc.o
 
 obj-$(CONFIG_BLK_DEV_UMEM)     += umem.o
 obj-$(CONFIG_BLK_DEV_NBD)      += nbd.o
diff --git a/drivers/block/sunvdc.c b/drivers/block/sunvdc.c
new file mode 100644 (file)
index 0000000..8dbbeac
--- /dev/null
@@ -0,0 +1,970 @@
+/* sunvdc.c: Sun LDOM Virtual Disk Client.
+ *
+ * Copyright (C) 2007 David S. Miller <davem@davemloft.net>
+ */
+
+#include <linux/module.h>
+#include <linux/kernel.h>
+#include <linux/types.h>
+#include <linux/blkdev.h>
+#include <linux/hdreg.h>
+#include <linux/genhd.h>
+#include <linux/slab.h>
+#include <linux/spinlock.h>
+#include <linux/completion.h>
+#include <linux/delay.h>
+#include <linux/init.h>
+#include <linux/list.h>
+
+#include <asm/vio.h>
+#include <asm/ldc.h>
+
+#define DRV_MODULE_NAME                "sunvdc"
+#define PFX DRV_MODULE_NAME    ": "
+#define DRV_MODULE_VERSION     "1.0"
+#define DRV_MODULE_RELDATE     "June 25, 2007"
+
+static char version[] __devinitdata =
+       DRV_MODULE_NAME ".c:v" DRV_MODULE_VERSION " (" DRV_MODULE_RELDATE ")\n";
+MODULE_AUTHOR("David S. Miller (davem@davemloft.net)");
+MODULE_DESCRIPTION("Sun LDOM virtual disk client driver");
+MODULE_LICENSE("GPL");
+MODULE_VERSION(DRV_MODULE_VERSION);
+
+#define VDC_TX_RING_SIZE       256
+
+#define WAITING_FOR_LINK_UP    0x01
+#define WAITING_FOR_TX_SPACE   0x02
+#define WAITING_FOR_GEN_CMD    0x04
+#define WAITING_FOR_ANY                -1
+
+struct vdc_req_entry {
+       struct request          *req;
+};
+
+struct vdc_port {
+       struct vio_driver_state vio;
+
+       struct vdc              *vp;
+
+       struct gendisk          *disk;
+
+       struct vdc_completion   *cmp;
+
+       u64                     req_id;
+       u64                     seq;
+       struct vdc_req_entry    rq_arr[VDC_TX_RING_SIZE];
+
+       unsigned long           ring_cookies;
+
+       u64                     max_xfer_size;
+       u32                     vdisk_block_size;
+
+       /* The server fills these in for us in the disk attribute
+        * ACK packet.
+        */
+       u64                     operations;
+       u32                     vdisk_size;
+       u8                      vdisk_type;
+       u8                      dev_no;
+
+       char                    disk_name[32];
+
+       struct vio_disk_geom    geom;
+       struct vio_disk_vtoc    label;
+
+       struct list_head        list;
+};
+
+static inline struct vdc_port *to_vdc_port(struct vio_driver_state *vio)
+{
+       return container_of(vio, struct vdc_port, vio);
+}
+
+struct vdc {
+       /* Protects prot_list.  */
+       spinlock_t              lock;
+
+       struct vio_dev          *dev;
+
+       struct list_head        port_list;
+};
+
+/* Ordered from largest major to lowest */
+static struct vio_version vdc_versions[] = {
+       { .major = 1, .minor = 0 },
+};
+
+#define VDCBLK_NAME    "vdisk"
+static int vdc_major;
+#define PARTITION_SHIFT        3
+
+static inline u32 vdc_tx_dring_avail(struct vio_dring_state *dr)
+{
+       return vio_dring_avail(dr, VDC_TX_RING_SIZE);
+}
+
+static int vdc_getgeo(struct block_device *bdev, struct hd_geometry *geo)
+{
+       struct gendisk *disk = bdev->bd_disk;
+       struct vdc_port *port = disk->private_data;
+
+       geo->heads = (u8) port->geom.num_hd;
+       geo->sectors = (u8) port->geom.num_sec;
+       geo->cylinders = port->geom.num_cyl;
+
+       return 0;
+}
+
+static struct block_device_operations vdc_fops = {
+       .owner          = THIS_MODULE,
+       .getgeo         = vdc_getgeo,
+};
+
+static void vdc_finish(struct vio_driver_state *vio, int err, int waiting_for)
+{
+       if (vio->cmp &&
+           (waiting_for == -1 ||
+            vio->cmp->waiting_for == waiting_for)) {
+               vio->cmp->err = err;
+               complete(&vio->cmp->com);
+               vio->cmp = NULL;
+       }
+}
+
+static void vdc_handshake_complete(struct vio_driver_state *vio)
+{
+       vdc_finish(vio, 0, WAITING_FOR_LINK_UP);
+}
+
+static int vdc_handle_unknown(struct vdc_port *port, void *arg)
+{
+       struct vio_msg_tag *pkt = arg;
+
+       printk(KERN_ERR PFX "Received unknown msg [%02x:%02x:%04x:%08x]\n",
+              pkt->type, pkt->stype, pkt->stype_env, pkt->sid);
+       printk(KERN_ERR PFX "Resetting connection.\n");
+
+       ldc_disconnect(port->vio.lp);
+
+       return -ECONNRESET;
+}
+
+static int vdc_send_attr(struct vio_driver_state *vio)
+{
+       struct vdc_port *port = to_vdc_port(vio);
+       struct vio_disk_attr_info pkt;
+
+       memset(&pkt, 0, sizeof(pkt));
+
+       pkt.tag.type = VIO_TYPE_CTRL;
+       pkt.tag.stype = VIO_SUBTYPE_INFO;
+       pkt.tag.stype_env = VIO_ATTR_INFO;
+       pkt.tag.sid = vio_send_sid(vio);
+
+       pkt.xfer_mode = VIO_DRING_MODE;
+       pkt.vdisk_block_size = port->vdisk_block_size;
+       pkt.max_xfer_size = port->max_xfer_size;
+
+       viodbg(HS, "SEND ATTR xfer_mode[0x%x] blksz[%u] max_xfer[%lu]\n",
+              pkt.xfer_mode, pkt.vdisk_block_size, pkt.max_xfer_size);
+
+       return vio_ldc_send(&port->vio, &pkt, sizeof(pkt));
+}
+
+static int vdc_handle_attr(struct vio_driver_state *vio, void *arg)
+{
+       struct vdc_port *port = to_vdc_port(vio);
+       struct vio_disk_attr_info *pkt = arg;
+
+       viodbg(HS, "GOT ATTR stype[0x%x] ops[%lx] disk_size[%lu] disk_type[%x] "
+              "xfer_mode[0x%x] blksz[%u] max_xfer[%lu]\n",
+              pkt->tag.stype, pkt->operations,
+              pkt->vdisk_size, pkt->vdisk_type,
+              pkt->xfer_mode, pkt->vdisk_block_size,
+              pkt->max_xfer_size);
+
+       if (pkt->tag.stype == VIO_SUBTYPE_ACK) {
+               switch (pkt->vdisk_type) {
+               case VD_DISK_TYPE_DISK:
+               case VD_DISK_TYPE_SLICE:
+                       break;
+
+               default:
+                       printk(KERN_ERR PFX "%s: Bogus vdisk_type 0x%x\n",
+                              vio->name, pkt->vdisk_type);
+                       return -ECONNRESET;
+               }
+
+               if (pkt->vdisk_block_size > port->vdisk_block_size) {
+                       printk(KERN_ERR PFX "%s: BLOCK size increased "
+                              "%u --> %u\n",
+                              vio->name,
+                              port->vdisk_block_size, pkt->vdisk_block_size);
+                       return -ECONNRESET;
+               }
+
+               port->operations = pkt->operations;
+               port->vdisk_size = pkt->vdisk_size;
+               port->vdisk_type = pkt->vdisk_type;
+               if (pkt->max_xfer_size < port->max_xfer_size)
+                       port->max_xfer_size = pkt->max_xfer_size;
+               port->vdisk_block_size = pkt->vdisk_block_size;
+               return 0;
+       } else {
+               printk(KERN_ERR PFX "%s: Attribute NACK\n", vio->name);
+
+               return -ECONNRESET;
+       }
+}
+
+static void vdc_end_special(struct vdc_port *port, struct vio_disk_desc *desc)
+{
+       int err = desc->status;
+
+       vdc_finish(&port->vio, -err, WAITING_FOR_GEN_CMD);
+}
+
+static void vdc_end_request(struct request *req, int uptodate, int num_sectors)
+{
+       if (end_that_request_first(req, uptodate, num_sectors))
+               return;
+       add_disk_randomness(req->rq_disk);
+       end_that_request_last(req, uptodate);
+}
+
+static void vdc_end_one(struct vdc_port *port, struct vio_dring_state *dr,
+                       unsigned int index)
+{
+       struct vio_disk_desc *desc = vio_dring_entry(dr, index);
+       struct vdc_req_entry *rqe = &port->rq_arr[index];
+       struct request *req;
+
+       if (unlikely(desc->hdr.state != VIO_DESC_DONE))
+               return;
+
+       ldc_unmap(port->vio.lp, desc->cookies, desc->ncookies);
+       desc->hdr.state = VIO_DESC_FREE;
+       dr->cons = (index + 1) & (VDC_TX_RING_SIZE - 1);
+
+       req = rqe->req;
+       if (req == NULL) {
+               vdc_end_special(port, desc);
+               return;
+       }
+
+       rqe->req = NULL;
+
+       vdc_end_request(req, !desc->status, desc->size >> 9);
+
+       if (blk_queue_stopped(port->disk->queue))
+               blk_start_queue(port->disk->queue);
+}
+
+static int vdc_ack(struct vdc_port *port, void *msgbuf)
+{
+       struct vio_dring_state *dr = &port->vio.drings[VIO_DRIVER_TX_RING];
+       struct vio_dring_data *pkt = msgbuf;
+
+       if (unlikely(pkt->dring_ident != dr->ident ||
+                    pkt->start_idx != pkt->end_idx ||
+                    pkt->start_idx >= VDC_TX_RING_SIZE))
+               return 0;
+
+       vdc_end_one(port, dr, pkt->start_idx);
+
+       return 0;
+}
+
+static int vdc_nack(struct vdc_port *port, void *msgbuf)
+{
+       /* XXX Implement me XXX */
+       return 0;
+}
+
+static void vdc_event(void *arg, int event)
+{
+       struct vdc_port *port = arg;
+       struct vio_driver_state *vio = &port->vio;
+       unsigned long flags;
+       int err;
+
+       spin_lock_irqsave(&vio->lock, flags);
+
+       if (unlikely(event == LDC_EVENT_RESET ||
+                    event == LDC_EVENT_UP)) {
+               vio_link_state_change(vio, event);
+               spin_unlock_irqrestore(&vio->lock, flags);
+               return;
+       }
+
+       if (unlikely(event != LDC_EVENT_DATA_READY)) {
+               printk(KERN_WARNING PFX "Unexpected LDC event %d\n", event);
+               spin_unlock_irqrestore(&vio->lock, flags);
+               return;
+       }
+
+       err = 0;
+       while (1) {
+               union {
+                       struct vio_msg_tag tag;
+                       u64 raw[8];
+               } msgbuf;
+
+               err = ldc_read(vio->lp, &msgbuf, sizeof(msgbuf));
+               if (unlikely(err < 0)) {
+                       if (err == -ECONNRESET)
+                               vio_conn_reset(vio);
+                       break;
+               }
+               if (err == 0)
+                       break;
+               viodbg(DATA, "TAG [%02x:%02x:%04x:%08x]\n",
+                      msgbuf.tag.type,
+                      msgbuf.tag.stype,
+                      msgbuf.tag.stype_env,
+                      msgbuf.tag.sid);
+               err = vio_validate_sid(vio, &msgbuf.tag);
+               if (err < 0)
+                       break;
+
+               if (likely(msgbuf.tag.type == VIO_TYPE_DATA)) {
+                       if (msgbuf.tag.stype == VIO_SUBTYPE_ACK)
+                               err = vdc_ack(port, &msgbuf);
+                       else if (msgbuf.tag.stype == VIO_SUBTYPE_NACK)
+                               err = vdc_nack(port, &msgbuf);
+                       else
+                               err = vdc_handle_unknown(port, &msgbuf);
+               } else if (msgbuf.tag.type == VIO_TYPE_CTRL) {
+                       err = vio_control_pkt_engine(vio, &msgbuf);
+               } else {
+                       err = vdc_handle_unknown(port, &msgbuf);
+               }
+               if (err < 0)
+                       break;
+       }
+       if (err < 0)
+               vdc_finish(&port->vio, err, WAITING_FOR_ANY);
+       spin_unlock_irqrestore(&vio->lock, flags);
+}
+
+static int __vdc_tx_trigger(struct vdc_port *port)
+{
+       struct vio_dring_state *dr = &port->vio.drings[VIO_DRIVER_TX_RING];
+       struct vio_dring_data hdr = {
+               .tag = {
+                       .type           = VIO_TYPE_DATA,
+                       .stype          = VIO_SUBTYPE_INFO,
+                       .stype_env      = VIO_DRING_DATA,
+                       .sid            = vio_send_sid(&port->vio),
+               },
+               .dring_ident            = dr->ident,
+               .start_idx              = dr->prod,
+               .end_idx                = dr->prod,
+       };
+       int err, delay;
+
+       hdr.seq = dr->snd_nxt;
+       delay = 1;
+       do {
+               err = vio_ldc_send(&port->vio, &hdr, sizeof(hdr));
+               if (err > 0) {
+                       dr->snd_nxt++;
+                       break;
+               }
+               udelay(delay);
+               if ((delay <<= 1) > 128)
+                       delay = 128;
+       } while (err == -EAGAIN);
+
+       return err;
+}
+
+static int __send_request(struct request *req)
+{
+       struct vdc_port *port = req->rq_disk->private_data;
+       struct vio_dring_state *dr = &port->vio.drings[VIO_DRIVER_TX_RING];
+       struct scatterlist sg[port->ring_cookies];
+       struct vdc_req_entry *rqe;
+       struct vio_disk_desc *desc;
+       unsigned int map_perm;
+       int nsg, err, i;
+       u64 len;
+       u8 op;
+
+       map_perm = LDC_MAP_SHADOW | LDC_MAP_DIRECT | LDC_MAP_IO;
+
+       if (rq_data_dir(req) == READ) {
+               map_perm |= LDC_MAP_W;
+               op = VD_OP_BREAD;
+       } else {
+               map_perm |= LDC_MAP_R;
+               op = VD_OP_BWRITE;
+       }
+
+       nsg = blk_rq_map_sg(req->q, req, sg);
+
+       len = 0;
+       for (i = 0; i < nsg; i++)
+               len += sg[i].length;
+
+       if (unlikely(vdc_tx_dring_avail(dr) < 1)) {
+               blk_stop_queue(port->disk->queue);
+               err = -ENOMEM;
+               goto out;
+       }
+
+       desc = vio_dring_cur(dr);
+
+       err = ldc_map_sg(port->vio.lp, sg, nsg,
+                        desc->cookies, port->ring_cookies,
+                        map_perm);
+       if (err < 0) {
+               printk(KERN_ERR PFX "ldc_map_sg() failure, err=%d.\n", err);
+               return err;
+       }
+
+       rqe = &port->rq_arr[dr->prod];
+       rqe->req = req;
+
+       desc->hdr.ack = VIO_ACK_ENABLE;
+       desc->req_id = port->req_id;
+       desc->operation = op;
+       if (port->vdisk_type == VD_DISK_TYPE_DISK) {
+               desc->slice = 2;
+       } else {
+               desc->slice = 0;
+       }
+       desc->status = ~0;
+       desc->offset = (req->sector << 9) / port->vdisk_block_size;
+       desc->size = len;
+       desc->ncookies = err;
+
+       /* This has to be a non-SMP write barrier because we are writing
+        * to memory which is shared with the peer LDOM.
+        */
+       wmb();
+       desc->hdr.state = VIO_DESC_READY;
+
+       err = __vdc_tx_trigger(port);
+       if (err < 0) {
+               printk(KERN_ERR PFX "vdc_tx_trigger() failure, err=%d\n", err);
+       } else {
+               port->req_id++;
+               dr->prod = (dr->prod + 1) & (VDC_TX_RING_SIZE - 1);
+       }
+out:
+
+       return err;
+}
+
+static void do_vdc_request(request_queue_t *q)
+{
+       while (1) {
+               struct request *req = elv_next_request(q);
+
+               if (!req)
+                       break;
+
+               blkdev_dequeue_request(req);
+               if (__send_request(req) < 0)
+                       vdc_end_request(req, 0, req->hard_nr_sectors);
+       }
+}
+
+static int generic_request(struct vdc_port *port, u8 op, void *buf, int len)
+{
+       struct vio_dring_state *dr;
+       struct vio_completion comp;
+       struct vio_disk_desc *desc;
+       unsigned int map_perm;
+       unsigned long flags;
+       int op_len, err;
+       void *req_buf;
+
+       if (!(((u64)1 << ((u64)op - 1)) & port->operations))
+               return -EOPNOTSUPP;
+
+       switch (op) {
+       case VD_OP_BREAD:
+       case VD_OP_BWRITE:
+       default:
+               return -EINVAL;
+
+       case VD_OP_FLUSH:
+               op_len = 0;
+               map_perm = 0;
+               break;
+
+       case VD_OP_GET_WCE:
+               op_len = sizeof(u32);
+               map_perm = LDC_MAP_W;
+               break;
+
+       case VD_OP_SET_WCE:
+               op_len = sizeof(u32);
+               map_perm = LDC_MAP_R;
+               break;
+
+       case VD_OP_GET_VTOC:
+               op_len = sizeof(struct vio_disk_vtoc);
+               map_perm = LDC_MAP_W;
+               break;
+
+       case VD_OP_SET_VTOC:
+               op_len = sizeof(struct vio_disk_vtoc);
+               map_perm = LDC_MAP_R;
+               break;
+
+       case VD_OP_GET_DISKGEOM:
+               op_len = sizeof(struct vio_disk_geom);
+               map_perm = LDC_MAP_W;
+               break;
+
+       case VD_OP_SET_DISKGEOM:
+               op_len = sizeof(struct vio_disk_geom);
+               map_perm = LDC_MAP_R;
+               break;
+
+       case VD_OP_SCSICMD:
+               op_len = 16;
+               map_perm = LDC_MAP_RW;
+               break;
+
+       case VD_OP_GET_DEVID:
+               op_len = sizeof(struct vio_disk_devid);
+               map_perm = LDC_MAP_W;
+               break;
+
+       case VD_OP_GET_EFI:
+       case VD_OP_SET_EFI:
+               return -EOPNOTSUPP;
+               break;
+       };
+
+       map_perm |= LDC_MAP_SHADOW | LDC_MAP_DIRECT | LDC_MAP_IO;
+
+       op_len = (op_len + 7) & ~7;
+       req_buf = kzalloc(op_len, GFP_KERNEL);
+       if (!req_buf)
+               return -ENOMEM;
+
+       if (len > op_len)
+               len = op_len;
+
+       if (map_perm & LDC_MAP_R)
+               memcpy(req_buf, buf, len);
+
+       spin_lock_irqsave(&port->vio.lock, flags);
+
+       dr = &port->vio.drings[VIO_DRIVER_TX_RING];
+
+       /* XXX If we want to use this code generically we have to
+        * XXX handle TX ring exhaustion etc.
+        */
+       desc = vio_dring_cur(dr);
+
+       err = ldc_map_single(port->vio.lp, req_buf, op_len,
+                            desc->cookies, port->ring_cookies,
+                            map_perm);
+       if (err < 0) {
+               spin_unlock_irqrestore(&port->vio.lock, flags);
+               kfree(req_buf);
+               return err;
+       }
+
+       init_completion(&comp.com);
+       comp.waiting_for = WAITING_FOR_GEN_CMD;
+       port->vio.cmp = &comp;
+
+       desc->hdr.ack = VIO_ACK_ENABLE;
+       desc->req_id = port->req_id;
+       desc->operation = op;
+       desc->slice = 0;
+       desc->status = ~0;
+       desc->offset = 0;
+       desc->size = op_len;
+       desc->ncookies = err;
+
+       /* This has to be a non-SMP write barrier because we are writing
+        * to memory which is shared with the peer LDOM.
+        */
+       wmb();
+       desc->hdr.state = VIO_DESC_READY;
+
+       err = __vdc_tx_trigger(port);
+       if (err >= 0) {
+               port->req_id++;
+               dr->prod = (dr->prod + 1) & (VDC_TX_RING_SIZE - 1);
+               spin_unlock_irqrestore(&port->vio.lock, flags);
+
+               wait_for_completion(&comp.com);
+               err = comp.err;
+       } else {
+               port->vio.cmp = NULL;
+               spin_unlock_irqrestore(&port->vio.lock, flags);
+       }
+
+       if (map_perm & LDC_MAP_W)
+               memcpy(buf, req_buf, len);
+
+       kfree(req_buf);
+
+       return err;
+}
+
+static int __devinit vdc_alloc_tx_ring(struct vdc_port *port)
+{
+       struct vio_dring_state *dr = &port->vio.drings[VIO_DRIVER_TX_RING];
+       unsigned long len, entry_size;
+       int ncookies;
+       void *dring;
+
+       entry_size = sizeof(struct vio_disk_desc) +
+               (sizeof(struct ldc_trans_cookie) * port->ring_cookies);
+       len = (VDC_TX_RING_SIZE * entry_size);
+
+       ncookies = VIO_MAX_RING_COOKIES;
+       dring = ldc_alloc_exp_dring(port->vio.lp, len,
+                                   dr->cookies, &ncookies,
+                                   (LDC_MAP_SHADOW |
+                                    LDC_MAP_DIRECT |
+                                    LDC_MAP_RW));
+       if (IS_ERR(dring))
+               return PTR_ERR(dring);
+
+       dr->base = dring;
+       dr->entry_size = entry_size;
+       dr->num_entries = VDC_TX_RING_SIZE;
+       dr->prod = dr->cons = 0;
+       dr->pending = VDC_TX_RING_SIZE;
+       dr->ncookies = ncookies;
+
+       return 0;
+}
+
+static void vdc_free_tx_ring(struct vdc_port *port)
+{
+       struct vio_dring_state *dr = &port->vio.drings[VIO_DRIVER_TX_RING];
+
+       if (dr->base) {
+               ldc_free_exp_dring(port->vio.lp, dr->base,
+                                  (dr->entry_size * dr->num_entries),
+                                  dr->cookies, dr->ncookies);
+               dr->base = NULL;
+               dr->entry_size = 0;
+               dr->num_entries = 0;
+               dr->pending = 0;
+               dr->ncookies = 0;
+       }
+}
+
+static int probe_disk(struct vdc_port *port)
+{
+       struct vio_completion comp;
+       struct request_queue *q;
+       struct gendisk *g;
+       int err;
+
+       init_completion(&comp.com);
+       comp.err = 0;
+       comp.waiting_for = WAITING_FOR_LINK_UP;
+       port->vio.cmp = &comp;
+
+       vio_port_up(&port->vio);
+
+       wait_for_completion(&comp.com);
+       if (comp.err)
+               return comp.err;
+
+       err = generic_request(port, VD_OP_GET_VTOC,
+                             &port->label, sizeof(port->label));
+       if (err < 0) {
+               printk(KERN_ERR PFX "VD_OP_GET_VTOC returns error %d\n", err);
+               return err;
+       }
+
+       err = generic_request(port, VD_OP_GET_DISKGEOM,
+                             &port->geom, sizeof(port->geom));
+       if (err < 0) {
+               printk(KERN_ERR PFX "VD_OP_GET_DISKGEOM returns "
+                      "error %d\n", err);
+               return err;
+       }
+
+       port->vdisk_size = ((u64)port->geom.num_cyl *
+                           (u64)port->geom.num_hd *
+                           (u64)port->geom.num_sec);
+
+       q = blk_init_queue(do_vdc_request, &port->vio.lock);
+       if (!q) {
+               printk(KERN_ERR PFX "%s: Could not allocate queue.\n",
+                      port->vio.name);
+               return -ENOMEM;
+       }
+       g = alloc_disk(1 << PARTITION_SHIFT);
+       if (!g) {
+               printk(KERN_ERR PFX "%s: Could not allocate gendisk.\n",
+                      port->vio.name);
+               blk_cleanup_queue(q);
+               return -ENOMEM;
+       }
+
+       port->disk = g;
+
+       blk_queue_max_hw_segments(q, port->ring_cookies);
+       blk_queue_max_phys_segments(q, port->ring_cookies);
+       blk_queue_max_sectors(q, port->max_xfer_size);
+       g->major = vdc_major;
+       g->first_minor = port->dev_no << PARTITION_SHIFT;
+       strcpy(g->disk_name, port->disk_name);
+
+       g->fops = &vdc_fops;
+       g->queue = q;
+       g->private_data = port;
+       g->driverfs_dev = &port->vio.vdev->dev;
+
+       set_capacity(g, port->vdisk_size);
+
+       printk(KERN_INFO PFX "%s: %u sectors (%u MB)\n",
+              g->disk_name,
+              port->vdisk_size, (port->vdisk_size >> (20 - 9)));
+
+       add_disk(g);
+
+       return 0;
+}
+
+static struct ldc_channel_config vdc_ldc_cfg = {
+       .event          = vdc_event,
+       .mtu            = 64,
+       .mode           = LDC_MODE_UNRELIABLE,
+};
+
+static struct vio_driver_ops vdc_vio_ops = {
+       .send_attr              = vdc_send_attr,
+       .handle_attr            = vdc_handle_attr,
+       .handshake_complete     = vdc_handshake_complete,
+};
+
+static int __devinit vdc_port_probe(struct vio_dev *vdev,
+                                   const struct vio_device_id *id)
+{
+       struct mdesc_node *endp;
+       struct vdc_port *port;
+       unsigned long flags;
+       struct vdc *vp;
+       const u64 *port_id;
+       int err;
+
+       vp = dev_get_drvdata(vdev->dev.parent);
+       if (!vp) {
+               printk(KERN_ERR PFX "Cannot find port parent vdc.\n");
+               return -ENODEV;
+       }
+
+       endp = vio_find_endpoint(vdev);
+       if (!endp) {
+               printk(KERN_ERR PFX "Port lacks channel-endpoint.\n");
+               return -ENODEV;
+       }
+
+       port_id = md_get_property(vdev->mp, "id", NULL);
+       if (!port_id) {
+               printk(KERN_ERR PFX "Port lacks id property.\n");
+               return -ENODEV;
+       }
+       if ((*port_id << PARTITION_SHIFT) & ~(u64)MINORMASK) {
+               printk(KERN_ERR PFX "Port id [%lu] too large.\n", *port_id);
+               return -ENODEV;
+       }
+
+       port = kzalloc(sizeof(*port), GFP_KERNEL);
+       if (!port) {
+               printk(KERN_ERR PFX "Cannot allocate vdc_port.\n");
+               return -ENOMEM;
+       }
+
+       port->vp = vp;
+       port->dev_no = *port_id;
+
+       if (port->dev_no >= 26)
+               snprintf(port->disk_name, sizeof(port->disk_name),
+                        VDCBLK_NAME "%c%c",
+                        'a' + (port->dev_no / 26) - 1,
+                        'a' + (port->dev_no % 26));
+       else
+               snprintf(port->disk_name, sizeof(port->disk_name),
+                        VDCBLK_NAME "%c", 'a' + (port->dev_no % 26));
+
+       err = vio_driver_init(&port->vio, vdev, VDEV_DISK, endp,
+                             vdc_versions, ARRAY_SIZE(vdc_versions),
+                             &vdc_vio_ops, port->disk_name);
+       if (err)
+               goto err_out_free_port;
+
+       port->vdisk_block_size = 512;
+       port->max_xfer_size = ((128 * 1024) / port->vdisk_block_size);
+       port->ring_cookies = ((port->max_xfer_size *
+                              port->vdisk_block_size) / PAGE_SIZE) + 2;
+
+       err = vio_ldc_alloc(&port->vio, &vdc_ldc_cfg, port);
+       if (err)
+               goto err_out_free_port;
+
+       err = vdc_alloc_tx_ring(port);
+       if (err)
+               goto err_out_free_ldc;
+
+       err = probe_disk(port);
+       if (err)
+               goto err_out_free_tx_ring;
+
+       INIT_LIST_HEAD(&port->list);
+
+       spin_lock_irqsave(&vp->lock, flags);
+       list_add(&port->list, &vp->port_list);
+       spin_unlock_irqrestore(&vp->lock, flags);
+
+       dev_set_drvdata(&vdev->dev, port);
+
+       return 0;
+
+err_out_free_tx_ring:
+       vdc_free_tx_ring(port);
+
+err_out_free_ldc:
+       vio_ldc_free(&port->vio);
+
+err_out_free_port:
+       kfree(port);
+
+       return err;
+}
+
+static int vdc_port_remove(struct vio_dev *vdev)
+{
+       struct vdc_port *port = dev_get_drvdata(&vdev->dev);
+
+       if (port) {
+               del_timer_sync(&port->vio.timer);
+
+               vdc_free_tx_ring(port);
+               vio_ldc_free(&port->vio);
+
+               dev_set_drvdata(&vdev->dev, NULL);
+
+               kfree(port);
+       }
+       return 0;
+}
+
+static struct vio_device_id vdc_port_match[] = {
+       {
+               .type = "vdc-port",
+       },
+       {},
+};
+MODULE_DEVICE_TABLE(vio, vdc_match);
+
+static struct vio_driver vdc_port_driver = {
+       .id_table       = vdc_port_match,
+       .probe          = vdc_port_probe,
+       .remove         = vdc_port_remove,
+       .driver         = {
+               .name   = "vdc_port",
+               .owner  = THIS_MODULE,
+       }
+};
+
+static int __devinit vdc_probe(struct vio_dev *vdev,
+                              const struct vio_device_id *id)
+{
+       static int vdc_version_printed;
+       struct vdc *vp;
+
+       if (vdc_version_printed++ == 0)
+               printk(KERN_INFO "%s", version);
+
+       vp = kzalloc(sizeof(struct vdc), GFP_KERNEL);
+       if (!vp)
+               return -ENOMEM;
+
+       spin_lock_init(&vp->lock);
+       vp->dev = vdev;
+       INIT_LIST_HEAD(&vp->port_list);
+
+       dev_set_drvdata(&vdev->dev, vp);
+
+       return 0;
+}
+
+static int vdc_remove(struct vio_dev *vdev)
+{
+
+       struct vdc *vp = dev_get_drvdata(&vdev->dev);
+
+       if (vp) {
+               kfree(vp);
+               dev_set_drvdata(&vdev->dev, NULL);
+       }
+       return 0;
+}
+
+static struct vio_device_id vdc_match[] = {
+       {
+               .type = "block",
+       },
+       {},
+};
+MODULE_DEVICE_TABLE(vio, vdc_match);
+
+static struct vio_driver vdc_driver = {
+       .id_table       = vdc_match,
+       .probe          = vdc_probe,
+       .remove         = vdc_remove,
+       .driver         = {
+               .name   = "vdc",
+               .owner  = THIS_MODULE,
+       }
+};
+
+static int __init vdc_init(void)
+{
+       int err;
+
+       err = register_blkdev(0, VDCBLK_NAME);
+       if (err < 0)
+               goto out_err;
+
+       vdc_major = err;
+       err = vio_register_driver(&vdc_driver);
+       if (err)
+               goto out_unregister_blkdev;
+
+       err = vio_register_driver(&vdc_port_driver);
+       if (err)
+               goto out_unregister_vdc;
+
+       return 0;
+
+out_unregister_vdc:
+       vio_unregister_driver(&vdc_driver);
+
+out_unregister_blkdev:
+       unregister_blkdev(vdc_major, VDCBLK_NAME);
+       vdc_major = 0;
+
+out_err:
+       return err;
+}
+
+static void __exit vdc_exit(void)
+{
+       vio_unregister_driver(&vdc_port_driver);
+       vio_unregister_driver(&vdc_driver);
+       unregister_blkdev(vdc_major, VDCBLK_NAME);
+}
+
+module_init(vdc_init);
+module_exit(vdc_exit);