sunvdc.c

来自「linux 内核源代码」· C语言 代码 · 共 890 行 · 第 1/2 页

/* sunvdc.c: Sun LDOM Virtual Disk Client.
 *
 * Copyright (C) 2007 David S. Miller <[e-mail]>
 */

#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 <linux/scatterlist.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 ([e-mail])");
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 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;

	char			disk_name[32];

	struct vio_disk_geom	geom;
	struct vio_disk_vtoc	label;
};

static inline struct vdc_port *to_vdc_port(struct vio_driver_state *vio)
{
	return container_of(vio, struct vdc_port, vio);
}

/* 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;
	}

	sg_init_table(sg, port->ring_cookies);
	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 = 0xff;
	} 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: