pktcdvd.c

来自「linux 内核源代码」· C语言 代码 · 共 2,596 行 · 第 1/5 页

 * a) The number of required segments for the write bio is minimized, which
 *    is necessary for some scsi controllers.
 * b) The data can be used as cache to avoid read requests if we receive a
 *    new write request for the same zone.
 */
static void pkt_make_local_copy(struct packet_data *pkt, struct bio_vec *bvec)
{
	int f, p, offs;

	/* Copy all data to pkt->pages[] */
	p = 0;
	offs = 0;
	for (f = 0; f < pkt->frames; f++) {
		if (bvec[f].bv_page != pkt->pages[p]) {
			void *vfrom = kmap_atomic(bvec[f].bv_page, KM_USER0) + bvec[f].bv_offset;
			void *vto = page_address(pkt->pages[p]) + offs;
			memcpy(vto, vfrom, CD_FRAMESIZE);
			kunmap_atomic(vfrom, KM_USER0);
			bvec[f].bv_page = pkt->pages[p];
			bvec[f].bv_offset = offs;
		} else {
			BUG_ON(bvec[f].bv_offset != offs);
		}
		offs += CD_FRAMESIZE;
		if (offs >= PAGE_SIZE) {
			offs = 0;
			p++;
		}
	}
}

static void pkt_end_io_read(struct bio *bio, int err)
{
	struct packet_data *pkt = bio->bi_private;
	struct pktcdvd_device *pd = pkt->pd;
	BUG_ON(!pd);

	VPRINTK("pkt_end_io_read: bio=%p sec0=%llx sec=%llx err=%d\n", bio,
		(unsigned long long)pkt->sector, (unsigned long long)bio->bi_sector, err);

	if (err)
		atomic_inc(&pkt->io_errors);
	if (atomic_dec_and_test(&pkt->io_wait)) {
		atomic_inc(&pkt->run_sm);
		wake_up(&pd->wqueue);
	}
	pkt_bio_finished(pd);
}

static void pkt_end_io_packet_write(struct bio *bio, int err)
{
	struct packet_data *pkt = bio->bi_private;
	struct pktcdvd_device *pd = pkt->pd;
	BUG_ON(!pd);

	VPRINTK("pkt_end_io_packet_write: id=%d, err=%d\n", pkt->id, err);

	pd->stats.pkt_ended++;

	pkt_bio_finished(pd);
	atomic_dec(&pkt->io_wait);
	atomic_inc(&pkt->run_sm);
	wake_up(&pd->wqueue);
}

/*
 * Schedule reads for the holes in a packet
 */
static void pkt_gather_data(struct pktcdvd_device *pd, struct packet_data *pkt)
{
	int frames_read = 0;
	struct bio *bio;
	int f;
	char written[PACKET_MAX_SIZE];

	BUG_ON(!pkt->orig_bios);

	atomic_set(&pkt->io_wait, 0);
	atomic_set(&pkt->io_errors, 0);

	/*
	 * Figure out which frames we need to read before we can write.
	 */
	memset(written, 0, sizeof(written));
	spin_lock(&pkt->lock);
	for (bio = pkt->orig_bios; bio; bio = bio->bi_next) {
		int first_frame = (bio->bi_sector - pkt->sector) / (CD_FRAMESIZE >> 9);
		int num_frames = bio->bi_size / CD_FRAMESIZE;
		pd->stats.secs_w += num_frames * (CD_FRAMESIZE >> 9);
		BUG_ON(first_frame < 0);
		BUG_ON(first_frame + num_frames > pkt->frames);
		for (f = first_frame; f < first_frame + num_frames; f++)
			written[f] = 1;
	}
	spin_unlock(&pkt->lock);

	if (pkt->cache_valid) {
		VPRINTK("pkt_gather_data: zone %llx cached\n",
			(unsigned long long)pkt->sector);
		goto out_account;
	}

	/*
	 * Schedule reads for missing parts of the packet.
	 */
	for (f = 0; f < pkt->frames; f++) {
		struct bio_vec *vec;

		int p, offset;
		if (written[f])
			continue;
		bio = pkt->r_bios[f];
		vec = bio->bi_io_vec;
		bio_init(bio);
		bio->bi_max_vecs = 1;
		bio->bi_sector = pkt->sector + f * (CD_FRAMESIZE >> 9);
		bio->bi_bdev = pd->bdev;
		bio->bi_end_io = pkt_end_io_read;
		bio->bi_private = pkt;
		bio->bi_io_vec = vec;
		bio->bi_destructor = pkt_bio_destructor;

		p = (f * CD_FRAMESIZE) / PAGE_SIZE;
		offset = (f * CD_FRAMESIZE) % PAGE_SIZE;
		VPRINTK("pkt_gather_data: Adding frame %d, page:%p offs:%d\n",
			f, pkt->pages[p], offset);
		if (!bio_add_page(bio, pkt->pages[p], CD_FRAMESIZE, offset))
			BUG();

		atomic_inc(&pkt->io_wait);
		bio->bi_rw = READ;
		pkt_queue_bio(pd, bio);
		frames_read++;
	}

out_account:
	VPRINTK("pkt_gather_data: need %d frames for zone %llx\n",
		frames_read, (unsigned long long)pkt->sector);
	pd->stats.pkt_started++;
	pd->stats.secs_rg += frames_read * (CD_FRAMESIZE >> 9);
}

/*
 * Find a packet matching zone, or the least recently used packet if
 * there is no match.
 */
static struct packet_data *pkt_get_packet_data(struct pktcdvd_device *pd, int zone)
{
	struct packet_data *pkt;

	list_for_each_entry(pkt, &pd->cdrw.pkt_free_list, list) {
		if (pkt->sector == zone || pkt->list.next == &pd->cdrw.pkt_free_list) {
			list_del_init(&pkt->list);
			if (pkt->sector != zone)
				pkt->cache_valid = 0;
			return pkt;
		}
	}
	BUG();
	return NULL;
}

static void pkt_put_packet_data(struct pktcdvd_device *pd, struct packet_data *pkt)
{
	if (pkt->cache_valid) {
		list_add(&pkt->list, &pd->cdrw.pkt_free_list);
	} else {
		list_add_tail(&pkt->list, &pd->cdrw.pkt_free_list);
	}
}

/*
 * recover a failed write, query for relocation if possible
 *
 * returns 1 if recovery is possible, or 0 if not
 *
 */
static int pkt_start_recovery(struct packet_data *pkt)
{
	/*
	 * FIXME. We need help from the file system to implement
	 * recovery handling.
	 */
	return 0;
#if 0
	struct request *rq = pkt->rq;
	struct pktcdvd_device *pd = rq->rq_disk->private_data;
	struct block_device *pkt_bdev;
	struct super_block *sb = NULL;
	unsigned long old_block, new_block;
	sector_t new_sector;

	pkt_bdev = bdget(kdev_t_to_nr(pd->pkt_dev));
	if (pkt_bdev) {
		sb = get_super(pkt_bdev);
		bdput(pkt_bdev);
	}

	if (!sb)
		return 0;

	if (!sb->s_op || !sb->s_op->relocate_blocks)
		goto out;

	old_block = pkt->sector / (CD_FRAMESIZE >> 9);
	if (sb->s_op->relocate_blocks(sb, old_block, &new_block))
		goto out;

	new_sector = new_block * (CD_FRAMESIZE >> 9);
	pkt->sector = new_sector;

	pkt->bio->bi_sector = new_sector;
	pkt->bio->bi_next = NULL;
	pkt->bio->bi_flags = 1 << BIO_UPTODATE;
	pkt->bio->bi_idx = 0;

	BUG_ON(pkt->bio->bi_rw != (1 << BIO_RW));
	BUG_ON(pkt->bio->bi_vcnt != pkt->frames);
	BUG_ON(pkt->bio->bi_size != pkt->frames * CD_FRAMESIZE);
	BUG_ON(pkt->bio->bi_end_io != pkt_end_io_packet_write);
	BUG_ON(pkt->bio->bi_private != pkt);

	drop_super(sb);
	return 1;

out:
	drop_super(sb);
	return 0;
#endif
}

static inline void pkt_set_state(struct packet_data *pkt, enum packet_data_state state)
{
#if PACKET_DEBUG > 1
	static const char *state_name[] = {
		"IDLE", "WAITING", "READ_WAIT", "WRITE_WAIT", "RECOVERY", "FINISHED"
	};
	enum packet_data_state old_state = pkt->state;
	VPRINTK("pkt %2d : s=%6llx %s -> %s\n", pkt->id, (unsigned long long)pkt->sector,
		state_name[old_state], state_name[state]);
#endif
	pkt->state = state;
}

/*
 * Scan the work queue to see if we can start a new packet.
 * returns non-zero if any work was done.
 */
static int pkt_handle_queue(struct pktcdvd_device *pd)
{
	struct packet_data *pkt, *p;
	struct bio *bio = NULL;
	sector_t zone = 0; /* Suppress gcc warning */
	struct pkt_rb_node *node, *first_node;
	struct rb_node *n;
	int wakeup;

	VPRINTK("handle_queue\n");

	atomic_set(&pd->scan_queue, 0);

	if (list_empty(&pd->cdrw.pkt_free_list)) {
		VPRINTK("handle_queue: no pkt\n");
		return 0;
	}

	/*
	 * Try to find a zone we are not already working on.
	 */
	spin_lock(&pd->lock);
	first_node = pkt_rbtree_find(pd, pd->current_sector);
	if (!first_node) {
		n = rb_first(&pd->bio_queue);
		if (n)
			first_node = rb_entry(n, struct pkt_rb_node, rb_node);
	}
	node = first_node;
	while (node) {
		bio = node->bio;
		zone = ZONE(bio->bi_sector, pd);
		list_for_each_entry(p, &pd->cdrw.pkt_active_list, list) {
			if (p->sector == zone) {
				bio = NULL;
				goto try_next_bio;
			}
		}
		break;
try_next_bio:
		node = pkt_rbtree_next(node);
		if (!node) {
			n = rb_first(&pd->bio_queue);
			if (n)
				node = rb_entry(n, struct pkt_rb_node, rb_node);
		}
		if (node == first_node)
			node = NULL;
	}
	spin_unlock(&pd->lock);
	if (!bio) {
		VPRINTK("handle_queue: no bio\n");
		return 0;
	}

	pkt = pkt_get_packet_data(pd, zone);

	pd->current_sector = zone + pd->settings.size;
	pkt->sector = zone;
	BUG_ON(pkt->frames != pd->settings.size >> 2);
	pkt->write_size = 0;

	/*
	 * Scan work queue for bios in the same zone and link them
	 * to this packet.
	 */
	spin_lock(&pd->lock);
	VPRINTK("pkt_handle_queue: looking for zone %llx\n", (unsigned long long)zone);
	while ((node = pkt_rbtree_find(pd, zone)) != NULL) {
		bio = node->bio;
		VPRINTK("pkt_handle_queue: found zone=%llx\n",
			(unsigned long long)ZONE(bio->bi_sector, pd));
		if (ZONE(bio->bi_sector, pd) != zone)
			break;
		pkt_rbtree_erase(pd, node);
		spin_lock(&pkt->lock);
		pkt_add_list_last(bio, &pkt->orig_bios, &pkt->orig_bios_tail);
		pkt->write_size += bio->bi_size / CD_FRAMESIZE;
		spin_unlock(&pkt->lock);
	}
	/* check write congestion marks, and if bio_queue_size is
	   below, wake up any waiters */
	wakeup = (pd->write_congestion_on > 0
	 		&& pd->bio_queue_size <= pd->write_congestion_off);
	spin_unlock(&pd->lock);
	if (wakeup)
		clear_bdi_congested(&pd->disk->queue->backing_dev_info, WRITE);

	pkt->sleep_time = max(PACKET_WAIT_TIME, 1);
	pkt_set_state(pkt, PACKET_WAITING_STATE);
	atomic_set(&pkt->run_sm, 1);

	spin_lock(&pd->cdrw.active_list_lock);
	list_add(&pkt->list, &pd->cdrw.pkt_active_list);
	spin_unlock(&pd->cdrw.active_list_lock);

	return 1;
}

/*
 * Assemble a bio to write one packet and queue the bio for processing
 * by the underlying block device.
 */
static void pkt_start_write(struct pktcdvd_device *pd, struct packet_data *pkt)
{
	struct bio *bio;
	int f;
	int frames_write;
	struct bio_vec *bvec = pkt->w_bio->bi_io_vec;

	for (f = 0; f < pkt->frames; f++) {
		bvec[f].bv_page = pkt->pages[(f * CD_FRAMESIZE) / PAGE_SIZE];
		bvec[f].bv_offset = (f * CD_FRAMESIZE) % PAGE_SIZE;
	}

	/*
	 * Fill-in bvec with data from orig_bios.
	 */
	frames_write = 0;
	spin_lock(&pkt->lock);
	for (bio = pkt->orig_bios; bio; bio = bio->bi_next) {
		int segment = bio->bi_idx;
		int src_offs = 0;
		int first_frame = (bio->bi_sector - pkt->sector) / (CD_FRAMESIZE >> 9);
		int num_frames = bio->bi_size / CD_FRAMESIZE;
		BUG_ON(first_frame < 0);
		BUG_ON(first_frame + num_frames > pkt->frames);
		for (f = first_frame; f < first_frame + num_frames; f++) {
			struct bio_vec *src_bvl = bio_iovec_idx(bio, segment);

			while (src_offs >= src_bvl->bv_len) {
				src_offs -= src_bvl->bv_len;
				segment++;
				BUG_ON(segment >= bio->bi_vcnt);
				src_bvl = bio_iovec_idx(bio, segment);
			}

			if (src_bvl->bv_len - src_offs >= CD_FRAMESIZE) {
				bvec[f].bv_page = src_bvl->bv_page;
				bvec[f].bv_offset = src_bvl->bv_offset + src_offs;
			} else {
				pkt_copy_bio_data(bio, segment, src_offs,
						  bvec[f].bv_page, bvec[f].bv_offset);
			}
			src_offs += CD_FRAMESIZE;
			frames_write++;
		}
	}
	pkt_set_state(pkt, PACKET_WRITE_WAIT_STATE);
	spin_unlock(&pkt->lock);

	VPRINTK("pkt_start_write: Writing %d frames for zone %llx\n",
		frames_write, (unsigned long long)pkt->sector);
	BUG_ON(frames_write != pkt->write_size);

	if (test_bit(PACKET_MERGE_SEGS, &pd->flags) || (pkt->write_size < pkt->frames)) {
		pkt_make_local_copy(pkt, bvec);
		pkt->cache_valid = 1;
	} else {
		pkt->cache_valid = 0;
	}

	/* Start the write request */
	bio_init(pkt->w_bio);
	pkt->w_bio->bi_max_vecs = PACKET_MAX_SIZE;
	pkt->w_bio->bi_sector = pkt->sector;
	pkt->w_bio->bi_bdev = pd->bdev;
	pkt->w_bio->bi_end_io = pkt_end_io_packet_write;
	pkt->w_bio->bi_private = pkt;
	pkt->w_bio->bi_io_vec = bvec;
	pkt->w_bio->bi_destructor = pkt_bio_destructor;
	for (f = 0; f < pkt->frames; f++)
		if (!bio_add_page(pkt->w_bio, bvec[f].bv_page, CD_FRAMESIZE, bvec[f].bv_offset))
			BUG();
	VPRINTK(DRIVER_NAME": vcnt=%d\n", pkt->w_bio->bi_vcnt);

	atomic_set(&pkt->io_wait, 1);
	pkt->w_bio->bi_rw = WRITE;
	pkt_queue_bio(pd, pkt->w_bio);
}

static void pkt_finish_packet(struct packet_data *pkt, int uptodate)
{
	struct bio *bio, *next;

	if (!uptodate)
		pkt->cache_valid = 0;

	/* Finish all bios corresponding to this packet */
	bio = pkt->orig_bios;
	while (bio) {
		next = bio->bi_next;
		bio->bi_next = NULL;
		bio_endio(bio, uptodate ? 0 : -EIO);
		bio = next;
	}
	pkt->orig_bios = pkt->orig_bios_tail = NULL;
}

static void pkt_run_state_machine(struct pktcdvd_device *pd, struct packet_data *pkt)
{
	int uptodate;

	VPRINTK("run_state_machine: pkt %d\n", pkt->id);

	for (;;) {
		switch (pkt->state) {
		case PACKET_WAITING_STATE:
			if ((pkt->write_size < pkt->frames) && (pkt->sleep_time > 0))
				return;

			pkt->sleep_time = 0;
			pkt_gather_data(pd, pkt);
			pkt_set_state(pkt, PACKET_READ_WAIT_STATE);
			break;

		case PACKET_READ_WAIT_STATE:
			if (atomic_read(&pkt->io_wait) > 0)
				return;

			if (atomic_read(&pkt->io_errors) > 0) {
				pkt_set_state(pkt, PACKET_RECOVERY_STATE);
			} else {
				pkt_start_write(pd, pkt);
			}
			break;

		case PACKET_WRITE_WAIT_STATE:
			if (atomic_read(&pkt->io_wait) > 0)
				return;

			if (test_bit(BIO_UPTODATE, &pkt->w_bio->bi_flags)) {
				pkt_set_state(pkt, PACKET_FINISHED_STATE);
			} else {
				pkt_set_state(pkt, PACKET_RECOVERY_STATE);
			}
			break;

		case PACKET_RECOVERY_STATE:
			if (pkt_start_recovery(pkt)) {
				pkt_start_write(pd, pkt);
			} else {
				VPRINTK("No recovery possible\n");
				pkt_set_state(pkt, PACKET_FINISHED_STATE);
			}
			break;

		case PACKET_FINISHED_STATE:
			uptodate = test_bit(BIO_UPTODATE, &pkt->w_bio->bi_flags);
			pkt_finish_packet(pkt, uptodate);
			return;

		default:
			BUG();
			break;
		}
	}
}

static void pkt_handle_packets(struct pktcdvd_device *pd)
{
	struct packet_data *pkt, *next;

	VPRINTK("pkt_handle_packets\n");

	/*
	 * Run state machine for active packets
	 */
	list_for_each_entry(pkt, &pd->cdrw.pkt_active_list, list) {
		if (atomic_read(&pkt->run_sm) > 0) {
			atomic_set(&pkt->run_sm, 0);
			pkt_run_state_machine(pd, pkt);