ubd_kern.c

来自「linux 内核源代码」· C语言 代码 · 共 1,485 行 · 第 1/3 页

		ubd_close_dev(ubd_dev);
	return 0;
}

static void cowify_bitmap(__u64 io_offset, int length, unsigned long *cow_mask,
			  __u64 *cow_offset, unsigned long *bitmap,
			  __u64 bitmap_offset, unsigned long *bitmap_words,
			  __u64 bitmap_len)
{
	__u64 sector = io_offset >> 9;
	int i, update_bitmap = 0;

	for(i = 0; i < length >> 9; i++){
		if(cow_mask != NULL)
			ubd_set_bit(i, (unsigned char *) cow_mask);
		if(ubd_test_bit(sector + i, (unsigned char *) bitmap))
			continue;

		update_bitmap = 1;
		ubd_set_bit(sector + i, (unsigned char *) bitmap);
	}

	if(!update_bitmap)
		return;

	*cow_offset = sector / (sizeof(unsigned long) * 8);

	/* This takes care of the case where we're exactly at the end of the
	 * device, and *cow_offset + 1 is off the end.  So, just back it up
	 * by one word.  Thanks to Lynn Kerby for the fix and James McMechan
	 * for the original diagnosis.
	 */
	if(*cow_offset == ((bitmap_len + sizeof(unsigned long) - 1) /
			   sizeof(unsigned long) - 1))
		(*cow_offset)--;

	bitmap_words[0] = bitmap[*cow_offset];
	bitmap_words[1] = bitmap[*cow_offset + 1];

	*cow_offset *= sizeof(unsigned long);
	*cow_offset += bitmap_offset;
}

static void cowify_req(struct io_thread_req *req, unsigned long *bitmap,
		       __u64 bitmap_offset, __u64 bitmap_len)
{
	__u64 sector = req->offset >> 9;
	int i;

	if(req->length > (sizeof(req->sector_mask) * 8) << 9)
		panic("Operation too long");

	if(req->op == UBD_READ) {
		for(i = 0; i < req->length >> 9; i++){
			if(ubd_test_bit(sector + i, (unsigned char *) bitmap))
				ubd_set_bit(i, (unsigned char *)
					    &req->sector_mask);
		}
	}
	else cowify_bitmap(req->offset, req->length, &req->sector_mask,
			   &req->cow_offset, bitmap, bitmap_offset,
			   req->bitmap_words, bitmap_len);
}

/* Called with dev->lock held */
static void prepare_request(struct request *req, struct io_thread_req *io_req,
			    unsigned long long offset, int page_offset,
			    int len, struct page *page)
{
	struct gendisk *disk = req->rq_disk;
	struct ubd *ubd_dev = disk->private_data;

	io_req->req = req;
	io_req->fds[0] = (ubd_dev->cow.file != NULL) ? ubd_dev->cow.fd :
		ubd_dev->fd;
	io_req->fds[1] = ubd_dev->fd;
	io_req->cow_offset = -1;
	io_req->offset = offset;
	io_req->length = len;
	io_req->error = 0;
	io_req->sector_mask = 0;

	io_req->op = (rq_data_dir(req) == READ) ? UBD_READ : UBD_WRITE;
	io_req->offsets[0] = 0;
	io_req->offsets[1] = ubd_dev->cow.data_offset;
	io_req->buffer = page_address(page) + page_offset;
	io_req->sectorsize = 1 << 9;

	if(ubd_dev->cow.file != NULL)
		cowify_req(io_req, ubd_dev->cow.bitmap,
			   ubd_dev->cow.bitmap_offset, ubd_dev->cow.bitmap_len);

}

/* Called with dev->lock held */
static void do_ubd_request(struct request_queue *q)
{
	struct io_thread_req *io_req;
	struct request *req;
	int n, last_sectors;

	while(1){
		struct ubd *dev = q->queuedata;
		if(dev->end_sg == 0){
			struct request *req = elv_next_request(q);
			if(req == NULL)
				return;

			dev->request = req;
			blkdev_dequeue_request(req);
			dev->start_sg = 0;
			dev->end_sg = blk_rq_map_sg(q, req, dev->sg);
		}

		req = dev->request;
		last_sectors = 0;
		while(dev->start_sg < dev->end_sg){
			struct scatterlist *sg = &dev->sg[dev->start_sg];

			req->sector += last_sectors;
			io_req = kmalloc(sizeof(struct io_thread_req),
					 GFP_ATOMIC);
			if(io_req == NULL){
				if(list_empty(&dev->restart))
					list_add(&dev->restart, &restart);
				return;
			}
			prepare_request(req, io_req,
					(unsigned long long) req->sector << 9,
					sg->offset, sg->length, sg_page(sg));

			last_sectors = sg->length >> 9;
			n = os_write_file(thread_fd, &io_req,
					  sizeof(struct io_thread_req *));
			if(n != sizeof(struct io_thread_req *)){
				if(n != -EAGAIN)
					printk("write to io thread failed, "
					       "errno = %d\n", -n);
				else if(list_empty(&dev->restart))
					list_add(&dev->restart, &restart);
				kfree(io_req);
				return;
			}

			dev->start_sg++;
		}
		dev->end_sg = 0;
		dev->request = NULL;
	}
}

static int ubd_getgeo(struct block_device *bdev, struct hd_geometry *geo)
{
	struct ubd *ubd_dev = bdev->bd_disk->private_data;

	geo->heads = 128;
	geo->sectors = 32;
	geo->cylinders = ubd_dev->size / (128 * 32 * 512);
	return 0;
}

static int ubd_ioctl(struct inode * inode, struct file * file,
		     unsigned int cmd, unsigned long arg)
{
	struct ubd *ubd_dev = inode->i_bdev->bd_disk->private_data;
	struct hd_driveid ubd_id = {
		.cyls		= 0,
		.heads		= 128,
		.sectors	= 32,
	};

	switch (cmd) {
		struct cdrom_volctrl volume;
	case HDIO_GET_IDENTITY:
		ubd_id.cyls = ubd_dev->size / (128 * 32 * 512);
		if(copy_to_user((char __user *) arg, (char *) &ubd_id,
				 sizeof(ubd_id)))
			return -EFAULT;
		return 0;

	case CDROMVOLREAD:
		if(copy_from_user(&volume, (char __user *) arg, sizeof(volume)))
			return -EFAULT;
		volume.channel0 = 255;
		volume.channel1 = 255;
		volume.channel2 = 255;
		volume.channel3 = 255;
		if(copy_to_user((char __user *) arg, &volume, sizeof(volume)))
			return -EFAULT;
		return 0;
	}
	return -EINVAL;
}

static int path_requires_switch(char *from_cmdline, char *from_cow, char *cow)
{
	struct uml_stat buf1, buf2;
	int err;

	if(from_cmdline == NULL)
		return 0;
	if(!strcmp(from_cmdline, from_cow))
		return 0;

	err = os_stat_file(from_cmdline, &buf1);
	if(err < 0){
		printk("Couldn't stat '%s', err = %d\n", from_cmdline, -err);
		return 0;
	}
	err = os_stat_file(from_cow, &buf2);
	if(err < 0){
		printk("Couldn't stat '%s', err = %d\n", from_cow, -err);
		return 1;
	}
	if((buf1.ust_dev == buf2.ust_dev) && (buf1.ust_ino == buf2.ust_ino))
		return 0;

	printk("Backing file mismatch - \"%s\" requested,\n"
	       "\"%s\" specified in COW header of \"%s\"\n",
	       from_cmdline, from_cow, cow);
	return 1;
}

static int backing_file_mismatch(char *file, __u64 size, time_t mtime)
{
	unsigned long modtime;
	unsigned long long actual;
	int err;

	err = os_file_modtime(file, &modtime);
	if(err < 0){
		printk("Failed to get modification time of backing file "
		       "\"%s\", err = %d\n", file, -err);
		return err;
	}

	err = os_file_size(file, &actual);
	if(err < 0){
		printk("Failed to get size of backing file \"%s\", "
		       "err = %d\n", file, -err);
		return err;
	}

	if(actual != size){
		/*__u64 can be a long on AMD64 and with %lu GCC complains; so
		 * the typecast.*/
		printk("Size mismatch (%llu vs %llu) of COW header vs backing "
		       "file\n", (unsigned long long) size, actual);
		return -EINVAL;
	}
	if(modtime != mtime){
		printk("mtime mismatch (%ld vs %ld) of COW header vs backing "
		       "file\n", mtime, modtime);
		return -EINVAL;
	}
	return 0;
}

int read_cow_bitmap(int fd, void *buf, int offset, int len)
{
	int err;

	err = os_seek_file(fd, offset);
	if(err < 0)
		return err;

	err = os_read_file(fd, buf, len);
	if(err < 0)
		return err;

	return 0;
}

int open_ubd_file(char *file, struct openflags *openflags, int shared,
		  char **backing_file_out, int *bitmap_offset_out,
		  unsigned long *bitmap_len_out, int *data_offset_out,
		  int *create_cow_out)
{
	time_t mtime;
	unsigned long long size;
	__u32 version, align;
	char *backing_file;
	int fd, err, sectorsize, asked_switch, mode = 0644;

	fd = os_open_file(file, *openflags, mode);
	if (fd < 0) {
		if ((fd == -ENOENT) && (create_cow_out != NULL))
			*create_cow_out = 1;
		if (!openflags->w ||
		    ((fd != -EROFS) && (fd != -EACCES)))
			return fd;
		openflags->w = 0;
		fd = os_open_file(file, *openflags, mode);
		if (fd < 0)
			return fd;
	}

	if(shared)
		printk("Not locking \"%s\" on the host\n", file);
	else {
		err = os_lock_file(fd, openflags->w);
		if(err < 0){
			printk("Failed to lock '%s', err = %d\n", file, -err);
			goto out_close;
		}
	}

	/* Successful return case! */
	if(backing_file_out == NULL)
		return fd;

	err = read_cow_header(file_reader, &fd, &version, &backing_file, &mtime,
			      &size, &sectorsize, &align, bitmap_offset_out);
	if(err && (*backing_file_out != NULL)){
		printk("Failed to read COW header from COW file \"%s\", "
		       "errno = %d\n", file, -err);
		goto out_close;
	}
	if(err)
		return fd;

	asked_switch = path_requires_switch(*backing_file_out, backing_file, file);

	/* Allow switching only if no mismatch. */
	if (asked_switch && !backing_file_mismatch(*backing_file_out, size, mtime)) {
		printk("Switching backing file to '%s'\n", *backing_file_out);
		err = write_cow_header(file, fd, *backing_file_out,
				       sectorsize, align, &size);
		if (err) {
			printk("Switch failed, errno = %d\n", -err);
			goto out_close;
		}
	} else {
		*backing_file_out = backing_file;
		err = backing_file_mismatch(*backing_file_out, size, mtime);
		if (err)
			goto out_close;
	}

	cow_sizes(version, size, sectorsize, align, *bitmap_offset_out,
		  bitmap_len_out, data_offset_out);

	return fd;
 out_close:
	os_close_file(fd);
	return err;
}

int create_cow_file(char *cow_file, char *backing_file, struct openflags flags,
		    int sectorsize, int alignment, int *bitmap_offset_out,
		    unsigned long *bitmap_len_out, int *data_offset_out)
{
	int err, fd;

	flags.c = 1;
	fd = open_ubd_file(cow_file, &flags, 0, NULL, NULL, NULL, NULL, NULL);
	if(fd < 0){
		err = fd;
		printk("Open of COW file '%s' failed, errno = %d\n", cow_file,
		       -err);
		goto out;
	}

	err = init_cow_file(fd, cow_file, backing_file, sectorsize, alignment,
			    bitmap_offset_out, bitmap_len_out,
			    data_offset_out);
	if(!err)
		return fd;
	os_close_file(fd);
 out:
	return err;
}

static int update_bitmap(struct io_thread_req *req)
{
	int n;

	if(req->cow_offset == -1)
		return 0;

	n = os_seek_file(req->fds[1], req->cow_offset);
	if(n < 0){
		printk("do_io - bitmap lseek failed : err = %d\n", -n);
		return 1;
	}

	n = os_write_file(req->fds[1], &req->bitmap_words,
			  sizeof(req->bitmap_words));
	if(n != sizeof(req->bitmap_words)){
		printk("do_io - bitmap update failed, err = %d fd = %d\n", -n,
		       req->fds[1]);
		return 1;
	}

	return 0;
}

void do_io(struct io_thread_req *req)
{
	char *buf;
	unsigned long len;
	int n, nsectors, start, end, bit;
	int err;
	__u64 off;

	nsectors = req->length / req->sectorsize;
	start = 0;
	do {
		bit = ubd_test_bit(start, (unsigned char *) &req->sector_mask);
		end = start;
		while((end < nsectors) &&
		      (ubd_test_bit(end, (unsigned char *)
				    &req->sector_mask) == bit))
			end++;

		off = req->offset + req->offsets[bit] +
			start * req->sectorsize;
		len = (end - start) * req->sectorsize;
		buf = &req->buffer[start * req->sectorsize];

		err = os_seek_file(req->fds[bit], off);
		if(err < 0){
			printk("do_io - lseek failed : err = %d\n", -err);
			req->error = 1;
			return;
		}
		if(req->op == UBD_READ){
			n = 0;
			do {
				buf = &buf[n];
				len -= n;
				n = os_read_file(req->fds[bit], buf, len);
				if (n < 0) {
					printk("do_io - read failed, err = %d "
					       "fd = %d\n", -n, req->fds[bit]);
					req->error = 1;
					return;
				}
			} while((n < len) && (n != 0));
			if (n < len) memset(&buf[n], 0, len - n);
		} else {
			n = os_write_file(req->fds[bit], buf, len);
			if(n != len){
				printk("do_io - write failed err = %d "
				       "fd = %d\n", -n, req->fds[bit]);
				req->error = 1;
				return;
			}
		}

		start = end;
	} while(start < nsectors);

	req->error = update_bitmap(req);
}

/* Changed in start_io_thread, which is serialized by being called only
 * from ubd_init, which is an initcall.
 */
int kernel_fd = -1;

/* Only changed by the io thread. XXX: currently unused. */
static int io_count = 0;

int io_thread(void *arg)
{
	struct io_thread_req *req;
	int n;

	ignore_sigwinch_sig();
	while(1){
		n = os_read_file(kernel_fd, &req,
				 sizeof(struct io_thread_req *));
		if(n != sizeof(struct io_thread_req *)){
			if(n < 0)
				printk("io_thread - read failed, fd = %d, "
				       "err = %d\n", kernel_fd, -n);
			else {
				printk("io_thread - short read, fd = %d, "
				       "length = %d\n", kernel_fd, n);
			}
			continue;
		}
		io_count++;
		do_io(req);
		n = os_write_file(kernel_fd, &req,
				  sizeof(struct io_thread_req *));
		if(n != sizeof(struct io_thread_req *))
			printk("io_thread - write failed, fd = %d, err = %d\n",
			       kernel_fd, -n);
	}

	return 0;
}