ubd_kern.c 21.5 KB
Newer Older
1 2 3 4 5 6 7 8
/* 
 * Copyright (C) 2000 Jeff Dike (jdike@karaya.com)
 * Licensed under the GPL
 */

/* 2001-09-28...2002-04-17
 * Partition stuff by James_McMechan@hotmail.com
 * old style ubd by setting UBD_SHIFT to 0
9 10
 * 2002-09-27...2002-10-18 massive tinkering for 2.5
 * partitions have changed in 2.5
11 12 13 14 15 16
 */

#define MAJOR_NR UBD_MAJOR
#define UBD_SHIFT 4

#include "linux/config.h"
17
#include "linux/module.h"
18 19 20 21 22 23 24 25 26 27 28 29 30
#include "linux/blk.h"
#include "linux/blkdev.h"
#include "linux/hdreg.h"
#include "linux/init.h"
#include "linux/devfs_fs_kernel.h"
#include "linux/cdrom.h"
#include "linux/proc_fs.h"
#include "linux/ctype.h"
#include "linux/capability.h"
#include "linux/mm.h"
#include "linux/vmalloc.h"
#include "linux/blkpg.h"
#include "linux/genhd.h"
31
#include "linux/spinlock.h"
32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47
#include "asm/segment.h"
#include "asm/uaccess.h"
#include "asm/irq.h"
#include "asm/types.h"
#include "asm/tlbflush.h"
#include "user_util.h"
#include "mem_user.h"
#include "kern_util.h"
#include "kern.h"
#include "mconsole_kern.h"
#include "init.h"
#include "irq_user.h"
#include "ubd_user.h"
#include "2_5compat.h"
#include "os.h"

48 49 50
static spinlock_t ubd_io_lock = SPIN_LOCK_UNLOCKED;
static spinlock_t ubd_lock = SPIN_LOCK_UNLOCKED;

51 52 53 54 55 56 57 58 59 60 61 62 63
static void (*do_ubd)(void);

static int ubd_open(struct inode * inode, struct file * filp);
static int ubd_release(struct inode * inode, struct file * file);
static int ubd_ioctl(struct inode * inode, struct file * file,
		     unsigned int cmd, unsigned long arg);

#define MAX_DEV (8)
#define MAX_MINOR (MAX_DEV << UBD_SHIFT)

#define DEVICE_NR(n) (minor(n) >> UBD_SHIFT)

static struct block_device_operations ubd_blops = {
64 65 66 67
        .owner		= THIS_MODULE,
        .open		= ubd_open,
        .release	= ubd_release,
        .ioctl		= ubd_ioctl,
68 69
};

70
/* Protected by the queue_lock */
71
static request_queue_t ubd_queue;
72

73
/* Protected by ubd_lock */
74
static int fake_major = 0;
75

76 77
static struct gendisk *ubd_gendisk[MAX_DEV];
static struct gendisk *fake_gendisk[MAX_DEV];
78
 
79
#ifdef CONFIG_BLK_DEV_UBD_SYNC
80
#define OPEN_FLAGS ((struct openflags) { .r = 1, .w = 1, .s = 1, .c = 0 })
81
#else
82
#define OPEN_FLAGS ((struct openflags) { .r = 1, .w = 1, .s = 0, .c = 0 })
83 84
#endif

85 86 87
/* Not protected - changed only in ubd_setup_common and then only to
 * to enable O_SYNC.
 */
88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112
static struct openflags global_openflags = OPEN_FLAGS;

struct cow {
	char *file;
	int fd;
	unsigned long *bitmap;
	unsigned long bitmap_len;
	int bitmap_offset;
        int data_offset;
};

struct ubd {
	char *file;
	int is_dir;
	int count;
	int fd;
	__u64 size;
	struct openflags boot_openflags;
	struct openflags openflags;
	devfs_handle_t real;
	devfs_handle_t fake;
	struct cow cow;
};

#define DEFAULT_COW { \
113 114 115 116 117
	.file =			NULL, \
        .fd =			-1, \
        .bitmap =		NULL, \
	.bitmap_offset =	0, \
        .data_offset =		0, \
118 119 120
}

#define DEFAULT_UBD { \
121 122 123 124 125 126 127 128 129 130
	.file = 		NULL, \
	.is_dir =		0, \
	.count =		0, \
	.fd =			-1, \
	.size =			-1, \
	.boot_openflags =	OPEN_FLAGS, \
	.openflags =		OPEN_FLAGS, \
	.real =			NULL, \
	.fake =			NULL, \
        .cow =			DEFAULT_COW, \
131 132 133 134 135 136 137 138 139 140 141 142 143
}

struct ubd ubd_dev[MAX_DEV] = { [ 0 ... MAX_DEV - 1 ] = DEFAULT_UBD };

static int ubd0_init(void)
{
	if(ubd_dev[0].file == NULL)
		ubd_dev[0].file = "root_fs";
	return(0);
}

__initcall(ubd0_init);

144
/* Only changed by fake_ide_setup which is a setup */
145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177
static int fake_ide = 0;
static struct proc_dir_entry *proc_ide_root = NULL;
static struct proc_dir_entry *proc_ide = NULL;

static void make_proc_ide(void)
{
	proc_ide_root = proc_mkdir("ide", 0);
	proc_ide = proc_mkdir("ide0", proc_ide_root);
}

static int proc_ide_read_media(char *page, char **start, off_t off, int count,
			       int *eof, void *data)
{
	int len;

	strcpy(page, "disk\n");
	len = strlen("disk\n");
	len -= off;
	if (len < count){
		*eof = 1;
		if (len <= 0) return 0;
	}
	else len = count;
	*start = page + off;
	return len;
}

static void make_ide_entries(char *dev_name)
{
	struct proc_dir_entry *dir, *ent;
	char name[64];

	if(proc_ide_root == NULL) make_proc_ide();
178

179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206
	dir = proc_mkdir(dev_name, proc_ide);
	ent = create_proc_entry("media", S_IFREG|S_IRUGO, dir);
	if(!ent) return;
	ent->nlink = 1;
	ent->data = NULL;
	ent->read_proc = proc_ide_read_media;
	ent->write_proc = NULL;
	sprintf(name,"ide0/%s", dev_name);
	proc_symlink(dev_name, proc_ide_root, name);
}

static int fake_ide_setup(char *str)
{
	fake_ide = 1;
	return(1);
}

__setup("fake_ide", fake_ide_setup);

__uml_help(fake_ide_setup,
"fake_ide\n"
"    Create ide0 entries that map onto ubd devices.\n\n"
);

static int ubd_setup_common(char *str, int *index_out)
{
	struct openflags flags = global_openflags;
	char *backing_file;
207
	int n, err;
208 209 210 211

	if(index_out) *index_out = -1;
	n = *str++;
	if(n == '='){
212
		static int fake_major_allowed = 1;
213 214 215 216 217 218 219 220 221 222 223 224 225
		char *end;
		int major;

		if(!strcmp(str, "sync")){
			global_openflags.s = 1;
			return(0);
		}
		major = simple_strtoul(str, &end, 0);
		if(*end != '\0'){
			printk(KERN_ERR 
			       "ubd_setup : didn't parse major number\n");
			return(1);
		}
226 227 228 229 230 231

		if(!fake_major_allowed){
			printk(KERN_ERR "Can't assign a fake major twice\n");
			return(1);
		}

232 233 234 235 236 237 238 239
		err = 1;
 		spin_lock(&ubd_lock);
 		if(!fake_major_allowed){
 			printk(KERN_ERR "Can't assign a fake major twice\n");
 			goto out1;
 		}
 
 		fake_major = major;
240 241
		fake_major_allowed = 0;

242 243
		printk(KERN_INFO "Setting extra ubd major number to %d\n",
		       major);
244 245 246 247
 		err = 0;
 	out1:
 		spin_unlock(&ubd_lock);
		return(err);
248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264
	}

	if(n < '0'){
		printk(KERN_ERR "ubd_setup : index out of range\n"); }

	if((n >= '0') && (n <= '9')) n -= '0';
	else if((n >= 'a') && (n <= 'z')) n -= 'a';
	else {
		printk(KERN_ERR "ubd_setup : device syntax invalid\n");
		return(1);
	}
	if(n >= MAX_DEV){
		printk(KERN_ERR "ubd_setup : index out of range "
		       "(%d devices)\n", MAX_DEV);	
		return(1);
	}

265 266 267
	err = 1;
	spin_lock(&ubd_lock);

268 269
	if(ubd_dev[n].file != NULL){
		printk(KERN_ERR "ubd_setup : device already configured\n");
270
		goto out2;
271 272 273 274 275 276 277 278 279 280 281 282 283 284
	}

	if(index_out) *index_out = n;

	if (*str == 'r'){
		flags.w = 0;
		str++;
	}
	if (*str == 's'){
		flags.s = 1;
		str++;
	}
	if(*str++ != '='){
		printk(KERN_ERR "ubd_setup : Expected '='\n");
285
		goto out2;
286
	}
287 288

	err = 0;
289 290 291 292 293 294 295 296 297 298
	backing_file = strchr(str, ',');
	if(backing_file){
		*backing_file = '\0';
		backing_file++;
	}
	ubd_dev[n].file = str;
	if(ubd_is_dir(ubd_dev[n].file))
		ubd_dev[n].is_dir = 1;
	ubd_dev[n].cow.file = backing_file;
	ubd_dev[n].boot_openflags = flags;
299 300 301
 out2:
	spin_unlock(&ubd_lock);
	return(err);
302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324
}

static int ubd_setup(char *str)
{
	ubd_setup_common(str, NULL);
	return(1);
}

__setup("ubd", ubd_setup);
__uml_help(ubd_setup,
"ubd<n>=<filename>\n"
"    This is used to associate a device with a file in the underlying\n"
"    filesystem. Usually, there is a filesystem in the file, but \n"
"    that's not required. Swap devices containing swap files can be\n"
"    specified like this. Also, a file which doesn't contain a\n"
"    filesystem can have its contents read in the virtual \n"
"    machine by running dd on the device. n must be in the range\n"
"    0 to 7. Appending an 'r' to the number will cause that device\n"
"    to be mounted read-only. For example ubd1r=./ext_fs. Appending\n"
"    an 's' (has to be _after_ 'r', if there is one) will cause data\n"
"    to be written to disk on the host immediately.\n\n"
);

Alexander Viro's avatar
Alexander Viro committed
325
static int fakehd_set = 0;
326 327 328
static int fakehd(char *str)
{
	printk(KERN_INFO 
Alexander Viro's avatar
Alexander Viro committed
329 330 331
	       "fakehd : Changing ubd name to \"hd\".\n");
	fakehd_set = 1;
	return 1;
332 333 334 335 336 337 338 339 340 341
}

__setup("fakehd", fakehd);
__uml_help(fakehd,
"fakehd\n"
"    Change the ubd device name to \"hd\".\n\n"
);

static void do_ubd_request(request_queue_t * q);

342
/* Only changed by ubd_init, which is an initcall. */
343 344
int thread_fd = -1;

345 346 347
/* Changed by ubd_handler, which is serialized because interrupts only
 * happen on CPU 0.
 */
348 349
int intr_count = 0;

350
static void ubd_finish(struct request *req, int error)
351 352 353 354
{
	int nsect;

	if(error){
355
 		spin_lock(&ubd_io_lock);
356
		end_request(req, 0);
357
 		spin_unlock(&ubd_io_lock);
358 359
		return;
	}
360 361 362 363 364 365
	nsect = req->current_nr_sectors;
	req->sector += nsect;
	req->buffer += nsect << 9;
	req->errors = 0;
	req->nr_sectors -= nsect;
	req->current_nr_sectors = 0;
366
	spin_lock(&ubd_io_lock);
367
	end_request(req, 1);
368
	spin_unlock(&ubd_io_lock);
369 370 371 372 373
}

static void ubd_handler(void)
{
	struct io_thread_req req;
374
	struct request *rq = elv_next_request(&ubd_queue);
375 376 377 378 379 380 381 382
	int n;

	do_ubd = NULL;
	intr_count++;
	n = read_ubd_fs(thread_fd, &req, sizeof(req));
	if(n != sizeof(req)){
		printk(KERN_ERR "Pid %d - spurious interrupt in ubd_handler, "
		       "errno = %d\n", os_getpid(), -n);
383
		spin_lock(&ubd_io_lock);
384
		end_request(rq, 0);
385
		spin_unlock(&ubd_io_lock);
386 387 388
		return;
	}
        
389 390
        if((req.offset != ((__u64) (rq->sector)) << 9) ||
	   (req.length != (rq->current_nr_sectors) << 9))
391 392
		panic("I/O op mismatch");
	
393
	ubd_finish(rq, req.error);
394
	reactivate_fd(thread_fd, UBD_IRQ);	
395
	do_ubd_request(&ubd_queue);
396 397 398 399 400 401 402
}

static void ubd_intr(int irq, void *dev, struct pt_regs *unused)
{
	ubd_handler();
}

403
/* Only changed by ubd_init, which is an initcall. */
404 405 406 407 408 409 410 411 412
static int io_pid = -1;

void kill_io_thread(void)
{
	if(io_pid != -1) kill(io_pid, SIGKILL);
}

__uml_exitcall(kill_io_thread);

Alexander Viro's avatar
Alexander Viro committed
413 414 415 416 417 418 419 420
static int ubd_file_size(struct ubd *dev, __u64 *size_out)
{
	char *file;

	file = dev->cow.file ? dev->cow.file : dev->file;
	return(os_file_size(file, size_out));
}

421 422 423 424 425
static void ubd_close(struct ubd *dev)
{
	os_close_file(dev->fd);
	if(dev->cow.file == NULL)
		return;
426

427 428 429 430 431 432
	os_close_file(dev->cow.fd);
	vfree(dev->cow.bitmap);
	dev->cow.bitmap = NULL;
}

static int ubd_open_dev(struct ubd *dev)
433
{
434 435
	struct openflags flags;
	int err, n, create_cow, *create_ptr;
Alexander Viro's avatar
Alexander Viro committed
436

437 438 439 440 441
	create_cow = 0;
	create_ptr = (dev->cow.file != NULL) ? &create_cow : NULL;
	dev->fd = open_ubd_file(dev->file, &dev->openflags, &dev->cow.file,
				&dev->cow.bitmap_offset, &dev->cow.bitmap_len, 
				&dev->cow.data_offset, create_ptr);
Alexander Viro's avatar
Alexander Viro committed
442

443 444 445 446 447 448 449 450 451 452
	if((dev->fd == -ENOENT) && create_cow){
		n = dev - ubd_dev;
		dev->fd = create_cow_file(dev->file, dev->cow.file, 
					  dev->openflags, 1 << 9,
					  &dev->cow.bitmap_offset, 
					  &dev->cow.bitmap_len,
					  &dev->cow.data_offset);
		if(dev->fd >= 0){
			printk(KERN_INFO "Creating \"%s\" as COW file for "
			       "\"%s\"\n", dev->file, dev->cow.file);
453
		}
Alexander Viro's avatar
Alexander Viro committed
454 455
	}

456
	if(dev->fd < 0) return(dev->fd);
457

458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474
	if(dev->cow.file != NULL){
		err = -ENOMEM;
		dev->cow.bitmap = (void *) vmalloc(dev->cow.bitmap_len);
		if(dev->cow.bitmap == NULL) goto error;
		flush_tlb_kernel_vm();

		err = read_cow_bitmap(dev->fd, dev->cow.bitmap, 
				      dev->cow.bitmap_offset, 
				      dev->cow.bitmap_len);
		if(err) goto error;

		flags = dev->openflags;
		flags.w = 0;
		err = open_ubd_file(dev->cow.file, &flags, NULL, NULL, NULL, 
				    NULL, NULL);
		if(err < 0) goto error;
		dev->cow.fd = err;
475
	}
476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507
	return(0);
 error:
	os_close_file(dev->fd);
	return(err);
}

static int ubd_new_disk(int major, u64 size, char *name, int unit,
			struct gendisk **disk_out, devfs_handle_t dir_handle,
			devfs_handle_t *handle_out)
{
	char devfs_name[sizeof("nnnnnn\0")];
	struct gendisk *disk;
	int minor = unit << UBD_SHIFT;

	disk = alloc_disk(1 << UBD_SHIFT);
	if(disk == NULL)
		return(-ENOMEM);

	disk->major = major;
	disk->first_minor = minor;
	disk->fops = &ubd_blops;
	set_capacity(disk, size / 512);
	/* needs to be ubd -> /dev/ubd/discX/disc */
	sprintf(disk->disk_name, "ubd");
	*disk_out = disk;

	/* /dev/ubd/N style names */
	sprintf(devfs_name, "%d", unit);
	*handle_out = devfs_register(dir_handle, devfs_name,
				     DEVFS_FL_REMOVABLE, major, minor,
				     S_IFBLK | S_IRUSR | S_IWUSR | S_IRGRP |
				     S_IWGRP, &ubd_blops, NULL);
508 509
	disk->private_data = &ubd_dev[unit];
	disk->queue = &ubd_queue;
510
	add_disk(disk);
511
	return(0);
512
}
513

514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549
/* Initialized in an initcall, and unchanged thereafter */
devfs_handle_t ubd_dir_handle;
devfs_handle_t ubd_fake_dir_handle;

static int ubd_add(int n)
{
	struct ubd *dev = &ubd_dev[n];
	int err;

	if (!dev->file || dev->is_dir)
		return(-ENODEV);

	if (ubd_open_dev(dev))
		return(-ENODEV);

	err = ubd_file_size(dev, &dev->size);
	if(err)
		return(err);

	err = ubd_new_disk(MAJOR_NR, dev->size, "ubd", n, &ubd_gendisk[n], 
			   ubd_dir_handle, &dev->real);
	if(err) 
		return(err);
 
	if(fake_major)
		ubd_new_disk(fake_major, dev->size, "ubd%d", n, 
			     &fake_gendisk[n], ubd_fake_dir_handle, 
			     &dev->fake);

	/* perhaps this should also be under the "if (fake_major)" above */
	/* using the fake_disk->disk_name and also the fakehd_set name */
	if (fake_ide)
		make_ide_entries(ubd_gendisk[n]->disk_name);

	ubd_close(dev);
	return 0;
550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567
}

static int ubd_config(char *str)
{
	int n, err;

	str = uml_strdup(str);
	if(str == NULL){
		printk(KERN_ERR "ubd_config failed to strdup string\n");
		return(1);
	}
	err = ubd_setup_common(str, &n);
	if(err){
		kfree(str);
		return(-1);
	}
	if(n == -1) return(0);

568
 	spin_lock(&ubd_lock);
569
	err = ubd_add(n);
570
	if(err)
571
		ubd_dev[n].file = NULL;
572
 	spin_unlock(&ubd_lock);
573

574
	return(err);
575 576
}

577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614
static int ubd_get_config(char *dev, char *str, int size, char **error_out)
{
	struct ubd *ubd;
	char *end;
	int major, n = 0;

	major = simple_strtoul(dev, &end, 0);
	if((*end != '\0') || (end == dev)){
		*error_out = "ubd_get_config : didn't parse major number";
		return(-1);
	}

	if((major >= MAX_DEV) || (major < 0)){
		*error_out = "ubd_get_config : major number out of range";
		return(-1);
	}

	ubd = &ubd_dev[major];
	spin_lock(&ubd_lock);

	if(ubd->file == NULL){
		CONFIG_CHUNK(str, size, n, "", 1);
		goto out;
	}

	CONFIG_CHUNK(str, size, n, ubd->file, 0);

	if(ubd->cow.file != NULL){
		CONFIG_CHUNK(str, size, n, ",", 0);
		CONFIG_CHUNK(str, size, n, ubd->cow.file, 1);
	}
	else CONFIG_CHUNK(str, size, n, "", 1);

 out:
	spin_unlock(&ubd_lock);
	return(n);
}

615 616 617
static int ubd_remove(char *str)
{
	struct ubd *dev;
618 619 620 621
	int n, err = -ENODEV;

	if(!isdigit(*str))
		return(err);	/* it should be a number 0-7/a-h */
622 623

	n = *str - '0';
624
	if(n > MAX_DEV) 
625 626
		return(err);

627
	dev = &ubd_dev[n];
628 629
	if(dev->count > 0)
		return(-EBUSY);	/* you cannot remove a open disk */
630 631 632

	err = 0;
 	spin_lock(&ubd_lock);
633 634 635 636

	if(ubd_gendisk[n] == NULL)
		goto out;

637 638 639
	del_gendisk(ubd_gendisk[n]);
	put_disk(ubd_gendisk[n]);
	ubd_gendisk[n] = NULL;
640 641 642 643
	if(dev->real != NULL) 
		devfs_unregister(dev->real);

	if(fake_gendisk[n] != NULL){
644 645 646
		del_gendisk(fake_gendisk[n]);
		put_disk(fake_gendisk[n]);
		fake_gendisk[n] = NULL;
647 648
		if(dev->fake != NULL) 
			devfs_unregister(dev->fake);
649
	}
650

651
	*dev = ((struct ubd) DEFAULT_UBD);
652 653 654 655
	err = 0;
 out:
 	spin_unlock(&ubd_lock);
	return(err);
656 657 658
}

static struct mc_device ubd_mc = {
659 660
	.name		= "ubd",
	.config		= ubd_config,
661
 	.get_config	= ubd_get_config,
662
	.remove		= ubd_remove,
663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681
};

static int ubd_mc_init(void)
{
	mconsole_register_dev(&ubd_mc);
	return(0);
}

__initcall(ubd_mc_init);

int ubd_init(void)
{
        int i;

	ubd_dir_handle = devfs_mk_dir (NULL, "ubd", NULL);
	if(register_blkdev(MAJOR_NR, "ubd", &ubd_blops)){
		printk(KERN_ERR "ubd: unable to get major %d\n", MAJOR_NR);
		return -1;
	}
682 683 684 685

	blk_init_queue(&ubd_queue, do_ubd_request, &ubd_io_lock);
	elevator_init(&ubd_queue, &elevator_noop);

686 687 688 689 690 691 692 693 694 695 696
	if(fake_major != 0){
		char name[sizeof("ubd_nnn\0")];

		snprintf(name, sizeof(name), "ubd_%d", fake_major);
		ubd_fake_dir_handle = devfs_mk_dir(NULL, name, NULL);
		if(register_blkdev(fake_major, "ubd", &ubd_blops)){
			printk(KERN_ERR "ubd: unable to get major %d\n",
			       fake_major);
			return -1;
		}
	}
697 698
	for(i = 0; i < MAX_DEV; i++) 
		ubd_add(i);
699 700 701 702 703 704 705 706 707
	return(0);
}

late_initcall(ubd_init);

int ubd_driver_init(void){
	unsigned long stack;
	int err;

708
	if(global_openflags.s){
709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731
		printk(KERN_INFO "ubd : Synchronous mode\n");
		return(0);
	}
	stack = alloc_stack(0, 0);
	io_pid = start_io_thread(stack + PAGE_SIZE - sizeof(void *), 
				 &thread_fd);
	if(io_pid < 0){
		printk(KERN_ERR 
		       "ubd : Failed to start I/O thread (errno = %d) - "
		       "falling back to synchronous I/O\n", -io_pid);
		return(0);
	}
	err = um_request_irq(UBD_IRQ, thread_fd, IRQ_READ, ubd_intr, 
			     SA_INTERRUPT, "ubd", ubd_dev);
	if(err != 0) printk(KERN_ERR 
			    "um_request_irq failed - errno = %d\n", -err);
	return(err);
}

device_initcall(ubd_driver_init);

static int ubd_open(struct inode *inode, struct file *filp)
{
732 733
	struct gendisk *disk = inode->i_bdev->bd_disk;
	struct ubd *dev = disk->private_data;
Jeff Dike's avatar
Jeff Dike committed
734
	int err = -EISDIR;
735

736
	if(dev->is_dir == 1)
737
		goto out;
738

Jeff Dike's avatar
Jeff Dike committed
739
	err = 0;
740 741 742 743 744
	if(dev->count == 0){
		dev->openflags = dev->boot_openflags;

		err = ubd_open_dev(dev);
		if(err){
745 746
			printk(KERN_ERR "%s: Can't open \"%s\": errno = %d\n",
			       disk->disk_name, dev->file, -err);
747
			goto out;
748 749 750 751 752
		}
	}
	dev->count++;
	if((filp->f_mode & FMODE_WRITE) && !dev->openflags.w){
	        if(--dev->count == 0) ubd_close(dev);
753
	        err = -EROFS;
754
	}
755 756
 out:
	return(err);
757 758 759 760
}

static int ubd_release(struct inode * inode, struct file * file)
{
761 762 763 764 765
	struct gendisk *disk = inode->i_bdev->bd_disk;
	struct ubd *dev = disk->private_data;

	if(--dev->count == 0)
		ubd_close(dev);
766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808
	return(0);
}

void cowify_req(struct io_thread_req *req, struct ubd *dev)
{
        int i, update_bitmap, sector = req->offset >> 9;

	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 *) 
					dev->cow.bitmap)){
				ubd_set_bit(i, (unsigned char *) 
					    &req->sector_mask);
			}
                }
        } 
        else {
		update_bitmap = 0;
		for(i = 0; i < req->length >> 9; i++){
			ubd_set_bit(i, (unsigned char *) 
				    &req->sector_mask);
			if(!ubd_test_bit(sector + i, (unsigned char *) 
					 dev->cow.bitmap))
				update_bitmap = 1;
			ubd_set_bit(sector + i, (unsigned char *) 
				    dev->cow.bitmap);
		}
		if(update_bitmap){
			req->cow_offset = sector / (sizeof(unsigned long) * 8);
			req->bitmap_words[0] = 
				dev->cow.bitmap[req->cow_offset];
			req->bitmap_words[1] = 
				dev->cow.bitmap[req->cow_offset + 1];
			req->cow_offset *= sizeof(unsigned long);
			req->cow_offset += dev->cow.bitmap_offset;
		}
	}
}

static int prepare_request(struct request *req, struct io_thread_req *io_req)
{
809 810
	struct gendisk *disk = req->rq_disk;
	struct ubd *dev = disk->private_data;
811
	__u64 block;
812
	int nsect;
813 814 815 816 817 818

	if(req->rq_status == RQ_INACTIVE) return(1);

	if(dev->is_dir){
		strcpy(req->buffer, "HOSTFS:");
		strcat(req->buffer, dev->file);
819
 		spin_lock(&ubd_io_lock);
820
		end_request(req, 1);
821
 		spin_unlock(&ubd_io_lock);
822 823 824 825
		return(1);
	}

	if((rq_data_dir(req) == WRITE) && !dev->openflags.w){
826 827
		printk("Write attempted on readonly ubd device %s\n", 
		       disk->disk_name);
828
 		spin_lock(&ubd_io_lock);
829
		end_request(req, 0);
830
 		spin_unlock(&ubd_io_lock);
831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864 865
		return(1);
	}

        block = req->sector;
        nsect = req->current_nr_sectors;

	io_req->op = rq_data_dir(req) == READ ? UBD_READ : UBD_WRITE;
	io_req->fds[0] = (dev->cow.file != NULL) ? dev->cow.fd : dev->fd;
	io_req->fds[1] = dev->fd;
	io_req->offsets[0] = 0;
	io_req->offsets[1] = dev->cow.data_offset;
	io_req->offset = ((__u64) block) << 9;
	io_req->length = nsect << 9;
	io_req->buffer = req->buffer;
	io_req->sectorsize = 1 << 9;
	io_req->sector_mask = 0;
	io_req->cow_offset = -1;
	io_req->error = 0;

        if(dev->cow.file != NULL) cowify_req(io_req, dev);
	return(0);
}

static void do_ubd_request(request_queue_t *q)
{
	struct io_thread_req io_req;
	struct request *req;
	int err, n;

	if(thread_fd == -1){
		while(!list_empty(&q->queue_head)){
			req = elv_next_request(q);
			err = prepare_request(req, &io_req);
			if(!err){
				do_io(&io_req);
866
				ubd_finish(req, io_req.error);
867 868 869 870 871 872 873 874 875 876 877 878 879 880 881 882 883 884 885 886 887 888
			}
		}
	}
	else {
		if(do_ubd || list_empty(&q->queue_head)) return;
		req = elv_next_request(q);
		err = prepare_request(req, &io_req);
		if(!err){
			do_ubd = ubd_handler;
			n = write_ubd_fs(thread_fd, (char *) &io_req, 
					 sizeof(io_req));
			if(n != sizeof(io_req))
				printk("write to io thread failed, "
				       "errno = %d\n", -n);
		}
	}
}

static int ubd_ioctl(struct inode * inode, struct file * file,
		     unsigned int cmd, unsigned long arg)
{
	struct hd_geometry *loc = (struct hd_geometry *) arg;
889 890
	struct ubd *dev = inode->i_bdev->bd_disk->private_data;
	int err;
891 892 893 894 895 896
	struct hd_driveid ubd_id = {
		.cyls =		0,
		.heads =	128,
		.sectors =	32,
	};

897 898 899 900 901 902 903
	switch (cmd) {
	        struct hd_geometry g;
		struct cdrom_volctrl volume;
	case HDIO_GETGEO:
		if(!loc) return(-EINVAL);
		g.heads = 128;
		g.sectors = 32;
Alexander Viro's avatar
Alexander Viro committed
904
		g.cylinders = dev->size / (128 * 32 * 512);
905 906 907 908 909
		g.start = 2;
		return(copy_to_user(loc, &g, sizeof(g)) ? -EFAULT : 0);

	case HDIO_SET_UNMASKINTR:
		if(!capable(CAP_SYS_ADMIN)) return(-EACCES);
910
		if((arg > 1) || (inode->i_bdev->bd_contains != inode->i_bdev))
911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927 928 929
			return(-EINVAL);
		return(0);

	case HDIO_GET_UNMASKINTR:
		if(!arg)  return(-EINVAL);
		err = verify_area(VERIFY_WRITE, (long *) arg, sizeof(long));
		if(err)
			return(err);
		return(0);

	case HDIO_GET_MULTCOUNT:
		if(!arg)  return(-EINVAL);
		err = verify_area(VERIFY_WRITE, (long *) arg, sizeof(long));
		if(err)
			return(err);
		return(0);

	case HDIO_SET_MULTCOUNT:
		if(!capable(CAP_SYS_ADMIN)) return(-EACCES);
930
		if(inode->i_bdev->bd_contains != inode->i_bdev)
931 932 933 934
			return(-EINVAL);
		return(0);

	case HDIO_GET_IDENTITY:
Alexander Viro's avatar
Alexander Viro committed
935
		ubd_id.cyls = dev->size / (128 * 32 * 512);
936 937 938 939 940 941 942 943 944 945 946 947 948 949 950 951
		if(copy_to_user((char *) arg, (char *) &ubd_id, 
				 sizeof(ubd_id)))
			return(-EFAULT);
		return(0);
		
	case CDROMVOLREAD:
		if(copy_from_user(&volume, (char *) arg, sizeof(volume)))
			return(-EFAULT);
		volume.channel0 = 255;
		volume.channel1 = 255;
		volume.channel2 = 255;
		volume.channel3 = 255;
		if(copy_to_user((char *) arg, &volume, sizeof(volume)))
			return(-EFAULT);
		return(0);
	}
952
	return(-EINVAL);
953 954 955 956 957 958 959 960 961 962 963 964
}

/*
 * Overrides for Emacs so that we follow Linus's tabbing style.
 * Emacs will notice this stuff at the end of the file and automatically
 * adjust the settings for this buffer only.  This must remain at the end
 * of the file.
 * ---------------------------------------------------------------------------
 * Local variables:
 * c-file-style: "linux"
 * End:
 */