kexec.c 27.5 KB
Newer Older
1 2 3 4 5 6 7 8
/*
 * kexec.c - kexec system call
 * Copyright (C) 2002-2004 Eric Biederman  <ebiederm@xmission.com>
 *
 * This source code is licensed under the GNU General Public License,
 * Version 2.  See the file COPYING for more details.
 */

9
#include <linux/capability.h>
10 11 12 13 14 15 16 17 18 19 20 21
#include <linux/mm.h>
#include <linux/file.h>
#include <linux/slab.h>
#include <linux/fs.h>
#include <linux/kexec.h>
#include <linux/spinlock.h>
#include <linux/list.h>
#include <linux/highmem.h>
#include <linux/syscalls.h>
#include <linux/reboot.h>
#include <linux/syscalls.h>
#include <linux/ioport.h>
22 23
#include <linux/hardirq.h>

24 25 26 27 28 29
#include <asm/page.h>
#include <asm/uaccess.h>
#include <asm/io.h>
#include <asm/system.h>
#include <asm/semaphore.h>

30 31 32
/* Per cpu memory for storing cpu states in case of system crash. */
note_buf_t* crash_notes;

33 34 35 36 37 38 39 40
/* Location of the reserved area for the crash kernel */
struct resource crashk_res = {
	.name  = "Crash kernel",
	.start = 0,
	.end   = 0,
	.flags = IORESOURCE_BUSY | IORESOURCE_MEM
};

41 42
int kexec_should_crash(struct task_struct *p)
{
43
	if (in_interrupt() || !p->pid || is_init(p) || panic_on_oops)
44 45 46 47
		return 1;
	return 0;
}

48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93
/*
 * When kexec transitions to the new kernel there is a one-to-one
 * mapping between physical and virtual addresses.  On processors
 * where you can disable the MMU this is trivial, and easy.  For
 * others it is still a simple predictable page table to setup.
 *
 * In that environment kexec copies the new kernel to its final
 * resting place.  This means I can only support memory whose
 * physical address can fit in an unsigned long.  In particular
 * addresses where (pfn << PAGE_SHIFT) > ULONG_MAX cannot be handled.
 * If the assembly stub has more restrictive requirements
 * KEXEC_SOURCE_MEMORY_LIMIT and KEXEC_DEST_MEMORY_LIMIT can be
 * defined more restrictively in <asm/kexec.h>.
 *
 * The code for the transition from the current kernel to the
 * the new kernel is placed in the control_code_buffer, whose size
 * is given by KEXEC_CONTROL_CODE_SIZE.  In the best case only a single
 * page of memory is necessary, but some architectures require more.
 * Because this memory must be identity mapped in the transition from
 * virtual to physical addresses it must live in the range
 * 0 - TASK_SIZE, as only the user space mappings are arbitrarily
 * modifiable.
 *
 * The assembly stub in the control code buffer is passed a linked list
 * of descriptor pages detailing the source pages of the new kernel,
 * and the destination addresses of those source pages.  As this data
 * structure is not used in the context of the current OS, it must
 * be self-contained.
 *
 * The code has been made to work with highmem pages and will use a
 * destination page in its final resting place (if it happens
 * to allocate it).  The end product of this is that most of the
 * physical address space, and most of RAM can be used.
 *
 * Future directions include:
 *  - allocating a page table with the control code buffer identity
 *    mapped, to simplify machine_kexec and make kexec_on_panic more
 *    reliable.
 */

/*
 * KIMAGE_NO_DEST is an impossible destination address..., for
 * allocating pages whose destination address we do not care about.
 */
#define KIMAGE_NO_DEST (-1UL)

M
Maneesh Soni 已提交
94 95 96
static int kimage_is_destination_range(struct kimage *image,
				       unsigned long start, unsigned long end);
static struct page *kimage_alloc_page(struct kimage *image,
A
Al Viro 已提交
97
				       gfp_t gfp_mask,
M
Maneesh Soni 已提交
98
				       unsigned long dest);
99 100

static int do_kimage_alloc(struct kimage **rimage, unsigned long entry,
M
Maneesh Soni 已提交
101 102
	                    unsigned long nr_segments,
                            struct kexec_segment __user *segments)
103 104 105 106 107 108 109 110 111
{
	size_t segment_bytes;
	struct kimage *image;
	unsigned long i;
	int result;

	/* Allocate a controlling structure */
	result = -ENOMEM;
	image = kmalloc(sizeof(*image), GFP_KERNEL);
M
Maneesh Soni 已提交
112
	if (!image)
113
		goto out;
M
Maneesh Soni 已提交
114

115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154
	memset(image, 0, sizeof(*image));
	image->head = 0;
	image->entry = &image->head;
	image->last_entry = &image->head;
	image->control_page = ~0; /* By default this does not apply */
	image->start = entry;
	image->type = KEXEC_TYPE_DEFAULT;

	/* Initialize the list of control pages */
	INIT_LIST_HEAD(&image->control_pages);

	/* Initialize the list of destination pages */
	INIT_LIST_HEAD(&image->dest_pages);

	/* Initialize the list of unuseable pages */
	INIT_LIST_HEAD(&image->unuseable_pages);

	/* Read in the segments */
	image->nr_segments = nr_segments;
	segment_bytes = nr_segments * sizeof(*segments);
	result = copy_from_user(image->segment, segments, segment_bytes);
	if (result)
		goto out;

	/*
	 * Verify we have good destination addresses.  The caller is
	 * responsible for making certain we don't attempt to load
	 * the new image into invalid or reserved areas of RAM.  This
	 * just verifies it is an address we can use.
	 *
	 * Since the kernel does everything in page size chunks ensure
	 * the destination addreses are page aligned.  Too many
	 * special cases crop of when we don't do this.  The most
	 * insidious is getting overlapping destination addresses
	 * simply because addresses are changed to page size
	 * granularity.
	 */
	result = -EADDRNOTAVAIL;
	for (i = 0; i < nr_segments; i++) {
		unsigned long mstart, mend;
M
Maneesh Soni 已提交
155

156 157 158 159 160 161 162 163 164 165 166 167 168 169
		mstart = image->segment[i].mem;
		mend   = mstart + image->segment[i].memsz;
		if ((mstart & ~PAGE_MASK) || (mend & ~PAGE_MASK))
			goto out;
		if (mend >= KEXEC_DESTINATION_MEMORY_LIMIT)
			goto out;
	}

	/* Verify our destination addresses do not overlap.
	 * If we alloed overlapping destination addresses
	 * through very weird things can happen with no
	 * easy explanation as one segment stops on another.
	 */
	result = -EINVAL;
M
Maneesh Soni 已提交
170
	for (i = 0; i < nr_segments; i++) {
171 172
		unsigned long mstart, mend;
		unsigned long j;
M
Maneesh Soni 已提交
173

174 175
		mstart = image->segment[i].mem;
		mend   = mstart + image->segment[i].memsz;
M
Maneesh Soni 已提交
176
		for (j = 0; j < i; j++) {
177 178 179 180 181 182 183 184 185 186 187 188 189 190 191
			unsigned long pstart, pend;
			pstart = image->segment[j].mem;
			pend   = pstart + image->segment[j].memsz;
			/* Do the segments overlap ? */
			if ((mend > pstart) && (mstart < pend))
				goto out;
		}
	}

	/* Ensure our buffer sizes are strictly less than
	 * our memory sizes.  This should always be the case,
	 * and it is easier to check up front than to be surprised
	 * later on.
	 */
	result = -EINVAL;
M
Maneesh Soni 已提交
192
	for (i = 0; i < nr_segments; i++) {
193 194 195 196 197
		if (image->segment[i].bufsz > image->segment[i].memsz)
			goto out;
	}

	result = 0;
M
Maneesh Soni 已提交
198 199
out:
	if (result == 0)
200
		*rimage = image;
M
Maneesh Soni 已提交
201
	else
202
		kfree(image);
M
Maneesh Soni 已提交
203

204 205 206 207 208
	return result;

}

static int kimage_normal_alloc(struct kimage **rimage, unsigned long entry,
M
Maneesh Soni 已提交
209 210
				unsigned long nr_segments,
				struct kexec_segment __user *segments)
211 212 213 214 215 216 217
{
	int result;
	struct kimage *image;

	/* Allocate and initialize a controlling structure */
	image = NULL;
	result = do_kimage_alloc(&image, entry, nr_segments, segments);
M
Maneesh Soni 已提交
218
	if (result)
219
		goto out;
M
Maneesh Soni 已提交
220

221 222 223 224 225 226 227 228 229
	*rimage = image;

	/*
	 * Find a location for the control code buffer, and add it
	 * the vector of segments so that it's pages will also be
	 * counted as destination pages.
	 */
	result = -ENOMEM;
	image->control_code_page = kimage_alloc_control_pages(image,
M
Maneesh Soni 已提交
230
					   get_order(KEXEC_CONTROL_CODE_SIZE));
231 232 233 234 235 236 237
	if (!image->control_code_page) {
		printk(KERN_ERR "Could not allocate control_code_buffer\n");
		goto out;
	}

	result = 0;
 out:
M
Maneesh Soni 已提交
238
	if (result == 0)
239
		*rimage = image;
M
Maneesh Soni 已提交
240
	else
241
		kfree(image);
M
Maneesh Soni 已提交
242

243 244 245 246
	return result;
}

static int kimage_crash_alloc(struct kimage **rimage, unsigned long entry,
M
Maneesh Soni 已提交
247
				unsigned long nr_segments,
248
				struct kexec_segment __user *segments)
249 250 251 252 253 254 255 256 257 258 259 260 261 262
{
	int result;
	struct kimage *image;
	unsigned long i;

	image = NULL;
	/* Verify we have a valid entry point */
	if ((entry < crashk_res.start) || (entry > crashk_res.end)) {
		result = -EADDRNOTAVAIL;
		goto out;
	}

	/* Allocate and initialize a controlling structure */
	result = do_kimage_alloc(&image, entry, nr_segments, segments);
M
Maneesh Soni 已提交
263
	if (result)
264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283
		goto out;

	/* Enable the special crash kernel control page
	 * allocation policy.
	 */
	image->control_page = crashk_res.start;
	image->type = KEXEC_TYPE_CRASH;

	/*
	 * Verify we have good destination addresses.  Normally
	 * the caller is responsible for making certain we don't
	 * attempt to load the new image into invalid or reserved
	 * areas of RAM.  But crash kernels are preloaded into a
	 * reserved area of ram.  We must ensure the addresses
	 * are in the reserved area otherwise preloading the
	 * kernel could corrupt things.
	 */
	result = -EADDRNOTAVAIL;
	for (i = 0; i < nr_segments; i++) {
		unsigned long mstart, mend;
M
Maneesh Soni 已提交
284

285
		mstart = image->segment[i].mem;
286
		mend = mstart + image->segment[i].memsz - 1;
287 288 289 290 291 292 293 294 295 296 297 298
		/* Ensure we are within the crash kernel limits */
		if ((mstart < crashk_res.start) || (mend > crashk_res.end))
			goto out;
	}

	/*
	 * Find a location for the control code buffer, and add
	 * the vector of segments so that it's pages will also be
	 * counted as destination pages.
	 */
	result = -ENOMEM;
	image->control_code_page = kimage_alloc_control_pages(image,
M
Maneesh Soni 已提交
299
					   get_order(KEXEC_CONTROL_CODE_SIZE));
300 301 302 303 304 305
	if (!image->control_code_page) {
		printk(KERN_ERR "Could not allocate control_code_buffer\n");
		goto out;
	}

	result = 0;
M
Maneesh Soni 已提交
306 307
out:
	if (result == 0)
308
		*rimage = image;
M
Maneesh Soni 已提交
309
	else
310
		kfree(image);
M
Maneesh Soni 已提交
311

312 313 314
	return result;
}

M
Maneesh Soni 已提交
315 316 317
static int kimage_is_destination_range(struct kimage *image,
					unsigned long start,
					unsigned long end)
318 319 320 321 322
{
	unsigned long i;

	for (i = 0; i < image->nr_segments; i++) {
		unsigned long mstart, mend;
M
Maneesh Soni 已提交
323

324
		mstart = image->segment[i].mem;
M
Maneesh Soni 已提交
325 326
		mend = mstart + image->segment[i].memsz;
		if ((end > mstart) && (start < mend))
327 328
			return 1;
	}
M
Maneesh Soni 已提交
329

330 331 332
	return 0;
}

A
Al Viro 已提交
333
static struct page *kimage_alloc_pages(gfp_t gfp_mask, unsigned int order)
334 335
{
	struct page *pages;
M
Maneesh Soni 已提交
336

337 338 339 340
	pages = alloc_pages(gfp_mask, order);
	if (pages) {
		unsigned int count, i;
		pages->mapping = NULL;
H
Hugh Dickins 已提交
341
		set_page_private(pages, order);
342
		count = 1 << order;
M
Maneesh Soni 已提交
343
		for (i = 0; i < count; i++)
344 345
			SetPageReserved(pages + i);
	}
M
Maneesh Soni 已提交
346

347 348 349 350 351 352
	return pages;
}

static void kimage_free_pages(struct page *page)
{
	unsigned int order, count, i;
M
Maneesh Soni 已提交
353

H
Hugh Dickins 已提交
354
	order = page_private(page);
355
	count = 1 << order;
M
Maneesh Soni 已提交
356
	for (i = 0; i < count; i++)
357 358 359 360 361 362 363
		ClearPageReserved(page + i);
	__free_pages(page, order);
}

static void kimage_free_page_list(struct list_head *list)
{
	struct list_head *pos, *next;
M
Maneesh Soni 已提交
364

365 366 367 368 369 370 371 372 373
	list_for_each_safe(pos, next, list) {
		struct page *page;

		page = list_entry(pos, struct page, lru);
		list_del(&page->lru);
		kimage_free_pages(page);
	}
}

M
Maneesh Soni 已提交
374 375
static struct page *kimage_alloc_normal_control_pages(struct kimage *image,
							unsigned int order)
376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401
{
	/* Control pages are special, they are the intermediaries
	 * that are needed while we copy the rest of the pages
	 * to their final resting place.  As such they must
	 * not conflict with either the destination addresses
	 * or memory the kernel is already using.
	 *
	 * The only case where we really need more than one of
	 * these are for architectures where we cannot disable
	 * the MMU and must instead generate an identity mapped
	 * page table for all of the memory.
	 *
	 * At worst this runs in O(N) of the image size.
	 */
	struct list_head extra_pages;
	struct page *pages;
	unsigned int count;

	count = 1 << order;
	INIT_LIST_HEAD(&extra_pages);

	/* Loop while I can allocate a page and the page allocated
	 * is a destination page.
	 */
	do {
		unsigned long pfn, epfn, addr, eaddr;
M
Maneesh Soni 已提交
402

403 404 405 406 407 408 409 410
		pages = kimage_alloc_pages(GFP_KERNEL, order);
		if (!pages)
			break;
		pfn   = page_to_pfn(pages);
		epfn  = pfn + count;
		addr  = pfn << PAGE_SHIFT;
		eaddr = epfn << PAGE_SHIFT;
		if ((epfn >= (KEXEC_CONTROL_MEMORY_LIMIT >> PAGE_SHIFT)) ||
M
Maneesh Soni 已提交
411
			      kimage_is_destination_range(image, addr, eaddr)) {
412 413 414
			list_add(&pages->lru, &extra_pages);
			pages = NULL;
		}
M
Maneesh Soni 已提交
415 416
	} while (!pages);

417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436
	if (pages) {
		/* Remember the allocated page... */
		list_add(&pages->lru, &image->control_pages);

		/* Because the page is already in it's destination
		 * location we will never allocate another page at
		 * that address.  Therefore kimage_alloc_pages
		 * will not return it (again) and we don't need
		 * to give it an entry in image->segment[].
		 */
	}
	/* Deal with the destination pages I have inadvertently allocated.
	 *
	 * Ideally I would convert multi-page allocations into single
	 * page allocations, and add everyting to image->dest_pages.
	 *
	 * For now it is simpler to just free the pages.
	 */
	kimage_free_page_list(&extra_pages);

M
Maneesh Soni 已提交
437
	return pages;
438 439
}

M
Maneesh Soni 已提交
440 441
static struct page *kimage_alloc_crash_control_pages(struct kimage *image,
						      unsigned int order)
442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465
{
	/* Control pages are special, they are the intermediaries
	 * that are needed while we copy the rest of the pages
	 * to their final resting place.  As such they must
	 * not conflict with either the destination addresses
	 * or memory the kernel is already using.
	 *
	 * Control pages are also the only pags we must allocate
	 * when loading a crash kernel.  All of the other pages
	 * are specified by the segments and we just memcpy
	 * into them directly.
	 *
	 * The only case where we really need more than one of
	 * these are for architectures where we cannot disable
	 * the MMU and must instead generate an identity mapped
	 * page table for all of the memory.
	 *
	 * Given the low demand this implements a very simple
	 * allocator that finds the first hole of the appropriate
	 * size in the reserved memory region, and allocates all
	 * of the memory up to and including the hole.
	 */
	unsigned long hole_start, hole_end, size;
	struct page *pages;
M
Maneesh Soni 已提交
466

467 468 469 470
	pages = NULL;
	size = (1 << order) << PAGE_SHIFT;
	hole_start = (image->control_page + (size - 1)) & ~(size - 1);
	hole_end   = hole_start + size - 1;
M
Maneesh Soni 已提交
471
	while (hole_end <= crashk_res.end) {
472
		unsigned long i;
M
Maneesh Soni 已提交
473 474

		if (hole_end > KEXEC_CONTROL_MEMORY_LIMIT)
475
			break;
M
Maneesh Soni 已提交
476
		if (hole_end > crashk_res.end)
477 478
			break;
		/* See if I overlap any of the segments */
M
Maneesh Soni 已提交
479
		for (i = 0; i < image->nr_segments; i++) {
480
			unsigned long mstart, mend;
M
Maneesh Soni 已提交
481

482 483 484 485 486 487 488 489 490 491 492 493 494 495 496
			mstart = image->segment[i].mem;
			mend   = mstart + image->segment[i].memsz - 1;
			if ((hole_end >= mstart) && (hole_start <= mend)) {
				/* Advance the hole to the end of the segment */
				hole_start = (mend + (size - 1)) & ~(size - 1);
				hole_end   = hole_start + size - 1;
				break;
			}
		}
		/* If I don't overlap any segments I have found my hole! */
		if (i == image->nr_segments) {
			pages = pfn_to_page(hole_start >> PAGE_SHIFT);
			break;
		}
	}
M
Maneesh Soni 已提交
497
	if (pages)
498
		image->control_page = hole_end;
M
Maneesh Soni 已提交
499

500 501 502 503
	return pages;
}


M
Maneesh Soni 已提交
504 505
struct page *kimage_alloc_control_pages(struct kimage *image,
					 unsigned int order)
506 507
{
	struct page *pages = NULL;
M
Maneesh Soni 已提交
508 509

	switch (image->type) {
510 511 512 513 514 515 516
	case KEXEC_TYPE_DEFAULT:
		pages = kimage_alloc_normal_control_pages(image, order);
		break;
	case KEXEC_TYPE_CRASH:
		pages = kimage_alloc_crash_control_pages(image, order);
		break;
	}
M
Maneesh Soni 已提交
517

518 519 520 521 522
	return pages;
}

static int kimage_add_entry(struct kimage *image, kimage_entry_t entry)
{
M
Maneesh Soni 已提交
523
	if (*image->entry != 0)
524
		image->entry++;
M
Maneesh Soni 已提交
525

526 527 528
	if (image->entry == image->last_entry) {
		kimage_entry_t *ind_page;
		struct page *page;
M
Maneesh Soni 已提交
529

530
		page = kimage_alloc_page(image, GFP_KERNEL, KIMAGE_NO_DEST);
M
Maneesh Soni 已提交
531
		if (!page)
532
			return -ENOMEM;
M
Maneesh Soni 已提交
533

534 535 536
		ind_page = page_address(page);
		*image->entry = virt_to_phys(ind_page) | IND_INDIRECTION;
		image->entry = ind_page;
M
Maneesh Soni 已提交
537 538
		image->last_entry = ind_page +
				      ((PAGE_SIZE/sizeof(kimage_entry_t)) - 1);
539 540 541 542
	}
	*image->entry = entry;
	image->entry++;
	*image->entry = 0;
M
Maneesh Soni 已提交
543

544 545 546
	return 0;
}

M
Maneesh Soni 已提交
547 548
static int kimage_set_destination(struct kimage *image,
				   unsigned long destination)
549 550 551 552 553
{
	int result;

	destination &= PAGE_MASK;
	result = kimage_add_entry(image, destination | IND_DESTINATION);
M
Maneesh Soni 已提交
554
	if (result == 0)
555
		image->destination = destination;
M
Maneesh Soni 已提交
556

557 558 559 560 561 562 563 564 565 566
	return result;
}


static int kimage_add_page(struct kimage *image, unsigned long page)
{
	int result;

	page &= PAGE_MASK;
	result = kimage_add_entry(image, page | IND_SOURCE);
M
Maneesh Soni 已提交
567
	if (result == 0)
568
		image->destination += PAGE_SIZE;
M
Maneesh Soni 已提交
569

570 571 572 573 574 575 576 577 578 579 580 581 582 583 584
	return result;
}


static void kimage_free_extra_pages(struct kimage *image)
{
	/* Walk through and free any extra destination pages I may have */
	kimage_free_page_list(&image->dest_pages);

	/* Walk through and free any unuseable pages I have cached */
	kimage_free_page_list(&image->unuseable_pages);

}
static int kimage_terminate(struct kimage *image)
{
M
Maneesh Soni 已提交
585
	if (*image->entry != 0)
586
		image->entry++;
M
Maneesh Soni 已提交
587

588
	*image->entry = IND_DONE;
M
Maneesh Soni 已提交
589

590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612
	return 0;
}

#define for_each_kimage_entry(image, ptr, entry) \
	for (ptr = &image->head; (entry = *ptr) && !(entry & IND_DONE); \
		ptr = (entry & IND_INDIRECTION)? \
			phys_to_virt((entry & PAGE_MASK)): ptr +1)

static void kimage_free_entry(kimage_entry_t entry)
{
	struct page *page;

	page = pfn_to_page(entry >> PAGE_SHIFT);
	kimage_free_pages(page);
}

static void kimage_free(struct kimage *image)
{
	kimage_entry_t *ptr, entry;
	kimage_entry_t ind = 0;

	if (!image)
		return;
M
Maneesh Soni 已提交
613

614 615 616 617
	kimage_free_extra_pages(image);
	for_each_kimage_entry(image, ptr, entry) {
		if (entry & IND_INDIRECTION) {
			/* Free the previous indirection page */
M
Maneesh Soni 已提交
618
			if (ind & IND_INDIRECTION)
619 620 621 622 623 624
				kimage_free_entry(ind);
			/* Save this indirection page until we are
			 * done with it.
			 */
			ind = entry;
		}
M
Maneesh Soni 已提交
625
		else if (entry & IND_SOURCE)
626 627 628
			kimage_free_entry(entry);
	}
	/* Free the final indirection page */
M
Maneesh Soni 已提交
629
	if (ind & IND_INDIRECTION)
630 631 632 633 634 635 636 637 638 639
		kimage_free_entry(ind);

	/* Handle any machine specific cleanup */
	machine_kexec_cleanup(image);

	/* Free the kexec control pages... */
	kimage_free_page_list(&image->control_pages);
	kfree(image);
}

M
Maneesh Soni 已提交
640 641
static kimage_entry_t *kimage_dst_used(struct kimage *image,
					unsigned long page)
642 643 644 645 646
{
	kimage_entry_t *ptr, entry;
	unsigned long destination = 0;

	for_each_kimage_entry(image, ptr, entry) {
M
Maneesh Soni 已提交
647
		if (entry & IND_DESTINATION)
648 649
			destination = entry & PAGE_MASK;
		else if (entry & IND_SOURCE) {
M
Maneesh Soni 已提交
650
			if (page == destination)
651 652 653 654
				return ptr;
			destination += PAGE_SIZE;
		}
	}
M
Maneesh Soni 已提交
655

656
	return NULL;
657 658
}

M
Maneesh Soni 已提交
659
static struct page *kimage_alloc_page(struct kimage *image,
A
Al Viro 已提交
660
					gfp_t gfp_mask,
M
Maneesh Soni 已提交
661
					unsigned long destination)
662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700
{
	/*
	 * Here we implement safeguards to ensure that a source page
	 * is not copied to its destination page before the data on
	 * the destination page is no longer useful.
	 *
	 * To do this we maintain the invariant that a source page is
	 * either its own destination page, or it is not a
	 * destination page at all.
	 *
	 * That is slightly stronger than required, but the proof
	 * that no problems will not occur is trivial, and the
	 * implementation is simply to verify.
	 *
	 * When allocating all pages normally this algorithm will run
	 * in O(N) time, but in the worst case it will run in O(N^2)
	 * time.   If the runtime is a problem the data structures can
	 * be fixed.
	 */
	struct page *page;
	unsigned long addr;

	/*
	 * Walk through the list of destination pages, and see if I
	 * have a match.
	 */
	list_for_each_entry(page, &image->dest_pages, lru) {
		addr = page_to_pfn(page) << PAGE_SHIFT;
		if (addr == destination) {
			list_del(&page->lru);
			return page;
		}
	}
	page = NULL;
	while (1) {
		kimage_entry_t *old;

		/* Allocate a page, if we run out of memory give up */
		page = kimage_alloc_pages(gfp_mask, 0);
M
Maneesh Soni 已提交
701
		if (!page)
702
			return NULL;
703
		/* If the page cannot be used file it away */
M
Maneesh Soni 已提交
704 705
		if (page_to_pfn(page) >
				(KEXEC_SOURCE_MEMORY_LIMIT >> PAGE_SHIFT)) {
706 707 708 709 710 711 712 713 714 715
			list_add(&page->lru, &image->unuseable_pages);
			continue;
		}
		addr = page_to_pfn(page) << PAGE_SHIFT;

		/* If it is the destination page we want use it */
		if (addr == destination)
			break;

		/* If the page is not a destination page use it */
M
Maneesh Soni 已提交
716 717
		if (!kimage_is_destination_range(image, addr,
						  addr + PAGE_SIZE))
718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749
			break;

		/*
		 * I know that the page is someones destination page.
		 * See if there is already a source page for this
		 * destination page.  And if so swap the source pages.
		 */
		old = kimage_dst_used(image, addr);
		if (old) {
			/* If so move it */
			unsigned long old_addr;
			struct page *old_page;

			old_addr = *old & PAGE_MASK;
			old_page = pfn_to_page(old_addr >> PAGE_SHIFT);
			copy_highpage(page, old_page);
			*old = addr | (*old & ~PAGE_MASK);

			/* The old page I have found cannot be a
			 * destination page, so return it.
			 */
			addr = old_addr;
			page = old_page;
			break;
		}
		else {
			/* Place the page on the destination list I
			 * will use it later.
			 */
			list_add(&page->lru, &image->dest_pages);
		}
	}
M
Maneesh Soni 已提交
750

751 752 753 754
	return page;
}

static int kimage_load_normal_segment(struct kimage *image,
M
Maneesh Soni 已提交
755
					 struct kexec_segment *segment)
756 757 758 759
{
	unsigned long maddr;
	unsigned long ubytes, mbytes;
	int result;
760
	unsigned char __user *buf;
761 762 763 764 765 766 767 768

	result = 0;
	buf = segment->buf;
	ubytes = segment->bufsz;
	mbytes = segment->memsz;
	maddr = segment->mem;

	result = kimage_set_destination(image, maddr);
M
Maneesh Soni 已提交
769
	if (result < 0)
770
		goto out;
M
Maneesh Soni 已提交
771 772

	while (mbytes) {
773 774 775
		struct page *page;
		char *ptr;
		size_t uchunk, mchunk;
M
Maneesh Soni 已提交
776

777 778 779 780 781
		page = kimage_alloc_page(image, GFP_HIGHUSER, maddr);
		if (page == 0) {
			result  = -ENOMEM;
			goto out;
		}
M
Maneesh Soni 已提交
782 783 784
		result = kimage_add_page(image, page_to_pfn(page)
								<< PAGE_SHIFT);
		if (result < 0)
785
			goto out;
M
Maneesh Soni 已提交
786

787 788 789 790 791
		ptr = kmap(page);
		/* Start with a clear page */
		memset(ptr, 0, PAGE_SIZE);
		ptr += maddr & ~PAGE_MASK;
		mchunk = PAGE_SIZE - (maddr & ~PAGE_MASK);
M
Maneesh Soni 已提交
792
		if (mchunk > mbytes)
793
			mchunk = mbytes;
M
Maneesh Soni 已提交
794

795
		uchunk = mchunk;
M
Maneesh Soni 已提交
796
		if (uchunk > ubytes)
797
			uchunk = ubytes;
M
Maneesh Soni 已提交
798

799 800 801 802 803 804 805 806 807 808 809
		result = copy_from_user(ptr, buf, uchunk);
		kunmap(page);
		if (result) {
			result = (result < 0) ? result : -EIO;
			goto out;
		}
		ubytes -= uchunk;
		maddr  += mchunk;
		buf    += mchunk;
		mbytes -= mchunk;
	}
M
Maneesh Soni 已提交
810
out:
811 812 813 814
	return result;
}

static int kimage_load_crash_segment(struct kimage *image,
M
Maneesh Soni 已提交
815
					struct kexec_segment *segment)
816 817 818 819 820 821 822 823
{
	/* For crash dumps kernels we simply copy the data from
	 * user space to it's destination.
	 * We do things a page at a time for the sake of kmap.
	 */
	unsigned long maddr;
	unsigned long ubytes, mbytes;
	int result;
824
	unsigned char __user *buf;
825 826 827 828 829 830

	result = 0;
	buf = segment->buf;
	ubytes = segment->bufsz;
	mbytes = segment->memsz;
	maddr = segment->mem;
M
Maneesh Soni 已提交
831
	while (mbytes) {
832 833 834
		struct page *page;
		char *ptr;
		size_t uchunk, mchunk;
M
Maneesh Soni 已提交
835

836 837 838 839 840 841 842 843
		page = pfn_to_page(maddr >> PAGE_SHIFT);
		if (page == 0) {
			result  = -ENOMEM;
			goto out;
		}
		ptr = kmap(page);
		ptr += maddr & ~PAGE_MASK;
		mchunk = PAGE_SIZE - (maddr & ~PAGE_MASK);
M
Maneesh Soni 已提交
844
		if (mchunk > mbytes)
845
			mchunk = mbytes;
M
Maneesh Soni 已提交
846

847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863
		uchunk = mchunk;
		if (uchunk > ubytes) {
			uchunk = ubytes;
			/* Zero the trailing part of the page */
			memset(ptr + uchunk, 0, mchunk - uchunk);
		}
		result = copy_from_user(ptr, buf, uchunk);
		kunmap(page);
		if (result) {
			result = (result < 0) ? result : -EIO;
			goto out;
		}
		ubytes -= uchunk;
		maddr  += mchunk;
		buf    += mchunk;
		mbytes -= mchunk;
	}
M
Maneesh Soni 已提交
864
out:
865 866 867 868
	return result;
}

static int kimage_load_segment(struct kimage *image,
M
Maneesh Soni 已提交
869
				struct kexec_segment *segment)
870 871
{
	int result = -ENOMEM;
M
Maneesh Soni 已提交
872 873

	switch (image->type) {
874 875 876 877 878 879 880
	case KEXEC_TYPE_DEFAULT:
		result = kimage_load_normal_segment(image, segment);
		break;
	case KEXEC_TYPE_CRASH:
		result = kimage_load_crash_segment(image, segment);
		break;
	}
M
Maneesh Soni 已提交
881

882 883 884 885 886 887 888 889 890 891 892 893 894 895 896 897 898 899 900 901 902 903 904
	return result;
}

/*
 * Exec Kernel system call: for obvious reasons only root may call it.
 *
 * This call breaks up into three pieces.
 * - A generic part which loads the new kernel from the current
 *   address space, and very carefully places the data in the
 *   allocated pages.
 *
 * - A generic part that interacts with the kernel and tells all of
 *   the devices to shut down.  Preventing on-going dmas, and placing
 *   the devices in a consistent state so a later kernel can
 *   reinitialize them.
 *
 * - A machine specific part that includes the syscall number
 *   and the copies the image to it's final destination.  And
 *   jumps into the image at entry.
 *
 * kexec does not sync, or unmount filesystems so if you need
 * that to happen you need to do that yourself.
 */
905 906
struct kimage *kexec_image;
struct kimage *kexec_crash_image;
907 908 909 910 911
/*
 * A home grown binary mutex.
 * Nothing can wait so this mutex is safe to use
 * in interrupt context :)
 */
912
static int kexec_lock;
913

M
Maneesh Soni 已提交
914 915 916
asmlinkage long sys_kexec_load(unsigned long entry, unsigned long nr_segments,
				struct kexec_segment __user *segments,
				unsigned long flags)
917 918 919 920 921 922 923 924 925 926 927 928 929 930 931 932 933 934 935 936 937 938 939 940 941 942 943 944 945 946 947 948 949 950 951 952 953 954 955
{
	struct kimage **dest_image, *image;
	int locked;
	int result;

	/* We only trust the superuser with rebooting the system. */
	if (!capable(CAP_SYS_BOOT))
		return -EPERM;

	/*
	 * Verify we have a legal set of flags
	 * This leaves us room for future extensions.
	 */
	if ((flags & KEXEC_FLAGS) != (flags & ~KEXEC_ARCH_MASK))
		return -EINVAL;

	/* Verify we are on the appropriate architecture */
	if (((flags & KEXEC_ARCH_MASK) != KEXEC_ARCH) &&
		((flags & KEXEC_ARCH_MASK) != KEXEC_ARCH_DEFAULT))
		return -EINVAL;

	/* Put an artificial cap on the number
	 * of segments passed to kexec_load.
	 */
	if (nr_segments > KEXEC_SEGMENT_MAX)
		return -EINVAL;

	image = NULL;
	result = 0;

	/* Because we write directly to the reserved memory
	 * region when loading crash kernels we need a mutex here to
	 * prevent multiple crash  kernels from attempting to load
	 * simultaneously, and to prevent a crash kernel from loading
	 * over the top of a in use crash kernel.
	 *
	 * KISS: always take the mutex.
	 */
	locked = xchg(&kexec_lock, 1);
M
Maneesh Soni 已提交
956
	if (locked)
957
		return -EBUSY;
M
Maneesh Soni 已提交
958

959
	dest_image = &kexec_image;
M
Maneesh Soni 已提交
960
	if (flags & KEXEC_ON_CRASH)
961 962 963
		dest_image = &kexec_crash_image;
	if (nr_segments > 0) {
		unsigned long i;
M
Maneesh Soni 已提交
964

965
		/* Loading another kernel to reboot into */
M
Maneesh Soni 已提交
966 967 968
		if ((flags & KEXEC_ON_CRASH) == 0)
			result = kimage_normal_alloc(&image, entry,
							nr_segments, segments);
969 970 971 972 973 974
		/* Loading another kernel to switch to if this one crashes */
		else if (flags & KEXEC_ON_CRASH) {
			/* Free any current crash dump kernel before
			 * we corrupt it.
			 */
			kimage_free(xchg(&kexec_crash_image, NULL));
M
Maneesh Soni 已提交
975 976
			result = kimage_crash_alloc(&image, entry,
						     nr_segments, segments);
977
		}
M
Maneesh Soni 已提交
978
		if (result)
979
			goto out;
M
Maneesh Soni 已提交
980

981
		result = machine_kexec_prepare(image);
M
Maneesh Soni 已提交
982
		if (result)
983
			goto out;
M
Maneesh Soni 已提交
984 985

		for (i = 0; i < nr_segments; i++) {
986
			result = kimage_load_segment(image, &image->segment[i]);
M
Maneesh Soni 已提交
987
			if (result)
988 989 990
				goto out;
		}
		result = kimage_terminate(image);
M
Maneesh Soni 已提交
991
		if (result)
992 993 994 995 996
			goto out;
	}
	/* Install the new kernel, and  Uninstall the old */
	image = xchg(dest_image, image);

M
Maneesh Soni 已提交
997
out:
998 999
	xchg(&kexec_lock, 0); /* Release the mutex */
	kimage_free(image);
M
Maneesh Soni 已提交
1000

1001 1002 1003 1004 1005
	return result;
}

#ifdef CONFIG_COMPAT
asmlinkage long compat_sys_kexec_load(unsigned long entry,
M
Maneesh Soni 已提交
1006 1007 1008
				unsigned long nr_segments,
				struct compat_kexec_segment __user *segments,
				unsigned long flags)
1009 1010 1011 1012 1013 1014 1015 1016
{
	struct compat_kexec_segment in;
	struct kexec_segment out, __user *ksegments;
	unsigned long i, result;

	/* Don't allow clients that don't understand the native
	 * architecture to do anything.
	 */
M
Maneesh Soni 已提交
1017
	if ((flags & KEXEC_ARCH_MASK) == KEXEC_ARCH_DEFAULT)
1018 1019
		return -EINVAL;

M
Maneesh Soni 已提交
1020
	if (nr_segments > KEXEC_SEGMENT_MAX)
1021 1022 1023 1024 1025
		return -EINVAL;

	ksegments = compat_alloc_user_space(nr_segments * sizeof(out));
	for (i=0; i < nr_segments; i++) {
		result = copy_from_user(&in, &segments[i], sizeof(in));
M
Maneesh Soni 已提交
1026
		if (result)
1027 1028 1029 1030 1031 1032 1033 1034
			return -EFAULT;

		out.buf   = compat_ptr(in.buf);
		out.bufsz = in.bufsz;
		out.mem   = in.mem;
		out.memsz = in.memsz;

		result = copy_to_user(&ksegments[i], &out, sizeof(out));
M
Maneesh Soni 已提交
1035
		if (result)
1036 1037 1038 1039 1040 1041 1042
			return -EFAULT;
	}

	return sys_kexec_load(entry, nr_segments, ksegments, flags);
}
#endif

1043
void crash_kexec(struct pt_regs *regs)
1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056 1057
{
	int locked;


	/* Take the kexec_lock here to prevent sys_kexec_load
	 * running on one cpu from replacing the crash kernel
	 * we are using after a panic on a different cpu.
	 *
	 * If the crash kernel was not located in a fixed area
	 * of memory the xchg(&kexec_crash_image) would be
	 * sufficient.  But since I reuse the memory...
	 */
	locked = xchg(&kexec_lock, 1);
	if (!locked) {
1058
		if (kexec_crash_image) {
1059 1060 1061
			struct pt_regs fixed_regs;
			crash_setup_regs(&fixed_regs, regs);
			machine_crash_shutdown(&fixed_regs);
1062
			machine_kexec(kexec_crash_image);
1063 1064 1065 1066
		}
		xchg(&kexec_lock, 0);
	}
}
1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079

static int __init crash_notes_memory_init(void)
{
	/* Allocate memory for saving cpu registers. */
	crash_notes = alloc_percpu(note_buf_t);
	if (!crash_notes) {
		printk("Kexec: Memory allocation for saving cpu register"
		" states failed\n");
		return -ENOMEM;
	}
	return 0;
}
module_init(crash_notes_memory_init)