module.c 92.1 KB
Newer Older
1
/*
L
Linus Torvalds 已提交
2
   Copyright (C) 2002 Richard Henderson
3
   Copyright (C) 2001 Rusty Russell, 2002, 2010 Rusty Russell IBM.
L
Linus Torvalds 已提交
4 5 6 7 8 9 10 11 12 13 14 15 16 17 18

    This program is free software; you can redistribute it and/or modify
    it under the terms of the GNU General Public License as published by
    the Free Software Foundation; either version 2 of the License, or
    (at your option) any later version.

    This program is distributed in the hope that it will be useful,
    but WITHOUT ANY WARRANTY; without even the implied warranty of
    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
    GNU General Public License for more details.

    You should have received a copy of the GNU General Public License
    along with this program; if not, write to the Free Software
    Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
*/
19
#include <linux/export.h>
L
Linus Torvalds 已提交
20
#include <linux/moduleloader.h>
21
#include <linux/ftrace_event.h>
L
Linus Torvalds 已提交
22
#include <linux/init.h>
23
#include <linux/kallsyms.h>
24
#include <linux/fs.h>
25
#include <linux/sysfs.h>
26
#include <linux/kernel.h>
L
Linus Torvalds 已提交
27 28 29
#include <linux/slab.h>
#include <linux/vmalloc.h>
#include <linux/elf.h>
30
#include <linux/proc_fs.h>
L
Linus Torvalds 已提交
31 32 33 34
#include <linux/seq_file.h>
#include <linux/syscalls.h>
#include <linux/fcntl.h>
#include <linux/rcupdate.h>
35
#include <linux/capability.h>
L
Linus Torvalds 已提交
36 37 38 39 40 41
#include <linux/cpu.h>
#include <linux/moduleparam.h>
#include <linux/errno.h>
#include <linux/err.h>
#include <linux/vermagic.h>
#include <linux/notifier.h>
42
#include <linux/sched.h>
L
Linus Torvalds 已提交
43 44
#include <linux/stop_machine.h>
#include <linux/device.h>
45
#include <linux/string.h>
46
#include <linux/mutex.h>
47
#include <linux/rculist.h>
L
Linus Torvalds 已提交
48 49
#include <asm/uaccess.h>
#include <asm/cacheflush.h>
50
#include <asm/mmu_context.h>
51
#include <linux/license.h>
52
#include <asm/sections.h>
53
#include <linux/tracepoint.h>
54
#include <linux/ftrace.h>
55
#include <linux/async.h>
56
#include <linux/percpu.h>
57
#include <linux/kmemleak.h>
58
#include <linux/jump_label.h>
59
#include <linux/pfn.h>
60
#include <linux/bsearch.h>
D
David Howells 已提交
61
#include <linux/fips.h>
62
#include "module-internal.h"
L
Linus Torvalds 已提交
63

64 65 66
#define CREATE_TRACE_POINTS
#include <trace/events/module.h>

L
Linus Torvalds 已提交
67 68 69 70
#ifndef ARCH_SHF_SMALL
#define ARCH_SHF_SMALL 0
#endif

71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90
/*
 * Modules' sections will be aligned on page boundaries
 * to ensure complete separation of code and data, but
 * only when CONFIG_DEBUG_SET_MODULE_RONX=y
 */
#ifdef CONFIG_DEBUG_SET_MODULE_RONX
# define debug_align(X) ALIGN(X, PAGE_SIZE)
#else
# define debug_align(X) (X)
#endif

/*
 * Given BASE and SIZE this macro calculates the number of pages the
 * memory regions occupies
 */
#define MOD_NUMBER_OF_PAGES(BASE, SIZE) (((SIZE) > 0) ?		\
		(PFN_DOWN((unsigned long)(BASE) + (SIZE) - 1) -	\
			 PFN_DOWN((unsigned long)BASE) + 1)	\
		: (0UL))

L
Linus Torvalds 已提交
91 92 93
/* If this is set, the section belongs in the init part of the module */
#define INIT_OFFSET_MASK (1UL << (BITS_PER_LONG-1))

94 95 96 97 98
/*
 * Mutex protects:
 * 1) List of modules (also safely readable with preempt_disable),
 * 2) module_use links,
 * 3) module_addr_min/module_addr_max.
99
 * (delete uses stop_machine/add uses RCU list operations). */
100 101
DEFINE_MUTEX(module_mutex);
EXPORT_SYMBOL_GPL(module_mutex);
L
Linus Torvalds 已提交
102
static LIST_HEAD(modules);
103 104 105 106
#ifdef CONFIG_KGDB_KDB
struct list_head *kdb_modules = &modules; /* kdb needs the list of modules */
#endif /* CONFIG_KGDB_KDB */

107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143
#ifdef CONFIG_MODULE_SIG
#ifdef CONFIG_MODULE_SIG_FORCE
static bool sig_enforce = true;
#else
static bool sig_enforce = false;

static int param_set_bool_enable_only(const char *val,
				      const struct kernel_param *kp)
{
	int err;
	bool test;
	struct kernel_param dummy_kp = *kp;

	dummy_kp.arg = &test;

	err = param_set_bool(val, &dummy_kp);
	if (err)
		return err;

	/* Don't let them unset it once it's set! */
	if (!test && sig_enforce)
		return -EROFS;

	if (test)
		sig_enforce = true;
	return 0;
}

static const struct kernel_param_ops param_ops_bool_enable_only = {
	.set = param_set_bool_enable_only,
	.get = param_get_bool,
};
#define param_check_bool_enable_only param_check_bool

module_param(sig_enforce, bool_enable_only, 0644);
#endif /* !CONFIG_MODULE_SIG_FORCE */
#endif /* CONFIG_MODULE_SIG */
L
Linus Torvalds 已提交
144

145 146
/* Block module loading/unloading? */
int modules_disabled = 0;
147
core_param(nomodule, modules_disabled, bint, 0);
148

149 150 151
/* Waiting for a module to finish initializing? */
static DECLARE_WAIT_QUEUE_HEAD(module_wq);

152
static BLOCKING_NOTIFIER_HEAD(module_notify_list);
L
Linus Torvalds 已提交
153

154 155
/* Bounds of module allocation, for speeding __module_address.
 * Protected by module_mutex. */
156 157
static unsigned long module_addr_min = -1UL, module_addr_max = 0;

L
Linus Torvalds 已提交
158 159
int register_module_notifier(struct notifier_block * nb)
{
160
	return blocking_notifier_chain_register(&module_notify_list, nb);
L
Linus Torvalds 已提交
161 162 163 164 165
}
EXPORT_SYMBOL(register_module_notifier);

int unregister_module_notifier(struct notifier_block * nb)
{
166
	return blocking_notifier_chain_unregister(&module_notify_list, nb);
L
Linus Torvalds 已提交
167 168 169
}
EXPORT_SYMBOL(unregister_module_notifier);

170 171 172 173
struct load_info {
	Elf_Ehdr *hdr;
	unsigned long len;
	Elf_Shdr *sechdrs;
174
	char *secstrings, *strtab;
175
	unsigned long symoffs, stroffs;
176 177
	struct _ddebug *debug;
	unsigned int num_debug;
178
	bool sig_ok;
179 180 181 182 183
	struct {
		unsigned int sym, str, mod, vers, info, pcpu;
	} index;
};

184 185
/* We require a truly strong try_module_get(): 0 means failure due to
   ongoing or failed initialization etc. */
L
Linus Torvalds 已提交
186 187 188
static inline int strong_try_module_get(struct module *mod)
{
	if (mod && mod->state == MODULE_STATE_COMING)
189 190
		return -EBUSY;
	if (try_module_get(mod))
L
Linus Torvalds 已提交
191
		return 0;
192 193
	else
		return -ENOENT;
L
Linus Torvalds 已提交
194 195
}

196 197 198
static inline void add_taint_module(struct module *mod, unsigned flag)
{
	add_taint(flag);
199
	mod->taints |= (1U << flag);
200 201
}

202 203 204
/*
 * A thread that wants to hold a reference to a module only while it
 * is running can call this to safely exit.  nfsd and lockd use this.
L
Linus Torvalds 已提交
205 206 207 208 209 210 211
 */
void __module_put_and_exit(struct module *mod, long code)
{
	module_put(mod);
	do_exit(code);
}
EXPORT_SYMBOL(__module_put_and_exit);
212

L
Linus Torvalds 已提交
213
/* Find a module section: 0 means not found. */
214
static unsigned int find_sec(const struct load_info *info, const char *name)
L
Linus Torvalds 已提交
215 216 217
{
	unsigned int i;

218 219
	for (i = 1; i < info->hdr->e_shnum; i++) {
		Elf_Shdr *shdr = &info->sechdrs[i];
L
Linus Torvalds 已提交
220
		/* Alloc bit cleared means "ignore it." */
221 222
		if ((shdr->sh_flags & SHF_ALLOC)
		    && strcmp(info->secstrings + shdr->sh_name, name) == 0)
L
Linus Torvalds 已提交
223
			return i;
224
	}
L
Linus Torvalds 已提交
225 226 227
	return 0;
}

228
/* Find a module section, or NULL. */
229
static void *section_addr(const struct load_info *info, const char *name)
230 231
{
	/* Section 0 has sh_addr 0. */
232
	return (void *)info->sechdrs[find_sec(info, name)].sh_addr;
233 234 235
}

/* Find a module section, or NULL.  Fill in number of "objects" in section. */
236
static void *section_objs(const struct load_info *info,
237 238 239 240
			  const char *name,
			  size_t object_size,
			  unsigned int *num)
{
241
	unsigned int sec = find_sec(info, name);
242 243

	/* Section 0 has sh_addr 0 and sh_size 0. */
244 245
	*num = info->sechdrs[sec].sh_size / object_size;
	return (void *)info->sechdrs[sec].sh_addr;
246 247
}

L
Linus Torvalds 已提交
248 249 250 251 252
/* Provided by the linker */
extern const struct kernel_symbol __start___ksymtab[];
extern const struct kernel_symbol __stop___ksymtab[];
extern const struct kernel_symbol __start___ksymtab_gpl[];
extern const struct kernel_symbol __stop___ksymtab_gpl[];
253 254
extern const struct kernel_symbol __start___ksymtab_gpl_future[];
extern const struct kernel_symbol __stop___ksymtab_gpl_future[];
L
Linus Torvalds 已提交
255 256
extern const unsigned long __start___kcrctab[];
extern const unsigned long __start___kcrctab_gpl[];
257
extern const unsigned long __start___kcrctab_gpl_future[];
258 259 260 261 262
#ifdef CONFIG_UNUSED_SYMBOLS
extern const struct kernel_symbol __start___ksymtab_unused[];
extern const struct kernel_symbol __stop___ksymtab_unused[];
extern const struct kernel_symbol __start___ksymtab_unused_gpl[];
extern const struct kernel_symbol __stop___ksymtab_unused_gpl[];
263 264
extern const unsigned long __start___kcrctab_unused[];
extern const unsigned long __start___kcrctab_unused_gpl[];
265
#endif
L
Linus Torvalds 已提交
266 267 268 269

#ifndef CONFIG_MODVERSIONS
#define symversion(base, idx) NULL
#else
270
#define symversion(base, idx) ((base != NULL) ? ((base) + (idx)) : NULL)
L
Linus Torvalds 已提交
271 272
#endif

273 274 275 276 277
static bool each_symbol_in_section(const struct symsearch *arr,
				   unsigned int arrsize,
				   struct module *owner,
				   bool (*fn)(const struct symsearch *syms,
					      struct module *owner,
278
					      void *data),
279
				   void *data)
280
{
281
	unsigned int j;
282

283
	for (j = 0; j < arrsize; j++) {
284 285
		if (fn(&arr[j], owner, data))
			return true;
286
	}
287 288

	return false;
289 290
}

291
/* Returns true as soon as fn returns true, otherwise false. */
292 293 294 295
bool each_symbol_section(bool (*fn)(const struct symsearch *arr,
				    struct module *owner,
				    void *data),
			 void *data)
296 297
{
	struct module *mod;
298
	static const struct symsearch arr[] = {
299
		{ __start___ksymtab, __stop___ksymtab, __start___kcrctab,
300
		  NOT_GPL_ONLY, false },
301
		{ __start___ksymtab_gpl, __stop___ksymtab_gpl,
302 303
		  __start___kcrctab_gpl,
		  GPL_ONLY, false },
304
		{ __start___ksymtab_gpl_future, __stop___ksymtab_gpl_future,
305 306
		  __start___kcrctab_gpl_future,
		  WILL_BE_GPL_ONLY, false },
307
#ifdef CONFIG_UNUSED_SYMBOLS
308
		{ __start___ksymtab_unused, __stop___ksymtab_unused,
309 310
		  __start___kcrctab_unused,
		  NOT_GPL_ONLY, true },
311
		{ __start___ksymtab_unused_gpl, __stop___ksymtab_unused_gpl,
312 313
		  __start___kcrctab_unused_gpl,
		  GPL_ONLY, true },
314
#endif
315
	};
316

317 318
	if (each_symbol_in_section(arr, ARRAY_SIZE(arr), NULL, fn, data))
		return true;
319

320
	list_for_each_entry_rcu(mod, &modules, list) {
321 322
		struct symsearch arr[] = {
			{ mod->syms, mod->syms + mod->num_syms, mod->crcs,
323
			  NOT_GPL_ONLY, false },
324
			{ mod->gpl_syms, mod->gpl_syms + mod->num_gpl_syms,
325 326
			  mod->gpl_crcs,
			  GPL_ONLY, false },
327 328
			{ mod->gpl_future_syms,
			  mod->gpl_future_syms + mod->num_gpl_future_syms,
329 330
			  mod->gpl_future_crcs,
			  WILL_BE_GPL_ONLY, false },
331
#ifdef CONFIG_UNUSED_SYMBOLS
332 333
			{ mod->unused_syms,
			  mod->unused_syms + mod->num_unused_syms,
334 335
			  mod->unused_crcs,
			  NOT_GPL_ONLY, true },
336 337
			{ mod->unused_gpl_syms,
			  mod->unused_gpl_syms + mod->num_unused_gpl_syms,
338 339
			  mod->unused_gpl_crcs,
			  GPL_ONLY, true },
340
#endif
341 342
		};

343 344 345 346 347
		if (each_symbol_in_section(arr, ARRAY_SIZE(arr), mod, fn, data))
			return true;
	}
	return false;
}
348
EXPORT_SYMBOL_GPL(each_symbol_section);
349 350 351 352 353 354 355 356 357 358

struct find_symbol_arg {
	/* Input */
	const char *name;
	bool gplok;
	bool warn;

	/* Output */
	struct module *owner;
	const unsigned long *crc;
359
	const struct kernel_symbol *sym;
360 361
};

362 363 364
static bool check_symbol(const struct symsearch *syms,
				 struct module *owner,
				 unsigned int symnum, void *data)
365 366 367 368 369 370 371 372 373 374 375 376 377
{
	struct find_symbol_arg *fsa = data;

	if (!fsa->gplok) {
		if (syms->licence == GPL_ONLY)
			return false;
		if (syms->licence == WILL_BE_GPL_ONLY && fsa->warn) {
			printk(KERN_WARNING "Symbol %s is being used "
			       "by a non-GPL module, which will not "
			       "be allowed in the future\n", fsa->name);
			printk(KERN_WARNING "Please see the file "
			       "Documentation/feature-removal-schedule.txt "
			       "in the kernel source tree for more details.\n");
378
		}
L
Linus Torvalds 已提交
379
	}
380

381
#ifdef CONFIG_UNUSED_SYMBOLS
382 383 384 385 386 387 388 389 390 391 392
	if (syms->unused && fsa->warn) {
		printk(KERN_WARNING "Symbol %s is marked as UNUSED, "
		       "however this module is using it.\n", fsa->name);
		printk(KERN_WARNING
		       "This symbol will go away in the future.\n");
		printk(KERN_WARNING
		       "Please evalute if this is the right api to use and if "
		       "it really is, submit a report the linux kernel "
		       "mailinglist together with submitting your code for "
		       "inclusion.\n");
	}
393
#endif
394 395 396

	fsa->owner = owner;
	fsa->crc = symversion(syms->crcs, symnum);
397
	fsa->sym = &syms->start[symnum];
398 399 400
	return true;
}

401 402 403 404 405 406 407 408
static int cmp_name(const void *va, const void *vb)
{
	const char *a;
	const struct kernel_symbol *b;
	a = va; b = vb;
	return strcmp(a, b->name);
}

409 410 411 412 413
static bool find_symbol_in_section(const struct symsearch *syms,
				   struct module *owner,
				   void *data)
{
	struct find_symbol_arg *fsa = data;
414 415 416 417 418 419 420
	struct kernel_symbol *sym;

	sym = bsearch(fsa->name, syms->start, syms->stop - syms->start,
			sizeof(struct kernel_symbol), cmp_name);

	if (sym != NULL && check_symbol(syms, owner, sym - syms->start, data))
		return true;
421 422 423 424

	return false;
}

425
/* Find a symbol and return it, along with, (optional) crc and
426
 * (optional) module which owns it.  Needs preempt disabled or module_mutex. */
427 428 429 430 431
const struct kernel_symbol *find_symbol(const char *name,
					struct module **owner,
					const unsigned long **crc,
					bool gplok,
					bool warn)
432 433 434 435 436 437 438
{
	struct find_symbol_arg fsa;

	fsa.name = name;
	fsa.gplok = gplok;
	fsa.warn = warn;

439
	if (each_symbol_section(find_symbol_in_section, &fsa)) {
440 441 442 443
		if (owner)
			*owner = fsa.owner;
		if (crc)
			*crc = fsa.crc;
444
		return fsa.sym;
445 446
	}

447
	pr_debug("Failed to find symbol %s\n", name);
448
	return NULL;
L
Linus Torvalds 已提交
449
}
450
EXPORT_SYMBOL_GPL(find_symbol);
L
Linus Torvalds 已提交
451 452

/* Search for module by name: must hold module_mutex. */
453
struct module *find_module(const char *name)
L
Linus Torvalds 已提交
454 455 456 457 458 459 460 461 462
{
	struct module *mod;

	list_for_each_entry(mod, &modules, list) {
		if (strcmp(mod->name, name) == 0)
			return mod;
	}
	return NULL;
}
463
EXPORT_SYMBOL_GPL(find_module);
L
Linus Torvalds 已提交
464 465

#ifdef CONFIG_SMP
466

467
static inline void __percpu *mod_percpu(struct module *mod)
468
{
469 470
	return mod->percpu;
}
471

472 473 474
static int percpu_modalloc(struct module *mod,
			   unsigned long size, unsigned long align)
{
475 476
	if (align > PAGE_SIZE) {
		printk(KERN_WARNING "%s: per-cpu alignment %li > %li\n",
477
		       mod->name, align, PAGE_SIZE);
478 479 480
		align = PAGE_SIZE;
	}

481 482
	mod->percpu = __alloc_reserved_percpu(size, align);
	if (!mod->percpu) {
483
		printk(KERN_WARNING
484 485
		       "%s: Could not allocate %lu bytes percpu data\n",
		       mod->name, size);
486 487 488 489
		return -ENOMEM;
	}
	mod->percpu_size = size;
	return 0;
490 491
}

492
static void percpu_modfree(struct module *mod)
493
{
494
	free_percpu(mod->percpu);
495 496
}

497
static unsigned int find_pcpusec(struct load_info *info)
498
{
499
	return find_sec(info, ".data..percpu");
500 501
}

502 503
static void percpu_modcopy(struct module *mod,
			   const void *from, unsigned long size)
504 505 506 507
{
	int cpu;

	for_each_possible_cpu(cpu)
508
		memcpy(per_cpu_ptr(mod->percpu, cpu), from, size);
509 510
}

511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542
/**
 * is_module_percpu_address - test whether address is from module static percpu
 * @addr: address to test
 *
 * Test whether @addr belongs to module static percpu area.
 *
 * RETURNS:
 * %true if @addr is from module static percpu area
 */
bool is_module_percpu_address(unsigned long addr)
{
	struct module *mod;
	unsigned int cpu;

	preempt_disable();

	list_for_each_entry_rcu(mod, &modules, list) {
		if (!mod->percpu_size)
			continue;
		for_each_possible_cpu(cpu) {
			void *start = per_cpu_ptr(mod->percpu, cpu);

			if ((void *)addr >= start &&
			    (void *)addr < start + mod->percpu_size) {
				preempt_enable();
				return true;
			}
		}
	}

	preempt_enable();
	return false;
543 544
}

L
Linus Torvalds 已提交
545
#else /* ... !CONFIG_SMP */
546

547
static inline void __percpu *mod_percpu(struct module *mod)
L
Linus Torvalds 已提交
548 549 550
{
	return NULL;
}
551 552 553 554 555 556
static inline int percpu_modalloc(struct module *mod,
				  unsigned long size, unsigned long align)
{
	return -ENOMEM;
}
static inline void percpu_modfree(struct module *mod)
L
Linus Torvalds 已提交
557 558
{
}
559
static unsigned int find_pcpusec(struct load_info *info)
L
Linus Torvalds 已提交
560 561 562
{
	return 0;
}
563 564
static inline void percpu_modcopy(struct module *mod,
				  const void *from, unsigned long size)
L
Linus Torvalds 已提交
565 566 567 568
{
	/* pcpusec should be 0, and size of that section should be 0. */
	BUG_ON(size != 0);
}
569 570 571 572
bool is_module_percpu_address(unsigned long addr)
{
	return false;
}
573

L
Linus Torvalds 已提交
574 575
#endif /* CONFIG_SMP */

576 577 578 579 580 581
#define MODINFO_ATTR(field)	\
static void setup_modinfo_##field(struct module *mod, const char *s)  \
{                                                                     \
	mod->field = kstrdup(s, GFP_KERNEL);                          \
}                                                                     \
static ssize_t show_modinfo_##field(struct module_attribute *mattr,   \
582
			struct module_kobject *mk, char *buffer)      \
583
{                                                                     \
584
	return sprintf(buffer, "%s\n", mk->mod->field);               \
585 586 587 588 589 590 591
}                                                                     \
static int modinfo_##field##_exists(struct module *mod)               \
{                                                                     \
	return mod->field != NULL;                                    \
}                                                                     \
static void free_modinfo_##field(struct module *mod)                  \
{                                                                     \
592 593
	kfree(mod->field);                                            \
	mod->field = NULL;                                            \
594 595
}                                                                     \
static struct module_attribute modinfo_##field = {                    \
596
	.attr = { .name = __stringify(field), .mode = 0444 },         \
597 598 599 600 601 602 603 604 605
	.show = show_modinfo_##field,                                 \
	.setup = setup_modinfo_##field,                               \
	.test = modinfo_##field##_exists,                             \
	.free = free_modinfo_##field,                                 \
};

MODINFO_ATTR(version);
MODINFO_ATTR(srcversion);

606 607
static char last_unloaded_module[MODULE_NAME_LEN+1];

608
#ifdef CONFIG_MODULE_UNLOAD
609 610 611

EXPORT_TRACEPOINT_SYMBOL(module_get);

L
Linus Torvalds 已提交
612
/* Init the unload section of the module. */
613
static int module_unload_init(struct module *mod)
L
Linus Torvalds 已提交
614
{
615 616 617 618
	mod->refptr = alloc_percpu(struct module_ref);
	if (!mod->refptr)
		return -ENOMEM;

619 620
	INIT_LIST_HEAD(&mod->source_list);
	INIT_LIST_HEAD(&mod->target_list);
621

L
Linus Torvalds 已提交
622
	/* Hold reference count during initialization. */
623
	__this_cpu_write(mod->refptr->incs, 1);
L
Linus Torvalds 已提交
624 625
	/* Backwards compatibility macros put refcount during init. */
	mod->waiter = current;
626 627

	return 0;
L
Linus Torvalds 已提交
628 629 630 631 632 633 634
}

/* Does a already use b? */
static int already_uses(struct module *a, struct module *b)
{
	struct module_use *use;

635 636
	list_for_each_entry(use, &b->source_list, source_list) {
		if (use->source == a) {
637
			pr_debug("%s uses %s!\n", a->name, b->name);
L
Linus Torvalds 已提交
638 639 640
			return 1;
		}
	}
641
	pr_debug("%s does not use %s!\n", a->name, b->name);
L
Linus Torvalds 已提交
642 643 644
	return 0;
}

645 646 647 648 649 650 651 652 653 654 655
/*
 * Module a uses b
 *  - we add 'a' as a "source", 'b' as a "target" of module use
 *  - the module_use is added to the list of 'b' sources (so
 *    'b' can walk the list to see who sourced them), and of 'a'
 *    targets (so 'a' can see what modules it targets).
 */
static int add_module_usage(struct module *a, struct module *b)
{
	struct module_use *use;

656
	pr_debug("Allocating new usage for %s.\n", a->name);
657 658 659 660 661 662 663 664 665 666 667 668 669
	use = kmalloc(sizeof(*use), GFP_ATOMIC);
	if (!use) {
		printk(KERN_WARNING "%s: out of memory loading\n", a->name);
		return -ENOMEM;
	}

	use->source = a;
	use->target = b;
	list_add(&use->source_list, &b->source_list);
	list_add(&use->target_list, &a->target_list);
	return 0;
}

670
/* Module a uses b: caller needs module_mutex() */
671
int ref_module(struct module *a, struct module *b)
L
Linus Torvalds 已提交
672
{
673
	int err;
K
Kay Sievers 已提交
674

675
	if (b == NULL || already_uses(a, b))
676 677
		return 0;

678 679
	/* If module isn't available, we fail. */
	err = strong_try_module_get(b);
680
	if (err)
681
		return err;
L
Linus Torvalds 已提交
682

683 684
	err = add_module_usage(a, b);
	if (err) {
L
Linus Torvalds 已提交
685
		module_put(b);
686
		return err;
L
Linus Torvalds 已提交
687
	}
688
	return 0;
L
Linus Torvalds 已提交
689
}
690
EXPORT_SYMBOL_GPL(ref_module);
L
Linus Torvalds 已提交
691 692 693 694

/* Clear the unload stuff of the module. */
static void module_unload_free(struct module *mod)
{
695
	struct module_use *use, *tmp;
L
Linus Torvalds 已提交
696

697
	mutex_lock(&module_mutex);
698 699
	list_for_each_entry_safe(use, tmp, &mod->target_list, target_list) {
		struct module *i = use->target;
700
		pr_debug("%s unusing %s\n", mod->name, i->name);
701 702 703 704
		module_put(i);
		list_del(&use->source_list);
		list_del(&use->target_list);
		kfree(use);
L
Linus Torvalds 已提交
705
	}
706
	mutex_unlock(&module_mutex);
707 708

	free_percpu(mod->refptr);
L
Linus Torvalds 已提交
709 710 711
}

#ifdef CONFIG_MODULE_FORCE_UNLOAD
712
static inline int try_force_unload(unsigned int flags)
L
Linus Torvalds 已提交
713 714 715
{
	int ret = (flags & O_TRUNC);
	if (ret)
716
		add_taint(TAINT_FORCED_RMMOD);
L
Linus Torvalds 已提交
717 718 719
	return ret;
}
#else
720
static inline int try_force_unload(unsigned int flags)
L
Linus Torvalds 已提交
721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737
{
	return 0;
}
#endif /* CONFIG_MODULE_FORCE_UNLOAD */

struct stopref
{
	struct module *mod;
	int flags;
	int *forced;
};

/* Whole machine is stopped with interrupts off when this runs. */
static int __try_stop_module(void *_sref)
{
	struct stopref *sref = _sref;

738 739
	/* If it's not unused, quit unless we're forcing. */
	if (module_refcount(sref->mod) != 0) {
740
		if (!(*sref->forced = try_force_unload(sref->flags)))
L
Linus Torvalds 已提交
741 742 743 744 745 746 747 748 749 750
			return -EWOULDBLOCK;
	}

	/* Mark it as dying. */
	sref->mod->state = MODULE_STATE_GOING;
	return 0;
}

static int try_stop_module(struct module *mod, int flags, int *forced)
{
751 752
	if (flags & O_NONBLOCK) {
		struct stopref sref = { mod, flags, forced };
L
Linus Torvalds 已提交
753

754
		return stop_machine(__try_stop_module, &sref, NULL);
755 756 757 758 759 760
	} else {
		/* We don't need to stop the machine for this. */
		mod->state = MODULE_STATE_GOING;
		synchronize_sched();
		return 0;
	}
L
Linus Torvalds 已提交
761 762
}

763
unsigned long module_refcount(struct module *mod)
L
Linus Torvalds 已提交
764
{
765
	unsigned long incs = 0, decs = 0;
766
	int cpu;
L
Linus Torvalds 已提交
767

768
	for_each_possible_cpu(cpu)
769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786
		decs += per_cpu_ptr(mod->refptr, cpu)->decs;
	/*
	 * ensure the incs are added up after the decs.
	 * module_put ensures incs are visible before decs with smp_wmb.
	 *
	 * This 2-count scheme avoids the situation where the refcount
	 * for CPU0 is read, then CPU0 increments the module refcount,
	 * then CPU1 drops that refcount, then the refcount for CPU1 is
	 * read. We would record a decrement but not its corresponding
	 * increment so we would see a low count (disaster).
	 *
	 * Rare situation? But module_refcount can be preempted, and we
	 * might be tallying up 4096+ CPUs. So it is not impossible.
	 */
	smp_rmb();
	for_each_possible_cpu(cpu)
		incs += per_cpu_ptr(mod->refptr, cpu)->incs;
	return incs - decs;
L
Linus Torvalds 已提交
787 788 789 790 791 792 793 794
}
EXPORT_SYMBOL(module_refcount);

/* This exists whether we can unload or not */
static void free_module(struct module *mod);

static void wait_for_zero_refcount(struct module *mod)
{
795
	/* Since we might sleep for some time, release the mutex first */
796
	mutex_unlock(&module_mutex);
L
Linus Torvalds 已提交
797
	for (;;) {
798
		pr_debug("Looking at refcount...\n");
L
Linus Torvalds 已提交
799 800 801 802 803 804
		set_current_state(TASK_UNINTERRUPTIBLE);
		if (module_refcount(mod) == 0)
			break;
		schedule();
	}
	current->state = TASK_RUNNING;
805
	mutex_lock(&module_mutex);
L
Linus Torvalds 已提交
806 807
}

808 809
SYSCALL_DEFINE2(delete_module, const char __user *, name_user,
		unsigned int, flags)
L
Linus Torvalds 已提交
810 811
{
	struct module *mod;
812
	char name[MODULE_NAME_LEN];
L
Linus Torvalds 已提交
813 814
	int ret, forced = 0;

815
	if (!capable(CAP_SYS_MODULE) || modules_disabled)
816 817 818 819 820 821
		return -EPERM;

	if (strncpy_from_user(name, name_user, MODULE_NAME_LEN-1) < 0)
		return -EFAULT;
	name[MODULE_NAME_LEN-1] = '\0';

822 823
	if (mutex_lock_interruptible(&module_mutex) != 0)
		return -EINTR;
L
Linus Torvalds 已提交
824 825 826 827 828 829 830

	mod = find_module(name);
	if (!mod) {
		ret = -ENOENT;
		goto out;
	}

831
	if (!list_empty(&mod->source_list)) {
L
Linus Torvalds 已提交
832 833 834 835 836 837 838 839 840
		/* Other modules depend on us: get rid of them first. */
		ret = -EWOULDBLOCK;
		goto out;
	}

	/* Doing init or already dying? */
	if (mod->state != MODULE_STATE_LIVE) {
		/* FIXME: if (force), slam module count and wake up
                   waiter --RR */
841
		pr_debug("%s already dying\n", mod->name);
L
Linus Torvalds 已提交
842 843 844 845 846
		ret = -EBUSY;
		goto out;
	}

	/* If it has an init func, it must have an exit func to unload */
847
	if (mod->init && !mod->exit) {
848
		forced = try_force_unload(flags);
L
Linus Torvalds 已提交
849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864 865 866 867
		if (!forced) {
			/* This module can't be removed */
			ret = -EBUSY;
			goto out;
		}
	}

	/* Set this up before setting mod->state */
	mod->waiter = current;

	/* Stop the machine so refcounts can't move and disable module. */
	ret = try_stop_module(mod, flags, &forced);
	if (ret != 0)
		goto out;

	/* Never wait if forced. */
	if (!forced && module_refcount(mod) != 0)
		wait_for_zero_refcount(mod);

868
	mutex_unlock(&module_mutex);
L
Lucas De Marchi 已提交
869
	/* Final destruction now no one is using it. */
870
	if (mod->exit != NULL)
L
Linus Torvalds 已提交
871
		mod->exit();
872 873
	blocking_notifier_call_chain(&module_notify_list,
				     MODULE_STATE_GOING, mod);
874
	async_synchronize_full();
875

876
	/* Store the name of the last unloaded module for diagnostic purposes */
877
	strlcpy(last_unloaded_module, mod->name, sizeof(last_unloaded_module));
L
Linus Torvalds 已提交
878

879 880 881
	free_module(mod);
	return 0;
out:
882
	mutex_unlock(&module_mutex);
L
Linus Torvalds 已提交
883 884 885
	return ret;
}

886
static inline void print_unload_info(struct seq_file *m, struct module *mod)
L
Linus Torvalds 已提交
887 888 889 890
{
	struct module_use *use;
	int printed_something = 0;

891
	seq_printf(m, " %lu ", module_refcount(mod));
L
Linus Torvalds 已提交
892 893 894

	/* Always include a trailing , so userspace can differentiate
           between this and the old multi-field proc format. */
895
	list_for_each_entry(use, &mod->source_list, source_list) {
L
Linus Torvalds 已提交
896
		printed_something = 1;
897
		seq_printf(m, "%s,", use->source->name);
L
Linus Torvalds 已提交
898 899 900 901 902 903 904 905 906 907 908 909 910 911 912
	}

	if (mod->init != NULL && mod->exit == NULL) {
		printed_something = 1;
		seq_printf(m, "[permanent],");
	}

	if (!printed_something)
		seq_printf(m, "-");
}

void __symbol_put(const char *symbol)
{
	struct module *owner;

913
	preempt_disable();
914
	if (!find_symbol(symbol, &owner, NULL, true, false))
L
Linus Torvalds 已提交
915 916
		BUG();
	module_put(owner);
917
	preempt_enable();
L
Linus Torvalds 已提交
918 919 920
}
EXPORT_SYMBOL(__symbol_put);

921
/* Note this assumes addr is a function, which it currently always is. */
L
Linus Torvalds 已提交
922 923
void symbol_put_addr(void *addr)
{
924
	struct module *modaddr;
925
	unsigned long a = (unsigned long)dereference_function_descriptor(addr);
L
Linus Torvalds 已提交
926

927
	if (core_kernel_text(a))
928
		return;
L
Linus Torvalds 已提交
929

930 931
	/* module_text_address is safe here: we're supposed to have reference
	 * to module from symbol_get, so it can't go away. */
932
	modaddr = __module_text_address(a);
933
	BUG_ON(!modaddr);
934
	module_put(modaddr);
L
Linus Torvalds 已提交
935 936 937 938
}
EXPORT_SYMBOL_GPL(symbol_put_addr);

static ssize_t show_refcnt(struct module_attribute *mattr,
939
			   struct module_kobject *mk, char *buffer)
L
Linus Torvalds 已提交
940
{
941
	return sprintf(buffer, "%lu\n", module_refcount(mk->mod));
L
Linus Torvalds 已提交
942 943
}

944 945
static struct module_attribute modinfo_refcnt =
	__ATTR(refcnt, 0444, show_refcnt, NULL);
L
Linus Torvalds 已提交
946

947 948 949 950 951 952 953 954 955 956 957 958 959 960 961 962 963 964 965 966 967 968 969 970 971 972 973 974 975 976
void __module_get(struct module *module)
{
	if (module) {
		preempt_disable();
		__this_cpu_inc(module->refptr->incs);
		trace_module_get(module, _RET_IP_);
		preempt_enable();
	}
}
EXPORT_SYMBOL(__module_get);

bool try_module_get(struct module *module)
{
	bool ret = true;

	if (module) {
		preempt_disable();

		if (likely(module_is_live(module))) {
			__this_cpu_inc(module->refptr->incs);
			trace_module_get(module, _RET_IP_);
		} else
			ret = false;

		preempt_enable();
	}
	return ret;
}
EXPORT_SYMBOL(try_module_get);

977 978 979
void module_put(struct module *module)
{
	if (module) {
980
		preempt_disable();
981 982
		smp_wmb(); /* see comment in module_refcount */
		__this_cpu_inc(module->refptr->decs);
983

984
		trace_module_put(module, _RET_IP_);
985 986 987
		/* Maybe they're waiting for us to drop reference? */
		if (unlikely(!module_is_live(module)))
			wake_up_process(module->waiter);
988
		preempt_enable();
989 990 991 992
	}
}
EXPORT_SYMBOL(module_put);

L
Linus Torvalds 已提交
993
#else /* !CONFIG_MODULE_UNLOAD */
994
static inline void print_unload_info(struct seq_file *m, struct module *mod)
L
Linus Torvalds 已提交
995 996 997 998 999 1000 1001 1002 1003
{
	/* We don't know the usage count, or what modules are using. */
	seq_printf(m, " - -");
}

static inline void module_unload_free(struct module *mod)
{
}

1004
int ref_module(struct module *a, struct module *b)
L
Linus Torvalds 已提交
1005
{
1006
	return strong_try_module_get(b);
L
Linus Torvalds 已提交
1007
}
1008
EXPORT_SYMBOL_GPL(ref_module);
L
Linus Torvalds 已提交
1009

1010
static inline int module_unload_init(struct module *mod)
L
Linus Torvalds 已提交
1011
{
1012
	return 0;
L
Linus Torvalds 已提交
1013 1014 1015
}
#endif /* CONFIG_MODULE_UNLOAD */

1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035
static size_t module_flags_taint(struct module *mod, char *buf)
{
	size_t l = 0;

	if (mod->taints & (1 << TAINT_PROPRIETARY_MODULE))
		buf[l++] = 'P';
	if (mod->taints & (1 << TAINT_OOT_MODULE))
		buf[l++] = 'O';
	if (mod->taints & (1 << TAINT_FORCED_MODULE))
		buf[l++] = 'F';
	if (mod->taints & (1 << TAINT_CRAP))
		buf[l++] = 'C';
	/*
	 * TAINT_FORCED_RMMOD: could be added.
	 * TAINT_UNSAFE_SMP, TAINT_MACHINE_CHECK, TAINT_BAD_PAGE don't
	 * apply to modules.
	 */
	return l;
}

1036
static ssize_t show_initstate(struct module_attribute *mattr,
1037
			      struct module_kobject *mk, char *buffer)
1038 1039 1040
{
	const char *state = "unknown";

1041
	switch (mk->mod->state) {
1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054
	case MODULE_STATE_LIVE:
		state = "live";
		break;
	case MODULE_STATE_COMING:
		state = "coming";
		break;
	case MODULE_STATE_GOING:
		state = "going";
		break;
	}
	return sprintf(buffer, "%s\n", state);
}

1055 1056
static struct module_attribute modinfo_initstate =
	__ATTR(initstate, 0444, show_initstate, NULL);
1057

1058 1059 1060 1061 1062 1063 1064 1065 1066 1067 1068
static ssize_t store_uevent(struct module_attribute *mattr,
			    struct module_kobject *mk,
			    const char *buffer, size_t count)
{
	enum kobject_action action;

	if (kobject_action_type(buffer, count, &action) == 0)
		kobject_uevent(&mk->kobj, action);
	return count;
}

1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 1101
struct module_attribute module_uevent =
	__ATTR(uevent, 0200, NULL, store_uevent);

static ssize_t show_coresize(struct module_attribute *mattr,
			     struct module_kobject *mk, char *buffer)
{
	return sprintf(buffer, "%u\n", mk->mod->core_size);
}

static struct module_attribute modinfo_coresize =
	__ATTR(coresize, 0444, show_coresize, NULL);

static ssize_t show_initsize(struct module_attribute *mattr,
			     struct module_kobject *mk, char *buffer)
{
	return sprintf(buffer, "%u\n", mk->mod->init_size);
}

static struct module_attribute modinfo_initsize =
	__ATTR(initsize, 0444, show_initsize, NULL);

static ssize_t show_taint(struct module_attribute *mattr,
			  struct module_kobject *mk, char *buffer)
{
	size_t l;

	l = module_flags_taint(mk->mod, buffer);
	buffer[l++] = '\n';
	return l;
}

static struct module_attribute modinfo_taint =
	__ATTR(taint, 0444, show_taint, NULL);
1102

1103
static struct module_attribute *modinfo_attrs[] = {
1104
	&module_uevent,
1105 1106
	&modinfo_version,
	&modinfo_srcversion,
1107 1108 1109 1110
	&modinfo_initstate,
	&modinfo_coresize,
	&modinfo_initsize,
	&modinfo_taint,
1111
#ifdef CONFIG_MODULE_UNLOAD
1112
	&modinfo_refcnt,
1113 1114 1115 1116
#endif
	NULL,
};

L
Linus Torvalds 已提交
1117 1118
static const char vermagic[] = VERMAGIC_STRING;

1119
static int try_to_force_load(struct module *mod, const char *reason)
1120 1121
{
#ifdef CONFIG_MODULE_FORCE_LOAD
1122
	if (!test_taint(TAINT_FORCED_MODULE))
1123 1124
		printk(KERN_WARNING "%s: %s: kernel tainted.\n",
		       mod->name, reason);
1125 1126 1127 1128 1129 1130 1131
	add_taint_module(mod, TAINT_FORCED_MODULE);
	return 0;
#else
	return -ENOEXEC;
#endif
}

L
Linus Torvalds 已提交
1132
#ifdef CONFIG_MODVERSIONS
1133 1134 1135 1136 1137 1138 1139 1140 1141 1142 1143
/* If the arch applies (non-zero) relocations to kernel kcrctab, unapply it. */
static unsigned long maybe_relocated(unsigned long crc,
				     const struct module *crc_owner)
{
#ifdef ARCH_RELOCATES_KCRCTAB
	if (crc_owner == NULL)
		return crc - (unsigned long)reloc_start;
#endif
	return crc;
}

L
Linus Torvalds 已提交
1144 1145 1146 1147
static int check_version(Elf_Shdr *sechdrs,
			 unsigned int versindex,
			 const char *symname,
			 struct module *mod, 
1148 1149
			 const unsigned long *crc,
			 const struct module *crc_owner)
L
Linus Torvalds 已提交
1150 1151 1152 1153 1154 1155 1156 1157
{
	unsigned int i, num_versions;
	struct modversion_info *versions;

	/* Exporting module didn't supply crcs?  OK, we're already tainted. */
	if (!crc)
		return 1;

1158 1159 1160 1161
	/* No versions at all?  modprobe --force does this. */
	if (versindex == 0)
		return try_to_force_load(mod, symname) == 0;

L
Linus Torvalds 已提交
1162 1163 1164 1165 1166 1167 1168 1169
	versions = (void *) sechdrs[versindex].sh_addr;
	num_versions = sechdrs[versindex].sh_size
		/ sizeof(struct modversion_info);

	for (i = 0; i < num_versions; i++) {
		if (strcmp(versions[i].name, symname) != 0)
			continue;

1170
		if (versions[i].crc == maybe_relocated(*crc, crc_owner))
L
Linus Torvalds 已提交
1171
			return 1;
1172
		pr_debug("Found checksum %lX vs module %lX\n",
1173
		       maybe_relocated(*crc, crc_owner), versions[i].crc);
1174
		goto bad_version;
L
Linus Torvalds 已提交
1175
	}
1176

1177 1178 1179
	printk(KERN_WARNING "%s: no symbol version for %s\n",
	       mod->name, symname);
	return 0;
1180 1181 1182 1183 1184

bad_version:
	printk("%s: disagrees about version of symbol %s\n",
	       mod->name, symname);
	return 0;
L
Linus Torvalds 已提交
1185 1186 1187 1188 1189 1190 1191 1192
}

static inline int check_modstruct_version(Elf_Shdr *sechdrs,
					  unsigned int versindex,
					  struct module *mod)
{
	const unsigned long *crc;

1193 1194
	/* Since this should be found in kernel (which can't be removed),
	 * no locking is necessary. */
1195 1196
	if (!find_symbol(MODULE_SYMBOL_PREFIX "module_layout", NULL,
			 &crc, true, false))
L
Linus Torvalds 已提交
1197
		BUG();
1198 1199
	return check_version(sechdrs, versindex, "module_layout", mod, crc,
			     NULL);
L
Linus Torvalds 已提交
1200 1201
}

1202 1203 1204
/* First part is kernel version, which we ignore if module has crcs. */
static inline int same_magic(const char *amagic, const char *bmagic,
			     bool has_crcs)
L
Linus Torvalds 已提交
1205
{
1206 1207 1208 1209
	if (has_crcs) {
		amagic += strcspn(amagic, " ");
		bmagic += strcspn(bmagic, " ");
	}
L
Linus Torvalds 已提交
1210 1211 1212 1213 1214 1215 1216
	return strcmp(amagic, bmagic) == 0;
}
#else
static inline int check_version(Elf_Shdr *sechdrs,
				unsigned int versindex,
				const char *symname,
				struct module *mod, 
1217 1218
				const unsigned long *crc,
				const struct module *crc_owner)
L
Linus Torvalds 已提交
1219 1220 1221 1222 1223 1224 1225 1226 1227 1228 1229
{
	return 1;
}

static inline int check_modstruct_version(Elf_Shdr *sechdrs,
					  unsigned int versindex,
					  struct module *mod)
{
	return 1;
}

1230 1231
static inline int same_magic(const char *amagic, const char *bmagic,
			     bool has_crcs)
L
Linus Torvalds 已提交
1232 1233 1234 1235 1236
{
	return strcmp(amagic, bmagic) == 0;
}
#endif /* CONFIG_MODVERSIONS */

1237
/* Resolve a symbol for this module.  I.e. if we find one, record usage. */
1238 1239
static const struct kernel_symbol *resolve_symbol(struct module *mod,
						  const struct load_info *info,
1240
						  const char *name,
1241
						  char ownername[])
L
Linus Torvalds 已提交
1242 1243
{
	struct module *owner;
1244
	const struct kernel_symbol *sym;
L
Linus Torvalds 已提交
1245
	const unsigned long *crc;
1246
	int err;
L
Linus Torvalds 已提交
1247

1248
	mutex_lock(&module_mutex);
1249
	sym = find_symbol(name, &owner, &crc,
1250
			  !(mod->taints & (1 << TAINT_PROPRIETARY_MODULE)), true);
1251 1252 1253
	if (!sym)
		goto unlock;

1254 1255
	if (!check_version(info->sechdrs, info->index.vers, name, mod, crc,
			   owner)) {
1256 1257
		sym = ERR_PTR(-EINVAL);
		goto getname;
L
Linus Torvalds 已提交
1258
	}
1259 1260 1261 1262 1263 1264 1265 1266 1267 1268 1269

	err = ref_module(mod, owner);
	if (err) {
		sym = ERR_PTR(err);
		goto getname;
	}

getname:
	/* We must make copy under the lock if we failed to get ref. */
	strncpy(ownername, module_name(owner), MODULE_NAME_LEN);
unlock:
1270
	mutex_unlock(&module_mutex);
1271
	return sym;
L
Linus Torvalds 已提交
1272 1273
}

1274 1275 1276 1277
static const struct kernel_symbol *
resolve_symbol_wait(struct module *mod,
		    const struct load_info *info,
		    const char *name)
1278 1279
{
	const struct kernel_symbol *ksym;
1280
	char owner[MODULE_NAME_LEN];
1281 1282

	if (wait_event_interruptible_timeout(module_wq,
1283 1284
			!IS_ERR(ksym = resolve_symbol(mod, info, name, owner))
			|| PTR_ERR(ksym) != -EBUSY,
1285 1286
					     30 * HZ) <= 0) {
		printk(KERN_WARNING "%s: gave up waiting for init of module %s.\n",
1287
		       mod->name, owner);
1288 1289 1290 1291
	}
	return ksym;
}

L
Linus Torvalds 已提交
1292 1293 1294 1295
/*
 * /sys/module/foo/sections stuff
 * J. Corbet <corbet@lwn.net>
 */
R
Rusty Russell 已提交
1296
#ifdef CONFIG_SYSFS
1297

R
Rusty Russell 已提交
1298
#ifdef CONFIG_KALLSYMS
1299 1300 1301 1302 1303
static inline bool sect_empty(const Elf_Shdr *sect)
{
	return !(sect->sh_flags & SHF_ALLOC) || sect->sh_size == 0;
}

1304 1305 1306 1307 1308 1309 1310 1311 1312 1313 1314 1315 1316 1317
struct module_sect_attr
{
	struct module_attribute mattr;
	char *name;
	unsigned long address;
};

struct module_sect_attrs
{
	struct attribute_group grp;
	unsigned int nsections;
	struct module_sect_attr attrs[0];
};

L
Linus Torvalds 已提交
1318
static ssize_t module_sect_show(struct module_attribute *mattr,
1319
				struct module_kobject *mk, char *buf)
L
Linus Torvalds 已提交
1320 1321 1322
{
	struct module_sect_attr *sattr =
		container_of(mattr, struct module_sect_attr, mattr);
1323
	return sprintf(buf, "0x%pK\n", (void *)sattr->address);
L
Linus Torvalds 已提交
1324 1325
}

1326 1327
static void free_sect_attrs(struct module_sect_attrs *sect_attrs)
{
1328
	unsigned int section;
1329 1330 1331 1332 1333 1334

	for (section = 0; section < sect_attrs->nsections; section++)
		kfree(sect_attrs->attrs[section].name);
	kfree(sect_attrs);
}

R
Rusty Russell 已提交
1335
static void add_sect_attrs(struct module *mod, const struct load_info *info)
L
Linus Torvalds 已提交
1336 1337 1338 1339 1340
{
	unsigned int nloaded = 0, i, size[2];
	struct module_sect_attrs *sect_attrs;
	struct module_sect_attr *sattr;
	struct attribute **gattr;
1341

L
Linus Torvalds 已提交
1342
	/* Count loaded sections and allocate structures */
R
Rusty Russell 已提交
1343 1344
	for (i = 0; i < info->hdr->e_shnum; i++)
		if (!sect_empty(&info->sechdrs[i]))
L
Linus Torvalds 已提交
1345 1346 1347 1348 1349
			nloaded++;
	size[0] = ALIGN(sizeof(*sect_attrs)
			+ nloaded * sizeof(sect_attrs->attrs[0]),
			sizeof(sect_attrs->grp.attrs[0]));
	size[1] = (nloaded + 1) * sizeof(sect_attrs->grp.attrs[0]);
1350 1351
	sect_attrs = kzalloc(size[0] + size[1], GFP_KERNEL);
	if (sect_attrs == NULL)
L
Linus Torvalds 已提交
1352 1353 1354 1355 1356 1357
		return;

	/* Setup section attributes. */
	sect_attrs->grp.name = "sections";
	sect_attrs->grp.attrs = (void *)sect_attrs + size[0];

1358
	sect_attrs->nsections = 0;
L
Linus Torvalds 已提交
1359 1360
	sattr = &sect_attrs->attrs[0];
	gattr = &sect_attrs->grp.attrs[0];
R
Rusty Russell 已提交
1361 1362 1363
	for (i = 0; i < info->hdr->e_shnum; i++) {
		Elf_Shdr *sec = &info->sechdrs[i];
		if (sect_empty(sec))
1364
			continue;
R
Rusty Russell 已提交
1365 1366
		sattr->address = sec->sh_addr;
		sattr->name = kstrdup(info->secstrings + sec->sh_name,
1367 1368 1369 1370
					GFP_KERNEL);
		if (sattr->name == NULL)
			goto out;
		sect_attrs->nsections++;
1371
		sysfs_attr_init(&sattr->mattr.attr);
L
Linus Torvalds 已提交
1372 1373 1374 1375 1376 1377 1378 1379 1380 1381 1382 1383 1384 1385
		sattr->mattr.show = module_sect_show;
		sattr->mattr.store = NULL;
		sattr->mattr.attr.name = sattr->name;
		sattr->mattr.attr.mode = S_IRUGO;
		*(gattr++) = &(sattr++)->mattr.attr;
	}
	*gattr = NULL;

	if (sysfs_create_group(&mod->mkobj.kobj, &sect_attrs->grp))
		goto out;

	mod->sect_attrs = sect_attrs;
	return;
  out:
1386
	free_sect_attrs(sect_attrs);
L
Linus Torvalds 已提交
1387 1388 1389 1390 1391 1392 1393 1394 1395
}

static void remove_sect_attrs(struct module *mod)
{
	if (mod->sect_attrs) {
		sysfs_remove_group(&mod->mkobj.kobj,
				   &mod->sect_attrs->grp);
		/* We are positive that no one is using any sect attrs
		 * at this point.  Deallocate immediately. */
1396
		free_sect_attrs(mod->sect_attrs);
L
Linus Torvalds 已提交
1397 1398 1399 1400
		mod->sect_attrs = NULL;
	}
}

1401 1402 1403 1404 1405 1406 1407 1408 1409 1410
/*
 * /sys/module/foo/notes/.section.name gives contents of SHT_NOTE sections.
 */

struct module_notes_attrs {
	struct kobject *dir;
	unsigned int notes;
	struct bin_attribute attrs[0];
};

1411
static ssize_t module_notes_read(struct file *filp, struct kobject *kobj,
1412 1413 1414 1415 1416 1417 1418 1419 1420 1421 1422 1423 1424 1425 1426 1427 1428
				 struct bin_attribute *bin_attr,
				 char *buf, loff_t pos, size_t count)
{
	/*
	 * The caller checked the pos and count against our size.
	 */
	memcpy(buf, bin_attr->private + pos, count);
	return count;
}

static void free_notes_attrs(struct module_notes_attrs *notes_attrs,
			     unsigned int i)
{
	if (notes_attrs->dir) {
		while (i-- > 0)
			sysfs_remove_bin_file(notes_attrs->dir,
					      &notes_attrs->attrs[i]);
1429
		kobject_put(notes_attrs->dir);
1430 1431 1432 1433
	}
	kfree(notes_attrs);
}

R
Rusty Russell 已提交
1434
static void add_notes_attrs(struct module *mod, const struct load_info *info)
1435 1436 1437 1438 1439
{
	unsigned int notes, loaded, i;
	struct module_notes_attrs *notes_attrs;
	struct bin_attribute *nattr;

1440 1441 1442 1443
	/* failed to create section attributes, so can't create notes */
	if (!mod->sect_attrs)
		return;

1444 1445
	/* Count notes sections and allocate structures.  */
	notes = 0;
R
Rusty Russell 已提交
1446 1447 1448
	for (i = 0; i < info->hdr->e_shnum; i++)
		if (!sect_empty(&info->sechdrs[i]) &&
		    (info->sechdrs[i].sh_type == SHT_NOTE))
1449 1450 1451 1452 1453 1454 1455 1456 1457 1458 1459 1460 1461
			++notes;

	if (notes == 0)
		return;

	notes_attrs = kzalloc(sizeof(*notes_attrs)
			      + notes * sizeof(notes_attrs->attrs[0]),
			      GFP_KERNEL);
	if (notes_attrs == NULL)
		return;

	notes_attrs->notes = notes;
	nattr = &notes_attrs->attrs[0];
R
Rusty Russell 已提交
1462 1463
	for (loaded = i = 0; i < info->hdr->e_shnum; ++i) {
		if (sect_empty(&info->sechdrs[i]))
1464
			continue;
R
Rusty Russell 已提交
1465
		if (info->sechdrs[i].sh_type == SHT_NOTE) {
1466
			sysfs_bin_attr_init(nattr);
1467 1468
			nattr->attr.name = mod->sect_attrs->attrs[loaded].name;
			nattr->attr.mode = S_IRUGO;
R
Rusty Russell 已提交
1469 1470
			nattr->size = info->sechdrs[i].sh_size;
			nattr->private = (void *) info->sechdrs[i].sh_addr;
1471 1472 1473 1474 1475 1476
			nattr->read = module_notes_read;
			++nattr;
		}
		++loaded;
	}

1477
	notes_attrs->dir = kobject_create_and_add("notes", &mod->mkobj.kobj);
1478 1479 1480 1481 1482 1483 1484 1485 1486 1487 1488 1489 1490 1491 1492 1493 1494 1495 1496 1497 1498
	if (!notes_attrs->dir)
		goto out;

	for (i = 0; i < notes; ++i)
		if (sysfs_create_bin_file(notes_attrs->dir,
					  &notes_attrs->attrs[i]))
			goto out;

	mod->notes_attrs = notes_attrs;
	return;

  out:
	free_notes_attrs(notes_attrs, i);
}

static void remove_notes_attrs(struct module *mod)
{
	if (mod->notes_attrs)
		free_notes_attrs(mod->notes_attrs, mod->notes_attrs->notes);
}

L
Linus Torvalds 已提交
1499
#else
1500

R
Rusty Russell 已提交
1501 1502
static inline void add_sect_attrs(struct module *mod,
				  const struct load_info *info)
L
Linus Torvalds 已提交
1503 1504 1505 1506 1507 1508
{
}

static inline void remove_sect_attrs(struct module *mod)
{
}
1509

R
Rusty Russell 已提交
1510 1511
static inline void add_notes_attrs(struct module *mod,
				   const struct load_info *info)
1512 1513 1514 1515 1516 1517
{
}

static inline void remove_notes_attrs(struct module *mod)
{
}
R
Rusty Russell 已提交
1518
#endif /* CONFIG_KALLSYMS */
L
Linus Torvalds 已提交
1519

1520 1521 1522 1523 1524 1525
static void add_usage_links(struct module *mod)
{
#ifdef CONFIG_MODULE_UNLOAD
	struct module_use *use;
	int nowarn;

1526
	mutex_lock(&module_mutex);
1527 1528 1529 1530
	list_for_each_entry(use, &mod->target_list, target_list) {
		nowarn = sysfs_create_link(use->target->holders_dir,
					   &mod->mkobj.kobj, mod->name);
	}
1531
	mutex_unlock(&module_mutex);
1532 1533 1534 1535 1536 1537 1538 1539
#endif
}

static void del_usage_links(struct module *mod)
{
#ifdef CONFIG_MODULE_UNLOAD
	struct module_use *use;

1540
	mutex_lock(&module_mutex);
1541 1542
	list_for_each_entry(use, &mod->target_list, target_list)
		sysfs_remove_link(use->target->holders_dir, mod->name);
1543
	mutex_unlock(&module_mutex);
1544 1545 1546
#endif
}

1547
static int module_add_modinfo_attrs(struct module *mod)
1548 1549
{
	struct module_attribute *attr;
1550
	struct module_attribute *temp_attr;
1551 1552 1553
	int error = 0;
	int i;

1554 1555 1556 1557 1558 1559 1560
	mod->modinfo_attrs = kzalloc((sizeof(struct module_attribute) *
					(ARRAY_SIZE(modinfo_attrs) + 1)),
					GFP_KERNEL);
	if (!mod->modinfo_attrs)
		return -ENOMEM;

	temp_attr = mod->modinfo_attrs;
1561 1562
	for (i = 0; (attr = modinfo_attrs[i]) && !error; i++) {
		if (!attr->test ||
1563 1564
		    (attr->test && attr->test(mod))) {
			memcpy(temp_attr, attr, sizeof(*temp_attr));
1565
			sysfs_attr_init(&temp_attr->attr);
1566 1567 1568
			error = sysfs_create_file(&mod->mkobj.kobj,&temp_attr->attr);
			++temp_attr;
		}
1569 1570 1571 1572
	}
	return error;
}

1573
static void module_remove_modinfo_attrs(struct module *mod)
1574 1575 1576 1577
{
	struct module_attribute *attr;
	int i;

1578 1579 1580 1581
	for (i = 0; (attr = &mod->modinfo_attrs[i]); i++) {
		/* pick a field to test for end of list */
		if (!attr->attr.name)
			break;
1582
		sysfs_remove_file(&mod->mkobj.kobj,&attr->attr);
1583 1584
		if (attr->free)
			attr->free(mod);
1585
	}
1586
	kfree(mod->modinfo_attrs);
1587
}
L
Linus Torvalds 已提交
1588

1589
static int mod_sysfs_init(struct module *mod)
L
Linus Torvalds 已提交
1590 1591
{
	int err;
1592
	struct kobject *kobj;
L
Linus Torvalds 已提交
1593

1594 1595
	if (!module_sysfs_initialized) {
		printk(KERN_ERR "%s: module sysfs not initialized\n",
1596 1597 1598 1599
		       mod->name);
		err = -EINVAL;
		goto out;
	}
1600 1601 1602 1603 1604 1605 1606 1607 1608

	kobj = kset_find_obj(module_kset, mod->name);
	if (kobj) {
		printk(KERN_ERR "%s: module is already loaded\n", mod->name);
		kobject_put(kobj);
		err = -EINVAL;
		goto out;
	}

L
Linus Torvalds 已提交
1609
	mod->mkobj.mod = mod;
1610

1611 1612 1613 1614 1615 1616
	memset(&mod->mkobj.kobj, 0, sizeof(mod->mkobj.kobj));
	mod->mkobj.kobj.kset = module_kset;
	err = kobject_init_and_add(&mod->mkobj.kobj, &module_ktype, NULL,
				   "%s", mod->name);
	if (err)
		kobject_put(&mod->mkobj.kobj);
K
Kay Sievers 已提交
1617

1618
	/* delay uevent until full sysfs population */
K
Kay Sievers 已提交
1619 1620 1621 1622
out:
	return err;
}

1623
static int mod_sysfs_setup(struct module *mod,
R
Rusty Russell 已提交
1624
			   const struct load_info *info,
K
Kay Sievers 已提交
1625 1626 1627 1628 1629
			   struct kernel_param *kparam,
			   unsigned int num_params)
{
	int err;

1630 1631 1632 1633
	err = mod_sysfs_init(mod);
	if (err)
		goto out;

1634
	mod->holders_dir = kobject_create_and_add("holders", &mod->mkobj.kobj);
1635 1636
	if (!mod->holders_dir) {
		err = -ENOMEM;
K
Kay Sievers 已提交
1637
		goto out_unreg;
1638
	}
K
Kay Sievers 已提交
1639

L
Linus Torvalds 已提交
1640 1641
	err = module_param_sysfs_setup(mod, kparam, num_params);
	if (err)
K
Kay Sievers 已提交
1642
		goto out_unreg_holders;
L
Linus Torvalds 已提交
1643

1644 1645
	err = module_add_modinfo_attrs(mod);
	if (err)
1646
		goto out_unreg_param;
1647

1648
	add_usage_links(mod);
R
Rusty Russell 已提交
1649 1650
	add_sect_attrs(mod, info);
	add_notes_attrs(mod, info);
1651

1652
	kobject_uevent(&mod->mkobj.kobj, KOBJ_ADD);
L
Linus Torvalds 已提交
1653 1654
	return 0;

1655 1656
out_unreg_param:
	module_param_sysfs_remove(mod);
K
Kay Sievers 已提交
1657
out_unreg_holders:
1658
	kobject_put(mod->holders_dir);
K
Kay Sievers 已提交
1659
out_unreg:
1660
	kobject_put(&mod->mkobj.kobj);
1661
out:
L
Linus Torvalds 已提交
1662 1663
	return err;
}
1664 1665 1666

static void mod_sysfs_fini(struct module *mod)
{
R
Rusty Russell 已提交
1667 1668
	remove_notes_attrs(mod);
	remove_sect_attrs(mod);
1669 1670 1671
	kobject_put(&mod->mkobj.kobj);
}

R
Rusty Russell 已提交
1672
#else /* !CONFIG_SYSFS */
1673

R
Rusty Russell 已提交
1674 1675
static int mod_sysfs_setup(struct module *mod,
			   const struct load_info *info,
1676 1677 1678 1679 1680 1681
			   struct kernel_param *kparam,
			   unsigned int num_params)
{
	return 0;
}

1682 1683 1684 1685
static void mod_sysfs_fini(struct module *mod)
{
}

1686 1687 1688 1689
static void module_remove_modinfo_attrs(struct module *mod)
{
}

1690 1691 1692 1693
static void del_usage_links(struct module *mod)
{
}

1694
#endif /* CONFIG_SYSFS */
L
Linus Torvalds 已提交
1695

1696
static void mod_sysfs_teardown(struct module *mod)
L
Linus Torvalds 已提交
1697
{
1698
	del_usage_links(mod);
1699
	module_remove_modinfo_attrs(mod);
L
Linus Torvalds 已提交
1700
	module_param_sysfs_remove(mod);
1701 1702
	kobject_put(mod->mkobj.drivers_dir);
	kobject_put(mod->holders_dir);
1703
	mod_sysfs_fini(mod);
L
Linus Torvalds 已提交
1704 1705 1706 1707 1708 1709 1710 1711 1712 1713
}

/*
 * unlink the module with the whole machine is stopped with interrupts off
 * - this defends against kallsyms not taking locks
 */
static int __unlink_module(void *_mod)
{
	struct module *mod = _mod;
	list_del(&mod->list);
1714
	module_bug_cleanup(mod);
L
Linus Torvalds 已提交
1715 1716 1717
	return 0;
}

1718 1719 1720 1721 1722 1723 1724 1725 1726 1727 1728 1729 1730 1731 1732 1733 1734 1735 1736 1737 1738 1739 1740 1741 1742 1743 1744 1745 1746 1747 1748 1749 1750 1751 1752 1753 1754 1755 1756 1757 1758 1759 1760 1761
#ifdef CONFIG_DEBUG_SET_MODULE_RONX
/*
 * LKM RO/NX protection: protect module's text/ro-data
 * from modification and any data from execution.
 */
void set_page_attributes(void *start, void *end, int (*set)(unsigned long start, int num_pages))
{
	unsigned long begin_pfn = PFN_DOWN((unsigned long)start);
	unsigned long end_pfn = PFN_DOWN((unsigned long)end);

	if (end_pfn > begin_pfn)
		set(begin_pfn << PAGE_SHIFT, end_pfn - begin_pfn);
}

static void set_section_ro_nx(void *base,
			unsigned long text_size,
			unsigned long ro_size,
			unsigned long total_size)
{
	/* begin and end PFNs of the current subsection */
	unsigned long begin_pfn;
	unsigned long end_pfn;

	/*
	 * Set RO for module text and RO-data:
	 * - Always protect first page.
	 * - Do not protect last partial page.
	 */
	if (ro_size > 0)
		set_page_attributes(base, base + ro_size, set_memory_ro);

	/*
	 * Set NX permissions for module data:
	 * - Do not protect first partial page.
	 * - Always protect last page.
	 */
	if (total_size > text_size) {
		begin_pfn = PFN_UP((unsigned long)base + text_size);
		end_pfn = PFN_UP((unsigned long)base + total_size);
		if (end_pfn > begin_pfn)
			set_memory_nx(begin_pfn << PAGE_SHIFT, end_pfn - begin_pfn);
	}
}

1762 1763 1764 1765 1766 1767 1768 1769 1770 1771 1772 1773 1774 1775 1776 1777 1778 1779
static void unset_module_core_ro_nx(struct module *mod)
{
	set_page_attributes(mod->module_core + mod->core_text_size,
		mod->module_core + mod->core_size,
		set_memory_x);
	set_page_attributes(mod->module_core,
		mod->module_core + mod->core_ro_size,
		set_memory_rw);
}

static void unset_module_init_ro_nx(struct module *mod)
{
	set_page_attributes(mod->module_init + mod->init_text_size,
		mod->module_init + mod->init_size,
		set_memory_x);
	set_page_attributes(mod->module_init,
		mod->module_init + mod->init_ro_size,
		set_memory_rw);
1780 1781 1782
}

/* Iterate through all modules and set each module's text as RW */
1783
void set_all_modules_text_rw(void)
1784 1785 1786 1787 1788 1789 1790 1791 1792 1793 1794 1795 1796 1797 1798 1799 1800 1801 1802 1803
{
	struct module *mod;

	mutex_lock(&module_mutex);
	list_for_each_entry_rcu(mod, &modules, list) {
		if ((mod->module_core) && (mod->core_text_size)) {
			set_page_attributes(mod->module_core,
						mod->module_core + mod->core_text_size,
						set_memory_rw);
		}
		if ((mod->module_init) && (mod->init_text_size)) {
			set_page_attributes(mod->module_init,
						mod->module_init + mod->init_text_size,
						set_memory_rw);
		}
	}
	mutex_unlock(&module_mutex);
}

/* Iterate through all modules and set each module's text as RO */
1804
void set_all_modules_text_ro(void)
1805 1806 1807 1808 1809 1810 1811 1812 1813 1814 1815 1816 1817 1818 1819 1820 1821 1822 1823 1824
{
	struct module *mod;

	mutex_lock(&module_mutex);
	list_for_each_entry_rcu(mod, &modules, list) {
		if ((mod->module_core) && (mod->core_text_size)) {
			set_page_attributes(mod->module_core,
						mod->module_core + mod->core_text_size,
						set_memory_ro);
		}
		if ((mod->module_init) && (mod->init_text_size)) {
			set_page_attributes(mod->module_init,
						mod->module_init + mod->init_text_size,
						set_memory_ro);
		}
	}
	mutex_unlock(&module_mutex);
}
#else
static inline void set_section_ro_nx(void *base, unsigned long text_size, unsigned long ro_size, unsigned long total_size) { }
1825 1826
static void unset_module_core_ro_nx(struct module *mod) { }
static void unset_module_init_ro_nx(struct module *mod) { }
1827 1828
#endif

1829 1830 1831 1832 1833 1834 1835 1836 1837
void __weak module_free(struct module *mod, void *module_region)
{
	vfree(module_region);
}

void __weak module_arch_cleanup(struct module *mod)
{
}

1838
/* Free a module, remove from lists, etc. */
L
Linus Torvalds 已提交
1839 1840
static void free_module(struct module *mod)
{
1841 1842
	trace_module_free(mod);

L
Linus Torvalds 已提交
1843
	/* Delete from various lists */
1844
	mutex_lock(&module_mutex);
1845
	stop_machine(__unlink_module, mod, NULL);
1846
	mutex_unlock(&module_mutex);
1847
	mod_sysfs_teardown(mod);
L
Linus Torvalds 已提交
1848

1849 1850 1851
	/* Remove dynamic debug info */
	ddebug_remove_module(mod->name);

L
Linus Torvalds 已提交
1852 1853 1854 1855 1856 1857
	/* Arch-specific cleanup. */
	module_arch_cleanup(mod);

	/* Module unload stuff */
	module_unload_free(mod);

1858 1859 1860
	/* Free any allocated parameters. */
	destroy_params(mod->kp, mod->num_kp);

L
Linus Torvalds 已提交
1861
	/* This may be NULL, but that's OK */
1862
	unset_module_init_ro_nx(mod);
L
Linus Torvalds 已提交
1863 1864
	module_free(mod, mod->module_init);
	kfree(mod->args);
1865
	percpu_modfree(mod);
1866

I
Ingo Molnar 已提交
1867 1868 1869
	/* Free lock-classes: */
	lockdep_free_key_range(mod->module_core, mod->core_size);

L
Linus Torvalds 已提交
1870
	/* Finally, free the core (containing the module structure) */
1871
	unset_module_core_ro_nx(mod);
L
Linus Torvalds 已提交
1872
	module_free(mod, mod->module_core);
1873 1874 1875 1876

#ifdef CONFIG_MPU
	update_protections(current->mm);
#endif
L
Linus Torvalds 已提交
1877 1878 1879 1880 1881
}

void *__symbol_get(const char *symbol)
{
	struct module *owner;
1882
	const struct kernel_symbol *sym;
L
Linus Torvalds 已提交
1883

1884
	preempt_disable();
1885 1886 1887
	sym = find_symbol(symbol, &owner, NULL, true, true);
	if (sym && strong_try_module_get(owner))
		sym = NULL;
1888
	preempt_enable();
L
Linus Torvalds 已提交
1889

1890
	return sym ? (void *)sym->value : NULL;
L
Linus Torvalds 已提交
1891 1892 1893
}
EXPORT_SYMBOL_GPL(__symbol_get);

1894 1895
/*
 * Ensure that an exported symbol [global namespace] does not already exist
1896
 * in the kernel or in some other module's exported symbol table.
1897 1898
 *
 * You must hold the module_mutex.
1899 1900 1901
 */
static int verify_export_symbols(struct module *mod)
{
1902
	unsigned int i;
1903
	struct module *owner;
1904 1905 1906 1907 1908 1909 1910 1911
	const struct kernel_symbol *s;
	struct {
		const struct kernel_symbol *sym;
		unsigned int num;
	} arr[] = {
		{ mod->syms, mod->num_syms },
		{ mod->gpl_syms, mod->num_gpl_syms },
		{ mod->gpl_future_syms, mod->num_gpl_future_syms },
1912
#ifdef CONFIG_UNUSED_SYMBOLS
1913 1914
		{ mod->unused_syms, mod->num_unused_syms },
		{ mod->unused_gpl_syms, mod->num_unused_gpl_syms },
1915
#endif
1916
	};
1917

1918 1919
	for (i = 0; i < ARRAY_SIZE(arr); i++) {
		for (s = arr[i].sym; s < arr[i].sym + arr[i].num; s++) {
1920
			if (find_symbol(s->name, &owner, NULL, true, false)) {
1921 1922 1923 1924 1925 1926
				printk(KERN_ERR
				       "%s: exports duplicate symbol %s"
				       " (owned by %s)\n",
				       mod->name, s->name, module_name(owner));
				return -ENOEXEC;
			}
1927
		}
1928 1929
	}
	return 0;
1930 1931
}

1932
/* Change all symbols so that st_value encodes the pointer directly. */
1933 1934 1935 1936
static int simplify_symbols(struct module *mod, const struct load_info *info)
{
	Elf_Shdr *symsec = &info->sechdrs[info->index.sym];
	Elf_Sym *sym = (void *)symsec->sh_addr;
L
Linus Torvalds 已提交
1937
	unsigned long secbase;
1938
	unsigned int i;
L
Linus Torvalds 已提交
1939
	int ret = 0;
1940
	const struct kernel_symbol *ksym;
L
Linus Torvalds 已提交
1941

1942 1943 1944
	for (i = 1; i < symsec->sh_size / sizeof(Elf_Sym); i++) {
		const char *name = info->strtab + sym[i].st_name;

L
Linus Torvalds 已提交
1945 1946 1947 1948
		switch (sym[i].st_shndx) {
		case SHN_COMMON:
			/* We compiled with -fno-common.  These are not
			   supposed to happen.  */
1949
			pr_debug("Common symbol: %s\n", name);
L
Linus Torvalds 已提交
1950 1951 1952 1953 1954 1955 1956
			printk("%s: please compile with -fno-common\n",
			       mod->name);
			ret = -ENOEXEC;
			break;

		case SHN_ABS:
			/* Don't need to do anything */
1957
			pr_debug("Absolute symbol: 0x%08lx\n",
L
Linus Torvalds 已提交
1958 1959 1960 1961
			       (long)sym[i].st_value);
			break;

		case SHN_UNDEF:
1962
			ksym = resolve_symbol_wait(mod, info, name);
L
Linus Torvalds 已提交
1963
			/* Ok if resolved.  */
1964
			if (ksym && !IS_ERR(ksym)) {
1965
				sym[i].st_value = ksym->value;
L
Linus Torvalds 已提交
1966
				break;
1967 1968
			}

L
Linus Torvalds 已提交
1969
			/* Ok if weak.  */
1970
			if (!ksym && ELF_ST_BIND(sym[i].st_info) == STB_WEAK)
L
Linus Torvalds 已提交
1971 1972
				break;

1973
			printk(KERN_WARNING "%s: Unknown symbol %s (err %li)\n",
1974
			       mod->name, name, PTR_ERR(ksym));
1975
			ret = PTR_ERR(ksym) ?: -ENOENT;
L
Linus Torvalds 已提交
1976 1977 1978 1979
			break;

		default:
			/* Divert to percpu allocation if a percpu var. */
1980
			if (sym[i].st_shndx == info->index.pcpu)
1981
				secbase = (unsigned long)mod_percpu(mod);
L
Linus Torvalds 已提交
1982
			else
1983
				secbase = info->sechdrs[sym[i].st_shndx].sh_addr;
L
Linus Torvalds 已提交
1984 1985 1986 1987 1988 1989 1990 1991
			sym[i].st_value += secbase;
			break;
		}
	}

	return ret;
}

1992
static int apply_relocations(struct module *mod, const struct load_info *info)
1993 1994 1995 1996 1997
{
	unsigned int i;
	int err = 0;

	/* Now do relocations. */
1998 1999
	for (i = 1; i < info->hdr->e_shnum; i++) {
		unsigned int infosec = info->sechdrs[i].sh_info;
2000 2001

		/* Not a valid relocation section? */
2002
		if (infosec >= info->hdr->e_shnum)
2003 2004 2005
			continue;

		/* Don't bother with non-allocated sections */
2006
		if (!(info->sechdrs[infosec].sh_flags & SHF_ALLOC))
2007 2008
			continue;

2009 2010 2011 2012 2013 2014
		if (info->sechdrs[i].sh_type == SHT_REL)
			err = apply_relocate(info->sechdrs, info->strtab,
					     info->index.sym, i, mod);
		else if (info->sechdrs[i].sh_type == SHT_RELA)
			err = apply_relocate_add(info->sechdrs, info->strtab,
						 info->index.sym, i, mod);
2015 2016 2017 2018 2019 2020
		if (err < 0)
			break;
	}
	return err;
}

2021 2022 2023 2024 2025 2026 2027 2028
/* Additional bytes needed by arch in front of individual sections */
unsigned int __weak arch_mod_section_prepend(struct module *mod,
					     unsigned int section)
{
	/* default implementation just returns zero */
	return 0;
}

L
Linus Torvalds 已提交
2029
/* Update size with this section: return offset. */
2030 2031
static long get_offset(struct module *mod, unsigned int *size,
		       Elf_Shdr *sechdr, unsigned int section)
L
Linus Torvalds 已提交
2032 2033 2034
{
	long ret;

2035
	*size += arch_mod_section_prepend(mod, section);
L
Linus Torvalds 已提交
2036 2037 2038 2039 2040 2041 2042 2043 2044
	ret = ALIGN(*size, sechdr->sh_addralign ?: 1);
	*size = ret + sechdr->sh_size;
	return ret;
}

/* Lay out the SHF_ALLOC sections in a way not dissimilar to how ld
   might -- code, read-only data, read-write data, small data.  Tally
   sizes, and place the offsets into sh_entsize fields: high bit means it
   belongs in init. */
2045
static void layout_sections(struct module *mod, struct load_info *info)
L
Linus Torvalds 已提交
2046 2047 2048 2049 2050 2051 2052 2053 2054 2055 2056 2057
{
	static unsigned long const masks[][2] = {
		/* NOTE: all executable code must be the first section
		 * in this array; otherwise modify the text_size
		 * finder in the two loops below */
		{ SHF_EXECINSTR | SHF_ALLOC, ARCH_SHF_SMALL },
		{ SHF_ALLOC, SHF_WRITE | ARCH_SHF_SMALL },
		{ SHF_WRITE | SHF_ALLOC, ARCH_SHF_SMALL },
		{ ARCH_SHF_SMALL | SHF_ALLOC, 0 }
	};
	unsigned int m, i;

2058 2059
	for (i = 0; i < info->hdr->e_shnum; i++)
		info->sechdrs[i].sh_entsize = ~0UL;
L
Linus Torvalds 已提交
2060

2061
	pr_debug("Core section allocation order:\n");
L
Linus Torvalds 已提交
2062
	for (m = 0; m < ARRAY_SIZE(masks); ++m) {
2063 2064 2065
		for (i = 0; i < info->hdr->e_shnum; ++i) {
			Elf_Shdr *s = &info->sechdrs[i];
			const char *sname = info->secstrings + s->sh_name;
L
Linus Torvalds 已提交
2066 2067 2068 2069

			if ((s->sh_flags & masks[m][0]) != masks[m][0]
			    || (s->sh_flags & masks[m][1])
			    || s->sh_entsize != ~0UL
2070
			    || strstarts(sname, ".init"))
L
Linus Torvalds 已提交
2071
				continue;
2072
			s->sh_entsize = get_offset(mod, &mod->core_size, s, i);
2073
			pr_debug("\t%s\n", sname);
L
Linus Torvalds 已提交
2074
		}
2075 2076 2077
		switch (m) {
		case 0: /* executable */
			mod->core_size = debug_align(mod->core_size);
L
Linus Torvalds 已提交
2078
			mod->core_text_size = mod->core_size;
2079 2080 2081 2082 2083 2084 2085 2086 2087
			break;
		case 1: /* RO: text and ro-data */
			mod->core_size = debug_align(mod->core_size);
			mod->core_ro_size = mod->core_size;
			break;
		case 3: /* whole core */
			mod->core_size = debug_align(mod->core_size);
			break;
		}
L
Linus Torvalds 已提交
2088 2089
	}

2090
	pr_debug("Init section allocation order:\n");
L
Linus Torvalds 已提交
2091
	for (m = 0; m < ARRAY_SIZE(masks); ++m) {
2092 2093 2094
		for (i = 0; i < info->hdr->e_shnum; ++i) {
			Elf_Shdr *s = &info->sechdrs[i];
			const char *sname = info->secstrings + s->sh_name;
L
Linus Torvalds 已提交
2095 2096 2097 2098

			if ((s->sh_flags & masks[m][0]) != masks[m][0]
			    || (s->sh_flags & masks[m][1])
			    || s->sh_entsize != ~0UL
2099
			    || !strstarts(sname, ".init"))
L
Linus Torvalds 已提交
2100
				continue;
2101
			s->sh_entsize = (get_offset(mod, &mod->init_size, s, i)
L
Linus Torvalds 已提交
2102
					 | INIT_OFFSET_MASK);
2103
			pr_debug("\t%s\n", sname);
L
Linus Torvalds 已提交
2104
		}
2105 2106 2107
		switch (m) {
		case 0: /* executable */
			mod->init_size = debug_align(mod->init_size);
L
Linus Torvalds 已提交
2108
			mod->init_text_size = mod->init_size;
2109 2110 2111 2112 2113 2114 2115 2116 2117
			break;
		case 1: /* RO: text and ro-data */
			mod->init_size = debug_align(mod->init_size);
			mod->init_ro_size = mod->init_size;
			break;
		case 3: /* whole init */
			mod->init_size = debug_align(mod->init_size);
			break;
		}
L
Linus Torvalds 已提交
2118 2119 2120 2121 2122 2123 2124 2125
	}
}

static void set_license(struct module *mod, const char *license)
{
	if (!license)
		license = "unspecified";

2126
	if (!license_is_gpl_compatible(license)) {
2127
		if (!test_taint(TAINT_PROPRIETARY_MODULE))
2128
			printk(KERN_WARNING "%s: module license '%s' taints "
2129 2130
				"kernel.\n", mod->name, license);
		add_taint_module(mod, TAINT_PROPRIETARY_MODULE);
L
Linus Torvalds 已提交
2131 2132 2133 2134 2135 2136 2137 2138 2139 2140 2141 2142 2143 2144 2145 2146 2147 2148 2149 2150 2151 2152
	}
}

/* Parse tag=value strings from .modinfo section */
static char *next_string(char *string, unsigned long *secsize)
{
	/* Skip non-zero chars */
	while (string[0]) {
		string++;
		if ((*secsize)-- <= 1)
			return NULL;
	}

	/* Skip any zero padding. */
	while (!string[0]) {
		string++;
		if ((*secsize)-- <= 1)
			return NULL;
	}
	return string;
}

2153
static char *get_modinfo(struct load_info *info, const char *tag)
L
Linus Torvalds 已提交
2154 2155 2156
{
	char *p;
	unsigned int taglen = strlen(tag);
2157 2158
	Elf_Shdr *infosec = &info->sechdrs[info->index.info];
	unsigned long size = infosec->sh_size;
L
Linus Torvalds 已提交
2159

2160
	for (p = (char *)infosec->sh_addr; p; p = next_string(p, &size)) {
L
Linus Torvalds 已提交
2161 2162 2163 2164 2165 2166
		if (strncmp(p, tag, taglen) == 0 && p[taglen] == '=')
			return p + taglen + 1;
	}
	return NULL;
}

2167
static void setup_modinfo(struct module *mod, struct load_info *info)
2168 2169 2170 2171 2172 2173
{
	struct module_attribute *attr;
	int i;

	for (i = 0; (attr = modinfo_attrs[i]); i++) {
		if (attr->setup)
2174
			attr->setup(mod, get_modinfo(info, attr->attr.name));
2175 2176 2177
	}
}

2178 2179 2180 2181 2182 2183 2184 2185 2186 2187 2188
static void free_modinfo(struct module *mod)
{
	struct module_attribute *attr;
	int i;

	for (i = 0; (attr = modinfo_attrs[i]); i++) {
		if (attr->free)
			attr->free(mod);
	}
}

L
Linus Torvalds 已提交
2189
#ifdef CONFIG_KALLSYMS
2190 2191 2192 2193 2194 2195

/* lookup symbol in given range of kernel_symbols */
static const struct kernel_symbol *lookup_symbol(const char *name,
	const struct kernel_symbol *start,
	const struct kernel_symbol *stop)
{
2196 2197
	return bsearch(name, start, stop - start,
			sizeof(struct kernel_symbol), cmp_name);
2198 2199
}

2200 2201
static int is_exported(const char *name, unsigned long value,
		       const struct module *mod)
L
Linus Torvalds 已提交
2202
{
2203 2204 2205
	const struct kernel_symbol *ks;
	if (!mod)
		ks = lookup_symbol(name, __start___ksymtab, __stop___ksymtab);
2206
	else
2207 2208
		ks = lookup_symbol(name, mod->syms, mod->syms + mod->num_syms);
	return ks != NULL && ks->value == value;
L
Linus Torvalds 已提交
2209 2210 2211
}

/* As per nm */
2212
static char elf_type(const Elf_Sym *sym, const struct load_info *info)
L
Linus Torvalds 已提交
2213
{
2214 2215
	const Elf_Shdr *sechdrs = info->sechdrs;

L
Linus Torvalds 已提交
2216 2217 2218 2219 2220 2221 2222 2223 2224 2225 2226 2227 2228 2229 2230 2231 2232 2233 2234 2235 2236 2237 2238 2239 2240 2241 2242 2243 2244
	if (ELF_ST_BIND(sym->st_info) == STB_WEAK) {
		if (ELF_ST_TYPE(sym->st_info) == STT_OBJECT)
			return 'v';
		else
			return 'w';
	}
	if (sym->st_shndx == SHN_UNDEF)
		return 'U';
	if (sym->st_shndx == SHN_ABS)
		return 'a';
	if (sym->st_shndx >= SHN_LORESERVE)
		return '?';
	if (sechdrs[sym->st_shndx].sh_flags & SHF_EXECINSTR)
		return 't';
	if (sechdrs[sym->st_shndx].sh_flags & SHF_ALLOC
	    && sechdrs[sym->st_shndx].sh_type != SHT_NOBITS) {
		if (!(sechdrs[sym->st_shndx].sh_flags & SHF_WRITE))
			return 'r';
		else if (sechdrs[sym->st_shndx].sh_flags & ARCH_SHF_SMALL)
			return 'g';
		else
			return 'd';
	}
	if (sechdrs[sym->st_shndx].sh_type == SHT_NOBITS) {
		if (sechdrs[sym->st_shndx].sh_flags & ARCH_SHF_SMALL)
			return 's';
		else
			return 'b';
	}
2245 2246
	if (strstarts(info->secstrings + sechdrs[sym->st_shndx].sh_name,
		      ".debug")) {
L
Linus Torvalds 已提交
2247
		return 'n';
2248
	}
L
Linus Torvalds 已提交
2249 2250 2251
	return '?';
}

2252 2253 2254 2255 2256 2257 2258 2259 2260 2261 2262 2263 2264 2265 2266 2267 2268 2269 2270 2271 2272
static bool is_core_symbol(const Elf_Sym *src, const Elf_Shdr *sechdrs,
                           unsigned int shnum)
{
	const Elf_Shdr *sec;

	if (src->st_shndx == SHN_UNDEF
	    || src->st_shndx >= shnum
	    || !src->st_name)
		return false;

	sec = sechdrs + src->st_shndx;
	if (!(sec->sh_flags & SHF_ALLOC)
#ifndef CONFIG_KALLSYMS_ALL
	    || !(sec->sh_flags & SHF_EXECINSTR)
#endif
	    || (sec->sh_entsize & INIT_OFFSET_MASK))
		return false;

	return true;
}

2273 2274 2275 2276 2277 2278 2279
/*
 * We only allocate and copy the strings needed by the parts of symtab
 * we keep.  This is simple, but has the effect of making multiple
 * copies of duplicates.  We could be more sophisticated, see
 * linux-kernel thread starting with
 * <73defb5e4bca04a6431392cc341112b1@localhost>.
 */
2280
static void layout_symtab(struct module *mod, struct load_info *info)
2281
{
2282 2283
	Elf_Shdr *symsect = info->sechdrs + info->index.sym;
	Elf_Shdr *strsect = info->sechdrs + info->index.str;
2284
	const Elf_Sym *src;
2285
	unsigned int i, nsrc, ndst, strtab_size;
2286 2287 2288 2289

	/* Put symbol section at end of init part of module. */
	symsect->sh_flags |= SHF_ALLOC;
	symsect->sh_entsize = get_offset(mod, &mod->init_size, symsect,
2290
					 info->index.sym) | INIT_OFFSET_MASK;
2291
	pr_debug("\t%s\n", info->secstrings + symsect->sh_name);
2292

2293
	src = (void *)info->hdr + symsect->sh_offset;
2294
	nsrc = symsect->sh_size / sizeof(*src);
2295

2296 2297
	/* Compute total space required for the core symbols' strtab. */
	for (ndst = i = strtab_size = 1; i < nsrc; ++i, ++src)
2298
		if (is_core_symbol(src, info->sechdrs, info->hdr->e_shnum)) {
2299 2300
			strtab_size += strlen(&info->strtab[src->st_name]) + 1;
			ndst++;
2301
		}
2302 2303

	/* Append room for core symbols at end of core part. */
2304
	info->symoffs = ALIGN(mod->core_size, symsect->sh_addralign ?: 1);
2305 2306
	info->stroffs = mod->core_size = info->symoffs + ndst * sizeof(Elf_Sym);
	mod->core_size += strtab_size;
2307

2308 2309 2310
	/* Put string table section at end of init part of module. */
	strsect->sh_flags |= SHF_ALLOC;
	strsect->sh_entsize = get_offset(mod, &mod->init_size, strsect,
2311
					 info->index.str) | INIT_OFFSET_MASK;
2312
	pr_debug("\t%s\n", info->secstrings + strsect->sh_name);
2313 2314
}

2315
static void add_kallsyms(struct module *mod, const struct load_info *info)
L
Linus Torvalds 已提交
2316
{
2317 2318 2319
	unsigned int i, ndst;
	const Elf_Sym *src;
	Elf_Sym *dst;
2320
	char *s;
2321
	Elf_Shdr *symsec = &info->sechdrs[info->index.sym];
L
Linus Torvalds 已提交
2322

2323 2324
	mod->symtab = (void *)symsec->sh_addr;
	mod->num_symtab = symsec->sh_size / sizeof(Elf_Sym);
2325 2326
	/* Make sure we get permanent strtab: don't use info->strtab. */
	mod->strtab = (void *)info->sechdrs[info->index.str].sh_addr;
L
Linus Torvalds 已提交
2327 2328 2329

	/* Set types up while we still have access to sections. */
	for (i = 0; i < mod->num_symtab; i++)
2330
		mod->symtab[i].st_info = elf_type(&mod->symtab[i], info);
2331

2332
	mod->core_symtab = dst = mod->module_core + info->symoffs;
2333
	mod->core_strtab = s = mod->module_core + info->stroffs;
2334 2335
	src = mod->symtab;
	*dst = *src;
2336
	*s++ = 0;
2337
	for (ndst = i = 1; i < mod->num_symtab; ++i, ++src) {
2338
		if (!is_core_symbol(src, info->sechdrs, info->hdr->e_shnum))
2339
			continue;
2340

2341
		dst[ndst] = *src;
2342 2343
		dst[ndst++].st_name = s - mod->core_strtab;
		s += strlcpy(s, &mod->strtab[src->st_name], KSYM_NAME_LEN) + 1;
2344 2345
	}
	mod->core_num_syms = ndst;
L
Linus Torvalds 已提交
2346 2347
}
#else
2348
static inline void layout_symtab(struct module *mod, struct load_info *info)
2349 2350
{
}
2351

2352
static void add_kallsyms(struct module *mod, const struct load_info *info)
L
Linus Torvalds 已提交
2353 2354 2355 2356
{
}
#endif /* CONFIG_KALLSYMS */

2357
static void dynamic_debug_setup(struct _ddebug *debug, unsigned int num)
2358
{
2359 2360
	if (!debug)
		return;
2361 2362 2363 2364 2365
#ifdef CONFIG_DYNAMIC_DEBUG
	if (ddebug_add_module(debug, num, debug->modname))
		printk(KERN_ERR "dynamic debug error adding module: %s\n",
					debug->modname);
#endif
2366
}
2367

2368 2369 2370 2371 2372 2373
static void dynamic_debug_remove(struct _ddebug *debug)
{
	if (debug)
		ddebug_remove_module(debug->modname);
}

2374 2375 2376 2377 2378
void * __weak module_alloc(unsigned long size)
{
	return size == 0 ? NULL : vmalloc_exec(size);
}

2379 2380 2381 2382 2383
static void *module_alloc_update_bounds(unsigned long size)
{
	void *ret = module_alloc(size);

	if (ret) {
2384
		mutex_lock(&module_mutex);
2385 2386 2387 2388 2389
		/* Update module bounds. */
		if ((unsigned long)ret < module_addr_min)
			module_addr_min = (unsigned long)ret;
		if ((unsigned long)ret + size > module_addr_max)
			module_addr_max = (unsigned long)ret + size;
2390
		mutex_unlock(&module_mutex);
2391 2392 2393 2394
	}
	return ret;
}

2395
#ifdef CONFIG_DEBUG_KMEMLEAK
2396 2397
static void kmemleak_load_module(const struct module *mod,
				 const struct load_info *info)
2398 2399 2400 2401
{
	unsigned int i;

	/* only scan the sections containing data */
2402
	kmemleak_scan_area(mod, sizeof(struct module), GFP_KERNEL);
2403

2404 2405 2406
	for (i = 1; i < info->hdr->e_shnum; i++) {
		const char *name = info->secstrings + info->sechdrs[i].sh_name;
		if (!(info->sechdrs[i].sh_flags & SHF_ALLOC))
2407
			continue;
2408
		if (!strstarts(name, ".data") && !strstarts(name, ".bss"))
2409 2410
			continue;

2411 2412
		kmemleak_scan_area((void *)info->sechdrs[i].sh_addr,
				   info->sechdrs[i].sh_size, GFP_KERNEL);
2413 2414 2415
	}
}
#else
2416 2417
static inline void kmemleak_load_module(const struct module *mod,
					const struct load_info *info)
2418 2419 2420 2421
{
}
#endif

2422 2423 2424 2425 2426 2427 2428 2429 2430 2431 2432 2433 2434 2435 2436 2437 2438 2439 2440 2441 2442 2443 2444 2445 2446 2447 2448 2449 2450
#ifdef CONFIG_MODULE_SIG
static int module_sig_check(struct load_info *info,
			    const void *mod, unsigned long *len)
{
	int err = -ENOKEY;
	const unsigned long markerlen = sizeof(MODULE_SIG_STRING) - 1;
	const void *p = mod, *end = mod + *len;

	/* Poor man's memmem. */
	while ((p = memchr(p, MODULE_SIG_STRING[0], end - p))) {
		if (p + markerlen > end)
			break;

		if (memcmp(p, MODULE_SIG_STRING, markerlen) == 0) {
			const void *sig = p + markerlen;
			/* Truncate module up to signature. */
			*len = p - mod;
			err = mod_verify_sig(mod, *len, sig, end - sig);
			break;
		}
		p++;
	}

	if (!err) {
		info->sig_ok = true;
		return 0;
	}

	/* Not having a signature is only an error if we're strict. */
D
David Howells 已提交
2451 2452 2453
	if (err < 0 && fips_enabled)
		panic("Module verification failed with error %d in FIPS mode\n",
		      err);
2454 2455 2456 2457 2458 2459 2460 2461 2462 2463 2464 2465 2466 2467
	if (err == -ENOKEY && !sig_enforce)
		err = 0;

	return err;
}
#else /* !CONFIG_MODULE_SIG */
static int module_sig_check(struct load_info *info,
			    void *mod, unsigned long *len)
{
	return 0;
}
#endif /* !CONFIG_MODULE_SIG */

/* Sets info->hdr, info->len and info->sig_ok. */
2468 2469 2470
static int copy_and_check(struct load_info *info,
			  const void __user *umod, unsigned long len,
			  const char __user *uargs)
2471 2472 2473 2474 2475 2476 2477 2478
{
	int err;
	Elf_Ehdr *hdr;

	if (len < sizeof(*hdr))
		return -ENOEXEC;

	/* Suck in entire file: we'll want most of it. */
2479
	if ((hdr = vmalloc(len)) == NULL)
2480 2481 2482 2483 2484 2485 2486
		return -ENOMEM;

	if (copy_from_user(hdr, umod, len) != 0) {
		err = -EFAULT;
		goto free_hdr;
	}

2487 2488 2489 2490
	err = module_sig_check(info, hdr, &len);
	if (err)
		goto free_hdr;

2491 2492 2493 2494 2495 2496 2497 2498 2499 2500
	/* Sanity checks against insmoding binaries or wrong arch,
	   weird elf version */
	if (memcmp(hdr->e_ident, ELFMAG, SELFMAG) != 0
	    || hdr->e_type != ET_REL
	    || !elf_check_arch(hdr)
	    || hdr->e_shentsize != sizeof(Elf_Shdr)) {
		err = -ENOEXEC;
		goto free_hdr;
	}

2501 2502
	if (hdr->e_shoff >= len ||
	    hdr->e_shnum * sizeof(Elf_Shdr) > len - hdr->e_shoff) {
2503 2504 2505
		err = -ENOEXEC;
		goto free_hdr;
	}
2506

L
Linus Torvalds 已提交
2507 2508
	info->hdr = hdr;
	info->len = len;
2509 2510 2511 2512 2513 2514 2515
	return 0;

free_hdr:
	vfree(hdr);
	return err;
}

2516 2517 2518 2519 2520
static void free_copy(struct load_info *info)
{
	vfree(info->hdr);
}

2521 2522 2523 2524 2525 2526 2527 2528 2529 2530 2531 2532 2533 2534 2535 2536 2537 2538 2539 2540 2541 2542 2543 2544 2545 2546
static int rewrite_section_headers(struct load_info *info)
{
	unsigned int i;

	/* This should always be true, but let's be sure. */
	info->sechdrs[0].sh_addr = 0;

	for (i = 1; i < info->hdr->e_shnum; i++) {
		Elf_Shdr *shdr = &info->sechdrs[i];
		if (shdr->sh_type != SHT_NOBITS
		    && info->len < shdr->sh_offset + shdr->sh_size) {
			printk(KERN_ERR "Module len %lu truncated\n",
			       info->len);
			return -ENOEXEC;
		}

		/* Mark all sections sh_addr with their address in the
		   temporary image. */
		shdr->sh_addr = (size_t)info->hdr + shdr->sh_offset;

#ifndef CONFIG_MODULE_UNLOAD
		/* Don't load .exit sections */
		if (strstarts(info->secstrings+shdr->sh_name, ".exit"))
			shdr->sh_flags &= ~(unsigned long)SHF_ALLOC;
#endif
	}
2547 2548

	/* Track but don't keep modinfo and version sections. */
2549 2550
	info->index.vers = find_sec(info, "__versions");
	info->index.info = find_sec(info, ".modinfo");
2551 2552
	info->sechdrs[info->index.info].sh_flags &= ~(unsigned long)SHF_ALLOC;
	info->sechdrs[info->index.vers].sh_flags &= ~(unsigned long)SHF_ALLOC;
2553 2554 2555
	return 0;
}

L
Linus Torvalds 已提交
2556 2557 2558 2559 2560 2561 2562 2563 2564 2565 2566
/*
 * Set up our basic convenience variables (pointers to section headers,
 * search for module section index etc), and do some basic section
 * verification.
 *
 * Return the temporary module pointer (we'll replace it with the final
 * one when we move the module sections around).
 */
static struct module *setup_load_info(struct load_info *info)
{
	unsigned int i;
2567
	int err;
L
Linus Torvalds 已提交
2568 2569 2570 2571
	struct module *mod;

	/* Set up the convenience variables */
	info->sechdrs = (void *)info->hdr + info->hdr->e_shoff;
2572 2573
	info->secstrings = (void *)info->hdr
		+ info->sechdrs[info->hdr->e_shstrndx].sh_offset;
L
Linus Torvalds 已提交
2574

2575 2576 2577
	err = rewrite_section_headers(info);
	if (err)
		return ERR_PTR(err);
L
Linus Torvalds 已提交
2578

2579 2580
	/* Find internal symbols and strings. */
	for (i = 1; i < info->hdr->e_shnum; i++) {
L
Linus Torvalds 已提交
2581 2582 2583
		if (info->sechdrs[i].sh_type == SHT_SYMTAB) {
			info->index.sym = i;
			info->index.str = info->sechdrs[i].sh_link;
2584 2585 2586
			info->strtab = (char *)info->hdr
				+ info->sechdrs[info->index.str].sh_offset;
			break;
L
Linus Torvalds 已提交
2587 2588 2589
		}
	}

2590
	info->index.mod = find_sec(info, ".gnu.linkonce.this_module");
L
Linus Torvalds 已提交
2591 2592 2593 2594 2595 2596 2597 2598 2599 2600 2601 2602 2603
	if (!info->index.mod) {
		printk(KERN_WARNING "No module found in object\n");
		return ERR_PTR(-ENOEXEC);
	}
	/* This is temporary: point mod into copy of data. */
	mod = (void *)info->sechdrs[info->index.mod].sh_addr;

	if (info->index.sym == 0) {
		printk(KERN_WARNING "%s: module has no symbols (stripped?)\n",
		       mod->name);
		return ERR_PTR(-ENOEXEC);
	}

2604
	info->index.pcpu = find_pcpusec(info);
L
Linus Torvalds 已提交
2605 2606 2607 2608 2609 2610 2611 2612

	/* Check module struct version now, before we try to use module. */
	if (!check_modstruct_version(info->sechdrs, info->index.vers, mod))
		return ERR_PTR(-ENOEXEC);

	return mod;
}

2613
static int check_modinfo(struct module *mod, struct load_info *info)
2614
{
2615
	const char *modmagic = get_modinfo(info, "vermagic");
2616 2617 2618 2619 2620 2621 2622
	int err;

	/* This is allowed: modprobe --force will invalidate it. */
	if (!modmagic) {
		err = try_to_force_load(mod, "bad vermagic");
		if (err)
			return err;
2623
	} else if (!same_magic(modmagic, vermagic, info->index.vers)) {
2624 2625 2626 2627 2628
		printk(KERN_ERR "%s: version magic '%s' should be '%s'\n",
		       mod->name, modmagic, vermagic);
		return -ENOEXEC;
	}

2629 2630 2631
	if (!get_modinfo(info, "intree"))
		add_taint_module(mod, TAINT_OOT_MODULE);

2632
	if (get_modinfo(info, "staging")) {
2633 2634 2635 2636 2637
		add_taint_module(mod, TAINT_CRAP);
		printk(KERN_WARNING "%s: module is from the staging directory,"
		       " the quality is unknown, you have been warned.\n",
		       mod->name);
	}
2638 2639

	/* Set up license info based on the info section */
2640
	set_license(mod, get_modinfo(info, "license"));
2641

2642 2643 2644
	return 0;
}

2645
static void find_module_sections(struct module *mod, struct load_info *info)
2646
{
2647
	mod->kp = section_objs(info, "__param",
2648
			       sizeof(*mod->kp), &mod->num_kp);
2649
	mod->syms = section_objs(info, "__ksymtab",
2650
				 sizeof(*mod->syms), &mod->num_syms);
2651 2652
	mod->crcs = section_addr(info, "__kcrctab");
	mod->gpl_syms = section_objs(info, "__ksymtab_gpl",
2653 2654
				     sizeof(*mod->gpl_syms),
				     &mod->num_gpl_syms);
2655 2656
	mod->gpl_crcs = section_addr(info, "__kcrctab_gpl");
	mod->gpl_future_syms = section_objs(info,
2657 2658 2659
					    "__ksymtab_gpl_future",
					    sizeof(*mod->gpl_future_syms),
					    &mod->num_gpl_future_syms);
2660
	mod->gpl_future_crcs = section_addr(info, "__kcrctab_gpl_future");
2661 2662

#ifdef CONFIG_UNUSED_SYMBOLS
2663
	mod->unused_syms = section_objs(info, "__ksymtab_unused",
2664 2665
					sizeof(*mod->unused_syms),
					&mod->num_unused_syms);
2666 2667
	mod->unused_crcs = section_addr(info, "__kcrctab_unused");
	mod->unused_gpl_syms = section_objs(info, "__ksymtab_unused_gpl",
2668 2669
					    sizeof(*mod->unused_gpl_syms),
					    &mod->num_unused_gpl_syms);
2670
	mod->unused_gpl_crcs = section_addr(info, "__kcrctab_unused_gpl");
2671 2672
#endif
#ifdef CONFIG_CONSTRUCTORS
2673
	mod->ctors = section_objs(info, ".ctors",
2674 2675 2676 2677
				  sizeof(*mod->ctors), &mod->num_ctors);
#endif

#ifdef CONFIG_TRACEPOINTS
2678 2679 2680
	mod->tracepoints_ptrs = section_objs(info, "__tracepoints_ptrs",
					     sizeof(*mod->tracepoints_ptrs),
					     &mod->num_tracepoints);
2681
#endif
2682 2683 2684 2685 2686
#ifdef HAVE_JUMP_LABEL
	mod->jump_entries = section_objs(info, "__jump_table",
					sizeof(*mod->jump_entries),
					&mod->num_jump_entries);
#endif
2687
#ifdef CONFIG_EVENT_TRACING
2688
	mod->trace_events = section_objs(info, "_ftrace_events",
2689 2690 2691 2692 2693 2694 2695 2696 2697
					 sizeof(*mod->trace_events),
					 &mod->num_trace_events);
	/*
	 * This section contains pointers to allocated objects in the trace
	 * code and not scanning it leads to false positives.
	 */
	kmemleak_scan_area(mod->trace_events, sizeof(*mod->trace_events) *
			   mod->num_trace_events, GFP_KERNEL);
#endif
2698 2699 2700 2701 2702 2703 2704 2705 2706 2707 2708 2709
#ifdef CONFIG_TRACING
	mod->trace_bprintk_fmt_start = section_objs(info, "__trace_printk_fmt",
					 sizeof(*mod->trace_bprintk_fmt_start),
					 &mod->num_trace_bprintk_fmt);
	/*
	 * This section contains pointers to allocated objects in the trace
	 * code and not scanning it leads to false positives.
	 */
	kmemleak_scan_area(mod->trace_bprintk_fmt_start,
			   sizeof(*mod->trace_bprintk_fmt_start) *
			   mod->num_trace_bprintk_fmt, GFP_KERNEL);
#endif
2710 2711
#ifdef CONFIG_FTRACE_MCOUNT_RECORD
	/* sechdrs[0].sh_size is always zero */
2712
	mod->ftrace_callsites = section_objs(info, "__mcount_loc",
2713 2714 2715
					     sizeof(*mod->ftrace_callsites),
					     &mod->num_ftrace_callsites);
#endif
2716

2717 2718 2719
	mod->extable = section_objs(info, "__ex_table",
				    sizeof(*mod->extable), &mod->num_exentries);

2720
	if (section_addr(info, "__obsparm"))
2721 2722
		printk(KERN_WARNING "%s: Ignoring obsolete parameters\n",
		       mod->name);
2723 2724 2725

	info->debug = section_objs(info, "__verbose",
				   sizeof(*info->debug), &info->num_debug);
2726 2727
}

2728
static int move_module(struct module *mod, struct load_info *info)
2729 2730 2731 2732 2733 2734 2735 2736 2737 2738 2739 2740 2741
{
	int i;
	void *ptr;

	/* Do the allocs. */
	ptr = module_alloc_update_bounds(mod->core_size);
	/*
	 * The pointer to this block is stored in the module structure
	 * which is inside the block. Just mark it as not being a
	 * leak.
	 */
	kmemleak_not_leak(ptr);
	if (!ptr)
2742
		return -ENOMEM;
2743 2744 2745 2746 2747 2748 2749 2750 2751 2752 2753 2754 2755 2756

	memset(ptr, 0, mod->core_size);
	mod->module_core = ptr;

	ptr = module_alloc_update_bounds(mod->init_size);
	/*
	 * The pointer to this block is stored in the module structure
	 * which is inside the block. This block doesn't need to be
	 * scanned as it contains data and code that will be freed
	 * after the module is initialized.
	 */
	kmemleak_ignore(ptr);
	if (!ptr && mod->init_size) {
		module_free(mod, mod->module_core);
2757
		return -ENOMEM;
2758 2759 2760 2761 2762
	}
	memset(ptr, 0, mod->init_size);
	mod->module_init = ptr;

	/* Transfer each section which specifies SHF_ALLOC */
2763
	pr_debug("final section addresses:\n");
2764
	for (i = 0; i < info->hdr->e_shnum; i++) {
2765
		void *dest;
2766
		Elf_Shdr *shdr = &info->sechdrs[i];
2767

2768
		if (!(shdr->sh_flags & SHF_ALLOC))
2769 2770
			continue;

2771
		if (shdr->sh_entsize & INIT_OFFSET_MASK)
2772
			dest = mod->module_init
2773
				+ (shdr->sh_entsize & ~INIT_OFFSET_MASK);
2774
		else
2775
			dest = mod->module_core + shdr->sh_entsize;
2776

2777 2778
		if (shdr->sh_type != SHT_NOBITS)
			memcpy(dest, (void *)shdr->sh_addr, shdr->sh_size);
2779
		/* Update sh_addr to point to copy in image. */
2780
		shdr->sh_addr = (unsigned long)dest;
2781 2782
		pr_debug("\t0x%lx %s\n",
			 (long)shdr->sh_addr, info->secstrings + shdr->sh_name);
2783
	}
2784 2785

	return 0;
2786 2787
}

2788
static int check_module_license_and_versions(struct module *mod)
2789 2790 2791 2792 2793 2794 2795 2796 2797 2798 2799 2800 2801
{
	/*
	 * ndiswrapper is under GPL by itself, but loads proprietary modules.
	 * Don't use add_taint_module(), as it would prevent ndiswrapper from
	 * using GPL-only symbols it needs.
	 */
	if (strcmp(mod->name, "ndiswrapper") == 0)
		add_taint(TAINT_PROPRIETARY_MODULE);

	/* driverloader was caught wrongly pretending to be under GPL */
	if (strcmp(mod->name, "driverloader") == 0)
		add_taint_module(mod, TAINT_PROPRIETARY_MODULE);

2802 2803 2804 2805
	/* lve claims to be GPL but upstream won't provide source */
	if (strcmp(mod->name, "lve") == 0)
		add_taint_module(mod, TAINT_PROPRIETARY_MODULE);

2806 2807 2808 2809 2810 2811 2812 2813 2814 2815 2816 2817 2818 2819 2820 2821 2822 2823 2824 2825 2826 2827 2828 2829 2830 2831 2832 2833 2834 2835 2836 2837 2838 2839 2840 2841 2842 2843 2844
#ifdef CONFIG_MODVERSIONS
	if ((mod->num_syms && !mod->crcs)
	    || (mod->num_gpl_syms && !mod->gpl_crcs)
	    || (mod->num_gpl_future_syms && !mod->gpl_future_crcs)
#ifdef CONFIG_UNUSED_SYMBOLS
	    || (mod->num_unused_syms && !mod->unused_crcs)
	    || (mod->num_unused_gpl_syms && !mod->unused_gpl_crcs)
#endif
		) {
		return try_to_force_load(mod,
					 "no versions for exported symbols");
	}
#endif
	return 0;
}

static void flush_module_icache(const struct module *mod)
{
	mm_segment_t old_fs;

	/* flush the icache in correct context */
	old_fs = get_fs();
	set_fs(KERNEL_DS);

	/*
	 * Flush the instruction cache, since we've played with text.
	 * Do it before processing of module parameters, so the module
	 * can provide parameter accessor functions of its own.
	 */
	if (mod->module_init)
		flush_icache_range((unsigned long)mod->module_init,
				   (unsigned long)mod->module_init
				   + mod->init_size);
	flush_icache_range((unsigned long)mod->module_core,
			   (unsigned long)mod->module_core + mod->core_size);

	set_fs(old_fs);
}

2845 2846 2847 2848 2849 2850 2851 2852
int __weak module_frob_arch_sections(Elf_Ehdr *hdr,
				     Elf_Shdr *sechdrs,
				     char *secstrings,
				     struct module *mod)
{
	return 0;
}

2853
static struct module *layout_and_allocate(struct load_info *info)
L
Linus Torvalds 已提交
2854
{
2855
	/* Module within temporary copy. */
L
Linus Torvalds 已提交
2856
	struct module *mod;
2857
	Elf_Shdr *pcpusec;
2858
	int err;
2859

2860 2861 2862
	mod = setup_load_info(info);
	if (IS_ERR(mod))
		return mod;
L
Linus Torvalds 已提交
2863

2864
	err = check_modinfo(mod, info);
2865 2866
	if (err)
		return ERR_PTR(err);
L
Linus Torvalds 已提交
2867 2868

	/* Allow arches to frob section contents and sizes.  */
2869 2870
	err = module_frob_arch_sections(info->hdr, info->sechdrs,
					info->secstrings, mod);
L
Linus Torvalds 已提交
2871
	if (err < 0)
2872
		goto out;
L
Linus Torvalds 已提交
2873

2874 2875
	pcpusec = &info->sechdrs[info->index.pcpu];
	if (pcpusec->sh_size) {
L
Linus Torvalds 已提交
2876
		/* We have a special allocation for this section. */
2877 2878
		err = percpu_modalloc(mod,
				      pcpusec->sh_size, pcpusec->sh_addralign);
2879
		if (err)
2880
			goto out;
2881
		pcpusec->sh_flags &= ~(unsigned long)SHF_ALLOC;
L
Linus Torvalds 已提交
2882 2883 2884 2885 2886
	}

	/* Determine total sizes, and put offsets in sh_entsize.  For now
	   this is done generically; there doesn't appear to be any
	   special cases for the architectures. */
2887 2888
	layout_sections(mod, info);
	layout_symtab(mod, info);
L
Linus Torvalds 已提交
2889

2890
	/* Allocate and move to the final place */
2891
	err = move_module(mod, info);
2892
	if (err)
2893
		goto free_percpu;
2894 2895 2896

	/* Module has been copied to its final place now: return it. */
	mod = (void *)info->sechdrs[info->index.mod].sh_addr;
2897
	kmemleak_load_module(mod, info);
2898 2899 2900 2901
	return mod;

free_percpu:
	percpu_modfree(mod);
2902
out:
2903 2904 2905 2906 2907 2908 2909 2910 2911 2912 2913
	return ERR_PTR(err);
}

/* mod is no longer valid after this! */
static void module_deallocate(struct module *mod, struct load_info *info)
{
	percpu_modfree(mod);
	module_free(mod, mod->module_init);
	module_free(mod, mod->module_core);
}

2914 2915 2916 2917 2918 2919 2920
int __weak module_finalize(const Elf_Ehdr *hdr,
			   const Elf_Shdr *sechdrs,
			   struct module *me)
{
	return 0;
}

2921 2922
static int post_relocation(struct module *mod, const struct load_info *info)
{
2923
	/* Sort exception table now relocations are done. */
2924 2925 2926 2927 2928 2929
	sort_extable(mod->extable, mod->extable + mod->num_exentries);

	/* Copy relocated percpu area over. */
	percpu_modcopy(mod, (void *)info->sechdrs[info->index.pcpu].sh_addr,
		       info->sechdrs[info->index.pcpu].sh_size);

2930
	/* Setup kallsyms-specific fields. */
2931 2932 2933 2934 2935 2936
	add_kallsyms(mod, info);

	/* Arch-specific module finalizing. */
	return module_finalize(info->hdr, info->sechdrs, mod);
}

2937 2938 2939 2940 2941 2942 2943 2944 2945 2946 2947 2948 2949 2950
/* Is this module of this name done loading?  No locks held. */
static bool finished_loading(const char *name)
{
	struct module *mod;
	bool ret;

	mutex_lock(&module_mutex);
	mod = find_module(name);
	ret = !mod || mod->state != MODULE_STATE_COMING;
	mutex_unlock(&module_mutex);

	return ret;
}

2951 2952
/* Allocate and load the module: note that size of section 0 is always
   zero, and we rely on this for optional sections. */
2953
static struct module *load_module(void __user *umod,
2954 2955 2956 2957
				  unsigned long len,
				  const char __user *uargs)
{
	struct load_info info = { NULL, };
2958
	struct module *mod, *old;
2959 2960
	long err;

2961
	pr_debug("load_module: umod=%p, len=%lu, uargs=%p\n",
2962 2963 2964 2965 2966 2967 2968 2969 2970
	       umod, len, uargs);

	/* Copy in the blobs from userspace, check they are vaguely sane. */
	err = copy_and_check(&info, umod, len, uargs);
	if (err)
		return ERR_PTR(err);

	/* Figure out module layout, and allocate all the memory. */
	mod = layout_and_allocate(&info);
2971 2972
	if (IS_ERR(mod)) {
		err = PTR_ERR(mod);
2973
		goto free_copy;
L
Linus Torvalds 已提交
2974 2975
	}

2976 2977 2978 2979 2980 2981
#ifdef CONFIG_MODULE_SIG
	mod->sig_ok = info.sig_ok;
	if (!mod->sig_ok)
		add_taint_module(mod, TAINT_FORCED_MODULE);
#endif

2982
	/* Now module is in final location, initialize linked lists, etc. */
2983 2984
	err = module_unload_init(mod);
	if (err)
2985
		goto free_module;
L
Linus Torvalds 已提交
2986

2987 2988
	/* Now we've got everything in the final locations, we can
	 * find optional sections. */
2989
	find_module_sections(mod, &info);
2990

2991
	err = check_module_license_and_versions(mod);
2992 2993
	if (err)
		goto free_unload;
2994

2995
	/* Set up MODINFO_ATTR fields */
2996
	setup_modinfo(mod, &info);
2997

L
Linus Torvalds 已提交
2998
	/* Fix up syms, so that st_value is a pointer to location. */
2999
	err = simplify_symbols(mod, &info);
L
Linus Torvalds 已提交
3000
	if (err < 0)
3001
		goto free_modinfo;
L
Linus Torvalds 已提交
3002

3003
	err = apply_relocations(mod, &info);
3004
	if (err < 0)
3005
		goto free_modinfo;
L
Linus Torvalds 已提交
3006

3007
	err = post_relocation(mod, &info);
L
Linus Torvalds 已提交
3008
	if (err < 0)
3009
		goto free_modinfo;
L
Linus Torvalds 已提交
3010

3011
	flush_module_icache(mod);
3012

3013 3014 3015 3016 3017 3018
	/* Now copy in args */
	mod->args = strndup_user(uargs, ~0UL >> 1);
	if (IS_ERR(mod->args)) {
		err = PTR_ERR(mod->args);
		goto free_arch_cleanup;
	}
3019

3020
	/* Mark state as coming so strong_try_module_get() ignores us. */
3021 3022
	mod->state = MODULE_STATE_COMING;

3023
	/* Now sew it into the lists so we can get lockdep and oops
L
Lucas De Marchi 已提交
3024
	 * info during argument parsing.  No one should access us, since
3025 3026 3027 3028 3029
	 * strong_try_module_get() will fail.
	 * lockdep/oops can run asynchronous, so use the RCU list insertion
	 * function to insert in a way safe to concurrent readers.
	 * The mutex protects against concurrent writers.
	 */
3030
again:
3031
	mutex_lock(&module_mutex);
3032 3033 3034 3035 3036 3037 3038 3039 3040 3041
	if ((old = find_module(mod->name)) != NULL) {
		if (old->state == MODULE_STATE_COMING) {
			/* Wait in case it fails to load. */
			mutex_unlock(&module_mutex);
			err = wait_event_interruptible(module_wq,
					       finished_loading(mod->name));
			if (err)
				goto free_arch_cleanup;
			goto again;
		}
3042
		err = -EEXIST;
3043
		goto unlock;
3044 3045
	}

3046
	/* This has to be done once we're sure module name is unique. */
3047
	dynamic_debug_setup(info.debug, info.num_debug);
3048

3049 3050 3051
	/* Find duplicate symbols */
	err = verify_export_symbols(mod);
	if (err < 0)
3052
		goto ddebug;
3053

3054
	module_bug_finalize(info.hdr, info.sechdrs, mod);
3055
	list_add_rcu(&mod->list, &modules);
3056
	mutex_unlock(&module_mutex);
3057

3058
	/* Module is ready to execute: parsing args may do that. */
3059
	err = parse_args(mod->name, mod->args, mod->kp, mod->num_kp,
3060
			 -32768, 32767, &ddebug_dyndbg_module_param_cb);
L
Linus Torvalds 已提交
3061
	if (err < 0)
3062
		goto unlink;
L
Linus Torvalds 已提交
3063

3064
	/* Link in to syfs. */
R
Rusty Russell 已提交
3065
	err = mod_sysfs_setup(mod, &info, mod->kp, mod->num_kp);
L
Linus Torvalds 已提交
3066
	if (err < 0)
3067
		goto unlink;
3068

3069
	/* Get rid of temporary copy. */
3070
	free_copy(&info);
L
Linus Torvalds 已提交
3071 3072

	/* Done! */
3073
	trace_module_load(mod);
L
Linus Torvalds 已提交
3074 3075
	return mod;

3076
 unlink:
3077
	mutex_lock(&module_mutex);
3078 3079
	/* Unlink carefully: kallsyms could be walking list. */
	list_del_rcu(&mod->list);
3080
	module_bug_cleanup(mod);
3081
	wake_up_all(&module_wq);
3082
 ddebug:
3083
	dynamic_debug_remove(info.debug);
3084
 unlock:
3085
	mutex_unlock(&module_mutex);
3086
	synchronize_sched();
3087 3088
	kfree(mod->args);
 free_arch_cleanup:
L
Linus Torvalds 已提交
3089
	module_arch_cleanup(mod);
3090
 free_modinfo:
3091
	free_modinfo(mod);
3092
 free_unload:
L
Linus Torvalds 已提交
3093
	module_unload_free(mod);
3094 3095 3096 3097
 free_module:
	module_deallocate(mod, &info);
 free_copy:
	free_copy(&info);
3098
	return ERR_PTR(err);
L
Linus Torvalds 已提交
3099 3100
}

3101 3102 3103 3104 3105 3106 3107 3108 3109 3110 3111
/* Call module constructors. */
static void do_mod_ctors(struct module *mod)
{
#ifdef CONFIG_CONSTRUCTORS
	unsigned long i;

	for (i = 0; i < mod->num_ctors; i++)
		mod->ctors[i]();
#endif
}

L
Linus Torvalds 已提交
3112
/* This is where the real work happens */
3113 3114
SYSCALL_DEFINE3(init_module, void __user *, umod,
		unsigned long, len, const char __user *, uargs)
L
Linus Torvalds 已提交
3115 3116 3117 3118 3119
{
	struct module *mod;
	int ret = 0;

	/* Must have permission */
3120
	if (!capable(CAP_SYS_MODULE) || modules_disabled)
L
Linus Torvalds 已提交
3121 3122 3123 3124
		return -EPERM;

	/* Do all the hard work */
	mod = load_module(umod, len, uargs);
3125
	if (IS_ERR(mod))
L
Linus Torvalds 已提交
3126 3127
		return PTR_ERR(mod);

3128 3129
	blocking_notifier_call_chain(&module_notify_list,
			MODULE_STATE_COMING, mod);
L
Linus Torvalds 已提交
3130

3131 3132 3133 3134 3135 3136 3137 3138 3139 3140 3141 3142
	/* Set RO and NX regions for core */
	set_section_ro_nx(mod->module_core,
				mod->core_text_size,
				mod->core_ro_size,
				mod->core_size);

	/* Set RO and NX regions for init */
	set_section_ro_nx(mod->module_init,
				mod->init_text_size,
				mod->init_ro_size,
				mod->init_size);

3143
	do_mod_ctors(mod);
L
Linus Torvalds 已提交
3144 3145
	/* Start the module */
	if (mod->init != NULL)
3146
		ret = do_one_initcall(mod->init);
L
Linus Torvalds 已提交
3147 3148 3149 3150
	if (ret < 0) {
		/* Init routine failed: abort.  Try to protect us from
                   buggy refcounters. */
		mod->state = MODULE_STATE_GOING;
3151
		synchronize_sched();
3152
		module_put(mod);
3153 3154
		blocking_notifier_call_chain(&module_notify_list,
					     MODULE_STATE_GOING, mod);
3155
		free_module(mod);
3156
		wake_up_all(&module_wq);
L
Linus Torvalds 已提交
3157 3158
		return ret;
	}
3159
	if (ret > 0) {
3160 3161 3162
		printk(KERN_WARNING
"%s: '%s'->init suspiciously returned %d, it should follow 0/-E convention\n"
"%s: loading module anyway...\n",
3163 3164 3165 3166
		       __func__, mod->name, ret,
		       __func__);
		dump_stack();
	}
L
Linus Torvalds 已提交
3167

3168
	/* Now it's a first class citizen! */
L
Linus Torvalds 已提交
3169
	mod->state = MODULE_STATE_LIVE;
3170 3171
	blocking_notifier_call_chain(&module_notify_list,
				     MODULE_STATE_LIVE, mod);
3172

3173 3174 3175
	/* We need to finish all async code before the module init sequence is done */
	async_synchronize_full();

3176
	mutex_lock(&module_mutex);
L
Linus Torvalds 已提交
3177 3178
	/* Drop initial reference. */
	module_put(mod);
3179
	trim_init_extable(mod);
3180 3181 3182
#ifdef CONFIG_KALLSYMS
	mod->num_symtab = mod->core_num_syms;
	mod->symtab = mod->core_symtab;
3183
	mod->strtab = mod->core_strtab;
3184
#endif
3185
	unset_module_init_ro_nx(mod);
L
Linus Torvalds 已提交
3186 3187 3188
	module_free(mod, mod->module_init);
	mod->module_init = NULL;
	mod->init_size = 0;
3189
	mod->init_ro_size = 0;
L
Linus Torvalds 已提交
3190
	mod->init_text_size = 0;
3191
	mutex_unlock(&module_mutex);
3192
	wake_up_all(&module_wq);
L
Linus Torvalds 已提交
3193 3194 3195 3196 3197 3198 3199 3200 3201 3202 3203 3204 3205 3206 3207 3208

	return 0;
}

static inline int within(unsigned long addr, void *start, unsigned long size)
{
	return ((void *)addr >= start && (void *)addr < start + size);
}

#ifdef CONFIG_KALLSYMS
/*
 * This ignores the intensely annoying "mapping symbols" found
 * in ARM ELF files: $a, $t and $d.
 */
static inline int is_arm_mapping_symbol(const char *str)
{
3209
	return str[0] == '$' && strchr("atd", str[1])
L
Linus Torvalds 已提交
3210 3211 3212 3213 3214 3215 3216 3217 3218 3219 3220 3221
	       && (str[2] == '\0' || str[2] == '.');
}

static const char *get_ksymbol(struct module *mod,
			       unsigned long addr,
			       unsigned long *size,
			       unsigned long *offset)
{
	unsigned int i, best = 0;
	unsigned long nextval;

	/* At worse, next value is at end of module */
3222
	if (within_module_init(addr, mod))
L
Linus Torvalds 已提交
3223
		nextval = (unsigned long)mod->module_init+mod->init_text_size;
3224
	else
L
Linus Torvalds 已提交
3225 3226
		nextval = (unsigned long)mod->module_core+mod->core_text_size;

L
Lucas De Marchi 已提交
3227
	/* Scan for closest preceding symbol, and next symbol. (ELF
3228
	   starts real symbols at 1). */
L
Linus Torvalds 已提交
3229 3230 3231 3232 3233 3234 3235 3236 3237 3238 3239 3240 3241 3242 3243 3244 3245 3246 3247 3248 3249
	for (i = 1; i < mod->num_symtab; i++) {
		if (mod->symtab[i].st_shndx == SHN_UNDEF)
			continue;

		/* We ignore unnamed symbols: they're uninformative
		 * and inserted at a whim. */
		if (mod->symtab[i].st_value <= addr
		    && mod->symtab[i].st_value > mod->symtab[best].st_value
		    && *(mod->strtab + mod->symtab[i].st_name) != '\0'
		    && !is_arm_mapping_symbol(mod->strtab + mod->symtab[i].st_name))
			best = i;
		if (mod->symtab[i].st_value > addr
		    && mod->symtab[i].st_value < nextval
		    && *(mod->strtab + mod->symtab[i].st_name) != '\0'
		    && !is_arm_mapping_symbol(mod->strtab + mod->symtab[i].st_name))
			nextval = mod->symtab[i].st_value;
	}

	if (!best)
		return NULL;

3250 3251 3252 3253
	if (size)
		*size = nextval - mod->symtab[best].st_value;
	if (offset)
		*offset = addr - mod->symtab[best].st_value;
L
Linus Torvalds 已提交
3254 3255 3256
	return mod->strtab + mod->symtab[best].st_name;
}

3257 3258
/* For kallsyms to ask for address resolution.  NULL means not found.  Careful
 * not to lock to avoid deadlock on oopses, simply disable preemption. */
3259
const char *module_address_lookup(unsigned long addr,
3260 3261 3262 3263
			    unsigned long *size,
			    unsigned long *offset,
			    char **modname,
			    char *namebuf)
L
Linus Torvalds 已提交
3264 3265
{
	struct module *mod;
3266
	const char *ret = NULL;
L
Linus Torvalds 已提交
3267

3268
	preempt_disable();
3269
	list_for_each_entry_rcu(mod, &modules, list) {
3270 3271
		if (within_module_init(addr, mod) ||
		    within_module_core(addr, mod)) {
3272 3273
			if (modname)
				*modname = mod->name;
3274 3275
			ret = get_ksymbol(mod, addr, size, offset);
			break;
L
Linus Torvalds 已提交
3276 3277
		}
	}
3278 3279 3280 3281 3282
	/* Make a copy in here where it's safe */
	if (ret) {
		strncpy(namebuf, ret, KSYM_NAME_LEN - 1);
		ret = namebuf;
	}
3283
	preempt_enable();
3284
	return ret;
L
Linus Torvalds 已提交
3285 3286
}

3287 3288 3289 3290
int lookup_module_symbol_name(unsigned long addr, char *symname)
{
	struct module *mod;

3291
	preempt_disable();
3292
	list_for_each_entry_rcu(mod, &modules, list) {
3293 3294
		if (within_module_init(addr, mod) ||
		    within_module_core(addr, mod)) {
3295 3296 3297 3298 3299
			const char *sym;

			sym = get_ksymbol(mod, addr, NULL, NULL);
			if (!sym)
				goto out;
3300
			strlcpy(symname, sym, KSYM_NAME_LEN);
3301
			preempt_enable();
3302 3303 3304 3305
			return 0;
		}
	}
out:
3306
	preempt_enable();
3307 3308 3309
	return -ERANGE;
}

3310 3311 3312 3313 3314
int lookup_module_symbol_attrs(unsigned long addr, unsigned long *size,
			unsigned long *offset, char *modname, char *name)
{
	struct module *mod;

3315
	preempt_disable();
3316
	list_for_each_entry_rcu(mod, &modules, list) {
3317 3318
		if (within_module_init(addr, mod) ||
		    within_module_core(addr, mod)) {
3319 3320 3321 3322 3323 3324
			const char *sym;

			sym = get_ksymbol(mod, addr, size, offset);
			if (!sym)
				goto out;
			if (modname)
3325
				strlcpy(modname, mod->name, MODULE_NAME_LEN);
3326
			if (name)
3327
				strlcpy(name, sym, KSYM_NAME_LEN);
3328
			preempt_enable();
3329 3330 3331 3332
			return 0;
		}
	}
out:
3333
	preempt_enable();
3334 3335 3336
	return -ERANGE;
}

3337 3338
int module_get_kallsym(unsigned int symnum, unsigned long *value, char *type,
			char *name, char *module_name, int *exported)
L
Linus Torvalds 已提交
3339 3340 3341
{
	struct module *mod;

3342
	preempt_disable();
3343
	list_for_each_entry_rcu(mod, &modules, list) {
L
Linus Torvalds 已提交
3344 3345 3346
		if (symnum < mod->num_symtab) {
			*value = mod->symtab[symnum].st_value;
			*type = mod->symtab[symnum].st_info;
3347
			strlcpy(name, mod->strtab + mod->symtab[symnum].st_name,
3348 3349
				KSYM_NAME_LEN);
			strlcpy(module_name, mod->name, MODULE_NAME_LEN);
3350
			*exported = is_exported(name, *value, mod);
3351
			preempt_enable();
3352
			return 0;
L
Linus Torvalds 已提交
3353 3354 3355
		}
		symnum -= mod->num_symtab;
	}
3356
	preempt_enable();
3357
	return -ERANGE;
L
Linus Torvalds 已提交
3358 3359 3360 3361 3362 3363 3364
}

static unsigned long mod_find_symname(struct module *mod, const char *name)
{
	unsigned int i;

	for (i = 0; i < mod->num_symtab; i++)
3365 3366
		if (strcmp(name, mod->strtab+mod->symtab[i].st_name) == 0 &&
		    mod->symtab[i].st_info != 'U')
L
Linus Torvalds 已提交
3367 3368 3369 3370 3371 3372 3373 3374 3375 3376 3377 3378
			return mod->symtab[i].st_value;
	return 0;
}

/* Look for this name: can be of form module:name. */
unsigned long module_kallsyms_lookup_name(const char *name)
{
	struct module *mod;
	char *colon;
	unsigned long ret = 0;

	/* Don't lock: we're in enough trouble already. */
3379
	preempt_disable();
L
Linus Torvalds 已提交
3380 3381 3382 3383 3384 3385
	if ((colon = strchr(name, ':')) != NULL) {
		*colon = '\0';
		if ((mod = find_module(name)) != NULL)
			ret = mod_find_symname(mod, colon+1);
		*colon = ':';
	} else {
3386
		list_for_each_entry_rcu(mod, &modules, list)
L
Linus Torvalds 已提交
3387 3388 3389
			if ((ret = mod_find_symname(mod, name)) != 0)
				break;
	}
3390
	preempt_enable();
L
Linus Torvalds 已提交
3391 3392
	return ret;
}
3393 3394 3395 3396 3397 3398 3399 3400 3401 3402 3403 3404 3405 3406 3407 3408 3409 3410 3411

int module_kallsyms_on_each_symbol(int (*fn)(void *, const char *,
					     struct module *, unsigned long),
				   void *data)
{
	struct module *mod;
	unsigned int i;
	int ret;

	list_for_each_entry(mod, &modules, list) {
		for (i = 0; i < mod->num_symtab; i++) {
			ret = fn(data, mod->strtab + mod->symtab[i].st_name,
				 mod, mod->symtab[i].st_value);
			if (ret != 0)
				return ret;
		}
	}
	return 0;
}
L
Linus Torvalds 已提交
3412 3413
#endif /* CONFIG_KALLSYMS */

3414
static char *module_flags(struct module *mod, char *buf)
3415 3416 3417
{
	int bx = 0;

3418 3419 3420
	if (mod->taints ||
	    mod->state == MODULE_STATE_GOING ||
	    mod->state == MODULE_STATE_COMING) {
3421
		buf[bx++] = '(';
3422
		bx += module_flags_taint(mod, buf + bx);
3423 3424 3425 3426 3427 3428
		/* Show a - for module-is-being-unloaded */
		if (mod->state == MODULE_STATE_GOING)
			buf[bx++] = '-';
		/* Show a + for module-is-being-loaded */
		if (mod->state == MODULE_STATE_COMING)
			buf[bx++] = '+';
3429 3430 3431 3432 3433 3434 3435
		buf[bx++] = ')';
	}
	buf[bx] = '\0';

	return buf;
}

3436 3437 3438 3439 3440 3441 3442 3443 3444 3445 3446 3447 3448 3449 3450 3451 3452 3453
#ifdef CONFIG_PROC_FS
/* Called by the /proc file system to return a list of modules. */
static void *m_start(struct seq_file *m, loff_t *pos)
{
	mutex_lock(&module_mutex);
	return seq_list_start(&modules, *pos);
}

static void *m_next(struct seq_file *m, void *p, loff_t *pos)
{
	return seq_list_next(p, &modules, pos);
}

static void m_stop(struct seq_file *m, void *p)
{
	mutex_unlock(&module_mutex);
}

L
Linus Torvalds 已提交
3454 3455 3456
static int m_show(struct seq_file *m, void *p)
{
	struct module *mod = list_entry(p, struct module, list);
3457 3458
	char buf[8];

3459
	seq_printf(m, "%s %u",
L
Linus Torvalds 已提交
3460 3461 3462 3463 3464 3465 3466 3467 3468
		   mod->name, mod->init_size + mod->core_size);
	print_unload_info(m, mod);

	/* Informative for users. */
	seq_printf(m, " %s",
		   mod->state == MODULE_STATE_GOING ? "Unloading":
		   mod->state == MODULE_STATE_COMING ? "Loading":
		   "Live");
	/* Used by oprofile and other similar tools. */
3469
	seq_printf(m, " 0x%pK", mod->module_core);
L
Linus Torvalds 已提交
3470

3471 3472
	/* Taints info */
	if (mod->taints)
3473
		seq_printf(m, " %s", module_flags(mod, buf));
3474

L
Linus Torvalds 已提交
3475 3476 3477 3478 3479 3480 3481 3482 3483
	seq_printf(m, "\n");
	return 0;
}

/* Format: modulename size refcount deps address

   Where refcount is a number or -, and deps is a comma-separated list
   of depends or -.
*/
3484
static const struct seq_operations modules_op = {
L
Linus Torvalds 已提交
3485 3486 3487 3488 3489 3490
	.start	= m_start,
	.next	= m_next,
	.stop	= m_stop,
	.show	= m_show
};

3491 3492 3493 3494 3495 3496 3497 3498 3499 3500 3501 3502 3503 3504 3505 3506 3507 3508 3509 3510
static int modules_open(struct inode *inode, struct file *file)
{
	return seq_open(file, &modules_op);
}

static const struct file_operations proc_modules_operations = {
	.open		= modules_open,
	.read		= seq_read,
	.llseek		= seq_lseek,
	.release	= seq_release,
};

static int __init proc_modules_init(void)
{
	proc_create("modules", 0, NULL, &proc_modules_operations);
	return 0;
}
module_init(proc_modules_init);
#endif

L
Linus Torvalds 已提交
3511 3512 3513 3514 3515 3516
/* Given an address, look for it in the module exception tables. */
const struct exception_table_entry *search_module_extables(unsigned long addr)
{
	const struct exception_table_entry *e = NULL;
	struct module *mod;

3517
	preempt_disable();
3518
	list_for_each_entry_rcu(mod, &modules, list) {
L
Linus Torvalds 已提交
3519 3520
		if (mod->num_exentries == 0)
			continue;
3521

L
Linus Torvalds 已提交
3522 3523 3524 3525 3526 3527
		e = search_extable(mod->extable,
				   mod->extable + mod->num_exentries - 1,
				   addr);
		if (e)
			break;
	}
3528
	preempt_enable();
L
Linus Torvalds 已提交
3529 3530

	/* Now, if we found one, we are running inside it now, hence
3531
	   we cannot unload the module, hence no refcnt needed. */
L
Linus Torvalds 已提交
3532 3533 3534
	return e;
}

3535
/*
R
Rusty Russell 已提交
3536 3537 3538 3539 3540
 * is_module_address - is this address inside a module?
 * @addr: the address to check.
 *
 * See is_module_text_address() if you simply want to see if the address
 * is code (not data).
3541
 */
R
Rusty Russell 已提交
3542
bool is_module_address(unsigned long addr)
3543
{
R
Rusty Russell 已提交
3544
	bool ret;
3545

3546
	preempt_disable();
R
Rusty Russell 已提交
3547
	ret = __module_address(addr) != NULL;
3548
	preempt_enable();
3549

R
Rusty Russell 已提交
3550
	return ret;
3551 3552
}

R
Rusty Russell 已提交
3553 3554 3555 3556 3557 3558 3559
/*
 * __module_address - get the module which contains an address.
 * @addr: the address.
 *
 * Must be called with preempt disabled or module mutex held so that
 * module doesn't get freed during this.
 */
3560
struct module *__module_address(unsigned long addr)
L
Linus Torvalds 已提交
3561 3562 3563
{
	struct module *mod;

3564 3565 3566
	if (addr < module_addr_min || addr > module_addr_max)
		return NULL;

3567
	list_for_each_entry_rcu(mod, &modules, list)
R
Rusty Russell 已提交
3568 3569
		if (within_module_core(addr, mod)
		    || within_module_init(addr, mod))
L
Linus Torvalds 已提交
3570 3571 3572
			return mod;
	return NULL;
}
3573
EXPORT_SYMBOL_GPL(__module_address);
L
Linus Torvalds 已提交
3574

R
Rusty Russell 已提交
3575 3576 3577 3578 3579 3580 3581 3582 3583 3584 3585 3586 3587 3588 3589 3590 3591 3592 3593 3594 3595 3596 3597 3598 3599 3600 3601 3602 3603 3604 3605 3606 3607 3608 3609 3610 3611
/*
 * is_module_text_address - is this address inside module code?
 * @addr: the address to check.
 *
 * See is_module_address() if you simply want to see if the address is
 * anywhere in a module.  See kernel_text_address() for testing if an
 * address corresponds to kernel or module code.
 */
bool is_module_text_address(unsigned long addr)
{
	bool ret;

	preempt_disable();
	ret = __module_text_address(addr) != NULL;
	preempt_enable();

	return ret;
}

/*
 * __module_text_address - get the module whose code contains an address.
 * @addr: the address.
 *
 * Must be called with preempt disabled or module mutex held so that
 * module doesn't get freed during this.
 */
struct module *__module_text_address(unsigned long addr)
{
	struct module *mod = __module_address(addr);
	if (mod) {
		/* Make sure it's within the text section. */
		if (!within(addr, mod->module_init, mod->init_text_size)
		    && !within(addr, mod->module_core, mod->core_text_size))
			mod = NULL;
	}
	return mod;
}
3612
EXPORT_SYMBOL_GPL(__module_text_address);
R
Rusty Russell 已提交
3613

L
Linus Torvalds 已提交
3614 3615 3616 3617
/* Don't grab lock, we're oopsing. */
void print_modules(void)
{
	struct module *mod;
3618
	char buf[8];
L
Linus Torvalds 已提交
3619

3620
	printk(KERN_DEFAULT "Modules linked in:");
3621 3622 3623
	/* Most callers should already have preempt disabled, but make sure */
	preempt_disable();
	list_for_each_entry_rcu(mod, &modules, list)
3624
		printk(" %s%s", mod->name, module_flags(mod, buf));
3625
	preempt_enable();
3626 3627
	if (last_unloaded_module[0])
		printk(" [last unloaded: %s]", last_unloaded_module);
L
Linus Torvalds 已提交
3628 3629 3630 3631
	printk("\n");
}

#ifdef CONFIG_MODVERSIONS
3632 3633 3634 3635 3636 3637
/* Generate the signature for all relevant module structures here.
 * If these change, we don't want to try to parse the module. */
void module_layout(struct module *mod,
		   struct modversion_info *ver,
		   struct kernel_param *kp,
		   struct kernel_symbol *ks,
3638
		   struct tracepoint * const *tp)
3639 3640 3641
{
}
EXPORT_SYMBOL(module_layout);
L
Linus Torvalds 已提交
3642
#endif
新手
引导
客服 返回
顶部