module.c 101.7 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/file.h>
25
#include <linux/fs.h>
26
#include <linux/sysfs.h>
27
#include <linux/kernel.h>
L
Linus Torvalds 已提交
28 29 30
#include <linux/slab.h>
#include <linux/vmalloc.h>
#include <linux/elf.h>
31
#include <linux/proc_fs.h>
32
#include <linux/security.h>
L
Linus Torvalds 已提交
33 34 35 36
#include <linux/seq_file.h>
#include <linux/syscalls.h>
#include <linux/fcntl.h>
#include <linux/rcupdate.h>
37
#include <linux/capability.h>
L
Linus Torvalds 已提交
38 39 40 41 42 43
#include <linux/cpu.h>
#include <linux/moduleparam.h>
#include <linux/errno.h>
#include <linux/err.h>
#include <linux/vermagic.h>
#include <linux/notifier.h>
44
#include <linux/sched.h>
L
Linus Torvalds 已提交
45
#include <linux/device.h>
46
#include <linux/string.h>
47
#include <linux/mutex.h>
48
#include <linux/rculist.h>
L
Linus Torvalds 已提交
49 50
#include <asm/uaccess.h>
#include <asm/cacheflush.h>
51
#include <asm/mmu_context.h>
52
#include <linux/license.h>
53
#include <asm/sections.h>
54
#include <linux/tracepoint.h>
55
#include <linux/ftrace.h>
56
#include <linux/async.h>
57
#include <linux/percpu.h>
58
#include <linux/kmemleak.h>
59
#include <linux/jump_label.h>
60
#include <linux/pfn.h>
61
#include <linux/bsearch.h>
62
#include <uapi/linux/module.h>
63
#include "module-internal.h"
L
Linus Torvalds 已提交
64

65 66 67
#define CREATE_TRACE_POINTS
#include <trace/events/module.h>

L
Linus Torvalds 已提交
68 69 70 71
#ifndef ARCH_SHF_SMALL
#define ARCH_SHF_SMALL 0
#endif

72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91
/*
 * 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
92 93 94
/* If this is set, the section belongs in the init part of the module */
#define INIT_OFFSET_MASK (1UL << (BITS_PER_LONG-1))

95 96 97 98 99
/*
 * Mutex protects:
 * 1) List of modules (also safely readable with preempt_disable),
 * 2) module_use links,
 * 3) module_addr_min/module_addr_max.
100
 * (delete and add uses RCU list operations). */
101 102
DEFINE_MUTEX(module_mutex);
EXPORT_SYMBOL_GPL(module_mutex);
L
Linus Torvalds 已提交
103
static LIST_HEAD(modules);
104

105 106
#ifdef CONFIG_MODULES_TREE_LOOKUP

107 108 109 110 111 112 113 114 115 116
/*
 * Use a latched RB-tree for __module_address(); this allows us to use
 * RCU-sched lookups of the address from any context.
 *
 * Because modules have two address ranges: init and core, we need two
 * latch_tree_nodes entries. Therefore we need the back-pointer from
 * mod_tree_node.
 *
 * Because init ranges are short lived we mark them unlikely and have placed
 * them outside the critical cacheline in struct module.
117 118 119 120
 *
 * This is conditional on PERF_EVENTS || TRACING because those can really hit
 * __module_address() hard by doing a lot of stack unwinding; potentially from
 * NMI context.
121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172
 */

static __always_inline unsigned long __mod_tree_val(struct latch_tree_node *n)
{
	struct mod_tree_node *mtn = container_of(n, struct mod_tree_node, node);
	struct module *mod = mtn->mod;

	if (unlikely(mtn == &mod->mtn_init))
		return (unsigned long)mod->module_init;

	return (unsigned long)mod->module_core;
}

static __always_inline unsigned long __mod_tree_size(struct latch_tree_node *n)
{
	struct mod_tree_node *mtn = container_of(n, struct mod_tree_node, node);
	struct module *mod = mtn->mod;

	if (unlikely(mtn == &mod->mtn_init))
		return (unsigned long)mod->init_size;

	return (unsigned long)mod->core_size;
}

static __always_inline bool
mod_tree_less(struct latch_tree_node *a, struct latch_tree_node *b)
{
	return __mod_tree_val(a) < __mod_tree_val(b);
}

static __always_inline int
mod_tree_comp(void *key, struct latch_tree_node *n)
{
	unsigned long val = (unsigned long)key;
	unsigned long start, end;

	start = __mod_tree_val(n);
	if (val < start)
		return -1;

	end = start + __mod_tree_size(n);
	if (val >= end)
		return 1;

	return 0;
}

static const struct latch_tree_ops mod_tree_ops = {
	.less = mod_tree_less,
	.comp = mod_tree_comp,
};

173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192
static struct mod_tree_root {
	struct latch_tree_root root;
	unsigned long addr_min;
	unsigned long addr_max;
} mod_tree __cacheline_aligned = {
	.addr_min = -1UL,
};

#define module_addr_min mod_tree.addr_min
#define module_addr_max mod_tree.addr_max

static noinline void __mod_tree_insert(struct mod_tree_node *node)
{
	latch_tree_insert(&node->node, &mod_tree.root, &mod_tree_ops);
}

static void __mod_tree_remove(struct mod_tree_node *node)
{
	latch_tree_erase(&node->node, &mod_tree.root, &mod_tree_ops);
}
193 194 195 196 197 198 199 200 201 202

/*
 * These modifications: insert, remove_init and remove; are serialized by the
 * module_mutex.
 */
static void mod_tree_insert(struct module *mod)
{
	mod->mtn_core.mod = mod;
	mod->mtn_init.mod = mod;

203
	__mod_tree_insert(&mod->mtn_core);
204
	if (mod->init_size)
205
		__mod_tree_insert(&mod->mtn_init);
206 207 208 209 210
}

static void mod_tree_remove_init(struct module *mod)
{
	if (mod->init_size)
211
		__mod_tree_remove(&mod->mtn_init);
212 213 214 215
}

static void mod_tree_remove(struct module *mod)
{
216
	__mod_tree_remove(&mod->mtn_core);
217 218 219
	mod_tree_remove_init(mod);
}

220
static struct module *mod_find(unsigned long addr)
221 222 223
{
	struct latch_tree_node *ltn;

224
	ltn = latch_tree_find((void *)addr, &mod_tree.root, &mod_tree_ops);
225 226 227 228 229 230
	if (!ltn)
		return NULL;

	return container_of(ltn, struct mod_tree_node, node)->mod;
}

231 232
#else /* MODULES_TREE_LOOKUP */

233 234
static unsigned long module_addr_min = -1UL, module_addr_max = 0;

235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252
static void mod_tree_insert(struct module *mod) { }
static void mod_tree_remove_init(struct module *mod) { }
static void mod_tree_remove(struct module *mod) { }

static struct module *mod_find(unsigned long addr)
{
	struct module *mod;

	list_for_each_entry_rcu(mod, &modules, list) {
		if (within_module(addr, mod))
			return mod;
	}

	return NULL;
}

#endif /* MODULES_TREE_LOOKUP */

253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274
/*
 * Bounds of module text, for speeding up __module_address.
 * Protected by module_mutex.
 */
static void __mod_update_bounds(void *base, unsigned int size)
{
	unsigned long min = (unsigned long)base;
	unsigned long max = min + size;

	if (min < module_addr_min)
		module_addr_min = min;
	if (max > module_addr_max)
		module_addr_max = max;
}

static void mod_update_bounds(struct module *mod)
{
	__mod_update_bounds(mod->module_core, mod->core_size);
	if (mod->init_size)
		__mod_update_bounds(mod->module_init, mod->init_size);
}

275 276 277 278
#ifdef CONFIG_KGDB_KDB
struct list_head *kdb_modules = &modules; /* kdb needs the list of modules */
#endif /* CONFIG_KGDB_KDB */

279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294
static void module_assert_mutex(void)
{
	lockdep_assert_held(&module_mutex);
}

static void module_assert_mutex_or_preempt(void)
{
#ifdef CONFIG_LOCKDEP
	if (unlikely(!debug_locks))
		return;

	WARN_ON(!rcu_read_lock_sched_held() &&
		!lockdep_is_held(&module_mutex));
#endif
}

295 296
static bool sig_enforce = IS_ENABLED(CONFIG_MODULE_SIG_FORCE);
#ifndef CONFIG_MODULE_SIG_FORCE
297 298
module_param(sig_enforce, bool_enable_only, 0644);
#endif /* !CONFIG_MODULE_SIG_FORCE */
L
Linus Torvalds 已提交
299

300 301
/* Block module loading/unloading? */
int modules_disabled = 0;
302
core_param(nomodule, modules_disabled, bint, 0);
303

304 305 306
/* Waiting for a module to finish initializing? */
static DECLARE_WAIT_QUEUE_HEAD(module_wq);

307
static BLOCKING_NOTIFIER_HEAD(module_notify_list);
L
Linus Torvalds 已提交
308

309
int register_module_notifier(struct notifier_block *nb)
L
Linus Torvalds 已提交
310
{
311
	return blocking_notifier_chain_register(&module_notify_list, nb);
L
Linus Torvalds 已提交
312 313 314
}
EXPORT_SYMBOL(register_module_notifier);

315
int unregister_module_notifier(struct notifier_block *nb)
L
Linus Torvalds 已提交
316
{
317
	return blocking_notifier_chain_unregister(&module_notify_list, nb);
L
Linus Torvalds 已提交
318 319 320
}
EXPORT_SYMBOL(unregister_module_notifier);

321 322 323 324
struct load_info {
	Elf_Ehdr *hdr;
	unsigned long len;
	Elf_Shdr *sechdrs;
325
	char *secstrings, *strtab;
326
	unsigned long symoffs, stroffs;
327 328
	struct _ddebug *debug;
	unsigned int num_debug;
329
	bool sig_ok;
330 331 332 333 334
	struct {
		unsigned int sym, str, mod, vers, info, pcpu;
	} index;
};

335 336
/* We require a truly strong try_module_get(): 0 means failure due to
   ongoing or failed initialization etc. */
L
Linus Torvalds 已提交
337 338
static inline int strong_try_module_get(struct module *mod)
{
339
	BUG_ON(mod && mod->state == MODULE_STATE_UNFORMED);
L
Linus Torvalds 已提交
340
	if (mod && mod->state == MODULE_STATE_COMING)
341 342
		return -EBUSY;
	if (try_module_get(mod))
L
Linus Torvalds 已提交
343
		return 0;
344 345
	else
		return -ENOENT;
L
Linus Torvalds 已提交
346 347
}

348 349
static inline void add_taint_module(struct module *mod, unsigned flag,
				    enum lockdep_ok lockdep_ok)
350
{
351
	add_taint(flag, lockdep_ok);
352
	mod->taints |= (1U << flag);
353 354
}

355 356 357
/*
 * 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
358 359 360 361 362 363 364
 */
void __module_put_and_exit(struct module *mod, long code)
{
	module_put(mod);
	do_exit(code);
}
EXPORT_SYMBOL(__module_put_and_exit);
365

L
Linus Torvalds 已提交
366
/* Find a module section: 0 means not found. */
367
static unsigned int find_sec(const struct load_info *info, const char *name)
L
Linus Torvalds 已提交
368 369 370
{
	unsigned int i;

371 372
	for (i = 1; i < info->hdr->e_shnum; i++) {
		Elf_Shdr *shdr = &info->sechdrs[i];
L
Linus Torvalds 已提交
373
		/* Alloc bit cleared means "ignore it." */
374 375
		if ((shdr->sh_flags & SHF_ALLOC)
		    && strcmp(info->secstrings + shdr->sh_name, name) == 0)
L
Linus Torvalds 已提交
376
			return i;
377
	}
L
Linus Torvalds 已提交
378 379 380
	return 0;
}

381
/* Find a module section, or NULL. */
382
static void *section_addr(const struct load_info *info, const char *name)
383 384
{
	/* Section 0 has sh_addr 0. */
385
	return (void *)info->sechdrs[find_sec(info, name)].sh_addr;
386 387 388
}

/* Find a module section, or NULL.  Fill in number of "objects" in section. */
389
static void *section_objs(const struct load_info *info,
390 391 392 393
			  const char *name,
			  size_t object_size,
			  unsigned int *num)
{
394
	unsigned int sec = find_sec(info, name);
395 396

	/* Section 0 has sh_addr 0 and sh_size 0. */
397 398
	*num = info->sechdrs[sec].sh_size / object_size;
	return (void *)info->sechdrs[sec].sh_addr;
399 400
}

L
Linus Torvalds 已提交
401 402 403 404 405
/* 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[];
406 407
extern const struct kernel_symbol __start___ksymtab_gpl_future[];
extern const struct kernel_symbol __stop___ksymtab_gpl_future[];
L
Linus Torvalds 已提交
408 409
extern const unsigned long __start___kcrctab[];
extern const unsigned long __start___kcrctab_gpl[];
410
extern const unsigned long __start___kcrctab_gpl_future[];
411 412 413 414 415
#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[];
416 417
extern const unsigned long __start___kcrctab_unused[];
extern const unsigned long __start___kcrctab_unused_gpl[];
418
#endif
L
Linus Torvalds 已提交
419 420 421 422

#ifndef CONFIG_MODVERSIONS
#define symversion(base, idx) NULL
#else
423
#define symversion(base, idx) ((base != NULL) ? ((base) + (idx)) : NULL)
L
Linus Torvalds 已提交
424 425
#endif

426 427 428 429 430
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,
431
					      void *data),
432
				   void *data)
433
{
434
	unsigned int j;
435

436
	for (j = 0; j < arrsize; j++) {
437 438
		if (fn(&arr[j], owner, data))
			return true;
439
	}
440 441

	return false;
442 443
}

444
/* Returns true as soon as fn returns true, otherwise false. */
445 446 447 448
bool each_symbol_section(bool (*fn)(const struct symsearch *arr,
				    struct module *owner,
				    void *data),
			 void *data)
449 450
{
	struct module *mod;
451
	static const struct symsearch arr[] = {
452
		{ __start___ksymtab, __stop___ksymtab, __start___kcrctab,
453
		  NOT_GPL_ONLY, false },
454
		{ __start___ksymtab_gpl, __stop___ksymtab_gpl,
455 456
		  __start___kcrctab_gpl,
		  GPL_ONLY, false },
457
		{ __start___ksymtab_gpl_future, __stop___ksymtab_gpl_future,
458 459
		  __start___kcrctab_gpl_future,
		  WILL_BE_GPL_ONLY, false },
460
#ifdef CONFIG_UNUSED_SYMBOLS
461
		{ __start___ksymtab_unused, __stop___ksymtab_unused,
462 463
		  __start___kcrctab_unused,
		  NOT_GPL_ONLY, true },
464
		{ __start___ksymtab_unused_gpl, __stop___ksymtab_unused_gpl,
465 466
		  __start___kcrctab_unused_gpl,
		  GPL_ONLY, true },
467
#endif
468
	};
469

470 471
	module_assert_mutex_or_preempt();

472 473
	if (each_symbol_in_section(arr, ARRAY_SIZE(arr), NULL, fn, data))
		return true;
474

475
	list_for_each_entry_rcu(mod, &modules, list) {
476 477
		struct symsearch arr[] = {
			{ mod->syms, mod->syms + mod->num_syms, mod->crcs,
478
			  NOT_GPL_ONLY, false },
479
			{ mod->gpl_syms, mod->gpl_syms + mod->num_gpl_syms,
480 481
			  mod->gpl_crcs,
			  GPL_ONLY, false },
482 483
			{ mod->gpl_future_syms,
			  mod->gpl_future_syms + mod->num_gpl_future_syms,
484 485
			  mod->gpl_future_crcs,
			  WILL_BE_GPL_ONLY, false },
486
#ifdef CONFIG_UNUSED_SYMBOLS
487 488
			{ mod->unused_syms,
			  mod->unused_syms + mod->num_unused_syms,
489 490
			  mod->unused_crcs,
			  NOT_GPL_ONLY, true },
491 492
			{ mod->unused_gpl_syms,
			  mod->unused_gpl_syms + mod->num_unused_gpl_syms,
493 494
			  mod->unused_gpl_crcs,
			  GPL_ONLY, true },
495
#endif
496 497
		};

498 499 500
		if (mod->state == MODULE_STATE_UNFORMED)
			continue;

501 502 503 504 505
		if (each_symbol_in_section(arr, ARRAY_SIZE(arr), mod, fn, data))
			return true;
	}
	return false;
}
506
EXPORT_SYMBOL_GPL(each_symbol_section);
507 508 509 510 511 512 513 514 515 516

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

	/* Output */
	struct module *owner;
	const unsigned long *crc;
517
	const struct kernel_symbol *sym;
518 519
};

520 521 522
static bool check_symbol(const struct symsearch *syms,
				 struct module *owner,
				 unsigned int symnum, void *data)
523 524 525 526 527 528 529
{
	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) {
530 531 532
			pr_warn("Symbol %s is being used by a non-GPL module, "
				"which will not be allowed in the future\n",
				fsa->name);
533
		}
L
Linus Torvalds 已提交
534
	}
535

536
#ifdef CONFIG_UNUSED_SYMBOLS
537
	if (syms->unused && fsa->warn) {
538 539 540
		pr_warn("Symbol %s is marked as UNUSED, however this module is "
			"using it.\n", fsa->name);
		pr_warn("This symbol will go away in the future.\n");
541 542 543
		pr_warn("Please evaluate if this is the right api to use and "
			"if it really is, submit a report to the linux kernel "
			"mailing list together with submitting your code for "
544
			"inclusion.\n");
545
	}
546
#endif
547 548 549

	fsa->owner = owner;
	fsa->crc = symversion(syms->crcs, symnum);
550
	fsa->sym = &syms->start[symnum];
551 552 553
	return true;
}

554 555 556 557 558 559 560 561
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);
}

562 563 564 565 566
static bool find_symbol_in_section(const struct symsearch *syms,
				   struct module *owner,
				   void *data)
{
	struct find_symbol_arg *fsa = data;
567 568 569 570 571 572 573
	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;
574 575 576 577

	return false;
}

578
/* Find a symbol and return it, along with, (optional) crc and
579
 * (optional) module which owns it.  Needs preempt disabled or module_mutex. */
580 581 582 583 584
const struct kernel_symbol *find_symbol(const char *name,
					struct module **owner,
					const unsigned long **crc,
					bool gplok,
					bool warn)
585 586 587 588 589 590 591
{
	struct find_symbol_arg fsa;

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

592
	if (each_symbol_section(find_symbol_in_section, &fsa)) {
593 594 595 596
		if (owner)
			*owner = fsa.owner;
		if (crc)
			*crc = fsa.crc;
597
		return fsa.sym;
598 599
	}

600
	pr_debug("Failed to find symbol %s\n", name);
601
	return NULL;
L
Linus Torvalds 已提交
602
}
603
EXPORT_SYMBOL_GPL(find_symbol);
L
Linus Torvalds 已提交
604 605

/* Search for module by name: must hold module_mutex. */
606
static struct module *find_module_all(const char *name, size_t len,
607
				      bool even_unformed)
L
Linus Torvalds 已提交
608 609 610
{
	struct module *mod;

611 612
	module_assert_mutex();

L
Linus Torvalds 已提交
613
	list_for_each_entry(mod, &modules, list) {
614 615
		if (!even_unformed && mod->state == MODULE_STATE_UNFORMED)
			continue;
616
		if (strlen(mod->name) == len && !memcmp(mod->name, name, len))
L
Linus Torvalds 已提交
617 618 619 620
			return mod;
	}
	return NULL;
}
621 622 623

struct module *find_module(const char *name)
{
624
	return find_module_all(name, strlen(name), false);
625
}
626
EXPORT_SYMBOL_GPL(find_module);
L
Linus Torvalds 已提交
627 628

#ifdef CONFIG_SMP
629

630
static inline void __percpu *mod_percpu(struct module *mod)
631
{
632 633
	return mod->percpu;
}
634

635
static int percpu_modalloc(struct module *mod, struct load_info *info)
636
{
637 638 639 640 641 642
	Elf_Shdr *pcpusec = &info->sechdrs[info->index.pcpu];
	unsigned long align = pcpusec->sh_addralign;

	if (!pcpusec->sh_size)
		return 0;

643
	if (align > PAGE_SIZE) {
644 645
		pr_warn("%s: per-cpu alignment %li > %li\n",
			mod->name, align, PAGE_SIZE);
646 647 648
		align = PAGE_SIZE;
	}

649
	mod->percpu = __alloc_reserved_percpu(pcpusec->sh_size, align);
650
	if (!mod->percpu) {
651 652
		pr_warn("%s: Could not allocate %lu bytes percpu data\n",
			mod->name, (unsigned long)pcpusec->sh_size);
653 654
		return -ENOMEM;
	}
655
	mod->percpu_size = pcpusec->sh_size;
656
	return 0;
657 658
}

659
static void percpu_modfree(struct module *mod)
660
{
661
	free_percpu(mod->percpu);
662 663
}

664
static unsigned int find_pcpusec(struct load_info *info)
665
{
666
	return find_sec(info, ".data..percpu");
667 668
}

669 670
static void percpu_modcopy(struct module *mod,
			   const void *from, unsigned long size)
671 672 673 674
{
	int cpu;

	for_each_possible_cpu(cpu)
675
		memcpy(per_cpu_ptr(mod->percpu, cpu), from, size);
676 677
}

678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694
/**
 * 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) {
695 696
		if (mod->state == MODULE_STATE_UNFORMED)
			continue;
697 698 699 700 701 702 703 704 705 706 707 708 709 710 711
		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;
712 713
}

L
Linus Torvalds 已提交
714
#else /* ... !CONFIG_SMP */
715

716
static inline void __percpu *mod_percpu(struct module *mod)
L
Linus Torvalds 已提交
717 718 719
{
	return NULL;
}
720
static int percpu_modalloc(struct module *mod, struct load_info *info)
721
{
722 723 724 725
	/* UP modules shouldn't have this section: ENOMEM isn't quite right */
	if (info->sechdrs[info->index.pcpu].sh_size != 0)
		return -ENOMEM;
	return 0;
726 727
}
static inline void percpu_modfree(struct module *mod)
L
Linus Torvalds 已提交
728 729
{
}
730
static unsigned int find_pcpusec(struct load_info *info)
L
Linus Torvalds 已提交
731 732 733
{
	return 0;
}
734 735
static inline void percpu_modcopy(struct module *mod,
				  const void *from, unsigned long size)
L
Linus Torvalds 已提交
736 737 738 739
{
	/* pcpusec should be 0, and size of that section should be 0. */
	BUG_ON(size != 0);
}
740 741 742 743
bool is_module_percpu_address(unsigned long addr)
{
	return false;
}
744

L
Linus Torvalds 已提交
745 746
#endif /* CONFIG_SMP */

747 748 749 750 751 752
#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,   \
753
			struct module_kobject *mk, char *buffer)      \
754
{                                                                     \
755
	return scnprintf(buffer, PAGE_SIZE, "%s\n", mk->mod->field);  \
756 757 758 759 760 761 762
}                                                                     \
static int modinfo_##field##_exists(struct module *mod)               \
{                                                                     \
	return mod->field != NULL;                                    \
}                                                                     \
static void free_modinfo_##field(struct module *mod)                  \
{                                                                     \
763 764
	kfree(mod->field);                                            \
	mod->field = NULL;                                            \
765 766
}                                                                     \
static struct module_attribute modinfo_##field = {                    \
767
	.attr = { .name = __stringify(field), .mode = 0444 },         \
768 769 770 771 772 773 774 775 776
	.show = show_modinfo_##field,                                 \
	.setup = setup_modinfo_##field,                               \
	.test = modinfo_##field##_exists,                             \
	.free = free_modinfo_##field,                                 \
};

MODINFO_ATTR(version);
MODINFO_ATTR(srcversion);

777 778
static char last_unloaded_module[MODULE_NAME_LEN+1];

779
#ifdef CONFIG_MODULE_UNLOAD
780 781 782

EXPORT_TRACEPOINT_SYMBOL(module_get);

783 784 785
/* MODULE_REF_BASE is the base reference count by kmodule loader. */
#define MODULE_REF_BASE	1

L
Linus Torvalds 已提交
786
/* Init the unload section of the module. */
787
static int module_unload_init(struct module *mod)
L
Linus Torvalds 已提交
788
{
789 790 791 792 793
	/*
	 * Initialize reference counter to MODULE_REF_BASE.
	 * refcnt == 0 means module is going.
	 */
	atomic_set(&mod->refcnt, MODULE_REF_BASE);
794

795 796
	INIT_LIST_HEAD(&mod->source_list);
	INIT_LIST_HEAD(&mod->target_list);
797

L
Linus Torvalds 已提交
798
	/* Hold reference count during initialization. */
799
	atomic_inc(&mod->refcnt);
800 801

	return 0;
L
Linus Torvalds 已提交
802 803 804 805 806 807 808
}

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

809 810
	list_for_each_entry(use, &b->source_list, source_list) {
		if (use->source == a) {
811
			pr_debug("%s uses %s!\n", a->name, b->name);
L
Linus Torvalds 已提交
812 813 814
			return 1;
		}
	}
815
	pr_debug("%s does not use %s!\n", a->name, b->name);
L
Linus Torvalds 已提交
816 817 818
	return 0;
}

819 820 821 822 823 824 825 826 827 828 829
/*
 * 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;

830
	pr_debug("Allocating new usage for %s.\n", a->name);
831 832
	use = kmalloc(sizeof(*use), GFP_ATOMIC);
	if (!use) {
833
		pr_warn("%s: out of memory loading\n", a->name);
834 835 836 837 838 839 840 841 842 843
		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;
}

844
/* Module a uses b: caller needs module_mutex() */
845
int ref_module(struct module *a, struct module *b)
L
Linus Torvalds 已提交
846
{
847
	int err;
K
Kay Sievers 已提交
848

849
	if (b == NULL || already_uses(a, b))
850 851
		return 0;

852 853
	/* If module isn't available, we fail. */
	err = strong_try_module_get(b);
854
	if (err)
855
		return err;
L
Linus Torvalds 已提交
856

857 858
	err = add_module_usage(a, b);
	if (err) {
L
Linus Torvalds 已提交
859
		module_put(b);
860
		return err;
L
Linus Torvalds 已提交
861
	}
862
	return 0;
L
Linus Torvalds 已提交
863
}
864
EXPORT_SYMBOL_GPL(ref_module);
L
Linus Torvalds 已提交
865 866 867 868

/* Clear the unload stuff of the module. */
static void module_unload_free(struct module *mod)
{
869
	struct module_use *use, *tmp;
L
Linus Torvalds 已提交
870

871
	mutex_lock(&module_mutex);
872 873
	list_for_each_entry_safe(use, tmp, &mod->target_list, target_list) {
		struct module *i = use->target;
874
		pr_debug("%s unusing %s\n", mod->name, i->name);
875 876 877 878
		module_put(i);
		list_del(&use->source_list);
		list_del(&use->target_list);
		kfree(use);
L
Linus Torvalds 已提交
879
	}
880
	mutex_unlock(&module_mutex);
L
Linus Torvalds 已提交
881 882 883
}

#ifdef CONFIG_MODULE_FORCE_UNLOAD
884
static inline int try_force_unload(unsigned int flags)
L
Linus Torvalds 已提交
885 886 887
{
	int ret = (flags & O_TRUNC);
	if (ret)
888
		add_taint(TAINT_FORCED_RMMOD, LOCKDEP_NOW_UNRELIABLE);
L
Linus Torvalds 已提交
889 890 891
	return ret;
}
#else
892
static inline int try_force_unload(unsigned int flags)
L
Linus Torvalds 已提交
893 894 895 896 897
{
	return 0;
}
#endif /* CONFIG_MODULE_FORCE_UNLOAD */

898 899
/* Try to release refcount of module, 0 means success. */
static int try_release_module_ref(struct module *mod)
L
Linus Torvalds 已提交
900
{
901
	int ret;
L
Linus Torvalds 已提交
902

903 904 905 906 907 908
	/* Try to decrement refcnt which we set at loading */
	ret = atomic_sub_return(MODULE_REF_BASE, &mod->refcnt);
	BUG_ON(ret < 0);
	if (ret)
		/* Someone can put this right now, recover with checking */
		ret = atomic_add_unless(&mod->refcnt, MODULE_REF_BASE, 0);
L
Linus Torvalds 已提交
909

910 911
	return ret;
}
L
Linus Torvalds 已提交
912

913 914
static int try_stop_module(struct module *mod, int flags, int *forced)
{
915
	/* If it's not unused, quit unless we're forcing. */
916 917 918
	if (try_release_module_ref(mod) != 0) {
		*forced = try_force_unload(flags);
		if (!(*forced))
L
Linus Torvalds 已提交
919 920 921 922
			return -EWOULDBLOCK;
	}

	/* Mark it as dying. */
923
	mod->state = MODULE_STATE_GOING;
L
Linus Torvalds 已提交
924

925
	return 0;
L
Linus Torvalds 已提交
926 927
}

928 929 930 931 932 933 934 935 936 937
/**
 * module_refcount - return the refcount or -1 if unloading
 *
 * @mod:	the module we're checking
 *
 * Returns:
 *	-1 if the module is in the process of unloading
 *	otherwise the number of references in the kernel to the module
 */
int module_refcount(struct module *mod)
L
Linus Torvalds 已提交
938
{
939
	return atomic_read(&mod->refcnt) - MODULE_REF_BASE;
L
Linus Torvalds 已提交
940 941 942 943 944 945
}
EXPORT_SYMBOL(module_refcount);

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

946 947
SYSCALL_DEFINE2(delete_module, const char __user *, name_user,
		unsigned int, flags)
L
Linus Torvalds 已提交
948 949
{
	struct module *mod;
950
	char name[MODULE_NAME_LEN];
L
Linus Torvalds 已提交
951 952
	int ret, forced = 0;

953
	if (!capable(CAP_SYS_MODULE) || modules_disabled)
954 955 956 957 958 959
		return -EPERM;

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

960 961
	if (mutex_lock_interruptible(&module_mutex) != 0)
		return -EINTR;
L
Linus Torvalds 已提交
962 963 964 965 966 967 968

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

969
	if (!list_empty(&mod->source_list)) {
L
Linus Torvalds 已提交
970 971 972 973 974 975 976
		/* 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) {
977
		/* FIXME: if (force), slam module count damn the torpedoes */
978
		pr_debug("%s already dying\n", mod->name);
L
Linus Torvalds 已提交
979 980 981 982 983
		ret = -EBUSY;
		goto out;
	}

	/* If it has an init func, it must have an exit func to unload */
984
	if (mod->init && !mod->exit) {
985
		forced = try_force_unload(flags);
L
Linus Torvalds 已提交
986 987 988 989 990 991 992 993 994 995 996 997
		if (!forced) {
			/* This module can't be removed */
			ret = -EBUSY;
			goto out;
		}
	}

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

998
	mutex_unlock(&module_mutex);
L
Lucas De Marchi 已提交
999
	/* Final destruction now no one is using it. */
1000
	if (mod->exit != NULL)
L
Linus Torvalds 已提交
1001
		mod->exit();
1002 1003
	blocking_notifier_call_chain(&module_notify_list,
				     MODULE_STATE_GOING, mod);
1004
	async_synchronize_full();
1005

1006
	/* Store the name of the last unloaded module for diagnostic purposes */
1007
	strlcpy(last_unloaded_module, mod->name, sizeof(last_unloaded_module));
L
Linus Torvalds 已提交
1008

1009 1010 1011
	free_module(mod);
	return 0;
out:
1012
	mutex_unlock(&module_mutex);
L
Linus Torvalds 已提交
1013 1014 1015
	return ret;
}

1016
static inline void print_unload_info(struct seq_file *m, struct module *mod)
L
Linus Torvalds 已提交
1017 1018 1019 1020
{
	struct module_use *use;
	int printed_something = 0;

1021
	seq_printf(m, " %i ", module_refcount(mod));
L
Linus Torvalds 已提交
1022

1023 1024 1025 1026
	/*
	 * Always include a trailing , so userspace can differentiate
	 * between this and the old multi-field proc format.
	 */
1027
	list_for_each_entry(use, &mod->source_list, source_list) {
L
Linus Torvalds 已提交
1028
		printed_something = 1;
1029
		seq_printf(m, "%s,", use->source->name);
L
Linus Torvalds 已提交
1030 1031 1032 1033
	}

	if (mod->init != NULL && mod->exit == NULL) {
		printed_something = 1;
1034
		seq_puts(m, "[permanent],");
L
Linus Torvalds 已提交
1035 1036 1037
	}

	if (!printed_something)
1038
		seq_puts(m, "-");
L
Linus Torvalds 已提交
1039 1040 1041 1042 1043 1044
}

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

1045
	preempt_disable();
1046
	if (!find_symbol(symbol, &owner, NULL, true, false))
L
Linus Torvalds 已提交
1047 1048
		BUG();
	module_put(owner);
1049
	preempt_enable();
L
Linus Torvalds 已提交
1050 1051 1052
}
EXPORT_SYMBOL(__symbol_put);

1053
/* Note this assumes addr is a function, which it currently always is. */
L
Linus Torvalds 已提交
1054 1055
void symbol_put_addr(void *addr)
{
1056
	struct module *modaddr;
1057
	unsigned long a = (unsigned long)dereference_function_descriptor(addr);
L
Linus Torvalds 已提交
1058

1059
	if (core_kernel_text(a))
1060
		return;
L
Linus Torvalds 已提交
1061

1062 1063
	/* module_text_address is safe here: we're supposed to have reference
	 * to module from symbol_get, so it can't go away. */
1064
	modaddr = __module_text_address(a);
1065
	BUG_ON(!modaddr);
1066
	module_put(modaddr);
L
Linus Torvalds 已提交
1067 1068 1069 1070
}
EXPORT_SYMBOL_GPL(symbol_put_addr);

static ssize_t show_refcnt(struct module_attribute *mattr,
1071
			   struct module_kobject *mk, char *buffer)
L
Linus Torvalds 已提交
1072
{
1073
	return sprintf(buffer, "%i\n", module_refcount(mk->mod));
L
Linus Torvalds 已提交
1074 1075
}

1076 1077
static struct module_attribute modinfo_refcnt =
	__ATTR(refcnt, 0444, show_refcnt, NULL);
L
Linus Torvalds 已提交
1078

1079 1080 1081 1082
void __module_get(struct module *module)
{
	if (module) {
		preempt_disable();
1083
		atomic_inc(&module->refcnt);
1084 1085 1086 1087 1088 1089 1090 1091 1092 1093 1094 1095
		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();
1096 1097 1098
		/* Note: here, we can fail to get a reference */
		if (likely(module_is_live(module) &&
			   atomic_inc_not_zero(&module->refcnt) != 0))
1099
			trace_module_get(module, _RET_IP_);
1100
		else
1101 1102 1103 1104 1105 1106 1107 1108
			ret = false;

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

1109 1110
void module_put(struct module *module)
{
1111 1112
	int ret;

1113
	if (module) {
1114
		preempt_disable();
1115 1116
		ret = atomic_dec_if_positive(&module->refcnt);
		WARN_ON(ret < 0);	/* Failed to put refcount */
1117
		trace_module_put(module, _RET_IP_);
1118
		preempt_enable();
1119 1120 1121 1122
	}
}
EXPORT_SYMBOL(module_put);

L
Linus Torvalds 已提交
1123
#else /* !CONFIG_MODULE_UNLOAD */
1124
static inline void print_unload_info(struct seq_file *m, struct module *mod)
L
Linus Torvalds 已提交
1125 1126
{
	/* We don't know the usage count, or what modules are using. */
1127
	seq_puts(m, " - -");
L
Linus Torvalds 已提交
1128 1129 1130 1131 1132 1133
}

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

1134
int ref_module(struct module *a, struct module *b)
L
Linus Torvalds 已提交
1135
{
1136
	return strong_try_module_get(b);
L
Linus Torvalds 已提交
1137
}
1138
EXPORT_SYMBOL_GPL(ref_module);
L
Linus Torvalds 已提交
1139

1140
static inline int module_unload_init(struct module *mod)
L
Linus Torvalds 已提交
1141
{
1142
	return 0;
L
Linus Torvalds 已提交
1143 1144 1145
}
#endif /* CONFIG_MODULE_UNLOAD */

1146 1147 1148 1149 1150 1151 1152 1153 1154 1155 1156 1157
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';
1158
	if (mod->taints & (1 << TAINT_UNSIGNED_MODULE))
1159
		buf[l++] = 'E';
1160 1161
	/*
	 * TAINT_FORCED_RMMOD: could be added.
1162
	 * TAINT_CPU_OUT_OF_SPEC, TAINT_MACHINE_CHECK, TAINT_BAD_PAGE don't
1163 1164 1165 1166 1167
	 * apply to modules.
	 */
	return l;
}

1168
static ssize_t show_initstate(struct module_attribute *mattr,
1169
			      struct module_kobject *mk, char *buffer)
1170 1171 1172
{
	const char *state = "unknown";

1173
	switch (mk->mod->state) {
1174 1175 1176 1177 1178 1179 1180 1181 1182
	case MODULE_STATE_LIVE:
		state = "live";
		break;
	case MODULE_STATE_COMING:
		state = "coming";
		break;
	case MODULE_STATE_GOING:
		state = "going";
		break;
1183 1184
	default:
		BUG();
1185 1186 1187 1188
	}
	return sprintf(buffer, "%s\n", state);
}

1189 1190
static struct module_attribute modinfo_initstate =
	__ATTR(initstate, 0444, show_initstate, NULL);
1191

1192 1193 1194 1195 1196 1197 1198 1199 1200 1201 1202
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;
}

1203 1204 1205 1206 1207 1208 1209 1210 1211 1212 1213 1214 1215 1216 1217 1218 1219 1220 1221 1222 1223 1224 1225 1226 1227 1228 1229 1230 1231 1232 1233 1234 1235
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);
1236

1237
static struct module_attribute *modinfo_attrs[] = {
1238
	&module_uevent,
1239 1240
	&modinfo_version,
	&modinfo_srcversion,
1241 1242 1243 1244
	&modinfo_initstate,
	&modinfo_coresize,
	&modinfo_initsize,
	&modinfo_taint,
1245
#ifdef CONFIG_MODULE_UNLOAD
1246
	&modinfo_refcnt,
1247 1248 1249 1250
#endif
	NULL,
};

L
Linus Torvalds 已提交
1251 1252
static const char vermagic[] = VERMAGIC_STRING;

1253
static int try_to_force_load(struct module *mod, const char *reason)
1254 1255
{
#ifdef CONFIG_MODULE_FORCE_LOAD
1256
	if (!test_taint(TAINT_FORCED_MODULE))
1257
		pr_warn("%s: %s: kernel tainted.\n", mod->name, reason);
1258
	add_taint_module(mod, TAINT_FORCED_MODULE, LOCKDEP_NOW_UNRELIABLE);
1259 1260 1261 1262 1263 1264
	return 0;
#else
	return -ENOEXEC;
#endif
}

L
Linus Torvalds 已提交
1265
#ifdef CONFIG_MODVERSIONS
1266 1267 1268 1269 1270 1271 1272 1273 1274 1275 1276
/* 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
1277 1278 1279
static int check_version(Elf_Shdr *sechdrs,
			 unsigned int versindex,
			 const char *symname,
1280
			 struct module *mod,
1281 1282
			 const unsigned long *crc,
			 const struct module *crc_owner)
L
Linus Torvalds 已提交
1283 1284 1285 1286 1287 1288 1289 1290
{
	unsigned int i, num_versions;
	struct modversion_info *versions;

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

1291 1292 1293 1294
	/* 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
1295 1296 1297 1298 1299 1300 1301 1302
	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;

1303
		if (versions[i].crc == maybe_relocated(*crc, crc_owner))
L
Linus Torvalds 已提交
1304
			return 1;
1305
		pr_debug("Found checksum %lX vs module %lX\n",
1306
		       maybe_relocated(*crc, crc_owner), versions[i].crc);
1307
		goto bad_version;
L
Linus Torvalds 已提交
1308
	}
1309

1310
	pr_warn("%s: no symbol version for %s\n", mod->name, symname);
1311
	return 0;
1312 1313

bad_version:
1314
	pr_warn("%s: disagrees about version of symbol %s\n",
1315 1316
	       mod->name, symname);
	return 0;
L
Linus Torvalds 已提交
1317 1318 1319 1320 1321 1322 1323 1324
}

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

1325 1326 1327 1328 1329
	/*
	 * Since this should be found in kernel (which can't be removed), no
	 * locking is necessary -- use preempt_disable() to placate lockdep.
	 */
	preempt_disable();
1330
	if (!find_symbol(VMLINUX_SYMBOL_STR(module_layout), NULL,
1331 1332
			 &crc, true, false)) {
		preempt_enable();
L
Linus Torvalds 已提交
1333
		BUG();
1334 1335
	}
	preempt_enable();
1336 1337
	return check_version(sechdrs, versindex,
			     VMLINUX_SYMBOL_STR(module_layout), mod, crc,
1338
			     NULL);
L
Linus Torvalds 已提交
1339 1340
}

1341 1342 1343
/* 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
1344
{
1345 1346 1347 1348
	if (has_crcs) {
		amagic += strcspn(amagic, " ");
		bmagic += strcspn(bmagic, " ");
	}
L
Linus Torvalds 已提交
1349 1350 1351 1352 1353 1354
	return strcmp(amagic, bmagic) == 0;
}
#else
static inline int check_version(Elf_Shdr *sechdrs,
				unsigned int versindex,
				const char *symname,
1355
				struct module *mod,
1356 1357
				const unsigned long *crc,
				const struct module *crc_owner)
L
Linus Torvalds 已提交
1358 1359 1360 1361 1362 1363 1364 1365 1366 1367 1368
{
	return 1;
}

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

1369 1370
static inline int same_magic(const char *amagic, const char *bmagic,
			     bool has_crcs)
L
Linus Torvalds 已提交
1371 1372 1373 1374 1375
{
	return strcmp(amagic, bmagic) == 0;
}
#endif /* CONFIG_MODVERSIONS */

1376
/* Resolve a symbol for this module.  I.e. if we find one, record usage. */
1377 1378
static const struct kernel_symbol *resolve_symbol(struct module *mod,
						  const struct load_info *info,
1379
						  const char *name,
1380
						  char ownername[])
L
Linus Torvalds 已提交
1381 1382
{
	struct module *owner;
1383
	const struct kernel_symbol *sym;
L
Linus Torvalds 已提交
1384
	const unsigned long *crc;
1385
	int err;
L
Linus Torvalds 已提交
1386

1387 1388 1389 1390 1391 1392
	/*
	 * The module_mutex should not be a heavily contended lock;
	 * if we get the occasional sleep here, we'll go an extra iteration
	 * in the wait_event_interruptible(), which is harmless.
	 */
	sched_annotate_sleep();
1393
	mutex_lock(&module_mutex);
1394
	sym = find_symbol(name, &owner, &crc,
1395
			  !(mod->taints & (1 << TAINT_PROPRIETARY_MODULE)), true);
1396 1397 1398
	if (!sym)
		goto unlock;

1399 1400
	if (!check_version(info->sechdrs, info->index.vers, name, mod, crc,
			   owner)) {
1401 1402
		sym = ERR_PTR(-EINVAL);
		goto getname;
L
Linus Torvalds 已提交
1403
	}
1404 1405 1406 1407 1408 1409 1410 1411 1412 1413 1414

	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:
1415
	mutex_unlock(&module_mutex);
1416
	return sym;
L
Linus Torvalds 已提交
1417 1418
}

1419 1420 1421 1422
static const struct kernel_symbol *
resolve_symbol_wait(struct module *mod,
		    const struct load_info *info,
		    const char *name)
1423 1424
{
	const struct kernel_symbol *ksym;
1425
	char owner[MODULE_NAME_LEN];
1426 1427

	if (wait_event_interruptible_timeout(module_wq,
1428 1429
			!IS_ERR(ksym = resolve_symbol(mod, info, name, owner))
			|| PTR_ERR(ksym) != -EBUSY,
1430
					     30 * HZ) <= 0) {
1431 1432
		pr_warn("%s: gave up waiting for init of module %s.\n",
			mod->name, owner);
1433 1434 1435 1436
	}
	return ksym;
}

L
Linus Torvalds 已提交
1437 1438 1439 1440
/*
 * /sys/module/foo/sections stuff
 * J. Corbet <corbet@lwn.net>
 */
R
Rusty Russell 已提交
1441
#ifdef CONFIG_SYSFS
1442

R
Rusty Russell 已提交
1443
#ifdef CONFIG_KALLSYMS
1444 1445 1446 1447 1448
static inline bool sect_empty(const Elf_Shdr *sect)
{
	return !(sect->sh_flags & SHF_ALLOC) || sect->sh_size == 0;
}

1449
struct module_sect_attr {
1450 1451 1452 1453 1454
	struct module_attribute mattr;
	char *name;
	unsigned long address;
};

1455
struct module_sect_attrs {
1456 1457 1458 1459 1460
	struct attribute_group grp;
	unsigned int nsections;
	struct module_sect_attr attrs[0];
};

L
Linus Torvalds 已提交
1461
static ssize_t module_sect_show(struct module_attribute *mattr,
1462
				struct module_kobject *mk, char *buf)
L
Linus Torvalds 已提交
1463 1464 1465
{
	struct module_sect_attr *sattr =
		container_of(mattr, struct module_sect_attr, mattr);
1466
	return sprintf(buf, "0x%pK\n", (void *)sattr->address);
L
Linus Torvalds 已提交
1467 1468
}

1469 1470
static void free_sect_attrs(struct module_sect_attrs *sect_attrs)
{
1471
	unsigned int section;
1472 1473 1474 1475 1476 1477

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

R
Rusty Russell 已提交
1478
static void add_sect_attrs(struct module *mod, const struct load_info *info)
L
Linus Torvalds 已提交
1479 1480 1481 1482 1483
{
	unsigned int nloaded = 0, i, size[2];
	struct module_sect_attrs *sect_attrs;
	struct module_sect_attr *sattr;
	struct attribute **gattr;
1484

L
Linus Torvalds 已提交
1485
	/* Count loaded sections and allocate structures */
R
Rusty Russell 已提交
1486 1487
	for (i = 0; i < info->hdr->e_shnum; i++)
		if (!sect_empty(&info->sechdrs[i]))
L
Linus Torvalds 已提交
1488 1489 1490 1491 1492
			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]);
1493 1494
	sect_attrs = kzalloc(size[0] + size[1], GFP_KERNEL);
	if (sect_attrs == NULL)
L
Linus Torvalds 已提交
1495 1496 1497 1498 1499 1500
		return;

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

1501
	sect_attrs->nsections = 0;
L
Linus Torvalds 已提交
1502 1503
	sattr = &sect_attrs->attrs[0];
	gattr = &sect_attrs->grp.attrs[0];
R
Rusty Russell 已提交
1504 1505 1506
	for (i = 0; i < info->hdr->e_shnum; i++) {
		Elf_Shdr *sec = &info->sechdrs[i];
		if (sect_empty(sec))
1507
			continue;
R
Rusty Russell 已提交
1508 1509
		sattr->address = sec->sh_addr;
		sattr->name = kstrdup(info->secstrings + sec->sh_name,
1510 1511 1512 1513
					GFP_KERNEL);
		if (sattr->name == NULL)
			goto out;
		sect_attrs->nsections++;
1514
		sysfs_attr_init(&sattr->mattr.attr);
L
Linus Torvalds 已提交
1515 1516 1517 1518 1519 1520 1521 1522 1523 1524 1525 1526 1527 1528
		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:
1529
	free_sect_attrs(sect_attrs);
L
Linus Torvalds 已提交
1530 1531 1532 1533 1534 1535 1536 1537 1538
}

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. */
1539
		free_sect_attrs(mod->sect_attrs);
L
Linus Torvalds 已提交
1540 1541 1542 1543
		mod->sect_attrs = NULL;
	}
}

1544 1545 1546 1547 1548 1549 1550 1551 1552 1553
/*
 * /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];
};

1554
static ssize_t module_notes_read(struct file *filp, struct kobject *kobj,
1555 1556 1557 1558 1559 1560 1561 1562 1563 1564 1565 1566 1567 1568 1569 1570 1571
				 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]);
1572
		kobject_put(notes_attrs->dir);
1573 1574 1575 1576
	}
	kfree(notes_attrs);
}

R
Rusty Russell 已提交
1577
static void add_notes_attrs(struct module *mod, const struct load_info *info)
1578 1579 1580 1581 1582
{
	unsigned int notes, loaded, i;
	struct module_notes_attrs *notes_attrs;
	struct bin_attribute *nattr;

1583 1584 1585 1586
	/* failed to create section attributes, so can't create notes */
	if (!mod->sect_attrs)
		return;

1587 1588
	/* Count notes sections and allocate structures.  */
	notes = 0;
R
Rusty Russell 已提交
1589 1590 1591
	for (i = 0; i < info->hdr->e_shnum; i++)
		if (!sect_empty(&info->sechdrs[i]) &&
		    (info->sechdrs[i].sh_type == SHT_NOTE))
1592 1593 1594 1595 1596 1597 1598 1599 1600 1601 1602 1603 1604
			++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
1605 1606
	for (loaded = i = 0; i < info->hdr->e_shnum; ++i) {
		if (sect_empty(&info->sechdrs[i]))
1607
			continue;
R
Rusty Russell 已提交
1608
		if (info->sechdrs[i].sh_type == SHT_NOTE) {
1609
			sysfs_bin_attr_init(nattr);
1610 1611
			nattr->attr.name = mod->sect_attrs->attrs[loaded].name;
			nattr->attr.mode = S_IRUGO;
R
Rusty Russell 已提交
1612 1613
			nattr->size = info->sechdrs[i].sh_size;
			nattr->private = (void *) info->sechdrs[i].sh_addr;
1614 1615 1616 1617 1618 1619
			nattr->read = module_notes_read;
			++nattr;
		}
		++loaded;
	}

1620
	notes_attrs->dir = kobject_create_and_add("notes", &mod->mkobj.kobj);
1621 1622 1623 1624 1625 1626 1627 1628 1629 1630 1631 1632 1633 1634 1635 1636 1637 1638 1639 1640 1641
	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
1642
#else
1643

R
Rusty Russell 已提交
1644 1645
static inline void add_sect_attrs(struct module *mod,
				  const struct load_info *info)
L
Linus Torvalds 已提交
1646 1647 1648 1649 1650 1651
{
}

static inline void remove_sect_attrs(struct module *mod)
{
}
1652

R
Rusty Russell 已提交
1653 1654
static inline void add_notes_attrs(struct module *mod,
				   const struct load_info *info)
1655 1656 1657 1658 1659 1660
{
}

static inline void remove_notes_attrs(struct module *mod)
{
}
R
Rusty Russell 已提交
1661
#endif /* CONFIG_KALLSYMS */
L
Linus Torvalds 已提交
1662

1663 1664 1665 1666 1667 1668
static void add_usage_links(struct module *mod)
{
#ifdef CONFIG_MODULE_UNLOAD
	struct module_use *use;
	int nowarn;

1669
	mutex_lock(&module_mutex);
1670 1671 1672 1673
	list_for_each_entry(use, &mod->target_list, target_list) {
		nowarn = sysfs_create_link(use->target->holders_dir,
					   &mod->mkobj.kobj, mod->name);
	}
1674
	mutex_unlock(&module_mutex);
1675 1676 1677 1678 1679 1680 1681 1682
#endif
}

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

1683
	mutex_lock(&module_mutex);
1684 1685
	list_for_each_entry(use, &mod->target_list, target_list)
		sysfs_remove_link(use->target->holders_dir, mod->name);
1686
	mutex_unlock(&module_mutex);
1687 1688 1689
#endif
}

1690
static int module_add_modinfo_attrs(struct module *mod)
1691 1692
{
	struct module_attribute *attr;
1693
	struct module_attribute *temp_attr;
1694 1695 1696
	int error = 0;
	int i;

1697 1698 1699 1700 1701 1702 1703
	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;
1704 1705
	for (i = 0; (attr = modinfo_attrs[i]) && !error; i++) {
		if (!attr->test ||
1706 1707
		    (attr->test && attr->test(mod))) {
			memcpy(temp_attr, attr, sizeof(*temp_attr));
1708
			sysfs_attr_init(&temp_attr->attr);
1709 1710
			error = sysfs_create_file(&mod->mkobj.kobj,
					&temp_attr->attr);
1711 1712
			++temp_attr;
		}
1713 1714 1715 1716
	}
	return error;
}

1717
static void module_remove_modinfo_attrs(struct module *mod)
1718 1719 1720 1721
{
	struct module_attribute *attr;
	int i;

1722 1723 1724 1725
	for (i = 0; (attr = &mod->modinfo_attrs[i]); i++) {
		/* pick a field to test for end of list */
		if (!attr->attr.name)
			break;
1726
		sysfs_remove_file(&mod->mkobj.kobj, &attr->attr);
1727 1728
		if (attr->free)
			attr->free(mod);
1729
	}
1730
	kfree(mod->modinfo_attrs);
1731
}
L
Linus Torvalds 已提交
1732

1733 1734 1735 1736 1737 1738 1739 1740
static void mod_kobject_put(struct module *mod)
{
	DECLARE_COMPLETION_ONSTACK(c);
	mod->mkobj.kobj_completion = &c;
	kobject_put(&mod->mkobj.kobj);
	wait_for_completion(&c);
}

1741
static int mod_sysfs_init(struct module *mod)
L
Linus Torvalds 已提交
1742 1743
{
	int err;
1744
	struct kobject *kobj;
L
Linus Torvalds 已提交
1745

1746
	if (!module_sysfs_initialized) {
1747
		pr_err("%s: module sysfs not initialized\n", mod->name);
1748 1749 1750
		err = -EINVAL;
		goto out;
	}
1751 1752 1753

	kobj = kset_find_obj(module_kset, mod->name);
	if (kobj) {
1754
		pr_err("%s: module is already loaded\n", mod->name);
1755 1756 1757 1758 1759
		kobject_put(kobj);
		err = -EINVAL;
		goto out;
	}

L
Linus Torvalds 已提交
1760
	mod->mkobj.mod = mod;
1761

1762 1763 1764 1765 1766
	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)
1767
		mod_kobject_put(mod);
K
Kay Sievers 已提交
1768

1769
	/* delay uevent until full sysfs population */
K
Kay Sievers 已提交
1770 1771 1772 1773
out:
	return err;
}

1774
static int mod_sysfs_setup(struct module *mod,
R
Rusty Russell 已提交
1775
			   const struct load_info *info,
K
Kay Sievers 已提交
1776 1777 1778 1779 1780
			   struct kernel_param *kparam,
			   unsigned int num_params)
{
	int err;

1781 1782 1783 1784
	err = mod_sysfs_init(mod);
	if (err)
		goto out;

1785
	mod->holders_dir = kobject_create_and_add("holders", &mod->mkobj.kobj);
1786 1787
	if (!mod->holders_dir) {
		err = -ENOMEM;
K
Kay Sievers 已提交
1788
		goto out_unreg;
1789
	}
K
Kay Sievers 已提交
1790

L
Linus Torvalds 已提交
1791 1792
	err = module_param_sysfs_setup(mod, kparam, num_params);
	if (err)
K
Kay Sievers 已提交
1793
		goto out_unreg_holders;
L
Linus Torvalds 已提交
1794

1795 1796
	err = module_add_modinfo_attrs(mod);
	if (err)
1797
		goto out_unreg_param;
1798

1799
	add_usage_links(mod);
R
Rusty Russell 已提交
1800 1801
	add_sect_attrs(mod, info);
	add_notes_attrs(mod, info);
1802

1803
	kobject_uevent(&mod->mkobj.kobj, KOBJ_ADD);
L
Linus Torvalds 已提交
1804 1805
	return 0;

1806 1807
out_unreg_param:
	module_param_sysfs_remove(mod);
K
Kay Sievers 已提交
1808
out_unreg_holders:
1809
	kobject_put(mod->holders_dir);
K
Kay Sievers 已提交
1810
out_unreg:
1811
	mod_kobject_put(mod);
1812
out:
L
Linus Torvalds 已提交
1813 1814
	return err;
}
1815 1816 1817

static void mod_sysfs_fini(struct module *mod)
{
R
Rusty Russell 已提交
1818 1819
	remove_notes_attrs(mod);
	remove_sect_attrs(mod);
1820
	mod_kobject_put(mod);
1821 1822
}

R
Rusty Russell 已提交
1823
#else /* !CONFIG_SYSFS */
1824

R
Rusty Russell 已提交
1825 1826
static int mod_sysfs_setup(struct module *mod,
			   const struct load_info *info,
1827 1828 1829 1830 1831 1832
			   struct kernel_param *kparam,
			   unsigned int num_params)
{
	return 0;
}

1833 1834 1835 1836
static void mod_sysfs_fini(struct module *mod)
{
}

1837 1838 1839 1840
static void module_remove_modinfo_attrs(struct module *mod)
{
}

1841 1842 1843 1844
static void del_usage_links(struct module *mod)
{
}

1845
#endif /* CONFIG_SYSFS */
L
Linus Torvalds 已提交
1846

1847
static void mod_sysfs_teardown(struct module *mod)
L
Linus Torvalds 已提交
1848
{
1849
	del_usage_links(mod);
1850
	module_remove_modinfo_attrs(mod);
L
Linus Torvalds 已提交
1851
	module_param_sysfs_remove(mod);
1852 1853
	kobject_put(mod->mkobj.drivers_dir);
	kobject_put(mod->holders_dir);
1854
	mod_sysfs_fini(mod);
L
Linus Torvalds 已提交
1855 1856
}

1857 1858 1859 1860 1861 1862 1863 1864 1865 1866 1867 1868 1869 1870 1871 1872 1873 1874 1875 1876 1877 1878 1879 1880 1881 1882 1883 1884 1885 1886 1887 1888 1889 1890 1891 1892 1893 1894 1895 1896 1897 1898 1899 1900
#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);
	}
}

1901 1902 1903 1904 1905 1906 1907 1908 1909 1910 1911 1912 1913 1914 1915 1916 1917 1918
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);
1919 1920 1921
}

/* Iterate through all modules and set each module's text as RW */
1922
void set_all_modules_text_rw(void)
1923 1924 1925 1926 1927
{
	struct module *mod;

	mutex_lock(&module_mutex);
	list_for_each_entry_rcu(mod, &modules, list) {
1928 1929
		if (mod->state == MODULE_STATE_UNFORMED)
			continue;
1930 1931 1932 1933 1934 1935 1936 1937 1938 1939 1940 1941 1942 1943 1944
		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 */
1945
void set_all_modules_text_ro(void)
1946 1947 1948 1949 1950
{
	struct module *mod;

	mutex_lock(&module_mutex);
	list_for_each_entry_rcu(mod, &modules, list) {
1951 1952
		if (mod->state == MODULE_STATE_UNFORMED)
			continue;
1953 1954 1955 1956 1957 1958 1959 1960 1961 1962 1963 1964 1965 1966 1967
		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) { }
1968 1969
static void unset_module_core_ro_nx(struct module *mod) { }
static void unset_module_init_ro_nx(struct module *mod) { }
1970 1971
#endif

1972
void __weak module_memfree(void *module_region)
1973 1974 1975 1976 1977 1978 1979 1980
{
	vfree(module_region);
}

void __weak module_arch_cleanup(struct module *mod)
{
}

1981 1982 1983 1984
void __weak module_arch_freeing_init(struct module *mod)
{
}

1985
/* Free a module, remove from lists, etc. */
L
Linus Torvalds 已提交
1986 1987
static void free_module(struct module *mod)
{
1988 1989
	trace_module_free(mod);

1990
	mod_sysfs_teardown(mod);
L
Linus Torvalds 已提交
1991

1992 1993
	/* We leave it in list to prevent duplicate loads, but make sure
	 * that noone uses it while it's being deconstructed. */
1994
	mutex_lock(&module_mutex);
1995
	mod->state = MODULE_STATE_UNFORMED;
1996
	mutex_unlock(&module_mutex);
1997

1998 1999 2000
	/* Remove dynamic debug info */
	ddebug_remove_module(mod->name);

L
Linus Torvalds 已提交
2001 2002 2003 2004 2005 2006
	/* Arch-specific cleanup. */
	module_arch_cleanup(mod);

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

2007 2008 2009
	/* Free any allocated parameters. */
	destroy_params(mod->kp, mod->num_kp);

2010 2011
	/* Now we can delete it from the lists */
	mutex_lock(&module_mutex);
2012 2013
	/* Unlink carefully: kallsyms could be walking list. */
	list_del_rcu(&mod->list);
2014
	mod_tree_remove(mod);
2015
	/* Remove this module from bug list, this uses list_del_rcu */
2016
	module_bug_cleanup(mod);
2017 2018
	/* Wait for RCU-sched synchronizing before releasing mod->list and buglist. */
	synchronize_sched();
2019 2020
	mutex_unlock(&module_mutex);

L
Linus Torvalds 已提交
2021
	/* This may be NULL, but that's OK */
2022
	unset_module_init_ro_nx(mod);
2023
	module_arch_freeing_init(mod);
2024
	module_memfree(mod->module_init);
L
Linus Torvalds 已提交
2025
	kfree(mod->args);
2026
	percpu_modfree(mod);
2027

2028
	/* Free lock-classes; relies on the preceding sync_rcu(). */
I
Ingo Molnar 已提交
2029 2030
	lockdep_free_key_range(mod->module_core, mod->core_size);

L
Linus Torvalds 已提交
2031
	/* Finally, free the core (containing the module structure) */
2032
	unset_module_core_ro_nx(mod);
2033
	module_memfree(mod->module_core);
2034 2035 2036 2037

#ifdef CONFIG_MPU
	update_protections(current->mm);
#endif
L
Linus Torvalds 已提交
2038 2039 2040 2041 2042
}

void *__symbol_get(const char *symbol)
{
	struct module *owner;
2043
	const struct kernel_symbol *sym;
L
Linus Torvalds 已提交
2044

2045
	preempt_disable();
2046 2047 2048
	sym = find_symbol(symbol, &owner, NULL, true, true);
	if (sym && strong_try_module_get(owner))
		sym = NULL;
2049
	preempt_enable();
L
Linus Torvalds 已提交
2050

2051
	return sym ? (void *)sym->value : NULL;
L
Linus Torvalds 已提交
2052 2053 2054
}
EXPORT_SYMBOL_GPL(__symbol_get);

2055 2056
/*
 * Ensure that an exported symbol [global namespace] does not already exist
2057
 * in the kernel or in some other module's exported symbol table.
2058 2059
 *
 * You must hold the module_mutex.
2060 2061 2062
 */
static int verify_export_symbols(struct module *mod)
{
2063
	unsigned int i;
2064
	struct module *owner;
2065 2066 2067 2068 2069 2070 2071 2072
	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 },
2073
#ifdef CONFIG_UNUSED_SYMBOLS
2074 2075
		{ mod->unused_syms, mod->num_unused_syms },
		{ mod->unused_gpl_syms, mod->num_unused_gpl_syms },
2076
#endif
2077
	};
2078

2079 2080
	for (i = 0; i < ARRAY_SIZE(arr); i++) {
		for (s = arr[i].sym; s < arr[i].sym + arr[i].num; s++) {
2081
			if (find_symbol(s->name, &owner, NULL, true, false)) {
2082
				pr_err("%s: exports duplicate symbol %s"
2083 2084 2085 2086
				       " (owned by %s)\n",
				       mod->name, s->name, module_name(owner));
				return -ENOEXEC;
			}
2087
		}
2088 2089
	}
	return 0;
2090 2091
}

2092
/* Change all symbols so that st_value encodes the pointer directly. */
2093 2094 2095 2096
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
2097
	unsigned long secbase;
2098
	unsigned int i;
L
Linus Torvalds 已提交
2099
	int ret = 0;
2100
	const struct kernel_symbol *ksym;
L
Linus Torvalds 已提交
2101

2102 2103 2104
	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
2105 2106
		switch (sym[i].st_shndx) {
		case SHN_COMMON:
2107 2108 2109 2110
			/* Ignore common symbols */
			if (!strncmp(name, "__gnu_lto", 9))
				break;

L
Linus Torvalds 已提交
2111 2112
			/* We compiled with -fno-common.  These are not
			   supposed to happen.  */
2113
			pr_debug("Common symbol: %s\n", name);
2114
			pr_warn("%s: please compile with -fno-common\n",
L
Linus Torvalds 已提交
2115 2116 2117 2118 2119 2120
			       mod->name);
			ret = -ENOEXEC;
			break;

		case SHN_ABS:
			/* Don't need to do anything */
2121
			pr_debug("Absolute symbol: 0x%08lx\n",
L
Linus Torvalds 已提交
2122 2123 2124 2125
			       (long)sym[i].st_value);
			break;

		case SHN_UNDEF:
2126
			ksym = resolve_symbol_wait(mod, info, name);
L
Linus Torvalds 已提交
2127
			/* Ok if resolved.  */
2128
			if (ksym && !IS_ERR(ksym)) {
2129
				sym[i].st_value = ksym->value;
L
Linus Torvalds 已提交
2130
				break;
2131 2132
			}

L
Linus Torvalds 已提交
2133
			/* Ok if weak.  */
2134
			if (!ksym && ELF_ST_BIND(sym[i].st_info) == STB_WEAK)
L
Linus Torvalds 已提交
2135 2136
				break;

2137 2138
			pr_warn("%s: Unknown symbol %s (err %li)\n",
				mod->name, name, PTR_ERR(ksym));
2139
			ret = PTR_ERR(ksym) ?: -ENOENT;
L
Linus Torvalds 已提交
2140 2141 2142 2143
			break;

		default:
			/* Divert to percpu allocation if a percpu var. */
2144
			if (sym[i].st_shndx == info->index.pcpu)
2145
				secbase = (unsigned long)mod_percpu(mod);
L
Linus Torvalds 已提交
2146
			else
2147
				secbase = info->sechdrs[sym[i].st_shndx].sh_addr;
L
Linus Torvalds 已提交
2148 2149 2150 2151 2152 2153 2154 2155
			sym[i].st_value += secbase;
			break;
		}
	}

	return ret;
}

2156
static int apply_relocations(struct module *mod, const struct load_info *info)
2157 2158 2159 2160 2161
{
	unsigned int i;
	int err = 0;

	/* Now do relocations. */
2162 2163
	for (i = 1; i < info->hdr->e_shnum; i++) {
		unsigned int infosec = info->sechdrs[i].sh_info;
2164 2165

		/* Not a valid relocation section? */
2166
		if (infosec >= info->hdr->e_shnum)
2167 2168 2169
			continue;

		/* Don't bother with non-allocated sections */
2170
		if (!(info->sechdrs[infosec].sh_flags & SHF_ALLOC))
2171 2172
			continue;

2173 2174 2175 2176 2177 2178
		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);
2179 2180 2181 2182 2183 2184
		if (err < 0)
			break;
	}
	return err;
}

2185 2186 2187 2188 2189 2190 2191 2192
/* 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
2193
/* Update size with this section: return offset. */
2194 2195
static long get_offset(struct module *mod, unsigned int *size,
		       Elf_Shdr *sechdr, unsigned int section)
L
Linus Torvalds 已提交
2196 2197 2198
{
	long ret;

2199
	*size += arch_mod_section_prepend(mod, section);
L
Linus Torvalds 已提交
2200 2201 2202 2203 2204 2205 2206 2207 2208
	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. */
2209
static void layout_sections(struct module *mod, struct load_info *info)
L
Linus Torvalds 已提交
2210 2211 2212 2213 2214 2215 2216 2217 2218 2219 2220 2221
{
	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;

2222 2223
	for (i = 0; i < info->hdr->e_shnum; i++)
		info->sechdrs[i].sh_entsize = ~0UL;
L
Linus Torvalds 已提交
2224

2225
	pr_debug("Core section allocation order:\n");
L
Linus Torvalds 已提交
2226
	for (m = 0; m < ARRAY_SIZE(masks); ++m) {
2227 2228 2229
		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
2230 2231 2232 2233

			if ((s->sh_flags & masks[m][0]) != masks[m][0]
			    || (s->sh_flags & masks[m][1])
			    || s->sh_entsize != ~0UL
2234
			    || strstarts(sname, ".init"))
L
Linus Torvalds 已提交
2235
				continue;
2236
			s->sh_entsize = get_offset(mod, &mod->core_size, s, i);
2237
			pr_debug("\t%s\n", sname);
L
Linus Torvalds 已提交
2238
		}
2239 2240 2241
		switch (m) {
		case 0: /* executable */
			mod->core_size = debug_align(mod->core_size);
L
Linus Torvalds 已提交
2242
			mod->core_text_size = mod->core_size;
2243 2244 2245 2246 2247 2248 2249 2250 2251
			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
2252 2253
	}

2254
	pr_debug("Init section allocation order:\n");
L
Linus Torvalds 已提交
2255
	for (m = 0; m < ARRAY_SIZE(masks); ++m) {
2256 2257 2258
		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
2259 2260 2261 2262

			if ((s->sh_flags & masks[m][0]) != masks[m][0]
			    || (s->sh_flags & masks[m][1])
			    || s->sh_entsize != ~0UL
2263
			    || !strstarts(sname, ".init"))
L
Linus Torvalds 已提交
2264
				continue;
2265
			s->sh_entsize = (get_offset(mod, &mod->init_size, s, i)
L
Linus Torvalds 已提交
2266
					 | INIT_OFFSET_MASK);
2267
			pr_debug("\t%s\n", sname);
L
Linus Torvalds 已提交
2268
		}
2269 2270 2271
		switch (m) {
		case 0: /* executable */
			mod->init_size = debug_align(mod->init_size);
L
Linus Torvalds 已提交
2272
			mod->init_text_size = mod->init_size;
2273 2274 2275 2276 2277 2278 2279 2280 2281
			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
2282 2283 2284 2285 2286 2287 2288 2289
	}
}

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

2290
	if (!license_is_gpl_compatible(license)) {
2291
		if (!test_taint(TAINT_PROPRIETARY_MODULE))
2292 2293
			pr_warn("%s: module license '%s' taints kernel.\n",
				mod->name, license);
2294 2295
		add_taint_module(mod, TAINT_PROPRIETARY_MODULE,
				 LOCKDEP_NOW_UNRELIABLE);
L
Linus Torvalds 已提交
2296 2297 2298 2299 2300 2301 2302 2303 2304 2305 2306 2307 2308 2309 2310 2311 2312 2313 2314 2315 2316 2317
	}
}

/* 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;
}

2318
static char *get_modinfo(struct load_info *info, const char *tag)
L
Linus Torvalds 已提交
2319 2320 2321
{
	char *p;
	unsigned int taglen = strlen(tag);
2322 2323
	Elf_Shdr *infosec = &info->sechdrs[info->index.info];
	unsigned long size = infosec->sh_size;
L
Linus Torvalds 已提交
2324

2325
	for (p = (char *)infosec->sh_addr; p; p = next_string(p, &size)) {
L
Linus Torvalds 已提交
2326 2327 2328 2329 2330 2331
		if (strncmp(p, tag, taglen) == 0 && p[taglen] == '=')
			return p + taglen + 1;
	}
	return NULL;
}

2332
static void setup_modinfo(struct module *mod, struct load_info *info)
2333 2334 2335 2336 2337 2338
{
	struct module_attribute *attr;
	int i;

	for (i = 0; (attr = modinfo_attrs[i]); i++) {
		if (attr->setup)
2339
			attr->setup(mod, get_modinfo(info, attr->attr.name));
2340 2341 2342
	}
}

2343 2344 2345 2346 2347 2348 2349 2350 2351 2352 2353
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
2354
#ifdef CONFIG_KALLSYMS
2355 2356 2357 2358 2359 2360

/* 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)
{
2361 2362
	return bsearch(name, start, stop - start,
			sizeof(struct kernel_symbol), cmp_name);
2363 2364
}

2365 2366
static int is_exported(const char *name, unsigned long value,
		       const struct module *mod)
L
Linus Torvalds 已提交
2367
{
2368 2369 2370
	const struct kernel_symbol *ks;
	if (!mod)
		ks = lookup_symbol(name, __start___ksymtab, __stop___ksymtab);
2371
	else
2372 2373
		ks = lookup_symbol(name, mod->syms, mod->syms + mod->num_syms);
	return ks != NULL && ks->value == value;
L
Linus Torvalds 已提交
2374 2375 2376
}

/* As per nm */
2377
static char elf_type(const Elf_Sym *sym, const struct load_info *info)
L
Linus Torvalds 已提交
2378
{
2379 2380
	const Elf_Shdr *sechdrs = info->sechdrs;

L
Linus Torvalds 已提交
2381 2382 2383 2384 2385 2386 2387 2388 2389 2390 2391 2392 2393 2394 2395 2396 2397 2398 2399 2400 2401 2402 2403 2404 2405 2406 2407 2408 2409
	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';
	}
2410 2411
	if (strstarts(info->secstrings + sechdrs[sym->st_shndx].sh_name,
		      ".debug")) {
L
Linus Torvalds 已提交
2412
		return 'n';
2413
	}
L
Linus Torvalds 已提交
2414 2415 2416
	return '?';
}

2417
static bool is_core_symbol(const Elf_Sym *src, const Elf_Shdr *sechdrs,
2418
			unsigned int shnum)
2419 2420 2421 2422 2423 2424 2425 2426 2427 2428 2429 2430 2431 2432 2433 2434 2435 2436 2437
{
	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;
}

2438 2439 2440 2441 2442 2443 2444
/*
 * 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>.
 */
2445
static void layout_symtab(struct module *mod, struct load_info *info)
2446
{
2447 2448
	Elf_Shdr *symsect = info->sechdrs + info->index.sym;
	Elf_Shdr *strsect = info->sechdrs + info->index.str;
2449
	const Elf_Sym *src;
2450
	unsigned int i, nsrc, ndst, strtab_size = 0;
2451 2452 2453 2454

	/* 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,
2455
					 info->index.sym) | INIT_OFFSET_MASK;
2456
	pr_debug("\t%s\n", info->secstrings + symsect->sh_name);
2457

2458
	src = (void *)info->hdr + symsect->sh_offset;
2459
	nsrc = symsect->sh_size / sizeof(*src);
2460

2461
	/* Compute total space required for the core symbols' strtab. */
2462 2463 2464 2465
	for (ndst = i = 0; i < nsrc; i++) {
		if (i == 0 ||
		    is_core_symbol(src+i, info->sechdrs, info->hdr->e_shnum)) {
			strtab_size += strlen(&info->strtab[src[i].st_name])+1;
2466
			ndst++;
2467
		}
2468
	}
2469 2470

	/* Append room for core symbols at end of core part. */
2471
	info->symoffs = ALIGN(mod->core_size, symsect->sh_addralign ?: 1);
2472 2473
	info->stroffs = mod->core_size = info->symoffs + ndst * sizeof(Elf_Sym);
	mod->core_size += strtab_size;
2474
	mod->core_size = debug_align(mod->core_size);
2475

2476 2477 2478
	/* 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,
2479
					 info->index.str) | INIT_OFFSET_MASK;
2480
	mod->init_size = debug_align(mod->init_size);
2481
	pr_debug("\t%s\n", info->secstrings + strsect->sh_name);
2482 2483
}

2484
static void add_kallsyms(struct module *mod, const struct load_info *info)
L
Linus Torvalds 已提交
2485
{
2486 2487 2488
	unsigned int i, ndst;
	const Elf_Sym *src;
	Elf_Sym *dst;
2489
	char *s;
2490
	Elf_Shdr *symsec = &info->sechdrs[info->index.sym];
L
Linus Torvalds 已提交
2491

2492 2493
	mod->symtab = (void *)symsec->sh_addr;
	mod->num_symtab = symsec->sh_size / sizeof(Elf_Sym);
2494 2495
	/* 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
2496 2497 2498

	/* Set types up while we still have access to sections. */
	for (i = 0; i < mod->num_symtab; i++)
2499
		mod->symtab[i].st_info = elf_type(&mod->symtab[i], info);
2500

2501
	mod->core_symtab = dst = mod->module_core + info->symoffs;
2502
	mod->core_strtab = s = mod->module_core + info->stroffs;
2503
	src = mod->symtab;
2504 2505 2506 2507 2508 2509 2510 2511
	for (ndst = i = 0; i < mod->num_symtab; i++) {
		if (i == 0 ||
		    is_core_symbol(src+i, info->sechdrs, info->hdr->e_shnum)) {
			dst[ndst] = src[i];
			dst[ndst++].st_name = s - mod->core_strtab;
			s += strlcpy(s, &mod->strtab[src[i].st_name],
				     KSYM_NAME_LEN) + 1;
		}
2512 2513
	}
	mod->core_num_syms = ndst;
L
Linus Torvalds 已提交
2514 2515
}
#else
2516
static inline void layout_symtab(struct module *mod, struct load_info *info)
2517 2518
{
}
2519

2520
static void add_kallsyms(struct module *mod, const struct load_info *info)
L
Linus Torvalds 已提交
2521 2522 2523 2524
{
}
#endif /* CONFIG_KALLSYMS */

2525
static void dynamic_debug_setup(struct _ddebug *debug, unsigned int num)
2526
{
2527 2528
	if (!debug)
		return;
2529 2530
#ifdef CONFIG_DYNAMIC_DEBUG
	if (ddebug_add_module(debug, num, debug->modname))
2531 2532
		pr_err("dynamic debug error adding module: %s\n",
			debug->modname);
2533
#endif
2534
}
2535

2536 2537 2538 2539 2540 2541
static void dynamic_debug_remove(struct _ddebug *debug)
{
	if (debug)
		ddebug_remove_module(debug->modname);
}

2542 2543
void * __weak module_alloc(unsigned long size)
{
2544
	return vmalloc_exec(size);
2545 2546
}

2547
#ifdef CONFIG_DEBUG_KMEMLEAK
2548 2549
static void kmemleak_load_module(const struct module *mod,
				 const struct load_info *info)
2550 2551 2552 2553
{
	unsigned int i;

	/* only scan the sections containing data */
2554
	kmemleak_scan_area(mod, sizeof(struct module), GFP_KERNEL);
2555

2556
	for (i = 1; i < info->hdr->e_shnum; i++) {
2557 2558 2559 2560
		/* Scan all writable sections that's not executable */
		if (!(info->sechdrs[i].sh_flags & SHF_ALLOC) ||
		    !(info->sechdrs[i].sh_flags & SHF_WRITE) ||
		    (info->sechdrs[i].sh_flags & SHF_EXECINSTR))
2561 2562
			continue;

2563 2564
		kmemleak_scan_area((void *)info->sechdrs[i].sh_addr,
				   info->sechdrs[i].sh_size, GFP_KERNEL);
2565 2566 2567
	}
}
#else
2568 2569
static inline void kmemleak_load_module(const struct module *mod,
					const struct load_info *info)
2570 2571 2572 2573
{
}
#endif

2574
#ifdef CONFIG_MODULE_SIG
2575
static int module_sig_check(struct load_info *info)
2576 2577
{
	int err = -ENOKEY;
2578 2579
	const unsigned long markerlen = sizeof(MODULE_SIG_STRING) - 1;
	const void *mod = info->hdr;
2580

2581 2582
	if (info->len > markerlen &&
	    memcmp(mod + info->len - markerlen, MODULE_SIG_STRING, markerlen) == 0) {
2583
		/* We truncate the module to discard the signature */
2584 2585
		info->len -= markerlen;
		err = mod_verify_sig(mod, &info->len);
2586 2587 2588 2589 2590 2591 2592 2593 2594 2595 2596 2597 2598 2599
	}

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

	/* Not having a signature is only an error if we're strict. */
	if (err == -ENOKEY && !sig_enforce)
		err = 0;

	return err;
}
#else /* !CONFIG_MODULE_SIG */
2600
static int module_sig_check(struct load_info *info)
2601 2602 2603 2604 2605
{
	return 0;
}
#endif /* !CONFIG_MODULE_SIG */

2606 2607
/* Sanity checks against invalid binaries, wrong arch, weird elf version. */
static int elf_header_check(struct load_info *info)
2608
{
2609 2610 2611 2612 2613 2614 2615 2616 2617 2618 2619 2620 2621
	if (info->len < sizeof(*(info->hdr)))
		return -ENOEXEC;

	if (memcmp(info->hdr->e_ident, ELFMAG, SELFMAG) != 0
	    || info->hdr->e_type != ET_REL
	    || !elf_check_arch(info->hdr)
	    || info->hdr->e_shentsize != sizeof(Elf_Shdr))
		return -ENOEXEC;

	if (info->hdr->e_shoff >= info->len
	    || (info->hdr->e_shnum * sizeof(Elf_Shdr) >
		info->len - info->hdr->e_shoff))
		return -ENOEXEC;
2622

2623 2624 2625
	return 0;
}

2626 2627 2628 2629 2630 2631 2632 2633 2634 2635 2636 2637 2638 2639 2640 2641 2642
#define COPY_CHUNK_SIZE (16*PAGE_SIZE)

static int copy_chunked_from_user(void *dst, const void __user *usrc, unsigned long len)
{
	do {
		unsigned long n = min(len, COPY_CHUNK_SIZE);

		if (copy_from_user(dst, usrc, n) != 0)
			return -EFAULT;
		cond_resched();
		dst += n;
		usrc += n;
		len -= n;
	} while (len);
	return 0;
}

2643 2644 2645
/* Sets info->hdr and info->len. */
static int copy_module_from_user(const void __user *umod, unsigned long len,
				  struct load_info *info)
2646 2647 2648
{
	int err;

2649 2650
	info->len = len;
	if (info->len < sizeof(*(info->hdr)))
2651 2652
		return -ENOEXEC;

2653 2654 2655 2656
	err = security_kernel_module_from_file(NULL);
	if (err)
		return err;

2657
	/* Suck in entire file: we'll want most of it. */
2658 2659
	info->hdr = __vmalloc(info->len,
			GFP_KERNEL | __GFP_HIGHMEM | __GFP_NOWARN, PAGE_KERNEL);
2660
	if (!info->hdr)
2661 2662
		return -ENOMEM;

2663
	if (copy_chunked_from_user(info->hdr, umod, info->len) != 0) {
2664 2665
		vfree(info->hdr);
		return -EFAULT;
2666 2667
	}

2668 2669 2670 2671 2672 2673
	return 0;
}

/* Sets info->hdr and info->len. */
static int copy_module_from_fd(int fd, struct load_info *info)
{
2674
	struct fd f = fdget(fd);
2675 2676 2677 2678 2679
	int err;
	struct kstat stat;
	loff_t pos;
	ssize_t bytes = 0;

2680
	if (!f.file)
2681 2682
		return -ENOEXEC;

2683
	err = security_kernel_module_from_file(f.file);
2684
	if (err)
2685
		goto out;
2686

2687
	err = vfs_getattr(&f.file->f_path, &stat);
2688
	if (err)
2689
		goto out;
2690

2691 2692 2693
	if (stat.size > INT_MAX) {
		err = -EFBIG;
		goto out;
2694
	}
2695 2696 2697 2698 2699 2700 2701

	/* Don't hand 0 to vmalloc, it whines. */
	if (stat.size == 0) {
		err = -EINVAL;
		goto out;
	}

2702 2703 2704 2705
	info->hdr = vmalloc(stat.size);
	if (!info->hdr) {
		err = -ENOMEM;
		goto out;
2706
	}
2707

2708 2709
	pos = 0;
	while (pos < stat.size) {
2710
		bytes = kernel_read(f.file, pos, (char *)(info->hdr) + pos,
2711 2712 2713 2714 2715 2716 2717 2718 2719 2720 2721
				    stat.size - pos);
		if (bytes < 0) {
			vfree(info->hdr);
			err = bytes;
			goto out;
		}
		if (bytes == 0)
			break;
		pos += bytes;
	}
	info->len = pos;
2722

2723
out:
2724
	fdput(f);
2725 2726 2727
	return err;
}

2728 2729 2730 2731 2732
static void free_copy(struct load_info *info)
{
	vfree(info->hdr);
}

2733
static int rewrite_section_headers(struct load_info *info, int flags)
2734 2735 2736 2737 2738 2739 2740 2741 2742 2743
{
	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) {
2744
			pr_err("Module len %lu truncated\n", info->len);
2745 2746 2747 2748 2749 2750 2751 2752 2753 2754 2755 2756 2757
			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
	}
2758 2759

	/* Track but don't keep modinfo and version sections. */
2760 2761 2762 2763
	if (flags & MODULE_INIT_IGNORE_MODVERSIONS)
		info->index.vers = 0; /* Pretend no __versions section! */
	else
		info->index.vers = find_sec(info, "__versions");
2764
	info->index.info = find_sec(info, ".modinfo");
2765 2766
	info->sechdrs[info->index.info].sh_flags &= ~(unsigned long)SHF_ALLOC;
	info->sechdrs[info->index.vers].sh_flags &= ~(unsigned long)SHF_ALLOC;
2767 2768 2769
	return 0;
}

L
Linus Torvalds 已提交
2770 2771 2772 2773 2774 2775 2776 2777
/*
 * 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).
 */
2778
static struct module *setup_load_info(struct load_info *info, int flags)
L
Linus Torvalds 已提交
2779 2780
{
	unsigned int i;
2781
	int err;
L
Linus Torvalds 已提交
2782 2783 2784 2785
	struct module *mod;

	/* Set up the convenience variables */
	info->sechdrs = (void *)info->hdr + info->hdr->e_shoff;
2786 2787
	info->secstrings = (void *)info->hdr
		+ info->sechdrs[info->hdr->e_shstrndx].sh_offset;
L
Linus Torvalds 已提交
2788

2789
	err = rewrite_section_headers(info, flags);
2790 2791
	if (err)
		return ERR_PTR(err);
L
Linus Torvalds 已提交
2792

2793 2794
	/* Find internal symbols and strings. */
	for (i = 1; i < info->hdr->e_shnum; i++) {
L
Linus Torvalds 已提交
2795 2796 2797
		if (info->sechdrs[i].sh_type == SHT_SYMTAB) {
			info->index.sym = i;
			info->index.str = info->sechdrs[i].sh_link;
2798 2799 2800
			info->strtab = (char *)info->hdr
				+ info->sechdrs[info->index.str].sh_offset;
			break;
L
Linus Torvalds 已提交
2801 2802 2803
		}
	}

2804
	info->index.mod = find_sec(info, ".gnu.linkonce.this_module");
L
Linus Torvalds 已提交
2805
	if (!info->index.mod) {
2806
		pr_warn("No module found in object\n");
L
Linus Torvalds 已提交
2807 2808 2809 2810 2811 2812
		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) {
2813
		pr_warn("%s: module has no symbols (stripped?)\n", mod->name);
L
Linus Torvalds 已提交
2814 2815 2816
		return ERR_PTR(-ENOEXEC);
	}

2817
	info->index.pcpu = find_pcpusec(info);
L
Linus Torvalds 已提交
2818 2819 2820 2821 2822 2823 2824 2825

	/* 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;
}

2826
static int check_modinfo(struct module *mod, struct load_info *info, int flags)
2827
{
2828
	const char *modmagic = get_modinfo(info, "vermagic");
2829 2830
	int err;

2831 2832 2833
	if (flags & MODULE_INIT_IGNORE_VERMAGIC)
		modmagic = NULL;

2834 2835 2836 2837 2838
	/* This is allowed: modprobe --force will invalidate it. */
	if (!modmagic) {
		err = try_to_force_load(mod, "bad vermagic");
		if (err)
			return err;
2839
	} else if (!same_magic(modmagic, vermagic, info->index.vers)) {
2840
		pr_err("%s: version magic '%s' should be '%s'\n",
2841 2842 2843 2844
		       mod->name, modmagic, vermagic);
		return -ENOEXEC;
	}

2845
	if (!get_modinfo(info, "intree"))
2846
		add_taint_module(mod, TAINT_OOT_MODULE, LOCKDEP_STILL_OK);
2847

2848
	if (get_modinfo(info, "staging")) {
2849
		add_taint_module(mod, TAINT_CRAP, LOCKDEP_STILL_OK);
2850 2851
		pr_warn("%s: module is from the staging directory, the quality "
			"is unknown, you have been warned.\n", mod->name);
2852
	}
2853 2854

	/* Set up license info based on the info section */
2855
	set_license(mod, get_modinfo(info, "license"));
2856

2857 2858 2859
	return 0;
}

2860
static int find_module_sections(struct module *mod, struct load_info *info)
2861
{
2862
	mod->kp = section_objs(info, "__param",
2863
			       sizeof(*mod->kp), &mod->num_kp);
2864
	mod->syms = section_objs(info, "__ksymtab",
2865
				 sizeof(*mod->syms), &mod->num_syms);
2866 2867
	mod->crcs = section_addr(info, "__kcrctab");
	mod->gpl_syms = section_objs(info, "__ksymtab_gpl",
2868 2869
				     sizeof(*mod->gpl_syms),
				     &mod->num_gpl_syms);
2870 2871
	mod->gpl_crcs = section_addr(info, "__kcrctab_gpl");
	mod->gpl_future_syms = section_objs(info,
2872 2873 2874
					    "__ksymtab_gpl_future",
					    sizeof(*mod->gpl_future_syms),
					    &mod->num_gpl_future_syms);
2875
	mod->gpl_future_crcs = section_addr(info, "__kcrctab_gpl_future");
2876 2877

#ifdef CONFIG_UNUSED_SYMBOLS
2878
	mod->unused_syms = section_objs(info, "__ksymtab_unused",
2879 2880
					sizeof(*mod->unused_syms),
					&mod->num_unused_syms);
2881 2882
	mod->unused_crcs = section_addr(info, "__kcrctab_unused");
	mod->unused_gpl_syms = section_objs(info, "__ksymtab_unused_gpl",
2883 2884
					    sizeof(*mod->unused_gpl_syms),
					    &mod->num_unused_gpl_syms);
2885
	mod->unused_gpl_crcs = section_addr(info, "__kcrctab_unused_gpl");
2886 2887
#endif
#ifdef CONFIG_CONSTRUCTORS
2888
	mod->ctors = section_objs(info, ".ctors",
2889
				  sizeof(*mod->ctors), &mod->num_ctors);
2890 2891 2892 2893 2894 2895 2896 2897
	if (!mod->ctors)
		mod->ctors = section_objs(info, ".init_array",
				sizeof(*mod->ctors), &mod->num_ctors);
	else if (find_sec(info, ".init_array")) {
		/*
		 * This shouldn't happen with same compiler and binutils
		 * building all parts of the module.
		 */
2898
		pr_warn("%s: has both .ctors and .init_array.\n",
2899 2900 2901
		       mod->name);
		return -EINVAL;
	}
2902 2903 2904
#endif

#ifdef CONFIG_TRACEPOINTS
2905 2906 2907
	mod->tracepoints_ptrs = section_objs(info, "__tracepoints_ptrs",
					     sizeof(*mod->tracepoints_ptrs),
					     &mod->num_tracepoints);
2908
#endif
2909 2910 2911 2912 2913
#ifdef HAVE_JUMP_LABEL
	mod->jump_entries = section_objs(info, "__jump_table",
					sizeof(*mod->jump_entries),
					&mod->num_jump_entries);
#endif
2914
#ifdef CONFIG_EVENT_TRACING
2915
	mod->trace_events = section_objs(info, "_ftrace_events",
2916 2917
					 sizeof(*mod->trace_events),
					 &mod->num_trace_events);
2918 2919 2920
	mod->trace_enums = section_objs(info, "_ftrace_enum_map",
					sizeof(*mod->trace_enums),
					&mod->num_trace_enums);
2921
#endif
2922 2923 2924 2925 2926
#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);
#endif
2927 2928
#ifdef CONFIG_FTRACE_MCOUNT_RECORD
	/* sechdrs[0].sh_size is always zero */
2929
	mod->ftrace_callsites = section_objs(info, "__mcount_loc",
2930 2931 2932
					     sizeof(*mod->ftrace_callsites),
					     &mod->num_ftrace_callsites);
#endif
2933

2934 2935 2936
	mod->extable = section_objs(info, "__ex_table",
				    sizeof(*mod->extable), &mod->num_exentries);

2937
	if (section_addr(info, "__obsparm"))
2938
		pr_warn("%s: Ignoring obsolete parameters\n", mod->name);
2939 2940 2941

	info->debug = section_objs(info, "__verbose",
				   sizeof(*info->debug), &info->num_debug);
2942 2943

	return 0;
2944 2945
}

2946
static int move_module(struct module *mod, struct load_info *info)
2947 2948 2949 2950 2951
{
	int i;
	void *ptr;

	/* Do the allocs. */
2952
	ptr = module_alloc(mod->core_size);
2953 2954 2955 2956 2957 2958 2959
	/*
	 * 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)
2960
		return -ENOMEM;
2961 2962 2963 2964

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

2965
	if (mod->init_size) {
2966
		ptr = module_alloc(mod->init_size);
2967 2968 2969 2970 2971 2972 2973 2974
		/*
		 * 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) {
2975
			module_memfree(mod->module_core);
2976 2977 2978 2979 2980 2981
			return -ENOMEM;
		}
		memset(ptr, 0, mod->init_size);
		mod->module_init = ptr;
	} else
		mod->module_init = NULL;
2982 2983

	/* Transfer each section which specifies SHF_ALLOC */
2984
	pr_debug("final section addresses:\n");
2985
	for (i = 0; i < info->hdr->e_shnum; i++) {
2986
		void *dest;
2987
		Elf_Shdr *shdr = &info->sechdrs[i];
2988

2989
		if (!(shdr->sh_flags & SHF_ALLOC))
2990 2991
			continue;

2992
		if (shdr->sh_entsize & INIT_OFFSET_MASK)
2993
			dest = mod->module_init
2994
				+ (shdr->sh_entsize & ~INIT_OFFSET_MASK);
2995
		else
2996
			dest = mod->module_core + shdr->sh_entsize;
2997

2998 2999
		if (shdr->sh_type != SHT_NOBITS)
			memcpy(dest, (void *)shdr->sh_addr, shdr->sh_size);
3000
		/* Update sh_addr to point to copy in image. */
3001
		shdr->sh_addr = (unsigned long)dest;
3002 3003
		pr_debug("\t0x%lx %s\n",
			 (long)shdr->sh_addr, info->secstrings + shdr->sh_name);
3004
	}
3005 3006

	return 0;
3007 3008
}

3009
static int check_module_license_and_versions(struct module *mod)
3010 3011 3012 3013 3014 3015 3016
{
	/*
	 * 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)
3017
		add_taint(TAINT_PROPRIETARY_MODULE, LOCKDEP_NOW_UNRELIABLE);
3018 3019 3020

	/* driverloader was caught wrongly pretending to be under GPL */
	if (strcmp(mod->name, "driverloader") == 0)
3021 3022
		add_taint_module(mod, TAINT_PROPRIETARY_MODULE,
				 LOCKDEP_NOW_UNRELIABLE);
3023

3024 3025
	/* lve claims to be GPL but upstream won't provide source */
	if (strcmp(mod->name, "lve") == 0)
3026 3027
		add_taint_module(mod, TAINT_PROPRIETARY_MODULE,
				 LOCKDEP_NOW_UNRELIABLE);
3028

3029 3030 3031 3032 3033 3034 3035 3036 3037 3038 3039 3040 3041 3042 3043 3044 3045 3046 3047 3048 3049 3050 3051 3052 3053 3054 3055 3056 3057 3058 3059 3060 3061 3062 3063 3064 3065 3066 3067
#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);
}

3068 3069 3070 3071 3072 3073 3074 3075
int __weak module_frob_arch_sections(Elf_Ehdr *hdr,
				     Elf_Shdr *sechdrs,
				     char *secstrings,
				     struct module *mod)
{
	return 0;
}

3076
static struct module *layout_and_allocate(struct load_info *info, int flags)
L
Linus Torvalds 已提交
3077
{
3078
	/* Module within temporary copy. */
L
Linus Torvalds 已提交
3079
	struct module *mod;
3080
	int err;
3081

3082
	mod = setup_load_info(info, flags);
3083 3084
	if (IS_ERR(mod))
		return mod;
L
Linus Torvalds 已提交
3085

3086
	err = check_modinfo(mod, info, flags);
3087 3088
	if (err)
		return ERR_PTR(err);
L
Linus Torvalds 已提交
3089 3090

	/* Allow arches to frob section contents and sizes.  */
3091 3092
	err = module_frob_arch_sections(info->hdr, info->sechdrs,
					info->secstrings, mod);
L
Linus Torvalds 已提交
3093
	if (err < 0)
3094
		return ERR_PTR(err);
L
Linus Torvalds 已提交
3095

3096 3097
	/* We will do a special allocation for per-cpu sections later. */
	info->sechdrs[info->index.pcpu].sh_flags &= ~(unsigned long)SHF_ALLOC;
L
Linus Torvalds 已提交
3098 3099 3100 3101

	/* 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. */
3102 3103
	layout_sections(mod, info);
	layout_symtab(mod, info);
L
Linus Torvalds 已提交
3104

3105
	/* Allocate and move to the final place */
3106
	err = move_module(mod, info);
3107
	if (err)
3108
		return ERR_PTR(err);
3109 3110 3111

	/* Module has been copied to its final place now: return it. */
	mod = (void *)info->sechdrs[info->index.mod].sh_addr;
3112
	kmemleak_load_module(mod, info);
3113 3114 3115 3116 3117 3118 3119
	return mod;
}

/* mod is no longer valid after this! */
static void module_deallocate(struct module *mod, struct load_info *info)
{
	percpu_modfree(mod);
3120
	module_arch_freeing_init(mod);
3121 3122
	module_memfree(mod->module_init);
	module_memfree(mod->module_core);
3123 3124
}

3125 3126 3127 3128 3129 3130 3131
int __weak module_finalize(const Elf_Ehdr *hdr,
			   const Elf_Shdr *sechdrs,
			   struct module *me)
{
	return 0;
}

3132 3133
static int post_relocation(struct module *mod, const struct load_info *info)
{
3134
	/* Sort exception table now relocations are done. */
3135 3136 3137 3138 3139 3140
	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);

3141
	/* Setup kallsyms-specific fields. */
3142 3143 3144 3145 3146 3147
	add_kallsyms(mod, info);

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

3148 3149 3150 3151 3152 3153
/* Is this module of this name done loading?  No locks held. */
static bool finished_loading(const char *name)
{
	struct module *mod;
	bool ret;

3154 3155 3156 3157 3158 3159
	/*
	 * The module_mutex should not be a heavily contended lock;
	 * if we get the occasional sleep here, we'll go an extra iteration
	 * in the wait_event_interruptible(), which is harmless.
	 */
	sched_annotate_sleep();
3160
	mutex_lock(&module_mutex);
3161
	mod = find_module_all(name, strlen(name), true);
3162 3163
	ret = !mod || mod->state == MODULE_STATE_LIVE
		|| mod->state == MODULE_STATE_GOING;
3164 3165 3166 3167 3168
	mutex_unlock(&module_mutex);

	return ret;
}

3169 3170 3171 3172 3173 3174 3175 3176 3177 3178 3179
/* 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
}

3180 3181 3182 3183 3184 3185 3186 3187 3188 3189 3190 3191 3192
/* For freeing module_init on success, in case kallsyms traversing */
struct mod_initfree {
	struct rcu_head rcu;
	void *module_init;
};

static void do_free_init(struct rcu_head *head)
{
	struct mod_initfree *m = container_of(head, struct mod_initfree, rcu);
	module_memfree(m->module_init);
	kfree(m);
}

3193 3194 3195 3196 3197 3198 3199
/*
 * This is where the real work happens.
 *
 * Keep it uninlined to provide a reliable breakpoint target, e.g. for the gdb
 * helper command 'lx-symbols'.
 */
static noinline int do_init_module(struct module *mod)
3200 3201
{
	int ret = 0;
3202 3203 3204 3205 3206 3207 3208 3209
	struct mod_initfree *freeinit;

	freeinit = kmalloc(sizeof(*freeinit), GFP_KERNEL);
	if (!freeinit) {
		ret = -ENOMEM;
		goto fail;
	}
	freeinit->module_init = mod->module_init;
3210

3211 3212 3213 3214 3215 3216
	/*
	 * We want to find out whether @mod uses async during init.  Clear
	 * PF_USED_ASYNC.  async_schedule*() will set it.
	 */
	current->flags &= ~PF_USED_ASYNC;

3217 3218 3219 3220 3221
	do_mod_ctors(mod);
	/* Start the module */
	if (mod->init != NULL)
		ret = do_one_initcall(mod->init);
	if (ret < 0) {
3222
		goto fail_free_freeinit;
3223 3224
	}
	if (ret > 0) {
3225 3226 3227 3228
		pr_warn("%s: '%s'->init suspiciously returned %d, it should "
			"follow 0/-E convention\n"
			"%s: loading module anyway...\n",
			__func__, mod->name, ret, __func__);
3229 3230 3231 3232 3233 3234 3235 3236
		dump_stack();
	}

	/* Now it's a first class citizen! */
	mod->state = MODULE_STATE_LIVE;
	blocking_notifier_call_chain(&module_notify_list,
				     MODULE_STATE_LIVE, mod);

3237 3238 3239 3240 3241 3242 3243 3244 3245 3246 3247 3248 3249 3250 3251 3252 3253 3254 3255
	/*
	 * We need to finish all async code before the module init sequence
	 * is done.  This has potential to deadlock.  For example, a newly
	 * detected block device can trigger request_module() of the
	 * default iosched from async probing task.  Once userland helper
	 * reaches here, async_synchronize_full() will wait on the async
	 * task waiting on request_module() and deadlock.
	 *
	 * This deadlock is avoided by perfomring async_synchronize_full()
	 * iff module init queued any async jobs.  This isn't a full
	 * solution as it will deadlock the same if module loading from
	 * async jobs nests more than once; however, due to the various
	 * constraints, this hack seems to be the best option for now.
	 * Please refer to the following thread for details.
	 *
	 * http://thread.gmane.org/gmane.linux.kernel/1420814
	 */
	if (current->flags & PF_USED_ASYNC)
		async_synchronize_full();
3256 3257 3258 3259 3260 3261 3262 3263 3264 3265

	mutex_lock(&module_mutex);
	/* Drop initial reference. */
	module_put(mod);
	trim_init_extable(mod);
#ifdef CONFIG_KALLSYMS
	mod->num_symtab = mod->core_num_syms;
	mod->symtab = mod->core_symtab;
	mod->strtab = mod->core_strtab;
#endif
3266
	mod_tree_remove_init(mod);
3267
	unset_module_init_ro_nx(mod);
3268
	module_arch_freeing_init(mod);
3269 3270 3271 3272
	mod->module_init = NULL;
	mod->init_size = 0;
	mod->init_ro_size = 0;
	mod->init_text_size = 0;
3273 3274
	/*
	 * We want to free module_init, but be aware that kallsyms may be
3275 3276 3277
	 * walking this with preempt disabled.  In all the failure paths, we
	 * call synchronize_sched(), but we don't want to slow down the success
	 * path, so use actual RCU here.
3278
	 */
3279
	call_rcu_sched(&freeinit->rcu, do_free_init);
3280 3281 3282 3283
	mutex_unlock(&module_mutex);
	wake_up_all(&module_wq);

	return 0;
3284 3285 3286 3287 3288 3289 3290 3291 3292 3293 3294 3295 3296

fail_free_freeinit:
	kfree(freeinit);
fail:
	/* Try to protect us from buggy refcounters. */
	mod->state = MODULE_STATE_GOING;
	synchronize_sched();
	module_put(mod);
	blocking_notifier_call_chain(&module_notify_list,
				     MODULE_STATE_GOING, mod);
	free_module(mod);
	wake_up_all(&module_wq);
	return ret;
3297 3298 3299 3300 3301 3302 3303 3304 3305 3306
}

static int may_init_module(void)
{
	if (!capable(CAP_SYS_MODULE) || modules_disabled)
		return -EPERM;

	return 0;
}

3307 3308 3309 3310 3311 3312 3313 3314 3315 3316 3317 3318 3319 3320
/*
 * We try to place it in the list now to make sure it's unique before
 * we dedicate too many resources.  In particular, temporary percpu
 * memory exhaustion.
 */
static int add_unformed_module(struct module *mod)
{
	int err;
	struct module *old;

	mod->state = MODULE_STATE_UNFORMED;

again:
	mutex_lock(&module_mutex);
3321 3322
	old = find_module_all(mod->name, strlen(mod->name), true);
	if (old != NULL) {
3323 3324 3325 3326
		if (old->state == MODULE_STATE_COMING
		    || old->state == MODULE_STATE_UNFORMED) {
			/* Wait in case it fails to load. */
			mutex_unlock(&module_mutex);
3327 3328
			err = wait_event_interruptible(module_wq,
					       finished_loading(mod->name));
3329 3330 3331 3332 3333 3334 3335
			if (err)
				goto out_unlocked;
			goto again;
		}
		err = -EEXIST;
		goto out;
	}
3336
	mod_update_bounds(mod);
3337
	list_add_rcu(&mod->list, &modules);
3338
	mod_tree_insert(mod);
3339 3340 3341 3342 3343 3344 3345 3346 3347 3348 3349 3350 3351 3352 3353 3354 3355 3356 3357 3358 3359 3360
	err = 0;

out:
	mutex_unlock(&module_mutex);
out_unlocked:
	return err;
}

static int complete_formation(struct module *mod, struct load_info *info)
{
	int err;

	mutex_lock(&module_mutex);

	/* Find duplicate symbols (must be called under lock). */
	err = verify_export_symbols(mod);
	if (err < 0)
		goto out;

	/* This relies on module_mutex for list integrity. */
	module_bug_finalize(info->hdr, info->sechdrs, mod);

3361 3362 3363 3364 3365 3366 3367 3368 3369 3370 3371 3372
	/* 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);

3373 3374 3375
	/* Mark state as coming so strong_try_module_get() ignores us,
	 * but kallsyms etc. can see us. */
	mod->state = MODULE_STATE_COMING;
3376 3377 3378 3379 3380
	mutex_unlock(&module_mutex);

	blocking_notifier_call_chain(&module_notify_list,
				     MODULE_STATE_COMING, mod);
	return 0;
3381 3382 3383 3384 3385 3386

out:
	mutex_unlock(&module_mutex);
	return err;
}

3387 3388
static int unknown_module_param_cb(char *param, char *val, const char *modname)
{
3389
	/* Check for magic 'dyndbg' arg */
3390
	int ret = ddebug_dyndbg_module_param_cb(param, val, modname);
3391 3392
	if (ret != 0)
		pr_warn("%s: unknown parameter '%s' ignored\n", modname, param);
3393 3394 3395
	return 0;
}

3396 3397
/* Allocate and load the module: note that size of section 0 is always
   zero, and we rely on this for optional sections. */
3398 3399
static int load_module(struct load_info *info, const char __user *uargs,
		       int flags)
3400
{
3401
	struct module *mod;
3402
	long err;
3403
	char *after_dashes;
3404

3405 3406 3407
	err = module_sig_check(info);
	if (err)
		goto free_copy;
3408

3409
	err = elf_header_check(info);
3410
	if (err)
3411
		goto free_copy;
3412 3413

	/* Figure out module layout, and allocate all the memory. */
3414
	mod = layout_and_allocate(info, flags);
3415 3416
	if (IS_ERR(mod)) {
		err = PTR_ERR(mod);
3417
		goto free_copy;
L
Linus Torvalds 已提交
3418 3419
	}

3420 3421 3422
	/* Reserve our place in the list. */
	err = add_unformed_module(mod);
	if (err)
3423 3424
		goto free_module;

3425
#ifdef CONFIG_MODULE_SIG
3426
	mod->sig_ok = info->sig_ok;
3427
	if (!mod->sig_ok) {
3428
		pr_notice_once("%s: module verification failed: signature "
3429
			       "and/or required key missing - tainting "
3430
			       "kernel\n", mod->name);
3431
		add_taint_module(mod, TAINT_UNSIGNED_MODULE, LOCKDEP_STILL_OK);
3432
	}
3433 3434
#endif

3435
	/* To avoid stressing percpu allocator, do this once we're unique. */
3436
	err = percpu_modalloc(mod, info);
3437 3438 3439
	if (err)
		goto unlink_mod;

3440
	/* Now module is in final location, initialize linked lists, etc. */
3441 3442
	err = module_unload_init(mod);
	if (err)
3443
		goto unlink_mod;
L
Linus Torvalds 已提交
3444

3445 3446
	/* Now we've got everything in the final locations, we can
	 * find optional sections. */
3447 3448 3449
	err = find_module_sections(mod, info);
	if (err)
		goto free_unload;
3450

3451
	err = check_module_license_and_versions(mod);
3452 3453
	if (err)
		goto free_unload;
3454

3455
	/* Set up MODINFO_ATTR fields */
3456
	setup_modinfo(mod, info);
3457

L
Linus Torvalds 已提交
3458
	/* Fix up syms, so that st_value is a pointer to location. */
3459
	err = simplify_symbols(mod, info);
L
Linus Torvalds 已提交
3460
	if (err < 0)
3461
		goto free_modinfo;
L
Linus Torvalds 已提交
3462

3463
	err = apply_relocations(mod, info);
3464
	if (err < 0)
3465
		goto free_modinfo;
L
Linus Torvalds 已提交
3466

3467
	err = post_relocation(mod, info);
L
Linus Torvalds 已提交
3468
	if (err < 0)
3469
		goto free_modinfo;
L
Linus Torvalds 已提交
3470

3471
	flush_module_icache(mod);
3472

3473 3474 3475 3476 3477 3478
	/* 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;
	}
3479

3480
	dynamic_debug_setup(info->debug, info->num_debug);
3481

3482 3483 3484
	/* Ftrace init must be called in the MODULE_STATE_UNFORMED state */
	ftrace_module_init(mod);

3485 3486 3487
	/* Finally it's fully formed, ready to start executing. */
	err = complete_formation(mod, info);
	if (err)
3488
		goto ddebug_cleanup;
3489

3490
	/* Module is ready to execute: parsing args may do that. */
3491 3492 3493 3494
	after_dashes = parse_args(mod->name, mod->args, mod->kp, mod->num_kp,
				  -32768, 32767, unknown_module_param_cb);
	if (IS_ERR(after_dashes)) {
		err = PTR_ERR(after_dashes);
3495
		goto bug_cleanup;
3496 3497 3498 3499
	} else if (after_dashes) {
		pr_warn("%s: parameters '%s' after `--' ignored\n",
		       mod->name, after_dashes);
	}
L
Linus Torvalds 已提交
3500

3501
	/* Link in to syfs. */
3502
	err = mod_sysfs_setup(mod, info, mod->kp, mod->num_kp);
L
Linus Torvalds 已提交
3503
	if (err < 0)
3504
		goto bug_cleanup;
3505

3506
	/* Get rid of temporary copy. */
3507
	free_copy(info);
L
Linus Torvalds 已提交
3508 3509

	/* Done! */
3510
	trace_module_load(mod);
3511 3512

	return do_init_module(mod);
L
Linus Torvalds 已提交
3513

3514 3515
 bug_cleanup:
	/* module_bug_cleanup needs module_mutex protection */
3516
	mutex_lock(&module_mutex);
3517
	module_bug_cleanup(mod);
3518
	mutex_unlock(&module_mutex);
3519 3520 3521 3522 3523

	/* we can't deallocate the module until we clear memory protection */
	unset_module_init_ro_nx(mod);
	unset_module_core_ro_nx(mod);

3524
 ddebug_cleanup:
3525
	dynamic_debug_remove(info->debug);
3526
	synchronize_sched();
3527 3528
	kfree(mod->args);
 free_arch_cleanup:
L
Linus Torvalds 已提交
3529
	module_arch_cleanup(mod);
3530
 free_modinfo:
3531
	free_modinfo(mod);
3532
 free_unload:
L
Linus Torvalds 已提交
3533
	module_unload_free(mod);
3534 3535 3536 3537 3538
 unlink_mod:
	mutex_lock(&module_mutex);
	/* Unlink carefully: kallsyms could be walking list. */
	list_del_rcu(&mod->list);
	wake_up_all(&module_wq);
3539 3540
	/* Wait for RCU-sched synchronizing before releasing mod->list. */
	synchronize_sched();
3541
	mutex_unlock(&module_mutex);
3542
 free_module:
3543 3544 3545
	/* Free lock-classes; relies on the preceding sync_rcu() */
	lockdep_free_key_range(mod->module_core, mod->core_size);

3546
	module_deallocate(mod, info);
3547
 free_copy:
3548 3549
	free_copy(info);
	return err;
3550 3551
}

3552 3553
SYSCALL_DEFINE3(init_module, void __user *, umod,
		unsigned long, len, const char __user *, uargs)
L
Linus Torvalds 已提交
3554
{
3555 3556
	int err;
	struct load_info info = { };
L
Linus Torvalds 已提交
3557

3558 3559 3560
	err = may_init_module();
	if (err)
		return err;
L
Linus Torvalds 已提交
3561

3562 3563
	pr_debug("init_module: umod=%p, len=%lu, uargs=%p\n",
	       umod, len, uargs);
L
Linus Torvalds 已提交
3564

3565 3566 3567
	err = copy_module_from_user(umod, len, &info);
	if (err)
		return err;
L
Linus Torvalds 已提交
3568

3569
	return load_module(&info, uargs, 0);
3570
}
3571

3572
SYSCALL_DEFINE3(finit_module, int, fd, const char __user *, uargs, int, flags)
3573 3574 3575
{
	int err;
	struct load_info info = { };
3576

3577 3578 3579
	err = may_init_module();
	if (err)
		return err;
L
Linus Torvalds 已提交
3580

3581
	pr_debug("finit_module: fd=%d, uargs=%p, flags=%i\n", fd, uargs, flags);
3582

3583 3584 3585
	if (flags & ~(MODULE_INIT_IGNORE_MODVERSIONS
		      |MODULE_INIT_IGNORE_VERMAGIC))
		return -EINVAL;
3586

3587 3588 3589
	err = copy_module_from_fd(fd, &info);
	if (err)
		return err;
L
Linus Torvalds 已提交
3590

3591
	return load_module(&info, uargs, flags);
L
Linus Torvalds 已提交
3592 3593 3594 3595 3596 3597 3598 3599 3600 3601 3602 3603 3604 3605
}

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 $d.
 */
static inline int is_arm_mapping_symbol(const char *str)
{
3606 3607
	if (str[0] == '.' && str[1] == 'L')
		return true;
3608
	return str[0] == '$' && strchr("axtd", str[1])
L
Linus Torvalds 已提交
3609 3610 3611 3612 3613 3614 3615 3616 3617 3618 3619 3620
	       && (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 */
3621
	if (within_module_init(addr, mod))
L
Linus Torvalds 已提交
3622
		nextval = (unsigned long)mod->module_init+mod->init_text_size;
3623
	else
L
Linus Torvalds 已提交
3624 3625
		nextval = (unsigned long)mod->module_core+mod->core_text_size;

L
Lucas De Marchi 已提交
3626
	/* Scan for closest preceding symbol, and next symbol. (ELF
3627
	   starts real symbols at 1). */
L
Linus Torvalds 已提交
3628 3629 3630 3631 3632 3633 3634 3635 3636 3637 3638 3639 3640 3641 3642 3643 3644 3645 3646 3647 3648
	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;

3649 3650 3651 3652
	if (size)
		*size = nextval - mod->symtab[best].st_value;
	if (offset)
		*offset = addr - mod->symtab[best].st_value;
L
Linus Torvalds 已提交
3653 3654 3655
	return mod->strtab + mod->symtab[best].st_name;
}

3656 3657
/* For kallsyms to ask for address resolution.  NULL means not found.  Careful
 * not to lock to avoid deadlock on oopses, simply disable preemption. */
3658
const char *module_address_lookup(unsigned long addr,
3659 3660 3661 3662
			    unsigned long *size,
			    unsigned long *offset,
			    char **modname,
			    char *namebuf)
L
Linus Torvalds 已提交
3663
{
3664
	const char *ret = NULL;
3665
	struct module *mod;
L
Linus Torvalds 已提交
3666

3667
	preempt_disable();
3668 3669 3670 3671 3672
	mod = __module_address(addr);
	if (mod) {
		if (modname)
			*modname = mod->name;
		ret = get_ksymbol(mod, addr, size, offset);
L
Linus Torvalds 已提交
3673
	}
3674 3675 3676 3677 3678
	/* Make a copy in here where it's safe */
	if (ret) {
		strncpy(namebuf, ret, KSYM_NAME_LEN - 1);
		ret = namebuf;
	}
3679
	preempt_enable();
3680

3681
	return ret;
L
Linus Torvalds 已提交
3682 3683
}

3684 3685 3686 3687
int lookup_module_symbol_name(unsigned long addr, char *symname)
{
	struct module *mod;

3688
	preempt_disable();
3689
	list_for_each_entry_rcu(mod, &modules, list) {
3690 3691
		if (mod->state == MODULE_STATE_UNFORMED)
			continue;
3692
		if (within_module(addr, mod)) {
3693 3694 3695 3696 3697
			const char *sym;

			sym = get_ksymbol(mod, addr, NULL, NULL);
			if (!sym)
				goto out;
3698
			strlcpy(symname, sym, KSYM_NAME_LEN);
3699
			preempt_enable();
3700 3701 3702 3703
			return 0;
		}
	}
out:
3704
	preempt_enable();
3705 3706 3707
	return -ERANGE;
}

3708 3709 3710 3711 3712
int lookup_module_symbol_attrs(unsigned long addr, unsigned long *size,
			unsigned long *offset, char *modname, char *name)
{
	struct module *mod;

3713
	preempt_disable();
3714
	list_for_each_entry_rcu(mod, &modules, list) {
3715 3716
		if (mod->state == MODULE_STATE_UNFORMED)
			continue;
3717
		if (within_module(addr, mod)) {
3718 3719 3720 3721 3722 3723
			const char *sym;

			sym = get_ksymbol(mod, addr, size, offset);
			if (!sym)
				goto out;
			if (modname)
3724
				strlcpy(modname, mod->name, MODULE_NAME_LEN);
3725
			if (name)
3726
				strlcpy(name, sym, KSYM_NAME_LEN);
3727
			preempt_enable();
3728 3729 3730 3731
			return 0;
		}
	}
out:
3732
	preempt_enable();
3733 3734 3735
	return -ERANGE;
}

3736 3737
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
3738 3739 3740
{
	struct module *mod;

3741
	preempt_disable();
3742
	list_for_each_entry_rcu(mod, &modules, list) {
3743 3744
		if (mod->state == MODULE_STATE_UNFORMED)
			continue;
L
Linus Torvalds 已提交
3745 3746 3747
		if (symnum < mod->num_symtab) {
			*value = mod->symtab[symnum].st_value;
			*type = mod->symtab[symnum].st_info;
3748
			strlcpy(name, mod->strtab + mod->symtab[symnum].st_name,
3749 3750
				KSYM_NAME_LEN);
			strlcpy(module_name, mod->name, MODULE_NAME_LEN);
3751
			*exported = is_exported(name, *value, mod);
3752
			preempt_enable();
3753
			return 0;
L
Linus Torvalds 已提交
3754 3755 3756
		}
		symnum -= mod->num_symtab;
	}
3757
	preempt_enable();
3758
	return -ERANGE;
L
Linus Torvalds 已提交
3759 3760 3761 3762 3763 3764 3765
}

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

	for (i = 0; i < mod->num_symtab; i++)
3766 3767
		if (strcmp(name, mod->strtab+mod->symtab[i].st_name) == 0 &&
		    mod->symtab[i].st_info != 'U')
L
Linus Torvalds 已提交
3768 3769 3770 3771 3772 3773 3774 3775 3776 3777 3778 3779
			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. */
3780
	preempt_disable();
L
Linus Torvalds 已提交
3781
	if ((colon = strchr(name, ':')) != NULL) {
3782
		if ((mod = find_module_all(name, colon - name, false)) != NULL)
L
Linus Torvalds 已提交
3783 3784
			ret = mod_find_symname(mod, colon+1);
	} else {
3785 3786 3787
		list_for_each_entry_rcu(mod, &modules, list) {
			if (mod->state == MODULE_STATE_UNFORMED)
				continue;
L
Linus Torvalds 已提交
3788 3789
			if ((ret = mod_find_symname(mod, name)) != 0)
				break;
3790
		}
L
Linus Torvalds 已提交
3791
	}
3792
	preempt_enable();
L
Linus Torvalds 已提交
3793 3794
	return ret;
}
3795 3796 3797 3798 3799 3800 3801 3802 3803

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;

3804 3805
	module_assert_mutex();

3806
	list_for_each_entry(mod, &modules, list) {
3807 3808
		if (mod->state == MODULE_STATE_UNFORMED)
			continue;
3809 3810 3811 3812 3813 3814 3815 3816 3817
		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
3818 3819
#endif /* CONFIG_KALLSYMS */

3820
static char *module_flags(struct module *mod, char *buf)
3821 3822 3823
{
	int bx = 0;

3824
	BUG_ON(mod->state == MODULE_STATE_UNFORMED);
3825 3826 3827
	if (mod->taints ||
	    mod->state == MODULE_STATE_GOING ||
	    mod->state == MODULE_STATE_COMING) {
3828
		buf[bx++] = '(';
3829
		bx += module_flags_taint(mod, buf + bx);
3830 3831 3832 3833 3834 3835
		/* 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++] = '+';
3836 3837 3838 3839 3840 3841 3842
		buf[bx++] = ')';
	}
	buf[bx] = '\0';

	return buf;
}

3843 3844 3845 3846 3847 3848 3849 3850 3851 3852 3853 3854 3855 3856 3857 3858 3859 3860
#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
3861 3862 3863
static int m_show(struct seq_file *m, void *p)
{
	struct module *mod = list_entry(p, struct module, list);
3864 3865
	char buf[8];

3866 3867 3868 3869
	/* We always ignore unformed modules. */
	if (mod->state == MODULE_STATE_UNFORMED)
		return 0;

3870
	seq_printf(m, "%s %u",
L
Linus Torvalds 已提交
3871 3872 3873 3874 3875
		   mod->name, mod->init_size + mod->core_size);
	print_unload_info(m, mod);

	/* Informative for users. */
	seq_printf(m, " %s",
3876 3877
		   mod->state == MODULE_STATE_GOING ? "Unloading" :
		   mod->state == MODULE_STATE_COMING ? "Loading" :
L
Linus Torvalds 已提交
3878 3879
		   "Live");
	/* Used by oprofile and other similar tools. */
3880
	seq_printf(m, " 0x%pK", mod->module_core);
L
Linus Torvalds 已提交
3881

3882 3883
	/* Taints info */
	if (mod->taints)
3884
		seq_printf(m, " %s", module_flags(mod, buf));
3885

3886
	seq_puts(m, "\n");
L
Linus Torvalds 已提交
3887 3888 3889 3890 3891 3892 3893 3894
	return 0;
}

/* Format: modulename size refcount deps address

   Where refcount is a number or -, and deps is a comma-separated list
   of depends or -.
*/
3895
static const struct seq_operations modules_op = {
L
Linus Torvalds 已提交
3896 3897 3898 3899 3900 3901
	.start	= m_start,
	.next	= m_next,
	.stop	= m_stop,
	.show	= m_show
};

3902 3903 3904 3905 3906 3907 3908 3909 3910 3911 3912 3913 3914 3915 3916 3917 3918 3919 3920 3921
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
3922 3923 3924 3925 3926 3927
/* 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;

3928
	preempt_disable();
3929
	list_for_each_entry_rcu(mod, &modules, list) {
3930 3931
		if (mod->state == MODULE_STATE_UNFORMED)
			continue;
L
Linus Torvalds 已提交
3932 3933
		if (mod->num_exentries == 0)
			continue;
3934

L
Linus Torvalds 已提交
3935 3936 3937 3938 3939 3940
		e = search_extable(mod->extable,
				   mod->extable + mod->num_exentries - 1,
				   addr);
		if (e)
			break;
	}
3941
	preempt_enable();
L
Linus Torvalds 已提交
3942 3943

	/* Now, if we found one, we are running inside it now, hence
3944
	   we cannot unload the module, hence no refcnt needed. */
L
Linus Torvalds 已提交
3945 3946 3947
	return e;
}

3948
/*
R
Rusty Russell 已提交
3949 3950 3951 3952 3953
 * 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).
3954
 */
R
Rusty Russell 已提交
3955
bool is_module_address(unsigned long addr)
3956
{
R
Rusty Russell 已提交
3957
	bool ret;
3958

3959
	preempt_disable();
R
Rusty Russell 已提交
3960
	ret = __module_address(addr) != NULL;
3961
	preempt_enable();
3962

R
Rusty Russell 已提交
3963
	return ret;
3964 3965
}

R
Rusty Russell 已提交
3966 3967 3968 3969 3970 3971 3972
/*
 * __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.
 */
3973
struct module *__module_address(unsigned long addr)
L
Linus Torvalds 已提交
3974 3975 3976
{
	struct module *mod;

3977 3978 3979
	if (addr < module_addr_min || addr > module_addr_max)
		return NULL;

3980 3981
	module_assert_mutex_or_preempt();

3982
	mod = mod_find(addr);
3983 3984
	if (mod) {
		BUG_ON(!within_module(addr, mod));
3985
		if (mod->state == MODULE_STATE_UNFORMED)
3986
			mod = NULL;
3987
	}
3988
	return mod;
L
Linus Torvalds 已提交
3989
}
3990
EXPORT_SYMBOL_GPL(__module_address);
L
Linus Torvalds 已提交
3991

R
Rusty Russell 已提交
3992 3993 3994 3995 3996 3997 3998 3999 4000 4001 4002 4003 4004 4005 4006 4007 4008 4009 4010 4011 4012 4013 4014 4015 4016 4017 4018 4019 4020 4021 4022 4023 4024 4025 4026 4027 4028
/*
 * 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;
}
4029
EXPORT_SYMBOL_GPL(__module_text_address);
R
Rusty Russell 已提交
4030

L
Linus Torvalds 已提交
4031 4032 4033 4034
/* Don't grab lock, we're oopsing. */
void print_modules(void)
{
	struct module *mod;
4035
	char buf[8];
L
Linus Torvalds 已提交
4036

4037
	printk(KERN_DEFAULT "Modules linked in:");
4038 4039
	/* Most callers should already have preempt disabled, but make sure */
	preempt_disable();
4040 4041 4042
	list_for_each_entry_rcu(mod, &modules, list) {
		if (mod->state == MODULE_STATE_UNFORMED)
			continue;
J
Jiri Slaby 已提交
4043
		pr_cont(" %s%s", mod->name, module_flags(mod, buf));
4044
	}
4045
	preempt_enable();
4046
	if (last_unloaded_module[0])
J
Jiri Slaby 已提交
4047 4048
		pr_cont(" [last unloaded: %s]", last_unloaded_module);
	pr_cont("\n");
L
Linus Torvalds 已提交
4049 4050 4051
}

#ifdef CONFIG_MODVERSIONS
4052 4053 4054 4055 4056 4057
/* 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,
4058
		   struct tracepoint * const *tp)
4059 4060 4061
{
}
EXPORT_SYMBOL(module_layout);
L
Linus Torvalds 已提交
4062
#endif
反馈
建议
客服 返回
顶部