module.c 109.5 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>
20
#include <linux/extable.h>
L
Linus Torvalds 已提交
21
#include <linux/moduleloader.h>
22
#include <linux/trace_events.h>
L
Linus Torvalds 已提交
23
#include <linux/init.h>
24
#include <linux/kallsyms.h>
25
#include <linux/file.h>
26
#include <linux/fs.h>
27
#include <linux/sysfs.h>
28
#include <linux/kernel.h>
L
Linus Torvalds 已提交
29 30 31
#include <linux/slab.h>
#include <linux/vmalloc.h>
#include <linux/elf.h>
32
#include <linux/proc_fs.h>
33
#include <linux/security.h>
L
Linus Torvalds 已提交
34 35 36 37
#include <linux/seq_file.h>
#include <linux/syscalls.h>
#include <linux/fcntl.h>
#include <linux/rcupdate.h>
38
#include <linux/capability.h>
L
Linus Torvalds 已提交
39 40 41 42 43 44
#include <linux/cpu.h>
#include <linux/moduleparam.h>
#include <linux/errno.h>
#include <linux/err.h>
#include <linux/vermagic.h>
#include <linux/notifier.h>
45
#include <linux/sched.h>
L
Linus Torvalds 已提交
46
#include <linux/device.h>
47
#include <linux/string.h>
48
#include <linux/mutex.h>
49
#include <linux/rculist.h>
50
#include <linux/uaccess.h>
L
Linus Torvalds 已提交
51
#include <asm/cacheflush.h>
52
#include <asm/mmu_context.h>
53
#include <linux/license.h>
54
#include <asm/sections.h>
55
#include <linux/tracepoint.h>
56
#include <linux/ftrace.h>
57
#include <linux/livepatch.h>
58
#include <linux/async.h>
59
#include <linux/percpu.h>
60
#include <linux/kmemleak.h>
61
#include <linux/jump_label.h>
62
#include <linux/pfn.h>
63
#include <linux/bsearch.h>
64
#include <linux/dynamic_debug.h>
65
#include <linux/audit.h>
66
#include <uapi/linux/module.h>
67
#include "module-internal.h"
L
Linus Torvalds 已提交
68

69 70 71
#define CREATE_TRACE_POINTS
#include <trace/events/module.h>

L
Linus Torvalds 已提交
72 73 74 75
#ifndef ARCH_SHF_SMALL
#define ARCH_SHF_SMALL 0
#endif

76 77 78 79 80 81 82 83 84 85 86
/*
 * 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

L
Linus Torvalds 已提交
87 88 89
/* If this is set, the section belongs in the init part of the module */
#define INIT_OFFSET_MASK (1UL << (BITS_PER_LONG-1))

90 91 92 93 94
/*
 * Mutex protects:
 * 1) List of modules (also safely readable with preempt_disable),
 * 2) module_use links,
 * 3) module_addr_min/module_addr_max.
95
 * (delete and add uses RCU list operations). */
96 97
DEFINE_MUTEX(module_mutex);
EXPORT_SYMBOL_GPL(module_mutex);
L
Linus Torvalds 已提交
98
static LIST_HEAD(modules);
99

100
#ifdef CONFIG_MODULES_TREE_LOOKUP
101

102 103 104 105
/*
 * Use a latched RB-tree for __module_address(); this allows us to use
 * RCU-sched lookups of the address from any context.
 *
106 107 108
 * 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.
109 110 111
 */

static __always_inline unsigned long __mod_tree_val(struct latch_tree_node *n)
112
{
113
	struct module_layout *layout = container_of(n, struct module_layout, mtn.node);
114

115
	return (unsigned long)layout->base;
116 117 118 119
}

static __always_inline unsigned long __mod_tree_size(struct latch_tree_node *n)
{
120
	struct module_layout *layout = container_of(n, struct module_layout, mtn.node);
121

122
	return (unsigned long)layout->size;
123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143
}

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;
144 145 146 147

	return 0;
}

148 149 150 151 152
static const struct latch_tree_ops mod_tree_ops = {
	.less = mod_tree_less,
	.comp = mod_tree_comp,
};

153 154 155 156 157 158
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,
159 160
};

161 162 163 164 165 166 167 168 169 170 171 172
#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);
}
173 174 175 176 177 178 179

/*
 * These modifications: insert, remove_init and remove; are serialized by the
 * module_mutex.
 */
static void mod_tree_insert(struct module *mod)
{
180 181
	mod->core_layout.mtn.mod = mod;
	mod->init_layout.mtn.mod = mod;
182

183 184 185
	__mod_tree_insert(&mod->core_layout.mtn);
	if (mod->init_layout.size)
		__mod_tree_insert(&mod->init_layout.mtn);
186 187 188 189
}

static void mod_tree_remove_init(struct module *mod)
{
190 191
	if (mod->init_layout.size)
		__mod_tree_remove(&mod->init_layout.mtn);
192 193 194 195
}

static void mod_tree_remove(struct module *mod)
{
196
	__mod_tree_remove(&mod->core_layout.mtn);
197 198 199
	mod_tree_remove_init(mod);
}

200
static struct module *mod_find(unsigned long addr)
201 202 203
{
	struct latch_tree_node *ltn;

204
	ltn = latch_tree_find((void *)addr, &mod_tree.root, &mod_tree_ops);
205 206 207 208 209 210
	if (!ltn)
		return NULL;

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

211 212
#else /* MODULES_TREE_LOOKUP */

213 214
static unsigned long module_addr_min = -1UL, module_addr_max = 0;

215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232
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 */

233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249
/*
 * 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)
{
250 251 252
	__mod_update_bounds(mod->core_layout.base, mod->core_layout.size);
	if (mod->init_layout.size)
		__mod_update_bounds(mod->init_layout.base, mod->init_layout.size);
253 254
}

255 256 257 258
#ifdef CONFIG_KGDB_KDB
struct list_head *kdb_modules = &modules; /* kdb needs the list of modules */
#endif /* CONFIG_KGDB_KDB */

259 260 261 262 263 264 265 266 267 268 269
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;

270
	WARN_ON_ONCE(!rcu_read_lock_sched_held() &&
271 272 273 274
		!lockdep_is_held(&module_mutex));
#endif
}

275 276
static bool sig_enforce = IS_ENABLED(CONFIG_MODULE_SIG_FORCE);
#ifndef CONFIG_MODULE_SIG_FORCE
277 278
module_param(sig_enforce, bool_enable_only, 0644);
#endif /* !CONFIG_MODULE_SIG_FORCE */
L
Linus Torvalds 已提交
279

280 281
/* Block module loading/unloading? */
int modules_disabled = 0;
282
core_param(nomodule, modules_disabled, bint, 0);
283

284 285 286
/* Waiting for a module to finish initializing? */
static DECLARE_WAIT_QUEUE_HEAD(module_wq);

287
static BLOCKING_NOTIFIER_HEAD(module_notify_list);
L
Linus Torvalds 已提交
288

289
int register_module_notifier(struct notifier_block *nb)
L
Linus Torvalds 已提交
290
{
291
	return blocking_notifier_chain_register(&module_notify_list, nb);
L
Linus Torvalds 已提交
292 293 294
}
EXPORT_SYMBOL(register_module_notifier);

295
int unregister_module_notifier(struct notifier_block *nb)
L
Linus Torvalds 已提交
296
{
297
	return blocking_notifier_chain_unregister(&module_notify_list, nb);
L
Linus Torvalds 已提交
298 299 300
}
EXPORT_SYMBOL(unregister_module_notifier);

301 302 303 304
struct load_info {
	Elf_Ehdr *hdr;
	unsigned long len;
	Elf_Shdr *sechdrs;
305
	char *secstrings, *strtab;
306
	unsigned long symoffs, stroffs;
307 308
	struct _ddebug *debug;
	unsigned int num_debug;
309
	bool sig_ok;
310 311 312
#ifdef CONFIG_KALLSYMS
	unsigned long mod_kallsyms_init_off;
#endif
313 314 315 316 317
	struct {
		unsigned int sym, str, mod, vers, info, pcpu;
	} index;
};

318 319 320 321 322
/*
 * We require a truly strong try_module_get(): 0 means success.
 * Otherwise an error is returned due to ongoing or failed
 * initialization etc.
 */
L
Linus Torvalds 已提交
323 324
static inline int strong_try_module_get(struct module *mod)
{
325
	BUG_ON(mod && mod->state == MODULE_STATE_UNFORMED);
L
Linus Torvalds 已提交
326
	if (mod && mod->state == MODULE_STATE_COMING)
327 328
		return -EBUSY;
	if (try_module_get(mod))
L
Linus Torvalds 已提交
329
		return 0;
330 331
	else
		return -ENOENT;
L
Linus Torvalds 已提交
332 333
}

334 335
static inline void add_taint_module(struct module *mod, unsigned flag,
				    enum lockdep_ok lockdep_ok)
336
{
337
	add_taint(flag, lockdep_ok);
338
	set_bit(flag, &mod->taints);
339 340
}

341 342 343
/*
 * 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
344
 */
345
void __noreturn __module_put_and_exit(struct module *mod, long code)
L
Linus Torvalds 已提交
346 347 348 349 350
{
	module_put(mod);
	do_exit(code);
}
EXPORT_SYMBOL(__module_put_and_exit);
351

L
Linus Torvalds 已提交
352
/* Find a module section: 0 means not found. */
353
static unsigned int find_sec(const struct load_info *info, const char *name)
L
Linus Torvalds 已提交
354 355 356
{
	unsigned int i;

357 358
	for (i = 1; i < info->hdr->e_shnum; i++) {
		Elf_Shdr *shdr = &info->sechdrs[i];
L
Linus Torvalds 已提交
359
		/* Alloc bit cleared means "ignore it." */
360 361
		if ((shdr->sh_flags & SHF_ALLOC)
		    && strcmp(info->secstrings + shdr->sh_name, name) == 0)
L
Linus Torvalds 已提交
362
			return i;
363
	}
L
Linus Torvalds 已提交
364 365 366
	return 0;
}

367
/* Find a module section, or NULL. */
368
static void *section_addr(const struct load_info *info, const char *name)
369 370
{
	/* Section 0 has sh_addr 0. */
371
	return (void *)info->sechdrs[find_sec(info, name)].sh_addr;
372 373 374
}

/* Find a module section, or NULL.  Fill in number of "objects" in section. */
375
static void *section_objs(const struct load_info *info,
376 377 378 379
			  const char *name,
			  size_t object_size,
			  unsigned int *num)
{
380
	unsigned int sec = find_sec(info, name);
381 382

	/* Section 0 has sh_addr 0 and sh_size 0. */
383 384
	*num = info->sechdrs[sec].sh_size / object_size;
	return (void *)info->sechdrs[sec].sh_addr;
385 386
}

L
Linus Torvalds 已提交
387 388 389 390 391
/* 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[];
392 393
extern const struct kernel_symbol __start___ksymtab_gpl_future[];
extern const struct kernel_symbol __stop___ksymtab_gpl_future[];
394 395 396
extern const s32 __start___kcrctab[];
extern const s32 __start___kcrctab_gpl[];
extern const s32 __start___kcrctab_gpl_future[];
397 398 399 400 401
#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[];
402 403
extern const s32 __start___kcrctab_unused[];
extern const s32 __start___kcrctab_unused_gpl[];
404
#endif
L
Linus Torvalds 已提交
405 406 407 408

#ifndef CONFIG_MODVERSIONS
#define symversion(base, idx) NULL
#else
409
#define symversion(base, idx) ((base != NULL) ? ((base) + (idx)) : NULL)
L
Linus Torvalds 已提交
410 411
#endif

412 413 414 415 416
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,
417
					      void *data),
418
				   void *data)
419
{
420
	unsigned int j;
421

422
	for (j = 0; j < arrsize; j++) {
423 424
		if (fn(&arr[j], owner, data))
			return true;
425
	}
426 427

	return false;
428 429
}

430
/* Returns true as soon as fn returns true, otherwise false. */
431 432 433 434
bool each_symbol_section(bool (*fn)(const struct symsearch *arr,
				    struct module *owner,
				    void *data),
			 void *data)
435 436
{
	struct module *mod;
437
	static const struct symsearch arr[] = {
438
		{ __start___ksymtab, __stop___ksymtab, __start___kcrctab,
439
		  NOT_GPL_ONLY, false },
440
		{ __start___ksymtab_gpl, __stop___ksymtab_gpl,
441 442
		  __start___kcrctab_gpl,
		  GPL_ONLY, false },
443
		{ __start___ksymtab_gpl_future, __stop___ksymtab_gpl_future,
444 445
		  __start___kcrctab_gpl_future,
		  WILL_BE_GPL_ONLY, false },
446
#ifdef CONFIG_UNUSED_SYMBOLS
447
		{ __start___ksymtab_unused, __stop___ksymtab_unused,
448 449
		  __start___kcrctab_unused,
		  NOT_GPL_ONLY, true },
450
		{ __start___ksymtab_unused_gpl, __stop___ksymtab_unused_gpl,
451 452
		  __start___kcrctab_unused_gpl,
		  GPL_ONLY, true },
453
#endif
454
	};
455

456 457
	module_assert_mutex_or_preempt();

458 459
	if (each_symbol_in_section(arr, ARRAY_SIZE(arr), NULL, fn, data))
		return true;
460

461
	list_for_each_entry_rcu(mod, &modules, list) {
462 463
		struct symsearch arr[] = {
			{ mod->syms, mod->syms + mod->num_syms, mod->crcs,
464
			  NOT_GPL_ONLY, false },
465
			{ mod->gpl_syms, mod->gpl_syms + mod->num_gpl_syms,
466 467
			  mod->gpl_crcs,
			  GPL_ONLY, false },
468 469
			{ mod->gpl_future_syms,
			  mod->gpl_future_syms + mod->num_gpl_future_syms,
470 471
			  mod->gpl_future_crcs,
			  WILL_BE_GPL_ONLY, false },
472
#ifdef CONFIG_UNUSED_SYMBOLS
473 474
			{ mod->unused_syms,
			  mod->unused_syms + mod->num_unused_syms,
475 476
			  mod->unused_crcs,
			  NOT_GPL_ONLY, true },
477 478
			{ mod->unused_gpl_syms,
			  mod->unused_gpl_syms + mod->num_unused_gpl_syms,
479 480
			  mod->unused_gpl_crcs,
			  GPL_ONLY, true },
481
#endif
482 483
		};

484 485 486
		if (mod->state == MODULE_STATE_UNFORMED)
			continue;

487 488 489 490 491
		if (each_symbol_in_section(arr, ARRAY_SIZE(arr), mod, fn, data))
			return true;
	}
	return false;
}
492
EXPORT_SYMBOL_GPL(each_symbol_section);
493 494 495 496 497 498 499 500 501

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

	/* Output */
	struct module *owner;
502
	const s32 *crc;
503
	const struct kernel_symbol *sym;
504 505
};

506 507 508
static bool check_symbol(const struct symsearch *syms,
				 struct module *owner,
				 unsigned int symnum, void *data)
509 510 511 512 513 514 515
{
	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) {
516 517 518
			pr_warn("Symbol %s is being used by a non-GPL module, "
				"which will not be allowed in the future\n",
				fsa->name);
519
		}
L
Linus Torvalds 已提交
520
	}
521

522
#ifdef CONFIG_UNUSED_SYMBOLS
523
	if (syms->unused && fsa->warn) {
524 525 526
		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");
527 528 529
		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 "
530
			"inclusion.\n");
531
	}
532
#endif
533 534 535

	fsa->owner = owner;
	fsa->crc = symversion(syms->crcs, symnum);
536
	fsa->sym = &syms->start[symnum];
537 538 539
	return true;
}

540 541 542 543 544 545 546 547
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);
}

548 549 550 551 552
static bool find_symbol_in_section(const struct symsearch *syms,
				   struct module *owner,
				   void *data)
{
	struct find_symbol_arg *fsa = data;
553 554 555 556 557 558 559
	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;
560 561 562 563

	return false;
}

564
/* Find a symbol and return it, along with, (optional) crc and
565
 * (optional) module which owns it.  Needs preempt disabled or module_mutex. */
566 567
const struct kernel_symbol *find_symbol(const char *name,
					struct module **owner,
568
					const s32 **crc,
569 570
					bool gplok,
					bool warn)
571 572 573 574 575 576 577
{
	struct find_symbol_arg fsa;

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

578
	if (each_symbol_section(find_symbol_in_section, &fsa)) {
579 580 581 582
		if (owner)
			*owner = fsa.owner;
		if (crc)
			*crc = fsa.crc;
583
		return fsa.sym;
584 585
	}

586
	pr_debug("Failed to find symbol %s\n", name);
587
	return NULL;
L
Linus Torvalds 已提交
588
}
589
EXPORT_SYMBOL_GPL(find_symbol);
L
Linus Torvalds 已提交
590

591 592 593 594
/*
 * Search for module by name: must hold module_mutex (or preempt disabled
 * for read-only access).
 */
595
static struct module *find_module_all(const char *name, size_t len,
596
				      bool even_unformed)
L
Linus Torvalds 已提交
597 598 599
{
	struct module *mod;

600
	module_assert_mutex_or_preempt();
601

L
Linus Torvalds 已提交
602
	list_for_each_entry(mod, &modules, list) {
603 604
		if (!even_unformed && mod->state == MODULE_STATE_UNFORMED)
			continue;
605
		if (strlen(mod->name) == len && !memcmp(mod->name, name, len))
L
Linus Torvalds 已提交
606 607 608 609
			return mod;
	}
	return NULL;
}
610 611 612

struct module *find_module(const char *name)
{
613
	module_assert_mutex();
614
	return find_module_all(name, strlen(name), false);
615
}
616
EXPORT_SYMBOL_GPL(find_module);
L
Linus Torvalds 已提交
617 618

#ifdef CONFIG_SMP
619

620
static inline void __percpu *mod_percpu(struct module *mod)
621
{
622 623
	return mod->percpu;
}
624

625
static int percpu_modalloc(struct module *mod, struct load_info *info)
626
{
627 628 629 630 631 632
	Elf_Shdr *pcpusec = &info->sechdrs[info->index.pcpu];
	unsigned long align = pcpusec->sh_addralign;

	if (!pcpusec->sh_size)
		return 0;

633
	if (align > PAGE_SIZE) {
634 635
		pr_warn("%s: per-cpu alignment %li > %li\n",
			mod->name, align, PAGE_SIZE);
636 637 638
		align = PAGE_SIZE;
	}

639
	mod->percpu = __alloc_reserved_percpu(pcpusec->sh_size, align);
640
	if (!mod->percpu) {
641 642
		pr_warn("%s: Could not allocate %lu bytes percpu data\n",
			mod->name, (unsigned long)pcpusec->sh_size);
643 644
		return -ENOMEM;
	}
645
	mod->percpu_size = pcpusec->sh_size;
646
	return 0;
647 648
}

649
static void percpu_modfree(struct module *mod)
650
{
651
	free_percpu(mod->percpu);
652 653
}

654
static unsigned int find_pcpusec(struct load_info *info)
655
{
656
	return find_sec(info, ".data..percpu");
657 658
}

659 660
static void percpu_modcopy(struct module *mod,
			   const void *from, unsigned long size)
661 662 663 664
{
	int cpu;

	for_each_possible_cpu(cpu)
665
		memcpy(per_cpu_ptr(mod->percpu, cpu), from, size);
666 667
}

668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684
/**
 * 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) {
685 686
		if (mod->state == MODULE_STATE_UNFORMED)
			continue;
687 688 689 690 691 692 693 694 695 696 697 698 699 700 701
		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;
702 703
}

L
Linus Torvalds 已提交
704
#else /* ... !CONFIG_SMP */
705

706
static inline void __percpu *mod_percpu(struct module *mod)
L
Linus Torvalds 已提交
707 708 709
{
	return NULL;
}
710
static int percpu_modalloc(struct module *mod, struct load_info *info)
711
{
712 713 714 715
	/* 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;
716 717
}
static inline void percpu_modfree(struct module *mod)
L
Linus Torvalds 已提交
718 719
{
}
720
static unsigned int find_pcpusec(struct load_info *info)
L
Linus Torvalds 已提交
721 722 723
{
	return 0;
}
724 725
static inline void percpu_modcopy(struct module *mod,
				  const void *from, unsigned long size)
L
Linus Torvalds 已提交
726 727 728 729
{
	/* pcpusec should be 0, and size of that section should be 0. */
	BUG_ON(size != 0);
}
730 731 732 733
bool is_module_percpu_address(unsigned long addr)
{
	return false;
}
734

L
Linus Torvalds 已提交
735 736
#endif /* CONFIG_SMP */

737 738 739 740 741 742
#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,   \
743
			struct module_kobject *mk, char *buffer)      \
744
{                                                                     \
745
	return scnprintf(buffer, PAGE_SIZE, "%s\n", mk->mod->field);  \
746 747 748 749 750 751 752
}                                                                     \
static int modinfo_##field##_exists(struct module *mod)               \
{                                                                     \
	return mod->field != NULL;                                    \
}                                                                     \
static void free_modinfo_##field(struct module *mod)                  \
{                                                                     \
753 754
	kfree(mod->field);                                            \
	mod->field = NULL;                                            \
755 756
}                                                                     \
static struct module_attribute modinfo_##field = {                    \
757
	.attr = { .name = __stringify(field), .mode = 0444 },         \
758 759 760 761 762 763 764 765 766
	.show = show_modinfo_##field,                                 \
	.setup = setup_modinfo_##field,                               \
	.test = modinfo_##field##_exists,                             \
	.free = free_modinfo_##field,                                 \
};

MODINFO_ATTR(version);
MODINFO_ATTR(srcversion);

767 768
static char last_unloaded_module[MODULE_NAME_LEN+1];

769
#ifdef CONFIG_MODULE_UNLOAD
770 771 772

EXPORT_TRACEPOINT_SYMBOL(module_get);

773 774 775
/* MODULE_REF_BASE is the base reference count by kmodule loader. */
#define MODULE_REF_BASE	1

L
Linus Torvalds 已提交
776
/* Init the unload section of the module. */
777
static int module_unload_init(struct module *mod)
L
Linus Torvalds 已提交
778
{
779 780 781 782 783
	/*
	 * Initialize reference counter to MODULE_REF_BASE.
	 * refcnt == 0 means module is going.
	 */
	atomic_set(&mod->refcnt, MODULE_REF_BASE);
784

785 786
	INIT_LIST_HEAD(&mod->source_list);
	INIT_LIST_HEAD(&mod->target_list);
787

L
Linus Torvalds 已提交
788
	/* Hold reference count during initialization. */
789
	atomic_inc(&mod->refcnt);
790 791

	return 0;
L
Linus Torvalds 已提交
792 793 794 795 796 797 798
}

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

799 800
	list_for_each_entry(use, &b->source_list, source_list) {
		if (use->source == a) {
801
			pr_debug("%s uses %s!\n", a->name, b->name);
L
Linus Torvalds 已提交
802 803 804
			return 1;
		}
	}
805
	pr_debug("%s does not use %s!\n", a->name, b->name);
L
Linus Torvalds 已提交
806 807 808
	return 0;
}

809 810 811 812 813 814 815 816 817 818 819
/*
 * 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;

820
	pr_debug("Allocating new usage for %s.\n", a->name);
821 822
	use = kmalloc(sizeof(*use), GFP_ATOMIC);
	if (!use) {
823
		pr_warn("%s: out of memory loading\n", a->name);
824 825 826 827 828 829 830 831 832 833
		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;
}

834
/* Module a uses b: caller needs module_mutex() */
835
int ref_module(struct module *a, struct module *b)
L
Linus Torvalds 已提交
836
{
837
	int err;
K
Kay Sievers 已提交
838

839
	if (b == NULL || already_uses(a, b))
840 841
		return 0;

842 843
	/* If module isn't available, we fail. */
	err = strong_try_module_get(b);
844
	if (err)
845
		return err;
L
Linus Torvalds 已提交
846

847 848
	err = add_module_usage(a, b);
	if (err) {
L
Linus Torvalds 已提交
849
		module_put(b);
850
		return err;
L
Linus Torvalds 已提交
851
	}
852
	return 0;
L
Linus Torvalds 已提交
853
}
854
EXPORT_SYMBOL_GPL(ref_module);
L
Linus Torvalds 已提交
855 856 857 858

/* Clear the unload stuff of the module. */
static void module_unload_free(struct module *mod)
{
859
	struct module_use *use, *tmp;
L
Linus Torvalds 已提交
860

861
	mutex_lock(&module_mutex);
862 863
	list_for_each_entry_safe(use, tmp, &mod->target_list, target_list) {
		struct module *i = use->target;
864
		pr_debug("%s unusing %s\n", mod->name, i->name);
865 866 867 868
		module_put(i);
		list_del(&use->source_list);
		list_del(&use->target_list);
		kfree(use);
L
Linus Torvalds 已提交
869
	}
870
	mutex_unlock(&module_mutex);
L
Linus Torvalds 已提交
871 872 873
}

#ifdef CONFIG_MODULE_FORCE_UNLOAD
874
static inline int try_force_unload(unsigned int flags)
L
Linus Torvalds 已提交
875 876 877
{
	int ret = (flags & O_TRUNC);
	if (ret)
878
		add_taint(TAINT_FORCED_RMMOD, LOCKDEP_NOW_UNRELIABLE);
L
Linus Torvalds 已提交
879 880 881
	return ret;
}
#else
882
static inline int try_force_unload(unsigned int flags)
L
Linus Torvalds 已提交
883 884 885 886 887
{
	return 0;
}
#endif /* CONFIG_MODULE_FORCE_UNLOAD */

888 889
/* Try to release refcount of module, 0 means success. */
static int try_release_module_ref(struct module *mod)
L
Linus Torvalds 已提交
890
{
891
	int ret;
L
Linus Torvalds 已提交
892

893 894 895 896 897 898
	/* 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
899

900 901
	return ret;
}
L
Linus Torvalds 已提交
902

903 904
static int try_stop_module(struct module *mod, int flags, int *forced)
{
905
	/* If it's not unused, quit unless we're forcing. */
906 907 908
	if (try_release_module_ref(mod) != 0) {
		*forced = try_force_unload(flags);
		if (!(*forced))
L
Linus Torvalds 已提交
909 910 911 912
			return -EWOULDBLOCK;
	}

	/* Mark it as dying. */
913
	mod->state = MODULE_STATE_GOING;
L
Linus Torvalds 已提交
914

915
	return 0;
L
Linus Torvalds 已提交
916 917
}

918 919 920 921 922 923 924 925 926 927
/**
 * 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
928
{
929
	return atomic_read(&mod->refcnt) - MODULE_REF_BASE;
L
Linus Torvalds 已提交
930 931 932 933 934 935
}
EXPORT_SYMBOL(module_refcount);

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

936 937
SYSCALL_DEFINE2(delete_module, const char __user *, name_user,
		unsigned int, flags)
L
Linus Torvalds 已提交
938 939
{
	struct module *mod;
940
	char name[MODULE_NAME_LEN];
L
Linus Torvalds 已提交
941 942
	int ret, forced = 0;

943
	if (!capable(CAP_SYS_MODULE) || modules_disabled)
944 945 946 947 948 949
		return -EPERM;

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

950 951
	if (mutex_lock_interruptible(&module_mutex) != 0)
		return -EINTR;
L
Linus Torvalds 已提交
952 953 954 955 956 957 958

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

959
	if (!list_empty(&mod->source_list)) {
L
Linus Torvalds 已提交
960 961 962 963 964 965 966
		/* 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) {
967
		/* FIXME: if (force), slam module count damn the torpedoes */
968
		pr_debug("%s already dying\n", mod->name);
L
Linus Torvalds 已提交
969 970 971 972 973
		ret = -EBUSY;
		goto out;
	}

	/* If it has an init func, it must have an exit func to unload */
974
	if (mod->init && !mod->exit) {
975
		forced = try_force_unload(flags);
L
Linus Torvalds 已提交
976 977 978 979 980 981 982 983 984 985 986 987
		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;

988
	mutex_unlock(&module_mutex);
L
Lucas De Marchi 已提交
989
	/* Final destruction now no one is using it. */
990
	if (mod->exit != NULL)
L
Linus Torvalds 已提交
991
		mod->exit();
992 993
	blocking_notifier_call_chain(&module_notify_list,
				     MODULE_STATE_GOING, mod);
994
	klp_module_going(mod);
995 996
	ftrace_release_mod(mod);

997
	async_synchronize_full();
998

999
	/* Store the name of the last unloaded module for diagnostic purposes */
1000
	strlcpy(last_unloaded_module, mod->name, sizeof(last_unloaded_module));
L
Linus Torvalds 已提交
1001

1002 1003 1004
	free_module(mod);
	return 0;
out:
1005
	mutex_unlock(&module_mutex);
L
Linus Torvalds 已提交
1006 1007 1008
	return ret;
}

1009
static inline void print_unload_info(struct seq_file *m, struct module *mod)
L
Linus Torvalds 已提交
1010 1011 1012 1013
{
	struct module_use *use;
	int printed_something = 0;

1014
	seq_printf(m, " %i ", module_refcount(mod));
L
Linus Torvalds 已提交
1015

1016 1017 1018 1019
	/*
	 * Always include a trailing , so userspace can differentiate
	 * between this and the old multi-field proc format.
	 */
1020
	list_for_each_entry(use, &mod->source_list, source_list) {
L
Linus Torvalds 已提交
1021
		printed_something = 1;
1022
		seq_printf(m, "%s,", use->source->name);
L
Linus Torvalds 已提交
1023 1024 1025 1026
	}

	if (mod->init != NULL && mod->exit == NULL) {
		printed_something = 1;
1027
		seq_puts(m, "[permanent],");
L
Linus Torvalds 已提交
1028 1029 1030
	}

	if (!printed_something)
1031
		seq_puts(m, "-");
L
Linus Torvalds 已提交
1032 1033 1034 1035 1036 1037
}

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

1038
	preempt_disable();
1039
	if (!find_symbol(symbol, &owner, NULL, true, false))
L
Linus Torvalds 已提交
1040 1041
		BUG();
	module_put(owner);
1042
	preempt_enable();
L
Linus Torvalds 已提交
1043 1044 1045
}
EXPORT_SYMBOL(__symbol_put);

1046
/* Note this assumes addr is a function, which it currently always is. */
L
Linus Torvalds 已提交
1047 1048
void symbol_put_addr(void *addr)
{
1049
	struct module *modaddr;
1050
	unsigned long a = (unsigned long)dereference_function_descriptor(addr);
L
Linus Torvalds 已提交
1051

1052
	if (core_kernel_text(a))
1053
		return;
L
Linus Torvalds 已提交
1054

1055 1056 1057 1058 1059
	/*
	 * Even though we hold a reference on the module; we still need to
	 * disable preemption in order to safely traverse the data structure.
	 */
	preempt_disable();
1060
	modaddr = __module_text_address(a);
1061
	BUG_ON(!modaddr);
1062
	module_put(modaddr);
1063
	preempt_enable();
L
Linus Torvalds 已提交
1064 1065 1066 1067
}
EXPORT_SYMBOL_GPL(symbol_put_addr);

static ssize_t show_refcnt(struct module_attribute *mattr,
1068
			   struct module_kobject *mk, char *buffer)
L
Linus Torvalds 已提交
1069
{
1070
	return sprintf(buffer, "%i\n", module_refcount(mk->mod));
L
Linus Torvalds 已提交
1071 1072
}

1073 1074
static struct module_attribute modinfo_refcnt =
	__ATTR(refcnt, 0444, show_refcnt, NULL);
L
Linus Torvalds 已提交
1075

1076 1077 1078 1079
void __module_get(struct module *module)
{
	if (module) {
		preempt_disable();
1080
		atomic_inc(&module->refcnt);
1081 1082 1083 1084 1085 1086 1087 1088 1089 1090 1091 1092
		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();
1093 1094 1095
		/* Note: here, we can fail to get a reference */
		if (likely(module_is_live(module) &&
			   atomic_inc_not_zero(&module->refcnt) != 0))
1096
			trace_module_get(module, _RET_IP_);
1097
		else
1098 1099 1100 1101 1102 1103 1104 1105
			ret = false;

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

1106 1107
void module_put(struct module *module)
{
1108 1109
	int ret;

1110
	if (module) {
1111
		preempt_disable();
1112 1113
		ret = atomic_dec_if_positive(&module->refcnt);
		WARN_ON(ret < 0);	/* Failed to put refcount */
1114
		trace_module_put(module, _RET_IP_);
1115
		preempt_enable();
1116 1117 1118 1119
	}
}
EXPORT_SYMBOL(module_put);

L
Linus Torvalds 已提交
1120
#else /* !CONFIG_MODULE_UNLOAD */
1121
static inline void print_unload_info(struct seq_file *m, struct module *mod)
L
Linus Torvalds 已提交
1122 1123
{
	/* We don't know the usage count, or what modules are using. */
1124
	seq_puts(m, " - -");
L
Linus Torvalds 已提交
1125 1126 1127 1128 1129 1130
}

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

1131
int ref_module(struct module *a, struct module *b)
L
Linus Torvalds 已提交
1132
{
1133
	return strong_try_module_get(b);
L
Linus Torvalds 已提交
1134
}
1135
EXPORT_SYMBOL_GPL(ref_module);
L
Linus Torvalds 已提交
1136

1137
static inline int module_unload_init(struct module *mod)
L
Linus Torvalds 已提交
1138
{
1139
	return 0;
L
Linus Torvalds 已提交
1140 1141 1142
}
#endif /* CONFIG_MODULE_UNLOAD */

1143 1144 1145
static size_t module_flags_taint(struct module *mod, char *buf)
{
	size_t l = 0;
1146 1147 1148 1149
	int i;

	for (i = 0; i < TAINT_FLAGS_COUNT; i++) {
		if (taint_flags[i].module && test_bit(i, &mod->taints))
1150
			buf[l++] = taint_flags[i].c_true;
1151
	}
1152 1153 1154 1155

	return l;
}

1156
static ssize_t show_initstate(struct module_attribute *mattr,
1157
			      struct module_kobject *mk, char *buffer)
1158 1159 1160
{
	const char *state = "unknown";

1161
	switch (mk->mod->state) {
1162 1163 1164 1165 1166 1167 1168 1169 1170
	case MODULE_STATE_LIVE:
		state = "live";
		break;
	case MODULE_STATE_COMING:
		state = "coming";
		break;
	case MODULE_STATE_GOING:
		state = "going";
		break;
1171 1172
	default:
		BUG();
1173 1174 1175 1176
	}
	return sprintf(buffer, "%s\n", state);
}

1177 1178
static struct module_attribute modinfo_initstate =
	__ATTR(initstate, 0444, show_initstate, NULL);
1179

1180 1181 1182 1183 1184 1185 1186 1187 1188 1189 1190
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;
}

1191 1192 1193 1194 1195 1196
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)
{
1197
	return sprintf(buffer, "%u\n", mk->mod->core_layout.size);
1198 1199 1200 1201 1202 1203 1204 1205
}

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)
{
1206
	return sprintf(buffer, "%u\n", mk->mod->init_layout.size);
1207 1208 1209 1210 1211 1212 1213 1214 1215 1216 1217 1218 1219 1220 1221 1222 1223
}

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);
1224

1225
static struct module_attribute *modinfo_attrs[] = {
1226
	&module_uevent,
1227 1228
	&modinfo_version,
	&modinfo_srcversion,
1229 1230 1231 1232
	&modinfo_initstate,
	&modinfo_coresize,
	&modinfo_initsize,
	&modinfo_taint,
1233
#ifdef CONFIG_MODULE_UNLOAD
1234
	&modinfo_refcnt,
1235 1236 1237 1238
#endif
	NULL,
};

L
Linus Torvalds 已提交
1239 1240
static const char vermagic[] = VERMAGIC_STRING;

1241
static int try_to_force_load(struct module *mod, const char *reason)
1242 1243
{
#ifdef CONFIG_MODULE_FORCE_LOAD
1244
	if (!test_taint(TAINT_FORCED_MODULE))
1245
		pr_warn("%s: %s: kernel tainted.\n", mod->name, reason);
1246
	add_taint_module(mod, TAINT_FORCED_MODULE, LOCKDEP_NOW_UNRELIABLE);
1247 1248 1249 1250 1251 1252
	return 0;
#else
	return -ENOEXEC;
#endif
}

L
Linus Torvalds 已提交
1253
#ifdef CONFIG_MODVERSIONS
1254 1255

static u32 resolve_rel_crc(const s32 *crc)
1256
{
1257
	return *(u32 *)((void *)crc + *crc);
1258 1259
}

L
Linus Torvalds 已提交
1260 1261 1262
static int check_version(Elf_Shdr *sechdrs,
			 unsigned int versindex,
			 const char *symname,
1263
			 struct module *mod,
1264
			 const s32 *crc)
L
Linus Torvalds 已提交
1265 1266 1267 1268 1269 1270 1271 1272
{
	unsigned int i, num_versions;
	struct modversion_info *versions;

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

1273 1274 1275 1276
	/* 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
1277 1278 1279 1280 1281
	versions = (void *) sechdrs[versindex].sh_addr;
	num_versions = sechdrs[versindex].sh_size
		/ sizeof(struct modversion_info);

	for (i = 0; i < num_versions; i++) {
1282 1283
		u32 crcval;

L
Linus Torvalds 已提交
1284 1285 1286
		if (strcmp(versions[i].name, symname) != 0)
			continue;

1287 1288 1289 1290 1291
		if (IS_ENABLED(CONFIG_MODULE_REL_CRCS))
			crcval = resolve_rel_crc(crc);
		else
			crcval = *crc;
		if (versions[i].crc == crcval)
L
Linus Torvalds 已提交
1292
			return 1;
1293 1294
		pr_debug("Found checksum %X vs module %lX\n",
			 crcval, versions[i].crc);
1295
		goto bad_version;
L
Linus Torvalds 已提交
1296
	}
1297

1298 1299 1300
	/* Broken toolchain. Warn once, then let it go.. */
	pr_warn_once("%s: no symbol version for %s\n", mod->name, symname);
	return 1;
1301 1302

bad_version:
1303
	pr_warn("%s: disagrees about version of symbol %s\n",
1304 1305
	       mod->name, symname);
	return 0;
L
Linus Torvalds 已提交
1306 1307 1308 1309 1310 1311
}

static inline int check_modstruct_version(Elf_Shdr *sechdrs,
					  unsigned int versindex,
					  struct module *mod)
{
1312
	const s32 *crc;
L
Linus Torvalds 已提交
1313

1314 1315 1316 1317 1318
	/*
	 * Since this should be found in kernel (which can't be removed), no
	 * locking is necessary -- use preempt_disable() to placate lockdep.
	 */
	preempt_disable();
1319
	if (!find_symbol(VMLINUX_SYMBOL_STR(module_layout), NULL,
1320 1321
			 &crc, true, false)) {
		preempt_enable();
L
Linus Torvalds 已提交
1322
		BUG();
1323 1324
	}
	preempt_enable();
1325
	return check_version(sechdrs, versindex,
1326
			     VMLINUX_SYMBOL_STR(module_layout), mod, crc);
L
Linus Torvalds 已提交
1327 1328
}

1329 1330 1331
/* 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
1332
{
1333 1334 1335 1336
	if (has_crcs) {
		amagic += strcspn(amagic, " ");
		bmagic += strcspn(bmagic, " ");
	}
L
Linus Torvalds 已提交
1337 1338 1339 1340 1341 1342
	return strcmp(amagic, bmagic) == 0;
}
#else
static inline int check_version(Elf_Shdr *sechdrs,
				unsigned int versindex,
				const char *symname,
1343
				struct module *mod,
1344
				const s32 *crc)
L
Linus Torvalds 已提交
1345 1346 1347 1348 1349 1350 1351 1352 1353 1354 1355
{
	return 1;
}

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

1356 1357
static inline int same_magic(const char *amagic, const char *bmagic,
			     bool has_crcs)
L
Linus Torvalds 已提交
1358 1359 1360 1361 1362
{
	return strcmp(amagic, bmagic) == 0;
}
#endif /* CONFIG_MODVERSIONS */

1363
/* Resolve a symbol for this module.  I.e. if we find one, record usage. */
1364 1365
static const struct kernel_symbol *resolve_symbol(struct module *mod,
						  const struct load_info *info,
1366
						  const char *name,
1367
						  char ownername[])
L
Linus Torvalds 已提交
1368 1369
{
	struct module *owner;
1370
	const struct kernel_symbol *sym;
1371
	const s32 *crc;
1372
	int err;
L
Linus Torvalds 已提交
1373

1374 1375 1376 1377 1378 1379
	/*
	 * 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();
1380
	mutex_lock(&module_mutex);
1381
	sym = find_symbol(name, &owner, &crc,
1382
			  !(mod->taints & (1 << TAINT_PROPRIETARY_MODULE)), true);
1383 1384 1385
	if (!sym)
		goto unlock;

1386
	if (!check_version(info->sechdrs, info->index.vers, name, mod, crc)) {
1387 1388
		sym = ERR_PTR(-EINVAL);
		goto getname;
L
Linus Torvalds 已提交
1389
	}
1390 1391 1392 1393 1394 1395 1396 1397 1398 1399 1400

	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:
1401
	mutex_unlock(&module_mutex);
1402
	return sym;
L
Linus Torvalds 已提交
1403 1404
}

1405 1406 1407 1408
static const struct kernel_symbol *
resolve_symbol_wait(struct module *mod,
		    const struct load_info *info,
		    const char *name)
1409 1410
{
	const struct kernel_symbol *ksym;
1411
	char owner[MODULE_NAME_LEN];
1412 1413

	if (wait_event_interruptible_timeout(module_wq,
1414 1415
			!IS_ERR(ksym = resolve_symbol(mod, info, name, owner))
			|| PTR_ERR(ksym) != -EBUSY,
1416
					     30 * HZ) <= 0) {
1417 1418
		pr_warn("%s: gave up waiting for init of module %s.\n",
			mod->name, owner);
1419 1420 1421 1422
	}
	return ksym;
}

L
Linus Torvalds 已提交
1423 1424 1425 1426
/*
 * /sys/module/foo/sections stuff
 * J. Corbet <corbet@lwn.net>
 */
R
Rusty Russell 已提交
1427
#ifdef CONFIG_SYSFS
1428

R
Rusty Russell 已提交
1429
#ifdef CONFIG_KALLSYMS
1430 1431 1432 1433 1434
static inline bool sect_empty(const Elf_Shdr *sect)
{
	return !(sect->sh_flags & SHF_ALLOC) || sect->sh_size == 0;
}

1435
struct module_sect_attr {
1436 1437 1438 1439 1440
	struct module_attribute mattr;
	char *name;
	unsigned long address;
};

1441
struct module_sect_attrs {
1442 1443 1444 1445 1446
	struct attribute_group grp;
	unsigned int nsections;
	struct module_sect_attr attrs[0];
};

L
Linus Torvalds 已提交
1447
static ssize_t module_sect_show(struct module_attribute *mattr,
1448
				struct module_kobject *mk, char *buf)
L
Linus Torvalds 已提交
1449 1450 1451
{
	struct module_sect_attr *sattr =
		container_of(mattr, struct module_sect_attr, mattr);
1452
	return sprintf(buf, "0x%pK\n", (void *)sattr->address);
L
Linus Torvalds 已提交
1453 1454
}

1455 1456
static void free_sect_attrs(struct module_sect_attrs *sect_attrs)
{
1457
	unsigned int section;
1458 1459 1460 1461 1462 1463

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

R
Rusty Russell 已提交
1464
static void add_sect_attrs(struct module *mod, const struct load_info *info)
L
Linus Torvalds 已提交
1465 1466 1467 1468 1469
{
	unsigned int nloaded = 0, i, size[2];
	struct module_sect_attrs *sect_attrs;
	struct module_sect_attr *sattr;
	struct attribute **gattr;
1470

L
Linus Torvalds 已提交
1471
	/* Count loaded sections and allocate structures */
R
Rusty Russell 已提交
1472 1473
	for (i = 0; i < info->hdr->e_shnum; i++)
		if (!sect_empty(&info->sechdrs[i]))
L
Linus Torvalds 已提交
1474 1475 1476 1477 1478
			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]);
1479 1480
	sect_attrs = kzalloc(size[0] + size[1], GFP_KERNEL);
	if (sect_attrs == NULL)
L
Linus Torvalds 已提交
1481 1482 1483 1484 1485 1486
		return;

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

1487
	sect_attrs->nsections = 0;
L
Linus Torvalds 已提交
1488 1489
	sattr = &sect_attrs->attrs[0];
	gattr = &sect_attrs->grp.attrs[0];
R
Rusty Russell 已提交
1490 1491 1492
	for (i = 0; i < info->hdr->e_shnum; i++) {
		Elf_Shdr *sec = &info->sechdrs[i];
		if (sect_empty(sec))
1493
			continue;
R
Rusty Russell 已提交
1494 1495
		sattr->address = sec->sh_addr;
		sattr->name = kstrdup(info->secstrings + sec->sh_name,
1496 1497 1498 1499
					GFP_KERNEL);
		if (sattr->name == NULL)
			goto out;
		sect_attrs->nsections++;
1500
		sysfs_attr_init(&sattr->mattr.attr);
L
Linus Torvalds 已提交
1501 1502 1503 1504 1505 1506 1507 1508 1509 1510 1511 1512 1513 1514
		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:
1515
	free_sect_attrs(sect_attrs);
L
Linus Torvalds 已提交
1516 1517 1518 1519 1520 1521 1522 1523 1524
}

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. */
1525
		free_sect_attrs(mod->sect_attrs);
L
Linus Torvalds 已提交
1526 1527 1528 1529
		mod->sect_attrs = NULL;
	}
}

1530 1531 1532 1533 1534 1535 1536 1537 1538 1539
/*
 * /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];
};

1540
static ssize_t module_notes_read(struct file *filp, struct kobject *kobj,
1541 1542 1543 1544 1545 1546 1547 1548 1549 1550 1551 1552 1553 1554 1555 1556 1557
				 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]);
1558
		kobject_put(notes_attrs->dir);
1559 1560 1561 1562
	}
	kfree(notes_attrs);
}

R
Rusty Russell 已提交
1563
static void add_notes_attrs(struct module *mod, const struct load_info *info)
1564 1565 1566 1567 1568
{
	unsigned int notes, loaded, i;
	struct module_notes_attrs *notes_attrs;
	struct bin_attribute *nattr;

1569 1570 1571 1572
	/* failed to create section attributes, so can't create notes */
	if (!mod->sect_attrs)
		return;

1573 1574
	/* Count notes sections and allocate structures.  */
	notes = 0;
R
Rusty Russell 已提交
1575 1576 1577
	for (i = 0; i < info->hdr->e_shnum; i++)
		if (!sect_empty(&info->sechdrs[i]) &&
		    (info->sechdrs[i].sh_type == SHT_NOTE))
1578 1579 1580 1581 1582 1583 1584 1585 1586 1587 1588 1589 1590
			++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
1591 1592
	for (loaded = i = 0; i < info->hdr->e_shnum; ++i) {
		if (sect_empty(&info->sechdrs[i]))
1593
			continue;
R
Rusty Russell 已提交
1594
		if (info->sechdrs[i].sh_type == SHT_NOTE) {
1595
			sysfs_bin_attr_init(nattr);
1596 1597
			nattr->attr.name = mod->sect_attrs->attrs[loaded].name;
			nattr->attr.mode = S_IRUGO;
R
Rusty Russell 已提交
1598 1599
			nattr->size = info->sechdrs[i].sh_size;
			nattr->private = (void *) info->sechdrs[i].sh_addr;
1600 1601 1602 1603 1604 1605
			nattr->read = module_notes_read;
			++nattr;
		}
		++loaded;
	}

1606
	notes_attrs->dir = kobject_create_and_add("notes", &mod->mkobj.kobj);
1607 1608 1609 1610 1611 1612 1613 1614 1615 1616 1617 1618 1619 1620 1621 1622 1623 1624 1625 1626 1627
	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
1628
#else
1629

R
Rusty Russell 已提交
1630 1631
static inline void add_sect_attrs(struct module *mod,
				  const struct load_info *info)
L
Linus Torvalds 已提交
1632 1633 1634 1635 1636 1637
{
}

static inline void remove_sect_attrs(struct module *mod)
{
}
1638

R
Rusty Russell 已提交
1639 1640
static inline void add_notes_attrs(struct module *mod,
				   const struct load_info *info)
1641 1642 1643 1644 1645 1646
{
}

static inline void remove_notes_attrs(struct module *mod)
{
}
R
Rusty Russell 已提交
1647
#endif /* CONFIG_KALLSYMS */
L
Linus Torvalds 已提交
1648

1649 1650 1651 1652 1653 1654
static void add_usage_links(struct module *mod)
{
#ifdef CONFIG_MODULE_UNLOAD
	struct module_use *use;
	int nowarn;

1655
	mutex_lock(&module_mutex);
1656 1657 1658 1659
	list_for_each_entry(use, &mod->target_list, target_list) {
		nowarn = sysfs_create_link(use->target->holders_dir,
					   &mod->mkobj.kobj, mod->name);
	}
1660
	mutex_unlock(&module_mutex);
1661 1662 1663 1664 1665 1666 1667 1668
#endif
}

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

1669
	mutex_lock(&module_mutex);
1670 1671
	list_for_each_entry(use, &mod->target_list, target_list)
		sysfs_remove_link(use->target->holders_dir, mod->name);
1672
	mutex_unlock(&module_mutex);
1673 1674 1675
#endif
}

1676
static int module_add_modinfo_attrs(struct module *mod)
1677 1678
{
	struct module_attribute *attr;
1679
	struct module_attribute *temp_attr;
1680 1681 1682
	int error = 0;
	int i;

1683 1684 1685 1686 1687 1688 1689
	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;
1690
	for (i = 0; (attr = modinfo_attrs[i]) && !error; i++) {
1691
		if (!attr->test || attr->test(mod)) {
1692
			memcpy(temp_attr, attr, sizeof(*temp_attr));
1693
			sysfs_attr_init(&temp_attr->attr);
1694 1695
			error = sysfs_create_file(&mod->mkobj.kobj,
					&temp_attr->attr);
1696 1697
			++temp_attr;
		}
1698 1699 1700 1701
	}
	return error;
}

1702
static void module_remove_modinfo_attrs(struct module *mod)
1703 1704 1705 1706
{
	struct module_attribute *attr;
	int i;

1707 1708 1709 1710
	for (i = 0; (attr = &mod->modinfo_attrs[i]); i++) {
		/* pick a field to test for end of list */
		if (!attr->attr.name)
			break;
1711
		sysfs_remove_file(&mod->mkobj.kobj, &attr->attr);
1712 1713
		if (attr->free)
			attr->free(mod);
1714
	}
1715
	kfree(mod->modinfo_attrs);
1716
}
L
Linus Torvalds 已提交
1717

1718 1719 1720 1721 1722 1723 1724 1725
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);
}

1726
static int mod_sysfs_init(struct module *mod)
L
Linus Torvalds 已提交
1727 1728
{
	int err;
1729
	struct kobject *kobj;
L
Linus Torvalds 已提交
1730

1731
	if (!module_sysfs_initialized) {
1732
		pr_err("%s: module sysfs not initialized\n", mod->name);
1733 1734 1735
		err = -EINVAL;
		goto out;
	}
1736 1737 1738

	kobj = kset_find_obj(module_kset, mod->name);
	if (kobj) {
1739
		pr_err("%s: module is already loaded\n", mod->name);
1740 1741 1742 1743 1744
		kobject_put(kobj);
		err = -EINVAL;
		goto out;
	}

L
Linus Torvalds 已提交
1745
	mod->mkobj.mod = mod;
1746

1747 1748 1749 1750 1751
	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)
1752
		mod_kobject_put(mod);
K
Kay Sievers 已提交
1753

1754
	/* delay uevent until full sysfs population */
K
Kay Sievers 已提交
1755 1756 1757 1758
out:
	return err;
}

1759
static int mod_sysfs_setup(struct module *mod,
R
Rusty Russell 已提交
1760
			   const struct load_info *info,
K
Kay Sievers 已提交
1761 1762 1763 1764 1765
			   struct kernel_param *kparam,
			   unsigned int num_params)
{
	int err;

1766 1767 1768 1769
	err = mod_sysfs_init(mod);
	if (err)
		goto out;

1770
	mod->holders_dir = kobject_create_and_add("holders", &mod->mkobj.kobj);
1771 1772
	if (!mod->holders_dir) {
		err = -ENOMEM;
K
Kay Sievers 已提交
1773
		goto out_unreg;
1774
	}
K
Kay Sievers 已提交
1775

L
Linus Torvalds 已提交
1776 1777
	err = module_param_sysfs_setup(mod, kparam, num_params);
	if (err)
K
Kay Sievers 已提交
1778
		goto out_unreg_holders;
L
Linus Torvalds 已提交
1779

1780 1781
	err = module_add_modinfo_attrs(mod);
	if (err)
1782
		goto out_unreg_param;
1783

1784
	add_usage_links(mod);
R
Rusty Russell 已提交
1785 1786
	add_sect_attrs(mod, info);
	add_notes_attrs(mod, info);
1787

1788
	kobject_uevent(&mod->mkobj.kobj, KOBJ_ADD);
L
Linus Torvalds 已提交
1789 1790
	return 0;

1791 1792
out_unreg_param:
	module_param_sysfs_remove(mod);
K
Kay Sievers 已提交
1793
out_unreg_holders:
1794
	kobject_put(mod->holders_dir);
K
Kay Sievers 已提交
1795
out_unreg:
1796
	mod_kobject_put(mod);
1797
out:
L
Linus Torvalds 已提交
1798 1799
	return err;
}
1800 1801 1802

static void mod_sysfs_fini(struct module *mod)
{
R
Rusty Russell 已提交
1803 1804
	remove_notes_attrs(mod);
	remove_sect_attrs(mod);
1805
	mod_kobject_put(mod);
1806 1807
}

1808 1809 1810 1811
static void init_param_lock(struct module *mod)
{
	mutex_init(&mod->param_lock);
}
R
Rusty Russell 已提交
1812
#else /* !CONFIG_SYSFS */
1813

R
Rusty Russell 已提交
1814 1815
static int mod_sysfs_setup(struct module *mod,
			   const struct load_info *info,
1816 1817 1818 1819 1820 1821
			   struct kernel_param *kparam,
			   unsigned int num_params)
{
	return 0;
}

1822 1823 1824 1825
static void mod_sysfs_fini(struct module *mod)
{
}

1826 1827 1828 1829
static void module_remove_modinfo_attrs(struct module *mod)
{
}

1830 1831 1832 1833
static void del_usage_links(struct module *mod)
{
}

1834 1835 1836
static void init_param_lock(struct module *mod)
{
}
1837
#endif /* CONFIG_SYSFS */
L
Linus Torvalds 已提交
1838

1839
static void mod_sysfs_teardown(struct module *mod)
L
Linus Torvalds 已提交
1840
{
1841
	del_usage_links(mod);
1842
	module_remove_modinfo_attrs(mod);
L
Linus Torvalds 已提交
1843
	module_param_sysfs_remove(mod);
1844 1845
	kobject_put(mod->mkobj.drivers_dir);
	kobject_put(mod->holders_dir);
1846
	mod_sysfs_fini(mod);
L
Linus Torvalds 已提交
1847 1848
}

1849 1850 1851 1852
#ifdef CONFIG_DEBUG_SET_MODULE_RONX
/*
 * LKM RO/NX protection: protect module's text/ro-data
 * from modification and any data from execution.
1853 1854
 *
 * General layout of module is:
1855 1856 1857 1858 1859
 *          [text] [read-only-data] [ro-after-init] [writable data]
 * text_size -----^                ^               ^               ^
 * ro_size ------------------------|               |               |
 * ro_after_init_size -----------------------------|               |
 * size -----------------------------------------------------------|
1860 1861
 *
 * These values are always page-aligned (as is base)
1862
 */
1863 1864
static void frob_text(const struct module_layout *layout,
		      int (*set_memory)(unsigned long start, int num_pages))
1865
{
1866 1867 1868 1869
	BUG_ON((unsigned long)layout->base & (PAGE_SIZE-1));
	BUG_ON((unsigned long)layout->text_size & (PAGE_SIZE-1));
	set_memory((unsigned long)layout->base,
		   layout->text_size >> PAGE_SHIFT);
1870 1871
}

1872 1873
static void frob_rodata(const struct module_layout *layout,
			int (*set_memory)(unsigned long start, int num_pages))
1874
{
1875 1876 1877 1878 1879
	BUG_ON((unsigned long)layout->base & (PAGE_SIZE-1));
	BUG_ON((unsigned long)layout->text_size & (PAGE_SIZE-1));
	BUG_ON((unsigned long)layout->ro_size & (PAGE_SIZE-1));
	set_memory((unsigned long)layout->base + layout->text_size,
		   (layout->ro_size - layout->text_size) >> PAGE_SHIFT);
1880 1881
}

1882 1883 1884 1885 1886 1887 1888 1889 1890 1891
static void frob_ro_after_init(const struct module_layout *layout,
				int (*set_memory)(unsigned long start, int num_pages))
{
	BUG_ON((unsigned long)layout->base & (PAGE_SIZE-1));
	BUG_ON((unsigned long)layout->ro_size & (PAGE_SIZE-1));
	BUG_ON((unsigned long)layout->ro_after_init_size & (PAGE_SIZE-1));
	set_memory((unsigned long)layout->base + layout->ro_size,
		   (layout->ro_after_init_size - layout->ro_size) >> PAGE_SHIFT);
}

1892 1893
static void frob_writable_data(const struct module_layout *layout,
			       int (*set_memory)(unsigned long start, int num_pages))
1894
{
1895
	BUG_ON((unsigned long)layout->base & (PAGE_SIZE-1));
1896
	BUG_ON((unsigned long)layout->ro_after_init_size & (PAGE_SIZE-1));
1897
	BUG_ON((unsigned long)layout->size & (PAGE_SIZE-1));
1898 1899
	set_memory((unsigned long)layout->base + layout->ro_after_init_size,
		   (layout->size - layout->ro_after_init_size) >> PAGE_SHIFT);
1900 1901
}

1902 1903
/* livepatching wants to disable read-only so it can frob module. */
void module_disable_ro(const struct module *mod)
1904
{
1905 1906 1907
	if (!rodata_enabled)
		return;

1908 1909
	frob_text(&mod->core_layout, set_memory_rw);
	frob_rodata(&mod->core_layout, set_memory_rw);
1910
	frob_ro_after_init(&mod->core_layout, set_memory_rw);
1911 1912
	frob_text(&mod->init_layout, set_memory_rw);
	frob_rodata(&mod->init_layout, set_memory_rw);
1913
}
1914

1915
void module_enable_ro(const struct module *mod, bool after_init)
1916
{
1917 1918 1919
	if (!rodata_enabled)
		return;

1920 1921 1922 1923
	frob_text(&mod->core_layout, set_memory_ro);
	frob_rodata(&mod->core_layout, set_memory_ro);
	frob_text(&mod->init_layout, set_memory_ro);
	frob_rodata(&mod->init_layout, set_memory_ro);
1924 1925 1926

	if (after_init)
		frob_ro_after_init(&mod->core_layout, set_memory_ro);
1927 1928
}

1929
static void module_enable_nx(const struct module *mod)
1930
{
1931
	frob_rodata(&mod->core_layout, set_memory_nx);
1932
	frob_ro_after_init(&mod->core_layout, set_memory_nx);
1933 1934 1935
	frob_writable_data(&mod->core_layout, set_memory_nx);
	frob_rodata(&mod->init_layout, set_memory_nx);
	frob_writable_data(&mod->init_layout, set_memory_nx);
1936 1937
}

1938
static void module_disable_nx(const struct module *mod)
1939
{
1940
	frob_rodata(&mod->core_layout, set_memory_x);
1941
	frob_ro_after_init(&mod->core_layout, set_memory_x);
1942 1943 1944
	frob_writable_data(&mod->core_layout, set_memory_x);
	frob_rodata(&mod->init_layout, set_memory_x);
	frob_writable_data(&mod->init_layout, set_memory_x);
1945 1946 1947
}

/* Iterate through all modules and set each module's text as RW */
1948
void set_all_modules_text_rw(void)
1949 1950 1951
{
	struct module *mod;

1952 1953 1954
	if (!rodata_enabled)
		return;

1955 1956
	mutex_lock(&module_mutex);
	list_for_each_entry_rcu(mod, &modules, list) {
1957 1958
		if (mod->state == MODULE_STATE_UNFORMED)
			continue;
1959 1960 1961

		frob_text(&mod->core_layout, set_memory_rw);
		frob_text(&mod->init_layout, set_memory_rw);
1962 1963 1964 1965 1966
	}
	mutex_unlock(&module_mutex);
}

/* Iterate through all modules and set each module's text as RO */
1967
void set_all_modules_text_ro(void)
1968 1969 1970
{
	struct module *mod;

1971 1972 1973
	if (!rodata_enabled)
		return;

1974 1975
	mutex_lock(&module_mutex);
	list_for_each_entry_rcu(mod, &modules, list) {
1976 1977 1978 1979 1980 1981 1982
		/*
		 * Ignore going modules since it's possible that ro
		 * protection has already been disabled, otherwise we'll
		 * run into protection faults at module deallocation.
		 */
		if (mod->state == MODULE_STATE_UNFORMED ||
			mod->state == MODULE_STATE_GOING)
1983
			continue;
1984 1985 1986

		frob_text(&mod->core_layout, set_memory_ro);
		frob_text(&mod->init_layout, set_memory_ro);
1987 1988 1989
	}
	mutex_unlock(&module_mutex);
}
1990 1991 1992

static void disable_ro_nx(const struct module_layout *layout)
{
1993 1994 1995 1996 1997
	if (rodata_enabled) {
		frob_text(layout, set_memory_rw);
		frob_rodata(layout, set_memory_rw);
		frob_ro_after_init(layout, set_memory_rw);
	}
1998
	frob_rodata(layout, set_memory_x);
1999
	frob_ro_after_init(layout, set_memory_x);
2000 2001 2002
	frob_writable_data(layout, set_memory_x);
}

2003
#else
2004 2005 2006
static void disable_ro_nx(const struct module_layout *layout) { }
static void module_enable_nx(const struct module *mod) { }
static void module_disable_nx(const struct module *mod) { }
2007 2008
#endif

2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029 2030 2031 2032 2033 2034 2035 2036 2037 2038 2039 2040 2041 2042 2043 2044 2045 2046 2047 2048 2049 2050 2051 2052 2053 2054 2055 2056 2057 2058 2059 2060 2061 2062 2063 2064 2065 2066 2067 2068 2069 2070 2071 2072 2073 2074 2075 2076 2077 2078 2079 2080 2081 2082 2083 2084 2085
#ifdef CONFIG_LIVEPATCH
/*
 * Persist Elf information about a module. Copy the Elf header,
 * section header table, section string table, and symtab section
 * index from info to mod->klp_info.
 */
static int copy_module_elf(struct module *mod, struct load_info *info)
{
	unsigned int size, symndx;
	int ret;

	size = sizeof(*mod->klp_info);
	mod->klp_info = kmalloc(size, GFP_KERNEL);
	if (mod->klp_info == NULL)
		return -ENOMEM;

	/* Elf header */
	size = sizeof(mod->klp_info->hdr);
	memcpy(&mod->klp_info->hdr, info->hdr, size);

	/* Elf section header table */
	size = sizeof(*info->sechdrs) * info->hdr->e_shnum;
	mod->klp_info->sechdrs = kmalloc(size, GFP_KERNEL);
	if (mod->klp_info->sechdrs == NULL) {
		ret = -ENOMEM;
		goto free_info;
	}
	memcpy(mod->klp_info->sechdrs, info->sechdrs, size);

	/* Elf section name string table */
	size = info->sechdrs[info->hdr->e_shstrndx].sh_size;
	mod->klp_info->secstrings = kmalloc(size, GFP_KERNEL);
	if (mod->klp_info->secstrings == NULL) {
		ret = -ENOMEM;
		goto free_sechdrs;
	}
	memcpy(mod->klp_info->secstrings, info->secstrings, size);

	/* Elf symbol section index */
	symndx = info->index.sym;
	mod->klp_info->symndx = symndx;

	/*
	 * For livepatch modules, core_kallsyms.symtab is a complete
	 * copy of the original symbol table. Adjust sh_addr to point
	 * to core_kallsyms.symtab since the copy of the symtab in module
	 * init memory is freed at the end of do_init_module().
	 */
	mod->klp_info->sechdrs[symndx].sh_addr = \
		(unsigned long) mod->core_kallsyms.symtab;

	return 0;

free_sechdrs:
	kfree(mod->klp_info->sechdrs);
free_info:
	kfree(mod->klp_info);
	return ret;
}

static void free_module_elf(struct module *mod)
{
	kfree(mod->klp_info->sechdrs);
	kfree(mod->klp_info->secstrings);
	kfree(mod->klp_info);
}
#else /* !CONFIG_LIVEPATCH */
static int copy_module_elf(struct module *mod, struct load_info *info)
{
	return 0;
}

static void free_module_elf(struct module *mod)
{
}
#endif /* CONFIG_LIVEPATCH */

2086
void __weak module_memfree(void *module_region)
2087 2088 2089 2090 2091 2092 2093 2094
{
	vfree(module_region);
}

void __weak module_arch_cleanup(struct module *mod)
{
}

2095 2096 2097 2098
void __weak module_arch_freeing_init(struct module *mod)
{
}

2099
/* Free a module, remove from lists, etc. */
L
Linus Torvalds 已提交
2100 2101
static void free_module(struct module *mod)
{
2102 2103
	trace_module_free(mod);

2104
	mod_sysfs_teardown(mod);
L
Linus Torvalds 已提交
2105

2106 2107
	/* We leave it in list to prevent duplicate loads, but make sure
	 * that noone uses it while it's being deconstructed. */
2108
	mutex_lock(&module_mutex);
2109
	mod->state = MODULE_STATE_UNFORMED;
2110
	mutex_unlock(&module_mutex);
2111

2112 2113 2114
	/* Remove dynamic debug info */
	ddebug_remove_module(mod->name);

L
Linus Torvalds 已提交
2115 2116 2117 2118 2119 2120
	/* Arch-specific cleanup. */
	module_arch_cleanup(mod);

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

2121 2122 2123
	/* Free any allocated parameters. */
	destroy_params(mod->kp, mod->num_kp);

2124 2125 2126
	if (is_livepatch_module(mod))
		free_module_elf(mod);

2127 2128
	/* Now we can delete it from the lists */
	mutex_lock(&module_mutex);
2129 2130
	/* Unlink carefully: kallsyms could be walking list. */
	list_del_rcu(&mod->list);
2131
	mod_tree_remove(mod);
2132
	/* Remove this module from bug list, this uses list_del_rcu */
2133
	module_bug_cleanup(mod);
2134 2135
	/* Wait for RCU-sched synchronizing before releasing mod->list and buglist. */
	synchronize_sched();
2136 2137
	mutex_unlock(&module_mutex);

2138 2139
	/* This may be empty, but that's OK */
	disable_ro_nx(&mod->init_layout);
2140
	module_arch_freeing_init(mod);
2141
	module_memfree(mod->init_layout.base);
L
Linus Torvalds 已提交
2142
	kfree(mod->args);
2143
	percpu_modfree(mod);
2144

2145
	/* Free lock-classes; relies on the preceding sync_rcu(). */
2146
	lockdep_free_key_range(mod->core_layout.base, mod->core_layout.size);
I
Ingo Molnar 已提交
2147

L
Linus Torvalds 已提交
2148
	/* Finally, free the core (containing the module structure) */
2149
	disable_ro_nx(&mod->core_layout);
2150
	module_memfree(mod->core_layout.base);
2151 2152 2153 2154

#ifdef CONFIG_MPU
	update_protections(current->mm);
#endif
L
Linus Torvalds 已提交
2155 2156 2157 2158 2159
}

void *__symbol_get(const char *symbol)
{
	struct module *owner;
2160
	const struct kernel_symbol *sym;
L
Linus Torvalds 已提交
2161

2162
	preempt_disable();
2163 2164 2165
	sym = find_symbol(symbol, &owner, NULL, true, true);
	if (sym && strong_try_module_get(owner))
		sym = NULL;
2166
	preempt_enable();
L
Linus Torvalds 已提交
2167

2168
	return sym ? (void *)sym->value : NULL;
L
Linus Torvalds 已提交
2169 2170 2171
}
EXPORT_SYMBOL_GPL(__symbol_get);

2172 2173
/*
 * Ensure that an exported symbol [global namespace] does not already exist
2174
 * in the kernel or in some other module's exported symbol table.
2175 2176
 *
 * You must hold the module_mutex.
2177 2178 2179
 */
static int verify_export_symbols(struct module *mod)
{
2180
	unsigned int i;
2181
	struct module *owner;
2182 2183 2184 2185 2186 2187 2188 2189
	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 },
2190
#ifdef CONFIG_UNUSED_SYMBOLS
2191 2192
		{ mod->unused_syms, mod->num_unused_syms },
		{ mod->unused_gpl_syms, mod->num_unused_gpl_syms },
2193
#endif
2194
	};
2195

2196 2197
	for (i = 0; i < ARRAY_SIZE(arr); i++) {
		for (s = arr[i].sym; s < arr[i].sym + arr[i].num; s++) {
2198
			if (find_symbol(s->name, &owner, NULL, true, false)) {
2199
				pr_err("%s: exports duplicate symbol %s"
2200 2201 2202 2203
				       " (owned by %s)\n",
				       mod->name, s->name, module_name(owner));
				return -ENOEXEC;
			}
2204
		}
2205 2206
	}
	return 0;
2207 2208
}

2209
/* Change all symbols so that st_value encodes the pointer directly. */
2210 2211 2212 2213
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
2214
	unsigned long secbase;
2215
	unsigned int i;
L
Linus Torvalds 已提交
2216
	int ret = 0;
2217
	const struct kernel_symbol *ksym;
L
Linus Torvalds 已提交
2218

2219 2220 2221
	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
2222 2223
		switch (sym[i].st_shndx) {
		case SHN_COMMON:
2224 2225 2226 2227
			/* Ignore common symbols */
			if (!strncmp(name, "__gnu_lto", 9))
				break;

L
Linus Torvalds 已提交
2228 2229
			/* We compiled with -fno-common.  These are not
			   supposed to happen.  */
2230
			pr_debug("Common symbol: %s\n", name);
2231
			pr_warn("%s: please compile with -fno-common\n",
L
Linus Torvalds 已提交
2232 2233 2234 2235 2236 2237
			       mod->name);
			ret = -ENOEXEC;
			break;

		case SHN_ABS:
			/* Don't need to do anything */
2238
			pr_debug("Absolute symbol: 0x%08lx\n",
L
Linus Torvalds 已提交
2239 2240 2241
			       (long)sym[i].st_value);
			break;

2242 2243 2244 2245
		case SHN_LIVEPATCH:
			/* Livepatch symbols are resolved by livepatch */
			break;

L
Linus Torvalds 已提交
2246
		case SHN_UNDEF:
2247
			ksym = resolve_symbol_wait(mod, info, name);
L
Linus Torvalds 已提交
2248
			/* Ok if resolved.  */
2249
			if (ksym && !IS_ERR(ksym)) {
2250
				sym[i].st_value = ksym->value;
L
Linus Torvalds 已提交
2251
				break;
2252 2253
			}

L
Linus Torvalds 已提交
2254
			/* Ok if weak.  */
2255
			if (!ksym && ELF_ST_BIND(sym[i].st_info) == STB_WEAK)
L
Linus Torvalds 已提交
2256 2257
				break;

2258 2259
			pr_warn("%s: Unknown symbol %s (err %li)\n",
				mod->name, name, PTR_ERR(ksym));
2260
			ret = PTR_ERR(ksym) ?: -ENOENT;
L
Linus Torvalds 已提交
2261 2262 2263 2264
			break;

		default:
			/* Divert to percpu allocation if a percpu var. */
2265
			if (sym[i].st_shndx == info->index.pcpu)
2266
				secbase = (unsigned long)mod_percpu(mod);
L
Linus Torvalds 已提交
2267
			else
2268
				secbase = info->sechdrs[sym[i].st_shndx].sh_addr;
L
Linus Torvalds 已提交
2269 2270 2271 2272 2273 2274 2275 2276
			sym[i].st_value += secbase;
			break;
		}
	}

	return ret;
}

2277
static int apply_relocations(struct module *mod, const struct load_info *info)
2278 2279 2280 2281 2282
{
	unsigned int i;
	int err = 0;

	/* Now do relocations. */
2283 2284
	for (i = 1; i < info->hdr->e_shnum; i++) {
		unsigned int infosec = info->sechdrs[i].sh_info;
2285 2286

		/* Not a valid relocation section? */
2287
		if (infosec >= info->hdr->e_shnum)
2288 2289 2290
			continue;

		/* Don't bother with non-allocated sections */
2291
		if (!(info->sechdrs[infosec].sh_flags & SHF_ALLOC))
2292 2293
			continue;

2294 2295 2296 2297
		/* Livepatch relocation sections are applied by livepatch */
		if (info->sechdrs[i].sh_flags & SHF_RELA_LIVEPATCH)
			continue;

2298 2299 2300 2301 2302 2303
		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);
2304 2305 2306 2307 2308 2309
		if (err < 0)
			break;
	}
	return err;
}

2310 2311 2312 2313 2314 2315 2316 2317
/* 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
2318
/* Update size with this section: return offset. */
2319 2320
static long get_offset(struct module *mod, unsigned int *size,
		       Elf_Shdr *sechdr, unsigned int section)
L
Linus Torvalds 已提交
2321 2322 2323
{
	long ret;

2324
	*size += arch_mod_section_prepend(mod, section);
L
Linus Torvalds 已提交
2325 2326 2327 2328 2329 2330 2331 2332 2333
	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. */
2334
static void layout_sections(struct module *mod, struct load_info *info)
L
Linus Torvalds 已提交
2335 2336 2337 2338 2339 2340 2341
{
	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 },
2342
		{ SHF_RO_AFTER_INIT | SHF_ALLOC, ARCH_SHF_SMALL },
L
Linus Torvalds 已提交
2343 2344 2345 2346 2347
		{ SHF_WRITE | SHF_ALLOC, ARCH_SHF_SMALL },
		{ ARCH_SHF_SMALL | SHF_ALLOC, 0 }
	};
	unsigned int m, i;

2348 2349
	for (i = 0; i < info->hdr->e_shnum; i++)
		info->sechdrs[i].sh_entsize = ~0UL;
L
Linus Torvalds 已提交
2350

2351
	pr_debug("Core section allocation order:\n");
L
Linus Torvalds 已提交
2352
	for (m = 0; m < ARRAY_SIZE(masks); ++m) {
2353 2354 2355
		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
2356 2357 2358 2359

			if ((s->sh_flags & masks[m][0]) != masks[m][0]
			    || (s->sh_flags & masks[m][1])
			    || s->sh_entsize != ~0UL
2360
			    || strstarts(sname, ".init"))
L
Linus Torvalds 已提交
2361
				continue;
2362
			s->sh_entsize = get_offset(mod, &mod->core_layout.size, s, i);
2363
			pr_debug("\t%s\n", sname);
L
Linus Torvalds 已提交
2364
		}
2365 2366
		switch (m) {
		case 0: /* executable */
2367 2368
			mod->core_layout.size = debug_align(mod->core_layout.size);
			mod->core_layout.text_size = mod->core_layout.size;
2369 2370
			break;
		case 1: /* RO: text and ro-data */
2371 2372
			mod->core_layout.size = debug_align(mod->core_layout.size);
			mod->core_layout.ro_size = mod->core_layout.size;
2373
			break;
2374 2375 2376 2377 2378
		case 2: /* RO after init */
			mod->core_layout.size = debug_align(mod->core_layout.size);
			mod->core_layout.ro_after_init_size = mod->core_layout.size;
			break;
		case 4: /* whole core */
2379
			mod->core_layout.size = debug_align(mod->core_layout.size);
2380 2381
			break;
		}
L
Linus Torvalds 已提交
2382 2383
	}

2384
	pr_debug("Init section allocation order:\n");
L
Linus Torvalds 已提交
2385
	for (m = 0; m < ARRAY_SIZE(masks); ++m) {
2386 2387 2388
		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
2389 2390 2391 2392

			if ((s->sh_flags & masks[m][0]) != masks[m][0]
			    || (s->sh_flags & masks[m][1])
			    || s->sh_entsize != ~0UL
2393
			    || !strstarts(sname, ".init"))
L
Linus Torvalds 已提交
2394
				continue;
2395
			s->sh_entsize = (get_offset(mod, &mod->init_layout.size, s, i)
L
Linus Torvalds 已提交
2396
					 | INIT_OFFSET_MASK);
2397
			pr_debug("\t%s\n", sname);
L
Linus Torvalds 已提交
2398
		}
2399 2400
		switch (m) {
		case 0: /* executable */
2401 2402
			mod->init_layout.size = debug_align(mod->init_layout.size);
			mod->init_layout.text_size = mod->init_layout.size;
2403 2404
			break;
		case 1: /* RO: text and ro-data */
2405 2406
			mod->init_layout.size = debug_align(mod->init_layout.size);
			mod->init_layout.ro_size = mod->init_layout.size;
2407
			break;
2408 2409 2410 2411 2412 2413 2414 2415
		case 2:
			/*
			 * RO after init doesn't apply to init_layout (only
			 * core_layout), so it just takes the value of ro_size.
			 */
			mod->init_layout.ro_after_init_size = mod->init_layout.ro_size;
			break;
		case 4: /* whole init */
2416
			mod->init_layout.size = debug_align(mod->init_layout.size);
2417 2418
			break;
		}
L
Linus Torvalds 已提交
2419 2420 2421 2422 2423 2424 2425 2426
	}
}

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

2427
	if (!license_is_gpl_compatible(license)) {
2428
		if (!test_taint(TAINT_PROPRIETARY_MODULE))
2429 2430
			pr_warn("%s: module license '%s' taints kernel.\n",
				mod->name, license);
2431 2432
		add_taint_module(mod, TAINT_PROPRIETARY_MODULE,
				 LOCKDEP_NOW_UNRELIABLE);
L
Linus Torvalds 已提交
2433 2434 2435 2436 2437 2438 2439 2440 2441 2442 2443 2444 2445 2446 2447 2448 2449 2450 2451 2452 2453 2454
	}
}

/* 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;
}

2455
static char *get_modinfo(struct load_info *info, const char *tag)
L
Linus Torvalds 已提交
2456 2457 2458
{
	char *p;
	unsigned int taglen = strlen(tag);
2459 2460
	Elf_Shdr *infosec = &info->sechdrs[info->index.info];
	unsigned long size = infosec->sh_size;
L
Linus Torvalds 已提交
2461

2462
	for (p = (char *)infosec->sh_addr; p; p = next_string(p, &size)) {
L
Linus Torvalds 已提交
2463 2464 2465 2466 2467 2468
		if (strncmp(p, tag, taglen) == 0 && p[taglen] == '=')
			return p + taglen + 1;
	}
	return NULL;
}

2469
static void setup_modinfo(struct module *mod, struct load_info *info)
2470 2471 2472 2473 2474 2475
{
	struct module_attribute *attr;
	int i;

	for (i = 0; (attr = modinfo_attrs[i]); i++) {
		if (attr->setup)
2476
			attr->setup(mod, get_modinfo(info, attr->attr.name));
2477 2478 2479
	}
}

2480 2481 2482 2483 2484 2485 2486 2487 2488 2489 2490
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
2491
#ifdef CONFIG_KALLSYMS
2492 2493 2494 2495 2496 2497

/* 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)
{
2498 2499
	return bsearch(name, start, stop - start,
			sizeof(struct kernel_symbol), cmp_name);
2500 2501
}

2502 2503
static int is_exported(const char *name, unsigned long value,
		       const struct module *mod)
L
Linus Torvalds 已提交
2504
{
2505 2506 2507
	const struct kernel_symbol *ks;
	if (!mod)
		ks = lookup_symbol(name, __start___ksymtab, __stop___ksymtab);
2508
	else
2509 2510
		ks = lookup_symbol(name, mod->syms, mod->syms + mod->num_syms);
	return ks != NULL && ks->value == value;
L
Linus Torvalds 已提交
2511 2512 2513
}

/* As per nm */
2514
static char elf_type(const Elf_Sym *sym, const struct load_info *info)
L
Linus Torvalds 已提交
2515
{
2516 2517
	const Elf_Shdr *sechdrs = info->sechdrs;

L
Linus Torvalds 已提交
2518 2519 2520 2521 2522 2523 2524 2525
	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';
2526
	if (sym->st_shndx == SHN_ABS || sym->st_shndx == info->index.pcpu)
L
Linus Torvalds 已提交
2527 2528 2529 2530 2531 2532 2533 2534 2535 2536 2537 2538 2539 2540 2541 2542 2543 2544 2545 2546
		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';
	}
2547 2548
	if (strstarts(info->secstrings + sechdrs[sym->st_shndx].sh_name,
		      ".debug")) {
L
Linus Torvalds 已提交
2549
		return 'n';
2550
	}
L
Linus Torvalds 已提交
2551 2552 2553
	return '?';
}

2554
static bool is_core_symbol(const Elf_Sym *src, const Elf_Shdr *sechdrs,
2555
			unsigned int shnum, unsigned int pcpundx)
2556 2557 2558 2559 2560 2561 2562 2563
{
	const Elf_Shdr *sec;

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

2564 2565 2566 2567 2568
#ifdef CONFIG_KALLSYMS_ALL
	if (src->st_shndx == pcpundx)
		return true;
#endif

2569 2570 2571 2572 2573 2574 2575 2576 2577 2578 2579
	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;
}

2580 2581 2582 2583 2584 2585 2586
/*
 * 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>.
 */
2587
static void layout_symtab(struct module *mod, struct load_info *info)
2588
{
2589 2590
	Elf_Shdr *symsect = info->sechdrs + info->index.sym;
	Elf_Shdr *strsect = info->sechdrs + info->index.str;
2591
	const Elf_Sym *src;
2592
	unsigned int i, nsrc, ndst, strtab_size = 0;
2593 2594 2595

	/* Put symbol section at end of init part of module. */
	symsect->sh_flags |= SHF_ALLOC;
2596
	symsect->sh_entsize = get_offset(mod, &mod->init_layout.size, symsect,
2597
					 info->index.sym) | INIT_OFFSET_MASK;
2598
	pr_debug("\t%s\n", info->secstrings + symsect->sh_name);
2599

2600
	src = (void *)info->hdr + symsect->sh_offset;
2601
	nsrc = symsect->sh_size / sizeof(*src);
2602

2603
	/* Compute total space required for the core symbols' strtab. */
2604
	for (ndst = i = 0; i < nsrc; i++) {
2605
		if (i == 0 || is_livepatch_module(mod) ||
2606 2607
		    is_core_symbol(src+i, info->sechdrs, info->hdr->e_shnum,
				   info->index.pcpu)) {
2608
			strtab_size += strlen(&info->strtab[src[i].st_name])+1;
2609
			ndst++;
2610
		}
2611
	}
2612 2613

	/* Append room for core symbols at end of core part. */
2614 2615 2616 2617
	info->symoffs = ALIGN(mod->core_layout.size, symsect->sh_addralign ?: 1);
	info->stroffs = mod->core_layout.size = info->symoffs + ndst * sizeof(Elf_Sym);
	mod->core_layout.size += strtab_size;
	mod->core_layout.size = debug_align(mod->core_layout.size);
2618

2619 2620
	/* Put string table section at end of init part of module. */
	strsect->sh_flags |= SHF_ALLOC;
2621
	strsect->sh_entsize = get_offset(mod, &mod->init_layout.size, strsect,
2622
					 info->index.str) | INIT_OFFSET_MASK;
2623
	pr_debug("\t%s\n", info->secstrings + strsect->sh_name);
2624 2625 2626 2627 2628 2629 2630

	/* We'll tack temporary mod_kallsyms on the end. */
	mod->init_layout.size = ALIGN(mod->init_layout.size,
				      __alignof__(struct mod_kallsyms));
	info->mod_kallsyms_init_off = mod->init_layout.size;
	mod->init_layout.size += sizeof(struct mod_kallsyms);
	mod->init_layout.size = debug_align(mod->init_layout.size);
2631 2632
}

2633 2634 2635 2636 2637
/*
 * We use the full symtab and strtab which layout_symtab arranged to
 * be appended to the init section.  Later we switch to the cut-down
 * core-only ones.
 */
2638
static void add_kallsyms(struct module *mod, const struct load_info *info)
L
Linus Torvalds 已提交
2639
{
2640 2641 2642
	unsigned int i, ndst;
	const Elf_Sym *src;
	Elf_Sym *dst;
2643
	char *s;
2644
	Elf_Shdr *symsec = &info->sechdrs[info->index.sym];
L
Linus Torvalds 已提交
2645

2646 2647 2648 2649 2650
	/* Set up to point into init section. */
	mod->kallsyms = mod->init_layout.base + info->mod_kallsyms_init_off;

	mod->kallsyms->symtab = (void *)symsec->sh_addr;
	mod->kallsyms->num_symtab = symsec->sh_size / sizeof(Elf_Sym);
2651
	/* Make sure we get permanent strtab: don't use info->strtab. */
2652
	mod->kallsyms->strtab = (void *)info->sechdrs[info->index.str].sh_addr;
L
Linus Torvalds 已提交
2653 2654

	/* Set types up while we still have access to sections. */
2655 2656 2657 2658 2659 2660 2661 2662 2663
	for (i = 0; i < mod->kallsyms->num_symtab; i++)
		mod->kallsyms->symtab[i].st_info
			= elf_type(&mod->kallsyms->symtab[i], info);

	/* Now populate the cut down core kallsyms for after init. */
	mod->core_kallsyms.symtab = dst = mod->core_layout.base + info->symoffs;
	mod->core_kallsyms.strtab = s = mod->core_layout.base + info->stroffs;
	src = mod->kallsyms->symtab;
	for (ndst = i = 0; i < mod->kallsyms->num_symtab; i++) {
2664
		if (i == 0 || is_livepatch_module(mod) ||
2665 2666
		    is_core_symbol(src+i, info->sechdrs, info->hdr->e_shnum,
				   info->index.pcpu)) {
2667
			dst[ndst] = src[i];
2668 2669
			dst[ndst++].st_name = s - mod->core_kallsyms.strtab;
			s += strlcpy(s, &mod->kallsyms->strtab[src[i].st_name],
2670 2671
				     KSYM_NAME_LEN) + 1;
		}
2672
	}
2673
	mod->core_kallsyms.num_symtab = ndst;
L
Linus Torvalds 已提交
2674 2675
}
#else
2676
static inline void layout_symtab(struct module *mod, struct load_info *info)
2677 2678
{
}
2679

2680
static void add_kallsyms(struct module *mod, const struct load_info *info)
L
Linus Torvalds 已提交
2681 2682 2683 2684
{
}
#endif /* CONFIG_KALLSYMS */

2685
static void dynamic_debug_setup(struct _ddebug *debug, unsigned int num)
2686
{
2687 2688
	if (!debug)
		return;
2689 2690
#ifdef CONFIG_DYNAMIC_DEBUG
	if (ddebug_add_module(debug, num, debug->modname))
2691 2692
		pr_err("dynamic debug error adding module: %s\n",
			debug->modname);
2693
#endif
2694
}
2695

2696 2697 2698 2699 2700 2701
static void dynamic_debug_remove(struct _ddebug *debug)
{
	if (debug)
		ddebug_remove_module(debug->modname);
}

2702 2703
void * __weak module_alloc(unsigned long size)
{
2704
	return vmalloc_exec(size);
2705 2706
}

2707
#ifdef CONFIG_DEBUG_KMEMLEAK
2708 2709
static void kmemleak_load_module(const struct module *mod,
				 const struct load_info *info)
2710 2711 2712 2713
{
	unsigned int i;

	/* only scan the sections containing data */
2714
	kmemleak_scan_area(mod, sizeof(struct module), GFP_KERNEL);
2715

2716
	for (i = 1; i < info->hdr->e_shnum; i++) {
2717 2718 2719 2720
		/* 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))
2721 2722
			continue;

2723 2724
		kmemleak_scan_area((void *)info->sechdrs[i].sh_addr,
				   info->sechdrs[i].sh_size, GFP_KERNEL);
2725 2726 2727
	}
}
#else
2728 2729
static inline void kmemleak_load_module(const struct module *mod,
					const struct load_info *info)
2730 2731 2732 2733
{
}
#endif

2734
#ifdef CONFIG_MODULE_SIG
2735
static int module_sig_check(struct load_info *info, int flags)
2736 2737
{
	int err = -ENOKEY;
2738 2739
	const unsigned long markerlen = sizeof(MODULE_SIG_STRING) - 1;
	const void *mod = info->hdr;
2740

2741 2742 2743 2744 2745 2746
	/*
	 * Require flags == 0, as a module with version information
	 * removed is no longer the module that was signed
	 */
	if (flags == 0 &&
	    info->len > markerlen &&
2747
	    memcmp(mod + info->len - markerlen, MODULE_SIG_STRING, markerlen) == 0) {
2748
		/* We truncate the module to discard the signature */
2749 2750
		info->len -= markerlen;
		err = mod_verify_sig(mod, &info->len);
2751 2752 2753 2754 2755 2756 2757 2758 2759 2760 2761 2762 2763 2764
	}

	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 */
2765
static int module_sig_check(struct load_info *info, int flags)
2766 2767 2768 2769 2770
{
	return 0;
}
#endif /* !CONFIG_MODULE_SIG */

2771 2772
/* Sanity checks against invalid binaries, wrong arch, weird elf version. */
static int elf_header_check(struct load_info *info)
2773
{
2774 2775 2776 2777 2778 2779 2780 2781 2782 2783 2784 2785 2786
	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;
2787

2788 2789 2790
	return 0;
}

2791 2792 2793 2794 2795 2796 2797 2798 2799 2800 2801 2802 2803 2804 2805 2806 2807
#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;
}

2808
#ifdef CONFIG_LIVEPATCH
2809
static int check_modinfo_livepatch(struct module *mod, struct load_info *info)
2810
{
2811 2812 2813 2814
	if (get_modinfo(info, "livepatch")) {
		mod->klp = true;
		add_taint_module(mod, TAINT_LIVEPATCH, LOCKDEP_STILL_OK);
	}
2815 2816 2817 2818

	return 0;
}
#else /* !CONFIG_LIVEPATCH */
2819
static int check_modinfo_livepatch(struct module *mod, struct load_info *info)
2820 2821 2822 2823 2824 2825 2826 2827 2828 2829 2830
{
	if (get_modinfo(info, "livepatch")) {
		pr_err("%s: module is marked as livepatch module, but livepatch support is disabled",
		       mod->name);
		return -ENOEXEC;
	}

	return 0;
}
#endif /* CONFIG_LIVEPATCH */

2831 2832 2833
/* Sets info->hdr and info->len. */
static int copy_module_from_user(const void __user *umod, unsigned long len,
				  struct load_info *info)
2834 2835 2836
{
	int err;

2837 2838
	info->len = len;
	if (info->len < sizeof(*(info->hdr)))
2839 2840
		return -ENOEXEC;

2841
	err = security_kernel_read_file(NULL, READING_MODULE);
2842 2843 2844
	if (err)
		return err;

2845
	/* Suck in entire file: we'll want most of it. */
2846 2847
	info->hdr = __vmalloc(info->len,
			GFP_KERNEL | __GFP_HIGHMEM | __GFP_NOWARN, PAGE_KERNEL);
2848
	if (!info->hdr)
2849 2850
		return -ENOMEM;

2851
	if (copy_chunked_from_user(info->hdr, umod, info->len) != 0) {
2852 2853
		vfree(info->hdr);
		return -EFAULT;
2854 2855
	}

2856 2857 2858
	return 0;
}

2859 2860 2861 2862 2863
static void free_copy(struct load_info *info)
{
	vfree(info->hdr);
}

2864
static int rewrite_section_headers(struct load_info *info, int flags)
2865 2866 2867 2868 2869 2870 2871 2872 2873 2874
{
	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) {
2875
			pr_err("Module len %lu truncated\n", info->len);
2876 2877 2878 2879 2880 2881 2882 2883 2884 2885 2886 2887 2888
			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
	}
2889 2890

	/* Track but don't keep modinfo and version sections. */
2891 2892 2893 2894
	if (flags & MODULE_INIT_IGNORE_MODVERSIONS)
		info->index.vers = 0; /* Pretend no __versions section! */
	else
		info->index.vers = find_sec(info, "__versions");
2895
	info->index.info = find_sec(info, ".modinfo");
2896 2897
	info->sechdrs[info->index.info].sh_flags &= ~(unsigned long)SHF_ALLOC;
	info->sechdrs[info->index.vers].sh_flags &= ~(unsigned long)SHF_ALLOC;
2898 2899 2900
	return 0;
}

L
Linus Torvalds 已提交
2901 2902 2903 2904 2905 2906 2907 2908
/*
 * 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).
 */
2909
static struct module *setup_load_info(struct load_info *info, int flags)
L
Linus Torvalds 已提交
2910 2911
{
	unsigned int i;
2912
	int err;
L
Linus Torvalds 已提交
2913 2914 2915 2916
	struct module *mod;

	/* Set up the convenience variables */
	info->sechdrs = (void *)info->hdr + info->hdr->e_shoff;
2917 2918
	info->secstrings = (void *)info->hdr
		+ info->sechdrs[info->hdr->e_shstrndx].sh_offset;
L
Linus Torvalds 已提交
2919

2920
	err = rewrite_section_headers(info, flags);
2921 2922
	if (err)
		return ERR_PTR(err);
L
Linus Torvalds 已提交
2923

2924 2925
	/* Find internal symbols and strings. */
	for (i = 1; i < info->hdr->e_shnum; i++) {
L
Linus Torvalds 已提交
2926 2927 2928
		if (info->sechdrs[i].sh_type == SHT_SYMTAB) {
			info->index.sym = i;
			info->index.str = info->sechdrs[i].sh_link;
2929 2930 2931
			info->strtab = (char *)info->hdr
				+ info->sechdrs[info->index.str].sh_offset;
			break;
L
Linus Torvalds 已提交
2932 2933 2934
		}
	}

2935
	info->index.mod = find_sec(info, ".gnu.linkonce.this_module");
L
Linus Torvalds 已提交
2936
	if (!info->index.mod) {
2937
		pr_warn("No module found in object\n");
L
Linus Torvalds 已提交
2938 2939 2940 2941 2942 2943
		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) {
2944
		pr_warn("%s: module has no symbols (stripped?)\n", mod->name);
L
Linus Torvalds 已提交
2945 2946 2947
		return ERR_PTR(-ENOEXEC);
	}

2948
	info->index.pcpu = find_pcpusec(info);
L
Linus Torvalds 已提交
2949 2950 2951 2952 2953 2954 2955 2956

	/* 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;
}

2957
static int check_modinfo(struct module *mod, struct load_info *info, int flags)
2958
{
2959
	const char *modmagic = get_modinfo(info, "vermagic");
2960 2961
	int err;

2962 2963 2964
	if (flags & MODULE_INIT_IGNORE_VERMAGIC)
		modmagic = NULL;

2965 2966 2967 2968 2969
	/* This is allowed: modprobe --force will invalidate it. */
	if (!modmagic) {
		err = try_to_force_load(mod, "bad vermagic");
		if (err)
			return err;
2970
	} else if (!same_magic(modmagic, vermagic, info->index.vers)) {
2971
		pr_err("%s: version magic '%s' should be '%s'\n",
2972 2973 2974 2975
		       mod->name, modmagic, vermagic);
		return -ENOEXEC;
	}

2976 2977 2978 2979
	if (!get_modinfo(info, "intree")) {
		if (!test_taint(TAINT_OOT_MODULE))
			pr_warn("%s: loading out-of-tree module taints kernel.\n",
				mod->name);
2980
		add_taint_module(mod, TAINT_OOT_MODULE, LOCKDEP_STILL_OK);
2981
	}
2982

2983
	if (get_modinfo(info, "staging")) {
2984
		add_taint_module(mod, TAINT_CRAP, LOCKDEP_STILL_OK);
2985 2986
		pr_warn("%s: module is from the staging directory, the quality "
			"is unknown, you have been warned.\n", mod->name);
2987
	}
2988

2989
	err = check_modinfo_livepatch(mod, info);
2990 2991 2992
	if (err)
		return err;

2993
	/* Set up license info based on the info section */
2994
	set_license(mod, get_modinfo(info, "license"));
2995

2996 2997 2998
	return 0;
}

2999
static int find_module_sections(struct module *mod, struct load_info *info)
3000
{
3001
	mod->kp = section_objs(info, "__param",
3002
			       sizeof(*mod->kp), &mod->num_kp);
3003
	mod->syms = section_objs(info, "__ksymtab",
3004
				 sizeof(*mod->syms), &mod->num_syms);
3005 3006
	mod->crcs = section_addr(info, "__kcrctab");
	mod->gpl_syms = section_objs(info, "__ksymtab_gpl",
3007 3008
				     sizeof(*mod->gpl_syms),
				     &mod->num_gpl_syms);
3009 3010
	mod->gpl_crcs = section_addr(info, "__kcrctab_gpl");
	mod->gpl_future_syms = section_objs(info,
3011 3012 3013
					    "__ksymtab_gpl_future",
					    sizeof(*mod->gpl_future_syms),
					    &mod->num_gpl_future_syms);
3014
	mod->gpl_future_crcs = section_addr(info, "__kcrctab_gpl_future");
3015 3016

#ifdef CONFIG_UNUSED_SYMBOLS
3017
	mod->unused_syms = section_objs(info, "__ksymtab_unused",
3018 3019
					sizeof(*mod->unused_syms),
					&mod->num_unused_syms);
3020 3021
	mod->unused_crcs = section_addr(info, "__kcrctab_unused");
	mod->unused_gpl_syms = section_objs(info, "__ksymtab_unused_gpl",
3022 3023
					    sizeof(*mod->unused_gpl_syms),
					    &mod->num_unused_gpl_syms);
3024
	mod->unused_gpl_crcs = section_addr(info, "__kcrctab_unused_gpl");
3025 3026
#endif
#ifdef CONFIG_CONSTRUCTORS
3027
	mod->ctors = section_objs(info, ".ctors",
3028
				  sizeof(*mod->ctors), &mod->num_ctors);
3029 3030 3031 3032 3033 3034 3035 3036
	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.
		 */
3037
		pr_warn("%s: has both .ctors and .init_array.\n",
3038 3039 3040
		       mod->name);
		return -EINVAL;
	}
3041 3042 3043
#endif

#ifdef CONFIG_TRACEPOINTS
3044 3045 3046
	mod->tracepoints_ptrs = section_objs(info, "__tracepoints_ptrs",
					     sizeof(*mod->tracepoints_ptrs),
					     &mod->num_tracepoints);
3047
#endif
3048 3049 3050 3051 3052
#ifdef HAVE_JUMP_LABEL
	mod->jump_entries = section_objs(info, "__jump_table",
					sizeof(*mod->jump_entries),
					&mod->num_jump_entries);
#endif
3053
#ifdef CONFIG_EVENT_TRACING
3054
	mod->trace_events = section_objs(info, "_ftrace_events",
3055 3056
					 sizeof(*mod->trace_events),
					 &mod->num_trace_events);
3057 3058 3059
	mod->trace_enums = section_objs(info, "_ftrace_enum_map",
					sizeof(*mod->trace_enums),
					&mod->num_trace_enums);
3060
#endif
3061 3062 3063 3064 3065
#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
3066 3067
#ifdef CONFIG_FTRACE_MCOUNT_RECORD
	/* sechdrs[0].sh_size is always zero */
3068
	mod->ftrace_callsites = section_objs(info, "__mcount_loc",
3069 3070 3071
					     sizeof(*mod->ftrace_callsites),
					     &mod->num_ftrace_callsites);
#endif
3072

3073 3074 3075
	mod->extable = section_objs(info, "__ex_table",
				    sizeof(*mod->extable), &mod->num_exentries);

3076
	if (section_addr(info, "__obsparm"))
3077
		pr_warn("%s: Ignoring obsolete parameters\n", mod->name);
3078 3079 3080

	info->debug = section_objs(info, "__verbose",
				   sizeof(*info->debug), &info->num_debug);
3081 3082

	return 0;
3083 3084
}

3085
static int move_module(struct module *mod, struct load_info *info)
3086 3087 3088 3089 3090
{
	int i;
	void *ptr;

	/* Do the allocs. */
3091
	ptr = module_alloc(mod->core_layout.size);
3092 3093 3094 3095 3096 3097 3098
	/*
	 * 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)
3099
		return -ENOMEM;
3100

3101 3102
	memset(ptr, 0, mod->core_layout.size);
	mod->core_layout.base = ptr;
3103

3104 3105
	if (mod->init_layout.size) {
		ptr = module_alloc(mod->init_layout.size);
3106 3107 3108 3109 3110 3111 3112 3113
		/*
		 * 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) {
3114
			module_memfree(mod->core_layout.base);
3115 3116
			return -ENOMEM;
		}
3117 3118
		memset(ptr, 0, mod->init_layout.size);
		mod->init_layout.base = ptr;
3119
	} else
3120
		mod->init_layout.base = NULL;
3121 3122

	/* Transfer each section which specifies SHF_ALLOC */
3123
	pr_debug("final section addresses:\n");
3124
	for (i = 0; i < info->hdr->e_shnum; i++) {
3125
		void *dest;
3126
		Elf_Shdr *shdr = &info->sechdrs[i];
3127

3128
		if (!(shdr->sh_flags & SHF_ALLOC))
3129 3130
			continue;

3131
		if (shdr->sh_entsize & INIT_OFFSET_MASK)
3132
			dest = mod->init_layout.base
3133
				+ (shdr->sh_entsize & ~INIT_OFFSET_MASK);
3134
		else
3135
			dest = mod->core_layout.base + shdr->sh_entsize;
3136

3137 3138
		if (shdr->sh_type != SHT_NOBITS)
			memcpy(dest, (void *)shdr->sh_addr, shdr->sh_size);
3139
		/* Update sh_addr to point to copy in image. */
3140
		shdr->sh_addr = (unsigned long)dest;
3141 3142
		pr_debug("\t0x%lx %s\n",
			 (long)shdr->sh_addr, info->secstrings + shdr->sh_name);
3143
	}
3144 3145

	return 0;
3146 3147
}

3148
static int check_module_license_and_versions(struct module *mod)
3149
{
3150 3151
	int prev_taint = test_taint(TAINT_PROPRIETARY_MODULE);

3152 3153 3154 3155 3156 3157
	/*
	 * 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)
3158
		add_taint(TAINT_PROPRIETARY_MODULE, LOCKDEP_NOW_UNRELIABLE);
3159 3160 3161

	/* driverloader was caught wrongly pretending to be under GPL */
	if (strcmp(mod->name, "driverloader") == 0)
3162 3163
		add_taint_module(mod, TAINT_PROPRIETARY_MODULE,
				 LOCKDEP_NOW_UNRELIABLE);
3164

3165 3166
	/* lve claims to be GPL but upstream won't provide source */
	if (strcmp(mod->name, "lve") == 0)
3167 3168
		add_taint_module(mod, TAINT_PROPRIETARY_MODULE,
				 LOCKDEP_NOW_UNRELIABLE);
3169

3170 3171 3172
	if (!prev_taint && test_taint(TAINT_PROPRIETARY_MODULE))
		pr_warn("%s: module license taints kernel.\n", mod->name);

3173 3174 3175 3176 3177 3178 3179 3180 3181 3182 3183 3184 3185 3186 3187 3188 3189 3190 3191 3192 3193 3194 3195 3196 3197 3198 3199 3200 3201
#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.
	 */
3202 3203 3204 3205 3206 3207
	if (mod->init_layout.base)
		flush_icache_range((unsigned long)mod->init_layout.base,
				   (unsigned long)mod->init_layout.base
				   + mod->init_layout.size);
	flush_icache_range((unsigned long)mod->core_layout.base,
			   (unsigned long)mod->core_layout.base + mod->core_layout.size);
3208 3209 3210 3211

	set_fs(old_fs);
}

3212 3213 3214 3215 3216 3217 3218 3219
int __weak module_frob_arch_sections(Elf_Ehdr *hdr,
				     Elf_Shdr *sechdrs,
				     char *secstrings,
				     struct module *mod)
{
	return 0;
}

3220 3221 3222 3223 3224 3225 3226 3227 3228 3229 3230 3231 3232 3233 3234 3235 3236 3237 3238 3239 3240
/* module_blacklist is a comma-separated list of module names */
static char *module_blacklist;
static bool blacklisted(char *module_name)
{
	const char *p;
	size_t len;

	if (!module_blacklist)
		return false;

	for (p = module_blacklist; *p; p += len) {
		len = strcspn(p, ",");
		if (strlen(module_name) == len && !memcmp(module_name, p, len))
			return true;
		if (p[len] == ',')
			len++;
	}
	return false;
}
core_param(module_blacklist, module_blacklist, charp, 0400);

3241
static struct module *layout_and_allocate(struct load_info *info, int flags)
L
Linus Torvalds 已提交
3242
{
3243
	/* Module within temporary copy. */
L
Linus Torvalds 已提交
3244
	struct module *mod;
3245
	unsigned int ndx;
3246
	int err;
3247

3248
	mod = setup_load_info(info, flags);
3249 3250
	if (IS_ERR(mod))
		return mod;
L
Linus Torvalds 已提交
3251

3252 3253 3254
	if (blacklisted(mod->name))
		return ERR_PTR(-EPERM);

3255
	err = check_modinfo(mod, info, flags);
3256 3257
	if (err)
		return ERR_PTR(err);
L
Linus Torvalds 已提交
3258 3259

	/* Allow arches to frob section contents and sizes.  */
3260 3261
	err = module_frob_arch_sections(info->hdr, info->sechdrs,
					info->secstrings, mod);
L
Linus Torvalds 已提交
3262
	if (err < 0)
3263
		return ERR_PTR(err);
L
Linus Torvalds 已提交
3264

3265 3266
	/* 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
3267

3268 3269 3270 3271 3272 3273 3274 3275 3276
	/*
	 * Mark ro_after_init section with SHF_RO_AFTER_INIT so that
	 * layout_sections() can put it in the right place.
	 * Note: ro_after_init sections also have SHF_{WRITE,ALLOC} set.
	 */
	ndx = find_sec(info, ".data..ro_after_init");
	if (ndx)
		info->sechdrs[ndx].sh_flags |= SHF_RO_AFTER_INIT;

L
Linus Torvalds 已提交
3277 3278 3279
	/* 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. */
3280 3281
	layout_sections(mod, info);
	layout_symtab(mod, info);
L
Linus Torvalds 已提交
3282

3283
	/* Allocate and move to the final place */
3284
	err = move_module(mod, info);
3285
	if (err)
3286
		return ERR_PTR(err);
3287 3288 3289

	/* Module has been copied to its final place now: return it. */
	mod = (void *)info->sechdrs[info->index.mod].sh_addr;
3290
	kmemleak_load_module(mod, info);
3291 3292 3293 3294 3295 3296 3297
	return mod;
}

/* mod is no longer valid after this! */
static void module_deallocate(struct module *mod, struct load_info *info)
{
	percpu_modfree(mod);
3298
	module_arch_freeing_init(mod);
3299 3300
	module_memfree(mod->init_layout.base);
	module_memfree(mod->core_layout.base);
3301 3302
}

3303 3304 3305 3306 3307 3308 3309
int __weak module_finalize(const Elf_Ehdr *hdr,
			   const Elf_Shdr *sechdrs,
			   struct module *me)
{
	return 0;
}

3310 3311
static int post_relocation(struct module *mod, const struct load_info *info)
{
3312
	/* Sort exception table now relocations are done. */
3313 3314 3315 3316 3317 3318
	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);

3319
	/* Setup kallsyms-specific fields. */
3320 3321 3322 3323 3324 3325
	add_kallsyms(mod, info);

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

3326 3327 3328 3329 3330 3331
/* Is this module of this name done loading?  No locks held. */
static bool finished_loading(const char *name)
{
	struct module *mod;
	bool ret;

3332 3333 3334 3335 3336 3337
	/*
	 * 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();
3338
	mutex_lock(&module_mutex);
3339
	mod = find_module_all(name, strlen(name), true);
3340 3341
	ret = !mod || mod->state == MODULE_STATE_LIVE
		|| mod->state == MODULE_STATE_GOING;
3342 3343 3344 3345 3346
	mutex_unlock(&module_mutex);

	return ret;
}

3347 3348 3349 3350 3351 3352 3353 3354 3355 3356 3357
/* 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
}

3358 3359 3360 3361 3362 3363 3364 3365 3366 3367 3368 3369 3370
/* 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);
}

3371 3372 3373 3374 3375 3376 3377
/*
 * 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)
3378 3379
{
	int ret = 0;
3380 3381 3382 3383 3384 3385 3386
	struct mod_initfree *freeinit;

	freeinit = kmalloc(sizeof(*freeinit), GFP_KERNEL);
	if (!freeinit) {
		ret = -ENOMEM;
		goto fail;
	}
3387
	freeinit->module_init = mod->init_layout.base;
3388

3389 3390 3391 3392 3393 3394
	/*
	 * 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;

3395 3396 3397 3398 3399
	do_mod_ctors(mod);
	/* Start the module */
	if (mod->init != NULL)
		ret = do_one_initcall(mod->init);
	if (ret < 0) {
3400
		goto fail_free_freeinit;
3401 3402
	}
	if (ret > 0) {
3403 3404 3405 3406
		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__);
3407 3408 3409 3410 3411 3412 3413 3414
		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);

3415 3416 3417 3418 3419 3420 3421 3422 3423 3424 3425 3426 3427 3428 3429 3430 3431
	/*
	 * 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
	 */
3432
	if (!mod->async_probe_requested && (current->flags & PF_USED_ASYNC))
3433
		async_synchronize_full();
3434 3435 3436 3437 3438 3439

	mutex_lock(&module_mutex);
	/* Drop initial reference. */
	module_put(mod);
	trim_init_extable(mod);
#ifdef CONFIG_KALLSYMS
3440 3441
	/* Switch to core kallsyms now init is done: kallsyms may be walking! */
	rcu_assign_pointer(mod->kallsyms, &mod->core_kallsyms);
3442
#endif
3443
	module_enable_ro(mod, true);
3444
	mod_tree_remove_init(mod);
3445
	disable_ro_nx(&mod->init_layout);
3446
	module_arch_freeing_init(mod);
3447 3448 3449
	mod->init_layout.base = NULL;
	mod->init_layout.size = 0;
	mod->init_layout.ro_size = 0;
3450
	mod->init_layout.ro_after_init_size = 0;
3451
	mod->init_layout.text_size = 0;
3452 3453
	/*
	 * We want to free module_init, but be aware that kallsyms may be
3454 3455 3456
	 * 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.
3457
	 */
3458
	call_rcu_sched(&freeinit->rcu, do_free_init);
3459 3460 3461 3462
	mutex_unlock(&module_mutex);
	wake_up_all(&module_wq);

	return 0;
3463 3464 3465 3466 3467 3468 3469 3470 3471 3472

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);
3473
	klp_module_going(mod);
3474
	ftrace_release_mod(mod);
3475 3476 3477
	free_module(mod);
	wake_up_all(&module_wq);
	return ret;
3478 3479 3480 3481 3482 3483 3484 3485 3486 3487
}

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

	return 0;
}

3488 3489 3490 3491 3492 3493 3494 3495 3496 3497 3498 3499 3500 3501
/*
 * 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);
3502 3503
	old = find_module_all(mod->name, strlen(mod->name), true);
	if (old != NULL) {
3504 3505 3506 3507
		if (old->state == MODULE_STATE_COMING
		    || old->state == MODULE_STATE_UNFORMED) {
			/* Wait in case it fails to load. */
			mutex_unlock(&module_mutex);
3508 3509
			err = wait_event_interruptible(module_wq,
					       finished_loading(mod->name));
3510 3511 3512 3513 3514 3515 3516
			if (err)
				goto out_unlocked;
			goto again;
		}
		err = -EEXIST;
		goto out;
	}
3517
	mod_update_bounds(mod);
3518
	list_add_rcu(&mod->list, &modules);
3519
	mod_tree_insert(mod);
3520 3521 3522 3523 3524 3525 3526 3527 3528 3529 3530 3531 3532 3533 3534 3535 3536 3537 3538 3539 3540 3541
	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);

3542
	module_enable_ro(mod, false);
3543
	module_enable_nx(mod);
3544

3545 3546 3547
	/* Mark state as coming so strong_try_module_get() ignores us,
	 * but kallsyms etc. can see us. */
	mod->state = MODULE_STATE_COMING;
3548 3549 3550
	mutex_unlock(&module_mutex);

	return 0;
3551 3552 3553 3554 3555 3556

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

3557 3558
static int prepare_coming_module(struct module *mod)
{
3559 3560
	int err;

3561
	ftrace_module_enable(mod);
3562 3563 3564 3565
	err = klp_module_coming(mod);
	if (err)
		return err;

3566 3567 3568 3569 3570
	blocking_notifier_call_chain(&module_notify_list,
				     MODULE_STATE_COMING, mod);
	return 0;
}

3571 3572
static int unknown_module_param_cb(char *param, char *val, const char *modname,
				   void *arg)
3573
{
3574 3575 3576 3577 3578 3579 3580 3581
	struct module *mod = arg;
	int ret;

	if (strcmp(param, "async_probe") == 0) {
		mod->async_probe_requested = true;
		return 0;
	}

3582
	/* Check for magic 'dyndbg' arg */
3583
	ret = ddebug_dyndbg_module_param_cb(param, val, modname);
3584 3585
	if (ret != 0)
		pr_warn("%s: unknown parameter '%s' ignored\n", modname, param);
3586 3587 3588
	return 0;
}

3589 3590
/* Allocate and load the module: note that size of section 0 is always
   zero, and we rely on this for optional sections. */
3591 3592
static int load_module(struct load_info *info, const char __user *uargs,
		       int flags)
3593
{
3594
	struct module *mod;
3595
	long err;
3596
	char *after_dashes;
3597

3598
	err = module_sig_check(info, flags);
3599 3600
	if (err)
		goto free_copy;
3601

3602
	err = elf_header_check(info);
3603
	if (err)
3604
		goto free_copy;
3605 3606

	/* Figure out module layout, and allocate all the memory. */
3607
	mod = layout_and_allocate(info, flags);
3608 3609
	if (IS_ERR(mod)) {
		err = PTR_ERR(mod);
3610
		goto free_copy;
L
Linus Torvalds 已提交
3611 3612
	}

3613 3614
	audit_log_kern_module(mod->name);

3615 3616 3617
	/* Reserve our place in the list. */
	err = add_unformed_module(mod);
	if (err)
3618 3619
		goto free_module;

3620
#ifdef CONFIG_MODULE_SIG
3621
	mod->sig_ok = info->sig_ok;
3622
	if (!mod->sig_ok) {
3623
		pr_notice_once("%s: module verification failed: signature "
3624
			       "and/or required key missing - tainting "
3625
			       "kernel\n", mod->name);
3626
		add_taint_module(mod, TAINT_UNSIGNED_MODULE, LOCKDEP_STILL_OK);
3627
	}
3628 3629
#endif

3630
	/* To avoid stressing percpu allocator, do this once we're unique. */
3631
	err = percpu_modalloc(mod, info);
3632 3633 3634
	if (err)
		goto unlink_mod;

3635
	/* Now module is in final location, initialize linked lists, etc. */
3636 3637
	err = module_unload_init(mod);
	if (err)
3638
		goto unlink_mod;
L
Linus Torvalds 已提交
3639

3640
	init_param_lock(mod);
3641

3642 3643
	/* Now we've got everything in the final locations, we can
	 * find optional sections. */
3644 3645 3646
	err = find_module_sections(mod, info);
	if (err)
		goto free_unload;
3647

3648
	err = check_module_license_and_versions(mod);
3649 3650
	if (err)
		goto free_unload;
3651

3652
	/* Set up MODINFO_ATTR fields */
3653
	setup_modinfo(mod, info);
3654

L
Linus Torvalds 已提交
3655
	/* Fix up syms, so that st_value is a pointer to location. */
3656
	err = simplify_symbols(mod, info);
L
Linus Torvalds 已提交
3657
	if (err < 0)
3658
		goto free_modinfo;
L
Linus Torvalds 已提交
3659

3660
	err = apply_relocations(mod, info);
3661
	if (err < 0)
3662
		goto free_modinfo;
L
Linus Torvalds 已提交
3663

3664
	err = post_relocation(mod, info);
L
Linus Torvalds 已提交
3665
	if (err < 0)
3666
		goto free_modinfo;
L
Linus Torvalds 已提交
3667

3668
	flush_module_icache(mod);
3669

3670 3671 3672 3673 3674 3675
	/* 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;
	}
3676

3677
	dynamic_debug_setup(info->debug, info->num_debug);
3678

3679 3680 3681
	/* Ftrace init must be called in the MODULE_STATE_UNFORMED state */
	ftrace_module_init(mod);

3682 3683 3684
	/* Finally it's fully formed, ready to start executing. */
	err = complete_formation(mod, info);
	if (err)
3685
		goto ddebug_cleanup;
3686

3687 3688 3689 3690
	err = prepare_coming_module(mod);
	if (err)
		goto bug_cleanup;

3691
	/* Module is ready to execute: parsing args may do that. */
3692
	after_dashes = parse_args(mod->name, mod->args, mod->kp, mod->num_kp,
3693
				  -32768, 32767, mod,
3694
				  unknown_module_param_cb);
3695 3696
	if (IS_ERR(after_dashes)) {
		err = PTR_ERR(after_dashes);
3697
		goto coming_cleanup;
3698 3699 3700 3701
	} else if (after_dashes) {
		pr_warn("%s: parameters '%s' after `--' ignored\n",
		       mod->name, after_dashes);
	}
L
Linus Torvalds 已提交
3702

3703
	/* Link in to sysfs. */
3704
	err = mod_sysfs_setup(mod, info, mod->kp, mod->num_kp);
L
Linus Torvalds 已提交
3705
	if (err < 0)
3706
		goto coming_cleanup;
3707

3708 3709 3710 3711 3712 3713
	if (is_livepatch_module(mod)) {
		err = copy_module_elf(mod, info);
		if (err < 0)
			goto sysfs_cleanup;
	}

3714
	/* Get rid of temporary copy. */
3715
	free_copy(info);
L
Linus Torvalds 已提交
3716 3717

	/* Done! */
3718
	trace_module_load(mod);
3719 3720

	return do_init_module(mod);
L
Linus Torvalds 已提交
3721

3722 3723
 sysfs_cleanup:
	mod_sysfs_teardown(mod);
3724
 coming_cleanup:
3725
	mod->state = MODULE_STATE_GOING;
3726 3727
	blocking_notifier_call_chain(&module_notify_list,
				     MODULE_STATE_GOING, mod);
3728
	klp_module_going(mod);
3729 3730
 bug_cleanup:
	/* module_bug_cleanup needs module_mutex protection */
3731
	mutex_lock(&module_mutex);
3732
	module_bug_cleanup(mod);
3733
	mutex_unlock(&module_mutex);
3734 3735

	/* we can't deallocate the module until we clear memory protection */
3736 3737
	module_disable_ro(mod);
	module_disable_nx(mod);
3738

3739
 ddebug_cleanup:
3740
	dynamic_debug_remove(info->debug);
3741
	synchronize_sched();
3742 3743
	kfree(mod->args);
 free_arch_cleanup:
L
Linus Torvalds 已提交
3744
	module_arch_cleanup(mod);
3745
 free_modinfo:
3746
	free_modinfo(mod);
3747
 free_unload:
L
Linus Torvalds 已提交
3748
	module_unload_free(mod);
3749 3750 3751 3752
 unlink_mod:
	mutex_lock(&module_mutex);
	/* Unlink carefully: kallsyms could be walking list. */
	list_del_rcu(&mod->list);
3753
	mod_tree_remove(mod);
3754
	wake_up_all(&module_wq);
3755 3756
	/* Wait for RCU-sched synchronizing before releasing mod->list. */
	synchronize_sched();
3757
	mutex_unlock(&module_mutex);
3758
 free_module:
3759 3760 3761 3762 3763 3764
	/*
	 * Ftrace needs to clean up what it initialized.
	 * This does nothing if ftrace_module_init() wasn't called,
	 * but it must be called outside of module_mutex.
	 */
	ftrace_release_mod(mod);
3765
	/* Free lock-classes; relies on the preceding sync_rcu() */
3766
	lockdep_free_key_range(mod->core_layout.base, mod->core_layout.size);
3767

3768
	module_deallocate(mod, info);
3769
 free_copy:
3770 3771
	free_copy(info);
	return err;
3772 3773
}

3774 3775
SYSCALL_DEFINE3(init_module, void __user *, umod,
		unsigned long, len, const char __user *, uargs)
L
Linus Torvalds 已提交
3776
{
3777 3778
	int err;
	struct load_info info = { };
L
Linus Torvalds 已提交
3779

3780 3781 3782
	err = may_init_module();
	if (err)
		return err;
L
Linus Torvalds 已提交
3783

3784 3785
	pr_debug("init_module: umod=%p, len=%lu, uargs=%p\n",
	       umod, len, uargs);
L
Linus Torvalds 已提交
3786

3787 3788 3789
	err = copy_module_from_user(umod, len, &info);
	if (err)
		return err;
L
Linus Torvalds 已提交
3790

3791
	return load_module(&info, uargs, 0);
3792
}
3793

3794
SYSCALL_DEFINE3(finit_module, int, fd, const char __user *, uargs, int, flags)
3795 3796
{
	struct load_info info = { };
3797 3798 3799
	loff_t size;
	void *hdr;
	int err;
3800

3801 3802 3803
	err = may_init_module();
	if (err)
		return err;
L
Linus Torvalds 已提交
3804

3805
	pr_debug("finit_module: fd=%d, uargs=%p, flags=%i\n", fd, uargs, flags);
3806

3807 3808 3809
	if (flags & ~(MODULE_INIT_IGNORE_MODVERSIONS
		      |MODULE_INIT_IGNORE_VERMAGIC))
		return -EINVAL;
3810

3811 3812
	err = kernel_read_file_from_fd(fd, &hdr, &size, INT_MAX,
				       READING_MODULE);
3813 3814
	if (err)
		return err;
3815 3816
	info.hdr = hdr;
	info.len = size;
L
Linus Torvalds 已提交
3817

3818
	return load_module(&info, uargs, flags);
L
Linus Torvalds 已提交
3819 3820 3821 3822 3823 3824 3825 3826 3827 3828 3829 3830 3831 3832
}

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 $d.
 */
static inline int is_arm_mapping_symbol(const char *str)
{
3833 3834
	if (str[0] == '.' && str[1] == 'L')
		return true;
3835
	return str[0] == '$' && strchr("axtd", str[1])
L
Linus Torvalds 已提交
3836 3837 3838
	       && (str[2] == '\0' || str[2] == '.');
}

3839
static const char *symname(struct mod_kallsyms *kallsyms, unsigned int symnum)
3840
{
3841
	return kallsyms->strtab + kallsyms->symtab[symnum].st_name;
3842 3843
}

L
Linus Torvalds 已提交
3844 3845 3846 3847 3848 3849 3850
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;
3851
	struct mod_kallsyms *kallsyms = rcu_dereference_sched(mod->kallsyms);
L
Linus Torvalds 已提交
3852 3853

	/* At worse, next value is at end of module */
3854
	if (within_module_init(addr, mod))
3855
		nextval = (unsigned long)mod->init_layout.base+mod->init_layout.text_size;
3856
	else
3857
		nextval = (unsigned long)mod->core_layout.base+mod->core_layout.text_size;
L
Linus Torvalds 已提交
3858

L
Lucas De Marchi 已提交
3859
	/* Scan for closest preceding symbol, and next symbol. (ELF
3860
	   starts real symbols at 1). */
3861 3862
	for (i = 1; i < kallsyms->num_symtab; i++) {
		if (kallsyms->symtab[i].st_shndx == SHN_UNDEF)
L
Linus Torvalds 已提交
3863 3864 3865 3866
			continue;

		/* We ignore unnamed symbols: they're uninformative
		 * and inserted at a whim. */
3867 3868
		if (*symname(kallsyms, i) == '\0'
		    || is_arm_mapping_symbol(symname(kallsyms, i)))
3869 3870
			continue;

3871 3872
		if (kallsyms->symtab[i].st_value <= addr
		    && kallsyms->symtab[i].st_value > kallsyms->symtab[best].st_value)
L
Linus Torvalds 已提交
3873
			best = i;
3874 3875 3876
		if (kallsyms->symtab[i].st_value > addr
		    && kallsyms->symtab[i].st_value < nextval)
			nextval = kallsyms->symtab[i].st_value;
L
Linus Torvalds 已提交
3877 3878 3879 3880 3881
	}

	if (!best)
		return NULL;

3882
	if (size)
3883
		*size = nextval - kallsyms->symtab[best].st_value;
3884
	if (offset)
3885 3886
		*offset = addr - kallsyms->symtab[best].st_value;
	return symname(kallsyms, best);
L
Linus Torvalds 已提交
3887 3888
}

3889 3890
/* For kallsyms to ask for address resolution.  NULL means not found.  Careful
 * not to lock to avoid deadlock on oopses, simply disable preemption. */
3891
const char *module_address_lookup(unsigned long addr,
3892 3893 3894 3895
			    unsigned long *size,
			    unsigned long *offset,
			    char **modname,
			    char *namebuf)
L
Linus Torvalds 已提交
3896
{
3897
	const char *ret = NULL;
3898
	struct module *mod;
L
Linus Torvalds 已提交
3899

3900
	preempt_disable();
3901 3902 3903 3904 3905
	mod = __module_address(addr);
	if (mod) {
		if (modname)
			*modname = mod->name;
		ret = get_ksymbol(mod, addr, size, offset);
L
Linus Torvalds 已提交
3906
	}
3907 3908 3909 3910 3911
	/* Make a copy in here where it's safe */
	if (ret) {
		strncpy(namebuf, ret, KSYM_NAME_LEN - 1);
		ret = namebuf;
	}
3912
	preempt_enable();
3913

3914
	return ret;
L
Linus Torvalds 已提交
3915 3916
}

3917 3918 3919 3920
int lookup_module_symbol_name(unsigned long addr, char *symname)
{
	struct module *mod;

3921
	preempt_disable();
3922
	list_for_each_entry_rcu(mod, &modules, list) {
3923 3924
		if (mod->state == MODULE_STATE_UNFORMED)
			continue;
3925
		if (within_module(addr, mod)) {
3926 3927 3928 3929 3930
			const char *sym;

			sym = get_ksymbol(mod, addr, NULL, NULL);
			if (!sym)
				goto out;
3931
			strlcpy(symname, sym, KSYM_NAME_LEN);
3932
			preempt_enable();
3933 3934 3935 3936
			return 0;
		}
	}
out:
3937
	preempt_enable();
3938 3939 3940
	return -ERANGE;
}

3941 3942 3943 3944 3945
int lookup_module_symbol_attrs(unsigned long addr, unsigned long *size,
			unsigned long *offset, char *modname, char *name)
{
	struct module *mod;

3946
	preempt_disable();
3947
	list_for_each_entry_rcu(mod, &modules, list) {
3948 3949
		if (mod->state == MODULE_STATE_UNFORMED)
			continue;
3950
		if (within_module(addr, mod)) {
3951 3952 3953 3954 3955 3956
			const char *sym;

			sym = get_ksymbol(mod, addr, size, offset);
			if (!sym)
				goto out;
			if (modname)
3957
				strlcpy(modname, mod->name, MODULE_NAME_LEN);
3958
			if (name)
3959
				strlcpy(name, sym, KSYM_NAME_LEN);
3960
			preempt_enable();
3961 3962 3963 3964
			return 0;
		}
	}
out:
3965
	preempt_enable();
3966 3967 3968
	return -ERANGE;
}

3969 3970
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
3971 3972 3973
{
	struct module *mod;

3974
	preempt_disable();
3975
	list_for_each_entry_rcu(mod, &modules, list) {
3976 3977
		struct mod_kallsyms *kallsyms;

3978 3979
		if (mod->state == MODULE_STATE_UNFORMED)
			continue;
3980 3981 3982 3983 3984
		kallsyms = rcu_dereference_sched(mod->kallsyms);
		if (symnum < kallsyms->num_symtab) {
			*value = kallsyms->symtab[symnum].st_value;
			*type = kallsyms->symtab[symnum].st_info;
			strlcpy(name, symname(kallsyms, symnum), KSYM_NAME_LEN);
3985
			strlcpy(module_name, mod->name, MODULE_NAME_LEN);
3986
			*exported = is_exported(name, *value, mod);
3987
			preempt_enable();
3988
			return 0;
L
Linus Torvalds 已提交
3989
		}
3990
		symnum -= kallsyms->num_symtab;
L
Linus Torvalds 已提交
3991
	}
3992
	preempt_enable();
3993
	return -ERANGE;
L
Linus Torvalds 已提交
3994 3995 3996 3997 3998
}

static unsigned long mod_find_symname(struct module *mod, const char *name)
{
	unsigned int i;
3999
	struct mod_kallsyms *kallsyms = rcu_dereference_sched(mod->kallsyms);
L
Linus Torvalds 已提交
4000

4001 4002 4003 4004
	for (i = 0; i < kallsyms->num_symtab; i++)
		if (strcmp(name, symname(kallsyms, i)) == 0 &&
		    kallsyms->symtab[i].st_info != 'U')
			return kallsyms->symtab[i].st_value;
L
Linus Torvalds 已提交
4005 4006 4007 4008 4009 4010 4011 4012 4013 4014 4015
	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. */
4016
	preempt_disable();
L
Linus Torvalds 已提交
4017
	if ((colon = strchr(name, ':')) != NULL) {
4018
		if ((mod = find_module_all(name, colon - name, false)) != NULL)
L
Linus Torvalds 已提交
4019 4020
			ret = mod_find_symname(mod, colon+1);
	} else {
4021 4022 4023
		list_for_each_entry_rcu(mod, &modules, list) {
			if (mod->state == MODULE_STATE_UNFORMED)
				continue;
L
Linus Torvalds 已提交
4024 4025
			if ((ret = mod_find_symname(mod, name)) != 0)
				break;
4026
		}
L
Linus Torvalds 已提交
4027
	}
4028
	preempt_enable();
L
Linus Torvalds 已提交
4029 4030
	return ret;
}
4031 4032 4033 4034 4035 4036 4037 4038 4039

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;

4040 4041
	module_assert_mutex();

4042
	list_for_each_entry(mod, &modules, list) {
4043 4044 4045
		/* We hold module_mutex: no need for rcu_dereference_sched */
		struct mod_kallsyms *kallsyms = mod->kallsyms;

4046 4047
		if (mod->state == MODULE_STATE_UNFORMED)
			continue;
4048 4049 4050
		for (i = 0; i < kallsyms->num_symtab; i++) {
			ret = fn(data, symname(kallsyms, i),
				 mod, kallsyms->symtab[i].st_value);
4051 4052 4053 4054 4055 4056
			if (ret != 0)
				return ret;
		}
	}
	return 0;
}
L
Linus Torvalds 已提交
4057 4058
#endif /* CONFIG_KALLSYMS */

4059 4060 4061 4062
/* Maximum number of characters written by module_flags() */
#define MODULE_FLAGS_BUF_SIZE (TAINT_FLAGS_COUNT + 4)

/* Keep in sync with MODULE_FLAGS_BUF_SIZE !!! */
4063
static char *module_flags(struct module *mod, char *buf)
4064 4065 4066
{
	int bx = 0;

4067
	BUG_ON(mod->state == MODULE_STATE_UNFORMED);
4068 4069 4070
	if (mod->taints ||
	    mod->state == MODULE_STATE_GOING ||
	    mod->state == MODULE_STATE_COMING) {
4071
		buf[bx++] = '(';
4072
		bx += module_flags_taint(mod, buf + bx);
4073 4074 4075 4076 4077 4078
		/* 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++] = '+';
4079 4080 4081 4082 4083 4084 4085
		buf[bx++] = ')';
	}
	buf[bx] = '\0';

	return buf;
}

4086 4087 4088 4089 4090 4091 4092 4093 4094 4095 4096 4097 4098 4099 4100 4101 4102 4103
#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
4104 4105 4106
static int m_show(struct seq_file *m, void *p)
{
	struct module *mod = list_entry(p, struct module, list);
4107
	char buf[MODULE_FLAGS_BUF_SIZE];
4108

4109 4110 4111 4112
	/* We always ignore unformed modules. */
	if (mod->state == MODULE_STATE_UNFORMED)
		return 0;

4113
	seq_printf(m, "%s %u",
4114
		   mod->name, mod->init_layout.size + mod->core_layout.size);
L
Linus Torvalds 已提交
4115 4116 4117 4118
	print_unload_info(m, mod);

	/* Informative for users. */
	seq_printf(m, " %s",
4119 4120
		   mod->state == MODULE_STATE_GOING ? "Unloading" :
		   mod->state == MODULE_STATE_COMING ? "Loading" :
L
Linus Torvalds 已提交
4121 4122
		   "Live");
	/* Used by oprofile and other similar tools. */
4123
	seq_printf(m, " 0x%pK", mod->core_layout.base);
L
Linus Torvalds 已提交
4124

4125 4126
	/* Taints info */
	if (mod->taints)
4127
		seq_printf(m, " %s", module_flags(mod, buf));
4128

4129
	seq_puts(m, "\n");
L
Linus Torvalds 已提交
4130 4131 4132 4133 4134 4135 4136 4137
	return 0;
}

/* Format: modulename size refcount deps address

   Where refcount is a number or -, and deps is a comma-separated list
   of depends or -.
*/
4138
static const struct seq_operations modules_op = {
L
Linus Torvalds 已提交
4139 4140 4141 4142 4143 4144
	.start	= m_start,
	.next	= m_next,
	.stop	= m_stop,
	.show	= m_show
};

4145 4146 4147 4148 4149 4150 4151 4152 4153 4154 4155 4156 4157 4158 4159 4160 4161 4162 4163 4164
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
4165 4166 4167 4168 4169 4170
/* 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;

4171
	preempt_disable();
4172
	list_for_each_entry_rcu(mod, &modules, list) {
4173 4174
		if (mod->state == MODULE_STATE_UNFORMED)
			continue;
L
Linus Torvalds 已提交
4175 4176
		if (mod->num_exentries == 0)
			continue;
4177

L
Linus Torvalds 已提交
4178 4179 4180 4181 4182 4183
		e = search_extable(mod->extable,
				   mod->extable + mod->num_exentries - 1,
				   addr);
		if (e)
			break;
	}
4184
	preempt_enable();
L
Linus Torvalds 已提交
4185 4186

	/* Now, if we found one, we are running inside it now, hence
4187
	   we cannot unload the module, hence no refcnt needed. */
L
Linus Torvalds 已提交
4188 4189 4190
	return e;
}

4191
/*
R
Rusty Russell 已提交
4192 4193 4194 4195 4196
 * 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).
4197
 */
R
Rusty Russell 已提交
4198
bool is_module_address(unsigned long addr)
4199
{
R
Rusty Russell 已提交
4200
	bool ret;
4201

4202
	preempt_disable();
R
Rusty Russell 已提交
4203
	ret = __module_address(addr) != NULL;
4204
	preempt_enable();
4205

R
Rusty Russell 已提交
4206
	return ret;
4207 4208
}

R
Rusty Russell 已提交
4209 4210 4211 4212 4213 4214 4215
/*
 * __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.
 */
4216
struct module *__module_address(unsigned long addr)
L
Linus Torvalds 已提交
4217 4218 4219
{
	struct module *mod;

4220 4221 4222
	if (addr < module_addr_min || addr > module_addr_max)
		return NULL;

4223 4224
	module_assert_mutex_or_preempt();

4225
	mod = mod_find(addr);
4226 4227
	if (mod) {
		BUG_ON(!within_module(addr, mod));
4228
		if (mod->state == MODULE_STATE_UNFORMED)
4229
			mod = NULL;
4230
	}
4231
	return mod;
L
Linus Torvalds 已提交
4232
}
4233
EXPORT_SYMBOL_GPL(__module_address);
L
Linus Torvalds 已提交
4234

R
Rusty Russell 已提交
4235 4236 4237 4238 4239 4240 4241 4242 4243 4244 4245 4246 4247 4248 4249 4250 4251 4252 4253 4254 4255 4256 4257 4258 4259 4260 4261 4262 4263 4264 4265
/*
 * 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. */
4266 4267
		if (!within(addr, mod->init_layout.base, mod->init_layout.text_size)
		    && !within(addr, mod->core_layout.base, mod->core_layout.text_size))
R
Rusty Russell 已提交
4268 4269 4270 4271
			mod = NULL;
	}
	return mod;
}
4272
EXPORT_SYMBOL_GPL(__module_text_address);
R
Rusty Russell 已提交
4273

L
Linus Torvalds 已提交
4274 4275 4276 4277
/* Don't grab lock, we're oopsing. */
void print_modules(void)
{
	struct module *mod;
4278
	char buf[MODULE_FLAGS_BUF_SIZE];
L
Linus Torvalds 已提交
4279

4280
	printk(KERN_DEFAULT "Modules linked in:");
4281 4282
	/* Most callers should already have preempt disabled, but make sure */
	preempt_disable();
4283 4284 4285
	list_for_each_entry_rcu(mod, &modules, list) {
		if (mod->state == MODULE_STATE_UNFORMED)
			continue;
J
Jiri Slaby 已提交
4286
		pr_cont(" %s%s", mod->name, module_flags(mod, buf));
4287
	}
4288
	preempt_enable();
4289
	if (last_unloaded_module[0])
J
Jiri Slaby 已提交
4290 4291
		pr_cont(" [last unloaded: %s]", last_unloaded_module);
	pr_cont("\n");
L
Linus Torvalds 已提交
4292 4293 4294
}

#ifdef CONFIG_MODVERSIONS
4295 4296 4297 4298 4299 4300
/* 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,
4301
		   struct tracepoint * const *tp)
4302 4303 4304
{
}
EXPORT_SYMBOL(module_layout);
L
Linus Torvalds 已提交
4305
#endif
反馈
建议
客服 返回
顶部