module.c 102.5 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>
L
Linus Torvalds 已提交
20
#include <linux/moduleloader.h>
21
#include <linux/ftrace_event.h>
L
Linus Torvalds 已提交
22
#include <linux/init.h>
23
#include <linux/kallsyms.h>
24
#include <linux/file.h>
25
#include <linux/fs.h>
R
Roland McGrath 已提交
26
#include <linux/sysfs.h>
27
#include <linux/kernel.h>
L
Linus Torvalds 已提交
28 29 30
#include <linux/slab.h>
#include <linux/vmalloc.h>
#include <linux/elf.h>
31
#include <linux/proc_fs.h>
32
#include <linux/security.h>
L
Linus Torvalds 已提交
33 34 35 36
#include <linux/seq_file.h>
#include <linux/syscalls.h>
#include <linux/fcntl.h>
#include <linux/rcupdate.h>
37
#include <linux/capability.h>
L
Linus Torvalds 已提交
38 39 40 41 42 43
#include <linux/cpu.h>
#include <linux/moduleparam.h>
#include <linux/errno.h>
#include <linux/err.h>
#include <linux/vermagic.h>
#include <linux/notifier.h>
A
Al Viro 已提交
44
#include <linux/sched.h>
L
Linus Torvalds 已提交
45
#include <linux/device.h>
46
#include <linux/string.h>
A
Arjan van de Ven 已提交
47
#include <linux/mutex.h>
48
#include <linux/rculist.h>
L
Linus Torvalds 已提交
49 50
#include <asm/uaccess.h>
#include <asm/cacheflush.h>
51
#include <asm/mmu_context.h>
52
#include <linux/license.h>
53
#include <asm/sections.h>
M
Mathieu Desnoyers 已提交
54
#include <linux/tracepoint.h>
55
#include <linux/ftrace.h>
56
#include <linux/async.h>
57
#include <linux/percpu.h>
C
Catalin Marinas 已提交
58
#include <linux/kmemleak.h>
59
#include <linux/jump_label.h>
60
#include <linux/pfn.h>
61
#include <linux/bsearch.h>
62
#include <uapi/linux/module.h>
R
Rusty Russell 已提交
63
#include "module-internal.h"
L
Linus Torvalds 已提交
64

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

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

72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91
/*
 * Modules' sections will be aligned on page boundaries
 * to ensure complete separation of code and data, but
 * only when CONFIG_DEBUG_SET_MODULE_RONX=y
 */
#ifdef CONFIG_DEBUG_SET_MODULE_RONX
# define debug_align(X) ALIGN(X, PAGE_SIZE)
#else
# define debug_align(X) (X)
#endif

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

L
Linus Torvalds 已提交
92 93 94
/* If this is set, the section belongs in the init part of the module */
#define INIT_OFFSET_MASK (1UL << (BITS_PER_LONG-1))

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

105 106
#ifdef CONFIG_MODULES_TREE_LOOKUP

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

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

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

	return (unsigned long)mod->module_core;
}

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

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

	return (unsigned long)mod->core_size;
}

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

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

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

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

	return 0;
}

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

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

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

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

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

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

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

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

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

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

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

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

231 232
#else /* MODULES_TREE_LOOKUP */

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

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

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

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

	return NULL;
}

#endif /* MODULES_TREE_LOOKUP */

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

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

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

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

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

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

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

R
Rusty Russell 已提交
295 296 297 298 299 300 301 302 303
#ifdef CONFIG_MODULE_SIG
#ifdef CONFIG_MODULE_SIG_FORCE
static bool sig_enforce = true;
#else
static bool sig_enforce = false;

static int param_set_bool_enable_only(const char *val,
				      const struct kernel_param *kp)
{
304 305 306
	int err = 0;
	bool new_value;
	bool orig_value = *(bool *)kp->arg;
R
Rusty Russell 已提交
307 308
	struct kernel_param dummy_kp = *kp;

309
	dummy_kp.arg = &new_value;
R
Rusty Russell 已提交
310 311 312 313 314 315

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

	/* Don't let them unset it once it's set! */
316
	if (!new_value && orig_value)
R
Rusty Russell 已提交
317 318
		return -EROFS;

319 320 321 322
	if (new_value)
		err = param_set_bool(val, kp);

	return err;
R
Rusty Russell 已提交
323 324 325
}

static const struct kernel_param_ops param_ops_bool_enable_only = {
326
	.flags = KERNEL_PARAM_OPS_FL_NOARG,
R
Rusty Russell 已提交
327 328 329 330 331 332 333 334
	.set = param_set_bool_enable_only,
	.get = param_get_bool,
};
#define param_check_bool_enable_only param_check_bool

module_param(sig_enforce, bool_enable_only, 0644);
#endif /* !CONFIG_MODULE_SIG_FORCE */
#endif /* CONFIG_MODULE_SIG */
L
Linus Torvalds 已提交
335

336 337
/* Block module loading/unloading? */
int modules_disabled = 0;
338
core_param(nomodule, modules_disabled, bint, 0);
339

340 341 342
/* Waiting for a module to finish initializing? */
static DECLARE_WAIT_QUEUE_HEAD(module_wq);

343
static BLOCKING_NOTIFIER_HEAD(module_notify_list);
L
Linus Torvalds 已提交
344

345
int register_module_notifier(struct notifier_block *nb)
L
Linus Torvalds 已提交
346
{
347
	return blocking_notifier_chain_register(&module_notify_list, nb);
L
Linus Torvalds 已提交
348 349 350
}
EXPORT_SYMBOL(register_module_notifier);

351
int unregister_module_notifier(struct notifier_block *nb)
L
Linus Torvalds 已提交
352
{
353
	return blocking_notifier_chain_unregister(&module_notify_list, nb);
L
Linus Torvalds 已提交
354 355 356
}
EXPORT_SYMBOL(unregister_module_notifier);

357 358 359 360
struct load_info {
	Elf_Ehdr *hdr;
	unsigned long len;
	Elf_Shdr *sechdrs;
361
	char *secstrings, *strtab;
R
Rusty Russell 已提交
362
	unsigned long symoffs, stroffs;
363 364
	struct _ddebug *debug;
	unsigned int num_debug;
R
Rusty Russell 已提交
365
	bool sig_ok;
366 367 368 369 370
	struct {
		unsigned int sym, str, mod, vers, info, pcpu;
	} index;
};

371 372
/* We require a truly strong try_module_get(): 0 means failure due to
   ongoing or failed initialization etc. */
L
Linus Torvalds 已提交
373 374
static inline int strong_try_module_get(struct module *mod)
{
375
	BUG_ON(mod && mod->state == MODULE_STATE_UNFORMED);
L
Linus Torvalds 已提交
376
	if (mod && mod->state == MODULE_STATE_COMING)
377 378
		return -EBUSY;
	if (try_module_get(mod))
L
Linus Torvalds 已提交
379
		return 0;
380 381
	else
		return -ENOENT;
L
Linus Torvalds 已提交
382 383
}

384 385
static inline void add_taint_module(struct module *mod, unsigned flag,
				    enum lockdep_ok lockdep_ok)
386
{
387
	add_taint(flag, lockdep_ok);
A
Andi Kleen 已提交
388
	mod->taints |= (1U << flag);
389 390
}

391 392 393
/*
 * 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
394 395 396 397 398 399 400
 */
void __module_put_and_exit(struct module *mod, long code)
{
	module_put(mod);
	do_exit(code);
}
EXPORT_SYMBOL(__module_put_and_exit);
D
Daniel Walker 已提交
401

L
Linus Torvalds 已提交
402
/* Find a module section: 0 means not found. */
403
static unsigned int find_sec(const struct load_info *info, const char *name)
L
Linus Torvalds 已提交
404 405 406
{
	unsigned int i;

407 408
	for (i = 1; i < info->hdr->e_shnum; i++) {
		Elf_Shdr *shdr = &info->sechdrs[i];
L
Linus Torvalds 已提交
409
		/* Alloc bit cleared means "ignore it." */
410 411
		if ((shdr->sh_flags & SHF_ALLOC)
		    && strcmp(info->secstrings + shdr->sh_name, name) == 0)
L
Linus Torvalds 已提交
412
			return i;
413
	}
L
Linus Torvalds 已提交
414 415 416
	return 0;
}

R
Rusty Russell 已提交
417
/* Find a module section, or NULL. */
418
static void *section_addr(const struct load_info *info, const char *name)
R
Rusty Russell 已提交
419 420
{
	/* Section 0 has sh_addr 0. */
421
	return (void *)info->sechdrs[find_sec(info, name)].sh_addr;
R
Rusty Russell 已提交
422 423 424
}

/* Find a module section, or NULL.  Fill in number of "objects" in section. */
425
static void *section_objs(const struct load_info *info,
R
Rusty Russell 已提交
426 427 428 429
			  const char *name,
			  size_t object_size,
			  unsigned int *num)
{
430
	unsigned int sec = find_sec(info, name);
R
Rusty Russell 已提交
431 432

	/* Section 0 has sh_addr 0 and sh_size 0. */
433 434
	*num = info->sechdrs[sec].sh_size / object_size;
	return (void *)info->sechdrs[sec].sh_addr;
R
Rusty Russell 已提交
435 436
}

L
Linus Torvalds 已提交
437 438 439 440 441
/* 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[];
442 443
extern const struct kernel_symbol __start___ksymtab_gpl_future[];
extern const struct kernel_symbol __stop___ksymtab_gpl_future[];
L
Linus Torvalds 已提交
444 445
extern const unsigned long __start___kcrctab[];
extern const unsigned long __start___kcrctab_gpl[];
446
extern const unsigned long __start___kcrctab_gpl_future[];
447 448 449 450 451
#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[];
452 453
extern const unsigned long __start___kcrctab_unused[];
extern const unsigned long __start___kcrctab_unused_gpl[];
454
#endif
L
Linus Torvalds 已提交
455 456 457 458

#ifndef CONFIG_MODVERSIONS
#define symversion(base, idx) NULL
#else
A
Andrew Morton 已提交
459
#define symversion(base, idx) ((base != NULL) ? ((base) + (idx)) : NULL)
L
Linus Torvalds 已提交
460 461
#endif

462 463 464 465 466
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,
467
					      void *data),
468
				   void *data)
469
{
470
	unsigned int j;
471

472
	for (j = 0; j < arrsize; j++) {
473 474
		if (fn(&arr[j], owner, data))
			return true;
475
	}
476 477

	return false;
478 479
}

480
/* Returns true as soon as fn returns true, otherwise false. */
481 482 483 484
bool each_symbol_section(bool (*fn)(const struct symsearch *arr,
				    struct module *owner,
				    void *data),
			 void *data)
485 486
{
	struct module *mod;
487
	static const struct symsearch arr[] = {
488
		{ __start___ksymtab, __stop___ksymtab, __start___kcrctab,
489
		  NOT_GPL_ONLY, false },
490
		{ __start___ksymtab_gpl, __stop___ksymtab_gpl,
491 492
		  __start___kcrctab_gpl,
		  GPL_ONLY, false },
493
		{ __start___ksymtab_gpl_future, __stop___ksymtab_gpl_future,
494 495
		  __start___kcrctab_gpl_future,
		  WILL_BE_GPL_ONLY, false },
496
#ifdef CONFIG_UNUSED_SYMBOLS
497
		{ __start___ksymtab_unused, __stop___ksymtab_unused,
498 499
		  __start___kcrctab_unused,
		  NOT_GPL_ONLY, true },
500
		{ __start___ksymtab_unused_gpl, __stop___ksymtab_unused_gpl,
501 502
		  __start___kcrctab_unused_gpl,
		  GPL_ONLY, true },
503
#endif
504
	};
505

506 507
	module_assert_mutex_or_preempt();

508 509
	if (each_symbol_in_section(arr, ARRAY_SIZE(arr), NULL, fn, data))
		return true;
510

511
	list_for_each_entry_rcu(mod, &modules, list) {
512 513
		struct symsearch arr[] = {
			{ mod->syms, mod->syms + mod->num_syms, mod->crcs,
514
			  NOT_GPL_ONLY, false },
515
			{ mod->gpl_syms, mod->gpl_syms + mod->num_gpl_syms,
516 517
			  mod->gpl_crcs,
			  GPL_ONLY, false },
518 519
			{ mod->gpl_future_syms,
			  mod->gpl_future_syms + mod->num_gpl_future_syms,
520 521
			  mod->gpl_future_crcs,
			  WILL_BE_GPL_ONLY, false },
522
#ifdef CONFIG_UNUSED_SYMBOLS
523 524
			{ mod->unused_syms,
			  mod->unused_syms + mod->num_unused_syms,
525 526
			  mod->unused_crcs,
			  NOT_GPL_ONLY, true },
527 528
			{ mod->unused_gpl_syms,
			  mod->unused_gpl_syms + mod->num_unused_gpl_syms,
529 530
			  mod->unused_gpl_crcs,
			  GPL_ONLY, true },
531
#endif
532 533
		};

534 535 536
		if (mod->state == MODULE_STATE_UNFORMED)
			continue;

537 538 539 540 541
		if (each_symbol_in_section(arr, ARRAY_SIZE(arr), mod, fn, data))
			return true;
	}
	return false;
}
542
EXPORT_SYMBOL_GPL(each_symbol_section);
543 544 545 546 547 548 549 550 551 552

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

	/* Output */
	struct module *owner;
	const unsigned long *crc;
553
	const struct kernel_symbol *sym;
554 555
};

556 557 558
static bool check_symbol(const struct symsearch *syms,
				 struct module *owner,
				 unsigned int symnum, void *data)
559 560 561 562 563 564 565
{
	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) {
A
Andrew Morton 已提交
566 567 568
			pr_warn("Symbol %s is being used by a non-GPL module, "
				"which will not be allowed in the future\n",
				fsa->name);
569
		}
L
Linus Torvalds 已提交
570
	}
571

572
#ifdef CONFIG_UNUSED_SYMBOLS
573
	if (syms->unused && fsa->warn) {
A
Andrew Morton 已提交
574 575 576
		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");
577 578 579
		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 "
A
Andrew Morton 已提交
580
			"inclusion.\n");
581
	}
582
#endif
583 584 585

	fsa->owner = owner;
	fsa->crc = symversion(syms->crcs, symnum);
586
	fsa->sym = &syms->start[symnum];
587 588 589
	return true;
}

590 591 592 593 594 595 596 597
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);
}

598 599 600 601 602
static bool find_symbol_in_section(const struct symsearch *syms,
				   struct module *owner,
				   void *data)
{
	struct find_symbol_arg *fsa = data;
603 604 605 606 607 608 609
	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;
610 611 612 613

	return false;
}

614
/* Find a symbol and return it, along with, (optional) crc and
615
 * (optional) module which owns it.  Needs preempt disabled or module_mutex. */
616 617 618 619 620
const struct kernel_symbol *find_symbol(const char *name,
					struct module **owner,
					const unsigned long **crc,
					bool gplok,
					bool warn)
621 622 623 624 625 626 627
{
	struct find_symbol_arg fsa;

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

628
	if (each_symbol_section(find_symbol_in_section, &fsa)) {
629 630 631 632
		if (owner)
			*owner = fsa.owner;
		if (crc)
			*crc = fsa.crc;
633
		return fsa.sym;
634 635
	}

636
	pr_debug("Failed to find symbol %s\n", name);
637
	return NULL;
L
Linus Torvalds 已提交
638
}
639
EXPORT_SYMBOL_GPL(find_symbol);
L
Linus Torvalds 已提交
640 641

/* Search for module by name: must hold module_mutex. */
642
static struct module *find_module_all(const char *name, size_t len,
643
				      bool even_unformed)
L
Linus Torvalds 已提交
644 645 646
{
	struct module *mod;

647 648
	module_assert_mutex();

L
Linus Torvalds 已提交
649
	list_for_each_entry(mod, &modules, list) {
650 651
		if (!even_unformed && mod->state == MODULE_STATE_UNFORMED)
			continue;
652
		if (strlen(mod->name) == len && !memcmp(mod->name, name, len))
L
Linus Torvalds 已提交
653 654 655 656
			return mod;
	}
	return NULL;
}
657 658 659

struct module *find_module(const char *name)
{
660
	return find_module_all(name, strlen(name), false);
661
}
662
EXPORT_SYMBOL_GPL(find_module);
L
Linus Torvalds 已提交
663 664

#ifdef CONFIG_SMP
665

666
static inline void __percpu *mod_percpu(struct module *mod)
667
{
668 669
	return mod->percpu;
}
670

R
Rusty Russell 已提交
671
static int percpu_modalloc(struct module *mod, struct load_info *info)
672
{
R
Rusty Russell 已提交
673 674 675 676 677 678
	Elf_Shdr *pcpusec = &info->sechdrs[info->index.pcpu];
	unsigned long align = pcpusec->sh_addralign;

	if (!pcpusec->sh_size)
		return 0;

679
	if (align > PAGE_SIZE) {
A
Andrew Morton 已提交
680 681
		pr_warn("%s: per-cpu alignment %li > %li\n",
			mod->name, align, PAGE_SIZE);
682 683 684
		align = PAGE_SIZE;
	}

R
Rusty Russell 已提交
685
	mod->percpu = __alloc_reserved_percpu(pcpusec->sh_size, align);
686
	if (!mod->percpu) {
A
Andrew Morton 已提交
687 688
		pr_warn("%s: Could not allocate %lu bytes percpu data\n",
			mod->name, (unsigned long)pcpusec->sh_size);
689 690
		return -ENOMEM;
	}
R
Rusty Russell 已提交
691
	mod->percpu_size = pcpusec->sh_size;
692
	return 0;
693 694
}

695
static void percpu_modfree(struct module *mod)
696
{
697
	free_percpu(mod->percpu);
698 699
}

700
static unsigned int find_pcpusec(struct load_info *info)
701
{
702
	return find_sec(info, ".data..percpu");
703 704
}

705 706
static void percpu_modcopy(struct module *mod,
			   const void *from, unsigned long size)
707 708 709 710
{
	int cpu;

	for_each_possible_cpu(cpu)
711
		memcpy(per_cpu_ptr(mod->percpu, cpu), from, size);
712 713
}

714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730
/**
 * 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) {
731 732
		if (mod->state == MODULE_STATE_UNFORMED)
			continue;
733 734 735 736 737 738 739 740 741 742 743 744 745 746 747
		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;
748 749
}

L
Linus Torvalds 已提交
750
#else /* ... !CONFIG_SMP */
751

752
static inline void __percpu *mod_percpu(struct module *mod)
L
Linus Torvalds 已提交
753 754 755
{
	return NULL;
}
R
Rusty Russell 已提交
756
static int percpu_modalloc(struct module *mod, struct load_info *info)
757
{
R
Rusty Russell 已提交
758 759 760 761
	/* 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;
762 763
}
static inline void percpu_modfree(struct module *mod)
L
Linus Torvalds 已提交
764 765
{
}
766
static unsigned int find_pcpusec(struct load_info *info)
L
Linus Torvalds 已提交
767 768 769
{
	return 0;
}
770 771
static inline void percpu_modcopy(struct module *mod,
				  const void *from, unsigned long size)
L
Linus Torvalds 已提交
772 773 774 775
{
	/* pcpusec should be 0, and size of that section should be 0. */
	BUG_ON(size != 0);
}
776 777 778 779
bool is_module_percpu_address(unsigned long addr)
{
	return false;
}
780

L
Linus Torvalds 已提交
781 782
#endif /* CONFIG_SMP */

783 784 785 786 787 788
#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,   \
789
			struct module_kobject *mk, char *buffer)      \
790
{                                                                     \
791
	return scnprintf(buffer, PAGE_SIZE, "%s\n", mk->mod->field);  \
792 793 794 795 796 797 798
}                                                                     \
static int modinfo_##field##_exists(struct module *mod)               \
{                                                                     \
	return mod->field != NULL;                                    \
}                                                                     \
static void free_modinfo_##field(struct module *mod)                  \
{                                                                     \
D
Daniel Walker 已提交
799 800
	kfree(mod->field);                                            \
	mod->field = NULL;                                            \
801 802
}                                                                     \
static struct module_attribute modinfo_##field = {                    \
803
	.attr = { .name = __stringify(field), .mode = 0444 },         \
804 805 806 807 808 809 810 811 812
	.show = show_modinfo_##field,                                 \
	.setup = setup_modinfo_##field,                               \
	.test = modinfo_##field##_exists,                             \
	.free = free_modinfo_##field,                                 \
};

MODINFO_ATTR(version);
MODINFO_ATTR(srcversion);

813 814
static char last_unloaded_module[MODULE_NAME_LEN+1];

815
#ifdef CONFIG_MODULE_UNLOAD
816 817 818

EXPORT_TRACEPOINT_SYMBOL(module_get);

819 820 821
/* MODULE_REF_BASE is the base reference count by kmodule loader. */
#define MODULE_REF_BASE	1

L
Linus Torvalds 已提交
822
/* Init the unload section of the module. */
823
static int module_unload_init(struct module *mod)
L
Linus Torvalds 已提交
824
{
825 826 827 828 829
	/*
	 * Initialize reference counter to MODULE_REF_BASE.
	 * refcnt == 0 means module is going.
	 */
	atomic_set(&mod->refcnt, MODULE_REF_BASE);
830

831 832
	INIT_LIST_HEAD(&mod->source_list);
	INIT_LIST_HEAD(&mod->target_list);
833

L
Linus Torvalds 已提交
834
	/* Hold reference count during initialization. */
835
	atomic_inc(&mod->refcnt);
836 837

	return 0;
L
Linus Torvalds 已提交
838 839 840 841 842 843 844
}

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

845 846
	list_for_each_entry(use, &b->source_list, source_list) {
		if (use->source == a) {
847
			pr_debug("%s uses %s!\n", a->name, b->name);
L
Linus Torvalds 已提交
848 849 850
			return 1;
		}
	}
851
	pr_debug("%s does not use %s!\n", a->name, b->name);
L
Linus Torvalds 已提交
852 853 854
	return 0;
}

855 856 857 858 859 860 861 862 863 864 865
/*
 * 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;

866
	pr_debug("Allocating new usage for %s.\n", a->name);
867 868
	use = kmalloc(sizeof(*use), GFP_ATOMIC);
	if (!use) {
A
Andrew Morton 已提交
869
		pr_warn("%s: out of memory loading\n", a->name);
870 871 872 873 874 875 876 877 878 879
		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;
}

880
/* Module a uses b: caller needs module_mutex() */
881
int ref_module(struct module *a, struct module *b)
L
Linus Torvalds 已提交
882
{
883
	int err;
K
Kay Sievers 已提交
884

885
	if (b == NULL || already_uses(a, b))
886 887
		return 0;

888 889
	/* If module isn't available, we fail. */
	err = strong_try_module_get(b);
890
	if (err)
891
		return err;
L
Linus Torvalds 已提交
892

893 894
	err = add_module_usage(a, b);
	if (err) {
L
Linus Torvalds 已提交
895
		module_put(b);
896
		return err;
L
Linus Torvalds 已提交
897
	}
898
	return 0;
L
Linus Torvalds 已提交
899
}
900
EXPORT_SYMBOL_GPL(ref_module);
L
Linus Torvalds 已提交
901 902 903 904

/* Clear the unload stuff of the module. */
static void module_unload_free(struct module *mod)
{
905
	struct module_use *use, *tmp;
L
Linus Torvalds 已提交
906

907
	mutex_lock(&module_mutex);
908 909
	list_for_each_entry_safe(use, tmp, &mod->target_list, target_list) {
		struct module *i = use->target;
910
		pr_debug("%s unusing %s\n", mod->name, i->name);
911 912 913 914
		module_put(i);
		list_del(&use->source_list);
		list_del(&use->target_list);
		kfree(use);
L
Linus Torvalds 已提交
915
	}
916
	mutex_unlock(&module_mutex);
L
Linus Torvalds 已提交
917 918 919
}

#ifdef CONFIG_MODULE_FORCE_UNLOAD
920
static inline int try_force_unload(unsigned int flags)
L
Linus Torvalds 已提交
921 922 923
{
	int ret = (flags & O_TRUNC);
	if (ret)
924
		add_taint(TAINT_FORCED_RMMOD, LOCKDEP_NOW_UNRELIABLE);
L
Linus Torvalds 已提交
925 926 927
	return ret;
}
#else
928
static inline int try_force_unload(unsigned int flags)
L
Linus Torvalds 已提交
929 930 931 932 933
{
	return 0;
}
#endif /* CONFIG_MODULE_FORCE_UNLOAD */

934 935
/* Try to release refcount of module, 0 means success. */
static int try_release_module_ref(struct module *mod)
L
Linus Torvalds 已提交
936
{
937
	int ret;
L
Linus Torvalds 已提交
938

939 940 941 942 943 944
	/* 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
945

946 947
	return ret;
}
L
Linus Torvalds 已提交
948

949 950
static int try_stop_module(struct module *mod, int flags, int *forced)
{
951
	/* If it's not unused, quit unless we're forcing. */
952 953 954
	if (try_release_module_ref(mod) != 0) {
		*forced = try_force_unload(flags);
		if (!(*forced))
L
Linus Torvalds 已提交
955 956 957 958
			return -EWOULDBLOCK;
	}

	/* Mark it as dying. */
959
	mod->state = MODULE_STATE_GOING;
L
Linus Torvalds 已提交
960

961
	return 0;
L
Linus Torvalds 已提交
962 963
}

964 965 966 967 968 969 970 971 972 973
/**
 * 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
974
{
975
	return atomic_read(&mod->refcnt) - MODULE_REF_BASE;
L
Linus Torvalds 已提交
976 977 978 979 980 981
}
EXPORT_SYMBOL(module_refcount);

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

982 983
SYSCALL_DEFINE2(delete_module, const char __user *, name_user,
		unsigned int, flags)
L
Linus Torvalds 已提交
984 985
{
	struct module *mod;
986
	char name[MODULE_NAME_LEN];
L
Linus Torvalds 已提交
987 988
	int ret, forced = 0;

989
	if (!capable(CAP_SYS_MODULE) || modules_disabled)
990 991 992 993 994 995
		return -EPERM;

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

996 997
	if (mutex_lock_interruptible(&module_mutex) != 0)
		return -EINTR;
L
Linus Torvalds 已提交
998 999 1000 1001 1002 1003 1004

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

1005
	if (!list_empty(&mod->source_list)) {
L
Linus Torvalds 已提交
1006 1007 1008 1009 1010 1011 1012
		/* 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) {
1013
		/* FIXME: if (force), slam module count damn the torpedoes */
1014
		pr_debug("%s already dying\n", mod->name);
L
Linus Torvalds 已提交
1015 1016 1017 1018 1019
		ret = -EBUSY;
		goto out;
	}

	/* If it has an init func, it must have an exit func to unload */
R
Rusty Russell 已提交
1020
	if (mod->init && !mod->exit) {
1021
		forced = try_force_unload(flags);
L
Linus Torvalds 已提交
1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033
		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;

1034
	mutex_unlock(&module_mutex);
L
Lucas De Marchi 已提交
1035
	/* Final destruction now no one is using it. */
1036
	if (mod->exit != NULL)
L
Linus Torvalds 已提交
1037
		mod->exit();
1038 1039
	blocking_notifier_call_chain(&module_notify_list,
				     MODULE_STATE_GOING, mod);
1040
	async_synchronize_full();
1041

1042
	/* Store the name of the last unloaded module for diagnostic purposes */
1043
	strlcpy(last_unloaded_module, mod->name, sizeof(last_unloaded_module));
L
Linus Torvalds 已提交
1044

1045 1046 1047
	free_module(mod);
	return 0;
out:
1048
	mutex_unlock(&module_mutex);
L
Linus Torvalds 已提交
1049 1050 1051
	return ret;
}

1052
static inline void print_unload_info(struct seq_file *m, struct module *mod)
L
Linus Torvalds 已提交
1053 1054 1055 1056
{
	struct module_use *use;
	int printed_something = 0;

1057
	seq_printf(m, " %i ", module_refcount(mod));
L
Linus Torvalds 已提交
1058

1059 1060 1061 1062
	/*
	 * Always include a trailing , so userspace can differentiate
	 * between this and the old multi-field proc format.
	 */
1063
	list_for_each_entry(use, &mod->source_list, source_list) {
L
Linus Torvalds 已提交
1064
		printed_something = 1;
1065
		seq_printf(m, "%s,", use->source->name);
L
Linus Torvalds 已提交
1066 1067 1068 1069
	}

	if (mod->init != NULL && mod->exit == NULL) {
		printed_something = 1;
1070
		seq_puts(m, "[permanent],");
L
Linus Torvalds 已提交
1071 1072 1073
	}

	if (!printed_something)
1074
		seq_puts(m, "-");
L
Linus Torvalds 已提交
1075 1076 1077 1078 1079 1080
}

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

R
Rusty Russell 已提交
1081
	preempt_disable();
1082
	if (!find_symbol(symbol, &owner, NULL, true, false))
L
Linus Torvalds 已提交
1083 1084
		BUG();
	module_put(owner);
R
Rusty Russell 已提交
1085
	preempt_enable();
L
Linus Torvalds 已提交
1086 1087 1088
}
EXPORT_SYMBOL(__symbol_put);

1089
/* Note this assumes addr is a function, which it currently always is. */
L
Linus Torvalds 已提交
1090 1091
void symbol_put_addr(void *addr)
{
1092
	struct module *modaddr;
1093
	unsigned long a = (unsigned long)dereference_function_descriptor(addr);
L
Linus Torvalds 已提交
1094

1095
	if (core_kernel_text(a))
1096
		return;
L
Linus Torvalds 已提交
1097

1098 1099
	/* module_text_address is safe here: we're supposed to have reference
	 * to module from symbol_get, so it can't go away. */
1100
	modaddr = __module_text_address(a);
1101
	BUG_ON(!modaddr);
1102
	module_put(modaddr);
L
Linus Torvalds 已提交
1103 1104 1105 1106
}
EXPORT_SYMBOL_GPL(symbol_put_addr);

static ssize_t show_refcnt(struct module_attribute *mattr,
1107
			   struct module_kobject *mk, char *buffer)
L
Linus Torvalds 已提交
1108
{
1109
	return sprintf(buffer, "%i\n", module_refcount(mk->mod));
L
Linus Torvalds 已提交
1110 1111
}

1112 1113
static struct module_attribute modinfo_refcnt =
	__ATTR(refcnt, 0444, show_refcnt, NULL);
L
Linus Torvalds 已提交
1114

1115 1116 1117 1118
void __module_get(struct module *module)
{
	if (module) {
		preempt_disable();
1119
		atomic_inc(&module->refcnt);
1120 1121 1122 1123 1124 1125 1126 1127 1128 1129 1130 1131
		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();
1132 1133 1134
		/* Note: here, we can fail to get a reference */
		if (likely(module_is_live(module) &&
			   atomic_inc_not_zero(&module->refcnt) != 0))
1135
			trace_module_get(module, _RET_IP_);
1136
		else
1137 1138 1139 1140 1141 1142 1143 1144
			ret = false;

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

A
Al Viro 已提交
1145 1146
void module_put(struct module *module)
{
1147 1148
	int ret;

A
Al Viro 已提交
1149
	if (module) {
1150
		preempt_disable();
1151 1152
		ret = atomic_dec_if_positive(&module->refcnt);
		WARN_ON(ret < 0);	/* Failed to put refcount */
1153
		trace_module_put(module, _RET_IP_);
1154
		preempt_enable();
A
Al Viro 已提交
1155 1156 1157 1158
	}
}
EXPORT_SYMBOL(module_put);

L
Linus Torvalds 已提交
1159
#else /* !CONFIG_MODULE_UNLOAD */
1160
static inline void print_unload_info(struct seq_file *m, struct module *mod)
L
Linus Torvalds 已提交
1161 1162
{
	/* We don't know the usage count, or what modules are using. */
1163
	seq_puts(m, " - -");
L
Linus Torvalds 已提交
1164 1165 1166 1167 1168 1169
}

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

1170
int ref_module(struct module *a, struct module *b)
L
Linus Torvalds 已提交
1171
{
1172
	return strong_try_module_get(b);
L
Linus Torvalds 已提交
1173
}
1174
EXPORT_SYMBOL_GPL(ref_module);
L
Linus Torvalds 已提交
1175

1176
static inline int module_unload_init(struct module *mod)
L
Linus Torvalds 已提交
1177
{
1178
	return 0;
L
Linus Torvalds 已提交
1179 1180 1181
}
#endif /* CONFIG_MODULE_UNLOAD */

1182 1183 1184 1185 1186 1187 1188 1189 1190 1191 1192 1193
static size_t module_flags_taint(struct module *mod, char *buf)
{
	size_t l = 0;

	if (mod->taints & (1 << TAINT_PROPRIETARY_MODULE))
		buf[l++] = 'P';
	if (mod->taints & (1 << TAINT_OOT_MODULE))
		buf[l++] = 'O';
	if (mod->taints & (1 << TAINT_FORCED_MODULE))
		buf[l++] = 'F';
	if (mod->taints & (1 << TAINT_CRAP))
		buf[l++] = 'C';
1194
	if (mod->taints & (1 << TAINT_UNSIGNED_MODULE))
1195
		buf[l++] = 'E';
1196 1197
	/*
	 * TAINT_FORCED_RMMOD: could be added.
1198
	 * TAINT_CPU_OUT_OF_SPEC, TAINT_MACHINE_CHECK, TAINT_BAD_PAGE don't
1199 1200 1201 1202 1203
	 * apply to modules.
	 */
	return l;
}

1204
static ssize_t show_initstate(struct module_attribute *mattr,
1205
			      struct module_kobject *mk, char *buffer)
1206 1207 1208
{
	const char *state = "unknown";

1209
	switch (mk->mod->state) {
1210 1211 1212 1213 1214 1215 1216 1217 1218
	case MODULE_STATE_LIVE:
		state = "live";
		break;
	case MODULE_STATE_COMING:
		state = "coming";
		break;
	case MODULE_STATE_GOING:
		state = "going";
		break;
1219 1220
	default:
		BUG();
1221 1222 1223 1224
	}
	return sprintf(buffer, "%s\n", state);
}

1225 1226
static struct module_attribute modinfo_initstate =
	__ATTR(initstate, 0444, show_initstate, NULL);
1227

1228 1229 1230 1231 1232 1233 1234 1235 1236 1237 1238
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;
}

1239 1240 1241 1242 1243 1244 1245 1246 1247 1248 1249 1250 1251 1252 1253 1254 1255 1256 1257 1258 1259 1260 1261 1262 1263 1264 1265 1266 1267 1268 1269 1270 1271
struct module_attribute module_uevent =
	__ATTR(uevent, 0200, NULL, store_uevent);

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

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

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

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

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

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

static struct module_attribute modinfo_taint =
	__ATTR(taint, 0444, show_taint, NULL);
1272

1273
static struct module_attribute *modinfo_attrs[] = {
1274
	&module_uevent,
1275 1276
	&modinfo_version,
	&modinfo_srcversion,
1277 1278 1279 1280
	&modinfo_initstate,
	&modinfo_coresize,
	&modinfo_initsize,
	&modinfo_taint,
1281
#ifdef CONFIG_MODULE_UNLOAD
1282
	&modinfo_refcnt,
1283 1284 1285 1286
#endif
	NULL,
};

L
Linus Torvalds 已提交
1287 1288
static const char vermagic[] = VERMAGIC_STRING;

1289
static int try_to_force_load(struct module *mod, const char *reason)
1290 1291
{
#ifdef CONFIG_MODULE_FORCE_LOAD
A
Andi Kleen 已提交
1292
	if (!test_taint(TAINT_FORCED_MODULE))
A
Andrew Morton 已提交
1293
		pr_warn("%s: %s: kernel tainted.\n", mod->name, reason);
1294
	add_taint_module(mod, TAINT_FORCED_MODULE, LOCKDEP_NOW_UNRELIABLE);
1295 1296 1297 1298 1299 1300
	return 0;
#else
	return -ENOEXEC;
#endif
}

L
Linus Torvalds 已提交
1301
#ifdef CONFIG_MODVERSIONS
1302 1303 1304 1305 1306 1307 1308 1309 1310 1311 1312
/* If the arch applies (non-zero) relocations to kernel kcrctab, unapply it. */
static unsigned long maybe_relocated(unsigned long crc,
				     const struct module *crc_owner)
{
#ifdef ARCH_RELOCATES_KCRCTAB
	if (crc_owner == NULL)
		return crc - (unsigned long)reloc_start;
#endif
	return crc;
}

L
Linus Torvalds 已提交
1313 1314 1315
static int check_version(Elf_Shdr *sechdrs,
			 unsigned int versindex,
			 const char *symname,
1316
			 struct module *mod,
1317 1318
			 const unsigned long *crc,
			 const struct module *crc_owner)
L
Linus Torvalds 已提交
1319 1320 1321 1322 1323 1324 1325 1326
{
	unsigned int i, num_versions;
	struct modversion_info *versions;

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

1327 1328 1329 1330
	/* 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
1331 1332 1333 1334 1335 1336 1337 1338
	versions = (void *) sechdrs[versindex].sh_addr;
	num_versions = sechdrs[versindex].sh_size
		/ sizeof(struct modversion_info);

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

1339
		if (versions[i].crc == maybe_relocated(*crc, crc_owner))
L
Linus Torvalds 已提交
1340
			return 1;
1341
		pr_debug("Found checksum %lX vs module %lX\n",
1342
		       maybe_relocated(*crc, crc_owner), versions[i].crc);
1343
		goto bad_version;
L
Linus Torvalds 已提交
1344
	}
1345

A
Andrew Morton 已提交
1346
	pr_warn("%s: no symbol version for %s\n", mod->name, symname);
1347
	return 0;
1348 1349

bad_version:
1350
	pr_warn("%s: disagrees about version of symbol %s\n",
1351 1352
	       mod->name, symname);
	return 0;
L
Linus Torvalds 已提交
1353 1354 1355 1356 1357 1358 1359 1360
}

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

1361 1362 1363 1364 1365
	/*
	 * Since this should be found in kernel (which can't be removed), no
	 * locking is necessary -- use preempt_disable() to placate lockdep.
	 */
	preempt_disable();
R
Rusty Russell 已提交
1366
	if (!find_symbol(VMLINUX_SYMBOL_STR(module_layout), NULL,
1367 1368
			 &crc, true, false)) {
		preempt_enable();
L
Linus Torvalds 已提交
1369
		BUG();
1370 1371
	}
	preempt_enable();
1372 1373
	return check_version(sechdrs, versindex,
			     VMLINUX_SYMBOL_STR(module_layout), mod, crc,
1374
			     NULL);
L
Linus Torvalds 已提交
1375 1376
}

1377 1378 1379
/* 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
1380
{
1381 1382 1383 1384
	if (has_crcs) {
		amagic += strcspn(amagic, " ");
		bmagic += strcspn(bmagic, " ");
	}
L
Linus Torvalds 已提交
1385 1386 1387 1388 1389 1390
	return strcmp(amagic, bmagic) == 0;
}
#else
static inline int check_version(Elf_Shdr *sechdrs,
				unsigned int versindex,
				const char *symname,
1391
				struct module *mod,
1392 1393
				const unsigned long *crc,
				const struct module *crc_owner)
L
Linus Torvalds 已提交
1394 1395 1396 1397 1398 1399 1400 1401 1402 1403 1404
{
	return 1;
}

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

1405 1406
static inline int same_magic(const char *amagic, const char *bmagic,
			     bool has_crcs)
L
Linus Torvalds 已提交
1407 1408 1409 1410 1411
{
	return strcmp(amagic, bmagic) == 0;
}
#endif /* CONFIG_MODVERSIONS */

1412
/* Resolve a symbol for this module.  I.e. if we find one, record usage. */
1413 1414
static const struct kernel_symbol *resolve_symbol(struct module *mod,
						  const struct load_info *info,
1415
						  const char *name,
1416
						  char ownername[])
L
Linus Torvalds 已提交
1417 1418
{
	struct module *owner;
1419
	const struct kernel_symbol *sym;
L
Linus Torvalds 已提交
1420
	const unsigned long *crc;
1421
	int err;
L
Linus Torvalds 已提交
1422

1423 1424 1425 1426 1427 1428
	/*
	 * 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();
1429
	mutex_lock(&module_mutex);
1430
	sym = find_symbol(name, &owner, &crc,
A
Andi Kleen 已提交
1431
			  !(mod->taints & (1 << TAINT_PROPRIETARY_MODULE)), true);
1432 1433 1434
	if (!sym)
		goto unlock;

1435 1436
	if (!check_version(info->sechdrs, info->index.vers, name, mod, crc,
			   owner)) {
1437 1438
		sym = ERR_PTR(-EINVAL);
		goto getname;
L
Linus Torvalds 已提交
1439
	}
1440 1441 1442 1443 1444 1445 1446 1447 1448 1449 1450

	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:
1451
	mutex_unlock(&module_mutex);
1452
	return sym;
L
Linus Torvalds 已提交
1453 1454
}

1455 1456 1457 1458
static const struct kernel_symbol *
resolve_symbol_wait(struct module *mod,
		    const struct load_info *info,
		    const char *name)
1459 1460
{
	const struct kernel_symbol *ksym;
1461
	char owner[MODULE_NAME_LEN];
1462 1463

	if (wait_event_interruptible_timeout(module_wq,
1464 1465
			!IS_ERR(ksym = resolve_symbol(mod, info, name, owner))
			|| PTR_ERR(ksym) != -EBUSY,
1466
					     30 * HZ) <= 0) {
A
Andrew Morton 已提交
1467 1468
		pr_warn("%s: gave up waiting for init of module %s.\n",
			mod->name, owner);
1469 1470 1471 1472
	}
	return ksym;
}

L
Linus Torvalds 已提交
1473 1474 1475 1476
/*
 * /sys/module/foo/sections stuff
 * J. Corbet <corbet@lwn.net>
 */
R
Rusty Russell 已提交
1477
#ifdef CONFIG_SYSFS
1478

R
Rusty Russell 已提交
1479
#ifdef CONFIG_KALLSYMS
1480 1481 1482 1483 1484
static inline bool sect_empty(const Elf_Shdr *sect)
{
	return !(sect->sh_flags & SHF_ALLOC) || sect->sh_size == 0;
}

1485
struct module_sect_attr {
1486 1487 1488 1489 1490
	struct module_attribute mattr;
	char *name;
	unsigned long address;
};

1491
struct module_sect_attrs {
1492 1493 1494 1495 1496
	struct attribute_group grp;
	unsigned int nsections;
	struct module_sect_attr attrs[0];
};

L
Linus Torvalds 已提交
1497
static ssize_t module_sect_show(struct module_attribute *mattr,
1498
				struct module_kobject *mk, char *buf)
L
Linus Torvalds 已提交
1499 1500 1501
{
	struct module_sect_attr *sattr =
		container_of(mattr, struct module_sect_attr, mattr);
1502
	return sprintf(buf, "0x%pK\n", (void *)sattr->address);
L
Linus Torvalds 已提交
1503 1504
}

1505 1506
static void free_sect_attrs(struct module_sect_attrs *sect_attrs)
{
1507
	unsigned int section;
1508 1509 1510 1511 1512 1513

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

R
Rusty Russell 已提交
1514
static void add_sect_attrs(struct module *mod, const struct load_info *info)
L
Linus Torvalds 已提交
1515 1516 1517 1518 1519
{
	unsigned int nloaded = 0, i, size[2];
	struct module_sect_attrs *sect_attrs;
	struct module_sect_attr *sattr;
	struct attribute **gattr;
D
Daniel Walker 已提交
1520

L
Linus Torvalds 已提交
1521
	/* Count loaded sections and allocate structures */
R
Rusty Russell 已提交
1522 1523
	for (i = 0; i < info->hdr->e_shnum; i++)
		if (!sect_empty(&info->sechdrs[i]))
L
Linus Torvalds 已提交
1524 1525 1526 1527 1528
			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]);
1529 1530
	sect_attrs = kzalloc(size[0] + size[1], GFP_KERNEL);
	if (sect_attrs == NULL)
L
Linus Torvalds 已提交
1531 1532 1533 1534 1535 1536
		return;

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

1537
	sect_attrs->nsections = 0;
L
Linus Torvalds 已提交
1538 1539
	sattr = &sect_attrs->attrs[0];
	gattr = &sect_attrs->grp.attrs[0];
R
Rusty Russell 已提交
1540 1541 1542
	for (i = 0; i < info->hdr->e_shnum; i++) {
		Elf_Shdr *sec = &info->sechdrs[i];
		if (sect_empty(sec))
1543
			continue;
R
Rusty Russell 已提交
1544 1545
		sattr->address = sec->sh_addr;
		sattr->name = kstrdup(info->secstrings + sec->sh_name,
1546 1547 1548 1549
					GFP_KERNEL);
		if (sattr->name == NULL)
			goto out;
		sect_attrs->nsections++;
1550
		sysfs_attr_init(&sattr->mattr.attr);
L
Linus Torvalds 已提交
1551 1552 1553 1554 1555 1556 1557 1558 1559 1560 1561 1562 1563 1564
		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:
1565
	free_sect_attrs(sect_attrs);
L
Linus Torvalds 已提交
1566 1567 1568 1569 1570 1571 1572 1573 1574
}

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. */
1575
		free_sect_attrs(mod->sect_attrs);
L
Linus Torvalds 已提交
1576 1577 1578 1579
		mod->sect_attrs = NULL;
	}
}

R
Roland McGrath 已提交
1580 1581 1582 1583 1584 1585 1586 1587 1588 1589
/*
 * /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];
};

1590
static ssize_t module_notes_read(struct file *filp, struct kobject *kobj,
R
Roland McGrath 已提交
1591 1592 1593 1594 1595 1596 1597 1598 1599 1600 1601 1602 1603 1604 1605 1606 1607
				 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]);
1608
		kobject_put(notes_attrs->dir);
R
Roland McGrath 已提交
1609 1610 1611 1612
	}
	kfree(notes_attrs);
}

R
Rusty Russell 已提交
1613
static void add_notes_attrs(struct module *mod, const struct load_info *info)
R
Roland McGrath 已提交
1614 1615 1616 1617 1618
{
	unsigned int notes, loaded, i;
	struct module_notes_attrs *notes_attrs;
	struct bin_attribute *nattr;

1619 1620 1621 1622
	/* failed to create section attributes, so can't create notes */
	if (!mod->sect_attrs)
		return;

R
Roland McGrath 已提交
1623 1624
	/* Count notes sections and allocate structures.  */
	notes = 0;
R
Rusty Russell 已提交
1625 1626 1627
	for (i = 0; i < info->hdr->e_shnum; i++)
		if (!sect_empty(&info->sechdrs[i]) &&
		    (info->sechdrs[i].sh_type == SHT_NOTE))
R
Roland McGrath 已提交
1628 1629 1630 1631 1632 1633 1634 1635 1636 1637 1638 1639 1640
			++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
1641 1642
	for (loaded = i = 0; i < info->hdr->e_shnum; ++i) {
		if (sect_empty(&info->sechdrs[i]))
R
Roland McGrath 已提交
1643
			continue;
R
Rusty Russell 已提交
1644
		if (info->sechdrs[i].sh_type == SHT_NOTE) {
1645
			sysfs_bin_attr_init(nattr);
R
Roland McGrath 已提交
1646 1647
			nattr->attr.name = mod->sect_attrs->attrs[loaded].name;
			nattr->attr.mode = S_IRUGO;
R
Rusty Russell 已提交
1648 1649
			nattr->size = info->sechdrs[i].sh_size;
			nattr->private = (void *) info->sechdrs[i].sh_addr;
R
Roland McGrath 已提交
1650 1651 1652 1653 1654 1655
			nattr->read = module_notes_read;
			++nattr;
		}
		++loaded;
	}

1656
	notes_attrs->dir = kobject_create_and_add("notes", &mod->mkobj.kobj);
R
Roland McGrath 已提交
1657 1658 1659 1660 1661 1662 1663 1664 1665 1666 1667 1668 1669 1670 1671 1672 1673 1674 1675 1676 1677
	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
1678
#else
1679

R
Rusty Russell 已提交
1680 1681
static inline void add_sect_attrs(struct module *mod,
				  const struct load_info *info)
L
Linus Torvalds 已提交
1682 1683 1684 1685 1686 1687
{
}

static inline void remove_sect_attrs(struct module *mod)
{
}
R
Roland McGrath 已提交
1688

R
Rusty Russell 已提交
1689 1690
static inline void add_notes_attrs(struct module *mod,
				   const struct load_info *info)
R
Roland McGrath 已提交
1691 1692 1693 1694 1695 1696
{
}

static inline void remove_notes_attrs(struct module *mod)
{
}
R
Rusty Russell 已提交
1697
#endif /* CONFIG_KALLSYMS */
L
Linus Torvalds 已提交
1698

1699 1700 1701 1702 1703 1704
static void add_usage_links(struct module *mod)
{
#ifdef CONFIG_MODULE_UNLOAD
	struct module_use *use;
	int nowarn;

1705
	mutex_lock(&module_mutex);
1706 1707 1708 1709
	list_for_each_entry(use, &mod->target_list, target_list) {
		nowarn = sysfs_create_link(use->target->holders_dir,
					   &mod->mkobj.kobj, mod->name);
	}
1710
	mutex_unlock(&module_mutex);
1711 1712 1713 1714 1715 1716 1717 1718
#endif
}

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

1719
	mutex_lock(&module_mutex);
1720 1721
	list_for_each_entry(use, &mod->target_list, target_list)
		sysfs_remove_link(use->target->holders_dir, mod->name);
1722
	mutex_unlock(&module_mutex);
1723 1724 1725
#endif
}

1726
static int module_add_modinfo_attrs(struct module *mod)
1727 1728
{
	struct module_attribute *attr;
1729
	struct module_attribute *temp_attr;
1730 1731 1732
	int error = 0;
	int i;

1733 1734 1735 1736 1737 1738 1739
	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;
1740 1741
	for (i = 0; (attr = modinfo_attrs[i]) && !error; i++) {
		if (!attr->test ||
1742 1743
		    (attr->test && attr->test(mod))) {
			memcpy(temp_attr, attr, sizeof(*temp_attr));
1744
			sysfs_attr_init(&temp_attr->attr);
1745 1746
			error = sysfs_create_file(&mod->mkobj.kobj,
					&temp_attr->attr);
1747 1748
			++temp_attr;
		}
1749 1750 1751 1752
	}
	return error;
}

1753
static void module_remove_modinfo_attrs(struct module *mod)
1754 1755 1756 1757
{
	struct module_attribute *attr;
	int i;

1758 1759 1760 1761
	for (i = 0; (attr = &mod->modinfo_attrs[i]); i++) {
		/* pick a field to test for end of list */
		if (!attr->attr.name)
			break;
1762
		sysfs_remove_file(&mod->mkobj.kobj, &attr->attr);
1763 1764
		if (attr->free)
			attr->free(mod);
1765
	}
1766
	kfree(mod->modinfo_attrs);
1767
}
L
Linus Torvalds 已提交
1768

1769 1770 1771 1772 1773 1774 1775 1776
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);
}

1777
static int mod_sysfs_init(struct module *mod)
L
Linus Torvalds 已提交
1778 1779
{
	int err;
1780
	struct kobject *kobj;
L
Linus Torvalds 已提交
1781

1782
	if (!module_sysfs_initialized) {
A
Andrew Morton 已提交
1783
		pr_err("%s: module sysfs not initialized\n", mod->name);
1784 1785 1786
		err = -EINVAL;
		goto out;
	}
1787 1788 1789

	kobj = kset_find_obj(module_kset, mod->name);
	if (kobj) {
A
Andrew Morton 已提交
1790
		pr_err("%s: module is already loaded\n", mod->name);
1791 1792 1793 1794 1795
		kobject_put(kobj);
		err = -EINVAL;
		goto out;
	}

L
Linus Torvalds 已提交
1796
	mod->mkobj.mod = mod;
1797

1798 1799 1800 1801 1802
	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)
1803
		mod_kobject_put(mod);
K
Kay Sievers 已提交
1804

1805
	/* delay uevent until full sysfs population */
K
Kay Sievers 已提交
1806 1807 1808 1809
out:
	return err;
}

1810
static int mod_sysfs_setup(struct module *mod,
R
Rusty Russell 已提交
1811
			   const struct load_info *info,
K
Kay Sievers 已提交
1812 1813 1814 1815 1816
			   struct kernel_param *kparam,
			   unsigned int num_params)
{
	int err;

1817 1818 1819 1820
	err = mod_sysfs_init(mod);
	if (err)
		goto out;

1821
	mod->holders_dir = kobject_create_and_add("holders", &mod->mkobj.kobj);
1822 1823
	if (!mod->holders_dir) {
		err = -ENOMEM;
K
Kay Sievers 已提交
1824
		goto out_unreg;
1825
	}
K
Kay Sievers 已提交
1826

L
Linus Torvalds 已提交
1827 1828
	err = module_param_sysfs_setup(mod, kparam, num_params);
	if (err)
K
Kay Sievers 已提交
1829
		goto out_unreg_holders;
L
Linus Torvalds 已提交
1830

1831 1832
	err = module_add_modinfo_attrs(mod);
	if (err)
1833
		goto out_unreg_param;
1834

1835
	add_usage_links(mod);
R
Rusty Russell 已提交
1836 1837
	add_sect_attrs(mod, info);
	add_notes_attrs(mod, info);
1838

1839
	kobject_uevent(&mod->mkobj.kobj, KOBJ_ADD);
L
Linus Torvalds 已提交
1840 1841
	return 0;

1842 1843
out_unreg_param:
	module_param_sysfs_remove(mod);
K
Kay Sievers 已提交
1844
out_unreg_holders:
1845
	kobject_put(mod->holders_dir);
K
Kay Sievers 已提交
1846
out_unreg:
1847
	mod_kobject_put(mod);
1848
out:
L
Linus Torvalds 已提交
1849 1850
	return err;
}
1851 1852 1853

static void mod_sysfs_fini(struct module *mod)
{
R
Rusty Russell 已提交
1854 1855
	remove_notes_attrs(mod);
	remove_sect_attrs(mod);
1856
	mod_kobject_put(mod);
1857 1858
}

R
Rusty Russell 已提交
1859
#else /* !CONFIG_SYSFS */
1860

R
Rusty Russell 已提交
1861 1862
static int mod_sysfs_setup(struct module *mod,
			   const struct load_info *info,
1863 1864 1865 1866 1867 1868
			   struct kernel_param *kparam,
			   unsigned int num_params)
{
	return 0;
}

1869 1870 1871 1872
static void mod_sysfs_fini(struct module *mod)
{
}

1873 1874 1875 1876
static void module_remove_modinfo_attrs(struct module *mod)
{
}

1877 1878 1879 1880
static void del_usage_links(struct module *mod)
{
}

1881
#endif /* CONFIG_SYSFS */
L
Linus Torvalds 已提交
1882

1883
static void mod_sysfs_teardown(struct module *mod)
L
Linus Torvalds 已提交
1884
{
1885
	del_usage_links(mod);
1886
	module_remove_modinfo_attrs(mod);
L
Linus Torvalds 已提交
1887
	module_param_sysfs_remove(mod);
1888 1889
	kobject_put(mod->mkobj.drivers_dir);
	kobject_put(mod->holders_dir);
1890
	mod_sysfs_fini(mod);
L
Linus Torvalds 已提交
1891 1892
}

1893 1894 1895 1896 1897 1898 1899 1900 1901 1902 1903 1904 1905 1906 1907 1908 1909 1910 1911 1912 1913 1914 1915 1916 1917 1918 1919 1920 1921 1922 1923 1924 1925 1926 1927 1928 1929 1930 1931 1932 1933 1934 1935 1936
#ifdef CONFIG_DEBUG_SET_MODULE_RONX
/*
 * LKM RO/NX protection: protect module's text/ro-data
 * from modification and any data from execution.
 */
void set_page_attributes(void *start, void *end, int (*set)(unsigned long start, int num_pages))
{
	unsigned long begin_pfn = PFN_DOWN((unsigned long)start);
	unsigned long end_pfn = PFN_DOWN((unsigned long)end);

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

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

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

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

1937 1938 1939 1940 1941 1942 1943 1944 1945 1946 1947 1948 1949 1950 1951 1952 1953 1954
static void unset_module_core_ro_nx(struct module *mod)
{
	set_page_attributes(mod->module_core + mod->core_text_size,
		mod->module_core + mod->core_size,
		set_memory_x);
	set_page_attributes(mod->module_core,
		mod->module_core + mod->core_ro_size,
		set_memory_rw);
}

static void unset_module_init_ro_nx(struct module *mod)
{
	set_page_attributes(mod->module_init + mod->init_text_size,
		mod->module_init + mod->init_size,
		set_memory_x);
	set_page_attributes(mod->module_init,
		mod->module_init + mod->init_ro_size,
		set_memory_rw);
1955 1956 1957
}

/* Iterate through all modules and set each module's text as RW */
1958
void set_all_modules_text_rw(void)
1959 1960 1961 1962 1963
{
	struct module *mod;

	mutex_lock(&module_mutex);
	list_for_each_entry_rcu(mod, &modules, list) {
1964 1965
		if (mod->state == MODULE_STATE_UNFORMED)
			continue;
1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980
		if ((mod->module_core) && (mod->core_text_size)) {
			set_page_attributes(mod->module_core,
						mod->module_core + mod->core_text_size,
						set_memory_rw);
		}
		if ((mod->module_init) && (mod->init_text_size)) {
			set_page_attributes(mod->module_init,
						mod->module_init + mod->init_text_size,
						set_memory_rw);
		}
	}
	mutex_unlock(&module_mutex);
}

/* Iterate through all modules and set each module's text as RO */
1981
void set_all_modules_text_ro(void)
1982 1983 1984 1985 1986
{
	struct module *mod;

	mutex_lock(&module_mutex);
	list_for_each_entry_rcu(mod, &modules, list) {
1987 1988
		if (mod->state == MODULE_STATE_UNFORMED)
			continue;
1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003
		if ((mod->module_core) && (mod->core_text_size)) {
			set_page_attributes(mod->module_core,
						mod->module_core + mod->core_text_size,
						set_memory_ro);
		}
		if ((mod->module_init) && (mod->init_text_size)) {
			set_page_attributes(mod->module_init,
						mod->module_init + mod->init_text_size,
						set_memory_ro);
		}
	}
	mutex_unlock(&module_mutex);
}
#else
static inline void set_section_ro_nx(void *base, unsigned long text_size, unsigned long ro_size, unsigned long total_size) { }
2004 2005
static void unset_module_core_ro_nx(struct module *mod) { }
static void unset_module_init_ro_nx(struct module *mod) { }
2006 2007
#endif

2008
void __weak module_memfree(void *module_region)
2009 2010 2011 2012 2013 2014 2015 2016
{
	vfree(module_region);
}

void __weak module_arch_cleanup(struct module *mod)
{
}

2017 2018 2019 2020
void __weak module_arch_freeing_init(struct module *mod)
{
}

2021
/* Free a module, remove from lists, etc. */
L
Linus Torvalds 已提交
2022 2023
static void free_module(struct module *mod)
{
2024 2025
	trace_module_free(mod);

2026
	mod_sysfs_teardown(mod);
L
Linus Torvalds 已提交
2027

2028 2029
	/* We leave it in list to prevent duplicate loads, but make sure
	 * that noone uses it while it's being deconstructed. */
2030
	mutex_lock(&module_mutex);
2031
	mod->state = MODULE_STATE_UNFORMED;
2032
	mutex_unlock(&module_mutex);
2033

2034 2035 2036
	/* Remove dynamic debug info */
	ddebug_remove_module(mod->name);

L
Linus Torvalds 已提交
2037 2038 2039 2040 2041 2042
	/* Arch-specific cleanup. */
	module_arch_cleanup(mod);

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

2043 2044 2045
	/* Free any allocated parameters. */
	destroy_params(mod->kp, mod->num_kp);

2046 2047
	/* Now we can delete it from the lists */
	mutex_lock(&module_mutex);
2048 2049
	/* Unlink carefully: kallsyms could be walking list. */
	list_del_rcu(&mod->list);
2050
	mod_tree_remove(mod);
2051
	/* Remove this module from bug list, this uses list_del_rcu */
2052
	module_bug_cleanup(mod);
2053 2054
	/* Wait for RCU-sched synchronizing before releasing mod->list and buglist. */
	synchronize_sched();
2055 2056
	mutex_unlock(&module_mutex);

L
Linus Torvalds 已提交
2057
	/* This may be NULL, but that's OK */
2058
	unset_module_init_ro_nx(mod);
2059
	module_arch_freeing_init(mod);
2060
	module_memfree(mod->module_init);
L
Linus Torvalds 已提交
2061
	kfree(mod->args);
2062
	percpu_modfree(mod);
2063

2064
	/* Free lock-classes; relies on the preceding sync_rcu(). */
I
Ingo Molnar 已提交
2065 2066
	lockdep_free_key_range(mod->module_core, mod->core_size);

L
Linus Torvalds 已提交
2067
	/* Finally, free the core (containing the module structure) */
2068
	unset_module_core_ro_nx(mod);
2069
	module_memfree(mod->module_core);
2070 2071 2072 2073

#ifdef CONFIG_MPU
	update_protections(current->mm);
#endif
L
Linus Torvalds 已提交
2074 2075 2076 2077 2078
}

void *__symbol_get(const char *symbol)
{
	struct module *owner;
2079
	const struct kernel_symbol *sym;
L
Linus Torvalds 已提交
2080

R
Rusty Russell 已提交
2081
	preempt_disable();
2082 2083 2084
	sym = find_symbol(symbol, &owner, NULL, true, true);
	if (sym && strong_try_module_get(owner))
		sym = NULL;
R
Rusty Russell 已提交
2085
	preempt_enable();
L
Linus Torvalds 已提交
2086

2087
	return sym ? (void *)sym->value : NULL;
L
Linus Torvalds 已提交
2088 2089 2090
}
EXPORT_SYMBOL_GPL(__symbol_get);

2091 2092
/*
 * Ensure that an exported symbol [global namespace] does not already exist
2093
 * in the kernel or in some other module's exported symbol table.
2094 2095
 *
 * You must hold the module_mutex.
2096 2097 2098
 */
static int verify_export_symbols(struct module *mod)
{
2099
	unsigned int i;
2100
	struct module *owner;
2101 2102 2103 2104 2105 2106 2107 2108
	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 },
2109
#ifdef CONFIG_UNUSED_SYMBOLS
2110 2111
		{ mod->unused_syms, mod->num_unused_syms },
		{ mod->unused_gpl_syms, mod->num_unused_gpl_syms },
2112
#endif
2113
	};
2114

2115 2116
	for (i = 0; i < ARRAY_SIZE(arr); i++) {
		for (s = arr[i].sym; s < arr[i].sym + arr[i].num; s++) {
2117
			if (find_symbol(s->name, &owner, NULL, true, false)) {
A
Andrew Morton 已提交
2118
				pr_err("%s: exports duplicate symbol %s"
2119 2120 2121 2122
				       " (owned by %s)\n",
				       mod->name, s->name, module_name(owner));
				return -ENOEXEC;
			}
2123
		}
2124 2125
	}
	return 0;
2126 2127
}

2128
/* Change all symbols so that st_value encodes the pointer directly. */
2129 2130 2131 2132
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
2133
	unsigned long secbase;
2134
	unsigned int i;
L
Linus Torvalds 已提交
2135
	int ret = 0;
2136
	const struct kernel_symbol *ksym;
L
Linus Torvalds 已提交
2137

2138 2139 2140
	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
2141 2142
		switch (sym[i].st_shndx) {
		case SHN_COMMON:
2143 2144 2145 2146
			/* Ignore common symbols */
			if (!strncmp(name, "__gnu_lto", 9))
				break;

L
Linus Torvalds 已提交
2147 2148
			/* We compiled with -fno-common.  These are not
			   supposed to happen.  */
2149
			pr_debug("Common symbol: %s\n", name);
2150
			pr_warn("%s: please compile with -fno-common\n",
L
Linus Torvalds 已提交
2151 2152 2153 2154 2155 2156
			       mod->name);
			ret = -ENOEXEC;
			break;

		case SHN_ABS:
			/* Don't need to do anything */
2157
			pr_debug("Absolute symbol: 0x%08lx\n",
L
Linus Torvalds 已提交
2158 2159 2160 2161
			       (long)sym[i].st_value);
			break;

		case SHN_UNDEF:
2162
			ksym = resolve_symbol_wait(mod, info, name);
L
Linus Torvalds 已提交
2163
			/* Ok if resolved.  */
2164
			if (ksym && !IS_ERR(ksym)) {
2165
				sym[i].st_value = ksym->value;
L
Linus Torvalds 已提交
2166
				break;
2167 2168
			}

L
Linus Torvalds 已提交
2169
			/* Ok if weak.  */
2170
			if (!ksym && ELF_ST_BIND(sym[i].st_info) == STB_WEAK)
L
Linus Torvalds 已提交
2171 2172
				break;

A
Andrew Morton 已提交
2173 2174
			pr_warn("%s: Unknown symbol %s (err %li)\n",
				mod->name, name, PTR_ERR(ksym));
2175
			ret = PTR_ERR(ksym) ?: -ENOENT;
L
Linus Torvalds 已提交
2176 2177 2178 2179
			break;

		default:
			/* Divert to percpu allocation if a percpu var. */
2180
			if (sym[i].st_shndx == info->index.pcpu)
2181
				secbase = (unsigned long)mod_percpu(mod);
L
Linus Torvalds 已提交
2182
			else
2183
				secbase = info->sechdrs[sym[i].st_shndx].sh_addr;
L
Linus Torvalds 已提交
2184 2185 2186 2187 2188 2189 2190 2191
			sym[i].st_value += secbase;
			break;
		}
	}

	return ret;
}

2192
static int apply_relocations(struct module *mod, const struct load_info *info)
2193 2194 2195 2196 2197
{
	unsigned int i;
	int err = 0;

	/* Now do relocations. */
2198 2199
	for (i = 1; i < info->hdr->e_shnum; i++) {
		unsigned int infosec = info->sechdrs[i].sh_info;
2200 2201

		/* Not a valid relocation section? */
2202
		if (infosec >= info->hdr->e_shnum)
2203 2204 2205
			continue;

		/* Don't bother with non-allocated sections */
2206
		if (!(info->sechdrs[infosec].sh_flags & SHF_ALLOC))
2207 2208
			continue;

2209 2210 2211 2212 2213 2214
		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);
2215 2216 2217 2218 2219 2220
		if (err < 0)
			break;
	}
	return err;
}

2221 2222 2223 2224 2225 2226 2227 2228
/* 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
2229
/* Update size with this section: return offset. */
2230 2231
static long get_offset(struct module *mod, unsigned int *size,
		       Elf_Shdr *sechdr, unsigned int section)
L
Linus Torvalds 已提交
2232 2233 2234
{
	long ret;

2235
	*size += arch_mod_section_prepend(mod, section);
L
Linus Torvalds 已提交
2236 2237 2238 2239 2240 2241 2242 2243 2244
	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. */
2245
static void layout_sections(struct module *mod, struct load_info *info)
L
Linus Torvalds 已提交
2246 2247 2248 2249 2250 2251 2252 2253 2254 2255 2256 2257
{
	static unsigned long const masks[][2] = {
		/* NOTE: all executable code must be the first section
		 * in this array; otherwise modify the text_size
		 * finder in the two loops below */
		{ SHF_EXECINSTR | SHF_ALLOC, ARCH_SHF_SMALL },
		{ SHF_ALLOC, SHF_WRITE | ARCH_SHF_SMALL },
		{ SHF_WRITE | SHF_ALLOC, ARCH_SHF_SMALL },
		{ ARCH_SHF_SMALL | SHF_ALLOC, 0 }
	};
	unsigned int m, i;

2258 2259
	for (i = 0; i < info->hdr->e_shnum; i++)
		info->sechdrs[i].sh_entsize = ~0UL;
L
Linus Torvalds 已提交
2260

2261
	pr_debug("Core section allocation order:\n");
L
Linus Torvalds 已提交
2262
	for (m = 0; m < ARRAY_SIZE(masks); ++m) {
2263 2264 2265
		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
2266 2267 2268 2269

			if ((s->sh_flags & masks[m][0]) != masks[m][0]
			    || (s->sh_flags & masks[m][1])
			    || s->sh_entsize != ~0UL
2270
			    || strstarts(sname, ".init"))
L
Linus Torvalds 已提交
2271
				continue;
2272
			s->sh_entsize = get_offset(mod, &mod->core_size, s, i);
2273
			pr_debug("\t%s\n", sname);
L
Linus Torvalds 已提交
2274
		}
2275 2276 2277
		switch (m) {
		case 0: /* executable */
			mod->core_size = debug_align(mod->core_size);
L
Linus Torvalds 已提交
2278
			mod->core_text_size = mod->core_size;
2279 2280 2281 2282 2283 2284 2285 2286 2287
			break;
		case 1: /* RO: text and ro-data */
			mod->core_size = debug_align(mod->core_size);
			mod->core_ro_size = mod->core_size;
			break;
		case 3: /* whole core */
			mod->core_size = debug_align(mod->core_size);
			break;
		}
L
Linus Torvalds 已提交
2288 2289
	}

2290
	pr_debug("Init section allocation order:\n");
L
Linus Torvalds 已提交
2291
	for (m = 0; m < ARRAY_SIZE(masks); ++m) {
2292 2293 2294
		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
2295 2296 2297 2298

			if ((s->sh_flags & masks[m][0]) != masks[m][0]
			    || (s->sh_flags & masks[m][1])
			    || s->sh_entsize != ~0UL
2299
			    || !strstarts(sname, ".init"))
L
Linus Torvalds 已提交
2300
				continue;
2301
			s->sh_entsize = (get_offset(mod, &mod->init_size, s, i)
L
Linus Torvalds 已提交
2302
					 | INIT_OFFSET_MASK);
2303
			pr_debug("\t%s\n", sname);
L
Linus Torvalds 已提交
2304
		}
2305 2306 2307
		switch (m) {
		case 0: /* executable */
			mod->init_size = debug_align(mod->init_size);
L
Linus Torvalds 已提交
2308
			mod->init_text_size = mod->init_size;
2309 2310 2311 2312 2313 2314 2315 2316 2317
			break;
		case 1: /* RO: text and ro-data */
			mod->init_size = debug_align(mod->init_size);
			mod->init_ro_size = mod->init_size;
			break;
		case 3: /* whole init */
			mod->init_size = debug_align(mod->init_size);
			break;
		}
L
Linus Torvalds 已提交
2318 2319 2320 2321 2322 2323 2324 2325
	}
}

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

2326
	if (!license_is_gpl_compatible(license)) {
A
Andi Kleen 已提交
2327
		if (!test_taint(TAINT_PROPRIETARY_MODULE))
A
Andrew Morton 已提交
2328 2329
			pr_warn("%s: module license '%s' taints kernel.\n",
				mod->name, license);
2330 2331
		add_taint_module(mod, TAINT_PROPRIETARY_MODULE,
				 LOCKDEP_NOW_UNRELIABLE);
L
Linus Torvalds 已提交
2332 2333 2334 2335 2336 2337 2338 2339 2340 2341 2342 2343 2344 2345 2346 2347 2348 2349 2350 2351 2352 2353
	}
}

/* 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;
}

2354
static char *get_modinfo(struct load_info *info, const char *tag)
L
Linus Torvalds 已提交
2355 2356 2357
{
	char *p;
	unsigned int taglen = strlen(tag);
2358 2359
	Elf_Shdr *infosec = &info->sechdrs[info->index.info];
	unsigned long size = infosec->sh_size;
L
Linus Torvalds 已提交
2360

2361
	for (p = (char *)infosec->sh_addr; p; p = next_string(p, &size)) {
L
Linus Torvalds 已提交
2362 2363 2364 2365 2366 2367
		if (strncmp(p, tag, taglen) == 0 && p[taglen] == '=')
			return p + taglen + 1;
	}
	return NULL;
}

2368
static void setup_modinfo(struct module *mod, struct load_info *info)
2369 2370 2371 2372 2373 2374
{
	struct module_attribute *attr;
	int i;

	for (i = 0; (attr = modinfo_attrs[i]); i++) {
		if (attr->setup)
2375
			attr->setup(mod, get_modinfo(info, attr->attr.name));
2376 2377 2378
	}
}

2379 2380 2381 2382 2383 2384 2385 2386 2387 2388 2389
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
2390
#ifdef CONFIG_KALLSYMS
2391 2392 2393 2394 2395 2396

/* 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)
{
2397 2398
	return bsearch(name, start, stop - start,
			sizeof(struct kernel_symbol), cmp_name);
2399 2400
}

2401 2402
static int is_exported(const char *name, unsigned long value,
		       const struct module *mod)
L
Linus Torvalds 已提交
2403
{
2404 2405 2406
	const struct kernel_symbol *ks;
	if (!mod)
		ks = lookup_symbol(name, __start___ksymtab, __stop___ksymtab);
2407
	else
2408 2409
		ks = lookup_symbol(name, mod->syms, mod->syms + mod->num_syms);
	return ks != NULL && ks->value == value;
L
Linus Torvalds 已提交
2410 2411 2412
}

/* As per nm */
2413
static char elf_type(const Elf_Sym *sym, const struct load_info *info)
L
Linus Torvalds 已提交
2414
{
2415 2416
	const Elf_Shdr *sechdrs = info->sechdrs;

L
Linus Torvalds 已提交
2417 2418 2419 2420 2421 2422 2423 2424 2425 2426 2427 2428 2429 2430 2431 2432 2433 2434 2435 2436 2437 2438 2439 2440 2441 2442 2443 2444 2445
	if (ELF_ST_BIND(sym->st_info) == STB_WEAK) {
		if (ELF_ST_TYPE(sym->st_info) == STT_OBJECT)
			return 'v';
		else
			return 'w';
	}
	if (sym->st_shndx == SHN_UNDEF)
		return 'U';
	if (sym->st_shndx == SHN_ABS)
		return 'a';
	if (sym->st_shndx >= SHN_LORESERVE)
		return '?';
	if (sechdrs[sym->st_shndx].sh_flags & SHF_EXECINSTR)
		return 't';
	if (sechdrs[sym->st_shndx].sh_flags & SHF_ALLOC
	    && sechdrs[sym->st_shndx].sh_type != SHT_NOBITS) {
		if (!(sechdrs[sym->st_shndx].sh_flags & SHF_WRITE))
			return 'r';
		else if (sechdrs[sym->st_shndx].sh_flags & ARCH_SHF_SMALL)
			return 'g';
		else
			return 'd';
	}
	if (sechdrs[sym->st_shndx].sh_type == SHT_NOBITS) {
		if (sechdrs[sym->st_shndx].sh_flags & ARCH_SHF_SMALL)
			return 's';
		else
			return 'b';
	}
2446 2447
	if (strstarts(info->secstrings + sechdrs[sym->st_shndx].sh_name,
		      ".debug")) {
L
Linus Torvalds 已提交
2448
		return 'n';
2449
	}
L
Linus Torvalds 已提交
2450 2451 2452
	return '?';
}

2453
static bool is_core_symbol(const Elf_Sym *src, const Elf_Shdr *sechdrs,
2454
			unsigned int shnum)
2455 2456 2457 2458 2459 2460 2461 2462 2463 2464 2465 2466 2467 2468 2469 2470 2471 2472 2473
{
	const Elf_Shdr *sec;

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

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

	return true;
}

2474 2475 2476 2477 2478 2479 2480
/*
 * 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>.
 */
2481
static void layout_symtab(struct module *mod, struct load_info *info)
2482
{
2483 2484
	Elf_Shdr *symsect = info->sechdrs + info->index.sym;
	Elf_Shdr *strsect = info->sechdrs + info->index.str;
2485
	const Elf_Sym *src;
2486
	unsigned int i, nsrc, ndst, strtab_size = 0;
2487 2488 2489 2490

	/* Put symbol section at end of init part of module. */
	symsect->sh_flags |= SHF_ALLOC;
	symsect->sh_entsize = get_offset(mod, &mod->init_size, symsect,
2491
					 info->index.sym) | INIT_OFFSET_MASK;
2492
	pr_debug("\t%s\n", info->secstrings + symsect->sh_name);
2493

2494
	src = (void *)info->hdr + symsect->sh_offset;
2495
	nsrc = symsect->sh_size / sizeof(*src);
2496

2497
	/* Compute total space required for the core symbols' strtab. */
2498 2499 2500 2501
	for (ndst = i = 0; i < nsrc; i++) {
		if (i == 0 ||
		    is_core_symbol(src+i, info->sechdrs, info->hdr->e_shnum)) {
			strtab_size += strlen(&info->strtab[src[i].st_name])+1;
2502
			ndst++;
2503
		}
2504
	}
2505 2506

	/* Append room for core symbols at end of core part. */
2507
	info->symoffs = ALIGN(mod->core_size, symsect->sh_addralign ?: 1);
2508 2509
	info->stroffs = mod->core_size = info->symoffs + ndst * sizeof(Elf_Sym);
	mod->core_size += strtab_size;
2510
	mod->core_size = debug_align(mod->core_size);
2511

2512 2513 2514
	/* Put string table section at end of init part of module. */
	strsect->sh_flags |= SHF_ALLOC;
	strsect->sh_entsize = get_offset(mod, &mod->init_size, strsect,
2515
					 info->index.str) | INIT_OFFSET_MASK;
2516
	mod->init_size = debug_align(mod->init_size);
2517
	pr_debug("\t%s\n", info->secstrings + strsect->sh_name);
2518 2519
}

2520
static void add_kallsyms(struct module *mod, const struct load_info *info)
L
Linus Torvalds 已提交
2521
{
2522 2523 2524
	unsigned int i, ndst;
	const Elf_Sym *src;
	Elf_Sym *dst;
2525
	char *s;
2526
	Elf_Shdr *symsec = &info->sechdrs[info->index.sym];
L
Linus Torvalds 已提交
2527

2528 2529
	mod->symtab = (void *)symsec->sh_addr;
	mod->num_symtab = symsec->sh_size / sizeof(Elf_Sym);
2530 2531
	/* Make sure we get permanent strtab: don't use info->strtab. */
	mod->strtab = (void *)info->sechdrs[info->index.str].sh_addr;
L
Linus Torvalds 已提交
2532 2533 2534

	/* Set types up while we still have access to sections. */
	for (i = 0; i < mod->num_symtab; i++)
2535
		mod->symtab[i].st_info = elf_type(&mod->symtab[i], info);
2536

R
Rusty Russell 已提交
2537
	mod->core_symtab = dst = mod->module_core + info->symoffs;
2538
	mod->core_strtab = s = mod->module_core + info->stroffs;
2539
	src = mod->symtab;
2540 2541 2542 2543 2544 2545 2546 2547
	for (ndst = i = 0; i < mod->num_symtab; i++) {
		if (i == 0 ||
		    is_core_symbol(src+i, info->sechdrs, info->hdr->e_shnum)) {
			dst[ndst] = src[i];
			dst[ndst++].st_name = s - mod->core_strtab;
			s += strlcpy(s, &mod->strtab[src[i].st_name],
				     KSYM_NAME_LEN) + 1;
		}
2548 2549
	}
	mod->core_num_syms = ndst;
L
Linus Torvalds 已提交
2550 2551
}
#else
2552
static inline void layout_symtab(struct module *mod, struct load_info *info)
2553 2554
{
}
2555

2556
static void add_kallsyms(struct module *mod, const struct load_info *info)
L
Linus Torvalds 已提交
2557 2558 2559 2560
{
}
#endif /* CONFIG_KALLSYMS */

2561
static void dynamic_debug_setup(struct _ddebug *debug, unsigned int num)
2562
{
2563 2564
	if (!debug)
		return;
2565 2566
#ifdef CONFIG_DYNAMIC_DEBUG
	if (ddebug_add_module(debug, num, debug->modname))
A
Andrew Morton 已提交
2567 2568
		pr_err("dynamic debug error adding module: %s\n",
			debug->modname);
2569
#endif
R
Rusty Russell 已提交
2570
}
2571

2572 2573 2574 2575 2576 2577
static void dynamic_debug_remove(struct _ddebug *debug)
{
	if (debug)
		ddebug_remove_module(debug->modname);
}

2578 2579
void * __weak module_alloc(unsigned long size)
{
R
Rusty Russell 已提交
2580
	return vmalloc_exec(size);
2581 2582
}

C
Catalin Marinas 已提交
2583
#ifdef CONFIG_DEBUG_KMEMLEAK
2584 2585
static void kmemleak_load_module(const struct module *mod,
				 const struct load_info *info)
C
Catalin Marinas 已提交
2586 2587 2588 2589
{
	unsigned int i;

	/* only scan the sections containing data */
2590
	kmemleak_scan_area(mod, sizeof(struct module), GFP_KERNEL);
C
Catalin Marinas 已提交
2591

2592
	for (i = 1; i < info->hdr->e_shnum; i++) {
2593 2594 2595 2596
		/* 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))
C
Catalin Marinas 已提交
2597 2598
			continue;

2599 2600
		kmemleak_scan_area((void *)info->sechdrs[i].sh_addr,
				   info->sechdrs[i].sh_size, GFP_KERNEL);
C
Catalin Marinas 已提交
2601 2602 2603
	}
}
#else
2604 2605
static inline void kmemleak_load_module(const struct module *mod,
					const struct load_info *info)
C
Catalin Marinas 已提交
2606 2607 2608 2609
{
}
#endif

R
Rusty Russell 已提交
2610
#ifdef CONFIG_MODULE_SIG
2611
static int module_sig_check(struct load_info *info)
R
Rusty Russell 已提交
2612 2613
{
	int err = -ENOKEY;
2614 2615
	const unsigned long markerlen = sizeof(MODULE_SIG_STRING) - 1;
	const void *mod = info->hdr;
2616

2617 2618
	if (info->len > markerlen &&
	    memcmp(mod + info->len - markerlen, MODULE_SIG_STRING, markerlen) == 0) {
2619
		/* We truncate the module to discard the signature */
2620 2621
		info->len -= markerlen;
		err = mod_verify_sig(mod, &info->len);
R
Rusty Russell 已提交
2622 2623 2624 2625 2626 2627 2628 2629 2630 2631 2632 2633 2634 2635
	}

	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 */
2636
static int module_sig_check(struct load_info *info)
R
Rusty Russell 已提交
2637 2638 2639 2640 2641
{
	return 0;
}
#endif /* !CONFIG_MODULE_SIG */

2642 2643
/* Sanity checks against invalid binaries, wrong arch, weird elf version. */
static int elf_header_check(struct load_info *info)
2644
{
2645 2646 2647 2648 2649 2650 2651 2652 2653 2654 2655 2656 2657
	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;
2658

2659 2660 2661
	return 0;
}

2662 2663 2664 2665 2666 2667 2668 2669 2670 2671 2672 2673 2674 2675 2676 2677 2678
#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;
}

2679 2680 2681
/* Sets info->hdr and info->len. */
static int copy_module_from_user(const void __user *umod, unsigned long len,
				  struct load_info *info)
2682 2683 2684
{
	int err;

2685 2686
	info->len = len;
	if (info->len < sizeof(*(info->hdr)))
2687 2688
		return -ENOEXEC;

2689 2690 2691 2692
	err = security_kernel_module_from_file(NULL);
	if (err)
		return err;

2693
	/* Suck in entire file: we'll want most of it. */
2694 2695
	info->hdr = __vmalloc(info->len,
			GFP_KERNEL | __GFP_HIGHMEM | __GFP_NOWARN, PAGE_KERNEL);
2696
	if (!info->hdr)
2697 2698
		return -ENOMEM;

2699
	if (copy_chunked_from_user(info->hdr, umod, info->len) != 0) {
2700 2701
		vfree(info->hdr);
		return -EFAULT;
2702 2703
	}

2704 2705 2706 2707 2708 2709
	return 0;
}

/* Sets info->hdr and info->len. */
static int copy_module_from_fd(int fd, struct load_info *info)
{
A
Al Viro 已提交
2710
	struct fd f = fdget(fd);
2711 2712 2713 2714 2715
	int err;
	struct kstat stat;
	loff_t pos;
	ssize_t bytes = 0;

A
Al Viro 已提交
2716
	if (!f.file)
2717 2718
		return -ENOEXEC;

A
Al Viro 已提交
2719
	err = security_kernel_module_from_file(f.file);
R
Rusty Russell 已提交
2720
	if (err)
2721
		goto out;
R
Rusty Russell 已提交
2722

A
Al Viro 已提交
2723
	err = vfs_getattr(&f.file->f_path, &stat);
R
Rusty Russell 已提交
2724
	if (err)
2725
		goto out;
2726

2727 2728 2729
	if (stat.size > INT_MAX) {
		err = -EFBIG;
		goto out;
2730
	}
2731 2732 2733 2734 2735 2736 2737

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

2738 2739 2740 2741
	info->hdr = vmalloc(stat.size);
	if (!info->hdr) {
		err = -ENOMEM;
		goto out;
2742
	}
R
Rusty Russell 已提交
2743

2744 2745
	pos = 0;
	while (pos < stat.size) {
A
Al Viro 已提交
2746
		bytes = kernel_read(f.file, pos, (char *)(info->hdr) + pos,
2747 2748 2749 2750 2751 2752 2753 2754 2755 2756 2757
				    stat.size - pos);
		if (bytes < 0) {
			vfree(info->hdr);
			err = bytes;
			goto out;
		}
		if (bytes == 0)
			break;
		pos += bytes;
	}
	info->len = pos;
2758

2759
out:
A
Al Viro 已提交
2760
	fdput(f);
2761 2762 2763
	return err;
}

R
Rusty Russell 已提交
2764 2765 2766 2767 2768
static void free_copy(struct load_info *info)
{
	vfree(info->hdr);
}

2769
static int rewrite_section_headers(struct load_info *info, int flags)
2770 2771 2772 2773 2774 2775 2776 2777 2778 2779
{
	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) {
A
Andrew Morton 已提交
2780
			pr_err("Module len %lu truncated\n", info->len);
2781 2782 2783 2784 2785 2786 2787 2788 2789 2790 2791 2792 2793
			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
	}
2794 2795

	/* Track but don't keep modinfo and version sections. */
2796 2797 2798 2799
	if (flags & MODULE_INIT_IGNORE_MODVERSIONS)
		info->index.vers = 0; /* Pretend no __versions section! */
	else
		info->index.vers = find_sec(info, "__versions");
2800
	info->index.info = find_sec(info, ".modinfo");
2801 2802
	info->sechdrs[info->index.info].sh_flags &= ~(unsigned long)SHF_ALLOC;
	info->sechdrs[info->index.vers].sh_flags &= ~(unsigned long)SHF_ALLOC;
2803 2804 2805
	return 0;
}

L
Linus Torvalds 已提交
2806 2807 2808 2809 2810 2811 2812 2813
/*
 * 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).
 */
2814
static struct module *setup_load_info(struct load_info *info, int flags)
L
Linus Torvalds 已提交
2815 2816
{
	unsigned int i;
2817
	int err;
L
Linus Torvalds 已提交
2818 2819 2820 2821
	struct module *mod;

	/* Set up the convenience variables */
	info->sechdrs = (void *)info->hdr + info->hdr->e_shoff;
2822 2823
	info->secstrings = (void *)info->hdr
		+ info->sechdrs[info->hdr->e_shstrndx].sh_offset;
L
Linus Torvalds 已提交
2824

2825
	err = rewrite_section_headers(info, flags);
2826 2827
	if (err)
		return ERR_PTR(err);
L
Linus Torvalds 已提交
2828

2829 2830
	/* Find internal symbols and strings. */
	for (i = 1; i < info->hdr->e_shnum; i++) {
L
Linus Torvalds 已提交
2831 2832 2833
		if (info->sechdrs[i].sh_type == SHT_SYMTAB) {
			info->index.sym = i;
			info->index.str = info->sechdrs[i].sh_link;
2834 2835 2836
			info->strtab = (char *)info->hdr
				+ info->sechdrs[info->index.str].sh_offset;
			break;
L
Linus Torvalds 已提交
2837 2838 2839
		}
	}

2840
	info->index.mod = find_sec(info, ".gnu.linkonce.this_module");
L
Linus Torvalds 已提交
2841
	if (!info->index.mod) {
A
Andrew Morton 已提交
2842
		pr_warn("No module found in object\n");
L
Linus Torvalds 已提交
2843 2844 2845 2846 2847 2848
		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) {
A
Andrew Morton 已提交
2849
		pr_warn("%s: module has no symbols (stripped?)\n", mod->name);
L
Linus Torvalds 已提交
2850 2851 2852
		return ERR_PTR(-ENOEXEC);
	}

2853
	info->index.pcpu = find_pcpusec(info);
L
Linus Torvalds 已提交
2854 2855 2856 2857 2858 2859 2860 2861

	/* 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;
}

2862
static int check_modinfo(struct module *mod, struct load_info *info, int flags)
2863
{
2864
	const char *modmagic = get_modinfo(info, "vermagic");
2865 2866
	int err;

2867 2868 2869
	if (flags & MODULE_INIT_IGNORE_VERMAGIC)
		modmagic = NULL;

2870 2871 2872 2873 2874
	/* This is allowed: modprobe --force will invalidate it. */
	if (!modmagic) {
		err = try_to_force_load(mod, "bad vermagic");
		if (err)
			return err;
2875
	} else if (!same_magic(modmagic, vermagic, info->index.vers)) {
A
Andrew Morton 已提交
2876
		pr_err("%s: version magic '%s' should be '%s'\n",
2877 2878 2879 2880
		       mod->name, modmagic, vermagic);
		return -ENOEXEC;
	}

2881
	if (!get_modinfo(info, "intree"))
2882
		add_taint_module(mod, TAINT_OOT_MODULE, LOCKDEP_STILL_OK);
2883

2884
	if (get_modinfo(info, "staging")) {
2885
		add_taint_module(mod, TAINT_CRAP, LOCKDEP_STILL_OK);
A
Andrew Morton 已提交
2886 2887
		pr_warn("%s: module is from the staging directory, the quality "
			"is unknown, you have been warned.\n", mod->name);
2888
	}
2889 2890

	/* Set up license info based on the info section */
2891
	set_license(mod, get_modinfo(info, "license"));
2892

2893 2894 2895
	return 0;
}

2896
static int find_module_sections(struct module *mod, struct load_info *info)
L
Linus Torvalds 已提交
2897
{
2898
	mod->kp = section_objs(info, "__param",
L
Linus Torvalds 已提交
2899
			       sizeof(*mod->kp), &mod->num_kp);
2900
	mod->syms = section_objs(info, "__ksymtab",
L
Linus Torvalds 已提交
2901
				 sizeof(*mod->syms), &mod->num_syms);
2902 2903
	mod->crcs = section_addr(info, "__kcrctab");
	mod->gpl_syms = section_objs(info, "__ksymtab_gpl",
L
Linus Torvalds 已提交
2904 2905
				     sizeof(*mod->gpl_syms),
				     &mod->num_gpl_syms);
2906 2907
	mod->gpl_crcs = section_addr(info, "__kcrctab_gpl");
	mod->gpl_future_syms = section_objs(info,
L
Linus Torvalds 已提交
2908 2909 2910
					    "__ksymtab_gpl_future",
					    sizeof(*mod->gpl_future_syms),
					    &mod->num_gpl_future_syms);
2911
	mod->gpl_future_crcs = section_addr(info, "__kcrctab_gpl_future");
L
Linus Torvalds 已提交
2912 2913

#ifdef CONFIG_UNUSED_SYMBOLS
2914
	mod->unused_syms = section_objs(info, "__ksymtab_unused",
L
Linus Torvalds 已提交
2915 2916
					sizeof(*mod->unused_syms),
					&mod->num_unused_syms);
2917 2918
	mod->unused_crcs = section_addr(info, "__kcrctab_unused");
	mod->unused_gpl_syms = section_objs(info, "__ksymtab_unused_gpl",
L
Linus Torvalds 已提交
2919 2920
					    sizeof(*mod->unused_gpl_syms),
					    &mod->num_unused_gpl_syms);
2921
	mod->unused_gpl_crcs = section_addr(info, "__kcrctab_unused_gpl");
L
Linus Torvalds 已提交
2922 2923
#endif
#ifdef CONFIG_CONSTRUCTORS
2924
	mod->ctors = section_objs(info, ".ctors",
L
Linus Torvalds 已提交
2925
				  sizeof(*mod->ctors), &mod->num_ctors);
2926 2927 2928 2929 2930 2931 2932 2933
	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.
		 */
2934
		pr_warn("%s: has both .ctors and .init_array.\n",
2935 2936 2937
		       mod->name);
		return -EINVAL;
	}
L
Linus Torvalds 已提交
2938 2939 2940
#endif

#ifdef CONFIG_TRACEPOINTS
2941 2942 2943
	mod->tracepoints_ptrs = section_objs(info, "__tracepoints_ptrs",
					     sizeof(*mod->tracepoints_ptrs),
					     &mod->num_tracepoints);
L
Linus Torvalds 已提交
2944
#endif
2945 2946 2947 2948 2949
#ifdef HAVE_JUMP_LABEL
	mod->jump_entries = section_objs(info, "__jump_table",
					sizeof(*mod->jump_entries),
					&mod->num_jump_entries);
#endif
L
Linus Torvalds 已提交
2950
#ifdef CONFIG_EVENT_TRACING
2951
	mod->trace_events = section_objs(info, "_ftrace_events",
L
Linus Torvalds 已提交
2952 2953
					 sizeof(*mod->trace_events),
					 &mod->num_trace_events);
2954 2955 2956
	mod->trace_enums = section_objs(info, "_ftrace_enum_map",
					sizeof(*mod->trace_enums),
					&mod->num_trace_enums);
L
Linus Torvalds 已提交
2957
#endif
2958 2959 2960 2961 2962
#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
L
Linus Torvalds 已提交
2963 2964
#ifdef CONFIG_FTRACE_MCOUNT_RECORD
	/* sechdrs[0].sh_size is always zero */
2965
	mod->ftrace_callsites = section_objs(info, "__mcount_loc",
L
Linus Torvalds 已提交
2966 2967 2968
					     sizeof(*mod->ftrace_callsites),
					     &mod->num_ftrace_callsites);
#endif
2969

2970 2971 2972
	mod->extable = section_objs(info, "__ex_table",
				    sizeof(*mod->extable), &mod->num_exentries);

2973
	if (section_addr(info, "__obsparm"))
A
Andrew Morton 已提交
2974
		pr_warn("%s: Ignoring obsolete parameters\n", mod->name);
2975 2976 2977

	info->debug = section_objs(info, "__verbose",
				   sizeof(*info->debug), &info->num_debug);
2978 2979

	return 0;
L
Linus Torvalds 已提交
2980 2981
}

2982
static int move_module(struct module *mod, struct load_info *info)
2983 2984 2985 2986 2987
{
	int i;
	void *ptr;

	/* Do the allocs. */
2988
	ptr = module_alloc(mod->core_size);
2989 2990 2991 2992 2993 2994 2995
	/*
	 * 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)
R
Rusty Russell 已提交
2996
		return -ENOMEM;
2997 2998 2999 3000

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

R
Rusty Russell 已提交
3001
	if (mod->init_size) {
3002
		ptr = module_alloc(mod->init_size);
R
Rusty Russell 已提交
3003 3004 3005 3006 3007 3008 3009 3010
		/*
		 * 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) {
3011
			module_memfree(mod->module_core);
R
Rusty Russell 已提交
3012 3013 3014 3015 3016 3017
			return -ENOMEM;
		}
		memset(ptr, 0, mod->init_size);
		mod->module_init = ptr;
	} else
		mod->module_init = NULL;
3018 3019

	/* Transfer each section which specifies SHF_ALLOC */
3020
	pr_debug("final section addresses:\n");
3021
	for (i = 0; i < info->hdr->e_shnum; i++) {
3022
		void *dest;
3023
		Elf_Shdr *shdr = &info->sechdrs[i];
3024

3025
		if (!(shdr->sh_flags & SHF_ALLOC))
3026 3027
			continue;

3028
		if (shdr->sh_entsize & INIT_OFFSET_MASK)
3029
			dest = mod->module_init
3030
				+ (shdr->sh_entsize & ~INIT_OFFSET_MASK);
3031
		else
3032
			dest = mod->module_core + shdr->sh_entsize;
3033

3034 3035
		if (shdr->sh_type != SHT_NOBITS)
			memcpy(dest, (void *)shdr->sh_addr, shdr->sh_size);
3036
		/* Update sh_addr to point to copy in image. */
3037
		shdr->sh_addr = (unsigned long)dest;
3038 3039
		pr_debug("\t0x%lx %s\n",
			 (long)shdr->sh_addr, info->secstrings + shdr->sh_name);
3040
	}
R
Rusty Russell 已提交
3041 3042

	return 0;
3043 3044
}

3045
static int check_module_license_and_versions(struct module *mod)
3046 3047 3048 3049 3050 3051 3052
{
	/*
	 * 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)
3053
		add_taint(TAINT_PROPRIETARY_MODULE, LOCKDEP_NOW_UNRELIABLE);
3054 3055 3056

	/* driverloader was caught wrongly pretending to be under GPL */
	if (strcmp(mod->name, "driverloader") == 0)
3057 3058
		add_taint_module(mod, TAINT_PROPRIETARY_MODULE,
				 LOCKDEP_NOW_UNRELIABLE);
3059

3060 3061
	/* lve claims to be GPL but upstream won't provide source */
	if (strcmp(mod->name, "lve") == 0)
3062 3063
		add_taint_module(mod, TAINT_PROPRIETARY_MODULE,
				 LOCKDEP_NOW_UNRELIABLE);
3064

3065 3066 3067 3068 3069 3070 3071 3072 3073 3074 3075 3076 3077 3078 3079 3080 3081 3082 3083 3084 3085 3086 3087 3088 3089 3090 3091 3092 3093 3094 3095 3096 3097 3098 3099 3100 3101 3102 3103
#ifdef CONFIG_MODVERSIONS
	if ((mod->num_syms && !mod->crcs)
	    || (mod->num_gpl_syms && !mod->gpl_crcs)
	    || (mod->num_gpl_future_syms && !mod->gpl_future_crcs)
#ifdef CONFIG_UNUSED_SYMBOLS
	    || (mod->num_unused_syms && !mod->unused_crcs)
	    || (mod->num_unused_gpl_syms && !mod->unused_gpl_crcs)
#endif
		) {
		return try_to_force_load(mod,
					 "no versions for exported symbols");
	}
#endif
	return 0;
}

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

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

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

	set_fs(old_fs);
}

3104 3105 3106 3107 3108 3109 3110 3111
int __weak module_frob_arch_sections(Elf_Ehdr *hdr,
				     Elf_Shdr *sechdrs,
				     char *secstrings,
				     struct module *mod)
{
	return 0;
}

3112
static struct module *layout_and_allocate(struct load_info *info, int flags)
L
Linus Torvalds 已提交
3113
{
R
Rusty Russell 已提交
3114
	/* Module within temporary copy. */
L
Linus Torvalds 已提交
3115
	struct module *mod;
R
Rusty Russell 已提交
3116
	int err;
3117

3118
	mod = setup_load_info(info, flags);
R
Rusty Russell 已提交
3119 3120
	if (IS_ERR(mod))
		return mod;
L
Linus Torvalds 已提交
3121

3122
	err = check_modinfo(mod, info, flags);
3123 3124
	if (err)
		return ERR_PTR(err);
L
Linus Torvalds 已提交
3125 3126

	/* Allow arches to frob section contents and sizes.  */
3127 3128
	err = module_frob_arch_sections(info->hdr, info->sechdrs,
					info->secstrings, mod);
L
Linus Torvalds 已提交
3129
	if (err < 0)
3130
		return ERR_PTR(err);
L
Linus Torvalds 已提交
3131

3132 3133
	/* 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
3134 3135 3136 3137

	/* 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. */
3138 3139
	layout_sections(mod, info);
	layout_symtab(mod, info);
L
Linus Torvalds 已提交
3140

3141
	/* Allocate and move to the final place */
3142
	err = move_module(mod, info);
R
Rusty Russell 已提交
3143
	if (err)
3144
		return ERR_PTR(err);
R
Rusty Russell 已提交
3145 3146 3147

	/* Module has been copied to its final place now: return it. */
	mod = (void *)info->sechdrs[info->index.mod].sh_addr;
3148
	kmemleak_load_module(mod, info);
R
Rusty Russell 已提交
3149 3150 3151 3152 3153 3154 3155
	return mod;
}

/* mod is no longer valid after this! */
static void module_deallocate(struct module *mod, struct load_info *info)
{
	percpu_modfree(mod);
3156
	module_arch_freeing_init(mod);
3157 3158
	module_memfree(mod->module_init);
	module_memfree(mod->module_core);
R
Rusty Russell 已提交
3159 3160
}

3161 3162 3163 3164 3165 3166 3167
int __weak module_finalize(const Elf_Ehdr *hdr,
			   const Elf_Shdr *sechdrs,
			   struct module *me)
{
	return 0;
}

3168 3169
static int post_relocation(struct module *mod, const struct load_info *info)
{
3170
	/* Sort exception table now relocations are done. */
3171 3172 3173 3174 3175 3176
	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);

3177
	/* Setup kallsyms-specific fields. */
3178 3179 3180 3181 3182 3183
	add_kallsyms(mod, info);

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

3184 3185 3186 3187 3188 3189
/* Is this module of this name done loading?  No locks held. */
static bool finished_loading(const char *name)
{
	struct module *mod;
	bool ret;

3190 3191 3192 3193 3194 3195
	/*
	 * 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();
3196
	mutex_lock(&module_mutex);
3197
	mod = find_module_all(name, strlen(name), true);
3198 3199
	ret = !mod || mod->state == MODULE_STATE_LIVE
		|| mod->state == MODULE_STATE_GOING;
3200 3201 3202 3203 3204
	mutex_unlock(&module_mutex);

	return ret;
}

3205 3206 3207 3208 3209 3210 3211 3212 3213 3214 3215
/* 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
}

3216 3217 3218 3219 3220 3221 3222 3223 3224 3225 3226 3227 3228
/* 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);
}

3229 3230 3231 3232 3233 3234 3235
/*
 * 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)
3236 3237
{
	int ret = 0;
3238 3239 3240 3241 3242 3243 3244 3245
	struct mod_initfree *freeinit;

	freeinit = kmalloc(sizeof(*freeinit), GFP_KERNEL);
	if (!freeinit) {
		ret = -ENOMEM;
		goto fail;
	}
	freeinit->module_init = mod->module_init;
3246

3247 3248 3249 3250 3251 3252
	/*
	 * 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;

3253 3254 3255 3256 3257
	do_mod_ctors(mod);
	/* Start the module */
	if (mod->init != NULL)
		ret = do_one_initcall(mod->init);
	if (ret < 0) {
3258
		goto fail_free_freeinit;
3259 3260
	}
	if (ret > 0) {
A
Andrew Morton 已提交
3261 3262 3263 3264
		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__);
3265 3266 3267 3268 3269 3270 3271 3272
		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);

3273 3274 3275 3276 3277 3278 3279 3280 3281 3282 3283 3284 3285 3286 3287 3288 3289 3290 3291
	/*
	 * We need to finish all async code before the module init sequence
	 * is done.  This has potential to deadlock.  For example, a newly
	 * detected block device can trigger request_module() of the
	 * default iosched from async probing task.  Once userland helper
	 * reaches here, async_synchronize_full() will wait on the async
	 * task waiting on request_module() and deadlock.
	 *
	 * This deadlock is avoided by perfomring async_synchronize_full()
	 * iff module init queued any async jobs.  This isn't a full
	 * solution as it will deadlock the same if module loading from
	 * async jobs nests more than once; however, due to the various
	 * constraints, this hack seems to be the best option for now.
	 * Please refer to the following thread for details.
	 *
	 * http://thread.gmane.org/gmane.linux.kernel/1420814
	 */
	if (current->flags & PF_USED_ASYNC)
		async_synchronize_full();
3292 3293 3294 3295 3296 3297 3298 3299 3300 3301

	mutex_lock(&module_mutex);
	/* Drop initial reference. */
	module_put(mod);
	trim_init_extable(mod);
#ifdef CONFIG_KALLSYMS
	mod->num_symtab = mod->core_num_syms;
	mod->symtab = mod->core_symtab;
	mod->strtab = mod->core_strtab;
#endif
3302
	mod_tree_remove_init(mod);
3303
	unset_module_init_ro_nx(mod);
3304
	module_arch_freeing_init(mod);
3305 3306 3307 3308
	mod->module_init = NULL;
	mod->init_size = 0;
	mod->init_ro_size = 0;
	mod->init_text_size = 0;
3309 3310
	/*
	 * We want to free module_init, but be aware that kallsyms may be
3311 3312 3313
	 * 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.
3314
	 */
3315
	call_rcu_sched(&freeinit->rcu, do_free_init);
3316 3317 3318 3319
	mutex_unlock(&module_mutex);
	wake_up_all(&module_wq);

	return 0;
3320 3321 3322 3323 3324 3325 3326 3327 3328 3329 3330 3331 3332

fail_free_freeinit:
	kfree(freeinit);
fail:
	/* Try to protect us from buggy refcounters. */
	mod->state = MODULE_STATE_GOING;
	synchronize_sched();
	module_put(mod);
	blocking_notifier_call_chain(&module_notify_list,
				     MODULE_STATE_GOING, mod);
	free_module(mod);
	wake_up_all(&module_wq);
	return ret;
3333 3334 3335 3336 3337 3338 3339 3340 3341 3342
}

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

	return 0;
}

3343 3344 3345 3346 3347 3348 3349 3350 3351 3352 3353 3354 3355 3356
/*
 * 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);
3357 3358
	old = find_module_all(mod->name, strlen(mod->name), true);
	if (old != NULL) {
3359 3360 3361 3362
		if (old->state == MODULE_STATE_COMING
		    || old->state == MODULE_STATE_UNFORMED) {
			/* Wait in case it fails to load. */
			mutex_unlock(&module_mutex);
3363 3364
			err = wait_event_interruptible(module_wq,
					       finished_loading(mod->name));
3365 3366 3367 3368 3369 3370 3371
			if (err)
				goto out_unlocked;
			goto again;
		}
		err = -EEXIST;
		goto out;
	}
3372
	mod_update_bounds(mod);
3373
	list_add_rcu(&mod->list, &modules);
3374
	mod_tree_insert(mod);
3375 3376 3377 3378 3379 3380 3381 3382 3383 3384 3385 3386 3387 3388 3389 3390 3391 3392 3393 3394 3395 3396
	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);

3397 3398 3399 3400 3401 3402 3403 3404 3405 3406 3407 3408
	/* Set RO and NX regions for core */
	set_section_ro_nx(mod->module_core,
				mod->core_text_size,
				mod->core_ro_size,
				mod->core_size);

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

3409 3410 3411
	/* Mark state as coming so strong_try_module_get() ignores us,
	 * but kallsyms etc. can see us. */
	mod->state = MODULE_STATE_COMING;
3412 3413 3414 3415 3416
	mutex_unlock(&module_mutex);

	blocking_notifier_call_chain(&module_notify_list,
				     MODULE_STATE_COMING, mod);
	return 0;
3417 3418 3419 3420 3421 3422

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

3423 3424
static int unknown_module_param_cb(char *param, char *val, const char *modname)
{
3425
	/* Check for magic 'dyndbg' arg */
3426
	int ret = ddebug_dyndbg_module_param_cb(param, val, modname);
A
Andrew Morton 已提交
3427 3428
	if (ret != 0)
		pr_warn("%s: unknown parameter '%s' ignored\n", modname, param);
3429 3430 3431
	return 0;
}

R
Rusty Russell 已提交
3432 3433
/* Allocate and load the module: note that size of section 0 is always
   zero, and we rely on this for optional sections. */
3434 3435
static int load_module(struct load_info *info, const char __user *uargs,
		       int flags)
R
Rusty Russell 已提交
3436
{
3437
	struct module *mod;
R
Rusty Russell 已提交
3438
	long err;
3439
	char *after_dashes;
R
Rusty Russell 已提交
3440

3441 3442 3443
	err = module_sig_check(info);
	if (err)
		goto free_copy;
R
Rusty Russell 已提交
3444

3445
	err = elf_header_check(info);
R
Rusty Russell 已提交
3446
	if (err)
3447
		goto free_copy;
R
Rusty Russell 已提交
3448 3449

	/* Figure out module layout, and allocate all the memory. */
3450
	mod = layout_and_allocate(info, flags);
3451 3452
	if (IS_ERR(mod)) {
		err = PTR_ERR(mod);
R
Rusty Russell 已提交
3453
		goto free_copy;
L
Linus Torvalds 已提交
3454 3455
	}

3456 3457 3458
	/* Reserve our place in the list. */
	err = add_unformed_module(mod);
	if (err)
3459 3460
		goto free_module;

R
Rusty Russell 已提交
3461
#ifdef CONFIG_MODULE_SIG
3462
	mod->sig_ok = info->sig_ok;
3463
	if (!mod->sig_ok) {
A
Andrew Morton 已提交
3464
		pr_notice_once("%s: module verification failed: signature "
3465
			       "and/or required key missing - tainting "
A
Andrew Morton 已提交
3466
			       "kernel\n", mod->name);
3467
		add_taint_module(mod, TAINT_UNSIGNED_MODULE, LOCKDEP_STILL_OK);
3468
	}
R
Rusty Russell 已提交
3469 3470
#endif

3471
	/* To avoid stressing percpu allocator, do this once we're unique. */
R
Rusty Russell 已提交
3472
	err = percpu_modalloc(mod, info);
3473 3474 3475
	if (err)
		goto unlink_mod;

3476
	/* Now module is in final location, initialize linked lists, etc. */
3477 3478
	err = module_unload_init(mod);
	if (err)
3479
		goto unlink_mod;
L
Linus Torvalds 已提交
3480

3481 3482
	/* Now we've got everything in the final locations, we can
	 * find optional sections. */
3483 3484 3485
	err = find_module_sections(mod, info);
	if (err)
		goto free_unload;
3486

3487
	err = check_module_license_and_versions(mod);
3488 3489
	if (err)
		goto free_unload;
3490

3491
	/* Set up MODINFO_ATTR fields */
3492
	setup_modinfo(mod, info);
3493

L
Linus Torvalds 已提交
3494
	/* Fix up syms, so that st_value is a pointer to location. */
3495
	err = simplify_symbols(mod, info);
L
Linus Torvalds 已提交
3496
	if (err < 0)
R
Rusty Russell 已提交
3497
		goto free_modinfo;
L
Linus Torvalds 已提交
3498

3499
	err = apply_relocations(mod, info);
3500
	if (err < 0)
R
Rusty Russell 已提交
3501
		goto free_modinfo;
L
Linus Torvalds 已提交
3502

3503
	err = post_relocation(mod, info);
L
Linus Torvalds 已提交
3504
	if (err < 0)
R
Rusty Russell 已提交
3505
		goto free_modinfo;
L
Linus Torvalds 已提交
3506

3507
	flush_module_icache(mod);
3508

3509 3510 3511 3512 3513 3514
	/* 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;
	}
R
Rusty Russell 已提交
3515

3516
	dynamic_debug_setup(info->debug, info->num_debug);
3517

3518 3519 3520
	/* Ftrace init must be called in the MODULE_STATE_UNFORMED state */
	ftrace_module_init(mod);

3521 3522 3523
	/* Finally it's fully formed, ready to start executing. */
	err = complete_formation(mod, info);
	if (err)
3524
		goto ddebug_cleanup;
3525

3526
	/* Module is ready to execute: parsing args may do that. */
3527 3528 3529 3530
	after_dashes = parse_args(mod->name, mod->args, mod->kp, mod->num_kp,
				  -32768, 32767, unknown_module_param_cb);
	if (IS_ERR(after_dashes)) {
		err = PTR_ERR(after_dashes);
3531
		goto bug_cleanup;
3532 3533 3534 3535
	} else if (after_dashes) {
		pr_warn("%s: parameters '%s' after `--' ignored\n",
		       mod->name, after_dashes);
	}
L
Linus Torvalds 已提交
3536

3537
	/* Link in to syfs. */
3538
	err = mod_sysfs_setup(mod, info, mod->kp, mod->num_kp);
L
Linus Torvalds 已提交
3539
	if (err < 0)
3540
		goto bug_cleanup;
3541

3542
	/* Get rid of temporary copy. */
3543
	free_copy(info);
L
Linus Torvalds 已提交
3544 3545

	/* Done! */
3546
	trace_module_load(mod);
3547 3548

	return do_init_module(mod);
L
Linus Torvalds 已提交
3549

3550 3551
 bug_cleanup:
	/* module_bug_cleanup needs module_mutex protection */
3552
	mutex_lock(&module_mutex);
3553
	module_bug_cleanup(mod);
3554
	mutex_unlock(&module_mutex);
3555 3556 3557 3558 3559

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

3560
 ddebug_cleanup:
3561
	dynamic_debug_remove(info->debug);
3562
	synchronize_sched();
3563 3564
	kfree(mod->args);
 free_arch_cleanup:
L
Linus Torvalds 已提交
3565
	module_arch_cleanup(mod);
R
Rusty Russell 已提交
3566
 free_modinfo:
3567
	free_modinfo(mod);
3568
 free_unload:
L
Linus Torvalds 已提交
3569
	module_unload_free(mod);
3570 3571 3572 3573 3574
 unlink_mod:
	mutex_lock(&module_mutex);
	/* Unlink carefully: kallsyms could be walking list. */
	list_del_rcu(&mod->list);
	wake_up_all(&module_wq);
3575 3576
	/* Wait for RCU-sched synchronizing before releasing mod->list. */
	synchronize_sched();
3577
	mutex_unlock(&module_mutex);
R
Rusty Russell 已提交
3578
 free_module:
3579 3580 3581
	/* Free lock-classes; relies on the preceding sync_rcu() */
	lockdep_free_key_range(mod->module_core, mod->core_size);

3582
	module_deallocate(mod, info);
R
Rusty Russell 已提交
3583
 free_copy:
3584 3585
	free_copy(info);
	return err;
3586 3587
}

3588 3589
SYSCALL_DEFINE3(init_module, void __user *, umod,
		unsigned long, len, const char __user *, uargs)
L
Linus Torvalds 已提交
3590
{
3591 3592
	int err;
	struct load_info info = { };
L
Linus Torvalds 已提交
3593

3594 3595 3596
	err = may_init_module();
	if (err)
		return err;
L
Linus Torvalds 已提交
3597

3598 3599
	pr_debug("init_module: umod=%p, len=%lu, uargs=%p\n",
	       umod, len, uargs);
L
Linus Torvalds 已提交
3600

3601 3602 3603
	err = copy_module_from_user(umod, len, &info);
	if (err)
		return err;
L
Linus Torvalds 已提交
3604

3605
	return load_module(&info, uargs, 0);
3606
}
3607

3608
SYSCALL_DEFINE3(finit_module, int, fd, const char __user *, uargs, int, flags)
3609 3610 3611
{
	int err;
	struct load_info info = { };
3612

3613 3614 3615
	err = may_init_module();
	if (err)
		return err;
L
Linus Torvalds 已提交
3616

3617
	pr_debug("finit_module: fd=%d, uargs=%p, flags=%i\n", fd, uargs, flags);
3618

3619 3620 3621
	if (flags & ~(MODULE_INIT_IGNORE_MODVERSIONS
		      |MODULE_INIT_IGNORE_VERMAGIC))
		return -EINVAL;
3622

3623 3624 3625
	err = copy_module_from_fd(fd, &info);
	if (err)
		return err;
L
Linus Torvalds 已提交
3626

3627
	return load_module(&info, uargs, flags);
L
Linus Torvalds 已提交
3628 3629 3630 3631 3632 3633 3634 3635 3636 3637 3638 3639 3640 3641
}

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 $d.
 */
static inline int is_arm_mapping_symbol(const char *str)
{
3642 3643
	if (str[0] == '.' && str[1] == 'L')
		return true;
K
Kyle McMartin 已提交
3644
	return str[0] == '$' && strchr("axtd", str[1])
L
Linus Torvalds 已提交
3645 3646 3647 3648 3649 3650 3651 3652 3653 3654 3655 3656
	       && (str[2] == '\0' || str[2] == '.');
}

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

	/* At worse, next value is at end of module */
3657
	if (within_module_init(addr, mod))
L
Linus Torvalds 已提交
3658
		nextval = (unsigned long)mod->module_init+mod->init_text_size;
D
Daniel Walker 已提交
3659
	else
L
Linus Torvalds 已提交
3660 3661
		nextval = (unsigned long)mod->module_core+mod->core_text_size;

L
Lucas De Marchi 已提交
3662
	/* Scan for closest preceding symbol, and next symbol. (ELF
D
Daniel Walker 已提交
3663
	   starts real symbols at 1). */
L
Linus Torvalds 已提交
3664 3665 3666 3667 3668 3669 3670 3671 3672 3673 3674 3675 3676 3677 3678 3679 3680 3681 3682 3683 3684
	for (i = 1; i < mod->num_symtab; i++) {
		if (mod->symtab[i].st_shndx == SHN_UNDEF)
			continue;

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

	if (!best)
		return NULL;

A
Alexey Dobriyan 已提交
3685 3686 3687 3688
	if (size)
		*size = nextval - mod->symtab[best].st_value;
	if (offset)
		*offset = addr - mod->symtab[best].st_value;
L
Linus Torvalds 已提交
3689 3690 3691
	return mod->strtab + mod->symtab[best].st_name;
}

3692 3693
/* For kallsyms to ask for address resolution.  NULL means not found.  Careful
 * not to lock to avoid deadlock on oopses, simply disable preemption. */
3694
const char *module_address_lookup(unsigned long addr,
3695 3696 3697 3698
			    unsigned long *size,
			    unsigned long *offset,
			    char **modname,
			    char *namebuf)
L
Linus Torvalds 已提交
3699
{
3700
	const char *ret = NULL;
3701
	struct module *mod;
L
Linus Torvalds 已提交
3702

3703
	preempt_disable();
3704 3705 3706 3707 3708
	mod = __module_address(addr);
	if (mod) {
		if (modname)
			*modname = mod->name;
		ret = get_ksymbol(mod, addr, size, offset);
L
Linus Torvalds 已提交
3709
	}
3710 3711 3712 3713 3714
	/* Make a copy in here where it's safe */
	if (ret) {
		strncpy(namebuf, ret, KSYM_NAME_LEN - 1);
		ret = namebuf;
	}
3715
	preempt_enable();
3716

3717
	return ret;
L
Linus Torvalds 已提交
3718 3719
}

3720 3721 3722 3723
int lookup_module_symbol_name(unsigned long addr, char *symname)
{
	struct module *mod;

3724
	preempt_disable();
3725
	list_for_each_entry_rcu(mod, &modules, list) {
3726 3727
		if (mod->state == MODULE_STATE_UNFORMED)
			continue;
3728
		if (within_module(addr, mod)) {
3729 3730 3731 3732 3733
			const char *sym;

			sym = get_ksymbol(mod, addr, NULL, NULL);
			if (!sym)
				goto out;
3734
			strlcpy(symname, sym, KSYM_NAME_LEN);
3735
			preempt_enable();
3736 3737 3738 3739
			return 0;
		}
	}
out:
3740
	preempt_enable();
3741 3742 3743
	return -ERANGE;
}

3744 3745 3746 3747 3748
int lookup_module_symbol_attrs(unsigned long addr, unsigned long *size,
			unsigned long *offset, char *modname, char *name)
{
	struct module *mod;

3749
	preempt_disable();
3750
	list_for_each_entry_rcu(mod, &modules, list) {
3751 3752
		if (mod->state == MODULE_STATE_UNFORMED)
			continue;
3753
		if (within_module(addr, mod)) {
3754 3755 3756 3757 3758 3759
			const char *sym;

			sym = get_ksymbol(mod, addr, size, offset);
			if (!sym)
				goto out;
			if (modname)
3760
				strlcpy(modname, mod->name, MODULE_NAME_LEN);
3761
			if (name)
3762
				strlcpy(name, sym, KSYM_NAME_LEN);
3763
			preempt_enable();
3764 3765 3766 3767
			return 0;
		}
	}
out:
3768
	preempt_enable();
3769 3770 3771
	return -ERANGE;
}

3772 3773
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
3774 3775 3776
{
	struct module *mod;

3777
	preempt_disable();
3778
	list_for_each_entry_rcu(mod, &modules, list) {
3779 3780
		if (mod->state == MODULE_STATE_UNFORMED)
			continue;
L
Linus Torvalds 已提交
3781 3782 3783
		if (symnum < mod->num_symtab) {
			*value = mod->symtab[symnum].st_value;
			*type = mod->symtab[symnum].st_info;
3784
			strlcpy(name, mod->strtab + mod->symtab[symnum].st_name,
3785 3786
				KSYM_NAME_LEN);
			strlcpy(module_name, mod->name, MODULE_NAME_LEN);
3787
			*exported = is_exported(name, *value, mod);
3788
			preempt_enable();
3789
			return 0;
L
Linus Torvalds 已提交
3790 3791 3792
		}
		symnum -= mod->num_symtab;
	}
3793
	preempt_enable();
3794
	return -ERANGE;
L
Linus Torvalds 已提交
3795 3796 3797 3798 3799 3800 3801
}

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

	for (i = 0; i < mod->num_symtab; i++)
3802 3803
		if (strcmp(name, mod->strtab+mod->symtab[i].st_name) == 0 &&
		    mod->symtab[i].st_info != 'U')
L
Linus Torvalds 已提交
3804 3805 3806 3807 3808 3809 3810 3811 3812 3813 3814 3815
			return mod->symtab[i].st_value;
	return 0;
}

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

	/* Don't lock: we're in enough trouble already. */
3816
	preempt_disable();
L
Linus Torvalds 已提交
3817
	if ((colon = strchr(name, ':')) != NULL) {
3818
		if ((mod = find_module_all(name, colon - name, false)) != NULL)
L
Linus Torvalds 已提交
3819 3820
			ret = mod_find_symname(mod, colon+1);
	} else {
3821 3822 3823
		list_for_each_entry_rcu(mod, &modules, list) {
			if (mod->state == MODULE_STATE_UNFORMED)
				continue;
L
Linus Torvalds 已提交
3824 3825
			if ((ret = mod_find_symname(mod, name)) != 0)
				break;
3826
		}
L
Linus Torvalds 已提交
3827
	}
3828
	preempt_enable();
L
Linus Torvalds 已提交
3829 3830
	return ret;
}
3831 3832 3833 3834 3835 3836 3837 3838 3839

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;

3840 3841
	module_assert_mutex();

3842
	list_for_each_entry(mod, &modules, list) {
3843 3844
		if (mod->state == MODULE_STATE_UNFORMED)
			continue;
3845 3846 3847 3848 3849 3850 3851 3852 3853
		for (i = 0; i < mod->num_symtab; i++) {
			ret = fn(data, mod->strtab + mod->symtab[i].st_name,
				 mod, mod->symtab[i].st_value);
			if (ret != 0)
				return ret;
		}
	}
	return 0;
}
L
Linus Torvalds 已提交
3854 3855
#endif /* CONFIG_KALLSYMS */

3856
static char *module_flags(struct module *mod, char *buf)
3857 3858 3859
{
	int bx = 0;

3860
	BUG_ON(mod->state == MODULE_STATE_UNFORMED);
3861 3862 3863
	if (mod->taints ||
	    mod->state == MODULE_STATE_GOING ||
	    mod->state == MODULE_STATE_COMING) {
3864
		buf[bx++] = '(';
3865
		bx += module_flags_taint(mod, buf + bx);
3866 3867 3868 3869 3870 3871
		/* 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++] = '+';
3872 3873 3874 3875 3876 3877 3878
		buf[bx++] = ')';
	}
	buf[bx] = '\0';

	return buf;
}

3879 3880 3881 3882 3883 3884 3885 3886 3887 3888 3889 3890 3891 3892 3893 3894 3895 3896
#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
3897 3898 3899
static int m_show(struct seq_file *m, void *p)
{
	struct module *mod = list_entry(p, struct module, list);
3900 3901
	char buf[8];

3902 3903 3904 3905
	/* We always ignore unformed modules. */
	if (mod->state == MODULE_STATE_UNFORMED)
		return 0;

3906
	seq_printf(m, "%s %u",
L
Linus Torvalds 已提交
3907 3908 3909 3910 3911
		   mod->name, mod->init_size + mod->core_size);
	print_unload_info(m, mod);

	/* Informative for users. */
	seq_printf(m, " %s",
3912 3913
		   mod->state == MODULE_STATE_GOING ? "Unloading" :
		   mod->state == MODULE_STATE_COMING ? "Loading" :
L
Linus Torvalds 已提交
3914 3915
		   "Live");
	/* Used by oprofile and other similar tools. */
3916
	seq_printf(m, " 0x%pK", mod->module_core);
L
Linus Torvalds 已提交
3917

3918 3919
	/* Taints info */
	if (mod->taints)
3920
		seq_printf(m, " %s", module_flags(mod, buf));
3921

3922
	seq_puts(m, "\n");
L
Linus Torvalds 已提交
3923 3924 3925 3926 3927 3928 3929 3930
	return 0;
}

/* Format: modulename size refcount deps address

   Where refcount is a number or -, and deps is a comma-separated list
   of depends or -.
*/
3931
static const struct seq_operations modules_op = {
L
Linus Torvalds 已提交
3932 3933 3934 3935 3936 3937
	.start	= m_start,
	.next	= m_next,
	.stop	= m_stop,
	.show	= m_show
};

3938 3939 3940 3941 3942 3943 3944 3945 3946 3947 3948 3949 3950 3951 3952 3953 3954 3955 3956 3957
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
3958 3959 3960 3961 3962 3963
/* 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;

R
Rusty Russell 已提交
3964
	preempt_disable();
3965
	list_for_each_entry_rcu(mod, &modules, list) {
3966 3967
		if (mod->state == MODULE_STATE_UNFORMED)
			continue;
L
Linus Torvalds 已提交
3968 3969
		if (mod->num_exentries == 0)
			continue;
D
Daniel Walker 已提交
3970

L
Linus Torvalds 已提交
3971 3972 3973 3974 3975 3976
		e = search_extable(mod->extable,
				   mod->extable + mod->num_exentries - 1,
				   addr);
		if (e)
			break;
	}
R
Rusty Russell 已提交
3977
	preempt_enable();
L
Linus Torvalds 已提交
3978 3979

	/* Now, if we found one, we are running inside it now, hence
D
Daniel Walker 已提交
3980
	   we cannot unload the module, hence no refcnt needed. */
L
Linus Torvalds 已提交
3981 3982 3983
	return e;
}

3984
/*
R
Rusty Russell 已提交
3985 3986 3987 3988 3989
 * 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).
3990
 */
R
Rusty Russell 已提交
3991
bool is_module_address(unsigned long addr)
3992
{
R
Rusty Russell 已提交
3993
	bool ret;
3994

R
Rusty Russell 已提交
3995
	preempt_disable();
R
Rusty Russell 已提交
3996
	ret = __module_address(addr) != NULL;
R
Rusty Russell 已提交
3997
	preempt_enable();
3998

R
Rusty Russell 已提交
3999
	return ret;
4000 4001
}

R
Rusty Russell 已提交
4002 4003 4004 4005 4006 4007 4008
/*
 * __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.
 */
4009
struct module *__module_address(unsigned long addr)
L
Linus Torvalds 已提交
4010 4011 4012
{
	struct module *mod;

4013 4014 4015
	if (addr < module_addr_min || addr > module_addr_max)
		return NULL;

4016 4017
	module_assert_mutex_or_preempt();

4018
	mod = mod_find(addr);
4019 4020
	if (mod) {
		BUG_ON(!within_module(addr, mod));
4021
		if (mod->state == MODULE_STATE_UNFORMED)
4022
			mod = NULL;
4023
	}
4024
	return mod;
L
Linus Torvalds 已提交
4025
}
4026
EXPORT_SYMBOL_GPL(__module_address);
L
Linus Torvalds 已提交
4027

R
Rusty Russell 已提交
4028 4029 4030 4031 4032 4033 4034 4035 4036 4037 4038 4039 4040 4041 4042 4043 4044 4045 4046 4047 4048 4049 4050 4051 4052 4053 4054 4055 4056 4057 4058 4059 4060 4061 4062 4063 4064
/*
 * is_module_text_address - is this address inside module code?
 * @addr: the address to check.
 *
 * See is_module_address() if you simply want to see if the address is
 * anywhere in a module.  See kernel_text_address() for testing if an
 * address corresponds to kernel or module code.
 */
bool is_module_text_address(unsigned long addr)
{
	bool ret;

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

	return ret;
}

/*
 * __module_text_address - get the module whose code contains an address.
 * @addr: the address.
 *
 * Must be called with preempt disabled or module mutex held so that
 * module doesn't get freed during this.
 */
struct module *__module_text_address(unsigned long addr)
{
	struct module *mod = __module_address(addr);
	if (mod) {
		/* Make sure it's within the text section. */
		if (!within(addr, mod->module_init, mod->init_text_size)
		    && !within(addr, mod->module_core, mod->core_text_size))
			mod = NULL;
	}
	return mod;
}
4065
EXPORT_SYMBOL_GPL(__module_text_address);
R
Rusty Russell 已提交
4066

L
Linus Torvalds 已提交
4067 4068 4069 4070
/* Don't grab lock, we're oopsing. */
void print_modules(void)
{
	struct module *mod;
4071
	char buf[8];
L
Linus Torvalds 已提交
4072

4073
	printk(KERN_DEFAULT "Modules linked in:");
4074 4075
	/* Most callers should already have preempt disabled, but make sure */
	preempt_disable();
4076 4077 4078
	list_for_each_entry_rcu(mod, &modules, list) {
		if (mod->state == MODULE_STATE_UNFORMED)
			continue;
J
Jiri Slaby 已提交
4079
		pr_cont(" %s%s", mod->name, module_flags(mod, buf));
4080
	}
4081
	preempt_enable();
4082
	if (last_unloaded_module[0])
J
Jiri Slaby 已提交
4083 4084
		pr_cont(" [last unloaded: %s]", last_unloaded_module);
	pr_cont("\n");
L
Linus Torvalds 已提交
4085 4086 4087
}

#ifdef CONFIG_MODVERSIONS
4088 4089 4090 4091 4092 4093
/* 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,
4094
		   struct tracepoint * const *tp)
4095 4096 4097
{
}
EXPORT_SYMBOL(module_layout);
L
Linus Torvalds 已提交
4098
#endif