compat.c 56.2 KB
Newer Older
L
Linus Torvalds 已提交
1 2 3 4 5 6 7 8 9 10
/*
 *  linux/fs/compat.c
 *
 *  Kernel compatibililty routines for e.g. 32 bit syscall support
 *  on 64 bit kernels.
 *
 *  Copyright (C) 2002       Stephen Rothwell, IBM Corporation
 *  Copyright (C) 1997-2000  Jakub Jelinek  (jakub@redhat.com)
 *  Copyright (C) 1998       Eddie C. Dost  (ecd@skynet.be)
 *  Copyright (C) 2001,2002  Andi Kleen, SuSE Labs 
P
Pavel Machek 已提交
11
 *  Copyright (C) 2003       Pavel Machek (pavel@ucw.cz)
L
Linus Torvalds 已提交
12 13 14 15 16 17
 *
 *  This program is free software; you can redistribute it and/or modify
 *  it under the terms of the GNU General Public License version 2 as
 *  published by the Free Software Foundation.
 */

18
#include <linux/stddef.h>
19
#include <linux/kernel.h>
L
Linus Torvalds 已提交
20 21 22 23 24 25 26 27
#include <linux/linkage.h>
#include <linux/compat.h>
#include <linux/errno.h>
#include <linux/time.h>
#include <linux/fs.h>
#include <linux/fcntl.h>
#include <linux/namei.h>
#include <linux/file.h>
A
Al Viro 已提交
28
#include <linux/fdtable.h>
L
Linus Torvalds 已提交
29 30 31 32 33 34
#include <linux/vfs.h>
#include <linux/ioctl.h>
#include <linux/init.h>
#include <linux/smb.h>
#include <linux/smb_mount.h>
#include <linux/ncp_mount.h>
35
#include <linux/nfs4_mount.h>
L
Linus Torvalds 已提交
36 37 38 39
#include <linux/syscalls.h>
#include <linux/ctype.h>
#include <linux/module.h>
#include <linux/dirent.h>
R
Robert Love 已提交
40
#include <linux/fsnotify.h>
L
Linus Torvalds 已提交
41 42 43 44
#include <linux/highuid.h>
#include <linux/nfsd/syscall.h>
#include <linux/personality.h>
#include <linux/rwsem.h>
45
#include <linux/tsacct_kern.h>
46
#include <linux/security.h>
A
Al Viro 已提交
47
#include <linux/highmem.h>
48
#include <linux/signal.h>
A
Al Viro 已提交
49
#include <linux/poll.h>
50
#include <linux/mm.h>
51
#include <linux/eventpoll.h>
A
Al Viro 已提交
52
#include <linux/fs_struct.h>
53
#include <linux/slab.h>
L
Linus Torvalds 已提交
54 55 56 57

#include <asm/uaccess.h>
#include <asm/mmu_context.h>
#include <asm/ioctls.h>
58
#include "internal.h"
59

60 61 62 63 64 65 66 67 68 69 70 71 72 73
int compat_log = 1;

int compat_printk(const char *fmt, ...)
{
	va_list ap;
	int ret;
	if (!compat_log)
		return 0;
	va_start(ap, fmt);
	ret = vprintk(fmt, ap);
	va_end(ap);
	return ret;
}

74 75
#include "read_write.h"

L
Linus Torvalds 已提交
76 77 78 79
/*
 * Not all architectures have sys_utime, so implement this in terms
 * of sys_utimes.
 */
80 81
asmlinkage long compat_sys_utime(const char __user *filename,
				 struct compat_utimbuf __user *t)
L
Linus Torvalds 已提交
82
{
U
Ulrich Drepper 已提交
83
	struct timespec tv[2];
L
Linus Torvalds 已提交
84 85 86 87 88

	if (t) {
		if (get_user(tv[0].tv_sec, &t->actime) ||
		    get_user(tv[1].tv_sec, &t->modtime))
			return -EFAULT;
U
Ulrich Drepper 已提交
89 90
		tv[0].tv_nsec = 0;
		tv[1].tv_nsec = 0;
L
Linus Torvalds 已提交
91
	}
U
Ulrich Drepper 已提交
92 93 94
	return do_utimes(AT_FDCWD, filename, t ? tv : NULL, 0);
}

95
asmlinkage long compat_sys_utimensat(unsigned int dfd, const char __user *filename, struct compat_timespec __user *t, int flags)
U
Ulrich Drepper 已提交
96 97 98 99 100 101 102 103 104 105 106 107
{
	struct timespec tv[2];

	if  (t) {
		if (get_compat_timespec(&tv[0], &t[0]) ||
		    get_compat_timespec(&tv[1], &t[1]))
			return -EFAULT;

		if (tv[0].tv_nsec == UTIME_OMIT && tv[1].tv_nsec == UTIME_OMIT)
			return 0;
	}
	return do_utimes(dfd, filename, t ? tv : NULL, flags);
L
Linus Torvalds 已提交
108 109
}

110
asmlinkage long compat_sys_futimesat(unsigned int dfd, const char __user *filename, struct compat_timeval __user *t)
L
Linus Torvalds 已提交
111
{
U
Ulrich Drepper 已提交
112
	struct timespec tv[2];
L
Linus Torvalds 已提交
113

114
	if (t) {
L
Linus Torvalds 已提交
115
		if (get_user(tv[0].tv_sec, &t[0].tv_sec) ||
U
Ulrich Drepper 已提交
116
		    get_user(tv[0].tv_nsec, &t[0].tv_usec) ||
L
Linus Torvalds 已提交
117
		    get_user(tv[1].tv_sec, &t[1].tv_sec) ||
U
Ulrich Drepper 已提交
118
		    get_user(tv[1].tv_nsec, &t[1].tv_usec))
119
			return -EFAULT;
U
Ulrich Drepper 已提交
120 121 122 123 124
		if (tv[0].tv_nsec >= 1000000 || tv[0].tv_nsec < 0 ||
		    tv[1].tv_nsec >= 1000000 || tv[1].tv_nsec < 0)
			return -EINVAL;
		tv[0].tv_nsec *= 1000;
		tv[1].tv_nsec *= 1000;
125
	}
U
Ulrich Drepper 已提交
126
	return do_utimes(dfd, filename, t ? tv : NULL, 0);
127 128
}

129
asmlinkage long compat_sys_utimes(const char __user *filename, struct compat_timeval __user *t)
130 131
{
	return compat_sys_futimesat(AT_FDCWD, filename, t);
L
Linus Torvalds 已提交
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 int cp_compat_stat(struct kstat *stat, struct compat_stat __user *ubuf)
{
	compat_ino_t ino = stat->ino;
	typeof(ubuf->st_uid) uid = 0;
	typeof(ubuf->st_gid) gid = 0;
	int err;

	SET_UID(uid, stat->uid);
	SET_GID(gid, stat->gid);

	if ((u64) stat->size > MAX_NON_LFS ||
	    !old_valid_dev(stat->dev) ||
	    !old_valid_dev(stat->rdev))
		return -EOVERFLOW;
	if (sizeof(ino) < sizeof(stat->ino) && ino != stat->ino)
		return -EOVERFLOW;

	if (clear_user(ubuf, sizeof(*ubuf)))
		return -EFAULT;

	err  = __put_user(old_encode_dev(stat->dev), &ubuf->st_dev);
	err |= __put_user(ino, &ubuf->st_ino);
	err |= __put_user(stat->mode, &ubuf->st_mode);
	err |= __put_user(stat->nlink, &ubuf->st_nlink);
	err |= __put_user(uid, &ubuf->st_uid);
	err |= __put_user(gid, &ubuf->st_gid);
	err |= __put_user(old_encode_dev(stat->rdev), &ubuf->st_rdev);
	err |= __put_user(stat->size, &ubuf->st_size);
	err |= __put_user(stat->atime.tv_sec, &ubuf->st_atime);
	err |= __put_user(stat->atime.tv_nsec, &ubuf->st_atime_nsec);
	err |= __put_user(stat->mtime.tv_sec, &ubuf->st_mtime);
	err |= __put_user(stat->mtime.tv_nsec, &ubuf->st_mtime_nsec);
	err |= __put_user(stat->ctime.tv_sec, &ubuf->st_ctime);
	err |= __put_user(stat->ctime.tv_nsec, &ubuf->st_ctime_nsec);
	err |= __put_user(stat->blksize, &ubuf->st_blksize);
	err |= __put_user(stat->blocks, &ubuf->st_blocks);
	return err;
}

173
asmlinkage long compat_sys_newstat(const char __user * filename,
L
Linus Torvalds 已提交
174 175 176
		struct compat_stat __user *statbuf)
{
	struct kstat stat;
177
	int error;
L
Linus Torvalds 已提交
178

179 180 181 182
	error = vfs_stat(filename, &stat);
	if (error)
		return error;
	return cp_compat_stat(&stat, statbuf);
L
Linus Torvalds 已提交
183 184
}

185
asmlinkage long compat_sys_newlstat(const char __user * filename,
L
Linus Torvalds 已提交
186 187 188
		struct compat_stat __user *statbuf)
{
	struct kstat stat;
189
	int error;
L
Linus Torvalds 已提交
190

191 192 193 194
	error = vfs_lstat(filename, &stat);
	if (error)
		return error;
	return cp_compat_stat(&stat, statbuf);
L
Linus Torvalds 已提交
195 196
}

197
#ifndef __ARCH_WANT_STAT64
198 199
asmlinkage long compat_sys_newfstatat(unsigned int dfd,
		const char __user *filename,
200 201 202
		struct compat_stat __user *statbuf, int flag)
{
	struct kstat stat;
203
	int error;
204

205 206 207 208
	error = vfs_fstatat(dfd, filename, &stat, flag);
	if (error)
		return error;
	return cp_compat_stat(&stat, statbuf);
209
}
210
#endif
211

L
Linus Torvalds 已提交
212 213 214 215 216 217 218 219 220 221 222 223 224 225 226
asmlinkage long compat_sys_newfstat(unsigned int fd,
		struct compat_stat __user * statbuf)
{
	struct kstat stat;
	int error = vfs_fstat(fd, &stat);

	if (!error)
		error = cp_compat_stat(&stat, statbuf);
	return error;
}

static int put_compat_statfs(struct compat_statfs __user *ubuf, struct kstatfs *kbuf)
{
	
	if (sizeof ubuf->f_blocks == 4) {
J
Jon Tollefson 已提交
227 228
		if ((kbuf->f_blocks | kbuf->f_bfree | kbuf->f_bavail |
		     kbuf->f_bsize | kbuf->f_frsize) & 0xffffffff00000000ULL)
L
Linus Torvalds 已提交
229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263
			return -EOVERFLOW;
		/* f_files and f_ffree may be -1; it's okay
		 * to stuff that into 32 bits */
		if (kbuf->f_files != 0xffffffffffffffffULL
		 && (kbuf->f_files & 0xffffffff00000000ULL))
			return -EOVERFLOW;
		if (kbuf->f_ffree != 0xffffffffffffffffULL
		 && (kbuf->f_ffree & 0xffffffff00000000ULL))
			return -EOVERFLOW;
	}
	if (!access_ok(VERIFY_WRITE, ubuf, sizeof(*ubuf)) ||
	    __put_user(kbuf->f_type, &ubuf->f_type) ||
	    __put_user(kbuf->f_bsize, &ubuf->f_bsize) ||
	    __put_user(kbuf->f_blocks, &ubuf->f_blocks) ||
	    __put_user(kbuf->f_bfree, &ubuf->f_bfree) ||
	    __put_user(kbuf->f_bavail, &ubuf->f_bavail) ||
	    __put_user(kbuf->f_files, &ubuf->f_files) ||
	    __put_user(kbuf->f_ffree, &ubuf->f_ffree) ||
	    __put_user(kbuf->f_namelen, &ubuf->f_namelen) ||
	    __put_user(kbuf->f_fsid.val[0], &ubuf->f_fsid.val[0]) ||
	    __put_user(kbuf->f_fsid.val[1], &ubuf->f_fsid.val[1]) ||
	    __put_user(kbuf->f_frsize, &ubuf->f_frsize) ||
	    __put_user(0, &ubuf->f_spare[0]) || 
	    __put_user(0, &ubuf->f_spare[1]) || 
	    __put_user(0, &ubuf->f_spare[2]) || 
	    __put_user(0, &ubuf->f_spare[3]) || 
	    __put_user(0, &ubuf->f_spare[4]))
		return -EFAULT;
	return 0;
}

/*
 * The following statfs calls are copies of code from fs/open.c and
 * should be checked against those from time to time
 */
264
asmlinkage long compat_sys_statfs(const char __user *pathname, struct compat_statfs __user *buf)
L
Linus Torvalds 已提交
265
{
266
	struct path path;
L
Linus Torvalds 已提交
267 268
	int error;

269
	error = user_path(pathname, &path);
L
Linus Torvalds 已提交
270 271
	if (!error) {
		struct kstatfs tmp;
272
		error = vfs_statfs(&path, &tmp);
273 274
		if (!error)
			error = put_compat_statfs(buf, &tmp);
275
		path_put(&path);
L
Linus Torvalds 已提交
276 277 278 279 280 281 282 283 284 285 286 287 288 289
	}
	return error;
}

asmlinkage long compat_sys_fstatfs(unsigned int fd, struct compat_statfs __user *buf)
{
	struct file * file;
	struct kstatfs tmp;
	int error;

	error = -EBADF;
	file = fget(fd);
	if (!file)
		goto out;
290
	error = vfs_statfs(&file->f_path, &tmp);
291 292
	if (!error)
		error = put_compat_statfs(buf, &tmp);
L
Linus Torvalds 已提交
293 294 295 296 297 298 299 300
	fput(file);
out:
	return error;
}

static int put_compat_statfs64(struct compat_statfs64 __user *ubuf, struct kstatfs *kbuf)
{
	if (sizeof ubuf->f_blocks == 4) {
J
Jon Tollefson 已提交
301 302
		if ((kbuf->f_blocks | kbuf->f_bfree | kbuf->f_bavail |
		     kbuf->f_bsize | kbuf->f_frsize) & 0xffffffff00000000ULL)
L
Linus Torvalds 已提交
303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328
			return -EOVERFLOW;
		/* f_files and f_ffree may be -1; it's okay
		 * to stuff that into 32 bits */
		if (kbuf->f_files != 0xffffffffffffffffULL
		 && (kbuf->f_files & 0xffffffff00000000ULL))
			return -EOVERFLOW;
		if (kbuf->f_ffree != 0xffffffffffffffffULL
		 && (kbuf->f_ffree & 0xffffffff00000000ULL))
			return -EOVERFLOW;
	}
	if (!access_ok(VERIFY_WRITE, ubuf, sizeof(*ubuf)) ||
	    __put_user(kbuf->f_type, &ubuf->f_type) ||
	    __put_user(kbuf->f_bsize, &ubuf->f_bsize) ||
	    __put_user(kbuf->f_blocks, &ubuf->f_blocks) ||
	    __put_user(kbuf->f_bfree, &ubuf->f_bfree) ||
	    __put_user(kbuf->f_bavail, &ubuf->f_bavail) ||
	    __put_user(kbuf->f_files, &ubuf->f_files) ||
	    __put_user(kbuf->f_ffree, &ubuf->f_ffree) ||
	    __put_user(kbuf->f_namelen, &ubuf->f_namelen) ||
	    __put_user(kbuf->f_fsid.val[0], &ubuf->f_fsid.val[0]) ||
	    __put_user(kbuf->f_fsid.val[1], &ubuf->f_fsid.val[1]) ||
	    __put_user(kbuf->f_frsize, &ubuf->f_frsize))
		return -EFAULT;
	return 0;
}

329
asmlinkage long compat_sys_statfs64(const char __user *pathname, compat_size_t sz, struct compat_statfs64 __user *buf)
L
Linus Torvalds 已提交
330
{
331
	struct path path;
L
Linus Torvalds 已提交
332 333 334 335 336
	int error;

	if (sz != sizeof(*buf))
		return -EINVAL;

337
	error = user_path(pathname, &path);
L
Linus Torvalds 已提交
338 339
	if (!error) {
		struct kstatfs tmp;
340
		error = vfs_statfs(&path, &tmp);
341 342
		if (!error)
			error = put_compat_statfs64(buf, &tmp);
343
		path_put(&path);
L
Linus Torvalds 已提交
344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360
	}
	return error;
}

asmlinkage long compat_sys_fstatfs64(unsigned int fd, compat_size_t sz, struct compat_statfs64 __user *buf)
{
	struct file * file;
	struct kstatfs tmp;
	int error;

	if (sz != sizeof(*buf))
		return -EINVAL;

	error = -EBADF;
	file = fget(fd);
	if (!file)
		goto out;
361
	error = vfs_statfs(&file->f_path, &tmp);
362 363
	if (!error)
		error = put_compat_statfs64(buf, &tmp);
L
Linus Torvalds 已提交
364 365 366 367 368
	fput(file);
out:
	return error;
}

C
Christoph Hellwig 已提交
369 370 371 372 373 374 375 376 377 378 379 380 381 382 383
/*
 * This is a copy of sys_ustat, just dealing with a structure layout.
 * Given how simple this syscall is that apporach is more maintainable
 * than the various conversion hacks.
 */
asmlinkage long compat_sys_ustat(unsigned dev, struct compat_ustat __user *u)
{
	struct super_block *sb;
	struct compat_ustat tmp;
	struct kstatfs sbuf;
	int err;

	sb = user_get_super(new_decode_dev(dev));
	if (!sb)
		return -EINVAL;
384
	err = statfs_by_dentry(sb->s_root, &sbuf);
C
Christoph Hellwig 已提交
385 386 387 388 389 390 391 392 393 394 395 396
	drop_super(sb);
	if (err)
		return err;

	memset(&tmp, 0, sizeof(struct compat_ustat));
	tmp.f_tfree = sbuf.f_bfree;
	tmp.f_tinode = sbuf.f_ffree;
	if (copy_to_user(u, &tmp, sizeof(struct compat_ustat)))
		return -EFAULT;
	return 0;
}

L
Linus Torvalds 已提交
397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467
static int get_compat_flock(struct flock *kfl, struct compat_flock __user *ufl)
{
	if (!access_ok(VERIFY_READ, ufl, sizeof(*ufl)) ||
	    __get_user(kfl->l_type, &ufl->l_type) ||
	    __get_user(kfl->l_whence, &ufl->l_whence) ||
	    __get_user(kfl->l_start, &ufl->l_start) ||
	    __get_user(kfl->l_len, &ufl->l_len) ||
	    __get_user(kfl->l_pid, &ufl->l_pid))
		return -EFAULT;
	return 0;
}

static int put_compat_flock(struct flock *kfl, struct compat_flock __user *ufl)
{
	if (!access_ok(VERIFY_WRITE, ufl, sizeof(*ufl)) ||
	    __put_user(kfl->l_type, &ufl->l_type) ||
	    __put_user(kfl->l_whence, &ufl->l_whence) ||
	    __put_user(kfl->l_start, &ufl->l_start) ||
	    __put_user(kfl->l_len, &ufl->l_len) ||
	    __put_user(kfl->l_pid, &ufl->l_pid))
		return -EFAULT;
	return 0;
}

#ifndef HAVE_ARCH_GET_COMPAT_FLOCK64
static int get_compat_flock64(struct flock *kfl, struct compat_flock64 __user *ufl)
{
	if (!access_ok(VERIFY_READ, ufl, sizeof(*ufl)) ||
	    __get_user(kfl->l_type, &ufl->l_type) ||
	    __get_user(kfl->l_whence, &ufl->l_whence) ||
	    __get_user(kfl->l_start, &ufl->l_start) ||
	    __get_user(kfl->l_len, &ufl->l_len) ||
	    __get_user(kfl->l_pid, &ufl->l_pid))
		return -EFAULT;
	return 0;
}
#endif

#ifndef HAVE_ARCH_PUT_COMPAT_FLOCK64
static int put_compat_flock64(struct flock *kfl, struct compat_flock64 __user *ufl)
{
	if (!access_ok(VERIFY_WRITE, ufl, sizeof(*ufl)) ||
	    __put_user(kfl->l_type, &ufl->l_type) ||
	    __put_user(kfl->l_whence, &ufl->l_whence) ||
	    __put_user(kfl->l_start, &ufl->l_start) ||
	    __put_user(kfl->l_len, &ufl->l_len) ||
	    __put_user(kfl->l_pid, &ufl->l_pid))
		return -EFAULT;
	return 0;
}
#endif

asmlinkage long compat_sys_fcntl64(unsigned int fd, unsigned int cmd,
		unsigned long arg)
{
	mm_segment_t old_fs;
	struct flock f;
	long ret;

	switch (cmd) {
	case F_GETLK:
	case F_SETLK:
	case F_SETLKW:
		ret = get_compat_flock(&f, compat_ptr(arg));
		if (ret != 0)
			break;
		old_fs = get_fs();
		set_fs(KERNEL_DS);
		ret = sys_fcntl(fd, cmd, (unsigned long)&f);
		set_fs(old_fs);
		if (cmd == F_GETLK && ret == 0) {
468
			/* GETLK was successful and we need to return the data...
469 470 471 472 473 474 475 476 477 478 479
			 * but it needs to fit in the compat structure.
			 * l_start shouldn't be too big, unless the original
			 * start + end is greater than COMPAT_OFF_T_MAX, in which
			 * case the app was asking for trouble, so we return
			 * -EOVERFLOW in that case.
			 * l_len could be too big, in which case we just truncate it,
			 * and only allow the app to see that part of the conflicting
			 * lock that might make sense to it anyway
			 */

			if (f.l_start > COMPAT_OFF_T_MAX)
L
Linus Torvalds 已提交
480
				ret = -EOVERFLOW;
481 482
			if (f.l_len > COMPAT_OFF_T_MAX)
				f.l_len = COMPAT_OFF_T_MAX;
L
Linus Torvalds 已提交
483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500
			if (ret == 0)
				ret = put_compat_flock(&f, compat_ptr(arg));
		}
		break;

	case F_GETLK64:
	case F_SETLK64:
	case F_SETLKW64:
		ret = get_compat_flock64(&f, compat_ptr(arg));
		if (ret != 0)
			break;
		old_fs = get_fs();
		set_fs(KERNEL_DS);
		ret = sys_fcntl(fd, (cmd == F_GETLK64) ? F_GETLK :
				((cmd == F_SETLK64) ? F_SETLK : F_SETLKW),
				(unsigned long)&f);
		set_fs(old_fs);
		if (cmd == F_GETLK64 && ret == 0) {
501 502
			/* need to return lock information - see above for commentary */
			if (f.l_start > COMPAT_LOFF_T_MAX)
L
Linus Torvalds 已提交
503
				ret = -EOVERFLOW;
504 505
			if (f.l_len > COMPAT_LOFF_T_MAX)
				f.l_len = COMPAT_LOFF_T_MAX;
L
Linus Torvalds 已提交
506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573
			if (ret == 0)
				ret = put_compat_flock64(&f, compat_ptr(arg));
		}
		break;

	default:
		ret = sys_fcntl(fd, cmd, arg);
		break;
	}
	return ret;
}

asmlinkage long compat_sys_fcntl(unsigned int fd, unsigned int cmd,
		unsigned long arg)
{
	if ((cmd == F_GETLK64) || (cmd == F_SETLK64) || (cmd == F_SETLKW64))
		return -EINVAL;
	return compat_sys_fcntl64(fd, cmd, arg);
}

asmlinkage long
compat_sys_io_setup(unsigned nr_reqs, u32 __user *ctx32p)
{
	long ret;
	aio_context_t ctx64;

	mm_segment_t oldfs = get_fs();
	if (unlikely(get_user(ctx64, ctx32p)))
		return -EFAULT;

	set_fs(KERNEL_DS);
	/* The __user pointer cast is valid because of the set_fs() */
	ret = sys_io_setup(nr_reqs, (aio_context_t __user *) &ctx64);
	set_fs(oldfs);
	/* truncating is ok because it's a user address */
	if (!ret)
		ret = put_user((u32) ctx64, ctx32p);
	return ret;
}

asmlinkage long
compat_sys_io_getevents(aio_context_t ctx_id,
				 unsigned long min_nr,
				 unsigned long nr,
				 struct io_event __user *events,
				 struct compat_timespec __user *timeout)
{
	long ret;
	struct timespec t;
	struct timespec __user *ut = NULL;

	ret = -EFAULT;
	if (unlikely(!access_ok(VERIFY_WRITE, events, 
				nr * sizeof(struct io_event))))
		goto out;
	if (timeout) {
		if (get_compat_timespec(&t, timeout))
			goto out;

		ut = compat_alloc_user_space(sizeof(*ut));
		if (copy_to_user(ut, &t, sizeof(t)) )
			goto out;
	} 
	ret = sys_io_getevents(ctx_id, min_nr, nr, events, ut);
out:
	return ret;
}

574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631
/* A write operation does a read from user space and vice versa */
#define vrfy_dir(type) ((type) == READ ? VERIFY_WRITE : VERIFY_READ)

ssize_t compat_rw_copy_check_uvector(int type,
		const struct compat_iovec __user *uvector, unsigned long nr_segs,
		unsigned long fast_segs, struct iovec *fast_pointer,
		struct iovec **ret_pointer)
{
	compat_ssize_t tot_len;
	struct iovec *iov = *ret_pointer = fast_pointer;
	ssize_t ret = 0;
	int seg;

	/*
	 * SuS says "The readv() function *may* fail if the iovcnt argument
	 * was less than or equal to 0, or greater than {IOV_MAX}.  Linux has
	 * traditionally returned zero for zero segments, so...
	 */
	if (nr_segs == 0)
		goto out;

	ret = -EINVAL;
	if (nr_segs > UIO_MAXIOV || nr_segs < 0)
		goto out;
	if (nr_segs > fast_segs) {
		ret = -ENOMEM;
		iov = kmalloc(nr_segs*sizeof(struct iovec), GFP_KERNEL);
		if (iov == NULL) {
			*ret_pointer = fast_pointer;
			goto out;
		}
	}
	*ret_pointer = iov;

	/*
	 * Single unix specification:
	 * We should -EINVAL if an element length is not >= 0 and fitting an
	 * ssize_t.  The total length is fitting an ssize_t
	 *
	 * Be careful here because iov_len is a size_t not an ssize_t
	 */
	tot_len = 0;
	ret = -EINVAL;
	for (seg = 0; seg < nr_segs; seg++) {
		compat_ssize_t tmp = tot_len;
		compat_uptr_t buf;
		compat_ssize_t len;

		if (__get_user(len, &uvector->iov_len) ||
		   __get_user(buf, &uvector->iov_base)) {
			ret = -EFAULT;
			goto out;
		}
		if (len < 0)	/* size_t not fitting in compat_ssize_t .. */
			goto out;
		tot_len += len;
		if (tot_len < tmp) /* maths overflow on the compat_ssize_t */
			goto out;
632
		if (!access_ok(vrfy_dir(type), compat_ptr(buf), len)) {
633 634 635 636 637 638 639 640 641 642 643 644 645 646
			ret = -EFAULT;
			goto out;
		}
		iov->iov_base = compat_ptr(buf);
		iov->iov_len = (compat_size_t) len;
		uvector++;
		iov++;
	}
	ret = tot_len;

out:
	return ret;
}

L
Linus Torvalds 已提交
647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678
static inline long
copy_iocb(long nr, u32 __user *ptr32, struct iocb __user * __user *ptr64)
{
	compat_uptr_t uptr;
	int i;

	for (i = 0; i < nr; ++i) {
		if (get_user(uptr, ptr32 + i))
			return -EFAULT;
		if (put_user(compat_ptr(uptr), ptr64 + i))
			return -EFAULT;
	}
	return 0;
}

#define MAX_AIO_SUBMITS 	(PAGE_SIZE/sizeof(struct iocb *))

asmlinkage long
compat_sys_io_submit(aio_context_t ctx_id, int nr, u32 __user *iocb)
{
	struct iocb __user * __user *iocb64; 
	long ret;

	if (unlikely(nr < 0))
		return -EINVAL;

	if (nr > MAX_AIO_SUBMITS)
		nr = MAX_AIO_SUBMITS;
	
	iocb64 = compat_alloc_user_space(nr * sizeof(*iocb64));
	ret = copy_iocb(nr, iocb, iocb64);
	if (!ret)
679
		ret = do_io_submit(ctx_id, nr, iocb64, 1);
L
Linus Torvalds 已提交
680 681 682 683 684 685
	return ret;
}

struct compat_ncp_mount_data {
	compat_int_t version;
	compat_uint_t ncp_fd;
686
	__compat_uid_t mounted_uid;
L
Linus Torvalds 已提交
687 688 689 690 691
	compat_pid_t wdog_pid;
	unsigned char mounted_vol[NCP_VOLNAME_LEN + 1];
	compat_uint_t time_out;
	compat_uint_t retry_count;
	compat_uint_t flags;
692 693
	__compat_uid_t uid;
	__compat_gid_t gid;
L
Linus Torvalds 已提交
694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749
	compat_mode_t file_mode;
	compat_mode_t dir_mode;
};

struct compat_ncp_mount_data_v4 {
	compat_int_t version;
	compat_ulong_t flags;
	compat_ulong_t mounted_uid;
	compat_long_t wdog_pid;
	compat_uint_t ncp_fd;
	compat_uint_t time_out;
	compat_uint_t retry_count;
	compat_ulong_t uid;
	compat_ulong_t gid;
	compat_ulong_t file_mode;
	compat_ulong_t dir_mode;
};

static void *do_ncp_super_data_conv(void *raw_data)
{
	int version = *(unsigned int *)raw_data;

	if (version == 3) {
		struct compat_ncp_mount_data *c_n = raw_data;
		struct ncp_mount_data *n = raw_data;

		n->dir_mode = c_n->dir_mode;
		n->file_mode = c_n->file_mode;
		n->gid = c_n->gid;
		n->uid = c_n->uid;
		memmove (n->mounted_vol, c_n->mounted_vol, (sizeof (c_n->mounted_vol) + 3 * sizeof (unsigned int)));
		n->wdog_pid = c_n->wdog_pid;
		n->mounted_uid = c_n->mounted_uid;
	} else if (version == 4) {
		struct compat_ncp_mount_data_v4 *c_n = raw_data;
		struct ncp_mount_data_v4 *n = raw_data;

		n->dir_mode = c_n->dir_mode;
		n->file_mode = c_n->file_mode;
		n->gid = c_n->gid;
		n->uid = c_n->uid;
		n->retry_count = c_n->retry_count;
		n->time_out = c_n->time_out;
		n->ncp_fd = c_n->ncp_fd;
		n->wdog_pid = c_n->wdog_pid;
		n->mounted_uid = c_n->mounted_uid;
		n->flags = c_n->flags;
	} else if (version != 5) {
		return NULL;
	}

	return raw_data;
}

struct compat_smb_mount_data {
	compat_int_t version;
750 751 752
	__compat_uid_t mounted_uid;
	__compat_uid_t uid;
	__compat_gid_t gid;
L
Linus Torvalds 已提交
753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772
	compat_mode_t file_mode;
	compat_mode_t dir_mode;
};

static void *do_smb_super_data_conv(void *raw_data)
{
	struct smb_mount_data *s = raw_data;
	struct compat_smb_mount_data *c_s = raw_data;

	if (c_s->version != SMB_MOUNT_OLDVERSION)
		goto out;
	s->dir_mode = c_s->dir_mode;
	s->file_mode = c_s->file_mode;
	s->gid = c_s->gid;
	s->uid = c_s->uid;
	s->mounted_uid = c_s->mounted_uid;
 out:
	return raw_data;
}

773 774
struct compat_nfs_string {
	compat_uint_t len;
775
	compat_uptr_t data;
776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799
};

static inline void compat_nfs_string(struct nfs_string *dst,
				     struct compat_nfs_string *src)
{
	dst->data = compat_ptr(src->data);
	dst->len = src->len;
}

struct compat_nfs4_mount_data_v1 {
	compat_int_t version;
	compat_int_t flags;
	compat_int_t rsize;
	compat_int_t wsize;
	compat_int_t timeo;
	compat_int_t retrans;
	compat_int_t acregmin;
	compat_int_t acregmax;
	compat_int_t acdirmin;
	compat_int_t acdirmax;
	struct compat_nfs_string client_addr;
	struct compat_nfs_string mnt_path;
	struct compat_nfs_string hostname;
	compat_uint_t host_addrlen;
800
	compat_uptr_t host_addr;
801 802
	compat_int_t proto;
	compat_int_t auth_flavourlen;
803
	compat_uptr_t auth_flavours;
804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837
};

static int do_nfs4_super_data_conv(void *raw_data)
{
	int version = *(compat_uint_t *) raw_data;

	if (version == 1) {
		struct compat_nfs4_mount_data_v1 *raw = raw_data;
		struct nfs4_mount_data *real = raw_data;

		/* copy the fields backwards */
		real->auth_flavours = compat_ptr(raw->auth_flavours);
		real->auth_flavourlen = raw->auth_flavourlen;
		real->proto = raw->proto;
		real->host_addr = compat_ptr(raw->host_addr);
		real->host_addrlen = raw->host_addrlen;
		compat_nfs_string(&real->hostname, &raw->hostname);
		compat_nfs_string(&real->mnt_path, &raw->mnt_path);
		compat_nfs_string(&real->client_addr, &raw->client_addr);
		real->acdirmax = raw->acdirmax;
		real->acdirmin = raw->acdirmin;
		real->acregmax = raw->acregmax;
		real->acregmin = raw->acregmin;
		real->retrans = raw->retrans;
		real->timeo = raw->timeo;
		real->wsize = raw->wsize;
		real->rsize = raw->rsize;
		real->flags = raw->flags;
		real->version = raw->version;
	}

	return 0;
}

L
Linus Torvalds 已提交
838 839
#define SMBFS_NAME      "smbfs"
#define NCPFS_NAME      "ncpfs"
840
#define NFS4_NAME	"nfs4"
L
Linus Torvalds 已提交
841

842 843 844 845
asmlinkage long compat_sys_mount(const char __user * dev_name,
				 const char __user * dir_name,
				 const char __user * type, unsigned long flags,
				 const void __user * data)
L
Linus Torvalds 已提交
846
{
847
	char *kernel_type;
L
Linus Torvalds 已提交
848
	unsigned long data_page;
849
	char *kernel_dev;
L
Linus Torvalds 已提交
850 851 852
	char *dir_page;
	int retval;

853
	retval = copy_mount_string(type, &kernel_type);
L
Linus Torvalds 已提交
854 855 856 857 858 859 860 861
	if (retval < 0)
		goto out;

	dir_page = getname(dir_name);
	retval = PTR_ERR(dir_page);
	if (IS_ERR(dir_page))
		goto out1;

862
	retval = copy_mount_string(dev_name, &kernel_dev);
L
Linus Torvalds 已提交
863 864 865
	if (retval < 0)
		goto out2;

866
	retval = copy_mount_options(data, &data_page);
L
Linus Torvalds 已提交
867 868 869 870 871
	if (retval < 0)
		goto out3;

	retval = -EINVAL;

872 873
	if (kernel_type && data_page) {
		if (!strcmp(kernel_type, SMBFS_NAME)) {
L
Linus Torvalds 已提交
874
			do_smb_super_data_conv((void *)data_page);
875
		} else if (!strcmp(kernel_type, NCPFS_NAME)) {
L
Linus Torvalds 已提交
876
			do_ncp_super_data_conv((void *)data_page);
877
		} else if (!strcmp(kernel_type, NFS4_NAME)) {
878 879
			if (do_nfs4_super_data_conv((void *) data_page))
				goto out4;
L
Linus Torvalds 已提交
880 881 882
		}
	}

883
	retval = do_mount(kernel_dev, dir_page, kernel_type,
L
Linus Torvalds 已提交
884 885
			flags, (void*)data_page);

886
 out4:
L
Linus Torvalds 已提交
887 888
	free_page(data_page);
 out3:
889
	kfree(kernel_dev);
L
Linus Torvalds 已提交
890 891 892
 out2:
	putname(dir_page);
 out1:
893
	kfree(kernel_type);
L
Linus Torvalds 已提交
894 895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910
 out:
	return retval;
}

struct compat_old_linux_dirent {
	compat_ulong_t	d_ino;
	compat_ulong_t	d_offset;
	unsigned short	d_namlen;
	char		d_name[1];
};

struct compat_readdir_callback {
	struct compat_old_linux_dirent __user *dirent;
	int result;
};

static int compat_fillonedir(void *__buf, const char *name, int namlen,
911
			loff_t offset, u64 ino, unsigned int d_type)
L
Linus Torvalds 已提交
912 913 914
{
	struct compat_readdir_callback *buf = __buf;
	struct compat_old_linux_dirent __user *dirent;
915
	compat_ulong_t d_ino;
L
Linus Torvalds 已提交
916 917 918

	if (buf->result)
		return -EINVAL;
919
	d_ino = ino;
920 921
	if (sizeof(d_ino) < sizeof(ino) && d_ino != ino) {
		buf->result = -EOVERFLOW;
922
		return -EOVERFLOW;
923
	}
L
Linus Torvalds 已提交
924 925 926 927 928 929
	buf->result++;
	dirent = buf->dirent;
	if (!access_ok(VERIFY_WRITE, dirent,
			(unsigned long)(dirent->d_name + namlen + 1) -
				(unsigned long)dirent))
		goto efault;
930
	if (	__put_user(d_ino, &dirent->d_ino) ||
L
Linus Torvalds 已提交
931 932 933 934 935 936 937 938 939 940 941 942 943 944 945 946 947 948 949 950 951 952 953 954 955 956 957
		__put_user(offset, &dirent->d_offset) ||
		__put_user(namlen, &dirent->d_namlen) ||
		__copy_to_user(dirent->d_name, name, namlen) ||
		__put_user(0, dirent->d_name + namlen))
		goto efault;
	return 0;
efault:
	buf->result = -EFAULT;
	return -EFAULT;
}

asmlinkage long compat_sys_old_readdir(unsigned int fd,
	struct compat_old_linux_dirent __user *dirent, unsigned int count)
{
	int error;
	struct file *file;
	struct compat_readdir_callback buf;

	error = -EBADF;
	file = fget(fd);
	if (!file)
		goto out;

	buf.result = 0;
	buf.dirent = dirent;

	error = vfs_readdir(file, compat_fillonedir, &buf);
958
	if (buf.result)
L
Linus Torvalds 已提交
959 960 961 962 963 964 965 966 967 968 969 970 971 972 973 974 975 976 977 978 979 980
		error = buf.result;

	fput(file);
out:
	return error;
}

struct compat_linux_dirent {
	compat_ulong_t	d_ino;
	compat_ulong_t	d_off;
	unsigned short	d_reclen;
	char		d_name[1];
};

struct compat_getdents_callback {
	struct compat_linux_dirent __user *current_dir;
	struct compat_linux_dirent __user *previous;
	int count;
	int error;
};

static int compat_filldir(void *__buf, const char *name, int namlen,
981
		loff_t offset, u64 ino, unsigned int d_type)
L
Linus Torvalds 已提交
982 983 984
{
	struct compat_linux_dirent __user * dirent;
	struct compat_getdents_callback *buf = __buf;
985
	compat_ulong_t d_ino;
986 987
	int reclen = ALIGN(offsetof(struct compat_linux_dirent, d_name) +
		namlen + 2, sizeof(compat_long_t));
L
Linus Torvalds 已提交
988 989 990 991

	buf->error = -EINVAL;	/* only used if we fail.. */
	if (reclen > buf->count)
		return -EINVAL;
992
	d_ino = ino;
993 994
	if (sizeof(d_ino) < sizeof(ino) && d_ino != ino) {
		buf->error = -EOVERFLOW;
995
		return -EOVERFLOW;
996
	}
L
Linus Torvalds 已提交
997 998 999 1000 1001 1002
	dirent = buf->previous;
	if (dirent) {
		if (__put_user(offset, &dirent->d_off))
			goto efault;
	}
	dirent = buf->current_dir;
1003
	if (__put_user(d_ino, &dirent->d_ino))
L
Linus Torvalds 已提交
1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043 1044 1045
		goto efault;
	if (__put_user(reclen, &dirent->d_reclen))
		goto efault;
	if (copy_to_user(dirent->d_name, name, namlen))
		goto efault;
	if (__put_user(0, dirent->d_name + namlen))
		goto efault;
	if (__put_user(d_type, (char  __user *) dirent + reclen - 1))
		goto efault;
	buf->previous = dirent;
	dirent = (void __user *)dirent + reclen;
	buf->current_dir = dirent;
	buf->count -= reclen;
	return 0;
efault:
	buf->error = -EFAULT;
	return -EFAULT;
}

asmlinkage long compat_sys_getdents(unsigned int fd,
		struct compat_linux_dirent __user *dirent, unsigned int count)
{
	struct file * file;
	struct compat_linux_dirent __user * lastdirent;
	struct compat_getdents_callback buf;
	int error;

	error = -EFAULT;
	if (!access_ok(VERIFY_WRITE, dirent, count))
		goto out;

	error = -EBADF;
	file = fget(fd);
	if (!file)
		goto out;

	buf.current_dir = dirent;
	buf.previous = NULL;
	buf.count = count;
	buf.error = 0;

	error = vfs_readdir(file, compat_filldir, &buf);
1046 1047
	if (error >= 0)
		error = buf.error;
L
Linus Torvalds 已提交
1048 1049 1050 1051 1052 1053 1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064 1065 1066 1067 1068 1069
	lastdirent = buf.previous;
	if (lastdirent) {
		if (put_user(file->f_pos, &lastdirent->d_off))
			error = -EFAULT;
		else
			error = count - buf.count;
	}
	fput(file);
out:
	return error;
}

#ifndef __ARCH_OMIT_COMPAT_SYS_GETDENTS64

struct compat_getdents_callback64 {
	struct linux_dirent64 __user *current_dir;
	struct linux_dirent64 __user *previous;
	int count;
	int error;
};

static int compat_filldir64(void * __buf, const char * name, int namlen, loff_t offset,
1070
		     u64 ino, unsigned int d_type)
L
Linus Torvalds 已提交
1071 1072 1073
{
	struct linux_dirent64 __user *dirent;
	struct compat_getdents_callback64 *buf = __buf;
1074 1075
	int reclen = ALIGN(offsetof(struct linux_dirent64, d_name) + namlen + 1,
		sizeof(u64));
L
Linus Torvalds 已提交
1076 1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 1102 1103 1104 1105 1106 1107 1108 1109 1110 1111 1112 1113 1114 1115 1116 1117 1118 1119 1120 1121 1122 1123 1124 1125 1126 1127 1128 1129 1130 1131 1132 1133
	u64 off;

	buf->error = -EINVAL;	/* only used if we fail.. */
	if (reclen > buf->count)
		return -EINVAL;
	dirent = buf->previous;

	if (dirent) {
		if (__put_user_unaligned(offset, &dirent->d_off))
			goto efault;
	}
	dirent = buf->current_dir;
	if (__put_user_unaligned(ino, &dirent->d_ino))
		goto efault;
	off = 0;
	if (__put_user_unaligned(off, &dirent->d_off))
		goto efault;
	if (__put_user(reclen, &dirent->d_reclen))
		goto efault;
	if (__put_user(d_type, &dirent->d_type))
		goto efault;
	if (copy_to_user(dirent->d_name, name, namlen))
		goto efault;
	if (__put_user(0, dirent->d_name + namlen))
		goto efault;
	buf->previous = dirent;
	dirent = (void __user *)dirent + reclen;
	buf->current_dir = dirent;
	buf->count -= reclen;
	return 0;
efault:
	buf->error = -EFAULT;
	return -EFAULT;
}

asmlinkage long compat_sys_getdents64(unsigned int fd,
		struct linux_dirent64 __user * dirent, unsigned int count)
{
	struct file * file;
	struct linux_dirent64 __user * lastdirent;
	struct compat_getdents_callback64 buf;
	int error;

	error = -EFAULT;
	if (!access_ok(VERIFY_WRITE, dirent, count))
		goto out;

	error = -EBADF;
	file = fget(fd);
	if (!file)
		goto out;

	buf.current_dir = dirent;
	buf.previous = NULL;
	buf.count = count;
	buf.error = 0;

	error = vfs_readdir(file, compat_filldir64, &buf);
1134 1135
	if (error >= 0)
		error = buf.error;
L
Linus Torvalds 已提交
1136 1137 1138
	lastdirent = buf.previous;
	if (lastdirent) {
		typeof(lastdirent->d_off) d_off = file->f_pos;
1139
		if (__put_user_unaligned(d_off, &lastdirent->d_off))
1140 1141 1142
			error = -EFAULT;
		else
			error = count - buf.count;
L
Linus Torvalds 已提交
1143 1144 1145 1146 1147 1148 1149 1150 1151 1152 1153 1154 1155
	}
	fput(file);
out:
	return error;
}
#endif /* ! __ARCH_OMIT_COMPAT_SYS_GETDENTS64 */

static ssize_t compat_do_readv_writev(int type, struct file *file,
			       const struct compat_iovec __user *uvector,
			       unsigned long nr_segs, loff_t *pos)
{
	compat_ssize_t tot_len;
	struct iovec iovstack[UIO_FASTIOV];
1156
	struct iovec *iov;
L
Linus Torvalds 已提交
1157 1158 1159 1160 1161 1162 1163
	ssize_t ret;
	io_fn_t fn;
	iov_fn_t fnv;

	ret = -EINVAL;
	if (!file->f_op)
		goto out;
1164

L
Linus Torvalds 已提交
1165 1166 1167 1168
	ret = -EFAULT;
	if (!access_ok(VERIFY_READ, uvector, nr_segs*sizeof(*uvector)))
		goto out;

1169 1170
	tot_len = compat_rw_copy_check_uvector(type, uvector, nr_segs,
					       UIO_FASTIOV, iovstack, &iov);
L
Linus Torvalds 已提交
1171 1172 1173 1174 1175 1176
	if (tot_len == 0) {
		ret = 0;
		goto out;
	}

	ret = rw_verify_area(type, file, pos, tot_len);
1177
	if (ret < 0)
L
Linus Torvalds 已提交
1178 1179 1180 1181 1182
		goto out;

	fnv = NULL;
	if (type == READ) {
		fn = file->f_op->read;
1183
		fnv = file->f_op->aio_read;
L
Linus Torvalds 已提交
1184 1185
	} else {
		fn = (io_fn_t)file->f_op->write;
1186
		fnv = file->f_op->aio_write;
L
Linus Torvalds 已提交
1187 1188
	}

1189 1190 1191 1192 1193
	if (fnv)
		ret = do_sync_readv_writev(file, iov, nr_segs, tot_len,
						pos, fnv);
	else
		ret = do_loop_readv_writev(file, iov, nr_segs, pos, fn);
L
Linus Torvalds 已提交
1194 1195 1196 1197

out:
	if (iov != iovstack)
		kfree(iov);
R
Robert Love 已提交
1198 1199
	if ((ret + (type == READ)) > 0) {
		if (type == READ)
1200
			fsnotify_access(file);
R
Robert Love 已提交
1201
		else
1202
			fsnotify_modify(file);
R
Robert Love 已提交
1203
	}
L
Linus Torvalds 已提交
1204 1205 1206
	return ret;
}

1207 1208 1209
static size_t compat_readv(struct file *file,
			   const struct compat_iovec __user *vec,
			   unsigned long vlen, loff_t *pos)
L
Linus Torvalds 已提交
1210 1211 1212 1213 1214 1215 1216
{
	ssize_t ret = -EBADF;

	if (!(file->f_mode & FMODE_READ))
		goto out;

	ret = -EINVAL;
1217
	if (!file->f_op || (!file->f_op->aio_read && !file->f_op->read))
L
Linus Torvalds 已提交
1218 1219
		goto out;

1220
	ret = compat_do_readv_writev(READ, file, vec, vlen, pos);
L
Linus Torvalds 已提交
1221 1222

out:
1223 1224 1225
	if (ret > 0)
		add_rchar(current, ret);
	inc_syscr(current);
1226 1227 1228 1229 1230 1231 1232 1233
	return ret;
}

asmlinkage ssize_t
compat_sys_readv(unsigned long fd, const struct compat_iovec __user *vec,
		 unsigned long vlen)
{
	struct file *file;
1234
	int fput_needed;
1235 1236
	ssize_t ret;

1237
	file = fget_light(fd, &fput_needed);
1238 1239 1240
	if (!file)
		return -EBADF;
	ret = compat_readv(file, vec, vlen, &file->f_pos);
1241
	fput_light(file, fput_needed);
L
Linus Torvalds 已提交
1242 1243 1244
	return ret;
}

1245 1246
asmlinkage ssize_t
compat_sys_preadv(unsigned long fd, const struct compat_iovec __user *vec,
1247
		  unsigned long vlen, u32 pos_low, u32 pos_high)
1248 1249 1250
{
	loff_t pos = ((loff_t)pos_high << 32) | pos_low;
	struct file *file;
1251
	int fput_needed;
1252 1253 1254 1255
	ssize_t ret;

	if (pos < 0)
		return -EINVAL;
1256
	file = fget_light(fd, &fput_needed);
1257 1258 1259
	if (!file)
		return -EBADF;
	ret = compat_readv(file, vec, vlen, &pos);
1260
	fput_light(file, fput_needed);
1261 1262 1263
	return ret;
}

1264 1265 1266
static size_t compat_writev(struct file *file,
			    const struct compat_iovec __user *vec,
			    unsigned long vlen, loff_t *pos)
L
Linus Torvalds 已提交
1267 1268 1269 1270 1271 1272 1273
{
	ssize_t ret = -EBADF;

	if (!(file->f_mode & FMODE_WRITE))
		goto out;

	ret = -EINVAL;
1274
	if (!file->f_op || (!file->f_op->aio_write && !file->f_op->write))
L
Linus Torvalds 已提交
1275 1276
		goto out;

1277
	ret = compat_do_readv_writev(WRITE, file, vec, vlen, pos);
L
Linus Torvalds 已提交
1278 1279

out:
1280 1281 1282
	if (ret > 0)
		add_wchar(current, ret);
	inc_syscw(current);
1283 1284 1285 1286 1287 1288 1289 1290
	return ret;
}

asmlinkage ssize_t
compat_sys_writev(unsigned long fd, const struct compat_iovec __user *vec,
		  unsigned long vlen)
{
	struct file *file;
1291
	int fput_needed;
1292 1293
	ssize_t ret;

1294
	file = fget_light(fd, &fput_needed);
1295 1296 1297
	if (!file)
		return -EBADF;
	ret = compat_writev(file, vec, vlen, &file->f_pos);
1298
	fput_light(file, fput_needed);
L
Linus Torvalds 已提交
1299 1300 1301
	return ret;
}

1302 1303
asmlinkage ssize_t
compat_sys_pwritev(unsigned long fd, const struct compat_iovec __user *vec,
1304
		   unsigned long vlen, u32 pos_low, u32 pos_high)
1305 1306 1307
{
	loff_t pos = ((loff_t)pos_high << 32) | pos_low;
	struct file *file;
1308
	int fput_needed;
1309 1310 1311 1312
	ssize_t ret;

	if (pos < 0)
		return -EINVAL;
1313
	file = fget_light(fd, &fput_needed);
1314 1315 1316
	if (!file)
		return -EBADF;
	ret = compat_writev(file, vec, vlen, &pos);
1317
	fput_light(file, fput_needed);
1318 1319 1320
	return ret;
}

1321 1322 1323 1324 1325
asmlinkage long
compat_sys_vmsplice(int fd, const struct compat_iovec __user *iov32,
		    unsigned int nr_segs, unsigned int flags)
{
	unsigned i;
1326
	struct iovec __user *iov;
1327
	if (nr_segs > UIO_MAXIOV)
1328 1329 1330 1331 1332 1333 1334 1335 1336 1337 1338 1339 1340
		return -EINVAL;
	iov = compat_alloc_user_space(nr_segs * sizeof(struct iovec));
	for (i = 0; i < nr_segs; i++) {
		struct compat_iovec v;
		if (get_user(v.iov_base, &iov32[i].iov_base) ||
		    get_user(v.iov_len, &iov32[i].iov_len) ||
		    put_user(compat_ptr(v.iov_base), &iov[i].iov_base) ||
		    put_user(v.iov_len, &iov[i].iov_len))
			return -EFAULT;
	}
	return sys_vmsplice(fd, iov, nr_segs, flags);
}

1341 1342 1343 1344 1345 1346 1347
/*
 * Exactly like fs/open.c:sys_open(), except that it doesn't set the
 * O_LARGEFILE flag.
 */
asmlinkage long
compat_sys_open(const char __user *filename, int flags, int mode)
{
1348 1349 1350 1351 1352 1353 1354 1355
	return do_sys_open(AT_FDCWD, filename, flags, mode);
}

/*
 * Exactly like fs/open.c:sys_openat(), except that it doesn't set the
 * O_LARGEFILE flag.
 */
asmlinkage long
1356
compat_sys_openat(unsigned int dfd, const char __user *filename, int flags, int mode)
1357 1358
{
	return do_sys_open(dfd, filename, flags, mode);
1359 1360
}

L
Linus Torvalds 已提交
1361 1362 1363 1364 1365 1366 1367 1368 1369 1370 1371 1372 1373 1374 1375 1376 1377 1378
/*
 * compat_count() counts the number of arguments/envelopes. It is basically
 * a copy of count() from fs/exec.c, except that it works with 32 bit argv
 * and envp pointers.
 */
static int compat_count(compat_uptr_t __user *argv, int max)
{
	int i = 0;

	if (argv != NULL) {
		for (;;) {
			compat_uptr_t p;

			if (get_user(p, argv))
				return -EFAULT;
			if (!p)
				break;
			argv++;
1379
			if (i++ >= max)
L
Linus Torvalds 已提交
1380 1381 1382 1383 1384 1385 1386 1387 1388 1389 1390 1391 1392 1393 1394
				return -E2BIG;
		}
	}
	return i;
}

/*
 * compat_copy_strings() is basically a copy of copy_strings() from fs/exec.c
 * except that it works with 32 bit argv and envp pointers.
 */
static int compat_copy_strings(int argc, compat_uptr_t __user *argv,
				struct linux_binprm *bprm)
{
	struct page *kmapped_page = NULL;
	char *kaddr = NULL;
1395
	unsigned long kpos = 0;
L
Linus Torvalds 已提交
1396 1397 1398 1399 1400 1401 1402 1403
	int ret;

	while (argc-- > 0) {
		compat_uptr_t str;
		int len;
		unsigned long pos;

		if (get_user(str, argv+argc) ||
1404
		    !(len = strnlen_user(compat_ptr(str), MAX_ARG_STRLEN))) {
L
Linus Torvalds 已提交
1405 1406 1407 1408
			ret = -EFAULT;
			goto out;
		}

1409
		if (len > MAX_ARG_STRLEN) {
L
Linus Torvalds 已提交
1410 1411 1412 1413
			ret = -E2BIG;
			goto out;
		}

1414
		/* We're going to work our way backwords. */
L
Linus Torvalds 已提交
1415
		pos = bprm->p;
1416 1417
		str += len;
		bprm->p -= len;
L
Linus Torvalds 已提交
1418 1419 1420 1421 1422

		while (len > 0) {
			int offset, bytes_to_copy;

			offset = pos % PAGE_SIZE;
1423 1424 1425 1426 1427 1428 1429 1430 1431 1432 1433 1434 1435 1436 1437 1438 1439 1440 1441 1442 1443 1444 1445 1446 1447 1448 1449 1450
			if (offset == 0)
				offset = PAGE_SIZE;

			bytes_to_copy = offset;
			if (bytes_to_copy > len)
				bytes_to_copy = len;

			offset -= bytes_to_copy;
			pos -= bytes_to_copy;
			str -= bytes_to_copy;
			len -= bytes_to_copy;

			if (!kmapped_page || kpos != (pos & PAGE_MASK)) {
				struct page *page;

#ifdef CONFIG_STACK_GROWSUP
				ret = expand_stack_downwards(bprm->vma, pos);
				if (ret < 0) {
					/* We've exceed the stack rlimit. */
					ret = -E2BIG;
					goto out;
				}
#endif
				ret = get_user_pages(current, bprm->mm, pos,
						     1, 1, 1, &page, NULL);
				if (ret <= 0) {
					/* We've exceed the stack rlimit. */
					ret = -E2BIG;
L
Linus Torvalds 已提交
1451 1452 1453
					goto out;
				}

1454 1455
				if (kmapped_page) {
					flush_kernel_dcache_page(kmapped_page);
L
Linus Torvalds 已提交
1456
					kunmap(kmapped_page);
1457 1458
					put_page(kmapped_page);
				}
L
Linus Torvalds 已提交
1459 1460
				kmapped_page = page;
				kaddr = kmap(kmapped_page);
1461 1462 1463
				kpos = pos & PAGE_MASK;
				flush_cache_page(bprm->vma, kpos,
						 page_to_pfn(kmapped_page));
L
Linus Torvalds 已提交
1464
			}
1465 1466
			if (copy_from_user(kaddr+offset, compat_ptr(str),
						bytes_to_copy)) {
L
Linus Torvalds 已提交
1467 1468 1469 1470 1471 1472 1473
				ret = -EFAULT;
				goto out;
			}
		}
	}
	ret = 0;
out:
1474 1475
	if (kmapped_page) {
		flush_kernel_dcache_page(kmapped_page);
L
Linus Torvalds 已提交
1476
		kunmap(kmapped_page);
1477
		put_page(kmapped_page);
L
Linus Torvalds 已提交
1478
	}
1479
	return ret;
L
Linus Torvalds 已提交
1480 1481 1482 1483 1484 1485 1486 1487 1488 1489 1490 1491 1492
}

/*
 * compat_do_execve() is mostly a copy of do_execve(), with the exception
 * that it processes 32 bit argv and envp pointers.
 */
int compat_do_execve(char * filename,
	compat_uptr_t __user *argv,
	compat_uptr_t __user *envp,
	struct pt_regs * regs)
{
	struct linux_binprm *bprm;
	struct file *file;
1493
	struct files_struct *displaced;
1494
	bool clear_in_exec;
L
Linus Torvalds 已提交
1495 1496
	int retval;

1497 1498 1499 1500
	retval = unshare_files(&displaced);
	if (retval)
		goto out_ret;

L
Linus Torvalds 已提交
1501
	retval = -ENOMEM;
1502
	bprm = kzalloc(sizeof(*bprm), GFP_KERNEL);
L
Linus Torvalds 已提交
1503
	if (!bprm)
1504
		goto out_files;
L
Linus Torvalds 已提交
1505

1506 1507
	retval = prepare_bprm_creds(bprm);
	if (retval)
1508
		goto out_free;
A
Al Viro 已提交
1509 1510

	retval = check_unsafe_exec(bprm);
1511
	if (retval < 0)
1512
		goto out_free;
1513
	clear_in_exec = retval;
1514
	current->in_execve = 1;
1515

L
Linus Torvalds 已提交
1516 1517 1518
	file = open_exec(filename);
	retval = PTR_ERR(file);
	if (IS_ERR(file))
A
Al Viro 已提交
1519
		goto out_unmark;
L
Linus Torvalds 已提交
1520 1521 1522 1523 1524 1525 1526

	sched_exec();

	bprm->file = file;
	bprm->filename = filename;
	bprm->interp = filename;

1527 1528 1529
	retval = bprm_mm_init(bprm);
	if (retval)
		goto out_file;
L
Linus Torvalds 已提交
1530

1531
	bprm->argc = compat_count(argv, MAX_ARG_STRINGS);
L
Linus Torvalds 已提交
1532
	if ((retval = bprm->argc) < 0)
1533
		goto out;
L
Linus Torvalds 已提交
1534

1535
	bprm->envc = compat_count(envp, MAX_ARG_STRINGS);
L
Linus Torvalds 已提交
1536 1537 1538 1539 1540 1541 1542 1543 1544 1545 1546 1547 1548 1549 1550 1551 1552 1553 1554 1555 1556
	if ((retval = bprm->envc) < 0)
		goto out;

	retval = prepare_binprm(bprm);
	if (retval < 0)
		goto out;

	retval = copy_strings_kernel(1, &bprm->filename, bprm);
	if (retval < 0)
		goto out;

	bprm->exec = bprm->p;
	retval = compat_copy_strings(bprm->envc, envp, bprm);
	if (retval < 0)
		goto out;

	retval = compat_copy_strings(bprm->argc, argv, bprm);
	if (retval < 0)
		goto out;

	retval = search_binary_handler(bprm, regs);
1557 1558
	if (retval < 0)
		goto out;
L
Linus Torvalds 已提交
1559

1560
	/* execve succeeded */
A
Al Viro 已提交
1561
	current->fs->in_exec = 0;
1562
	current->in_execve = 0;
1563 1564
	acct_update_integrals(current);
	free_bprm(bprm);
1565 1566
	if (displaced)
		put_files_struct(displaced);
1567
	return retval;
L
Linus Torvalds 已提交
1568

1569
out:
L
Linus Torvalds 已提交
1570
	if (bprm->mm)
1571
		mmput(bprm->mm);
L
Linus Torvalds 已提交
1572 1573 1574 1575 1576 1577 1578

out_file:
	if (bprm->file) {
		allow_write_access(bprm->file);
		fput(bprm->file);
	}

A
Al Viro 已提交
1579
out_unmark:
1580 1581
	if (clear_in_exec)
		current->fs->in_exec = 0;
1582
	current->in_execve = 0;
1583 1584

out_free:
1585
	free_bprm(bprm);
L
Linus Torvalds 已提交
1586

1587 1588 1589
out_files:
	if (displaced)
		reset_files_struct(displaced);
L
Linus Torvalds 已提交
1590 1591 1592 1593 1594 1595
out_ret:
	return retval;
}

#define __COMPAT_NFDBITS       (8 * sizeof(compat_ulong_t))

1596 1597 1598 1599 1600 1601 1602 1603 1604 1605 1606 1607 1608 1609 1610 1611 1612 1613 1614 1615 1616 1617 1618 1619 1620 1621 1622 1623 1624 1625 1626 1627 1628 1629 1630 1631 1632 1633 1634 1635 1636 1637 1638 1639 1640 1641 1642 1643 1644 1645 1646
static int poll_select_copy_remaining(struct timespec *end_time, void __user *p,
				      int timeval, int ret)
{
	struct timespec ts;

	if (!p)
		return ret;

	if (current->personality & STICKY_TIMEOUTS)
		goto sticky;

	/* No update for zero timeout */
	if (!end_time->tv_sec && !end_time->tv_nsec)
		return ret;

	ktime_get_ts(&ts);
	ts = timespec_sub(*end_time, ts);
	if (ts.tv_sec < 0)
		ts.tv_sec = ts.tv_nsec = 0;

	if (timeval) {
		struct compat_timeval rtv;

		rtv.tv_sec = ts.tv_sec;
		rtv.tv_usec = ts.tv_nsec / NSEC_PER_USEC;

		if (!copy_to_user(p, &rtv, sizeof(rtv)))
			return ret;
	} else {
		struct compat_timespec rts;

		rts.tv_sec = ts.tv_sec;
		rts.tv_nsec = ts.tv_nsec;

		if (!copy_to_user(p, &rts, sizeof(rts)))
			return ret;
	}
	/*
	 * If an application puts its timeval in read-only memory, we
	 * don't want the Linux-specific update to the timeval to
	 * cause a fault after the select has completed
	 * successfully. However, because we're not updating the
	 * timeval, we can't restart the system call.
	 */

sticky:
	if (ret == -ERESTARTNOHAND)
		ret = -EINTR;
	return ret;
}

L
Linus Torvalds 已提交
1647 1648 1649 1650
/*
 * Ooo, nasty.  We need here to frob 32-bit unsigned longs to
 * 64-bit unsigned longs.
 */
1651
static
L
Linus Torvalds 已提交
1652 1653 1654
int compat_get_fd_set(unsigned long nr, compat_ulong_t __user *ufdset,
			unsigned long *fdset)
{
1655
	nr = DIV_ROUND_UP(nr, __COMPAT_NFDBITS);
L
Linus Torvalds 已提交
1656 1657 1658 1659 1660 1661 1662 1663 1664 1665
	if (ufdset) {
		unsigned long odd;

		if (!access_ok(VERIFY_WRITE, ufdset, nr*sizeof(compat_ulong_t)))
			return -EFAULT;

		odd = nr & 1UL;
		nr &= ~1UL;
		while (nr) {
			unsigned long h, l;
1666 1667
			if (__get_user(l, ufdset) || __get_user(h, ufdset+1))
				return -EFAULT;
L
Linus Torvalds 已提交
1668 1669 1670 1671
			ufdset += 2;
			*fdset++ = h << 32 | l;
			nr -= 2;
		}
1672 1673
		if (odd && __get_user(*fdset, ufdset))
			return -EFAULT;
L
Linus Torvalds 已提交
1674 1675 1676 1677 1678 1679 1680 1681 1682 1683
	} else {
		/* Tricky, must clear full unsigned long in the
		 * kernel fdset at the end, this makes sure that
		 * actually happens.
		 */
		memset(fdset, 0, ((nr + 1) & ~1)*sizeof(compat_ulong_t));
	}
	return 0;
}

1684
static
1685 1686
int compat_set_fd_set(unsigned long nr, compat_ulong_t __user *ufdset,
		      unsigned long *fdset)
L
Linus Torvalds 已提交
1687 1688
{
	unsigned long odd;
1689
	nr = DIV_ROUND_UP(nr, __COMPAT_NFDBITS);
L
Linus Torvalds 已提交
1690 1691

	if (!ufdset)
1692
		return 0;
L
Linus Torvalds 已提交
1693 1694 1695 1696 1697 1698 1699

	odd = nr & 1UL;
	nr &= ~1UL;
	while (nr) {
		unsigned long h, l;
		l = *fdset++;
		h = l >> 32;
1700 1701
		if (__put_user(l, ufdset) || __put_user(h, ufdset+1))
			return -EFAULT;
L
Linus Torvalds 已提交
1702 1703 1704
		ufdset += 2;
		nr -= 2;
	}
1705 1706 1707
	if (odd && __put_user(*fdset, ufdset))
		return -EFAULT;
	return 0;
L
Linus Torvalds 已提交
1708 1709 1710 1711 1712 1713 1714 1715 1716 1717 1718 1719 1720 1721 1722 1723 1724 1725 1726
}


/*
 * This is a virtual copy of sys_select from fs/select.c and probably
 * should be compared to it from time to time
 */

/*
 * We can actually return ERESTARTSYS instead of EINTR, but I'd
 * like to be certain this leads to no problems. So I return
 * EINTR just for safety.
 *
 * Update: ERESTARTSYS breaks at least the xview clock binary, so
 * I'm trying ERESTARTNOHAND which restart only when you want to.
 */
#define MAX_SELECT_SECONDS \
	((unsigned long) (MAX_SCHEDULE_TIMEOUT / HZ)-1)

1727
int compat_core_sys_select(int n, compat_ulong_t __user *inp,
1728 1729
	compat_ulong_t __user *outp, compat_ulong_t __user *exp,
	struct timespec *end_time)
L
Linus Torvalds 已提交
1730 1731
{
	fd_set_bits fds;
1732
	void *bits;
1733
	int size, max_fds, ret = -EINVAL;
1734
	struct fdtable *fdt;
1735
	long stack_fds[SELECT_STACK_ALLOC/sizeof(long)];
L
Linus Torvalds 已提交
1736 1737 1738 1739

	if (n < 0)
		goto out_nofds;

1740
	/* max_fds can increase, so grab it once to avoid race */
L
Linus Torvalds 已提交
1741
	rcu_read_lock();
1742
	fdt = files_fdtable(current->files);
1743
	max_fds = fdt->max_fds;
L
Linus Torvalds 已提交
1744
	rcu_read_unlock();
1745 1746
	if (n > max_fds)
		n = max_fds;
L
Linus Torvalds 已提交
1747 1748 1749 1750 1751 1752 1753

	/*
	 * We need 6 bitmaps (in/out/ex for both incoming and outgoing),
	 * since we used fdset we need to allocate memory in units of
	 * long-words.
	 */
	size = FDS_BYTES(n);
1754 1755 1756 1757 1758 1759 1760
	bits = stack_fds;
	if (size > sizeof(stack_fds) / 6) {
		bits = kmalloc(6 * size, GFP_KERNEL);
		ret = -ENOMEM;
		if (!bits)
			goto out_nofds;
	}
L
Linus Torvalds 已提交
1761 1762 1763 1764 1765 1766 1767 1768 1769 1770 1771 1772 1773 1774 1775
	fds.in      = (unsigned long *)  bits;
	fds.out     = (unsigned long *) (bits +   size);
	fds.ex      = (unsigned long *) (bits + 2*size);
	fds.res_in  = (unsigned long *) (bits + 3*size);
	fds.res_out = (unsigned long *) (bits + 4*size);
	fds.res_ex  = (unsigned long *) (bits + 5*size);

	if ((ret = compat_get_fd_set(n, inp, fds.in)) ||
	    (ret = compat_get_fd_set(n, outp, fds.out)) ||
	    (ret = compat_get_fd_set(n, exp, fds.ex)))
		goto out;
	zero_fd_set(n, fds.res_in);
	zero_fd_set(n, fds.res_out);
	zero_fd_set(n, fds.res_ex);

1776
	ret = do_select(n, &fds, end_time);
L
Linus Torvalds 已提交
1777 1778 1779 1780 1781 1782 1783 1784 1785 1786

	if (ret < 0)
		goto out;
	if (!ret) {
		ret = -ERESTARTNOHAND;
		if (signal_pending(current))
			goto out;
		ret = 0;
	}

1787 1788 1789 1790
	if (compat_set_fd_set(n, inp, fds.res_in) ||
	    compat_set_fd_set(n, outp, fds.res_out) ||
	    compat_set_fd_set(n, exp, fds.res_ex))
		ret = -EFAULT;
L
Linus Torvalds 已提交
1791
out:
1792 1793
	if (bits != stack_fds)
		kfree(bits);
L
Linus Torvalds 已提交
1794 1795 1796 1797
out_nofds:
	return ret;
}

1798 1799 1800 1801
asmlinkage long compat_sys_select(int n, compat_ulong_t __user *inp,
	compat_ulong_t __user *outp, compat_ulong_t __user *exp,
	struct compat_timeval __user *tvp)
{
1802
	struct timespec end_time, *to = NULL;
1803 1804 1805 1806 1807 1808 1809
	struct compat_timeval tv;
	int ret;

	if (tvp) {
		if (copy_from_user(&tv, tvp, sizeof(tv)))
			return -EFAULT;

1810
		to = &end_time;
1811 1812 1813
		if (poll_select_set_timeout(to,
				tv.tv_sec + (tv.tv_usec / USEC_PER_SEC),
				(tv.tv_usec % USEC_PER_SEC) * NSEC_PER_USEC))
1814 1815 1816
			return -EINVAL;
	}

1817 1818
	ret = compat_core_sys_select(n, inp, outp, exp, to);
	ret = poll_select_copy_remaining(&end_time, tvp, 1, ret);
1819 1820 1821 1822

	return ret;
}

C
Christoph Hellwig 已提交
1823 1824 1825 1826 1827 1828 1829 1830 1831 1832 1833 1834 1835 1836 1837 1838 1839 1840
struct compat_sel_arg_struct {
	compat_ulong_t n;
	compat_uptr_t inp;
	compat_uptr_t outp;
	compat_uptr_t exp;
	compat_uptr_t tvp;
};

asmlinkage long compat_sys_old_select(struct compat_sel_arg_struct __user *arg)
{
	struct compat_sel_arg_struct a;

	if (copy_from_user(&a, arg, sizeof(a)))
		return -EFAULT;
	return compat_sys_select(a.n, compat_ptr(a.inp), compat_ptr(a.outp),
				 compat_ptr(a.exp), compat_ptr(a.tvp));
}

1841
#ifdef HAVE_SET_RESTORE_SIGMASK
1842
static long do_compat_pselect(int n, compat_ulong_t __user *inp,
1843 1844 1845 1846 1847 1848 1849
	compat_ulong_t __user *outp, compat_ulong_t __user *exp,
	struct compat_timespec __user *tsp, compat_sigset_t __user *sigmask,
	compat_size_t sigsetsize)
{
	compat_sigset_t ss32;
	sigset_t ksigmask, sigsaved;
	struct compat_timespec ts;
1850
	struct timespec end_time, *to = NULL;
1851 1852 1853 1854 1855 1856
	int ret;

	if (tsp) {
		if (copy_from_user(&ts, tsp, sizeof(ts)))
			return -EFAULT;

1857 1858
		to = &end_time;
		if (poll_select_set_timeout(to, ts.tv_sec, ts.tv_nsec))
1859 1860 1861 1862 1863 1864 1865 1866 1867 1868 1869 1870 1871 1872
			return -EINVAL;
	}

	if (sigmask) {
		if (sigsetsize != sizeof(compat_sigset_t))
			return -EINVAL;
		if (copy_from_user(&ss32, sigmask, sizeof(ss32)))
			return -EFAULT;
		sigset_from_compat(&ksigmask, &ss32);

		sigdelsetmask(&ksigmask, sigmask(SIGKILL)|sigmask(SIGSTOP));
		sigprocmask(SIG_SETMASK, &ksigmask, &sigsaved);
	}

1873 1874
	ret = compat_core_sys_select(n, inp, outp, exp, to);
	ret = poll_select_copy_remaining(&end_time, tsp, 0, ret);
1875 1876 1877 1878 1879 1880 1881 1882 1883 1884

	if (ret == -ERESTARTNOHAND) {
		/*
		 * Don't restore the signal mask yet. Let do_signal() deliver
		 * the signal on the way back to userspace, before the signal
		 * mask is restored.
		 */
		if (sigmask) {
			memcpy(&current->saved_sigmask, &sigsaved,
					sizeof(sigsaved));
1885
			set_restore_sigmask();
1886 1887 1888 1889 1890 1891 1892 1893 1894 1895 1896 1897 1898 1899 1900 1901 1902 1903 1904 1905 1906 1907
		}
	} else if (sigmask)
		sigprocmask(SIG_SETMASK, &sigsaved, NULL);

	return ret;
}

asmlinkage long compat_sys_pselect6(int n, compat_ulong_t __user *inp,
	compat_ulong_t __user *outp, compat_ulong_t __user *exp,
	struct compat_timespec __user *tsp, void __user *sig)
{
	compat_size_t sigsetsize = 0;
	compat_uptr_t up = 0;

	if (sig) {
		if (!access_ok(VERIFY_READ, sig,
				sizeof(compat_uptr_t)+sizeof(compat_size_t)) ||
		    	__get_user(up, (compat_uptr_t __user *)sig) ||
		    	__get_user(sigsetsize,
				(compat_size_t __user *)(sig+sizeof(up))))
			return -EFAULT;
	}
1908 1909
	return do_compat_pselect(n, inp, outp, exp, tsp, compat_ptr(up),
				 sigsetsize);
1910 1911 1912 1913 1914 1915 1916 1917 1918
}

asmlinkage long compat_sys_ppoll(struct pollfd __user *ufds,
	unsigned int nfds, struct compat_timespec __user *tsp,
	const compat_sigset_t __user *sigmask, compat_size_t sigsetsize)
{
	compat_sigset_t ss32;
	sigset_t ksigmask, sigsaved;
	struct compat_timespec ts;
1919
	struct timespec end_time, *to = NULL;
1920 1921 1922 1923 1924 1925
	int ret;

	if (tsp) {
		if (copy_from_user(&ts, tsp, sizeof(ts)))
			return -EFAULT;

1926 1927 1928
		to = &end_time;
		if (poll_select_set_timeout(to, ts.tv_sec, ts.tv_nsec))
			return -EINVAL;
1929 1930 1931
	}

	if (sigmask) {
1932
		if (sigsetsize != sizeof(compat_sigset_t))
1933 1934 1935 1936 1937 1938 1939 1940 1941
			return -EINVAL;
		if (copy_from_user(&ss32, sigmask, sizeof(ss32)))
			return -EFAULT;
		sigset_from_compat(&ksigmask, &ss32);

		sigdelsetmask(&ksigmask, sigmask(SIGKILL)|sigmask(SIGSTOP));
		sigprocmask(SIG_SETMASK, &ksigmask, &sigsaved);
	}

1942
	ret = do_sys_poll(ufds, nfds, to);
1943 1944 1945 1946 1947 1948 1949 1950 1951 1952 1953

	/* We can restart this syscall, usually */
	if (ret == -EINTR) {
		/*
		 * Don't restore the signal mask yet. Let do_signal() deliver
		 * the signal on the way back to userspace, before the signal
		 * mask is restored.
		 */
		if (sigmask) {
			memcpy(&current->saved_sigmask, &sigsaved,
				sizeof(sigsaved));
1954
			set_restore_sigmask();
1955 1956 1957 1958 1959
		}
		ret = -ERESTARTNOHAND;
	} else if (sigmask)
		sigprocmask(SIG_SETMASK, &sigsaved, NULL);

1960
	ret = poll_select_copy_remaining(&end_time, tsp, 0, ret);
1961 1962 1963

	return ret;
}
1964
#endif /* HAVE_SET_RESTORE_SIGMASK */
1965

L
Linus Torvalds 已提交
1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987
#if defined(CONFIG_NFSD) || defined(CONFIG_NFSD_MODULE)
/* Stuff for NFS server syscalls... */
struct compat_nfsctl_svc {
	u16			svc32_port;
	s32			svc32_nthreads;
};

struct compat_nfsctl_client {
	s8			cl32_ident[NFSCLNT_IDMAX+1];
	s32			cl32_naddr;
	struct in_addr		cl32_addrlist[NFSCLNT_ADDRMAX];
	s32			cl32_fhkeytype;
	s32			cl32_fhkeylen;
	u8			cl32_fhkey[NFSCLNT_KEYMAX];
};

struct compat_nfsctl_export {
	char		ex32_client[NFSCLNT_IDMAX+1];
	char		ex32_path[NFS_MAXPATHLEN+1];
	compat_dev_t	ex32_dev;
	compat_ino_t	ex32_ino;
	compat_int_t	ex32_flags;
1988 1989
	__compat_uid_t	ex32_anon_uid;
	__compat_gid_t	ex32_anon_gid;
L
Linus Torvalds 已提交
1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024
};

struct compat_nfsctl_fdparm {
	struct sockaddr		gd32_addr;
	s8			gd32_path[NFS_MAXPATHLEN+1];
	compat_int_t		gd32_version;
};

struct compat_nfsctl_fsparm {
	struct sockaddr		gd32_addr;
	s8			gd32_path[NFS_MAXPATHLEN+1];
	compat_int_t		gd32_maxlen;
};

struct compat_nfsctl_arg {
	compat_int_t		ca32_version;	/* safeguard */
	union {
		struct compat_nfsctl_svc	u32_svc;
		struct compat_nfsctl_client	u32_client;
		struct compat_nfsctl_export	u32_export;
		struct compat_nfsctl_fdparm	u32_getfd;
		struct compat_nfsctl_fsparm	u32_getfs;
	} u;
#define ca32_svc	u.u32_svc
#define ca32_client	u.u32_client
#define ca32_export	u.u32_export
#define ca32_getfd	u.u32_getfd
#define ca32_getfs	u.u32_getfs
};

union compat_nfsctl_res {
	__u8			cr32_getfh[NFS_FHSIZE];
	struct knfsd_fh		cr32_getfs;
};

2025 2026
static int compat_nfs_svc_trans(struct nfsctl_arg *karg,
				struct compat_nfsctl_arg __user *arg)
L
Linus Torvalds 已提交
2027
{
2028 2029 2030 2031 2032 2033 2034
	if (!access_ok(VERIFY_READ, &arg->ca32_svc, sizeof(arg->ca32_svc)) ||
		get_user(karg->ca_version, &arg->ca32_version) ||
		__get_user(karg->ca_svc.svc_port, &arg->ca32_svc.svc32_port) ||
		__get_user(karg->ca_svc.svc_nthreads,
				&arg->ca32_svc.svc32_nthreads))
		return -EFAULT;
	return 0;
L
Linus Torvalds 已提交
2035 2036
}

2037 2038
static int compat_nfs_clnt_trans(struct nfsctl_arg *karg,
				struct compat_nfsctl_arg __user *arg)
L
Linus Torvalds 已提交
2039
{
2040 2041 2042 2043 2044 2045 2046 2047 2048 2049 2050 2051 2052 2053 2054 2055 2056 2057 2058
	if (!access_ok(VERIFY_READ, &arg->ca32_client,
			sizeof(arg->ca32_client)) ||
		get_user(karg->ca_version, &arg->ca32_version) ||
		__copy_from_user(&karg->ca_client.cl_ident[0],
				&arg->ca32_client.cl32_ident[0],
				NFSCLNT_IDMAX) ||
		__get_user(karg->ca_client.cl_naddr,
				&arg->ca32_client.cl32_naddr) ||
		__copy_from_user(&karg->ca_client.cl_addrlist[0],
				&arg->ca32_client.cl32_addrlist[0],
				(sizeof(struct in_addr) * NFSCLNT_ADDRMAX)) ||
		__get_user(karg->ca_client.cl_fhkeytype,
				&arg->ca32_client.cl32_fhkeytype) ||
		__get_user(karg->ca_client.cl_fhkeylen,
				&arg->ca32_client.cl32_fhkeylen) ||
		__copy_from_user(&karg->ca_client.cl_fhkey[0],
				&arg->ca32_client.cl32_fhkey[0],
				NFSCLNT_KEYMAX))
		return -EFAULT;
L
Linus Torvalds 已提交
2059

2060
	return 0;
L
Linus Torvalds 已提交
2061 2062
}

2063 2064
static int compat_nfs_exp_trans(struct nfsctl_arg *karg,
				struct compat_nfsctl_arg __user *arg)
L
Linus Torvalds 已提交
2065
{
2066 2067 2068 2069 2070 2071 2072 2073 2074 2075 2076 2077 2078 2079 2080 2081 2082 2083 2084 2085
	if (!access_ok(VERIFY_READ, &arg->ca32_export,
				sizeof(arg->ca32_export)) ||
		get_user(karg->ca_version, &arg->ca32_version) ||
		__copy_from_user(&karg->ca_export.ex_client[0],
				&arg->ca32_export.ex32_client[0],
				NFSCLNT_IDMAX) ||
		__copy_from_user(&karg->ca_export.ex_path[0],
				&arg->ca32_export.ex32_path[0],
				NFS_MAXPATHLEN) ||
		__get_user(karg->ca_export.ex_dev,
				&arg->ca32_export.ex32_dev) ||
		__get_user(karg->ca_export.ex_ino,
				&arg->ca32_export.ex32_ino) ||
		__get_user(karg->ca_export.ex_flags,
				&arg->ca32_export.ex32_flags) ||
		__get_user(karg->ca_export.ex_anon_uid,
				&arg->ca32_export.ex32_anon_uid) ||
		__get_user(karg->ca_export.ex_anon_gid,
				&arg->ca32_export.ex32_anon_gid))
		return -EFAULT;
L
Linus Torvalds 已提交
2086 2087 2088
	SET_UID(karg->ca_export.ex_anon_uid, karg->ca_export.ex_anon_uid);
	SET_GID(karg->ca_export.ex_anon_gid, karg->ca_export.ex_anon_gid);

2089
	return 0;
L
Linus Torvalds 已提交
2090 2091
}

2092 2093
static int compat_nfs_getfd_trans(struct nfsctl_arg *karg,
				struct compat_nfsctl_arg __user *arg)
L
Linus Torvalds 已提交
2094
{
2095 2096 2097 2098 2099 2100 2101 2102 2103 2104 2105 2106
	if (!access_ok(VERIFY_READ, &arg->ca32_getfd,
			sizeof(arg->ca32_getfd)) ||
		get_user(karg->ca_version, &arg->ca32_version) ||
		__copy_from_user(&karg->ca_getfd.gd_addr,
				&arg->ca32_getfd.gd32_addr,
				(sizeof(struct sockaddr))) ||
		__copy_from_user(&karg->ca_getfd.gd_path,
				&arg->ca32_getfd.gd32_path,
				(NFS_MAXPATHLEN+1)) ||
		__get_user(karg->ca_getfd.gd_version,
				&arg->ca32_getfd.gd32_version))
		return -EFAULT;
L
Linus Torvalds 已提交
2107

2108
	return 0;
L
Linus Torvalds 已提交
2109 2110
}

2111 2112
static int compat_nfs_getfs_trans(struct nfsctl_arg *karg,
				struct compat_nfsctl_arg __user *arg)
L
Linus Torvalds 已提交
2113
{
2114 2115 2116 2117 2118 2119 2120 2121 2122 2123 2124
	if (!access_ok(VERIFY_READ,&arg->ca32_getfs,sizeof(arg->ca32_getfs)) ||
		get_user(karg->ca_version, &arg->ca32_version) ||
		__copy_from_user(&karg->ca_getfs.gd_addr,
				&arg->ca32_getfs.gd32_addr,
				(sizeof(struct sockaddr))) ||
		__copy_from_user(&karg->ca_getfs.gd_path,
				&arg->ca32_getfs.gd32_path,
				(NFS_MAXPATHLEN+1)) ||
		__get_user(karg->ca_getfs.gd_maxlen,
				&arg->ca32_getfs.gd32_maxlen))
		return -EFAULT;
L
Linus Torvalds 已提交
2125

2126
	return 0;
L
Linus Torvalds 已提交
2127 2128 2129 2130 2131
}

/* This really doesn't need translations, we are only passing
 * back a union which contains opaque nfs file handle data.
 */
2132 2133
static int compat_nfs_getfh_res_trans(union nfsctl_res *kres,
				union compat_nfsctl_res __user *res)
L
Linus Torvalds 已提交
2134 2135 2136 2137 2138 2139 2140 2141
{
	int err;

	err = copy_to_user(res, kres, sizeof(*res));

	return (err) ? -EFAULT : 0;
}

2142 2143 2144
asmlinkage long compat_sys_nfsservctl(int cmd,
				struct compat_nfsctl_arg __user *arg,
				union compat_nfsctl_res __user *res)
L
Linus Torvalds 已提交
2145 2146 2147 2148 2149 2150 2151 2152 2153 2154 2155 2156 2157 2158 2159 2160 2161 2162 2163 2164 2165 2166 2167 2168 2169 2170 2171 2172 2173 2174 2175 2176 2177 2178 2179 2180 2181 2182 2183 2184 2185
{
	struct nfsctl_arg *karg;
	union nfsctl_res *kres;
	mm_segment_t oldfs;
	int err;

	karg = kmalloc(sizeof(*karg), GFP_USER);
	kres = kmalloc(sizeof(*kres), GFP_USER);
	if(!karg || !kres) {
		err = -ENOMEM;
		goto done;
	}

	switch(cmd) {
	case NFSCTL_SVC:
		err = compat_nfs_svc_trans(karg, arg);
		break;

	case NFSCTL_ADDCLIENT:
		err = compat_nfs_clnt_trans(karg, arg);
		break;

	case NFSCTL_DELCLIENT:
		err = compat_nfs_clnt_trans(karg, arg);
		break;

	case NFSCTL_EXPORT:
	case NFSCTL_UNEXPORT:
		err = compat_nfs_exp_trans(karg, arg);
		break;

	case NFSCTL_GETFD:
		err = compat_nfs_getfd_trans(karg, arg);
		break;

	case NFSCTL_GETFS:
		err = compat_nfs_getfs_trans(karg, arg);
		break;

	default:
		err = -EINVAL;
2186
		break;
L
Linus Torvalds 已提交
2187 2188
	}

2189 2190 2191
	if (err)
		goto done;

L
Linus Torvalds 已提交
2192 2193 2194 2195 2196 2197 2198 2199 2200 2201 2202 2203 2204 2205 2206 2207 2208 2209 2210 2211 2212 2213 2214 2215
	oldfs = get_fs();
	set_fs(KERNEL_DS);
	/* The __user pointer casts are valid because of the set_fs() */
	err = sys_nfsservctl(cmd, (void __user *) karg, (void __user *) kres);
	set_fs(oldfs);

	if (err)
		goto done;

	if((cmd == NFSCTL_GETFD) ||
	   (cmd == NFSCTL_GETFS))
		err = compat_nfs_getfh_res_trans(kres, res);

done:
	kfree(karg);
	kfree(kres);
	return err;
}
#else /* !NFSD */
long asmlinkage compat_sys_nfsservctl(int cmd, void *notused, void *notused2)
{
	return sys_ni_syscall();
}
#endif
2216 2217 2218

#ifdef CONFIG_EPOLL

2219
#ifdef HAVE_SET_RESTORE_SIGMASK
2220 2221 2222 2223 2224 2225 2226 2227 2228 2229 2230 2231 2232 2233 2234 2235 2236 2237 2238 2239 2240 2241 2242 2243 2244 2245 2246 2247 2248 2249 2250 2251 2252 2253 2254 2255
asmlinkage long compat_sys_epoll_pwait(int epfd,
			struct compat_epoll_event __user *events,
			int maxevents, int timeout,
			const compat_sigset_t __user *sigmask,
			compat_size_t sigsetsize)
{
	long err;
	compat_sigset_t csigmask;
	sigset_t ksigmask, sigsaved;

	/*
	 * If the caller wants a certain signal mask to be set during the wait,
	 * we apply it here.
	 */
	if (sigmask) {
		if (sigsetsize != sizeof(compat_sigset_t))
			return -EINVAL;
		if (copy_from_user(&csigmask, sigmask, sizeof(csigmask)))
			return -EFAULT;
		sigset_from_compat(&ksigmask, &csigmask);
		sigdelsetmask(&ksigmask, sigmask(SIGKILL) | sigmask(SIGSTOP));
		sigprocmask(SIG_SETMASK, &ksigmask, &sigsaved);
	}

	err = sys_epoll_wait(epfd, events, maxevents, timeout);

	/*
	 * If we changed the signal mask, we need to restore the original one.
	 * In case we've got a signal while waiting, we do not restore the
	 * signal mask yet, and we allow do_signal() to deliver the signal on
	 * the way back to userspace, before the signal mask is restored.
	 */
	if (sigmask) {
		if (err == -EINTR) {
			memcpy(&current->saved_sigmask, &sigsaved,
			       sizeof(sigsaved));
2256
			set_restore_sigmask();
2257 2258 2259 2260 2261 2262
		} else
			sigprocmask(SIG_SETMASK, &sigsaved, NULL);
	}

	return err;
}
2263
#endif /* HAVE_SET_RESTORE_SIGMASK */
2264 2265

#endif /* CONFIG_EPOLL */
2266 2267 2268

#ifdef CONFIG_SIGNALFD

U
Ulrich Drepper 已提交
2269 2270 2271
asmlinkage long compat_sys_signalfd4(int ufd,
				     const compat_sigset_t __user *sigmask,
				     compat_size_t sigsetsize, int flags)
2272 2273 2274 2275 2276 2277 2278 2279 2280 2281 2282 2283 2284 2285
{
	compat_sigset_t ss32;
	sigset_t tmp;
	sigset_t __user *ksigmask;

	if (sigsetsize != sizeof(compat_sigset_t))
		return -EINVAL;
	if (copy_from_user(&ss32, sigmask, sizeof(ss32)))
		return -EFAULT;
	sigset_from_compat(&tmp, &ss32);
	ksigmask = compat_alloc_user_space(sizeof(sigset_t));
	if (copy_to_user(ksigmask, &tmp, sizeof(sigset_t)))
		return -EFAULT;

U
Ulrich Drepper 已提交
2286
	return sys_signalfd4(ufd, ksigmask, sizeof(sigset_t), flags);
2287 2288
}

U
Ulrich Drepper 已提交
2289 2290 2291 2292 2293 2294
asmlinkage long compat_sys_signalfd(int ufd,
				    const compat_sigset_t __user *sigmask,
				    compat_size_t sigsetsize)
{
	return compat_sys_signalfd4(ufd, sigmask, sigsetsize, 0);
}
2295 2296
#endif /* CONFIG_SIGNALFD */

2297 2298
#ifdef CONFIG_TIMERFD

D
Davide Libenzi 已提交
2299 2300 2301
asmlinkage long compat_sys_timerfd_settime(int ufd, int flags,
				   const struct compat_itimerspec __user *utmr,
				   struct compat_itimerspec __user *otmr)
2302
{
D
Davide Libenzi 已提交
2303
	int error;
2304 2305 2306 2307
	struct itimerspec t;
	struct itimerspec __user *ut;

	if (get_compat_itimerspec(&t, utmr))
H
Heiko Carstens 已提交
2308
		return -EFAULT;
D
Davide Libenzi 已提交
2309 2310
	ut = compat_alloc_user_space(2 * sizeof(struct itimerspec));
	if (copy_to_user(&ut[0], &t, sizeof(t)))
H
Heiko Carstens 已提交
2311
		return -EFAULT;
D
Davide Libenzi 已提交
2312 2313 2314 2315 2316 2317 2318 2319 2320 2321 2322 2323 2324 2325
	error = sys_timerfd_settime(ufd, flags, &ut[0], &ut[1]);
	if (!error && otmr)
		error = (copy_from_user(&t, &ut[1], sizeof(struct itimerspec)) ||
			 put_compat_itimerspec(otmr, &t)) ? -EFAULT: 0;

	return error;
}

asmlinkage long compat_sys_timerfd_gettime(int ufd,
				   struct compat_itimerspec __user *otmr)
{
	int error;
	struct itimerspec t;
	struct itimerspec __user *ut;
2326

D
Davide Libenzi 已提交
2327 2328 2329 2330 2331 2332 2333
	ut = compat_alloc_user_space(sizeof(struct itimerspec));
	error = sys_timerfd_gettime(ufd, ut);
	if (!error)
		error = (copy_from_user(&t, ut, sizeof(struct itimerspec)) ||
			 put_compat_itimerspec(otmr, &t)) ? -EFAULT: 0;

	return error;
2334 2335 2336
}

#endif /* CONFIG_TIMERFD */