1// SPDX-License-Identifier: GPL-2.0-or-later 2/* Common capabilities, needed by capability.o. 3 */ 4 5#include <linux/capability.h> 6#include <linux/audit.h> 7#include <linux/init.h> 8#include <linux/kernel.h> 9#include <linux/lsm_hooks.h> 10#include <linux/file.h> 11#include <linux/mm.h> 12#include <linux/mman.h> 13#include <linux/pagemap.h> 14#include <linux/swap.h> 15#include <linux/skbuff.h> 16#include <linux/netlink.h> 17#include <linux/ptrace.h> 18#include <linux/xattr.h> 19#include <linux/hugetlb.h> 20#include <linux/mount.h> 21#include <linux/sched.h> 22#include <linux/prctl.h> 23#include <linux/securebits.h> 24#include <linux/user_namespace.h> 25#include <linux/binfmts.h> 26#include <linux/personality.h> 27#include <linux/mnt_idmapping.h> 28 29/* 30 * If a non-root user executes a setuid-root binary in 31 * !secure(SECURE_NOROOT) mode, then we raise capabilities. 32 * However if fE is also set, then the intent is for only 33 * the file capabilities to be applied, and the setuid-root 34 * bit is left on either to change the uid (plausible) or 35 * to get full privilege on a kernel without file capabilities 36 * support. So in that case we do not raise capabilities. 37 * 38 * Warn if that happens, once per boot. 39 */ 40static void warn_setuid_and_fcaps_mixed(const char *fname) 41{ 42 static int warned; 43 if (!warned) { 44 printk(KERN_INFO "warning: `%s' has both setuid-root and" 45 " effective capabilities. Therefore not raising all" 46 " capabilities.\n", fname); 47 warned = 1; 48 } 49} 50 51/** 52 * cap_capable - Determine whether a task has a particular effective capability 53 * @cred: The credentials to use 54 * @targ_ns: The user namespace in which we need the capability 55 * @cap: The capability to check for 56 * @opts: Bitmask of options defined in include/linux/security.h 57 * 58 * Determine whether the nominated task has the specified capability amongst 59 * its effective set, returning 0 if it does, -ve if it does not. 60 * 61 * NOTE WELL: cap_has_capability() cannot be used like the kernel's capable() 62 * and has_capability() functions. That is, it has the reverse semantics: 63 * cap_has_capability() returns 0 when a task has a capability, but the 64 * kernel's capable() and has_capability() returns 1 for this case. 65 */ 66int cap_capable(const struct cred *cred, struct user_namespace *targ_ns, 67 int cap, unsigned int opts) 68{ 69 struct user_namespace *ns = targ_ns; 70 71 /* See if cred has the capability in the target user namespace 72 * by examining the target user namespace and all of the target 73 * user namespace's parents. 74 */ 75 for (;;) { 76 /* Do we have the necessary capabilities? */ 77 if (ns == cred->user_ns) 78 return cap_raised(cred->cap_effective, cap) ? 0 : -EPERM; 79 80 /* 81 * If we're already at a lower level than we're looking for, 82 * we're done searching. 83 */ 84 if (ns->level <= cred->user_ns->level) 85 return -EPERM; 86 87 /* 88 * The owner of the user namespace in the parent of the 89 * user namespace has all caps. 90 */ 91 if ((ns->parent == cred->user_ns) && uid_eq(ns->owner, cred->euid)) 92 return 0; 93 94 /* 95 * If you have a capability in a parent user ns, then you have 96 * it over all children user namespaces as well. 97 */ 98 ns = ns->parent; 99 } 100 101 /* We never get here */ 102} 103 104/** 105 * cap_settime - Determine whether the current process may set the system clock 106 * @ts: The time to set 107 * @tz: The timezone to set 108 * 109 * Determine whether the current process may set the system clock and timezone 110 * information, returning 0 if permission granted, -ve if denied. 111 */ 112int cap_settime(const struct timespec64 *ts, const struct timezone *tz) 113{ 114 if (!capable(CAP_SYS_TIME)) 115 return -EPERM; 116 return 0; 117} 118 119/** 120 * cap_ptrace_access_check - Determine whether the current process may access 121 * another 122 * @child: The process to be accessed 123 * @mode: The mode of attachment. 124 * 125 * If we are in the same or an ancestor user_ns and have all the target 126 * task's capabilities, then ptrace access is allowed. 127 * If we have the ptrace capability to the target user_ns, then ptrace 128 * access is allowed. 129 * Else denied. 130 * 131 * Determine whether a process may access another, returning 0 if permission 132 * granted, -ve if denied. 133 */ 134int cap_ptrace_access_check(struct task_struct *child, unsigned int mode) 135{ 136 int ret = 0; 137 const struct cred *cred, *child_cred; 138 const kernel_cap_t *caller_caps; 139 140 rcu_read_lock(); 141 cred = current_cred(); 142 child_cred = __task_cred(child); 143 if (mode & PTRACE_MODE_FSCREDS) 144 caller_caps = &cred->cap_effective; 145 else 146 caller_caps = &cred->cap_permitted; 147 if (cred->user_ns == child_cred->user_ns && 148 cap_issubset(child_cred->cap_permitted, *caller_caps)) 149 goto out; 150 if (ns_capable(child_cred->user_ns, CAP_SYS_PTRACE)) 151 goto out; 152 ret = -EPERM; 153out: 154 rcu_read_unlock(); 155 return ret; 156} 157 158/** 159 * cap_ptrace_traceme - Determine whether another process may trace the current 160 * @parent: The task proposed to be the tracer 161 * 162 * If parent is in the same or an ancestor user_ns and has all current's 163 * capabilities, then ptrace access is allowed. 164 * If parent has the ptrace capability to current's user_ns, then ptrace 165 * access is allowed. 166 * Else denied. 167 * 168 * Determine whether the nominated task is permitted to trace the current 169 * process, returning 0 if permission is granted, -ve if denied. 170 */ 171int cap_ptrace_traceme(struct task_struct *parent) 172{ 173 int ret = 0; 174 const struct cred *cred, *child_cred; 175 176 rcu_read_lock(); 177 cred = __task_cred(parent); 178 child_cred = current_cred(); 179 if (cred->user_ns == child_cred->user_ns && 180 cap_issubset(child_cred->cap_permitted, cred->cap_permitted)) 181 goto out; 182 if (has_ns_capability(parent, child_cred->user_ns, CAP_SYS_PTRACE)) 183 goto out; 184 ret = -EPERM; 185out: 186 rcu_read_unlock(); 187 return ret; 188} 189 190/** 191 * cap_capget - Retrieve a task's capability sets 192 * @target: The task from which to retrieve the capability sets 193 * @effective: The place to record the effective set 194 * @inheritable: The place to record the inheritable set 195 * @permitted: The place to record the permitted set 196 * 197 * This function retrieves the capabilities of the nominated task and returns 198 * them to the caller. 199 */ 200int cap_capget(const struct task_struct *target, kernel_cap_t *effective, 201 kernel_cap_t *inheritable, kernel_cap_t *permitted) 202{ 203 const struct cred *cred; 204 205 /* Derived from kernel/capability.c:sys_capget. */ 206 rcu_read_lock(); 207 cred = __task_cred(target); 208 *effective = cred->cap_effective; 209 *inheritable = cred->cap_inheritable; 210 *permitted = cred->cap_permitted; 211 rcu_read_unlock(); 212 return 0; 213} 214 215/* 216 * Determine whether the inheritable capabilities are limited to the old 217 * permitted set. Returns 1 if they are limited, 0 if they are not. 218 */ 219static inline int cap_inh_is_capped(void) 220{ 221 /* they are so limited unless the current task has the CAP_SETPCAP 222 * capability 223 */ 224 if (cap_capable(current_cred(), current_cred()->user_ns, 225 CAP_SETPCAP, CAP_OPT_NONE) == 0) 226 return 0; 227 return 1; 228} 229 230/** 231 * cap_capset - Validate and apply proposed changes to current's capabilities 232 * @new: The proposed new credentials; alterations should be made here 233 * @old: The current task's current credentials 234 * @effective: A pointer to the proposed new effective capabilities set 235 * @inheritable: A pointer to the proposed new inheritable capabilities set 236 * @permitted: A pointer to the proposed new permitted capabilities set 237 * 238 * This function validates and applies a proposed mass change to the current 239 * process's capability sets. The changes are made to the proposed new 240 * credentials, and assuming no error, will be committed by the caller of LSM. 241 */ 242int cap_capset(struct cred *new, 243 const struct cred *old, 244 const kernel_cap_t *effective, 245 const kernel_cap_t *inheritable, 246 const kernel_cap_t *permitted) 247{ 248 if (cap_inh_is_capped() && 249 !cap_issubset(*inheritable, 250 cap_combine(old->cap_inheritable, 251 old->cap_permitted))) 252 /* incapable of using this inheritable set */ 253 return -EPERM; 254 255 if (!cap_issubset(*inheritable, 256 cap_combine(old->cap_inheritable, 257 old->cap_bset))) 258 /* no new pI capabilities outside bounding set */ 259 return -EPERM; 260 261 /* verify restrictions on target's new Permitted set */ 262 if (!cap_issubset(*permitted, old->cap_permitted)) 263 return -EPERM; 264 265 /* verify the _new_Effective_ is a subset of the _new_Permitted_ */ 266 if (!cap_issubset(*effective, *permitted)) 267 return -EPERM; 268 269 new->cap_effective = *effective; 270 new->cap_inheritable = *inheritable; 271 new->cap_permitted = *permitted; 272 273 /* 274 * Mask off ambient bits that are no longer both permitted and 275 * inheritable. 276 */ 277 new->cap_ambient = cap_intersect(new->cap_ambient, 278 cap_intersect(*permitted, 279 *inheritable)); 280 if (WARN_ON(!cap_ambient_invariant_ok(new))) 281 return -EINVAL; 282 return 0; 283} 284 285/** 286 * cap_inode_need_killpriv - Determine if inode change affects privileges 287 * @dentry: The inode/dentry in being changed with change marked ATTR_KILL_PRIV 288 * 289 * Determine if an inode having a change applied that's marked ATTR_KILL_PRIV 290 * affects the security markings on that inode, and if it is, should 291 * inode_killpriv() be invoked or the change rejected. 292 * 293 * Return: 1 if security.capability has a value, meaning inode_killpriv() 294 * is required, 0 otherwise, meaning inode_killpriv() is not required. 295 */ 296int cap_inode_need_killpriv(struct dentry *dentry) 297{ 298 struct inode *inode = d_backing_inode(dentry); 299 int error; 300 301 error = __vfs_getxattr(dentry, inode, XATTR_NAME_CAPS, NULL, 0); 302 return error > 0; 303} 304 305/** 306 * cap_inode_killpriv - Erase the security markings on an inode 307 * 308 * @idmap: idmap of the mount the inode was found from 309 * @dentry: The inode/dentry to alter 310 * 311 * Erase the privilege-enhancing security markings on an inode. 312 * 313 * If the inode has been found through an idmapped mount the idmap of 314 * the vfsmount must be passed through @idmap. This function will then 315 * take care to map the inode according to @idmap before checking 316 * permissions. On non-idmapped mounts or if permission checking is to be 317 * performed on the raw inode simply pass @nop_mnt_idmap. 318 * 319 * Return: 0 if successful, -ve on error. 320 */ 321int cap_inode_killpriv(struct mnt_idmap *idmap, struct dentry *dentry) 322{ 323 int error; 324 325 error = __vfs_removexattr(idmap, dentry, XATTR_NAME_CAPS); 326 if (error == -EOPNOTSUPP) 327 error = 0; 328 return error; 329} 330 331static bool rootid_owns_currentns(vfsuid_t rootvfsuid) 332{ 333 struct user_namespace *ns; 334 kuid_t kroot; 335 336 if (!vfsuid_valid(rootvfsuid)) 337 return false; 338 339 kroot = vfsuid_into_kuid(rootvfsuid); 340 for (ns = current_user_ns();; ns = ns->parent) { 341 if (from_kuid(ns, kroot) == 0) 342 return true; 343 if (ns == &init_user_ns) 344 break; 345 } 346 347 return false; 348} 349 350static __u32 sansflags(__u32 m) 351{ 352 return m & ~VFS_CAP_FLAGS_EFFECTIVE; 353} 354 355static bool is_v2header(int size, const struct vfs_cap_data *cap) 356{ 357 if (size != XATTR_CAPS_SZ_2) 358 return false; 359 return sansflags(le32_to_cpu(cap->magic_etc)) == VFS_CAP_REVISION_2; 360} 361 362static bool is_v3header(int size, const struct vfs_cap_data *cap) 363{ 364 if (size != XATTR_CAPS_SZ_3) 365 return false; 366 return sansflags(le32_to_cpu(cap->magic_etc)) == VFS_CAP_REVISION_3; 367} 368 369/* 370 * getsecurity: We are called for security.* before any attempt to read the 371 * xattr from the inode itself. 372 * 373 * This gives us a chance to read the on-disk value and convert it. If we 374 * return -EOPNOTSUPP, then vfs_getxattr() will call the i_op handler. 375 * 376 * Note we are not called by vfs_getxattr_alloc(), but that is only called 377 * by the integrity subsystem, which really wants the unconverted values - 378 * so that's good. 379 */ 380int cap_inode_getsecurity(struct mnt_idmap *idmap, 381 struct inode *inode, const char *name, void **buffer, 382 bool alloc) 383{ 384 int size; 385 kuid_t kroot; 386 vfsuid_t vfsroot; 387 u32 nsmagic, magic; 388 uid_t root, mappedroot; 389 char *tmpbuf = NULL; 390 struct vfs_cap_data *cap; 391 struct vfs_ns_cap_data *nscap = NULL; 392 struct dentry *dentry; 393 struct user_namespace *fs_ns; 394 395 if (strcmp(name, "capability") != 0) 396 return -EOPNOTSUPP; 397 398 dentry = d_find_any_alias(inode); 399 if (!dentry) 400 return -EINVAL; 401 size = vfs_getxattr_alloc(idmap, dentry, XATTR_NAME_CAPS, &tmpbuf, 402 sizeof(struct vfs_ns_cap_data), GFP_NOFS); 403 dput(dentry); 404 /* gcc11 complains if we don't check for !tmpbuf */ 405 if (size < 0 || !tmpbuf) 406 goto out_free; 407 408 fs_ns = inode->i_sb->s_user_ns; 409 cap = (struct vfs_cap_data *) tmpbuf; 410 if (is_v2header(size, cap)) { 411 root = 0; 412 } else if (is_v3header(size, cap)) { 413 nscap = (struct vfs_ns_cap_data *) tmpbuf; 414 root = le32_to_cpu(nscap->rootid); 415 } else { 416 size = -EINVAL; 417 goto out_free; 418 } 419 420 kroot = make_kuid(fs_ns, root); 421 422 /* If this is an idmapped mount shift the kuid. */ 423 vfsroot = make_vfsuid(idmap, fs_ns, kroot); 424 425 /* If the root kuid maps to a valid uid in current ns, then return 426 * this as a nscap. */ 427 mappedroot = from_kuid(current_user_ns(), vfsuid_into_kuid(vfsroot)); 428 if (mappedroot != (uid_t)-1 && mappedroot != (uid_t)0) { 429 size = sizeof(struct vfs_ns_cap_data); 430 if (alloc) { 431 if (!nscap) { 432 /* v2 -> v3 conversion */ 433 nscap = kzalloc(size, GFP_ATOMIC); 434 if (!nscap) { 435 size = -ENOMEM; 436 goto out_free; 437 } 438 nsmagic = VFS_CAP_REVISION_3; 439 magic = le32_to_cpu(cap->magic_etc); 440 if (magic & VFS_CAP_FLAGS_EFFECTIVE) 441 nsmagic |= VFS_CAP_FLAGS_EFFECTIVE; 442 memcpy(&nscap->data, &cap->data, sizeof(__le32) * 2 * VFS_CAP_U32); 443 nscap->magic_etc = cpu_to_le32(nsmagic); 444 } else { 445 /* use allocated v3 buffer */ 446 tmpbuf = NULL; 447 } 448 nscap->rootid = cpu_to_le32(mappedroot); 449 *buffer = nscap; 450 } 451 goto out_free; 452 } 453 454 if (!rootid_owns_currentns(vfsroot)) { 455 size = -EOVERFLOW; 456 goto out_free; 457 } 458 459 /* This comes from a parent namespace. Return as a v2 capability */ 460 size = sizeof(struct vfs_cap_data); 461 if (alloc) { 462 if (nscap) { 463 /* v3 -> v2 conversion */ 464 cap = kzalloc(size, GFP_ATOMIC); 465 if (!cap) { 466 size = -ENOMEM; 467 goto out_free; 468 } 469 magic = VFS_CAP_REVISION_2; 470 nsmagic = le32_to_cpu(nscap->magic_etc); 471 if (nsmagic & VFS_CAP_FLAGS_EFFECTIVE) 472 magic |= VFS_CAP_FLAGS_EFFECTIVE; 473 memcpy(&cap->data, &nscap->data, sizeof(__le32) * 2 * VFS_CAP_U32); 474 cap->magic_etc = cpu_to_le32(magic); 475 } else { 476 /* use unconverted v2 */ 477 tmpbuf = NULL; 478 } 479 *buffer = cap; 480 } 481out_free: 482 kfree(tmpbuf); 483 return size; 484} 485 486/** 487 * rootid_from_xattr - translate root uid of vfs caps 488 * 489 * @value: vfs caps value which may be modified by this function 490 * @size: size of @ivalue 491 * @task_ns: user namespace of the caller 492 */ 493static vfsuid_t rootid_from_xattr(const void *value, size_t size, 494 struct user_namespace *task_ns) 495{ 496 const struct vfs_ns_cap_data *nscap = value; 497 uid_t rootid = 0; 498 499 if (size == XATTR_CAPS_SZ_3) 500 rootid = le32_to_cpu(nscap->rootid); 501 502 return VFSUIDT_INIT(make_kuid(task_ns, rootid)); 503} 504 505static bool validheader(size_t size, const struct vfs_cap_data *cap) 506{ 507 return is_v2header(size, cap) || is_v3header(size, cap); 508} 509 510/** 511 * cap_convert_nscap - check vfs caps 512 * 513 * @idmap: idmap of the mount the inode was found from 514 * @dentry: used to retrieve inode to check permissions on 515 * @ivalue: vfs caps value which may be modified by this function 516 * @size: size of @ivalue 517 * 518 * User requested a write of security.capability. If needed, update the 519 * xattr to change from v2 to v3, or to fixup the v3 rootid. 520 * 521 * If the inode has been found through an idmapped mount the idmap of 522 * the vfsmount must be passed through @idmap. This function will then 523 * take care to map the inode according to @idmap before checking 524 * permissions. On non-idmapped mounts or if permission checking is to be 525 * performed on the raw inode simply pass @nop_mnt_idmap. 526 * 527 * Return: On success, return the new size; on error, return < 0. 528 */ 529int cap_convert_nscap(struct mnt_idmap *idmap, struct dentry *dentry, 530 const void **ivalue, size_t size) 531{ 532 struct vfs_ns_cap_data *nscap; 533 uid_t nsrootid; 534 const struct vfs_cap_data *cap = *ivalue; 535 __u32 magic, nsmagic; 536 struct inode *inode = d_backing_inode(dentry); 537 struct user_namespace *task_ns = current_user_ns(), 538 *fs_ns = inode->i_sb->s_user_ns; 539 kuid_t rootid; 540 vfsuid_t vfsrootid; 541 size_t newsize; 542 543 if (!*ivalue) 544 return -EINVAL; 545 if (!validheader(size, cap)) 546 return -EINVAL; 547 if (!capable_wrt_inode_uidgid(idmap, inode, CAP_SETFCAP)) 548 return -EPERM; 549 if (size == XATTR_CAPS_SZ_2 && (idmap == &nop_mnt_idmap)) 550 if (ns_capable(inode->i_sb->s_user_ns, CAP_SETFCAP)) 551 /* user is privileged, just write the v2 */ 552 return size; 553 554 vfsrootid = rootid_from_xattr(*ivalue, size, task_ns); 555 if (!vfsuid_valid(vfsrootid)) 556 return -EINVAL; 557 558 rootid = from_vfsuid(idmap, fs_ns, vfsrootid); 559 if (!uid_valid(rootid)) 560 return -EINVAL; 561 562 nsrootid = from_kuid(fs_ns, rootid); 563 if (nsrootid == -1) 564 return -EINVAL; 565 566 newsize = sizeof(struct vfs_ns_cap_data); 567 nscap = kmalloc(newsize, GFP_ATOMIC); 568 if (!nscap) 569 return -ENOMEM; 570 nscap->rootid = cpu_to_le32(nsrootid); 571 nsmagic = VFS_CAP_REVISION_3; 572 magic = le32_to_cpu(cap->magic_etc); 573 if (magic & VFS_CAP_FLAGS_EFFECTIVE) 574 nsmagic |= VFS_CAP_FLAGS_EFFECTIVE; 575 nscap->magic_etc = cpu_to_le32(nsmagic); 576 memcpy(&nscap->data, &cap->data, sizeof(__le32) * 2 * VFS_CAP_U32); 577 578 *ivalue = nscap; 579 return newsize; 580} 581 582/* 583 * Calculate the new process capability sets from the capability sets attached 584 * to a file. 585 */ 586static inline int bprm_caps_from_vfs_caps(struct cpu_vfs_cap_data *caps, 587 struct linux_binprm *bprm, 588 bool *effective, 589 bool *has_fcap) 590{ 591 struct cred *new = bprm->cred; 592 int ret = 0; 593 594 if (caps->magic_etc & VFS_CAP_FLAGS_EFFECTIVE) 595 *effective = true; 596 597 if (caps->magic_etc & VFS_CAP_REVISION_MASK) 598 *has_fcap = true; 599 600 /* 601 * pP' = (X & fP) | (pI & fI) 602 * The addition of pA' is handled later. 603 */ 604 new->cap_permitted.val = 605 (new->cap_bset.val & caps->permitted.val) | 606 (new->cap_inheritable.val & caps->inheritable.val); 607 608 if (caps->permitted.val & ~new->cap_permitted.val) 609 /* insufficient to execute correctly */ 610 ret = -EPERM; 611 612 /* 613 * For legacy apps, with no internal support for recognizing they 614 * do not have enough capabilities, we return an error if they are 615 * missing some "forced" (aka file-permitted) capabilities. 616 */ 617 return *effective ? ret : 0; 618} 619 620/** 621 * get_vfs_caps_from_disk - retrieve vfs caps from disk 622 * 623 * @idmap: idmap of the mount the inode was found from 624 * @dentry: dentry from which @inode is retrieved 625 * @cpu_caps: vfs capabilities 626 * 627 * Extract the on-exec-apply capability sets for an executable file. 628 * 629 * If the inode has been found through an idmapped mount the idmap of 630 * the vfsmount must be passed through @idmap. This function will then 631 * take care to map the inode according to @idmap before checking 632 * permissions. On non-idmapped mounts or if permission checking is to be 633 * performed on the raw inode simply pass @nop_mnt_idmap. 634 */ 635int get_vfs_caps_from_disk(struct mnt_idmap *idmap, 636 const struct dentry *dentry, 637 struct cpu_vfs_cap_data *cpu_caps) 638{ 639 struct inode *inode = d_backing_inode(dentry); 640 __u32 magic_etc; 641 int size; 642 struct vfs_ns_cap_data data, *nscaps = &data; 643 struct vfs_cap_data *caps = (struct vfs_cap_data *) &data; 644 kuid_t rootkuid; 645 vfsuid_t rootvfsuid; 646 struct user_namespace *fs_ns; 647 648 memset(cpu_caps, 0, sizeof(struct cpu_vfs_cap_data)); 649 650 if (!inode) 651 return -ENODATA; 652 653 fs_ns = inode->i_sb->s_user_ns; 654 size = __vfs_getxattr((struct dentry *)dentry, inode, 655 XATTR_NAME_CAPS, &data, XATTR_CAPS_SZ); 656 if (size == -ENODATA || size == -EOPNOTSUPP) 657 /* no data, that's ok */ 658 return -ENODATA; 659 660 if (size < 0) 661 return size; 662 663 if (size < sizeof(magic_etc)) 664 return -EINVAL; 665 666 cpu_caps->magic_etc = magic_etc = le32_to_cpu(caps->magic_etc); 667 668 rootkuid = make_kuid(fs_ns, 0); 669 switch (magic_etc & VFS_CAP_REVISION_MASK) { 670 case VFS_CAP_REVISION_1: 671 if (size != XATTR_CAPS_SZ_1) 672 return -EINVAL; 673 break; 674 case VFS_CAP_REVISION_2: 675 if (size != XATTR_CAPS_SZ_2) 676 return -EINVAL; 677 break; 678 case VFS_CAP_REVISION_3: 679 if (size != XATTR_CAPS_SZ_3) 680 return -EINVAL; 681 rootkuid = make_kuid(fs_ns, le32_to_cpu(nscaps->rootid)); 682 break; 683 684 default: 685 return -EINVAL; 686 } 687 688 rootvfsuid = make_vfsuid(idmap, fs_ns, rootkuid); 689 if (!vfsuid_valid(rootvfsuid)) 690 return -ENODATA; 691 692 /* Limit the caps to the mounter of the filesystem 693 * or the more limited uid specified in the xattr. 694 */ 695 if (!rootid_owns_currentns(rootvfsuid)) 696 return -ENODATA; 697 698 cpu_caps->permitted.val = le32_to_cpu(caps->data[0].permitted); 699 cpu_caps->inheritable.val = le32_to_cpu(caps->data[0].inheritable); 700 701 /* 702 * Rev1 had just a single 32-bit word, later expanded 703 * to a second one for the high bits 704 */ 705 if ((magic_etc & VFS_CAP_REVISION_MASK) != VFS_CAP_REVISION_1) { 706 cpu_caps->permitted.val += (u64)le32_to_cpu(caps->data[1].permitted) << 32; 707 cpu_caps->inheritable.val += (u64)le32_to_cpu(caps->data[1].inheritable) << 32; 708 } 709 710 cpu_caps->permitted.val &= CAP_VALID_MASK; 711 cpu_caps->inheritable.val &= CAP_VALID_MASK; 712 713 cpu_caps->rootid = vfsuid_into_kuid(rootvfsuid); 714 715 return 0; 716} 717 718/* 719 * Attempt to get the on-exec apply capability sets for an executable file from 720 * its xattrs and, if present, apply them to the proposed credentials being 721 * constructed by execve(). 722 */ 723static int get_file_caps(struct linux_binprm *bprm, struct file *file, 724 bool *effective, bool *has_fcap) 725{ 726 int rc = 0; 727 struct cpu_vfs_cap_data vcaps; 728 729 cap_clear(bprm->cred->cap_permitted); 730 731 if (!file_caps_enabled) 732 return 0; 733 734 if (!mnt_may_suid(file->f_path.mnt)) 735 return 0; 736 737 /* 738 * This check is redundant with mnt_may_suid() but is kept to make 739 * explicit that capability bits are limited to s_user_ns and its 740 * descendants. 741 */ 742 if (!current_in_userns(file->f_path.mnt->mnt_sb->s_user_ns)) 743 return 0; 744 745 rc = get_vfs_caps_from_disk(file_mnt_idmap(file), 746 file->f_path.dentry, &vcaps); 747 if (rc < 0) { 748 if (rc == -EINVAL) 749 printk(KERN_NOTICE "Invalid argument reading file caps for %s\n", 750 bprm->filename); 751 else if (rc == -ENODATA) 752 rc = 0; 753 goto out; 754 } 755 756 rc = bprm_caps_from_vfs_caps(&vcaps, bprm, effective, has_fcap); 757 758out: 759 if (rc) 760 cap_clear(bprm->cred->cap_permitted); 761 762 return rc; 763} 764 765static inline bool root_privileged(void) { return !issecure(SECURE_NOROOT); } 766 767static inline bool __is_real(kuid_t uid, struct cred *cred) 768{ return uid_eq(cred->uid, uid); } 769 770static inline bool __is_eff(kuid_t uid, struct cred *cred) 771{ return uid_eq(cred->euid, uid); } 772 773static inline bool __is_suid(kuid_t uid, struct cred *cred) 774{ return !__is_real(uid, cred) && __is_eff(uid, cred); } 775 776/* 777 * handle_privileged_root - Handle case of privileged root 778 * @bprm: The execution parameters, including the proposed creds 779 * @has_fcap: Are any file capabilities set? 780 * @effective: Do we have effective root privilege? 781 * @root_uid: This namespace' root UID WRT initial USER namespace 782 * 783 * Handle the case where root is privileged and hasn't been neutered by 784 * SECURE_NOROOT. If file capabilities are set, they won't be combined with 785 * set UID root and nothing is changed. If we are root, cap_permitted is 786 * updated. If we have become set UID root, the effective bit is set. 787 */ 788static void handle_privileged_root(struct linux_binprm *bprm, bool has_fcap, 789 bool *effective, kuid_t root_uid) 790{ 791 const struct cred *old = current_cred(); 792 struct cred *new = bprm->cred; 793 794 if (!root_privileged()) 795 return; 796 /* 797 * If the legacy file capability is set, then don't set privs 798 * for a setuid root binary run by a non-root user. Do set it 799 * for a root user just to cause least surprise to an admin. 800 */ 801 if (has_fcap && __is_suid(root_uid, new)) { 802 warn_setuid_and_fcaps_mixed(bprm->filename); 803 return; 804 } 805 /* 806 * To support inheritance of root-permissions and suid-root 807 * executables under compatibility mode, we override the 808 * capability sets for the file. 809 */ 810 if (__is_eff(root_uid, new) || __is_real(root_uid, new)) { 811 /* pP' = (cap_bset & ~0) | (pI & ~0) */ 812 new->cap_permitted = cap_combine(old->cap_bset, 813 old->cap_inheritable); 814 } 815 /* 816 * If only the real uid is 0, we do not set the effective bit. 817 */ 818 if (__is_eff(root_uid, new)) 819 *effective = true; 820} 821 822#define __cap_gained(field, target, source) \ 823 !cap_issubset(target->cap_##field, source->cap_##field) 824#define __cap_grew(target, source, cred) \ 825 !cap_issubset(cred->cap_##target, cred->cap_##source) 826#define __cap_full(field, cred) \ 827 cap_issubset(CAP_FULL_SET, cred->cap_##field) 828 829static inline bool __is_setuid(struct cred *new, const struct cred *old) 830{ return !uid_eq(new->euid, old->uid); } 831 832static inline bool __is_setgid(struct cred *new, const struct cred *old) 833{ return !gid_eq(new->egid, old->gid); } 834 835/* 836 * 1) Audit candidate if current->cap_effective is set 837 * 838 * We do not bother to audit if 3 things are true: 839 * 1) cap_effective has all caps 840 * 2) we became root *OR* are were already root 841 * 3) root is supposed to have all caps (SECURE_NOROOT) 842 * Since this is just a normal root execing a process. 843 * 844 * Number 1 above might fail if you don't have a full bset, but I think 845 * that is interesting information to audit. 846 * 847 * A number of other conditions require logging: 848 * 2) something prevented setuid root getting all caps 849 * 3) non-setuid root gets fcaps 850 * 4) non-setuid root gets ambient 851 */ 852static inline bool nonroot_raised_pE(struct cred *new, const struct cred *old, 853 kuid_t root, bool has_fcap) 854{ 855 bool ret = false; 856 857 if ((__cap_grew(effective, ambient, new) && 858 !(__cap_full(effective, new) && 859 (__is_eff(root, new) || __is_real(root, new)) && 860 root_privileged())) || 861 (root_privileged() && 862 __is_suid(root, new) && 863 !__cap_full(effective, new)) || 864 (!__is_setuid(new, old) && 865 ((has_fcap && 866 __cap_gained(permitted, new, old)) || 867 __cap_gained(ambient, new, old)))) 868 869 ret = true; 870 871 return ret; 872} 873 874/** 875 * cap_bprm_creds_from_file - Set up the proposed credentials for execve(). 876 * @bprm: The execution parameters, including the proposed creds 877 * @file: The file to pull the credentials from 878 * 879 * Set up the proposed credentials for a new execution context being 880 * constructed by execve(). The proposed creds in @bprm->cred is altered, 881 * which won't take effect immediately. 882 * 883 * Return: 0 if successful, -ve on error. 884 */ 885int cap_bprm_creds_from_file(struct linux_binprm *bprm, struct file *file) 886{ 887 /* Process setpcap binaries and capabilities for uid 0 */ 888 const struct cred *old = current_cred(); 889 struct cred *new = bprm->cred; 890 bool effective = false, has_fcap = false, is_setid; 891 int ret; 892 kuid_t root_uid; 893 894 if (WARN_ON(!cap_ambient_invariant_ok(old))) 895 return -EPERM; 896 897 ret = get_file_caps(bprm, file, &effective, &has_fcap); 898 if (ret < 0) 899 return ret; 900 901 root_uid = make_kuid(new->user_ns, 0); 902 903 handle_privileged_root(bprm, has_fcap, &effective, root_uid); 904 905 /* if we have fs caps, clear dangerous personality flags */ 906 if (__cap_gained(permitted, new, old)) 907 bprm->per_clear |= PER_CLEAR_ON_SETID; 908 909 /* Don't let someone trace a set[ug]id/setpcap binary with the revised 910 * credentials unless they have the appropriate permit. 911 * 912 * In addition, if NO_NEW_PRIVS, then ensure we get no new privs. 913 */ 914 is_setid = __is_setuid(new, old) || __is_setgid(new, old); 915 916 if ((is_setid || __cap_gained(permitted, new, old)) && 917 ((bprm->unsafe & ~LSM_UNSAFE_PTRACE) || 918 !ptracer_capable(current, new->user_ns))) { 919 /* downgrade; they get no more than they had, and maybe less */ 920 if (!ns_capable(new->user_ns, CAP_SETUID) || 921 (bprm->unsafe & LSM_UNSAFE_NO_NEW_PRIVS)) { 922 new->euid = new->uid; 923 new->egid = new->gid; 924 } 925 new->cap_permitted = cap_intersect(new->cap_permitted, 926 old->cap_permitted); 927 } 928 929 new->suid = new->fsuid = new->euid; 930 new->sgid = new->fsgid = new->egid; 931 932 /* File caps or setid cancels ambient. */ 933 if (has_fcap || is_setid) 934 cap_clear(new->cap_ambient); 935 936 /* 937 * Now that we've computed pA', update pP' to give: 938 * pP' = (X & fP) | (pI & fI) | pA' 939 */ 940 new->cap_permitted = cap_combine(new->cap_permitted, new->cap_ambient); 941 942 /* 943 * Set pE' = (fE ? pP' : pA'). Because pA' is zero if fE is set, 944 * this is the same as pE' = (fE ? pP' : 0) | pA'. 945 */ 946 if (effective) 947 new->cap_effective = new->cap_permitted; 948 else 949 new->cap_effective = new->cap_ambient; 950 951 if (WARN_ON(!cap_ambient_invariant_ok(new))) 952 return -EPERM; 953 954 if (nonroot_raised_pE(new, old, root_uid, has_fcap)) { 955 ret = audit_log_bprm_fcaps(bprm, new, old); 956 if (ret < 0) 957 return ret; 958 } 959 960 new->securebits &= ~issecure_mask(SECURE_KEEP_CAPS); 961 962 if (WARN_ON(!cap_ambient_invariant_ok(new))) 963 return -EPERM; 964 965 /* Check for privilege-elevated exec. */ 966 if (is_setid || 967 (!__is_real(root_uid, new) && 968 (effective || 969 __cap_grew(permitted, ambient, new)))) 970 bprm->secureexec = 1; 971 972 return 0; 973} 974 975/** 976 * cap_inode_setxattr - Determine whether an xattr may be altered 977 * @dentry: The inode/dentry being altered 978 * @name: The name of the xattr to be changed 979 * @value: The value that the xattr will be changed to 980 * @size: The size of value 981 * @flags: The replacement flag 982 * 983 * Determine whether an xattr may be altered or set on an inode, returning 0 if 984 * permission is granted, -ve if denied. 985 * 986 * This is used to make sure security xattrs don't get updated or set by those 987 * who aren't privileged to do so. 988 */ 989int cap_inode_setxattr(struct dentry *dentry, const char *name, 990 const void *value, size_t size, int flags) 991{ 992 struct user_namespace *user_ns = dentry->d_sb->s_user_ns; 993 994 /* Ignore non-security xattrs */ 995 if (strncmp(name, XATTR_SECURITY_PREFIX, 996 XATTR_SECURITY_PREFIX_LEN) != 0) 997 return 0; 998 999 /* 1000 * For XATTR_NAME_CAPS the check will be done in 1001 * cap_convert_nscap(), called by setxattr() 1002 */ 1003 if (strcmp(name, XATTR_NAME_CAPS) == 0) 1004 return 0; 1005 1006 if (!ns_capable(user_ns, CAP_SYS_ADMIN)) 1007 return -EPERM; 1008 return 0; 1009} 1010 1011/** 1012 * cap_inode_removexattr - Determine whether an xattr may be removed 1013 * 1014 * @idmap: idmap of the mount the inode was found from 1015 * @dentry: The inode/dentry being altered 1016 * @name: The name of the xattr to be changed 1017 * 1018 * Determine whether an xattr may be removed from an inode, returning 0 if 1019 * permission is granted, -ve if denied. 1020 * 1021 * If the inode has been found through an idmapped mount the idmap of 1022 * the vfsmount must be passed through @idmap. This function will then 1023 * take care to map the inode according to @idmap before checking 1024 * permissions. On non-idmapped mounts or if permission checking is to be 1025 * performed on the raw inode simply pass @nop_mnt_idmap. 1026 * 1027 * This is used to make sure security xattrs don't get removed by those who 1028 * aren't privileged to remove them. 1029 */ 1030int cap_inode_removexattr(struct mnt_idmap *idmap, 1031 struct dentry *dentry, const char *name) 1032{ 1033 struct user_namespace *user_ns = dentry->d_sb->s_user_ns; 1034 1035 /* Ignore non-security xattrs */ 1036 if (strncmp(name, XATTR_SECURITY_PREFIX, 1037 XATTR_SECURITY_PREFIX_LEN) != 0) 1038 return 0; 1039 1040 if (strcmp(name, XATTR_NAME_CAPS) == 0) { 1041 /* security.capability gets namespaced */ 1042 struct inode *inode = d_backing_inode(dentry); 1043 if (!inode) 1044 return -EINVAL; 1045 if (!capable_wrt_inode_uidgid(idmap, inode, CAP_SETFCAP)) 1046 return -EPERM; 1047 return 0; 1048 } 1049 1050 if (!ns_capable(user_ns, CAP_SYS_ADMIN)) 1051 return -EPERM; 1052 return 0; 1053} 1054 1055/* 1056 * cap_emulate_setxuid() fixes the effective / permitted capabilities of 1057 * a process after a call to setuid, setreuid, or setresuid. 1058 * 1059 * 1) When set*uiding _from_ one of {r,e,s}uid == 0 _to_ all of 1060 * {r,e,s}uid != 0, the permitted and effective capabilities are 1061 * cleared. 1062 * 1063 * 2) When set*uiding _from_ euid == 0 _to_ euid != 0, the effective 1064 * capabilities of the process are cleared. 1065 * 1066 * 3) When set*uiding _from_ euid != 0 _to_ euid == 0, the effective 1067 * capabilities are set to the permitted capabilities. 1068 * 1069 * fsuid is handled elsewhere. fsuid == 0 and {r,e,s}uid!= 0 should 1070 * never happen. 1071 * 1072 * -astor 1073 * 1074 * cevans - New behaviour, Oct '99 1075 * A process may, via prctl(), elect to keep its capabilities when it 1076 * calls setuid() and switches away from uid==0. Both permitted and 1077 * effective sets will be retained. 1078 * Without this change, it was impossible for a daemon to drop only some 1079 * of its privilege. The call to setuid(!=0) would drop all privileges! 1080 * Keeping uid 0 is not an option because uid 0 owns too many vital 1081 * files.. 1082 * Thanks to Olaf Kirch and Peter Benie for spotting this. 1083 */ 1084static inline void cap_emulate_setxuid(struct cred *new, const struct cred *old) 1085{ 1086 kuid_t root_uid = make_kuid(old->user_ns, 0); 1087 1088 if ((uid_eq(old->uid, root_uid) || 1089 uid_eq(old->euid, root_uid) || 1090 uid_eq(old->suid, root_uid)) && 1091 (!uid_eq(new->uid, root_uid) && 1092 !uid_eq(new->euid, root_uid) && 1093 !uid_eq(new->suid, root_uid))) { 1094 if (!issecure(SECURE_KEEP_CAPS)) { 1095 cap_clear(new->cap_permitted); 1096 cap_clear(new->cap_effective); 1097 } 1098 1099 /* 1100 * Pre-ambient programs expect setresuid to nonroot followed 1101 * by exec to drop capabilities. We should make sure that 1102 * this remains the case. 1103 */ 1104 cap_clear(new->cap_ambient); 1105 } 1106 if (uid_eq(old->euid, root_uid) && !uid_eq(new->euid, root_uid)) 1107 cap_clear(new->cap_effective); 1108 if (!uid_eq(old->euid, root_uid) && uid_eq(new->euid, root_uid)) 1109 new->cap_effective = new->cap_permitted; 1110} 1111 1112/** 1113 * cap_task_fix_setuid - Fix up the results of setuid() call 1114 * @new: The proposed credentials 1115 * @old: The current task's current credentials 1116 * @flags: Indications of what has changed 1117 * 1118 * Fix up the results of setuid() call before the credential changes are 1119 * actually applied. 1120 * 1121 * Return: 0 to grant the changes, -ve to deny them. 1122 */ 1123int cap_task_fix_setuid(struct cred *new, const struct cred *old, int flags) 1124{ 1125 switch (flags) { 1126 case LSM_SETID_RE: 1127 case LSM_SETID_ID: 1128 case LSM_SETID_RES: 1129 /* juggle the capabilities to follow [RES]UID changes unless 1130 * otherwise suppressed */ 1131 if (!issecure(SECURE_NO_SETUID_FIXUP)) 1132 cap_emulate_setxuid(new, old); 1133 break; 1134 1135 case LSM_SETID_FS: 1136 /* juggle the capabilities to follow FSUID changes, unless 1137 * otherwise suppressed 1138 * 1139 * FIXME - is fsuser used for all CAP_FS_MASK capabilities? 1140 * if not, we might be a bit too harsh here. 1141 */ 1142 if (!issecure(SECURE_NO_SETUID_FIXUP)) { 1143 kuid_t root_uid = make_kuid(old->user_ns, 0); 1144 if (uid_eq(old->fsuid, root_uid) && !uid_eq(new->fsuid, root_uid)) 1145 new->cap_effective = 1146 cap_drop_fs_set(new->cap_effective); 1147 1148 if (!uid_eq(old->fsuid, root_uid) && uid_eq(new->fsuid, root_uid)) 1149 new->cap_effective = 1150 cap_raise_fs_set(new->cap_effective, 1151 new->cap_permitted); 1152 } 1153 break; 1154 1155 default: 1156 return -EINVAL; 1157 } 1158 1159 return 0; 1160} 1161 1162/* 1163 * Rationale: code calling task_setscheduler, task_setioprio, and 1164 * task_setnice, assumes that 1165 * . if capable(cap_sys_nice), then those actions should be allowed 1166 * . if not capable(cap_sys_nice), but acting on your own processes, 1167 * then those actions should be allowed 1168 * This is insufficient now since you can call code without suid, but 1169 * yet with increased caps. 1170 * So we check for increased caps on the target process. 1171 */ 1172static int cap_safe_nice(struct task_struct *p) 1173{ 1174 int is_subset, ret = 0; 1175 1176 rcu_read_lock(); 1177 is_subset = cap_issubset(__task_cred(p)->cap_permitted, 1178 current_cred()->cap_permitted); 1179 if (!is_subset && !ns_capable(__task_cred(p)->user_ns, CAP_SYS_NICE)) 1180 ret = -EPERM; 1181 rcu_read_unlock(); 1182 1183 return ret; 1184} 1185 1186/** 1187 * cap_task_setscheduler - Determine if scheduler policy change is permitted 1188 * @p: The task to affect 1189 * 1190 * Determine if the requested scheduler policy change is permitted for the 1191 * specified task. 1192 * 1193 * Return: 0 if permission is granted, -ve if denied. 1194 */ 1195int cap_task_setscheduler(struct task_struct *p) 1196{ 1197 return cap_safe_nice(p); 1198} 1199 1200/** 1201 * cap_task_setioprio - Determine if I/O priority change is permitted 1202 * @p: The task to affect 1203 * @ioprio: The I/O priority to set 1204 * 1205 * Determine if the requested I/O priority change is permitted for the specified 1206 * task. 1207 * 1208 * Return: 0 if permission is granted, -ve if denied. 1209 */ 1210int cap_task_setioprio(struct task_struct *p, int ioprio) 1211{ 1212 return cap_safe_nice(p); 1213} 1214 1215/** 1216 * cap_task_setnice - Determine if task priority change is permitted 1217 * @p: The task to affect 1218 * @nice: The nice value to set 1219 * 1220 * Determine if the requested task priority change is permitted for the 1221 * specified task. 1222 * 1223 * Return: 0 if permission is granted, -ve if denied. 1224 */ 1225int cap_task_setnice(struct task_struct *p, int nice) 1226{ 1227 return cap_safe_nice(p); 1228} 1229 1230/* 1231 * Implement PR_CAPBSET_DROP. Attempt to remove the specified capability from 1232 * the current task's bounding set. Returns 0 on success, -ve on error. 1233 */ 1234static int cap_prctl_drop(unsigned long cap) 1235{ 1236 struct cred *new; 1237 1238 if (!ns_capable(current_user_ns(), CAP_SETPCAP)) 1239 return -EPERM; 1240 if (!cap_valid(cap)) 1241 return -EINVAL; 1242 1243 new = prepare_creds(); 1244 if (!new) 1245 return -ENOMEM; 1246 cap_lower(new->cap_bset, cap); 1247 return commit_creds(new); 1248} 1249 1250/** 1251 * cap_task_prctl - Implement process control functions for this security module 1252 * @option: The process control function requested 1253 * @arg2: The argument data for this function 1254 * @arg3: The argument data for this function 1255 * @arg4: The argument data for this function 1256 * @arg5: The argument data for this function 1257 * 1258 * Allow process control functions (sys_prctl()) to alter capabilities; may 1259 * also deny access to other functions not otherwise implemented here. 1260 * 1261 * Return: 0 or +ve on success, -ENOSYS if this function is not implemented 1262 * here, other -ve on error. If -ENOSYS is returned, sys_prctl() and other LSM 1263 * modules will consider performing the function. 1264 */ 1265int cap_task_prctl(int option, unsigned long arg2, unsigned long arg3, 1266 unsigned long arg4, unsigned long arg5) 1267{ 1268 const struct cred *old = current_cred(); 1269 struct cred *new; 1270 1271 switch (option) { 1272 case PR_CAPBSET_READ: 1273 if (!cap_valid(arg2)) 1274 return -EINVAL; 1275 return !!cap_raised(old->cap_bset, arg2); 1276 1277 case PR_CAPBSET_DROP: 1278 return cap_prctl_drop(arg2); 1279 1280 /* 1281 * The next four prctl's remain to assist with transitioning a 1282 * system from legacy UID=0 based privilege (when filesystem 1283 * capabilities are not in use) to a system using filesystem 1284 * capabilities only - as the POSIX.1e draft intended. 1285 * 1286 * Note: 1287 * 1288 * PR_SET_SECUREBITS = 1289 * issecure_mask(SECURE_KEEP_CAPS_LOCKED) 1290 * | issecure_mask(SECURE_NOROOT) 1291 * | issecure_mask(SECURE_NOROOT_LOCKED) 1292 * | issecure_mask(SECURE_NO_SETUID_FIXUP) 1293 * | issecure_mask(SECURE_NO_SETUID_FIXUP_LOCKED) 1294 * 1295 * will ensure that the current process and all of its 1296 * children will be locked into a pure 1297 * capability-based-privilege environment. 1298 */ 1299 case PR_SET_SECUREBITS: 1300 if ((((old->securebits & SECURE_ALL_LOCKS) >> 1) 1301 & (old->securebits ^ arg2)) /*[1]*/ 1302 || ((old->securebits & SECURE_ALL_LOCKS & ~arg2)) /*[2]*/ 1303 || (arg2 & ~(SECURE_ALL_LOCKS | SECURE_ALL_BITS)) /*[3]*/ 1304 || (cap_capable(current_cred(), 1305 current_cred()->user_ns, 1306 CAP_SETPCAP, 1307 CAP_OPT_NONE) != 0) /*[4]*/ 1308 /* 1309 * [1] no changing of bits that are locked 1310 * [2] no unlocking of locks 1311 * [3] no setting of unsupported bits 1312 * [4] doing anything requires privilege (go read about 1313 * the "sendmail capabilities bug") 1314 */ 1315 ) 1316 /* cannot change a locked bit */ 1317 return -EPERM; 1318 1319 new = prepare_creds(); 1320 if (!new) 1321 return -ENOMEM; 1322 new->securebits = arg2; 1323 return commit_creds(new); 1324 1325 case PR_GET_SECUREBITS: 1326 return old->securebits; 1327 1328 case PR_GET_KEEPCAPS: 1329 return !!issecure(SECURE_KEEP_CAPS); 1330 1331 case PR_SET_KEEPCAPS: 1332 if (arg2 > 1) /* Note, we rely on arg2 being unsigned here */ 1333 return -EINVAL; 1334 if (issecure(SECURE_KEEP_CAPS_LOCKED)) 1335 return -EPERM; 1336 1337 new = prepare_creds(); 1338 if (!new) 1339 return -ENOMEM; 1340 if (arg2) 1341 new->securebits |= issecure_mask(SECURE_KEEP_CAPS); 1342 else 1343 new->securebits &= ~issecure_mask(SECURE_KEEP_CAPS); 1344 return commit_creds(new); 1345 1346 case PR_CAP_AMBIENT: 1347 if (arg2 == PR_CAP_AMBIENT_CLEAR_ALL) { 1348 if (arg3 | arg4 | arg5) 1349 return -EINVAL; 1350 1351 new = prepare_creds(); 1352 if (!new) 1353 return -ENOMEM; 1354 cap_clear(new->cap_ambient); 1355 return commit_creds(new); 1356 } 1357 1358 if (((!cap_valid(arg3)) | arg4 | arg5)) 1359 return -EINVAL; 1360 1361 if (arg2 == PR_CAP_AMBIENT_IS_SET) { 1362 return !!cap_raised(current_cred()->cap_ambient, arg3); 1363 } else if (arg2 != PR_CAP_AMBIENT_RAISE && 1364 arg2 != PR_CAP_AMBIENT_LOWER) { 1365 return -EINVAL; 1366 } else { 1367 if (arg2 == PR_CAP_AMBIENT_RAISE && 1368 (!cap_raised(current_cred()->cap_permitted, arg3) || 1369 !cap_raised(current_cred()->cap_inheritable, 1370 arg3) || 1371 issecure(SECURE_NO_CAP_AMBIENT_RAISE))) 1372 return -EPERM; 1373 1374 new = prepare_creds(); 1375 if (!new) 1376 return -ENOMEM; 1377 if (arg2 == PR_CAP_AMBIENT_RAISE) 1378 cap_raise(new->cap_ambient, arg3); 1379 else 1380 cap_lower(new->cap_ambient, arg3); 1381 return commit_creds(new); 1382 } 1383 1384 default: 1385 /* No functionality available - continue with default */ 1386 return -ENOSYS; 1387 } 1388} 1389 1390/** 1391 * cap_vm_enough_memory - Determine whether a new virtual mapping is permitted 1392 * @mm: The VM space in which the new mapping is to be made 1393 * @pages: The size of the mapping 1394 * 1395 * Determine whether the allocation of a new virtual mapping by the current 1396 * task is permitted. 1397 * 1398 * Return: 1 if permission is granted, 0 if not. 1399 */ 1400int cap_vm_enough_memory(struct mm_struct *mm, long pages) 1401{ 1402 int cap_sys_admin = 0; 1403 1404 if (cap_capable(current_cred(), &init_user_ns, 1405 CAP_SYS_ADMIN, CAP_OPT_NOAUDIT) == 0) 1406 cap_sys_admin = 1; 1407 1408 return cap_sys_admin; 1409} 1410 1411/** 1412 * cap_mmap_addr - check if able to map given addr 1413 * @addr: address attempting to be mapped 1414 * 1415 * If the process is attempting to map memory below dac_mmap_min_addr they need 1416 * CAP_SYS_RAWIO. The other parameters to this function are unused by the 1417 * capability security module. 1418 * 1419 * Return: 0 if this mapping should be allowed or -EPERM if not. 1420 */ 1421int cap_mmap_addr(unsigned long addr) 1422{ 1423 int ret = 0; 1424 1425 if (addr < dac_mmap_min_addr) { 1426 ret = cap_capable(current_cred(), &init_user_ns, CAP_SYS_RAWIO, 1427 CAP_OPT_NONE); 1428 /* set PF_SUPERPRIV if it turns out we allow the low mmap */ 1429 if (ret == 0) 1430 current->flags |= PF_SUPERPRIV; 1431 } 1432 return ret; 1433} 1434 1435int cap_mmap_file(struct file *file, unsigned long reqprot, 1436 unsigned long prot, unsigned long flags) 1437{ 1438 return 0; 1439} 1440 1441#ifdef CONFIG_SECURITY 1442 1443static struct security_hook_list capability_hooks[] __ro_after_init = { 1444 LSM_HOOK_INIT(capable, cap_capable), 1445 LSM_HOOK_INIT(settime, cap_settime), 1446 LSM_HOOK_INIT(ptrace_access_check, cap_ptrace_access_check), 1447 LSM_HOOK_INIT(ptrace_traceme, cap_ptrace_traceme), 1448 LSM_HOOK_INIT(capget, cap_capget), 1449 LSM_HOOK_INIT(capset, cap_capset), 1450 LSM_HOOK_INIT(bprm_creds_from_file, cap_bprm_creds_from_file), 1451 LSM_HOOK_INIT(inode_need_killpriv, cap_inode_need_killpriv), 1452 LSM_HOOK_INIT(inode_killpriv, cap_inode_killpriv), 1453 LSM_HOOK_INIT(inode_getsecurity, cap_inode_getsecurity), 1454 LSM_HOOK_INIT(mmap_addr, cap_mmap_addr), 1455 LSM_HOOK_INIT(mmap_file, cap_mmap_file), 1456 LSM_HOOK_INIT(task_fix_setuid, cap_task_fix_setuid), 1457 LSM_HOOK_INIT(task_prctl, cap_task_prctl), 1458 LSM_HOOK_INIT(task_setscheduler, cap_task_setscheduler), 1459 LSM_HOOK_INIT(task_setioprio, cap_task_setioprio), 1460 LSM_HOOK_INIT(task_setnice, cap_task_setnice), 1461 LSM_HOOK_INIT(vm_enough_memory, cap_vm_enough_memory), 1462}; 1463 1464static int __init capability_init(void) 1465{ 1466 security_add_hooks(capability_hooks, ARRAY_SIZE(capability_hooks), 1467 "capability"); 1468 return 0; 1469} 1470 1471DEFINE_LSM(capability) = { 1472 .name = "capability", 1473 .order = LSM_ORDER_FIRST, 1474 .init = capability_init, 1475}; 1476 1477#endif /* CONFIG_SECURITY */ 1478