linux-stable.git/security, branch v3.16.71 Linux kernel stable tree selinux: avoid silent denials in permissive mode under RCU walk 2019-07-09T21:03:45+00:00 Stephen Smalley sds@tycho.nsa.gov 2018-12-12T15:10:55+00:00 cf1e870707687a674608836cb9947d033cd255f7 commit 3a28cff3bd4bf43f02be0c4e7933aebf3dc8197e upstream. commit 0dc1ba24f7fff6 ("SELINUX: Make selinux cache VFS RCU walks safe") results in no audit messages at all if in permissive mode because the cache is updated during the rcu walk and thus no denial occurs on the subsequent ref walk. Fix this by not updating the cache when performing a non-blocking permission check. This only affects search and symlink read checks during rcu walk. Fixes: 0dc1ba24f7fff6 ("SELINUX: Make selinux cache VFS RCU walks safe") Reported-by: BMK <bmktuwien@gmail.com> Signed-off-by: Stephen Smalley <sds@tycho.nsa.gov> Signed-off-by: Paul Moore <paul@paul-moore.com> [bwh: Backported to 3.16: - Add flags parameter to avc_update_node(), done upstream in commit fa1aa143ac4a "selinux: extended permissions for ioctls" - Adjust context] Signed-off-by: Ben Hutchings <ben@decadent.org.uk> 
commit 3a28cff3bd4bf43f02be0c4e7933aebf3dc8197e upstream.

commit 0dc1ba24f7fff6 ("SELINUX: Make selinux cache VFS RCU walks safe")
results in no audit messages at all if in permissive mode because the
cache is updated during the rcu walk and thus no denial occurs on
the subsequent ref walk.  Fix this by not updating the cache when
performing a non-blocking permission check.  This only affects search
and symlink read checks during rcu walk.

Fixes: 0dc1ba24f7fff6 ("SELINUX: Make selinux cache VFS RCU walks safe")
Reported-by: BMK <bmktuwien@gmail.com>
Signed-off-by: Stephen Smalley <sds@tycho.nsa.gov>
Signed-off-by: Paul Moore <paul@paul-moore.com>
[bwh: Backported to 3.16:
 - Add flags parameter to avc_update_node(), done upstream in commit
   fa1aa143ac4a "selinux: extended permissions for ioctls"
 - Adjust context]
Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
security/selinux: pass 'flags' arg to avc_audit() and avc_has_perm_flags() 2019-07-09T21:03:45+00:00 NeilBrown neilb@suse.de 2015-03-23T02:37:39+00:00 872dfecfaf15655f4c77b51d0077792df7c26bcb commit 7b20ea2579238f5e0da4bc93276c1b63c960c9ef upstream. This allows MAY_NOT_BLOCK to be passed, in RCU-walk mode, through the new avc_has_perm_flags() to avc_audit() and thence the slow_avc_audit. Signed-off-by: NeilBrown <neilb@suse.de> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk> [bwh: Backported to 3.16 as dependency of commit 3a28cff3bd4b "selinux: avoid silent denials in permissive mode under RCU walk"] Signed-off-by: Ben Hutchings <ben@decadent.org.uk> 
commit 7b20ea2579238f5e0da4bc93276c1b63c960c9ef upstream.

This allows MAY_NOT_BLOCK to be passed, in RCU-walk mode, through
the new avc_has_perm_flags() to avc_audit() and thence the slow_avc_audit.

Signed-off-by: NeilBrown <neilb@suse.de>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
[bwh: Backported to 3.16 as dependency of commit 3a28cff3bd4b
 "selinux: avoid silent denials in permissive mode under RCU walk"]
Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
apparmor: provide userspace flag indicating binfmt_elf_mmap change 2019-05-02T20:42:07+00:00 John Johansen john.johansen@canonical.com 2017-01-16T08:42:43+00:00 a62ea6707949c2a49f8a4f62c5b5d4c1703451d3 commit 34c426acb75cc21bdf84685e106db0c1a3565057 upstream. Commit 9f834ec18def ("binfmt_elf: switch to new creds when switching to new mm") changed when the creds are installed by the binfmt_elf handler. This affects which creds are used to mmap the executable into the address space. Which can have an affect on apparmor policy. Add a flag to apparmor at /sys/kernel/security/apparmor/features/domain/fix_binfmt_elf_mmap to make it possible to detect this semantic change so that the userspace tools and the regression test suite can correctly deal with the change. BugLink: http://bugs.launchpad.net/bugs/1630069 Signed-off-by: John Johansen <john.johansen@canonical.com> Signed-off-by: Ben Hutchings <ben@decadent.org.uk> 
commit 34c426acb75cc21bdf84685e106db0c1a3565057 upstream.

Commit 9f834ec18def ("binfmt_elf: switch to new creds when switching to new mm")
changed when the creds are installed by the binfmt_elf handler. This
affects which creds are used to mmap the executable into the address
space. Which can have an affect on apparmor policy.

Add a flag to apparmor at
/sys/kernel/security/apparmor/features/domain/fix_binfmt_elf_mmap

to make it possible to detect this semantic change so that the userspace
tools and the regression test suite can correctly deal with the change.

BugLink: http://bugs.launchpad.net/bugs/1630069
Signed-off-by: John Johansen <john.johansen@canonical.com>
Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
KEYS: always initialize keyring_index_key::desc_len 2019-05-02T20:41:58+00:00 Eric Biggers ebiggers@google.com 2019-02-22T15:36:18+00:00 696bfa8c3d86e000b4a01fd391020e71d4fc1a9b commit ede0fa98a900e657d1fcd80b50920efc896c1a4c upstream. syzbot hit the 'BUG_ON(index_key->desc_len == 0);' in __key_link_begin() called from construct_alloc_key() during sys_request_key(), because the length of the key description was never calculated. The problem is that we rely on ->desc_len being initialized by search_process_keyrings(), specifically by search_nested_keyrings(). But, if the process isn't subscribed to any keyrings that never happens. Fix it by always initializing keyring_index_key::desc_len as soon as the description is set, like we already do in some places. The following program reproduces the BUG_ON() when it's run as root and no session keyring has been installed. If it doesn't work, try removing pam_keyinit.so from /etc/pam.d/login and rebooting. #include <stdlib.h> #include <unistd.h> #include <keyutils.h> int main(void) { int id = add_key("keyring", "syz", NULL, 0, KEY_SPEC_USER_KEYRING); keyctl_setperm(id, KEY_OTH_WRITE); setreuid(5000, 5000); request_key("user", "desc", "", id); } Reported-by: syzbot+ec24e95ea483de0a24da@syzkaller.appspotmail.com Fixes: b2a4df200d57 ("KEYS: Expand the capacity of a keyring") Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: David Howells <dhowells@redhat.com> Signed-off-by: James Morris <james.morris@microsoft.com> Signed-off-by: Ben Hutchings <ben@decadent.org.uk> 
commit ede0fa98a900e657d1fcd80b50920efc896c1a4c upstream.

syzbot hit the 'BUG_ON(index_key->desc_len == 0);' in __key_link_begin()
called from construct_alloc_key() during sys_request_key(), because the
length of the key description was never calculated.

The problem is that we rely on ->desc_len being initialized by
search_process_keyrings(), specifically by search_nested_keyrings().
But, if the process isn't subscribed to any keyrings that never happens.

Fix it by always initializing keyring_index_key::desc_len as soon as the
description is set, like we already do in some places.

The following program reproduces the BUG_ON() when it's run as root and
no session keyring has been installed.  If it doesn't work, try removing
pam_keyinit.so from /etc/pam.d/login and rebooting.

    #include <stdlib.h>
    #include <unistd.h>
    #include <keyutils.h>

    int main(void)
    {
            int id = add_key("keyring", "syz", NULL, 0, KEY_SPEC_USER_KEYRING);

            keyctl_setperm(id, KEY_OTH_WRITE);
            setreuid(5000, 5000);
            request_key("user", "desc", "", id);
    }

Reported-by: syzbot+ec24e95ea483de0a24da@syzkaller.appspotmail.com
Fixes: b2a4df200d57 ("KEYS: Expand the capacity of a keyring")
Signed-off-by: Eric Biggers <ebiggers@google.com>
Signed-off-by: David Howells <dhowells@redhat.com>
Signed-off-by: James Morris <james.morris@microsoft.com>
Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
KEYS: restrict /proc/keys by credentials at open time 2019-05-02T20:41:58+00:00 Eric Biggers ebiggers@google.com 2017-09-18T18:38:29+00:00 1a844ebde7e7de0970d028faca990507e4287104 commit 4aa68e07d845562561f5e73c04aa521376e95252 upstream. When checking for permission to view keys whilst reading from /proc/keys, we should use the credentials with which the /proc/keys file was opened. This is because, in a classic type of exploit, it can be possible to bypass checks for the *current* credentials by passing the file descriptor to a suid program. Following commit 34dbbcdbf633 ("Make file credentials available to the seqfile interfaces") we can finally fix it. So let's do it. Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: David Howells <dhowells@redhat.com> [bwh: Backported to 3.16: adjust context] Signed-off-by: Ben Hutchings <ben@decadent.org.uk> 
commit 4aa68e07d845562561f5e73c04aa521376e95252 upstream.

When checking for permission to view keys whilst reading from
/proc/keys, we should use the credentials with which the /proc/keys file
was opened.  This is because, in a classic type of exploit, it can be
possible to bypass checks for the *current* credentials by passing the
file descriptor to a suid program.

Following commit 34dbbcdbf633 ("Make file credentials available to the
seqfile interfaces") we can finally fix it.  So let's do it.

Signed-off-by: Eric Biggers <ebiggers@google.com>
Signed-off-by: David Howells <dhowells@redhat.com>
[bwh: Backported to 3.16: adjust context]
Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
KEYS: allow reaching the keys quotas exactly 2019-05-02T20:41:56+00:00 Eric Biggers ebiggers@google.com 2019-02-14T16:20:01+00:00 980e41e1b9db1a32b382efdd08c4b7c70b5dbd24 commit a08bf91ce28ed3ae7b6fef35d843fef8dc8c2cd9 upstream. If the sysctl 'kernel.keys.maxkeys' is set to some number n, then actually users can only add up to 'n - 1' keys. Likewise for 'kernel.keys.maxbytes' and the root_* versions of these sysctls. But these sysctls are apparently supposed to be *maximums*, as per their names and all documentation I could find -- the keyrings(7) man page, Documentation/security/keys/core.rst, and all the mentions of EDQUOT meaning that the key quota was *exceeded* (as opposed to reached). Thus, fix the code to allow reaching the quotas exactly. Fixes: 0b77f5bfb45c ("keys: make the keyring quotas controllable through /proc/sys") Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: David Howells <dhowells@redhat.com> Signed-off-by: James Morris <james.morris@microsoft.com> Signed-off-by: Ben Hutchings <ben@decadent.org.uk> 
commit a08bf91ce28ed3ae7b6fef35d843fef8dc8c2cd9 upstream.

If the sysctl 'kernel.keys.maxkeys' is set to some number n, then
actually users can only add up to 'n - 1' keys.  Likewise for
'kernel.keys.maxbytes' and the root_* versions of these sysctls.  But
these sysctls are apparently supposed to be *maximums*, as per their
names and all documentation I could find -- the keyrings(7) man page,
Documentation/security/keys/core.rst, and all the mentions of EDQUOT
meaning that the key quota was *exceeded* (as opposed to reached).

Thus, fix the code to allow reaching the quotas exactly.

Fixes: 0b77f5bfb45c ("keys: make the keyring quotas controllable through /proc/sys")
Signed-off-by: Eric Biggers <ebiggers@google.com>
Signed-off-by: David Howells <dhowells@redhat.com>
Signed-off-by: James Morris <james.morris@microsoft.com>
Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
Yama: Check for pid death before checking ancestry 2019-05-02T20:41:23+00:00 Kees Cook keescook@chromium.org 2019-01-16T18:31:09+00:00 0c8a56a6fa056db65e66873a1b597c51559049b2 commit 9474f4e7cd71a633fa1ef93b7daefd44bbdfd482 upstream. It's possible that a pid has died before we take the rcu lock, in which case we can't walk the ancestry list as it may be detached. Instead, check for death first before doing the walk. Reported-by: syzbot+a9ac39bf55329e206219@syzkaller.appspotmail.com Fixes: 2d514487faf1 ("security: Yama LSM") Suggested-by: Oleg Nesterov <oleg@redhat.com> Signed-off-by: Kees Cook <keescook@chromium.org> Signed-off-by: James Morris <james.morris@microsoft.com> Signed-off-by: Ben Hutchings <ben@decadent.org.uk> 
commit 9474f4e7cd71a633fa1ef93b7daefd44bbdfd482 upstream.

It's possible that a pid has died before we take the rcu lock, in which
case we can't walk the ancestry list as it may be detached. Instead, check
for death first before doing the walk.

Reported-by: syzbot+a9ac39bf55329e206219@syzkaller.appspotmail.com
Fixes: 2d514487faf1 ("security: Yama LSM")
Suggested-by: Oleg Nesterov <oleg@redhat.com>
Signed-off-by: Kees Cook <keescook@chromium.org>
Signed-off-by: James Morris <james.morris@microsoft.com>
Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
ima: fix showing large 'violations' or 'runtime_measurements_count' 2019-02-11T17:53:30+00:00 Eric Biggers ebiggers@google.com 2018-09-07T21:33:24+00:00 5d9cd4ee9bec4c27e80203655a16545f825d7127 commit 1e4c8dafbb6bf72fb5eca035b861e39c5896c2b7 upstream. The 12 character temporary buffer is not necessarily long enough to hold a 'long' value. Increase it. Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: Mimi Zohar <zohar@linux.ibm.com> Signed-off-by: Ben Hutchings <ben@decadent.org.uk> 
commit 1e4c8dafbb6bf72fb5eca035b861e39c5896c2b7 upstream.

The 12 character temporary buffer is not necessarily long enough to hold
a 'long' value.  Increase it.

Signed-off-by: Eric Biggers <ebiggers@google.com>
Signed-off-by: Mimi Zohar <zohar@linux.ibm.com>
Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
KEYS: encrypted: fix buffer overread in valid_master_desc() 2018-12-16T22:09:44+00:00 Eric Biggers ebiggers@google.com 2017-06-08T13:48:18+00:00 bfe535bf4ab73e41922c7a58d6a858a2c435ff29 commit 794b4bc292f5d31739d89c0202c54e7dc9bc3add upstream. With the 'encrypted' key type it was possible for userspace to provide a data blob ending with a master key description shorter than expected, e.g. 'keyctl add encrypted desc "new x" @s'. When validating such a master key description, validate_master_desc() could read beyond the end of the buffer. Fix this by using strncmp() instead of memcmp(). [Also clean up the code to deduplicate some logic.] Cc: Mimi Zohar <zohar@linux.vnet.ibm.com> Signed-off-by: Eric Biggers <ebiggers@google.com> Signed-off-by: David Howells <dhowells@redhat.com> Signed-off-by: James Morris <james.l.morris@oracle.com> Signed-off-by: Ben Hutchings <ben@decadent.org.uk> 
commit 794b4bc292f5d31739d89c0202c54e7dc9bc3add upstream.

With the 'encrypted' key type it was possible for userspace to provide a
data blob ending with a master key description shorter than expected,
e.g. 'keyctl add encrypted desc "new x" @s'.  When validating such a
master key description, validate_master_desc() could read beyond the end
of the buffer.  Fix this by using strncmp() instead of memcmp().  [Also
clean up the code to deduplicate some logic.]

Cc: Mimi Zohar <zohar@linux.vnet.ibm.com>
Signed-off-by: Eric Biggers <ebiggers@google.com>
Signed-off-by: David Howells <dhowells@redhat.com>
Signed-off-by: James Morris <james.l.morris@oracle.com>
Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
apparmor: remove no-op permission check in policy_unpack 2018-12-16T22:08:51+00:00 John Johansen john.johansen@canonical.com 2018-08-22T00:19:53+00:00 d05549029edc975305f396c48a832f768a93f7f8 commit c037bd615885f1d9d3bdb48531bace79fae1505d upstream. The patch 736ec752d95e: "AppArmor: policy routines for loading and unpacking policy" from Jul 29, 2010, leads to the following static checker warning: security/apparmor/policy_unpack.c:410 verify_accept() warn: bitwise AND condition is false here security/apparmor/policy_unpack.c:413 verify_accept() warn: bitwise AND condition is false here security/apparmor/policy_unpack.c 392 #define DFA_VALID_PERM_MASK 0xffffffff 393 #define DFA_VALID_PERM2_MASK 0xffffffff 394 395 /** 396 * verify_accept - verify the accept tables of a dfa 397 * @dfa: dfa to verify accept tables of (NOT NULL) 398 * @flags: flags governing dfa 399 * 400 * Returns: 1 if valid accept tables else 0 if error 401 */ 402 static bool verify_accept(struct aa_dfa *dfa, int flags) 403 { 404 int i; 405 406 /* verify accept permissions */ 407 for (i = 0; i < dfa->tables[YYTD_ID_ACCEPT]->td_lolen; i++) { 408 int mode = ACCEPT_TABLE(dfa)[i]; 409 410 if (mode & ~DFA_VALID_PERM_MASK) 411 return 0; 412 413 if (ACCEPT_TABLE2(dfa)[i] & ~DFA_VALID_PERM2_MASK) 414 return 0; fixes: 736ec752d95e ("AppArmor: policy routines for loading and unpacking policy") Reported-by: Dan Carpenter <dan.carpenter@oracle.com> Signed-off-by: John Johansen <john.johansen@canonical.com> Signed-off-by: Ben Hutchings <ben@decadent.org.uk> 
commit c037bd615885f1d9d3bdb48531bace79fae1505d upstream.

The patch 736ec752d95e: "AppArmor: policy routines for loading and
unpacking policy" from Jul 29, 2010, leads to the following static
checker warning:

    security/apparmor/policy_unpack.c:410 verify_accept()
    warn: bitwise AND condition is false here

    security/apparmor/policy_unpack.c:413 verify_accept()
    warn: bitwise AND condition is false here

security/apparmor/policy_unpack.c
   392  #define DFA_VALID_PERM_MASK             0xffffffff
   393  #define DFA_VALID_PERM2_MASK            0xffffffff
   394
   395  /**
   396   * verify_accept - verify the accept tables of a dfa
   397   * @dfa: dfa to verify accept tables of (NOT NULL)
   398   * @flags: flags governing dfa
   399   *
   400   * Returns: 1 if valid accept tables else 0 if error
   401   */
   402  static bool verify_accept(struct aa_dfa *dfa, int flags)
   403  {
   404          int i;
   405
   406          /* verify accept permissions */
   407          for (i = 0; i < dfa->tables[YYTD_ID_ACCEPT]->td_lolen; i++) {
   408                  int mode = ACCEPT_TABLE(dfa)[i];
   409
   410                  if (mode & ~DFA_VALID_PERM_MASK)
   411                          return 0;
   412
   413                  if (ACCEPT_TABLE2(dfa)[i] & ~DFA_VALID_PERM2_MASK)
   414                          return 0;

fixes: 736ec752d95e ("AppArmor: policy routines for loading and unpacking policy")
Reported-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: John Johansen <john.johansen@canonical.com>
Signed-off-by: Ben Hutchings <ben@decadent.org.uk>