]> git.proxmox.com Git - mirror_ubuntu-bionic-kernel.git/commit
ipc/mqueue.c: only perform resource calculation if user valid
authorKees Cook <keescook@chromium.org>
Tue, 16 Jul 2019 23:30:21 +0000 (16:30 -0700)
committerKleber Sacilotto de Souza <kleber.souza@canonical.com>
Wed, 14 Aug 2019 09:18:49 +0000 (11:18 +0200)
commit71379582d6a15f4c546409046520f259e3ceb88f
tree7869e84b9c1c54389cb8cff441e5e24a76990293
parentdac0c8ecb8f0bd63e3f7d9d62a7b88ed1af1d25e
ipc/mqueue.c: only perform resource calculation if user valid

BugLink: https://bugs.launchpad.net/bugs/1839376
[ Upstream commit a318f12ed8843cfac53198390c74a565c632f417 ]

Andreas Christoforou reported:

  UBSAN: Undefined behaviour in ipc/mqueue.c:414:49 signed integer overflow:
  9 * 2305843009213693951 cannot be represented in type 'long int'
  ...
  Call Trace:
    mqueue_evict_inode+0x8e7/0xa10 ipc/mqueue.c:414
    evict+0x472/0x8c0 fs/inode.c:558
    iput_final fs/inode.c:1547 [inline]
    iput+0x51d/0x8c0 fs/inode.c:1573
    mqueue_get_inode+0x8eb/0x1070 ipc/mqueue.c:320
    mqueue_create_attr+0x198/0x440 ipc/mqueue.c:459
    vfs_mkobj+0x39e/0x580 fs/namei.c:2892
    prepare_open ipc/mqueue.c:731 [inline]
    do_mq_open+0x6da/0x8e0 ipc/mqueue.c:771

Which could be triggered by:

        struct mq_attr attr = {
                .mq_flags = 0,
                .mq_maxmsg = 9,
                .mq_msgsize = 0x1fffffffffffffff,
                .mq_curmsgs = 0,
        };

        if (mq_open("/testing", 0x40, 3, &attr) == (mqd_t) -1)
                perror("mq_open");

mqueue_get_inode() was correctly rejecting the giant mq_msgsize, and
preparing to return -EINVAL.  During the cleanup, it calls
mqueue_evict_inode() which performed resource usage tracking math for
updating "user", before checking if there was a valid "user" at all
(which would indicate that the calculations would be sane).  Instead,
delay this check to after seeing a valid "user".

The overflow was real, but the results went unused, so while the flaw is
harmless, it's noisy for kernel fuzzers, so just fix it by moving the
calculation under the non-NULL "user" where it actually gets used.

Link: http://lkml.kernel.org/r/201906072207.ECB65450@keescook
Signed-off-by: Kees Cook <keescook@chromium.org>
Reported-by: Andreas Christoforou <andreaschristofo@gmail.com>
Acked-by: "Eric W. Biederman" <ebiederm@xmission.com>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: Arnd Bergmann <arnd@arndb.de>
Cc: Davidlohr Bueso <dave@stgolabs.net>
Cc: Manfred Spraul <manfred@colorfullife.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
Signed-off-by: Kamal Mostafa <kamal@canonical.com>
Signed-off-by: Khalid Elmously <khalid.elmously@canonical.com>
ipc/mqueue.c