Comment 0 for bug 1274349

Reported by pageexec

asmlinkage long compat_sys_recvmmsg(int fd, struct compat_mmsghdr __user *mmsg,
                                    unsigned int vlen, unsigned int flags,
                                    struct compat_timespec __user *timeout)
{
        int datagrams;
        struct timespec ktspec;

        if (flags & MSG_CMSG_COMPAT)
                return -EINVAL;

        if (COMPAT_USE_64BIT_TIME)
                return __sys_recvmmsg(fd, (struct mmsghdr __user *)mmsg, vlen,
                                      flags | MSG_CMSG_COMPAT,
                                      (struct timespec *) timeout);
/*...*/

The timeout pointer parameter is provided by userland (hence the
__user annotation) but for x32 syscalls it's simply cast to a kernel
pointer and is passed to __sys_recvmmsg which will eventually directly
dereference it for both reading and writing. Other callers to
__sys_recvmmsg properly copy from userland to the kernel first.

The impact is a sort of arbitrary kernel write-where-what primitive by
unprivileged users where the to-be-written area must contain valid
timespec data initially (the first 64 bit long field must be positive
and the second one must be < 1G).

The bug was introduced by commit
http://git.kernel.org/linus/ee4fa23c4b (other uses of
COMPAT_USE_64BIT_TIME seem fine) and should affect all kernels since
3.4 (and perhaps vendor kernels if they backported x32 support along
with this code). Note that CONFIG_X86_X32_ABI gets enabled at build
time and only if CONFIG_X86_X32 is enabled and ld can build x32
executables.