> -----Original Message-----
> From: stable-ow...@vger.kernel.org [mailto:stable-ow...@vger.kernel.org] On
> 4.4-stable review patch. If anyone has any objections, please let me know.
>
> ------------------
>
> From: Davidlohr Bueso <d...@stgolabs.net>
>
> commit a73ab244f0dad8fffb3291b905f73e2d3eaa7c00 upstream.
>
> Patch series "ipc/shm: shmat() fixes around nil-page".
Sorry for being a bit late (the pace is really fast here).
I have found a regression from 4.4.133-rc1 to 4.4.134-rc1 using Fuego LTP
wrapper.
4.4.134-rc1
tst_test.c:982: INFO: Timeout per run is 0h 05m 00s
cve-2017-5669.c:62: INFO: Attempting to attach shared memory to null
page
cve-2017-5669.c:74: INFO: Mapped shared memory to (nil)
cve-2017-5669.c:78: FAIL: We have mapped a VM address within the first
64Kb
cve-2017-5669.c:84: INFO: Touching shared memory to see if anything
strange happens
4.4.133-rc1:
tst_test.c:982: INFO: Timeout per run is 0h 05m 00s
cve-2017-5669.c:62: INFO: Attempting to attach shared memory to null
page
cve-2017-5669.c:67: PASS: shmat returned EINVAL
The culprits should be one or both of the two last commits to ipc/shm (one of
them a revert).
- ipc/shm: fix shmat() nil address after round-down when remapping
- Revert "ipc/shm: Fix shmat mmap nil-page protection"
I need to investigate the concrete reason, but for now I just wanted to report
it.
Thanks,
Daniel
>
> These patches fix two issues reported[1] a while back by Joe and Andrea
> around how shmat(2) behaves with nil-page.
>
> The first reverts a commit that it was incorrectly thought that mapping
> nil-page (address=0) was a no no with MAP_FIXED. This is not the case,
> with the exception of SHM_REMAP; which is address in the second patch.
>
> I chose two patches because it is easier to backport and it explicitly
> reverts bogus behaviour. Both patches ought to be in -stable and ltp
> testcases need updated (the added testcase around the cve can be
> modified to just test for SHM_RND|SHM_REMAP).
>
> [1] lkml.kernel.org/r/20180430172152.nfa564pvgpk3ut7p@linux-n805
>
> This patch (of 2):
>
> Commit 95e91b831f87 ("ipc/shm: Fix shmat mmap nil-page protection")
> worked on the idea that we should not be mapping as root addr=0 and
> MAP_FIXED. However, it was reported that this scenario is in fact
> valid, thus making the patch both bogus and breaks userspace as well.
>
> For example X11's libint10.so relies on shmat(1, SHM_RND) for lowmem
> initialization[1].
>
> [1]
> https://cgit.freedesktop.org/xorg/xserver/tree/hw/xfree86/os-support/linux/int1
> 0/linux.c#n347
> Link: http://lkml.kernel.org/r/20180503203243.15045-2-d...@stgolabs.net
> Fixes: 95e91b831f87 ("ipc/shm: Fix shmat mmap nil-page protection")
> Signed-off-by: Davidlohr Bueso <dbu...@suse.de>
> Reported-by: Joe Lawrence <joe.lawre...@redhat.com>
> Reported-by: Andrea Arcangeli <aarca...@redhat.com>
> Cc: Manfred Spraul <manf...@colorfullife.com>
> Cc: <sta...@vger.kernel.org>
> Signed-off-by: Andrew Morton <a...@linux-foundation.org>
> Signed-off-by: Linus Torvalds <torva...@linux-foundation.org>
> Signed-off-by: Greg Kroah-Hartman <gre...@linuxfoundation.org>
>
> ---
> ipc/shm.c | 9 ++-------
> 1 file changed, 2 insertions(+), 7 deletions(-)
>
> --- a/ipc/shm.c
> +++ b/ipc/shm.c
> @@ -1113,13 +1113,8 @@ long do_shmat(int shmid, char __user *sh
> goto out;
> else if ((addr = (ulong)shmaddr)) {
> if (addr & (shmlba - 1)) {
> - /*
> - * Round down to the nearest multiple of shmlba.
> - * For sane do_mmap_pgoff() parameters, avoid
> - * round downs that trigger nil-page and MAP_FIXED.
> - */
> - if ((shmflg & SHM_RND) && addr >= shmlba)
> - addr &= ~(shmlba - 1);
> + if (shmflg & SHM_RND)
> + addr &= ~(shmlba - 1); /* round down */
> else
> #ifndef __ARCH_FORCE_SHMLBA
> if (addr & ~PAGE_MASK)
>
>
