]> www.infradead.org Git - users/willy/xarray.git/commit
binfmt_elf: mseal address zero
authorJeff Xu <jeffxu@chromium.org>
Tue, 6 Aug 2024 21:49:27 +0000 (21:49 +0000)
committerKees Cook <kees@kernel.org>
Wed, 14 Aug 2024 16:56:48 +0000 (09:56 -0700)
commit44f65d900698278a8451988abe0d5ca37fd46882
treedd9e6a1cc68db8ca862015ee1108ebf5a4f1a535
parent7d442a33bfe817ab2a735f3d2e430e36305354ea
binfmt_elf: mseal address zero

In load_elf_binary as part of the execve(), when the current
task’s personality has MMAP_PAGE_ZERO set, the kernel allocates
one page at address 0. According to the comment:

/* Why this, you ask???  Well SVr4 maps page 0 as read-only,
    and some applications "depend" upon this behavior.
    Since we do not have the power to recompile these, we
     emulate the SVr4 behavior. Sigh. */

At one point, Linus suggested removing this [1].

Code search in debian didn't see much use of MMAP_PAGE_ZERO [2],
it exists in util and test (rr).

Sealing this is probably safe, the comment doesn't say
the app ever wanting to change the mapping to rwx. Sealing
also ensures that never happens.

If there is a complaint, we can make this configurable.

Link: https://lore.kernel.org/lkml/CAHk-=whVa=nm_GW=NVfPHqcxDbWt4JjjK1YWb0cLjO4ZSGyiDA@mail.gmail.com/
Link: https://codesearch.debian.net/search?q=MMAP_PAGE_ZERO&literal=1&perpkg=1&page=1
Signed-off-by: Jeff Xu <jeffxu@chromium.org>
Link: https://lore.kernel.org/r/20240806214931.2198172-2-jeffxu@google.com
Signed-off-by: Kees Cook <kees@kernel.org>
fs/binfmt_elf.c
include/linux/mm.h
mm/mseal.c