compat_do_execve should unshare_files
commit3d068235e6c8580431f47254b71f75b1e70e6b26
authorHugh Dickins <hugh@veritas.com>
Sat, 28 Mar 2009 23:16:03 +0000 (28 23:16 +0000)
committerGreg Kroah-Hartman <gregkh@suse.de>
Fri, 8 May 2009 22:45:05 +0000 (8 15:45 -0700)
tree78ccbe511165145e7a4cf832f355ec383d925a4f
parent669e1834c4c57a880260a0e720266b1077d7d355
compat_do_execve should unshare_files

commit 53e9309e01277ec99c38e84e0ca16921287cf470 upstream.

2.6.26's commit fd8328be874f4190a811c58cd4778ec2c74d2c05
"sanitize handling of shared descriptor tables in failing execve()"
moved the unshare_files() from flush_old_exec() and several binfmts
to the head of do_execve(); but forgot to make the same change to
compat_do_execve(), leaving a CLONE_FILES files_struct shared across
exec from a 32-bit process on a 64-bit kernel.

It's arguable whether the files_struct really ought to be unshared
across exec; but 2.6.1 made that so to stop the loading binary's fd
leaking into other threads, and a 32-bit process on a 64-bit kernel
ought to behave in the same way as 32 on 32 and 64 on 64.

Signed-off-by: Hugh Dickins <hugh@veritas.com>
Cc: stable@kernel.org
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
fs/compat.c