exec: delay address limit change until point of no return
commitc14276ca9b14f8a7f3b40bec3083c276754e012b
authorMathias Krause <minipli@googlemail.com>
Thu, 9 Jun 2011 18:05:18 +0000 (9 20:05 +0200)
committerGreg Kroah-Hartman <gregkh@suse.de>
Thu, 23 Jun 2011 22:05:36 +0000 (23 15:05 -0700)
treeac90706b7b44bdfeffa3f31ee58854b3fcce94c8
parent50091c12f224d7b767efd11478f6948812728c07
exec: delay address limit change until point of no return

commit dac853ae89043f1b7752875300faf614de43c74b upstream.

Unconditionally changing the address limit to USER_DS and not restoring
it to its old value in the error path is wrong because it prevents us
using kernel memory on repeated calls to this function.  This, in fact,
breaks the fallback of hard coded paths to the init program from being
ever successful if the first candidate fails to load.

With this patch applied switching to USER_DS is delayed until the point
of no return is reached which makes it possible to have a multi-arch
rootfs with one arch specific init binary for each of the (hard coded)
probed paths.

Since the address limit is already set to USER_DS when start_thread()
will be invoked, this redundancy can be safely removed.

Signed-off-by: Mathias Krause <minipli@googlemail.com>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
arch/x86/kernel/process_32.c
arch/x86/kernel/process_64.c
fs/exec.c