x86: Enable HAVE_ARCH_SECCOMP_FILTER
authorWill Drewry <wad@chromium.org>
Thu, 12 Apr 2012 21:48:03 +0000 (16:48 -0500)
committerJames Morris <james.l.morris@oracle.com>
Sat, 14 Apr 2012 01:13:21 +0000 (11:13 +1000)
commitc6cfbeb4029610c8c330c312dcf4d514cc067554
treed61f1431884e042dc47ceb020fbcfada0c66bb3b
parentfb0fadf9b213f55ca9368f3edafe51101d5d2deb
x86: Enable HAVE_ARCH_SECCOMP_FILTER

Enable support for seccomp filter on x86:
- syscall_get_arch()
- syscall_get_arguments()
- syscall_rollback()
- syscall_set_return_value()
- SIGSYS siginfo_t support
- secure_computing is called from a ptrace_event()-safe context
- secure_computing return value is checked (see below).

SECCOMP_RET_TRACE and SECCOMP_RET_TRAP may result in seccomp needing to
skip a system call without killing the process.  This is done by
returning a non-zero (-1) value from secure_computing.  This change
makes x86 respect that return value.

To ensure that minimal kernel code is exposed, a non-zero return value
results in an immediate return to user space (with an invalid syscall
number).

Signed-off-by: Will Drewry <wad@chromium.org>
Reviewed-by: H. Peter Anvin <hpa@zytor.com>
Acked-by: Eric Paris <eparis@redhat.com>
Reviewed-by: Kees Cook <keescook@chromium.org>
v18: rebase and tweaked change description, acked-by
v17: added reviewed by and rebased
v..: all rebases since original introduction.
Signed-off-by: James Morris <james.l.morris@oracle.com>
arch/x86/Kconfig
arch/x86/kernel/ptrace.c