[AUDIT] Don't allow ptrace to fool auditing, log arch of audited syscalls.
author <dwmw2@shinybook.infradead.org> <>
Fri, 29 Apr 2005 15:08:28 +0000 (16:08 +0100)
committer <dwmw2@shinybook.infradead.org> <>
Fri, 29 Apr 2005 15:08:28 +0000 (16:08 +0100)
commit2fd6f58ba6efc82ea2c9c2630f7ff5ed9eeaf34a
tree87cf236a78ad242ae01f1b71c289131e6d1c0662
parentea3834d9fb348fb1144ad3affea22df933eaf62e
[AUDIT] Don't allow ptrace to fool auditing, log arch of audited syscalls.

We were calling ptrace_notify() after auditing the syscall and arguments,
but the debugger could have _changed_ them before the syscall was actually
invoked. Reorder the calls to fix that.

While we're touching ever call to audit_syscall_entry(), we also make it
take an extra argument: the architecture of the syscall which was made,
because some architectures allow more than one type of syscall.

Also add an explicit success/failure flag to audit_syscall_exit(), for
the benefit of architectures which return that in a condition register
rather than only returning a single register.

Change type of syscall return value to 'long' not 'int'.

Signed-off-by: David Woodhouse <dwmw2@infradead.org>
arch/i386/kernel/ptrace.c
arch/ia64/kernel/ptrace.c
arch/mips/kernel/ptrace.c
arch/ppc64/kernel/ptrace.c
arch/s390/kernel/ptrace.c
arch/x86_64/kernel/ptrace.c
include/linux/audit.h
kernel/auditsc.c