On Mon, Jun 20, 2016 at 8:24 AM, Oleg Nesterov <o...@redhat.com> wrote: > On 06/19, Andy Lutomirski wrote: >> >> On Sat, Jun 18, 2016 at 10:02 AM, Andy Lutomirski <l...@amacapital.net> >> wrote: >> Step 1: for 4.7 and for -stable, introduce TS_I386_REGS_POKED. Set it >> in putreg32. Use it in syscall_get_error, get_nr_restart_syscall, >> etc. Clear it in do_signal. > > do_signal() won't be necessarily called...
True. But I should have said "clear it in prepare_exit_to_usermode", and the patch I'm just about to send does that. > >> I wonder if we could actually get away with doing syscall restart >> processing before ptrace invocation. > > How? this doesn't look possible or I misunderstood. > > How about the simple change below for now? IIRC 32-bit task can't use > "syscall" so if syscall_get_nr() >= 0 then even the wrong TS_COMPAT is > not that bad, even if it "leaks" to user-mode. Hmm. That should fix the minor security issue, but it will even further break cross-arch tracing: now a 32-bit tracer tracing a 64-bit task that does int $0x80 will malfunction even more than it would have. Also, it relies on bizarre arch details IMO. I think I prefer my version, coming momentarily. --Andy
