Messages in this thread | ![/](/images/icornerl.gif) | | From | KOSAKI Motohiro <> | Subject | Re: [PATCH 1/1][2nd resend] sys_unshare: remove the dead CLONE_THREAD/SIGHAND/VM code | Date | Sun, 14 Nov 2010 16:14:01 +0900 (JST) |
| |
> On 11/09, KOSAKI Motohiro wrote: > > > > > -static void check_unshare_flags(unsigned long *flags_ptr) > > > +static int check_unshare_flags(unsigned long unshare_flags) > > > { > > > + if (unshare_flags & ~(CLONE_THREAD|CLONE_FS|CLONE_NEWNS|CLONE_SIGHAND| > > > + CLONE_VM|CLONE_FILES|CLONE_SYSVSEM| > > > + CLONE_NEWUTS|CLONE_NEWIPC|CLONE_NEWNET)) > > > + return -EINVAL; > > > > Please put WARN_ON_ONCE() explicitly. That's good way to find hidden > > user if exist and getting better bug report. > > Perhaps... but this needs a separate change. > > Please note that this check was simply moved from sys_unshare(), this > patch shouldn't have any visible effect. > > Personally, I think it would be even better if, say, unshare(CLONE_THREAD) > returned -EINVAL unconditionally.
Ah, OK. you are right.
> > And, I've reveied this patch and I've found no fault. but I will not put > > my ack because I think I haven't understand original intention perhaps. > > Thanks! > > IIRC, the main (only?) motivation for sys_unshare() was unshare_fs(). > Most probably unshare_thread/vm were added as placeholders.
| ![\](/images/icornerr.gif) |