Messages in this thread | | | Date | Wed, 21 Oct 2015 22:44:19 +0200 | From | Oleg Nesterov <> | Subject | Re: [PATCH 1/2] wait/ptrace: always assume __WALL if the child is traced |
| |
On 10/21, Andrew Morton wrote: > > On Wed, 21 Oct 2015 19:41:50 +0200 Oleg Nesterov <oleg@redhat.com> wrote: > > > On 10/20, Andrew Morton wrote: > > > > > > On Tue, 20 Oct 2015 19:17:54 +0200 Oleg Nesterov <oleg@redhat.com> wrote: > > > > > > > This is not a kernel bug, at least in a sense that everything works as > > > > expected: debugger should reap a traced sub-thread before it can reap > > > > the leader, but without __WALL/__WCLONE do_wait() ignores sub-threads. > > > > > > > > Unfortunately, it seems that /sbin/init in most (all?) distributions > > > > doesn't use it and we have to change the kernel to avoid the problem. > > > > > > Well, to fix this a distro needs to roll out a new kernel. Or a new > > > init(8). Is there any reason to believe that distributing/deploying a > > > new kernel is significantly easier for everyone? Because fixing init > > > sounds like a much preferable solution to this problem. > > > > I will be happy if we decide that this is userpace problem and we should > > not fix the kernel. I simply do not know. > > The kernel patch sounds pretty sketchy - something we should avoid > doing if at all possible.
Yes, I agree.
> > However, please look at 2/2 which imho makes sense regardless and looks > > "obviously safe". Without this patch waitid() can not use __WALL, so if > > /sbin/init uses waitid() then the userspace fix won't be one-liner. And > > at least Fedora22 and Ubuntu use waitid(). > > 2/2 does look sensible (needs a better changelog if it's to be a > standalone thing),
Yes. Without 1/2 the changlelog should menetion that at least __WALL makes sense because /sbin/init has a good reason to use waitid(WALL).
Plus it should cc -stable.
> but if we're expecting distros to fix this with an > updated init(8) only, then they can't assume that the kernel's waitid() > has been altered.
Well, 2/2 looks safe for every kernel version... starting from git history at least.
> So init will need the not-one-liner version of the > fix.
Then I think this fix will stay forever ;)
> > So personally I'd prefer 2/2 + fix-init, not sure if this can work... > > I'm just guessing here. Are you (or someone) able to find out which > approach the distros will prefer, and why?
No, I have no idea, sorry.
> And what has to be done to init(8) to fix this bug when running current > kernels?
Say, http://git.busybox.net/busybox/tree/init/init.c at first glance it just needs
- wpid = waitpid(-1, NULL, maybe_WNOHANG); + wpid = waitpid(-1, NULL, maybe_WNOHANG | __WALL);
I have a testing machine running Fedora22, according to strace /bin/systemd does
waitid(P_ALL, 0, {}, WNOHANG|WEXITED|WNOWAIT, NULL); ... waitid(P_PID, 25558, {INFO}, WEXITED, NULL)
so it probably wants siginfo and thus it can't use waitpid. Without 2/2 systemd can probably just do something like
while (waitpid(-1, NULL, __WCLONE | WNOHANG) != ESRCH) { log("Dmitry Vyukov detected"); }
every time it does waitid() to reap the traced subthreads.
Unless of course systemd itself uses ptrace or forks a child with (clone_flags & CSIGNAL) != SIGCHLD. Unlikely, but who knows.
In any case I think the fix should be simple. 2/2 can help, most probably systemd can too just add __WALL to wait options.
Oleg.
| |