Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
MIME-Version: 1.0
References: <20190315180306.sq3z645p3hygrmt2@brauner.io> <20190315181324.GA248160@google.com>
 <20190315182426.sujcqbzhzw4llmsa@brauner.io> <20190315184903.GB248160@google.com>
 <CAJuCfpGp_9fE9MPGVCWjnTaeBE0K_Q22LS1pBqhp7zW2M=dbGw@mail.gmail.com>
 <CAKOZueuauUXRyrvhzBD0op6W4TAnydSx92bvrPN2VRWERX8iQg@mail.gmail.com>
 <20190316185726.jc53aqq5ph65ojpk@brauner.io> <CAJuCfpF-uYpUZ1RO99i2qEw5Ou4nSimSkiQvnNQ_rv8ogHKRfw@mail.gmail.com>
 <20190317015306.GA167393@google.com> <20190317114238.ab6tvvovpkpozld5@brauner.io>
 <20190317163505.GA9904@mail.hallyn.com>
In-Reply-To: <20190317163505.GA9904@mail.hallyn.com>
From:   Daniel Colascione <dancol@google.com>
Date:   Sun, 17 Mar 2019 10:11:10 -0700
Message-ID: <CAKOZuet+HCZoOgJBAUrcm8nxC-bQ00W7w+=k2SOh+dfXffMU4w@mail.gmail.com>
Subject: Re: [RFC] simple_lmk: Introduce Simple Low Memory Killer for Android
To:     "Serge E. Hallyn" <serge@hallyn.com>
Cc:     Christian Brauner <christian@brauner.io>,
        Joel Fernandes <joel@joelfernandes.org>,
        Suren Baghdasaryan <surenb@google.com>,
        Steven Rostedt <rostedt@goodmis.org>,
        Sultan Alsawaf <sultan@kerneltoast.com>,
        Tim Murray <timmurray@google.com>,
        Michal Hocko <mhocko@kernel.org>,
        Greg Kroah-Hartman <gregkh@linuxfoundation.org>,
        =?UTF-8?B?QXJ2ZSBIasO4bm5ldsOlZw==?= <arve@android.com>,
        Todd Kjos <tkjos@android.com>,
        Martijn Coenen <maco@android.com>,
        Ingo Molnar <mingo@redhat.com>,
        Peter Zijlstra <peterz@infradead.org>,
        LKML <linux-kernel@vger.kernel.org>,
        "open list:ANDROID DRIVERS" <devel@driverdev.osuosl.org>,
        linux-mm <linux-mm@kvack.org>,
        kernel-team <kernel-team@android.com>,
        Oleg Nesterov <oleg@redhat.com>,
        Andy Lutomirski <luto@amacapital.net>
Content-Type: text/plain; charset="UTF-8"
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

On Sun, Mar 17, 2019 at 9:35 AM Serge E. Hallyn <serge@hallyn.com> wrote:
>
> On Sun, Mar 17, 2019 at 12:42:40PM +0100, Christian Brauner wrote:
> > On Sat, Mar 16, 2019 at 09:53:06PM -0400, Joel Fernandes wrote:
> > > On Sat, Mar 16, 2019 at 12:37:18PM -0700, Suren Baghdasaryan wrote:
> > > > On Sat, Mar 16, 2019 at 11:57 AM Christian Brauner <christian@brauner.io> wrote:
> > > > >
> > > > > On Sat, Mar 16, 2019 at 11:00:10AM -0700, Daniel Colascione wrote:
> > > > > > On Sat, Mar 16, 2019 at 10:31 AM Suren Baghdasaryan <surenb@google.com> wrote:
> > > > > > >
> > > > > > > On Fri, Mar 15, 2019 at 11:49 AM Joel Fernandes <joel@joelfernandes.org> wrote:
> > > > > > > >
> > > > > > > > On Fri, Mar 15, 2019 at 07:24:28PM +0100, Christian Brauner wrote:
> > > > > > > > [..]
> > > > > > > > > > why do we want to add a new syscall (pidfd_wait) though? Why not just use
> > > > > > > > > > standard poll/epoll interface on the proc fd like Daniel was suggesting.
> > > > > > > > > > AFAIK, once the proc file is opened, the struct pid is essentially pinned
> > > > > > > > > > even though the proc number may be reused. Then the caller can just poll.
> > > > > > > > > > We can add a waitqueue to struct pid, and wake up any waiters on process
> > > > > > > > > > death (A quick look shows task_struct can be mapped to its struct pid) and
> > > > > > > > > > also possibly optimize it using Steve's TIF flag idea. No new syscall is
> > > > > > > > > > needed then, let me know if I missed something?
> > > > > > > > >
> > > > > > > > > Huh, I thought that Daniel was against the poll/epoll solution?
> > > > > > > >
> > > > > > > > Hmm, going through earlier threads, I believe so now. Here was Daniel's
> > > > > > > > reasoning about avoiding a notification about process death through proc
> > > > > > > > directory fd: http://lkml.iu.edu/hypermail/linux/kernel/1811.0/00232.html
> > > > > > > >
> > > > > > > > May be a dedicated syscall for this would be cleaner after all.
> > > > > > >
> > > > > > > Ah, I wish I've seen that discussion before...
> > > > > > > syscall makes sense and it can be non-blocking and we can use
> > > > > > > select/poll/epoll if we use eventfd.
> > > > > >
> > > > > > Thanks for taking a look.
> > > > > >
> > > > > > > I would strongly advocate for
> > > > > > > non-blocking version or at least to have a non-blocking option.
> > > > > >
> > > > > > Waiting for FD readiness is *already* blocking or non-blocking
> > > > > > according to the caller's desire --- users can pass options they want
> > > > > > to poll(2) or whatever. There's no need for any kind of special
> > > > > > configuration knob or non-blocking option. We already *have* a
> > > > > > non-blocking option that works universally for everything.
> > > > > >
> > > > > > As I mentioned in the linked thread, waiting for process exit should
> > > > > > work just like waiting for bytes to appear on a pipe. Process exit
> > > > > > status is just another blob of bytes that a process might receive. A
> > > > > > process exit handle ought to be just another information source. The
> > > > > > reason the unix process API is so awful is that for whatever reason
> > > > > > the original designers treated processes as some kind of special kind
> > > > > > of resource instead of fitting them into the otherwise general-purpose
> > > > > > unix data-handling API. Let's not repeat that mistake.
> > > > > >
> > > > > > > Something like this:
> > > > > > >
> > > > > > > evfd = eventfd(0, EFD_NONBLOCK | EFD_CLOEXEC);
> > > > > > > // register eventfd to receive death notification
> > > > > > > pidfd_wait(pid_to_kill, evfd);
> > > > > > > // kill the process
> > > > > > > pidfd_send_signal(pid_to_kill, ...)
> > > > > > > // tend to other things
> > > > > >
> > > > > > Now you've lost me. pidfd_wait should return a *new* FD, not wire up
> > > > > > an eventfd.
> > > > > >
> > > >
> > > > Ok, I probably misunderstood your post linked by Joel. I though your
> > > > original proposal was based on being able to poll a file under
> > > > /proc/pid and then you changed your mind to have a separate syscall
> > > > which I assumed would be a blocking one to wait for process exit.
> > > > Maybe you can describe the new interface you are thinking about in
> > > > terms of userspace usage like I did above? Several lines of code would
> > > > explain more than paragraphs of text.
> > >
> > > Hey, Thanks Suren for the eventfd idea. I agree with Daniel on this. The idea
> > > from Daniel here is to wait for process death and exit events by just
> > > referring to a stable fd, independent of whatever is going on in /proc.
> > >
> > > What is needed is something like this (in highly pseudo-code form):
> > >
> > > pidfd = opendir("/proc/<pid>",..);
> > > wait_fd = pidfd_wait(pidfd);
> > > read or poll wait_fd (non-blocking or blocking whichever)
> > >
> > > wait_fd will block until the task has either died or reaped. In both these
> > > cases, it can return a suitable string such as "dead" or "reaped" although an
> > > integer with some predefined meaning is also Ok.
> > >
> > > What that guarantees is, even if the task's PID has been reused, or the task
> > > has already died or already died + reaped, all of these events cannot race
> > > with the code above and the information passed to the user is race-free and
> > > stable / guaranteed.
> > >
> > > An eventfd seems to not fit well, because AFAICS passing the raw PID to
> > > eventfd as in your example would still race since the PID could have been
> > > reused by another process by the time the eventfd is created.
> > >
> > > Also Andy's idea in [1] seems to use poll flags to communicate various tihngs
> > > which is still not as explicit about the PID's status so that's a poor API
> > > choice compared to the explicit syscall.
> > >
> > > I am planning to work on a prototype patch based on Daniel's idea and post something
> > > soon (chatted with Daniel about it and will reference him in the posting as
> > > well), during this posting I will also summarize all the previous discussions
> > > and come up with some tests as well.  I hope to have something soon.
> >
> > Having pidfd_wait() return another fd will make the syscall harder to
> > swallow for a lot of people I reckon.
> > What exactly prevents us from making the pidfd itself readable/pollable
> > for the exit staus? They are "special" fds anyway. I would really like
> > to avoid polluting the api with multiple different types of fds if possible.
> >
> > ret = pidfd_wait(pidfd);
> > read or poll pidfd
>
> I'm not quite clear on what the two steps are doing here.  Is pidfd_wait()
> doing a waitpid(2), and the read gets exit status info?

pidfd_wait on an open pidfd returns a "wait handle" FD. The wait
handle works just like a pipe: you can select/epoll/whatever for
readability. read(2) on the wait handle (which blocks unless you set
O_NONBLOCK, just like a pipe) completes with a siginfo_t when the
process to which the wait handle is attached exits. Roughly,

int kill_and_wait_for_exit(int pidfd) {
  int wait_handle = pidfd_wait(pidfd);
  pidfd_send_signal(pidfd, ...);
  siginfo_t exit_info;
  read(wait_handle, &exit_info, sizeof(exit_info)); // Blocks because
we haven't configured non-blocking behavior, just like a pipe.
  close(wait_handle);
  return exit_info.si_status;
}

>
> > (Note that I'm traveling so my responses might be delayed quite a bit.)
> > (Ccing a few people that might have an opinion here.)
> >
> > Christian
>
> On its own, what you (Christian) show seems nicer.  But think about a main event
> loop (like in lxc), where we just loop over epoll_wait() on various descriptors.
> If we want to wait for any of several types of events - maybe a signalfd, socket
> traffic, or a process death - it would be nice if we can treat them all the same
> way, without having to setup a separate thread to watch the pidfd and send
> data over another fd.  Is there a nice way we can provide that with what you've
> got above?

Nobody is proposing any kind of mechanism that would require a
separate thread. What I'm proposing works with poll and read and
should be trivial to integrate into any existing event loop: from the
perspective of the event loop, it looks just like a pipe.