All of lore.kernel.org
 help / color / mirror / Atom feed
From: Hajime Tazaki<thehajime@gmail.com>
To: johannes@sipsolutions.net
Cc: linux-um@lists.infradead.org
Subject: Re: [PATCH 4/6] um: implement time-travel=ext
Date: Fri, 14 Feb 2020 14:34:33 +0900	[thread overview]
Message-ID: <m2d0ahwx3q.wl-thehajime@gmail.com> (raw)
In-Reply-To: <e3073fbe7298451467a3a484efad90611df19c9e.camel@sipsolutions.net>


Hello,

On Fri, 14 Feb 2020 00:52:39 +0900,
Johannes Berg wrote:

> > > The protocol for the unix domain socket is defined in the new file
> > > include/uapi/linux/um_timetravel.h.
> > 
> > I was wondering if the source code of the special application (central
> > scheduler?) is public (or not) so that we can test on a local machine.
> 
> So ... sadly, it's a bit of a sad story right now.
> 
> I have one, but it's part of a bigger testing framework that we wrote
> (and uses the infrastructure from it), that I haven't been able to
> release so far.

I understand.  Thanks for telling this.

> On this tool, conceptually, the code is very simple, just keep a sorted
> "next event" list, and tell the next one that it's their turn.

I can imagine how it looks like, as I was doing a similar integration
with kernel code and ns-3 network simulator (which isn't LKL).

This is just my 2 cents (and may be out of topic for the patch
review); it would be nice to have an example code of this external
program.  Since the program is independent from arch/um, this is a
great opportunity to extend in future.

One might wish to implement his/her own scheduler for particular
purpose; for instance, a randomized error injection into time progress
to exercise or fuzzing the Linux implementation, or alternate external
program that can handle multiple UML instances across different nodes
(it this already possible?).

In such case, having a reference implementation would be helpful.

> For a vhost-user device implementation, it's a bit more complex, as it
> needs to be able to handle scheduler messages while waiting for a vhost-
> user ACK message, similar to what you can see in the UML vhost-user
> code.

Let me try to look at the code.

-- Hajime

_______________________________________________
linux-um mailing list
linux-um@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-um


  reply	other threads:[~2020-02-14  5:34 UTC|newest]

Thread overview: 12+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-02-13 13:26 [PATCH 0/6] um: external time travel Johannes Berg
2020-02-13 13:26 ` [PATCH 1/6] um: move timer-internal.h to non-shared Johannes Berg
2020-02-13 13:26 ` [PATCH 2/6] um: time-travel: rewrite as an event scheduler Johannes Berg
2020-02-13 13:26 ` [PATCH 3/6] um: virtio: implement VHOST_USER_PROTOCOL_F_INBAND_NOTIFICATIONS Johannes Berg
2020-02-13 13:26 ` [PATCH 4/6] um: implement time-travel=ext Johannes Berg
2020-02-13 14:10   ` Hajime Tazaki
2020-02-13 15:52     ` Johannes Berg
2020-02-14  5:34       ` Hajime Tazaki [this message]
2020-02-14  8:14         ` Johannes Berg
2020-02-13 13:26 ` [PATCH 5/6] um: implement ndelay/udelay in time-travel mode Johannes Berg
2020-02-13 13:26 ` [PATCH 6/6] um: implement cpu_relax() as ndelay(1) for time-travel Johannes Berg
2020-03-29 21:48 ` [PATCH 0/6] um: external time travel Richard Weinberger

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=m2d0ahwx3q.wl-thehajime@gmail.com \
    --to=thehajime@gmail.com \
    --cc=johannes@sipsolutions.net \
    --cc=linux-um@lists.infradead.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body. 
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.