All of lore.kernel.org
 help / color / mirror / Atom feed
From: Ian Campbell <ian.campbell@citrix.com>
To: Samuel Thibault <samuel.thibault@ens-lyon.org>,
	Wei Liu <wei.liu2@citrix.com>
Cc: minios-devel@lists.xenproject.org,
	Xen-devel <xen-devel@lists.xenproject.org>,
	Ian Jackson <Ian.Jackson@eu.citrix.com>
Subject: Re: [PATCH MINI-OS v3 2/2] xenbus: workaround oxenstored short-write
Date: Mon, 16 Nov 2015 11:30:14 +0000	[thread overview]
Message-ID: <1447673414.27871.19.camel@citrix.com> (raw)
In-Reply-To: <20151027154740.GF2483@var.bordeaux.inria.fr>

On Tue, 2015-10-27 at 16:47 +0100, Samuel Thibault wrote:
> Wei Liu, le Tue 27 Oct 2015 15:43:29 +0000, a écrit :
> > Oxenstored has a behaviour that it only writes a contiguous piece of
> > data. When it writes across ring boundary it will return a short-write
> > while there is still room. That leads to mini-os stalling when it sees
> > there is not enough data in the ring.
> > 
> > Given that oxenstored is the default xenstored implementation we think
> > it would be useful to workaround this for the benefit of running mini-
> > os
> > (and unikernel based on it) on any Xen installation.
> > 
> > Signed-off-by: Wei Liu <wei.liu2@citrix.com>
> 
> As discussed in the previous thread, this is rather a workaround.

Given that I'm in the process of applying "tools/ocaml/xb: Correct
calculations of data/space the ring" which I think is the real fix here am
unsure if we still want to take this workaround.

Any thoughts?


>  Better
> be safe than sorry.
> 
> Acked-by: Samuel Thibault <samuel.thibault@ens-lyon.org>
> 
> And could too be backported to stable branches.
> 
> > ---
> >  xenbus/xenbus.c | 10 ++++++++--
> >  1 file changed, 8 insertions(+), 2 deletions(-)
> > 
> > diff --git a/xenbus/xenbus.c b/xenbus/xenbus.c
> > index 0ab387a..abf8b1b 100644
> > --- a/xenbus/xenbus.c
> > +++ b/xenbus/xenbus.c
> > @@ -205,8 +205,11 @@ static void xenbus_thread_func(void *ign)
> >              prod = xenstore_buf->rsp_prod;
> >              DEBUG("Rsp_cons %d, rsp_prod %d.\n", xenstore_buf-
> > >rsp_cons,
> >                      xenstore_buf->rsp_prod);
> > -            if (xenstore_buf->rsp_prod - xenstore_buf->rsp_cons <
> > sizeof(msg))
> > +            if (xenstore_buf->rsp_prod - xenstore_buf->rsp_cons <
> > sizeof(msg)) {
> > +                /* Work around oxenstored bug */
> > +                notify_remote_via_evtchn(start_info.store_evtchn);
> >                  break;
> > +            }
> >              rmb();
> >              memcpy_from_ring(xenstore_buf->rsp,
> >                      &msg,
> > @@ -217,8 +220,11 @@ static void xenbus_thread_func(void *ign)
> >                      xenstore_buf->rsp_prod - xenstore_buf->rsp_cons,
> >                      msg.req_id);
> >              if (xenstore_buf->rsp_prod - xenstore_buf->rsp_cons <
> > -                    sizeof(msg) + msg.len)
> > +                    sizeof(msg) + msg.len) {
> > +                /* Work around oxenstored bug */
> > +                notify_remote_via_evtchn(start_info.store_evtchn);
> >                  break;
> > +            }
> >  
> >              DEBUG("Message is good.\n");
> >  
> > -- 
> > 2.1.4
> > 
> 
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xen.org
http://lists.xen.org/xen-devel

  reply	other threads:[~2015-11-16 11:30 UTC|newest]

Thread overview: 9+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2015-10-27 15:43 [PATCH MINI-OS v3 0/2] Two mini-os xenbus patches Wei Liu
2015-10-27 15:43 ` [PATCH MINI-OS v3 1/2] xenbus: notify the other end when necessary Wei Liu
2015-10-27 15:46   ` Samuel Thibault
2015-10-27 15:55     ` Wei Liu
2015-11-16 12:10     ` Ian Campbell
2015-10-27 15:43 ` [PATCH MINI-OS v3 2/2] xenbus: workaround oxenstored short-write Wei Liu
2015-10-27 15:47   ` Samuel Thibault
2015-11-16 11:30     ` Ian Campbell [this message]
2015-11-16 11:47       ` Wei Liu

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=1447673414.27871.19.camel@citrix.com \
    --to=ian.campbell@citrix.com \
    --cc=Ian.Jackson@eu.citrix.com \
    --cc=minios-devel@lists.xenproject.org \
    --cc=samuel.thibault@ens-lyon.org \
    --cc=wei.liu2@citrix.com \
    --cc=xen-devel@lists.xenproject.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.