OpenBSD Journal

Call For Testers - NFS mbuf handling diff needs testing

Contributed by johan on from the help-us-help-you dept.

Bret Lambert (blambert@) wrote to tech@ asking users to test a diff that improves the mbuf handling.

Please help test this diff so that OpenBSD NFS performance can be improved.

Please read on for Bret's message:

List:       openbsd-tech
Subject:    call for NFS testers
From:       "Bret S. Lambert" 
Date:       2008-11-02 17:38:29

Currently, NFS mbuf handling is atrocious. In order for any serious work
on NFS to go forward, it needs to be fixed. This diff brings us about
halfway to where we need to be (and there's a lot more NFS cleanup, not
just of mbuf handling, that needs to be done before anything new can
even be considered for inclusion).

So, this is a call for those of you who would like to see improvements
in NFS to help us help you. If you run an NFS setup, please test the
following diff and report either my glorious success or ignominious
failure.

Thank you.

- Bert

PS - If you're also willing to test further NFS diffs, please contact
     me off-list.

(Comments are closed)

Comments

  1. By Anonymous Coward (84.18.252.136) on

    I think you meant tech@ instead of misc@.

    Comments

    1. By Ray Lai (ray) ray@cyth.net on http://cyth.net/~ray/

      > I think you meant tech@ instead of misc@.

      Fixed, thanks!

  2. By Anonymous Coward (74.13.41.221) on

    The new mos driver is worth mention too.

    Comments

    1. By Anonymous Coward (67.69.227.99) on

      > The new mos driver is worth mention too.

      ?

      Comments

      1. By Dunceor (194.237.142.21) on

        > > The new mos driver is worth mention too.
        >
        > ?

        You forgot to add a question before your '?'.

    2. By phessler (phessler) on http://theapt.org

      > The new mos driver is worth mention too.

      in an article about NFS patches?

  3. By Mayuresh Kathe (59.182.235.91) kathe.mayuresh@gmail.com on http://kathe.in/

    A while back there was some activity about NFSv4 under OpenBSD, it seems to have dried up.
    Edd Barrett told me that NFSv4 isn't part of OpenBSD, is it because it isn't well tested?

    Comments

    1. By Paul 'WEiRD' de Weerd (weerd) on http://www.weirdnet.nl/openbsd/

      > A while back there was some activity about NFSv4 under OpenBSD, it seems to have dried up.
      > Edd Barrett told me that NFSv4 isn't part of OpenBSD, is it because it isn't well tested?
      >

      If you read Brets post to tech@ carefully, you'll find this :

      "(and there's a lot more NFS cleanup, not just of mbuf handling, that needs to be done before anything new can even be considered for inclusion)"

    2. By Ian McWilliam (123.243.195.128) on

      > A while back there was some activity about NFSv4 under OpenBSD, it seems to have dried up.
      > Edd Barrett told me that NFSv4 isn't part of OpenBSD, is it because it isn't well tested?
      >

      I'm assuming your talking about this project.

      http://snowhite.cis.uoguelph.ca/nfsv4/

      anybody shed some light?

      Comments

      1. By Paul Frank Covello (76.29.212.15) on

        > > A while back there was some activity about NFSv4 under OpenBSD, it seems to have dried up.
        > > Edd Barrett told me that NFSv4 isn't part of OpenBSD, is it because it isn't well tested?
        > >
        >
        > I'm assuming your talking about this project.
        >
        > http://snowhite.cis.uoguelph.ca/nfsv4/
        >
        > anybody shed some light?

      2. By Anonymous Coward (213.10.204.118) on

        > > A while back there was some activity about NFSv4 under OpenBSD, it seems to have dried up.
        > > Edd Barrett told me that NFSv4 isn't part of OpenBSD, is it because it isn't well tested?
        > >
        >
        > I'm assuming your talking about this project.
        >
        > http://snowhite.cis.uoguelph.ca/nfsv4/
        >
        > anybody shed some light?

        This is an NFSv4 implementation by Rick Macklem, the original author of the NFS code.

      3. By Mayuresh Kathe (59.182.226.241) kathe.mayuresh@gmail.com on http://kathe.in/

        > > A while back there was some activity about NFSv4 under OpenBSD, it seems to have dried up.
        > > Edd Barrett told me that NFSv4 isn't part of OpenBSD, is it because it isn't well tested?
        > >
        >
        > I'm assuming your talking about this project.
        >
        > http://snowhite.cis.uoguelph.ca/nfsv4/
        >
        > anybody shed some light?

        Yes, I was talking about that project, its strange that none of the BSDs want to work at incorporating it.

        I'm not a C programmer, so I don't know the complexities of adopting Macklem's code into OpenBSD, but, wouldn't the effort be worth it?

        Comments

        1. By tedu (udet) on


          > I'm not a C programmer, so I don't know the complexities of adopting Macklem's code into OpenBSD, but, wouldn't the effort be worth it?

          Worth what? How can you even do a cost/benefit when you don't know what the cost is?

        2. By Anonymous Coward (213.10.204.118) on

          > > > A while back there was some activity about NFSv4 under OpenBSD, it seems to have dried up.
          > > > Edd Barrett told me that NFSv4 isn't part of OpenBSD, is it because it isn't well tested?
          > > >
          > >
          > > I'm assuming your talking about this project.
          > >
          > > http://snowhite.cis.uoguelph.ca/nfsv4/
          > >
          > > anybody shed some light?
          >
          > Yes, I was talking about that project, its strange that none of the BSDs want to work at incorporating it.
          >
          > I'm not a C programmer, so I don't know the complexities of adopting Macklem's code into OpenBSD, but, wouldn't the effort be worth it?

          Rick Macklem is the author of the original NFS code.

          That code is currently in the process of being removed.

          If it looks anything like the code that's being removed, then the answer is a resounding 'no'.

          Comments

          1. By Mayuresh Kathe (59.182.239.41) kathe.mayuresh@gmail.com on http://kathe.in/

            > > > > A while back there was some activity about NFSv4 under OpenBSD, it seems to have dried up.
            > > > > Edd Barrett told me that NFSv4 isn't part of OpenBSD, is it because it isn't well tested?
            > > > >
            > > >
            > > > I'm assuming your talking about this project.
            > > >
            > > > http://snowhite.cis.uoguelph.ca/nfsv4/
            > > >
            > > > anybody shed some light?
            > >
            > > Yes, I was talking about that project, its strange that none of the BSDs want to work at incorporating it.
            > >
            > > I'm not a C programmer, so I don't know the complexities of adopting Macklem's code into OpenBSD, but, wouldn't the effort be worth it?
            >
            > Rick Macklem is the author of the original NFS code.
            >
            > That code is currently in the process of being removed.
            >
            > If it looks anything like the code that's being removed, then the answer is a resounding 'no'.

            Got your point, thanks for the note... :-)

  4. By Anonymous Coward (213.168.66.183) on

    For the dumb of us, who are running a nfs setup and would be willing to test, how do we apply the patch and what differences should we conduct our attention to?
    - copy the patch from the mail to a textfile and run which command?
    - take benchmarks before and after applying the patch?

    Thanks,
    anon cow

    Comments

    1. By Anonymous Coward (213.10.204.118) on

      > For the dumb of us, who are running a nfs setup and would be willing to test, how do we apply the patch and what differences should we conduct our attention to?
      > - copy the patch from the mail to a textfile and run which command?
      > - take benchmarks before and after applying the patch?
      >
      > Thanks,
      > anon cow

      Concact me (blambert@) off-list, and I can run you through these steps.

Latest Articles

Credits

Copyright © - Daniel Hartmeier. All rights reserved. Articles and comments are copyright their respective authors, submission implies license to publish on this web site. Contents of the archive prior to as well as images and HTML templates were copied from the fabulous original deadly.org with Jose's and Jim's kind permission. This journal runs as CGI with httpd(8) on OpenBSD, the source code is BSD licensed. undeadly \Un*dead"ly\, a. Not subject to death; immortal. [Obs.]