OpenBSD Journal

Call for Testing: New AML parser for ACPI

Contributed by phessler on from the zombie-eats-aml-for-breakfast dept.

Earlier today, Marco re-enabled the new parser in ACPI. This fixes many previously broken machines, including my infamous 'zombie' machine that liked to break as much of the ACPI spec as possible. Now is time for our (i386 and amd64) readers to update to the latest snapshot (or build from source), and run with the new ACPI parser. Read on for Marco's email to misc@

(Update 2008-06-01: There is also a patch that fixes some PCI routing issues on additional machines but needs more testing. Test and report! 
A few minutes ago I re-enabled the new AML parser for ACPI.  This parser
fixes quite a few issues, including those pesky new HP boxes that were
crapping out with the setbufint panic.  Also ACPI interrupts in MP on
amd64 now work as well (at least on my machines).

Even though we think this helps a lot we want to make sure we do not
have any regressions.  So please check out a new tree and compile a new
kernel or download a snapshot and test acpi as much as you can.

What would help is to see if the dmesg changes.  So take a dmesg with
the old and the new kernel and run diff -uNp on it.  If anything changes
besides a little bit of memory usage please mail me the results with all
usual suspects (acpidump, dmesg old, dmesg new).

Do not wait on this, we need results now if this is to be included on
the next OpenBSD version.
We would like to add that all machines with ACPI (laptops, servers, desktops, workstations, etc) need testing.

(Comments are closed)

Comments

  1. By Anonymous Coward (81.169.167.206) on

    What exactly does this improve?
    I am currently building from src (current) but I would like to know more about what gets improved.

    F.e. I face "can't allocate I/O space" related to the Cardbus-System on a x61s.

    Furthermore there is a counting issue and I am unsure if it is a serious bug.

    I plugged the card into the laptop, removed it and attached it again and removed it again.

    com3 at pcmcia0 function 0: can't allocate i/o space
    com4 at pcmcia0 function 0: can't allocate i/o space

    As you can see "comX" gets incrased.
    I tried it with a old Netgear WLAN PCMICA Card (wi0).

    Any hints would be nice.

    Comments

    1. By okan (160.39.158.193) on

      > What exactly does this improve?
      > I am currently building from src (current) but I would like to know more about what gets improved.
      >
      > F.e. I face "can't allocate I/O space" related to the Cardbus-System on a x61s.
      >
      > Furthermore there is a counting issue and I am unsure if it is a serious bug.
      >
      > I plugged the card into the laptop, removed it and attached it again and removed it again.
      >
      > com3 at pcmcia0 function 0: can't allocate i/o space
      > com4 at pcmcia0 function 0: can't allocate i/o space
      >
      > As you can see "comX" gets incrased.
      > I tried it with a old Netgear WLAN PCMICA Card (wi0).
      >
      > Any hints would be nice.

      try -current with the latest patch on tech@ (link above). my x61s handles pcmcia now.

    2. By Anonymous Coward (213.41.185.88) on

      > What exactly does this improve?
      > I am currently building from src (current) but I would like to know more about what gets improved.

      You don't quite get it, I think.
      It's supposed to be a better parser, that will work with more laptops
      and situations. There's not a simple list of what it will improve.

      What it needs is tests, positive and negative tests.
      It means people have to get off their asses and try it out.
      Even if acpi works fine for them already ! especially if it works fine, because this might break some stuff for some people.

      The PC hardware, in general, is crap. Which is why acpi is progressing so slowly, because this is a minefield. Which is why it needs to be tested every step of the way, so that people see improvements all over the place, and not the sorry pattern `improvement for some, regression for others'.

  2. By Anonymous Coward (2001:6f8:94d:4:202:b3ff:feb7:54e8) on

    Starting which date (mtime) is it in the snapshots?

  3. By Anonymous Coward (83.226.152.62) on

    My Compaq Evo D310 now happily boots with acpi enabled, instead of crashing horribly. great work!

  4. By Remy Couture (208.111.87.229) on

    HP Compaq DC7100 is now booting. Thank you I needed it.

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.]