OpenBSD Journal

Heads up! Critical aml diff needs testing

Contributed by johan on from the mime dept.

Jordan Hargrave (jordan@) and Marco Pereeboom (marco@) posted a diff to the tech mailing list which fixes critical problems in the ACPI aml code.

The diff needs urgent review and testing, so please do try the diff and let Jordan and Marco know your results.

Please read on for Jordan's message.

List:       openbsd-tech
Subject:    Test fix for aml_xconvert die
From:       jordan () peereboom ! us
Date:       2009-05-19 5:27:09

For those of you who have been having problems with ACPI die/panic in
aml_xconvert (Could not convert!)

Can you try this patch against -current?  This is an attempt to handle
non-existing AML names in the namespace.

It will cancel execution of the current method and display an error message.

(Comments are closed)

Comments

  1. By Anonymous Coward (85.19.213.88) on

    > "For those of you who have been having problems with ACPI die/panic
    > in aml_xconvert ..."

    Well, I don't. Should I still test?

    Comments

    1. By Anonymous Coward (85.166.58.148) on

      > > "For those of you who have been having problems with ACPI die/panic
      > > in aml_xconvert ..."
      >
      > Well, I don't. Should I still test?
      >

      Yes, unless you like surprises.

      Comments

      1. By Anonymous Coward (85.19.213.88) on

        > > > "For those of you who have been having problems with ACPI die/panic
        > > > in aml_xconvert ..."
        > >
        > > Well, I don't. Should I still test?
        >
        > Yes, unless you like surprises.

        Then a simple "please test" would be sufficient. "For those of you"
        does not include me. It gave me the impression he only wanted reports
        from people actually having the problem described, rather than getting
        spammed with useless success reports from people not having a problem
        in the first place.

        If you want things tested, at least be precise in what you want.

        And yeah, I'll give the diff a spin ...

        Comments

        1. By David Chisnall (82.7.192.45) on

          > Then a simple "please test" would be sufficient. "For those of you"
          > does not include me. It gave me the impression he only wanted reports
          > from people actually having the problem described, rather than getting
          > spammed with useless success reports from people not having a problem
          > in the first place.

          Both of these are likely true. Reports of success from people who didn't previously have this problem are not useful. Reports of success from people who did show that the diff worked, reports of failures from people who didn't show that it introduced other bugs. Hopefully, if you test it and didn't have the problem, you will see no change, and so don't need to report anything, but no code has ever been made worse by thorough testing.

          Comments

          1. By sthen (2a01:348:108:100:230:18ff:fea0:6af6) on

            > > Then a simple "please test" would be sufficient. "For those of you"
            > > does not include me. It gave me the impression he only wanted reports
            > > from people actually having the problem described, rather than getting
            > > spammed with useless success reports from people not having a problem
            > > in the first place.
            >
            > Both of these are likely true. Reports of success from people who didn't previously have this problem are not useful. Reports of success from people who did show that the diff worked, reports of failures from people who didn't show that it introduced other bugs. Hopefully, if you test it and didn't have the problem, you will see no change, and so don't need to report anything, but no code has ever been made worse by thorough testing.

            I agree with most of this, but seeing how few things get tested, I think most people who ask for a diff to be tested will be delighted to have a quick mail (off-list is usually best), with a subject like "tested aml diff, no regressions" with a dmesg pasted into the body. (And, for the love of $DEITY, not mime encoded. :-p)

            Comments

            1. By Anonymous Coward (143.166.255.63) on

              Yes, sthen is right that is what we want

              Comments

              1. By Anonymous Coward (122.49.162.211) on

                Dr. Samuel Johnson's right about Olson Johnson being right!

            2. By Anonymous Coward (69.168.48.76) on

              > > > Then a simple "please test" would be sufficient. "For those of you"
              > > > does not include me. It gave me the impression he only wanted reports
              > > > from people actually having the problem described, rather than getting
              > > > spammed with useless success reports from people not having a problem
              > > > in the first place.
              > >
              > > Both of these are likely true. Reports of success from people who didn't previously have this problem are not useful. Reports of success from people who did show that the diff worked, reports of failures from people who didn't show that it introduced other bugs. Hopefully, if you test it and didn't have the problem, you will see no change, and so don't need to report anything, but no code has ever been made worse by thorough testing.
              >
              > I agree with most of this, but seeing how few things get tested, I think most people who ask for a diff to be tested will be delighted to have a quick mail (off-list is usually best), with a subject like "tested aml diff, no regressions" with a dmesg pasted into the body. (And, for the love of $DEITY, not mime encoded. :-p)


              Good point, I had no issues but didn't include any dmesg so i probably fell into the spam category. Oops.

    2. By Anonymous Coward (67.193.56.126) on

      > > "For those of you who have been having problems with ACPI die/panic
      > > in aml_xconvert ..."
      >
      > Well, I don't. Should I still test?
      >

      If you care about having a decent and stable OpenBSD experience then yes, you absolutely should test it. There is always a chance that a change could cause your currently working system to start flaking out.

  2. By Bayu Krisnawan (125.164.138.114) krisna@versalite.com on http://orchid.versalite.com

    I don't need to manually 'disable acpiec*' but still got error.
    --
    acpiec0 at acpi0### AML PARSE ERROR (0x3fa7): Undefined name: \\_SB_.C002.C121.C1BE
    error evaluating: \\_SB_.C002.C003.C005._REG
    : Failed to register address space
    ---

    Already sent dmesg to marco@ and jordan@

    Comments

    1. By Anonymous Coward (80.239.26.211) on

      > I don't need to manually 'disable acpiec*' but still got error.
      > --
      > acpiec0 at acpi0### AML PARSE ERROR (0x3fa7): Undefined name: \\_SB_.C002.C121.C1BE
      > error evaluating: \\_SB_.C002.C003.C005._REG
      > : Failed to register address space
      > ---
      >
      > Already sent dmesg to marco@ and jordan@

      And this is why those who don't like surprises should still test.

      Comments

      1. By Anonymous Coward (69.91.86.39) on

        This is actually the new code in action and working as expected.

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.]