Announcement

Collapse
No announcement yet.

Oracle's Kernel Test Framework Might Be Added To The Linux Kernel Tree

Collapse
X
  • Filter
  • Time
  • Show
Clear All
new posts

  • Oracle's Kernel Test Framework Might Be Added To The Linux Kernel Tree

    Phoronix: Oracle's Kernel Test Framework Might Be Added To The Linux Kernel Tree

    Knut Omang of Oracle is working on integrating the Kernel Test Framework into the Linux kernel source tree/repository...

    http://www.phoronix.com/scan.php?pag...-Framework-Git

  • #2
    So ORACLE, the company who tried to copyright fucking APIs id s trying to mainline some Linux based thing they made huh?

    Its stilk a little confusing as to what it does. Improves backporting?

    Comment


    • #3
      Originally posted by ElectricPrism View Post
      Its stilk a little confusing as to what it does. Improves backporting?
      It lets you run tests on the kernel. That's the sort of stuff that's supposed to prevent regressions during future development. It's generally a good thing as long as it's not filled with stupid useless tests that don't check anything useful.

      They want it in the tree to ease their maintenance burden for the framework. Unless there's something better in the kernel for dealing with it, test frameworks can add value. A few potential problems:

      1. Oracle is known for over-engineering fragile stuff and for avoiding responsibility for anything they can blame on the linux kernel/distro. Though those are separate departments (dev vs support) and the fragility in part can be blamed on years of opaque, poorly documented feature tests built for their database engine. I bet that uses a different framework, though.

      They do hire some good engineers sometimes, but everyone's going to be eyeing this like it's got some sort of hidden plague until it gets reviewed a bit. The person who sent the email to the ML seemed approachable, and somewhat knowledgeable, which seems good. At least it can start a dialog about testing procedures.

      2. It's hard-coded to use Googletest. Might have other assumptions about servers and stuff that make it harder for regular people to use it. If the kernel's going to accept stuff, it's gotta be usable by anyone. Whatever makes it in won't be under Oracle control, that's for sure.

      3. It adds a new protocol, debugfs stuff and many other things that could cause security issues if left enabled. Tests are supposedly done as kernel modules, so there'll probably be review and discussion of whether this is separated cleanly. Expect more than the usual LKML implementation bike-shedding. That sounds negative, but much of the technical criticism is well-meaning and insightful.

      Hopefully they'll offer to donate some cloud server time to CI testing as part of this initiative.

      Comment


      • #4
        Originally posted by Terrablit View Post
        2. It's hard-coded to use Googletest. Might have other assumptions about servers and stuff.
        You mean whole patch or just googletest part of it. GoogleTest is C++ unit test framework and has nothing to do with "servers". I wouldn't call it hardcoding thought. It's like saying Qt apps are hardcoded to use Qt framework. Of course they are, they are written in Qt to begin with.
        I see the value here and prefer GoogleTest (with GoogleMock, hopefully they generate stubs automatically though) over other unit test C/C++ frameworks.

        Comment


        • #5
          Originally posted by ElectricPrism View Post
          So ORACLE, the company who tried to copyright fucking APIs id s trying to mainline some Linux based thing they made huh?

          Its stilk a little confusing as to what it does. Improves backporting?
          disclaimer: I work for Oracle, but nothing to do with the Oracle Linux team, Java team etc.

          This might blow your mind a little: https://git.kernel.org/pub/scm/linux...r&q=oracle.com
          Oracle contributes a lot to the Linux kernel, and employs a number of the core maintainers, including Darrick Wong (XFS maintainer), and Konrad Wilk (Xen etc.)

          Comment

          Working...
          X