Source-Changes archive

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Old Index]

Re: CVS commit: src



> From: Jason Thorpe <thorpej%nas.nasa.gov@localhost>
>
> On Mon, 19 Apr 1999 23:16:36 -0700 (PDT) 
>  Ross Harvey <ross%netbsd.org@localhost> wrote:
>
>  > Modified Files:
>  >    src/sys/arch/alpha/conf: GENERIC
>  > Log Message:
>  > options INSECURE, like i386
>  > pckbc* -> pckbc0, like i386
>
> Why?!

Oh, I didn't think this would be controversial. I did mention it to Charles
and to our security-ocifer (hic!) last night. :-)  Our security-officer
noted that in his opinion securelevel=1 was annoying for no real security.

Anyway, I can change it back if you are sure about this, but I'll list my
reasoning below:

>
>       - i386 uses INSECURE so that the X server works "out of the box".
>         It's not necessary on the Alpha.

Right, that's what I figured, and since...

    * -E1 X is working again
    * we seem to have mostly multia users (poor guys :-)
    * the -E2 card is almost working
    * the X server is being accelerated
    * i'll be working on VGA X server stuff soon

    ... it seemed like it would be nice if X worked `out of the box'
    on alpha too. We will have 1.4 users for A LONG TIME, and I was kind
    of planning on building X against 1.4 in the coming year if we made
    worthwhile enhancements that didn't require kernel mods.

    Given the huge time between releases, and since kernel changes are
    necessary for real X support, I was hoping to provide a "1.4 patch"
    kit later this year, depending on point releases, 1.5 plans, etc...

>       - pckbc* is perfectly legitimate, and I think cloned devices
>         are more appropriate in GENERIC than wired devices.  The
>         only reason this wasn't cloned in the i386 port is because
>         it used to use __BROKEN_INDIRECT_CONFIG; it would actually work
>         now.

    Well, OK, but it hardly matters, and there are other reasons...

    * pckbc0 is exactly what the man page says. (And there is so
      @@#*@!!-little documentation for wscons, it seems a pity to
      violate about the only #@@#*@! thing that _is_ written down
      about it. :-)

    * the damn thing _is_ hardwired at compile-time for I/O port 60,
      you know .. and the match routine just checks that and bails if
      you try to do anything else. There is some partially written and
      certainly never-successfully-tested code to operate an aux port
      keyboard, but even that must be hardwired in the config file or
      it bails. (And I think you can't have both.)

Ross.Harvey%Computer.Org@localhost



Home | Main Index | Thread Index | Old Index