Re: Let's vote for PnP on 2.2 (Was: PnP in 2.2 ?)

Albert D. Cahalan (acahalan@cs.uml.edu)
Mon, 8 Dec 1997 13:00:22 -0500 (EST)


>> vote ? when have we started to vote for stuff in the kernel ?
>> i must have missed some things... the last, i've seen, was
>> Linus denying the GGI in the kernel and myself screaming because
>> this damn X server was locking my workstation again and i've to
>> login over the net to kill the process.. the guy, who said linux
>> is stable, was lieing.. sorry for this kernel-policital mail..
>
> So complain to the xfree people or whereever else you got your
> X server from. The kernel didn't barf => no bug in Linux.

The video (note: hardware) is useless => bug in the linux kernel.
Windows 3.1 ran on top of DOS, much like X runs on top of Linux.
In both cases, stability was/is poor. In a unix system, hardware
access is supposed to be through the kernel. See /dev and /proc.

Some of you are very lucky. You can telnet in and fix the video.
For other people, "fix the video" requires a _hard_ reboot.
Power switch only, not even rebooting to NT can fix the hardware.
No, restoretextmode won't work. No, DOSEMU won't work. Nothing!
The card can only be reset by knowing the current state.

In any case, XFree86 can't do DMA to the video card. Doesn't that
remind anybody of older IDE controllers that eat all the CPU time?
XFree86 can't use an IRQ to know when the video accelerator is
ready for another command. Instead, XFree86 has to poll the card.
Doesn't that remind anybody of the junk SCSI cards you sometimes
get "free" with a scanner or CD-ROM drive?

Back to the PnP subject:

Linux PnP is not MS PnP. It is a good feature. Mostly, it consists
of a resource manager to prevent conflicts. (last time I heard...)
PnP is the way that hardware should be -- only the transition is bad.