Subject: Re:1.4.1 kernels
To: Bernhard H. Buckel <buckel@wmad95.mathematik.uni-wuerzburg.de>
From: David Brownlee <abs@anim.dreamworks.com>
List: port-vax
Date: 07/14/1999 21:52:59
On Wed, 14 Jul 1999, Bernhard H. Buckel wrote:

> Great! At the moment I'm trying to compile a kernel on my VS2000
> without swap and using a SCSI drive attached to the tape port - it's
> slow and watching the interrupt load on top is "tremendous fun" :-)
> 
	I always liked 'systat vmstat' myself :)

> Rejoice! This is great news for us "lunchboxers" - I was planning to
> run the lunchbox as a small webserver.
> 
	In that case I'd imagine you'd want to throw out the gfx board
	support to save more memory. I'm goint o try a 'minimal KA630'
	only config. If it saves as much as I hope I'll do the same for
	the KA410.

>     David>   Note: If you are using the _POSTAGE_STAMP_ config to
>     David> configure and build your own kernel you will need to
>     David> uncomment VAX43 support, or add an extra #ifdef to locore.c
> 
> I learnt this from my first try to compile a kernel, isn't it related
> to the SCSI-code for the VS3100?
> 
	Nope - I think Ragge cleaned that up - this is just a missed
	#ifdef around the part that detects if a machine is a MicroVAX
	or VAXstation. I'll probably commit it to -current, but its
	not going to make 1.4.1.

>     David>  Note2: I have tested the generic kernel, but was unable to
>     David> test the netbsd.ps kernel - please let me know if it works
>     David> for you!
> 
> Uhmmm... Did it refuse to work or what was the reason for this?
> 
	I don't have a Qbus machine here - all my Qbus machines are on
	the other side of the Atlantic (though I can usually harass
	Brian Chase to test things on his many boxes :)

> No nagging intended, but is there a chance that DMA will be supported
> on the tape controller of the VS2K someday? I looked at the code
> (isn't free software great?) but being mainly a <GASP>Pascal and
> Fortran</GASP> man, I couldn't understand it that much :-)
> 
	I understand Ragge has been reworking all the drivers to use
	the bus_space(9) and bus_dma(9) machine independent interfaces,
	which should make SCSI DMA a much cleaner prospect. Until he
	has completed that (probably quite significant :) task, I don't
	think its worth anyone working on DMA (but I could be wrong :)

	Hmm - maybe people could increase the relative priority of SCSI
	DMA by getting together and having a barrel of good German beer
	shipped to Ragge? :)

> P.p.s.:
> What does the word "hensa" mean? I saw this sometimes when my kernel
> panicked while fsck'ing. Just curious.

	I'm afraid I have _no_ idea on that one! :)

		David/absolute

                 -=-  "You can have my absence of faith"  -=-