Apple Open Source: the drivers

Apple Open Source: the drivers

[Benjamin Herrenschmidt]

Ok, I finally found where the driver are hidden: In the kernel archive,
in bsd/dev/ppc. They are mostly objective C drivers. Among other things,
there are drivers for the PowerMac ATA with apparently all the timing
stuffs, and for the CMD646 controller in the blue G3.

There's a PMU driver with apparently up-to-date command length tables and
lots of comments about the commands ;-) An interesting thing that I've
not seen in the LinuxPPC driver: When sending read commands to
non-existing devices, the PMU may never answer, and so there's need for
some kind of timeout. I didn't find infos about putting the PowerBook to
sleep, maybe MacOS X Server cannot sleep...

Finally, the MESH driver is a piece of anthology, but I'll let interested
people discover it by themselves ;-)


-- 
           E-Mail: <mailto:bh40@calva.net>
BenH.      Web   : <http://calvaweb.calvacom.fr/bh40/>





[[ This message was sent via the linuxppc-dev mailing list.  Replies are ]]
[[ not  forced  back  to the list, so be sure to Cc linuxppc-dev if your ]]
[[ reply is of general interest. Please check http://lists.linuxppc.org/ ]]
[[ and http://www.linuxppc.org/ for useful information before posting.   ]]

Re: Apple Open Source: the drivers

[sean o'malley]

>There's a PMU driver with apparently up-to-date command length tables and
>lots of comments about the commands ;-) An interesting thing that I've
>not seen in the LinuxPPC driver: When sending read commands to
>non-existing devices, the PMU may never answer, and so there's need for
>some kind of timeout. I didn't find infos about putting the PowerBook to
>sleep, maybe MacOS X Server cannot sleep...

Hey maybe we will get the b/w g-3's (yMacs) running yet..

Macos X Server doesnt have any sleep code.. it wasnt written for that..
you also wont see _much_ USB (keyboard and mouse) support and it is still
lacking Firewire support from everything i have read.

Powerbooks arent supported at all..
Which makes me wonder why... Apple was giving major discounts to developers
on those things. unless they just wanted to give out information and
support linuxppc..





[[ This message was sent via the linuxppc-dev mailing list.  Replies are ]]
[[ not  forced  back  to the list, so be sure to Cc linuxppc-dev if your ]]
[[ reply is of general interest. Please check http://lists.linuxppc.org/ ]]
[[ and http://www.linuxppc.org/ for useful information before posting.   ]]

Re: Apple Open Source: the drivers

[Hollis R Blanchard]

On Wed, 17 Mar 1999, Benjamin Herrenschmidt wrote:
> 
> Ok, I finally found where the driver are hidden: In the kernel archive,
> in bsd/dev/ppc. They are mostly objective C drivers. Among other things,
> there are drivers for the PowerMac ATA with apparently all the timing
> stuffs, and for the CMD646 controller in the blue G3.

So they're useful? Last night I suddenly got worried because OS X is based on
Mach... so wouldn't any OS X drivers be more helpful to the MkLinux folks than
to monolithic Linux?

And another question: how is it that OS X, based on Mach, is still comparable
in speed to other non-microkernel operating systems? It was my (uneducated)
impression that one of the tradeoffs of a microkernel was a performance
penalty.

-Hollis


[[ This message was sent via the linuxppc-dev mailing list.  Replies are ]]
[[ not  forced  back  to the list, so be sure to Cc linuxppc-dev if your ]]
[[ reply is of general interest. Please check http://lists.linuxppc.org/ ]]
[[ and http://www.linuxppc.org/ for useful information before posting.   ]]

Re: Apple Open Source: the drivers

[David A. Gatwood]

On Wed, 17 Mar 1999, Hollis R Blanchard wrote:

> And another question: how is it that OS X, based on Mach, is still comparable
> in speed to other non-microkernel operating systems? It was my (uneducated)
> impression that one of the tradeoffs of a microkernel was a performance
> penalty.

There is a performance penalty, but not very much.  This is substantially
reduced when the server is collocated (loaded into kernel space), since
this eliminates the need for message passing.  The reason MkLinux has a
bigger speed tradeoff has less to do with the MK architecture and more to
do with slow legacy drivers. Many of MkLinux's drivers can be traced back
for a number of years (some, like the serial driver, may even date back to
pre-Mac days -- not positive, though).  Since MacOS X and X Server use
different drivers (with different driver frameworks, etc.), they should
perform better because of that alone.  Apple's compiler optimization no
doubt helps, too, though. :-)


Later,
David

David A. Gatwood                         Visit globegate's internet
dgatwood@globegate.utm.edu                  talker, Deep Space 36
http://globegate.utm.edu                telnet globegate.utm.edu:9624


[[ This message was sent via the linuxppc-dev mailing list.  Replies are ]]
[[ not  forced  back  to the list, so be sure to Cc linuxppc-dev if your ]]
[[ reply is of general interest. Please check http://lists.linuxppc.org/ ]]
[[ and http://www.linuxppc.org/ for useful information before posting.   ]]