* CPU server
@ 1996-03-17 1:27 Thomas
0 siblings, 0 replies; 6+ messages in thread
From: Thomas @ 1996-03-17 1:27 UTC (permalink / raw)
I'm sorry to clutter the bandwidth here... but since
the next thing that I have to deal with is setting up a CPU
server... is there a magic memory requirement for CPU servers
too?
-Tmo
---- Where theory and reality meet.
---- Thomas Riemer, triemer@wesleyan.edu
^ permalink raw reply [flat|nested] 6+ messages in thread
* CPU server
@ 1996-03-18 0:44 jim
0 siblings, 0 replies; 6+ messages in thread
From: jim @ 1996-03-18 0:44 UTC (permalink / raw)
There are no busmastering ethernet devices in the distribution;
the soundblaster and floppy use the generic DMA routines which
do the check and provide a buffer if necessary.
However, it's unlikely - Plan 9 does not do DMA to user
memory, only to kernel buffers. The space allocated by the
kernel at boot time for dynamic allocation is probably
within the first 16Mb.
------ forwarded message follows ------
>From cse.psu.edu!9fans-outgoing-owner Sun Mar 17 16:47:00 EST 1996
Received: from colossus.cse.psu.edu by plan9; Sun Mar 17 16:47:00 EST 1996
Received: by colossus.cse.psu.edu id <78848>; Sun, 17 Mar 1996 16:39:24 -0500
Received: from galapagos.cse.psu.edu ([130.203.2.12]) by colossus.cse.psu.edu with SMTP id <78847>; Sun, 17 Mar 1996 16:39:08 -0500
Received: from localhost by galapagos.cse.psu.edu with SMTP id <12690>; Sun, 17 Mar 1996 16:37:59 -0500
To: cse.psu.edu!9fans
Subject: Re: CPU server
In-reply-to: Your message of "Sun, 17 Mar 1996 15:03:05 EST."
<96Mar17.150802est.78842@colossus.cse.psu.edu>
Date: Sun, 17 Mar 1996 16:37:44 -0500
From: Scott Schwartz <galapagos.cse.psu.edu!schwartz>
Message-Id: <96Mar17.163759est.12690@galapagos.cse.psu.edu>
Sender: cse.psu.edu!owner-9fans
Precedence: bulk
Reply-To: cse.psu.edu!9fans
"jim mckie" <jmk@plan9.bell-labs.com> writes:
| The PC fileserver code has always been aware of this, the cpuserver/terminal
| code needs some fixes which went out recently.
What about devices other than disks? If I get more ram for my
terminal, will ethernet, soundblaster, etc, break?
^ permalink raw reply [flat|nested] 6+ messages in thread
* CPU server
@ 1996-03-17 21:37 Scott
0 siblings, 0 replies; 6+ messages in thread
From: Scott @ 1996-03-17 21:37 UTC (permalink / raw)
"jim mckie" <jmk@plan9.bell-labs.com> writes:
| The PC fileserver code has always been aware of this, the cpuserver/terminal
| code needs some fixes which went out recently.
What about devices other than disks? If I get more ram for my
terminal, will ethernet, soundblaster, etc, break?
^ permalink raw reply [flat|nested] 6+ messages in thread
* CPU server
@ 1996-03-17 20:03 jim
0 siblings, 0 replies; 6+ messages in thread
From: jim @ 1996-03-17 20:03 UTC (permalink / raw)
The PC fileserver code has always been aware of this, the cpuserver/terminal
code needs some fixes which went out recently.
>From all the problems which others are finding in the PC disc drivers it should
be apparent how much we actually use discs on pc's.
------ forwarded message follows ------
>From cse.psu.edu!9fans-outgoing-owner Sun Mar 17 14:11:37 EST 1996
Received: from colossus.cse.psu.edu by plan9; Sun Mar 17 14:11:37 EST 1996
Received: by colossus.cse.psu.edu id <78840>; Sun, 17 Mar 1996 14:05:11 -0500
Received: from galapagos.cse.psu.edu ([130.203.2.12]) by colossus.cse.psu.edu with SMTP id <78841>; Sun, 17 Mar 1996 14:04:55 -0500
Received: from localhost by galapagos.cse.psu.edu with SMTP id <12690>; Sun, 17 Mar 1996 14:03:34 -0500
To: cse.psu.edu!9fans
Subject: Re: CPU server
In-reply-to: Your message of "Sun, 17 Mar 1996 07:51:13 EST."
<96Mar17.074400est.78827@colossus.cse.psu.edu>
Date: Sun, 17 Mar 1996 14:03:24 -0500
From: Scott Schwartz <galapagos.cse.psu.edu!schwartz>
Message-Id: <96Mar17.140334est.12690@galapagos.cse.psu.edu>
Sender: cse.psu.edu!owner-9fans
Precedence: bulk
Reply-To: cse.psu.edu!9fans
forsyth@plan9.cs.york.ac.uk writes:
| with less than 16mbytes, the system will page dramatically when you
| link the PC kernel.
Speaking of that, don't pc's with isa devices need bounce buffers in
the device drivers in order to access more than 16M?
^ permalink raw reply [flat|nested] 6+ messages in thread
* CPU server
@ 1996-03-17 19:03 Scott
0 siblings, 0 replies; 6+ messages in thread
From: Scott @ 1996-03-17 19:03 UTC (permalink / raw)
forsyth@plan9.cs.york.ac.uk writes:
| with less than 16mbytes, the system will page dramatically when you
| link the PC kernel.
Speaking of that, don't pc's with isa devices need bounce buffers in
the device drivers in order to access more than 16M?
^ permalink raw reply [flat|nested] 6+ messages in thread
* CPU server
@ 1996-03-17 12:51 forsyth
0 siblings, 0 replies; 6+ messages in thread
From: forsyth @ 1996-03-17 12:51 UTC (permalink / raw)
>>server... is there a magic memory requirement for CPU servers
>>too?
8 mbytes and up is fine. i've used cpu servers with only 4mbytes for
specialised functions (and it's probably possible to squeeze that further),
but you won't want to run very much on anything under 8 mbytes.
with less than 16mbytes, the system will page dramatically when you
link the PC kernel.
^ permalink raw reply [flat|nested] 6+ messages in thread
end of thread, other threads:[~1996-03-18 0:44 UTC | newest]
Thread overview: 6+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
1996-03-17 1:27 CPU server Thomas
1996-03-17 12:51 forsyth
1996-03-17 19:03 Scott
1996-03-17 20:03 jim
1996-03-17 21:37 Scott
1996-03-18 0:44 jim
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).