9fans - fans of the OS Plan 9 from Bell Labs
 help / color / mirror / Atom feed
* 386 maximum segment size of 16Mb
@ 1995-11-02 16:21 G.David
  0 siblings, 0 replies; 5+ messages in thread
From: G.David @ 1995-11-02 16:21 UTC (permalink / raw)


Hello again,

Has anybody noticed the problem compiling /sys/src/libtiff/tif_fax3.c on
a 386?  It seems 8c needs more than 16Mb of data space to build this one.

The real question, has anybody looked into increasing this limit?  (look
in /sys/src/9/pc/mem.h, SEGMEMSIZE, I think, set to 16.)

Thanks.






^ permalink raw reply	[flat|nested] 5+ messages in thread

* 386 maximum segment size of 16Mb
@ 1995-11-03  8:33 Nigel
  0 siblings, 0 replies; 5+ messages in thread
From: Nigel @ 1995-11-03  8:33 UTC (permalink / raw)


> 
> changing SEGMAPSIZE works fine.
> It incurs overhead allocated Segment structures
> so 16 seemed reasonable for a terminal. We really
> dont use PC's to compile.
> 
> 

Well, wouldn't you know. I was going to post all this yesterday, and 
as we get up earlier in the morning this side of the pond, I would 
have got there first. I fixed SEGMAPSIZE to 32, so must be half a 
woos (?sp).

What you have to watch is the linker! It gets to 14 meg doing 
9pcdisk; it gets to 20mb doing ghostscript, even without any useful 
printer drivers added in (which was what I was doing). Anyone else 
tried compiling in, say, HP deskjet support?






^ permalink raw reply	[flat|nested] 5+ messages in thread

* 386 maximum segment size of 16Mb
@ 1995-11-02 18:38 philw
  0 siblings, 0 replies; 5+ messages in thread
From: philw @ 1995-11-02 18:38 UTC (permalink / raw)



changing SEGMAPSIZE works fine.
It incurs overhead allocated Segment structures
so 16 seemed reasonable for a terminal. We really
dont use PC's to compile.






^ permalink raw reply	[flat|nested] 5+ messages in thread

* 386 maximum segment size of 16Mb
@ 1995-11-02 18:28 G.David
  0 siblings, 0 replies; 5+ messages in thread
From: G.David @ 1995-11-02 18:28 UTC (permalink / raw)


>changing SEGMAPSIZE works fine.
>It incurs overhead allocated Segment structures
>so 16 seemed reasonable for a terminal.

I agree.

>                                        We really
>dont use PC's to compile.

I was compiling on my cpu server, which is just a bigger PC.

Thanks.






^ permalink raw reply	[flat|nested] 5+ messages in thread

* 386 maximum segment size of 16Mb
@ 1995-11-02 17:14 G.David
  0 siblings, 0 replies; 5+ messages in thread
From: G.David @ 1995-11-02 17:14 UTC (permalink / raw)


I hate replying to my own mail, but I guess I wasn't finished... :-)

>The real question, has anybody looked into increasing this limit?  (look
>in /sys/src/9/pc/mem.h, SEGMEMSIZE, I think, set to 16.)
Make that                SEGMAPSIZE.

Also several of the machines have this in mem.h:

#define SEGMAPSIZE 64 /* 16 is for wooses */

This would imply that all is necessary is to change 16 to 64 in pc/mem.h,
but that would be too simple...

Thanks again.






^ permalink raw reply	[flat|nested] 5+ messages in thread

end of thread, other threads:[~1995-11-03  8:33 UTC | newest]

Thread overview: 5+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
1995-11-02 16:21 386 maximum segment size of 16Mb G.David
1995-11-02 17:14 G.David
1995-11-02 18:28 G.David
1995-11-02 18:38 philw
1995-11-03  8:33 Nigel

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).