public inbox for cgen@sourceware.org
 help / color / mirror / Atom feed
* CGEN: Program vs Data Memory
@ 2000-09-15 13:02 Dave Brolley
  2000-09-15 13:10 ` Doug Evans
  0 siblings, 1 reply; 3+ messages in thread
From: Dave Brolley @ 2000-09-15 13:02 UTC (permalink / raw)
  To: cgen

Hi,

The port I'm working on has a harvard Architecture with separate program
and data memory, both addressed from 0 to 4M. Does CGEN have anything
for handling this, or should I just map references to data memory myself
by adding 4M? How do I communicate this to gdb?

Dave

^ permalink raw reply	[flat|nested] 3+ messages in thread

* CGEN: Program vs Data Memory
  2000-09-15 13:02 CGEN: Program vs Data Memory Dave Brolley
@ 2000-09-15 13:10 ` Doug Evans
  2000-09-15 13:49   ` Dave Brolley
  0 siblings, 1 reply; 3+ messages in thread
From: Doug Evans @ 2000-09-15 13:10 UTC (permalink / raw)
  To: Dave Brolley; +Cc: cgen

Dave Brolley writes:
 > The port I'm working on has a harvard Architecture with separate program
 > and data memory, both addressed from 0 to 4M. Does CGEN have anything
 > for handling this, or should I just map references to data memory myself
 > by adding 4M? How do I communicate this to gdb?

Depends, I guess.

There is the hardware "selector" facility which was added to
support things like sparc's ASI's.  Maybe that could be used
[not saying you won't run into some unimplemented things though].

You need to elaborate on your gdb question.
How is this relevant to cgen?

^ permalink raw reply	[flat|nested] 3+ messages in thread

* Re: CGEN: Program vs Data Memory
  2000-09-15 13:10 ` Doug Evans
@ 2000-09-15 13:49   ` Dave Brolley
  0 siblings, 0 replies; 3+ messages in thread
From: Dave Brolley @ 2000-09-15 13:49 UTC (permalink / raw)
  To: Doug Evans; +Cc: cgen

Doug Evans wrote:

> Dave Brolley writes:
>  > The port I'm working on has a harvard Architecture with separate program
>  > and data memory, both addressed from 0 to 4M. Does CGEN have anything
>  > for handling this, or should I just map references to data memory myself
>  > by adding 4M? How do I communicate this to gdb?
>
> Depends, I guess.
>
> There is the hardware "selector" facility which was added to
> support things like sparc's ASI's.  Maybe that could be used
> [not saying you won't run into some unimplemented things though].

I can't find anything in the documentation about this. Can you elaborate on
the 'selector' facility. Could you also explain to me what sparc's ASI's are?
:-)

> You need to elaborate on your gdb question.
> How is this relevant to cgen?

Maybe not directly related, but if I end up mapping memory myself, or using
some automatic mapping feature, gdb needs to know that program memory is
different from program memory when when setting breakpoints and
getting/setting variables. I'll ask on the gdb list, but I thought that if
somone here has implemented something like this, they would have to have
handled it already.

Dave

^ permalink raw reply	[flat|nested] 3+ messages in thread

end of thread, other threads:[~2000-09-15 13:49 UTC | newest]

Thread overview: 3+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2000-09-15 13:02 CGEN: Program vs Data Memory Dave Brolley
2000-09-15 13:10 ` Doug Evans
2000-09-15 13:49   ` Dave Brolley

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 read-only IMAP folder(s) and NNTP newsgroup(s).