On Tue, Dec 04, 2012 at 01:26:04PM +0100, Olof Kindgren wrote:
> 2012/12/4 Stefan Kristiansson <[email protected]>:
> > On Tue, Dec 04, 2012 at 02:10:42PM +0200, Stefan Kristiansson wrote:
> >> On Mon, Dec 03, 2012 at 03:17:26PM -0500, Peter Gavin wrote:
> >> > Another problem is that l.nop REPORT requires that register 3 be read by
> >> > the pipeline.  That's not possible without heavily altering the pipeline
> >> > design--the register addresses accessed by any given instruction need to 
> >> > be
> >> > available immediately without any decoding overhead.  So if we introduce 
> >> > a
> >> > debug UART we need to rethink the instructions used to output to it.  I
> >> > would suggest adding a separate instruction just for that purpose.
> >>
> >> How is that any different than any other instruction that is using 
> >> registers as
> >> source for their data?
> >> I agree that there will be some overhead if actually implemented in 
> >> hardware,
> >> but so would any other instruction outputting to a debug uart.
> >>
> >> Or am I completely missing your point here?
> >
> > Actually, reading your message again and giving this some more thought
> > I kind of see what you mean, if you want to have the register output
> > ready when you are doing the instruction decoding, you're kind of screwed
> > since you don't have the register address in the l.nop instruction.
> >
> > Stefan
> > _______________________________________________
> > Openrisc mailing list
> > [email protected]
> > http://lists.opencores.org/listinfo/openrisc
> 
> Would it help to use 5 bits in the nop immediate for a register number?
> 

No, but using the reserved bit [20:16] would.

Stefan
_______________________________________________
OpenRISC mailing list
[email protected]
http://lists.openrisc.net/listinfo/openrisc

Reply via email to