Patrick McNamara wrote:
> ----- Original Message ----
> From: Petter Urkedal <[EMAIL PROTECTED]>
> To: [email protected]
> Sent: Friday, July 28, 2006 11:05:41 AM
> Subject: [Open-graphics] Memory Controller
>
>  <snip>
>   
>>  Is it okay to commit the current work to Subversion?  I can use
>> rtl/mem_ctl/pnp (pipelined non-programmable) or the like, in case
>> there'll be more implementations.
>>     
>
>
>  
>  I'm certainly not the keeper of the SVN repository, but this does come 
> closer to my day job so I'll toss in my thoughts.  I would suggest we go 
> ahead and just put it at rtl/mem_ctl.  First implementation becomes the 
> default for the time being.  People making functional changes can do so on 
> branches of the code base.  Should we get to the point where we have to 
> viable memory controller options, we can re-arrange the directory and merge 
> the second, or third etc, verison back down to the main branch.
>   

Sorry, I committed before reading your mail.  We should maybe have a
policy for how to arrange the source tree.  Maybe even some common
makefile rules.  If I may extrapolate what you said, I'd suggest then
that we keep each core in a subdirectory directly under rtl, and if
there are two implementations of, say, a memory controllers, they just
have to be named differently (mem_ctl, ddrsdram_ctl, mem_ctrl_johndoe,
...).  Does that sound right?

In any case, we can rename my implementation to mem_ctl or mem_ctl_npi.

_______________________________________________
Open-graphics mailing list
[email protected]
http://lists.duskglow.com/mailman/listinfo/open-graphics
List service provided by Duskglow Consulting, LLC (www.duskglow.com)

Reply via email to