|
Message
From: Marko Mlinar<markom@o...>
Date: Fri Jan 23 08:12:41 CET 2004
Subject: [openrisc] ORP memory layout
Heiko,no, you should use exactly the same executable on or1ksim and your HW. What is usually done is than when system starts, MC (momory controller) is not inicialized and flash is seen all over the address map. Therefore or1k has no problems booting from 0x100. Next, or1k jumps to the address where flash will reside after the inicialization of MC. MC inicialization is done and RAM gets visible at 0. Next, .data and .bss and possibly some other sections as well gets copied and initialized. Then program jumps to "main" routine of C program or some other startup routine, which calls "main".
You have examples how this functions in basically all or1k programs, e.g. orpmon.
hope it helps, Marko
On Thursday 22 January 2004 16:12, Heiko Panther wrote: > ORP.txt says there is 1G RAM at address 0x0. OR1000 wants to find its > reset vector at addr 0x100. So obviously, steps must be taken to have > program code available in the RAM area after reset. > > The ORP HDL code mirrors the ROM temporarily to address 0x0. So that is > a possible solution. Still I don't understand why ORP places RAM at the > processors reset address. > > But that scheme doesn't work on the or1ksim, which doesn't do the > mirroring. Does that mean I have to build a seperate executable for use > with or1ksim? > > Heiko > _______________________________________________ > http://www.opencores.org/mailman/listinfo/openrisc
|
 |