Forget my little cry-baby whining ... more useful answere would be to; According to the HowTo "ALIM1631_M1535D" ... the correct config should set "docipl" to; mainboard/asus/cua/ipl.S
In that directory, however, there is a (mainboard/asus/cua/) there is "config.DoC" that says to set "docipl" to; northbridge/acer/m1631/ipl.S Which of these is the right one for the m1631? (there are a few diffs in there) < * $Id: ipl.S,v 1.1 2001/09/12 22:40:47 rminnich Exp $ --- > * $Id: ipl.S,v 1.25 2001/08/04 22:39:44 rminnich Exp $ ----- Original Message ----- >From: "Ian" <[EMAIL PROTECTED]> >To: "Ronald G Minnich" <[EMAIL PROTECTED]> >Subject: Re: RAM Init >Date: Wed, 20 Feb 2002 10:12:26 +1000 > > > Ok, from the snippet ... (of ~/m1631/raminit.inc) ... > > // if we are using DOC MIL, the initial setup is not needed. > #ifdef USE_DOC_MIL > // pull in the value of the current register. > inl %dx, %ecx > jmp oddslot > #endif > > <snippy snip> > > oddslot: > // OK, memory is on. Size is in @0. > // Turn on the odd slot and see if it's there. > orl $0x1800000, %ecx > WRITE_MCR0 > // ok, the odd bank is on. See if there's something there. > // put a 0 in, if there's something there, it will read back. > movl 0, %esi > movl $0, (%esi) > cmp $0, (%esi) > jz odd_slot_ok > andl (~$0x1800000), %ecx > WRITE_MCR0 > odd_slot_ok: > // pick the next memory register to configure. > // If done, drop out. > > > > What is "@0" in this language? I thought @ was a macro decleration (from my a86 >days)? > > Generally, I'm assuming that on DoC, due to space restrictions, you have some sort >of > predefined table for your specific RAM config ... which is why we're skirting around >all > of the autodetect / sizing guff ... > > ... but then you've got memory sizing twice ... once in ipl.S and again in >raminit.inc ... I > don't see where these pieces join ... ? > > > > ----- Original Message ----- > >From: "Ronald G Minnich" <[EMAIL PROTECTED]> > >To: "Ian" <[EMAIL PROTECTED]> > >Subject: Re: RAM Init > >Date: Tue, 19 Feb 2002 07:52:34 -0700 > > > > On Tue, 19 Feb 2002, Ian wrote: > > > > > I'm looking at RAM detection code here for at least three different types of >RAM, and I'm > > > simply not interested in it ... I don't have the space to do this properly. > > > > > > you can't just take the code and reduce it. We tried that once before. You > > have to really understand SDRAM startup and then write your own. > > > > I recommend you go line-by-line through the M1631 setup, and not write a > > single line of code until you know, for each line, what it does. > > > > ron > > > > > >
