[prev in list] [next in list] [prev in thread] [next in thread] 

List:       linux-ecc
Subject:    Re: [ecc] Re: 875p patch
From:       Joseph V Moss <jmoss () ichips ! intel ! com>
Date:       2004-05-04 18:15:28
Message-ID: 15611.1083694528 () ichips ! intel ! com
[Download RAW message or body]

> Hi, one-man show today :-/
> 
> > > Slava,
> > > Will I have to add a line
> > > { 0x8086, 0x254c, probe_e7500, "e7501" },
> > > to the cs_table structure?
> > 
> > Seems to work. Now I found that the machine has 4GB of memory ('free'
> > gives
> > 
> > Mem:       4011116    3498056     513060          0        244 2182140)
> > 
> > and of course 4GB would not fit into an unsigned long anymore, so the
> > test 
> > 		if (last_mem>mem_end) {
> > in ecc_proc_output will fail for the last (4th) bank.
> > This looks like a serious problem :-((( 
> 
> Not if one finds all the places to modify. Which I cannot guarantee, but
> at least it works now with my 7501 :-)
> 
> > Would it make sense to do all the computations in Kbytes instead of
> > bytes? Would give us 10 more bits to work with... 4TB of real memory
> > won't be there before end of next year... What do you think?
> 
> Please look at the patch attached. Slava, could you probably look twice?
> It might make sense to have this in the next major version...

I solved this problem a different way in the code we're using at Intel.  I
switched all memory addresses to long longs, i.e. using 64 bits to store them,
instead of 32 bits.   We've got lots of 4G systems and some with more than that,
so we were seeing lots of mis-reports.



 
Yahoo! Groups Links

<*> To visit your group on the web, go to:
     http://groups.yahoo.com/group/ecc/

<*> To unsubscribe from this group, send an email to:
     ecc-unsubscribe@yahoogroups.com

<*> Your use of Yahoo! Groups is subject to:
     http://docs.yahoo.com/info/terms/
 

[prev in list] [next in list] [prev in thread] [next in thread]