On Saturday 10 June 2006 14:16, Dieter wrote:
> > may I suggest that the write protect pin be
> > connected to a header, so a jumper plug can disable it?  That way, it can
> > be manually protected against runaway software.
>
> If the PROM contents are bad due to runaway software or whatever reason,
> does the board become a doorstop, or is there a fail-safe method to
> reprogram it?

If the PCI interface FPGA PROM is corrupted, it looks like you'll want to plug 
a JTAG configuration cable (from a different computer) onto J1 & J2[*] and 
blast a new core onto the FPGA, which will get the board working.  Now, if 
the broken FPGA configuration in the PROM has hung the PCI bus on the 
development box, how do you reprogram the PROM without rebooting the system 
and thus reloading the corrupted image?  The way Timothy's explained it 
seemed to indicate that you can only program the PROM over the PCI bus, so I 
must have misinterpreted him. :P

Although I would have preferred it if they had used a Molex 87832-1420 or 
equivalent for the JTAG socket, I have a nice little ribbon cable from Xilinx 
that fits straight onto one of those!

But don't forget that there are actually two PROMs, one for the little FPGA 
and one for the big FPGA.  Normally you won't touch the small PROM, and just 
play around with the big one.  So write-protect headers do make some 
sense. :)

Peter

[*] I can't see any silkscreen layer on the current artwork, but hopefully all 
the J1 & J2 will have the pins labelled TMS, TDI, TDO, TCK, or much confusion 
may well result.  Also, the current artwork seems to show J1 & J2 in a block 
rather than in a line... if they're in a line it makes it a lot easier to 
interface a cable onto them IMHO.

P.S. KPDF handles these big complex PDF files far better than Acrobat Reader 
does. :D

-- 
Quake II build tools maintainer             http://tinyurl.com/fkldd

v2sw6YShw7$ln5pr6ck3ma8u6/8Lw3+2m0l7Ci6e4+8t4Eb8Aen5+6g6Pa2Xs5MSr5p4
  hackerkey.com

Attachment: pgptnY046AOBq.pgp
Description: PGP signature

_______________________________________________
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