Timothy Normand Miller wrote: > > Is it worth checking the project files in? > > If not, we just need to understand what each of the files are and > collect them together.
Good plan. > Also, we should not host them in the SVN repo, because the project > file would be a non-ascii blob. We can put them on Traversal's web > site or something. I disagree with this. Please do put them in svn. Subversion has no problem with handling binaries. If the files are not automatically recognized as binary, just make sure to: svn add -kb subdir/* to let svn know that they are binary. It is extremely valuable to have every single file related to the project available in the repo. Especially files like these which may not be easily reproducable by a large part of the open source community. > > The image in the coreboot ROM for the OGD1 is a small serial > > monitor program that accepts a couple of 1 character commands. > > This has the effect than when I boot the victim PC I get a wait > > prompt when the OGD1 BIOS gets called. From this prompt I can > > then insert a real OGD1 BIOS and run that. > > This is very cool. Over in coreboot land we're also excited about SerialICE. :) There's a web site for it at http://serialice.org/ SerialICE uses Qemu to run a binary, but will forward IO and memory accesses to real hardware via serial, and produce a nice log. > I was worried about you and others not having a sufficient debug > environment. It's also good that you can do more extensive > debugging on your end so that we can leave it as last resort to ask > Howard to use the PCI bus analyzer. SerialICE can of course not help with PCI bus level analysis, but can be very useful for software debugging. > > The advantage of this setup is that the time from recompiling the > > BIOS to actually testing it is less than 20 seconds, This actually seems like fairly long, to me. ;) > > It should be possible to get a similar set-up working with the > > stock BIOS. The only snag is that the stock BIOS probably hasn't > > set-up the superIO at the time the VGA ROM gets called, so the > > serial port won't work by default. Are you sure? I think it might be. In any case, it might be possible to masquerade as a non-VGA type option ROM, which means not getting called until very late in the standard BIOS execution flow. > > If people fancy trying this then they should try to find a > > motherbaord that is supported by coreboot, and preferably one > > that is supported by serialice. Sadly coreboot doesn't support any easily available boards at the moment. We are waiting for AMD to release their code for the latest generation chipsets since many months now. :\ But several slightly older boards are well supported, and might still be possible to find used on eBay or similar. > > Maybe this is a bit off topic? > > No. It's totally on-topic. People need to be able to do this kind > of debugging, whether it's with OGD1 or something else. > Documenting your setup so people can duplicate it would be really > helpful to everyone. If you're using coreboot and SerialICE I think it would be fantastic to have the info in the coreboot wiki! //Peter _______________________________________________ Open-graphics mailing list [email protected] http://lists.duskglow.com/mailman/listinfo/open-graphics List service provided by Duskglow Consulting, LLC (www.duskglow.com)
