All such transactions end with master abort, meaning that no device claimed the access.
Anyhow, what we see is that if we plug in some other graphics card, we see hundreds of thousands of accesses to this address, but then that stops happening, and it boots. With OGD1, we see millions of accesses, and the system never boots up, or boots with no video, or other things we can't explain yet. Maybe this I/O address is a red herring. On 4/29/09, Tom Sylla <[email protected]> wrote: > It is probably being used in the same way as port 0xED. Some BIOSes > use that address a a "IO delay" port (or possibly pseudo-serializing). > They use this port instead of 0x80 so they don't contaminate their > port 80 cards display. A write to this port will give them a ~1us > delay. See some information in this thread: > > http://lkml.indiana.edu/hypermail/linux/kernel/0203.1/0826.html > > > > On Wed, Apr 29, 2009 at 4:29 PM, Timothy Normand Miller > <[email protected]> wrote: > > We're having a boot problem with OGD1, and we see millions of accesses > > to I/O space address 0xEB. I'm not aware of that mapping to anything > > graphics-related, and we're not sure that what we're seeing on the bus > > isn't bogus. What would 0xEB map to? > > > > -- > > Timothy Normand Miller > > http://www.cse.ohio-state.edu/~millerti > > Open Graphics Project > > > _______________________________________________ > > Open-graphics mailing list > > [email protected] > > http://lists.duskglow.com/mailman/listinfo/open-graphics > > List service provided by Duskglow Consulting, LLC (www.duskglow.com) > > > -- Timothy Normand Miller http://www.cse.ohio-state.edu/~millerti Open Graphics Project _______________________________________________ Open-graphics mailing list [email protected] http://lists.duskglow.com/mailman/listinfo/open-graphics List service provided by Duskglow Consulting, LLC (www.duskglow.com)
