On Thu, 30 Jun 2005, David Daney wrote: > It seems that it is a memory consistancy problem of some sort. By > placing wbflush() after all writes to NIC registers it works. This > leads me to think that either the driver is buggy WRT processors that > have write-back queues or my implementation (the default implementation) > of writeb() and friends is buggy on this processor. Now it could be > that all that is needed is wmb() before some of the register writes, but > since on my processor they both do the same thing (sync) it is hard to > tell. Most likely that code has only been ever used on i386 systems (who'd want to use such a weird Ethernet chip elsewhere?), so don't expect it to be terribly sane. Maciej
