> Hmm... so whats now?I've verified in sim that the crash in do_lookup() I was seeing (similar backtrace) was indeed caused by an interrupt coming between those two instructions and clobbering the saved GPRs. So I think it's very likely our problem. Ben. --
