On Sun, May 24, 2009 at 02:49:31PM +0300, Avi Kivity wrote: > Marcelo Tosatti wrote: >>> another one. This time the vm where booted sequentialy: >>> >>> [ 253.268993] kvm: 2907: cpu0 unimplemented perfctr wrmsr: >>> 0xc0010003 data 0x0 >>> [ 475.036542] rmap_remove: ffff8800cdb913b0 10 0->BUG >>> >> ^^^ >> >> So 0x10 is 1 bit different from 0x00, which would be the no present >> entry for AMD. >> >> Usually an indication of hardware problems. Perhaps you want to try >> the test Lucas mentioned. >> > > It's actually the PCD bit, not that it affects your analysis. Strange > that we see this pattern (1 bit differences on sptes) on both AMD and > Intel. Very suspicious. On Intel there was confirmation (through memtest86) that it was memory failure. The pattern was: 0xff7ffffffffff001 -- To unsubscribe from this list: send the line "unsubscribe kvm" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
