From: Phil F. <da...@xg...> - 2001-09-28 06:09:19
|
I guess it was a bochs bug; I got a new bochs and it seems to work just fine with the same image, yet the bochs i had continues to fail. On Fri, Sep 28, 2001 at 12:45:00AM -0500, Phil Frost wrote: > Late this morning i was working on hydrogen and started getting the mystery > bug again :( I'm using the latest bochs as of yesterday; i have not yet tried > the updates that were commited today. A tarball of the buggy code is at > http://xgs.dhs.org/uuu-xgs-yttrium-bug.tar.bz2. > > Execution seem to go wrong at EIP = 0x97A4; the byte at that offset is 0x11 > but it's something else when bochs gets to it; thus the jz jmps to the wrong > offset. I didn't get much of a chance to look at it but i tried putting a > write watchpoint on this location; it only caught the expected write when > the core is loaded from disk. I'm wondering if maybe this is that bochs bug; > the trace-on behaviour is diffrent from the trace-off. |