On Tuesday, February 28, 2012 at 9:55 PM, Michael Bloom wrote: > That's great news. > > I'm still curious as to the cause of gdb not being able to give a useful stack > trace on your system.
The lack of a stack trace was due to the useful part of the stack being overwritten by the code in a eth_devices which was called by eth_getname. The code in eth_devices was passed a pointer to a stack array declared in eth_getname. The code in eth_devices wrote beyond the bounds of that array and corrupted some of the return context on the stack. > For future reference, when you type "disas eth_open" in gdb, was the first > line displayed "push %ebp"? > > I'm hoping it's not. If so, a compile time option could make it be > so. But if it already is, then something else is the cause and future > debugging on your platform might prove to be difficult. > > Michael L. Umbricht mike at umbricht.org wrote: > > > Looks like the new version is working now! I still need to boot VMS and > > configure the interface for DECnet, but no more segfaults. > > > > Thank you very much! > > > > -mikeu > > _______________________________________________ > Simh mailing list > [email protected] > http://mailman.trailing-edge.com/mailman/listinfo/simh _______________________________________________ Simh mailing list [email protected] http://mailman.trailing-edge.com/mailman/listinfo/simh
