Re: Crash on -current on i386 (NOT amd64)

[John Nemeth <jnemeth%cue.bc.ca@localhost>]

On Dec 6,  6:36am, Paul Goyette wrote:
} On Sat, 5 Dec 2015, John Nemeth wrote:
} > On Dec 6,  6:12am, Paul Goyette wrote:
} > } On Sat, 5 Dec 2015, John Nemeth wrote:
} > }
} > } > } What's the backtrace? Ideally I'd like to remove the DIOCGPART code...
} > } >
} > } >     I don't think there was a backtrace. ...
} > }
} > } Correct.  I was unable to get a backtrace - ddb got a recursive fault
} > } when I tried to get the backtrace.
} >
} >     How about a dmesg then?  That would at least give some indication
} > of which driver was involved.
} 
} The problem did not occur during boot.
} 
} It happened long after sysinst had started.  We had already written
} new labels and boot blocks to the target root device, and were looking
} for the installation media.
} 
} The DIAGNOSTIC message and crash occurred as soon as I entered "a" for
} "CD-ROM".

     Which was probably the first time something tried to open the
CD-ROM.

} I was able to get gdb to disassemble the code at the %eip and this
} confirmed that the fault occurred at the point where your patch was
} fixing.

     The bug was in spec_vnopen().  Even though the crash didn't
happen during boot, a dmesg would still tell us what drivers were
involved.

}-- End of excerpt from Paul Goyette