On Fri, 21 May 1999, Matthew Dillon wrote: > :> entirely contained within the current stack trace. > :All my kernels are now DDB kernels :) But since I do almost all of > :my work remotely they are DDB_UNATTENDED, and the machine I am panic-ing > :is not on the serial console server (sorry). I do have another question > :about DDB, I unstalled -STABLE as of today (from releng3.fre...) and I > :compiled the kernel with DDB, and DDB_UNATTENDED per usual. Now when I > :C-A-E to get into the debugger and type 'panic' it drops me at another > :debugging prompt. If I type panic from that I get the real thing, any ideas? > : > :My next email will hopefully have the stack trace for this panic. > :-- > :David Cross | email: cro...@cs.rpi.edu > > Panic has a counter. The first time it is called it tries to drop into > the > debugger. The second time it is called it reboots the machine for real. > When you control-alt-escape, you have not yet called panic for the > first time, so when you panic manually from the DDB prompt it drops you > into the debugger again. Second time's the charm! > > Since the debugger repeats the previous command if you just > hit return, I've gotten used to simply typing > 'panic <return><return><return><return>...'
I use that too :-). For the alpha, I put in a 'halt' command (also linked to remote-gdb's kill operation) which drops the machine back to the firmware which is handy if you don't care about buffers not being synced. -- Doug Rabson Mail: d...@nlsystems.com Nonlinear Systems Ltd. Phone: +44 181 442 9037 To Unsubscribe: send mail to majord...@freebsd.org with "unsubscribe freebsd-hackers" in the body of the message