[dm-devel] dm panic debugging

Tim Burgess tim.burgess at anu.edu.au
Mon Feb 21 15:32:05 UTC 2005


Hi All,,

I'm trying to debug this panic that I emailed the list about the other 
day...  I've gotten a little way with KDB - just wondering if there's 
someone familiar with the internals of dm-raid1 that I can bounce 
things off as I find them (lest something I find jog inspiration with 
someone who actually knows what's going on!).

The crash occurs here in rh_dec(), called by mirror_end_io():

         if (atomic_dec_and_test(&reg->pending)) {

Because reg is a null pointer, returned by __rh_lookup()

That's as far as I have gotten so far :)

The crash is not complicated to reproduce but it does take quite a 
while doing heavy i/o before it crops up.  So it might be an hour or so 
before I can get back into my KDB session with the same problem (there 
is a deadlock problem as well triggered by the same steps).

Anyway - if someone is willing to lend a hand please let me know!!!

Cheers,
Tim

-- 
    tim.burgess at anu.edu.au         ANU Supercomputer Facility
    Phone: +61 2 6125 1431           and APAC National Facility
    Fax:   +61 2 6125 8199           Leonard Huxley Bldg (No. 56)
                                     Australian National University
                                     Canberra, ACT, 0200, Australia




More information about the dm-devel mailing list