On Mon, Jan 20, 2020 at 09:28:32AM -0800, Paul Goyette wrote: > On Mon, 20 Jan 2020, Patrick Welche wrote: > > > On Mon, Jan 20, 2020 at 12:51:00PM +0000, Andrew Doran wrote: > > > This also happened the last time I touched rw_downgrade(), and I backed > > > out > > > the change then, but both times I don't see the bug. I have some > > > questions: > > > > > > - Are you running DIAGNOSTIC and/or LOCKDEBUG? I would be very interested > > > to see what happens with a LOCKDEBUG kernel here. > > Hmmm, at least on x86, in the LOCKDEBUG case we don't use the assembler > stubs; we simply use the C versions. > > On IRC/ICB, mlelstv has indicated there's something wrong in the stubs, > but I don't see it.
Yup, I think it's the stubs choking on the RW_NODEBUG flag being set. Testing a change for that now. Andrew > > > One worked with the addition of LOCKDEBUG. The other didn't, but it seems > > to be unrelated: > > Yeah, that backtrace looks unrelated. > > > +--------------------+--------------------------+-----------------------+ > | Paul Goyette | PGP Key fingerprint: | E-mail addresses: | > | (Retired) | FA29 0E3B 35AF E8AE 6651 | p...@whooppee.com | > | Software Developer | 0786 F758 55DE 53BA 7731 | pgoye...@netbsd.org | > +--------------------+--------------------------+-----------------------+