:
:
:Hi Matt/All,
:
:Okay, so far so good... with the (properly applied) patch that
:Matt posted my system is once again stable..
:
:Thanks again for this Matt! And apologies to all for my
:earlier brain-cramp while applying it !!!
:
:Andrew.
Nah, don't worry about it. I actually think that your original
issue might have been a combination of the procfs AND the linux pathing
code, whereas the original patch I sent you just fixed the linux
pathing code.
What I think I am going to do now is commit both pieces and let the
chips fall where they may. If people start getting linux emulation
related crashes please report them and try to provide a kernel core
and/or gdb backtrace and I will stomp them flat.
Changing over the VFS locking from fake locks to real locks created
a bit more instability then I would have liked, though it wasn't as bad
as my original FP code (which caused actual FS corruption). I will
try to do better in the next VFS commit (in which I will attempt to
change one VFS call to pass a namecache pointer instead of a directory
vnode). And after that, namespace locking for directory operations
and wholesale replacement of dvp with namecache in VOP calls.
-Matt
Matthew Dillon
<dillon@backplane.com>