Current-Users archive

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Old Index]

Re: rump missing cv_has_waiters?

On Thu Jun 05 2008 at 21:14:03 +0200, Matthias Drochner wrote:
> said:
> > /Volumes/scratch/tls/newnb/sys/ufs/ffs/ffs_alloc.c:307: undefined
> > reference to `cv_has_waiters' 
> Since the cv_has_waiters thing doesn't seem to be worth to waste
> emulation code on I've helped myself with the appended patch.
> I'm not sure whether it is good to pollute the kernel sources
> with rump specific conditionals either...

Please don't use #ifdef _RUMPKERNEL in the kernel sources.  It's _always_
the wrong thing to do (*).  I originally added _RUMPKERNEL because
I couldn't figure out a way to get the kauth_cred_t type defined
properly for non-NetBSD hosts in rump.h.  A nicer way is to introduce
similar functionality to pthread.  <rant>Userspace and kernelspace
sychnoronization interfaces should be unified anyway.</rant>

But if someone wants to fix the build, I'm not going to jump up and
down objecting.  It's doubtful I'll get around to looking at the proper
fix before next week.

  - antti

*) Ok, I broke the rule once in kern_lock.c, but that was because
lockmunger is supposed to be going away and I didn't feel like waging
war against it.  Actually, maybe rump no longer needs kern_lock.c at all?

Home | Main Index | Thread Index | Old Index