[Lcdproc] lis driver fails to compile

[Rene Wagner]

On Sat, 2007-11-03 at 00:10 -0500, Daryl F wrote:
> Ah, I was just an innocent babe in the woods. Then out from behind a
> tree jumped the _Threads_Wolf_. After reading about threading and it's
> evolution in Un*x variants, esp. http://www.kegel.com/c10k.html it
> seems that I can't have just one implementation for all.
>  
> If libpthread is used it seems it will eventually be obsolete in Linux
> because of the progress on NPTL. And if I use NPTL only, it is
> unclear when, or if, FreeBSD will support it.

As far as Linux is concerned, you can use the POSIX threads interface
provided by libpthread with both the old LinuxThreads kernel/glibc
implementation and the new NPTL implementation doing the actual work for
you. In fact, LinuxThreads is now obsolete on most architectures and
distributions ship with NPTL nowadays.

So, on Linux there's no need to worry about pthread_create disappearing
or not working.

FreeBSD support is a different story. NPTL being Linux specific
kernel/glibc code doesn't make a lot of sense in FreeBSD (except 
perhaps for a binary compatibility layer). The Linux and FreeBSD
schedulers are both unique in their own ways, so a direct port seems
unlikely (and undesired due to different licensing).

I'm moderately sure though that there is POSIX threads support in
FreeBSD (regardless of the implementation behind it). If the
pthread_create variant compiles on FreeBSD, that will probably 
be good enough for now since the code can only be tested once there are
FreeBSD users using the corresponding hardware anyway.

Differences in OS implementation details are indeed evil, but usually
POSIX is there to protect you from the need to worry about them :)

Regards,

Rene