Re: 2.2: TaskID's for CLONE_PID? Proper signal handling?

Richard Gooch (rgooch@atnf.CSIRO.AU)
Thu, 1 Jan 1998 23:03:56 +1100

Ulrich Drepper writes:
> > Yes, We could probably rob one of the RT signals, and SIGUNUSED for
> > linuxthreads (on x86), and rob two of the RT signals for platforms that don't
> > have a SIGUNUSED available. I don't if there is a requirement on the number
> > of RT signals that are needed though.
> >
> > Below are the changes that I mean (I haven't tried them yet).
> >
> > Peeter
> >
> > diff --recursive -u --new-file linux.base/include/asm-alpha/signal.h linux/in\
> clude/asm-alpha/signal.h
> > --- linux.base/include/asm-alpha/signal.h Tue Dec 9 12:31:19 1997
> > +++ linux/include/asm-alpha/signal.h Tue Dec 30 01:53:21 1997
> > @@ -64,13 +64,15 @@
> > #define SIGINFO 29
> > #define SIGUSR1 30
> > #define SIGUSR2 31
> > +#define SIGTHD1 32
> > +#define SIGTHD2 33
>
> That's not necessary. glibc 2.1 will provide a easy and standard
> compliant method to solve this problem.

But it is necessary for those still using libc 5.

Regards,

Richard....