>
> On Sun, 18 Aug 1996, Bryn Paul Arnold Jones wrote:
>
> > You don't seem to understand, fork() _IS_ clone(0), not equilivent to, not
> > like, but IS. I don't see any point in doing fork() based threads now,
> > and clone based later, just get the clone stuff done. We could (should
> > IMHO) remove fork(2), and implement fork(3).
>
> alright.. howabout getting a clone(0,COPYVM|SIGCLD) threads lib going
> first and then finish up a clone(astackptr,SIGCLD) threads lib down the
> road..
>
Erm, I think you mean clone(CLONE_VM|CLONE_SIGHAND), and the second one
seems like a step back to me .... The flags are in <linux/sched.h> ;),
shouldn't these be moved out, or at least the kernel specfic bits moved
inside "#ifdef __kernel__"'s (anyone) ?
There is a clone based threads package avalable, which proberbly need
some debuging. Ie get, test, report bugs, ..., test, perfect -> hay presto,
one clone based threads lib ...
Bryn
--
PGP key pass phrase forgotten, \ Overload -- core meltdown sequence
again :( and I don't care ;) | initiated.
/ This space is intentionally left
| blank, apart from this text ;-)
\____________________________________