Subject: Re: Why no /usr/include/sys/kthread.h?
To: Matthew Orgass <email@example.com>
From: Jason R Thorpe <firstname.lastname@example.org>
Date: 06/05/2000 16:38:48
On Mon, Jun 05, 2000 at 07:18:09PM -0400, Matthew Orgass wrote:
> I hope this makes it in 1.5. I think it would be a good idea to also
> implement rfork, though. The non-RFPROC case could be used to detatch a
> thread's descriptors after a clone or rfork and the RFPROC case (also as
> sfork, which is the share everything possible equivalent of fork) is
> rather convenient for direct use in some situations where you do not
> really need a thread library and would otherwise just use fork.
I have a hard time calling something rfork(2) if it doens't work like the
original rfork(2) (as seen in Plan 9). IMO, the FreeBSD people choose a
poor name for thier __clone()-like system call.
> It would also be nice to have the MAP_STACK mmap flag as in FreeBSD (but
> specify that the created stack always goes the right direction), though
> that would probably be more work.
I see little value in MAP_STACK. The reference model for threads (Solaris)
doesn't need it. I don't see why we should. If you want it to avoid pre-
allocating swap space for stack pages, then again we can follow Solaris's
lead and implement MAP_NORESERVE.
-- Jason R. Thorpe <email@example.com>