[EMAIL PROTECTED] writes:
> This reminded me of something I meant to ask since it's been a _long_
> time since I've read up on this stuff. Are user-space threads
> (under linux) incapable of using more than one processor on a SMP system?
Most user-space thread libraries I've seen are implemented in terms of
cute tricks around setjmp()/longjmp() (for example, older versions of
Netscape). Systems like this can't take advantage of SMP machines.
Of course, on a Linux box, clone() is available, but I think it's
pretty clear that if you use that you're getting out of userland.
> If a one user-space thread blocks, do the others block also?
That depends on which syscall you're blocking in, and how it deals
with signals, how you've got your signal disposition set up, etc.
Unless you do a bit of setup, the answer is "probably yes".
If you have decent kernel-threads, my inclination would be to use
them.
--kevin
--
Kevin D. Clark ([EMAIL PROTECTED]) |
Cetacean Networks, Inc. | Give me a decent UNIX
Portsmouth, N.H. (USA) | and I can move the world
[EMAIL PROTECTED] (PGP Key Available) |
**********************************************************
To unsubscribe from this list, send mail to
[EMAIL PROTECTED] with the following text in the
*body* (*not* the subject line) of the letter:
unsubscribe gnhlug
**********************************************************
