[POSIX] implementation of pthread_create()

Carl Flippin carlf at photocarl.org
Fri Mar 6 00:15:45 UTC 2009

overflow_ <overflow_ at libero.it> writes:

> overflow_ wrote:
> Thanks for the help, I think I got it, even if I have still something
> not completely clear.
> - in the do_clone function there ISN'T ANY CALL TO the clone
> function!!!!!!!  there is just a INTERNAL_SYSCALL call.
> it uses just a macro ARCH_CLONE, that call the function __clone, but I
> dind't find where is it implemented I think its implementation changes
> with the hardware, but I didn't find where is it! Is there someone
> that know it?

This is expected. Ultimately, most of the glibc functions end up calling
a linux syscall. That's how glibc gets linux to actually do something. I
would expect the final implementation to make a syscall since all the
threading functions ultimately rely on the kernel. You will find this to
be true in most cases.

|Carl Flippin         | If debugging is the process of removing |
|carlf at photocarl.org  | software bugs, programming must be the  |
|http://photocarl.org | process of putting them in.             |

More information about the ubuntu-users mailing list