2003-09-22 03:07:30

by Albert Cahalan

Subject: CLONE_SIGHAND w/o CLONE_VM

Does CLONE_SIGHAND without CLONE_VM ever
make sense?

Note that the arch-specific kernel_thread()
implementations add CLONE_VM, so kernel_thread()
usage doesn't count unless you can point to an
arch that doesn't add the CLONE_VM flag. (BTW, the
user-mode port is missing CLONE_UNTRACED. Bug?)

2003-09-22 03:41:50

by Daniel Jacobowitz

Subject: Re: CLONE_SIGHAND w/o CLONE_VM

On Sun, Sep 21, 2003 at 10:54:05PM -0400, Albert Cahalan wrote:
> Does CLONE_SIGHAND without CLONE_VM ever
> make sense?
>
> Note that the arch-specific kernel_thread()
> implementations add CLONE_VM, so kernel_thread()
> usage doesn't count unless you can point to an
> arch that doesn't add the CLONE_VM flag. (BTW, the
> user-mode port is missing CLONE_UNTRACED. Bug?)

Minor bug, but yes, it's a bug. kernel threads should always be
CLONE_UNTRACED; UML wasn't in the tree when I added it.

--
Daniel Jacobowitz
MontaVista Software Debian GNU/Linux Developer