Re: [uml-devel] [PATCH 04/10] uml: make execvp safe for our usage

[Jeff Dike <jdike@addtoit.com>]

On Sat, Oct 21, 2006 at 02:11:28AM +0200, Blaisorblade wrote:
> > This is horriby ugly.
> 
> Detail why. The code of execvp()? Passing in the buffer?
> I'm not saying it's the brightest code around here, but it's ok for me.

My initial reaction was mostly due to the look of the code, which is
fixable.  I also don't like carrying around bits of libc (although we
do have setjmp/longjmp, but that's a special case).  However, it's
unlikely that it will need much maintenance, so this is more a taste
thing as well.

> I initially thought to design a two-steps API with a "which" operation (where
> memory allocation was used) to call later execvp(); when I saw the glibc 
> implementation (it allocates one single fixed-size buffer) I saw it was 
> simpler this way.

I think I still like the two-stage thing better.  If the 'which' part
finds something that doesn't exec, then we can just spit out a nice error.

> I'd not do that at boot, but just before the fork()+execve() - it is 
> conceivable that a given user will install a support binary after booting 
> UML.

I was envisioning it being part of bootup, but doing it just before
the exec would be OK, too.

				Jeff