Re: [PATCH] Re: sharing mutex between libc stub and libpthread

[Martin Husemann <martin%duskware.de@localhost>]

On Thu, Feb 21, 2013 at 06:37:52AM +0100, Emmanuel Dreyfus wrote:
> The advantage of this approach is that it is not an intrusive change and
> it should not have a big performance impact. On the negative side, it
> just cannot support calling pthread_create(), but I now beleive this
> cannot be addresses without moving almost all libpthread into libc. This
> could be a goal for netbsd-7 if it is decided desirable, but the present
> patch would be a nice fix for netbsd-6, IMO.

Ok, this has the nice property of only punishing threaded apps.

For moving libpthread into libc as a goal for -7: I still strongly like to
see good performance figures first, my guess is still that it will be a
no-go.

But, I'm sorry, my main problem is still that I just can not see how the 
new behaviour you are proposing (die on first pthread_create() call)
is better than the what we had before we added the noload flag --
and IMHO the much better aproach would be Joerg's fix to glib2 plus
fixing PAM to output proper diagnostic on dlopen() failures (which
needs to be fixed anyway).

Martin