Re: gcc done, eglibc problem

[Thorsten Glaser <tg@mirbsd.de>]

Finn Thain dixit:

>It doesn't matter, though, because with libc_MIN_KERNEL_SUPPORTED we 
>aren't changing dependencies. Setting --enable-kernel=2.6.16 simply makes 
>the libc more backwards compatible (to whatever extent that it can work 
>without TLS).

Ah, okay. This might actually be not a bad thing.

>it is a separate issue in that it doesn't relate to the m68k.mk changes 

Okay.

>I think this patch is what you are looking for.

Sure, this looks like it. I’ll implement that then. Thanks!

>The situation isn't ideal, because libc can still be installed in the 
>absence of TLS support, given a mainline kernel between 2.6.32 and .33.y.

Yes, I’ll just assume people use Debian provided kernels though,
everything else is SOL (even if “traditionally” m68k kernels have
been built outside of packaging, or so I read recently), since we
cannot cope with potentially everything. (We have backported TLS
in linux, gcc, eglibc – all three of them – so some assumptions
need to be done.)

Sorry for the confusion, and thanks a lot again for your help!
I’ll continue compiling.

//mira“still aching a bit”bilos
-- 
I believe no one can invent an algorithm. One just happens to hit upon it
when God enlightens him. Or only God invents algorithms, we merely copy them.
If you don't believe in God, just consider God as Nature if you won't deny
existence.		-- Coywolf Qi Hunt