Skip site navigation (1)Skip section navigation (2)
Date:      Thu, 10 May 2012 10:19:03 +0300
From:      Andriy Gapon <avg@FreeBSD.org>
To:        Alfred Bartsch <bartsch@dssgmbh.de>
Cc:        freebsd-stable@FreeBSD.org
Subject:   Re: FreeBSD 8 i386 gptboot corrupt - SOLVED
Message-ID:  <4FAB6BE7.9060500@FreeBSD.org>
In-Reply-To: <4FAB6A7B.9050500@dssgmbh.de>
References:  <4FAA3912.3030801@dssgmbh.de> <4FAA4A11.808@FreeBSD.org> <4FAA5E70.7030508@dssgmbh.de> <4FAA83BD.2030204@FreeBSD.org> <4FAB6A7B.9050500@dssgmbh.de>

next in thread | previous in thread | raw e-mail | index | archive | help
on 10/05/2012 10:12 Alfred Bartsch said the following:
> I got this stupid idea of a "16k limit" during testing. It was unobvious to
> me that the build process in a standard environment (i386) simply produces
> invalid code. In i386 (32-bit) hardware, we don't use zfs at all, so I
> can't tell anything about gptzfsboot. For now, modifying
> /sys/boot/i386/gptboot/Makefile completely solves this actual build
> problem.
> 
> IMHO the compiler should always know perfectly well in which hardware 
> environment it runs and for which target environment it produces code. So
> the build environment should be modified to fix this. I would certainly
> give it a try, but unfortunately this is far beyond my knowledge. :-(

That's an interesting theory.
What kind of hardware do you have?  Is it something non-mainstream or
sufficiently old?
As far as I can tell, our base GCC uses i686 target arch if none is explicitly
requested.

-- 
Andriy Gapon



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?4FAB6BE7.9060500>