Skip site navigation (1)Skip section navigation (2)
Date:      Mon, 18 May 2015 20:18:41 +0300
From:      Jukka Ukkonen <jau789@gmail.com>
To:        freebsd-ppc@freebsd.org
Subject:   Re: Is this a sign that nobody has tested 10.1-stable on a 32-bit ppc?
Message-ID:  <555A1EF1.6020904@gmail.com>
In-Reply-To: <555A0F2D.9050707@yahoo.com>
References:  <5559E371.9080602@gmail.com> <555A0F2D.9050707@yahoo.com>

next in thread | previous in thread | raw e-mail | index | archive | help
On 05/18/15 19:11, Benjamin Brink via freebsd-ppc wrote:
> On 5/18/15 6:04 AM, Jukka Ukkonen wrote:
>>
>> Find attached a trailing snippet of build time messages
>> while running buildkernel on a 32 bit ppc (PowerMac G4
>> Quicksilver).
>> During the last 2 months or so there have been some changes
>> to sys/dev/drm2/radeon/radeon_fence.c in 10.1-stable which
>> now break buildkernel. Previously I had rebuilt the whole
>> operating system on this ppc platform on Apr 23 and then
>> the build worked just fine.
>>
>> Now the whole oddity starts with an announcement...
>> "cc1: warnings being treated as errors"
>> and then begin warning messages about implicit declarations
>> of this and that atomic 64 bit functions and more warnings
>> about nested extern declarations of those same functions.
>> I guess that some of the code seen by the compiler was
>> never intended to be used on a 32 bit ppc system, but those
>> changes were never actually tested on a 32 bit ppc either.
>>
>> Has anyone else noticed this?
>>
> 
> FWIW, build failed locally on a new 10.1 32bit install for:
> 
> x11-wm/xfce4
> sysutils/screen
> 
> Am guessing compiler has issues with 32bit ppc when fed some 64bit code;
> I neglected to capture errors during build.
> 
> I bet making 64bit instruction sets back-compatible for a compiler
> building 32-bit ppc is a real hassle.
> 
> I'm willing to build again and capture errors if someone else wants to
> diagnose/fix.
> 
> ATM am powerpc-happy using getty/xterm;


So, the problem is not only in FreeBSD itself. At the moment
I am uncertain whether I should take that as good news or as
very bad news.
For some reason I am far more annoyed, though, by the fact
that I can no longer update my kernel and user space using
the latest "stable" source.
A few dysfunctional ports kits would be far easier to put
up with.
I can't help remembering my daughter's often heard comment
a few years ago when something went totally pear shape:
"Why me? Why me? Why is it always me?"

--jau




Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?555A1EF1.6020904>