> Randy McMurchy wrote:
>> So, bottom line for me is I don't see any errors other than the
>> math/test-double error in a 32bit multilib build. Both the 64bit
>> multilib and regular LFS (not JH branch) don't get any errors.
>
> Not really the bottom line (yet). Dennis is working on a PowerPC arch.
> Or, at least, the last I checked he was.
Yes Sir. Same machine and I am still working through the JH branch of
the LFS process. I did take a break from it for a few days but in
general the unit remains powered on and connected via serial console
cable.
[EMAIL PROTECTED]:~$ cat /proc/cpuinfo
processor : 0
cpu : G2_LE
clock : 396.000000MHz
revision : 1.4 (pvr 8082 2014)
bogomips : 65.53
timebase : 33000000
platform : Efika
machine : EFIKA5K2 CHRP PowerPC System
revision : 2B3
vendor : bplan
[EMAIL PROTECTED]:~$ uptime
2:22pm up 18 days 2:59, 2 users, load average: 1.25, 1.14, 1.10
[EMAIL PROTECTED]:~$
Right now it is working on GCC in Chap 6.
> I know that there are different expectations for what the tests will
> produce on PowerPC, but as to what is acceptable, I haven't had the time
> to investigate since the original test I ran several months back. Ken do
> you have any test data for PowerPC on a recent build?
I think that there is a lack of data here. I could begin to hunt down the
long double issues and perhaps there is a problem with the IEEE754 long
double implementation. A signaling NaN or some such that is not trapped and
handled correctly. Tough to say and it would take *considerable* work to
find. Since I am not about to break out Metrowerks Codewarrior JTAG debugger
( Freescale? ) to single step through this and watch for FPU activity I am
going to face east , pray for forgiveness over sins committed and install
with fingers crossed.
:-)
However, this is glibc and not my favorite video game. It deserves more than
a nod and thus I'll drop a note to David Woodhouse[1] ( whom I have spoken
with before ) and see if perhaps some PPC intelligence can be brought to
bear on the issue. It wouldn't be the first time that people managed to
track down the offense and then pushed a patch upstream.
Dennis
[1] and he does have a flair for the PowerPC
see : http://lkml.org/lkml/2007/5/24/127
--
http://linuxfromscratch.org/mailman/listinfo/lfs-dev
FAQ: http://www.linuxfromscratch.org/faq/
Unsubscribe: See the above information page
