Fighting gcc 3.2 again.

Bill Maltby - LFS Related lfsbill at wlmcs.com
Tue Oct 29 05:30:50 PST 2002


On Tue, 29 Oct 2002, Kaiyi Lin wrote:

>One question. If it is a CPU heat problem the PC should hang, right?
>Why just a compilation error?

Please remember I *know* nothing about hardware, so some more
knowledgeable people on the list would be better sources to answer your
question. But having said that... I will fill the room with hot air now.

:)

You are beyond my expertise with that question. But, from what little
understanding I have, the failure modes for electronic components are not
only one, like locking up. AFAIK, as the boundaries of reliabe operation
are approached, various types of "breakdown" in individual transistor
operations can occur. In not sure if this is a complete failure of the
node, or sporadic correct operation or what. Regardless, depending on
which transistors fail, and I *guess* how they fail, the result could be
other than lock up.

On very old equipment, I have seen invalid data results, but everything
looked like it was working OK. It was tough to find those errors. Didn't
know if it was disk, memory, software or what. And when we turn the
machine off and open it, everything cools down. :-\

Keep in mind, I only *think* it may be heat related. If you have good
indications that it is something else, you should investigate that because
you are on site and have access to the units.

One last thing. If memory serves, there is some "aging" of electronic
components that occurs over time. IIRC, it was mentioned for memory. From
what I remember, that could cause perfectly good components to degrade
just enough so that (for example) when the room was cool, all worked OK
and when the room warmed up just a few degrees, these components became
unreliable.

Please remember I *know* nothing about hardware, so some more
knowledgeable people on the list would be better sources to answer your
question.

>
>Bill Maltby - LFS Related wrote:
>
>> On Mon, 28 Oct 2002, kaiyi lin wrote:
>> 
>>>Na, just complete chapter5 and starting chapter6.
>>>After lucky feeling of long glibc compiling with great expectation
>>>to finish gcc 3.2 installation in chapter 6 today. But it seems that
>>>compiling gcc is the biggest problem in my LFS I got the following
>>>segmentation errors
>> 
>> Did you remember to apply the patch? I'm not sure if this is related or
>> not...
>> 
>>>
>>>./xgcc -B./ -B/usr/i586-pc-linux-gnu/bin/ -isystem
>>>/usr/i586-pc-linux-gnu/include -isystem
>>>/usr/i586-pc-linux-gnu/sys-include -O2  -DIN_GCC    -W -Wall
>>>-Wwrite-strings -Wstrict-prototypes -Wmissing-prototypes -isystem
>>>./include  -fPIC -g -DHAVE_GTHR_DEFAULT -DIN_LIBGCC2
>>>-D__GCC_FLOAT_NOT_NEEDED  -I. -I. -I../../gcc-3.2/gcc
>>>-I../../gcc-3.2/gcc/. -I../../gcc-3.2/gcc/config
>>>-I../../gcc-3.2/gcc/../include  -DL_lshrdi3 -c
>>>../../gcc-3.2/gcc/libgcc2.c -o libgcc/./_lshrdi3.o
>>>../../gcc-3.2/gcc/libgcc2.c: In function `__lshrdi3':
>>>../../gcc-3.2/gcc/libgcc2.c:266: internal error: Segmentation fault
>>>Please submit a full bug report,
>>>with preprocessed source if appropriate.
>>>See <URL:http://www.gnu.org/software/gcc/bugs.html> for instructions.
>>>make[3]: *** [libgcc/./_lshrdi3.o] Error 1
>>>make[3]: Leaving directory `/static/src/gcc-build/gcc'
>>>make[2]: *** [libgcc.a] Error 2
>>>make[2]: Leaving directory `/static/src/gcc-build/gcc'
>>>make[1]: *** [stage2_build] Error 2
>>>make[1]: Leaving directory `/static/src/gcc-build/gcc'
>>>make: *** [bootstrap] Error 2
>>>
>>>I have netscape running at the same time and one xterm windows while
>>>compiling.
>>>
>>>Running K6-2 550 286M RAM
>> 
>> If you just type the same make command it was running, it should pick up
>> with the correct module and continue. It may fail again, but see if it is
>> in a later module. If so, keep doing this and it will usually finish
>> eventually. This works if the fault is related to heat, "flaky" hardware
>> and so on. I had to do this on a 32MB AMD 5x86-133 one time. It worked.
>> Later there were a couple of other packages that also requeired this
>> procedure to make it through.
>> 
>
>

-- 
Bill Maltby
billm at wlmcs.com

-- 
Unsubscribe: send email to listar at linuxfromscratch.org
and put 'unsubscribe lfs-support' in the subject header of the message



More information about the lfs-support mailing list