Very Computer

Hello all,

I am trying to compile Linux 1.2.1 kernel (I know it's old, but it's
on the CD) and keep getting the following error message on make zdisk
and make zImage, after make config, make dep and make clean.

        gcc: Internal compiler error: progam cc1 got fatal signal 11

What does this mean? Using Infomagic March 1995 CD with Slakware 2.X
(kernel 1.2.1). gcc 2.6.3 486DX-33 w/ 8megs ram. Nothing else running.
I know of a possible intermittant memory problem. Could this be it?

After this error, make zdisk or make zImage gets further through
compilation but still dies on me.

Help and comments requested.
Regards,

Jeff Brown                                 __________
Associate Network Engineer                |   The    |
The Home Depot, USA, Inc.      ___________|Home Depot|__________
2727 Paces Ferry Rd           |###########|==========|##########|
Atlanta, GA 30339             |===========|    __    |==========|

Check /usr/include/linux/errno.h.  I believe that it's a segmentation
fault.

Steve

Quote:>Hello all,

>I am trying to compile Linux 1.2.1 kernel (I know it's old, but it's
>on the CD) and keep getting the following error message on make zdisk
>and make zImage, after make config, make dep and make clean.

>    gcc: Internal compiler error: progam cc1 got fatal signal 11

>What does this mean? Using Infomagic March 1995 CD with Slakware 2.X
>(kernel 1.2.1). gcc 2.6.3 486DX-33 w/ 8megs ram. Nothing else running.
>I know of a possible intermittant memory problem. Could this be it?
>After this error, make zdisk or make zImage gets further through
>compilation but still dies on me.

    I get a similar error when I compile anything.  My compiler seems to error
out at random intervals.  There is no real explanation because I can compile
the same file without a problem.  If I do a make clean, then start the compile
operation again, it might error on something different next time.

    I have found that I just need to either recompile the file that was being
compiled when it failed or delete the .o file that was created.  When I do
this, eventually I make it all the way through the compile process and the
executable is fine.

Myron

Quote:>    gcc: Internal compiler error: progam cc1 got fatal signal 11
>I know of a possible intermittant memory problem. Could this be it?

Most likely. Signal 11 seems to be what most people get when gcc
barfs on some bad memory.

(Now don't you wish you had ECC memory? :)

--

A bachelor is a selfish, undeserving guy who has cheated some woman out
of a divorce.
                -- Don Quinn

Bad memory, overclocked processor, bus speed too fast, bad cache chips,
overheated processor.  Sig11 errors are almost always hardware related.  In
my case it was the overclocked processor, in other cases it's been the other
things.

Check http://einstein.et.tudelft.nl:80/~wolff/sig11/ for a good summary of
what happens.

Skunky

--
-------------------------+-------------------------------------------
Jeff W. (Skunk) Schouten | Someday, when all the fighting ends we'll
PFC, USMC (Ret.)         | all realize just how stupid the things we


(Internet RT Asst. Sysop)| (I'm not a roadkill......yet!)

That's a segmentation fault.  I get those every once in a while too
but don't know why.  They're much less frequent with gcc 2.7.0 than
they were with 2.5.8.

Steve

: >Hello all,
: >
: >I am trying to compile Linux 1.2.1 kernel (I know it's old, but it's
: >on the CD) and keep getting the following error message on make zdisk
: >and make zImage, after make config, make dep and make clean.
: >
: >  gcc: Internal compiler error: progam cc1 got fatal signal 11
: >
: >What does this mean? Using Infomagic March 1995 CD with Slakware 2.X
: >(kernel 1.2.1). gcc 2.6.3 486DX-33 w/ 8megs ram. Nothing else running.
: >I know of a possible intermittant memory problem. Could this be it?
: >After this error, make zdisk or make zImage gets further through
: >compilation but still dies on me.

:     I get a similar error when I compile anything.  My compiler seems to error
: out at random intervals.  There is no real explanation because I can compile
: the same file without a problem.  If I do a make clean, then start the compile
: operation again, it might error on something different next time.

:     I have found that I just need to either recompile the file that was being
: compiled when it failed or delete the .o file that was created.  When I do
: this, eventually I make it all the way through the compile process and the
: executable is fine.

: Myron

My experince with sig 11......
I disabled cache ram (386-33 dx) and 100% compile sucess
cache ram enabled 90-100% failure rate...

FWIW....
--
The wealth of reality, cannot be seen from your locality.

Hi,

  Steve's right. Segmentation errors are usually memory related. If you
can add a wait state in your BIOS (slow memory access down a smidge) then
it might help.... I had to go back to my memory vendor and physically
swap out SIMMs.... put in some NEC 8Mb SIMMs and it solved my problem.....
Meanwhile, if you note the directory that make was working in when it
segfaulted, then you can cd to that directory and rm *.o and then go back
and do a make again, it'll eventually build a kernel...... Pretty painful
though.....
*************************************************************************

Schlumberger Information Network   *   Telephone:     1-713-275-7792
%Schlumberger Limited              *   Fax:           1-713-240-7204
205 Industrial Boulevard           *      
Sugar Land, Texas  77478           *        
USA                                *
************************************************************************

Quote:>>>>> Skunk Schouten writes:

SS> Bad memory, overclocked processor, bus speed too fast, bad cache chips,
SS> overheated processor.
[ ... ]

SS> http://einstein.et.tudelft.nl:80/~wolff/sig11/ for a good summary
SS> of what happens.

I can also report at least one (mine !) case of a ill-behaved EIDE hard
disk.
I use an ASUS p55tp4xe with a P133 and 256K standard PBurst (think it's
15ns but not checked) ... with 70ns SIMMs (albeit ASUS recommends 60ns
ones) and two 1G EIDE (WD Caviar 21) drives for RAID0 ("md", very good,
recommended !)

It gave very numerous flaws (gcc compilation was impossible) => I disabled
the PBurst cache and all went good.

But one of the EIDE drives was in fact bad (did not dynamically remap bad
sectors).
I have it changed and the entire thing now works no flawlessly, even
with PBurst on.

Compilations (kernel and gcc) and "memtest" does not reveal any more
problem in an entire afternoon.

I do not know the explanation. Perhaps bus-timings related problems, since
IDE-family drives 'controlers' are bus extensions (?)

But the box is more responsive. The Byte bench performance index is now
15.2, it's a 39% boost (thanks to the RAID0 and PBurst cache) !

[ Followup-To: comp.os.linux.hardware ]