This is the mail archive of the gcc-help@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]

Re: Signal 11 and RH6.2


Yes, I have done a kernel recompile. With egcs in rh6.2, I got signal 11; with

gcc 2.95.2 on Turbolinux, I compiled both 2.2.16 and 2.4.0-test2 kernels and I

am running both without a single problem. I reinstalled redhat couple of
times. Usually, sig11 only occured on kernel recompiles. However I recall
where I couldn't even compile ssh and was getting signal 11. GCC would not
compile using egcs and I got signal 11. However, if the hardware is at fault,
how do you determine which part and why is everything working flawlessly (as
far as I know at least) with gcc 2.95.2  and Turbo distro? Turbo, in my view,
is a complete redhat knockoff and seems to be almost identical to to a redhat
install (6.1). But turbo has gcc 2.95.2. I don't like this turbo distro
though, and would much rather run redhat but can't (for obvious reasons).

Am clueless here. What test can I run on my system to determine what would
cause signal 11 ???

I also ran OpenBSD before and didn't have any problems. Why would hardware
problems only show up now (in rh6.2) ? I would like to figure out if it is
indeed the compiler or maybe I do have a hardware issue. It seems if it is
hardware, it can be one of four things: CPU, Chipset, Memory, or motherboard
(some other component). CPU is not my first guess and I tried two different
sticks of memory, both of which are tested working memory. So I am left to
believe it's either a chipset problem or incompatibility of VIA MVP4 and linux

or linux egc 2.91.66 compiler. Anything I can do here? It's like a catch-22:
hardware points to software and software blames hardware.

Thanks for your help.

llewelly@dbritsch.dsl.xmission.com wrote:

> On Mon, 10 Jul 2000, Victor wrote:
>
> > Hi. I installed RH 6.2 on my machine (AMD K6-2 500, MVP4, Crucial ECC
> > 128Meg RAM, Maxtor HDD). The compilere version is egcs 2.91.66 Every
> > time I want to compile somthing I get an error in ccp or cpp and get a
> > Signal 11. I was told that signal 11 means either memory or hardware
> > problem. I changed memory sticks, still, same thing.
> >
> > I installed TurboLinux Server 6.0.2 (it has gcc 2.95.2) and I have
> > absolutely no problems.
>
> Have you done any large builds? (linux kernel size or bigger) Hardware
>   problems often do not show up for installs and small builds.
>
> >
> > As I see it, the problem seems to be the compiler. Anyone have similar
> > issues with EGCS on MVP4 or other via chipsets? What is the relationship
> > between egcs and gcc? Why does redhat use egcs instead of gcc? Is one
> > better? I get a feeling that the two were merged.
>
> Yes. egcs and gcc were merged to become the current gcc. See gcc.gnu.org
>   for more info.
>
> I believe redhat uses egcs-2.91.66 because gcc-2.95.2 uses a different ABI
>   for C++ programs, and redhat did not want an ABI breakage, even if it
>   only affected C++ programs.
>
> >
> > I want to run RH6.2 but how do I upgrade that compiler? When I wanted to
> > compile gcc2.95.2 I (obviously) got signal 11. How do you upgrade from
> > egcs to gcc? What I mean is how do I know that all the header files and
> > libraries are updated?
> >
> > Thanks for any info
> > Victor
>
> I did it by building gcc 2.95.2 with egcs-2.91.66 . If you get sig11 while
>   building gcc 2.95.2 with egcs-2.91.66, you probably have a hardware
>   problem.
>
> Note that gcc 2.95.x and egcs-2.91.66 have different C++ ABIs (due to an
>   MI bugfix, I belive). So you must ensure that all parts of a C++ program
>   are built with the same compiler. (this does not affect C.)


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]