nanogui: Thread: mipsel-linux-gcc keeps core dumping

Hello,

I keep this getting message when I get to the final part of my build.

I am trying to link in all the object files, but it spits out the
following message:

ld terminated with signal 11 [Segmentation fault], core dumped

Hints?

Rob

On Fri, 26 Jan 2001, Robert Hartley wrote:
> ld terminated with signal 11 [Segmentation fault], core dumped

A common cause of this when cross compiling is that you are trying to link
in some of the libraries from your host machine, and the cross linker does
not understand the format.

-- 
------- Alex Holden -------
http://www.linuxhacker.org/
 http://www.robogeeks.org/

Thanks for the tip.

The only libs I am linking in on the link line are for Nano X.

I stubbed out the functions I did not need to avoid linkning in the NanoX
library and I still get the same error.

My link line from 'make ' appears like this:

mipsel-linux-gcc obj/cache.o obj/class.o obj/fields.o obj/frame.o
obj/garbage.o obj/global.o obj/interpret.o obj/loader.o obj/main.o
obj/native.o obj/pool.o obj/thread.o obj/nativeCore.o obj/loaderFile.o
obj/nativeFunctionTableUnix.o obj/runtime_md.o obj/fakeStaticMemory.o
obj/hashtable.o obj/profiling.o obj/StartJVM.o obj/verifier.o obj/log.o
obj/jar.o obj/property.o obj/nativeSpotlet.o obj/nativeGraphics.o
obj/events.o obj/storage.o obj/storagePrim.o obj/network.o obj/networkPrim.o
obj/resource.o obj/ROMjavaUnix.o -o kvm


I ran mipsel-linux-strip on the object files.

Running 'file *.o' shows all of them with the following:
... ELF 32-bit LSB relocatable, MIPS R3000_BE - invalid byte order, version
1, stripped

Everything else is just an object file built with mipsel-linux-gcc

I tried the following inside my object file directory, but got the following
error message:

mipsel-linux-gcc \
 cache.o class.o fields.o frame.o garbage.o global.o \
 interpret.o loader.o main.o native.o pool.o thread.o \
 nativeCore.o loaderFile.o nativeFunctionTableUnix.o \
 runtime_md.o fakeStaticMemory.o  hashtable.o profiling.o \
 StartJVM.o verifier.o log.o jar.o property.o  \
 nativeSpotlet.o nativeGraphics.o events.o storage.o \
 storagePrim.o network.o  networkPrim.o resource.o \
 ROMjavaUnix.o \
 -o kvm
/opt/tvt/mipsel-linux/lib/crt1.o(.text+0x18): undefined reference to `main'
collect2: ld returned 1 exit status

As you can see, I am linking in no external libraries.

I have tried running the link step separate from gcc, but I get this error
message:

mipsel-linux-ld -L/opt/tvt/mipsel-linux/lib cache.o class.o fields.o frame.o
garbage.o global.o interpret.o loader.o main.o native.o pool.o thread.o
nativeCore.o loaderFile.o nativeFunctionTableUnix.o runtime_md.o
fakeStaticMemory.o hashtable.o profiling.o StartJVM.o verifier.o log.o jar.o
property.o nativeSpotlet.o nativeGraphics.o events.o storage.o storagePrim.o
network.o networkPrim.o resource.o ROMjavaUnix.o -o kvm -lc

mipsel-linux-ld: warning: cannot find entry symbol __start; defaulting to
0000000000412450
interpret.o: In function `Interpret':
../../../kvm/VmCommon/src/interpret.c(.text+0x288c): undefined reference to
`__divdi3'
../../../kvm/VmCommon/src/interpret.c(.text+0x2aec): undefined reference to
`__moddi3'
../../../kvm/VmCommon/src/interpret.c(.text+0x34d4): undefined reference to
`sitofp'
../../../kvm/VmCommon/src/interpret.c(.text+0x3504): undefined reference to
`litodp'
Segmentation fault


What am I doing wrong?
(This is driving me crazy...)


I am able to build an executable 'hello' program that run on the Helio using
the mipsel-linux-* tool chain, so I think they might be ok, at least if I am
using -static

Rob


Alex Holden wrote:

> On Fri, 26 Jan 2001, Robert Hartley wrote:
> > ld terminated with signal 11 [Segmentation fault], core dumped
>
> A common cause of this when cross compiling is that you are trying to link
> in some of the libraries from your host machine, and the cross linker does
> not understand the format.
>
> --
> ------- Alex Holden -------
> http://www.linuxhacker.org/
>  http://www.robogeeks.org/
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: ####@####.####
> For additional commands, e-mail: ####@####.####

--
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
=  Robert Hartley                 Mail:         201 Broadway        =
=  Central Region Systems Engineer              Cambridge, MA 02139 =
=  Integrated Computer            Email:        ####@####.####    =
=  Solutions, Inc.                Web Site:     www.ics.com         =
=  Tech Support: ####@####.####  Phone:        800-800-4271        =
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=

Visit the MotifZone (www.motifzone.org) for info on Motif!

On Sat, 27 Jan 2001, Robert Hartley wrote:
> I ran mipsel-linux-strip on the object files.

Don't do that- strip removes the symbols which are needed when linking the
objects together. You can strip the program after it's been linked
together.

-- 
------- Alex Holden -------
http://www.linuxhacker.org/
 http://www.robogeeks.org/

----- Original Message -----
From: Alex Holden ####@####.####
To: Robert Hartley ####@####.####
Cc: nanogui linuxhacker ####@####.####
Sent: Sunday, January 28, 2001 2:06 AM
Subject: Re: mipsel-linux-gcc keeps core dumping


> On Sat, 27 Jan 2001, Robert Hartley wrote:
> > I ran mipsel-linux-strip on the object files.
>
> Don't do that- strip removes the symbols which are needed when linking the
> objects together. You can strip the program after it's been linked
> together.
>
I doubt it. As far as I know, a striped library is available.
> --
> ------- Alex Holden -------
> http://www.linuxhacker.org/
>  http://www.robogeeks.org/
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: ####@####.####
> For additional commands, e-mail: ####@####.####
>
>

On Wed, 31 Jan 2001, Yu Xuanwei wrote:
> > Don't do that- strip removes the symbols which are needed when linking the
> > objects together. You can strip the program after it's been linked
> > together.
> I doubt it. As far as I know, a striped library is available.

[alex@hyperspace tea-total]# rm -f teatotal
[alex@hyperspace tea-total]# strip *.o
[alex@hyperspace tea-total]# make teatotal
gcc  -s -o teatotal teatotal.o teauntea.o tea-kgen.o randdev.o tead.o
teatun.o teash.o teacp.o teaprot.o huffman.o heap.o base64.o md5.o btea.o
readkey.o lineget.o util.o getarg.o
/usr/lib/crt1.o: In function `_start':
/usr/lib/crt1.o(.text+0x36): undefined reference to `main'
collect2: ld returned 1 exit status
make: *** [teatotal] Error 1
[alex@hyperspace tea-total]#

-- 
------- Alex Holden -------
http://www.linuxhacker.org/
 http://www.robogeeks.org/

----- Original Message -----
From: Alex Holden ####@####.####
To: Yu Xuanwei ####@####.####
Cc: nanogui linuxhacker ####@####.####
Sent: Wednesday, January 31, 2001 5:53 PM
Subject: Re: mipsel-linux-gcc keeps core dumping


> On Wed, 31 Jan 2001, Yu Xuanwei wrote:
> > > Don't do that- strip removes the symbols which are needed when linking
the
> > > objects together. You can strip the program after it's been linked
> > > together.
> > I doubt it. As far as I know, a striped library is available.
>
> [alex@hyperspace tea-total]# rm -f teatotal
> [alex@hyperspace tea-total]# strip *.o
> [alex@hyperspace tea-total]# make teatotal
> gcc  -s -o teatotal teatotal.o teauntea.o tea-kgen.o randdev.o tead.o
> teatun.o teash.o teacp.o teaprot.o huffman.o heap.o base64.o md5.o btea.o
> readkey.o lineget.o util.o getarg.o
> /usr/lib/crt1.o: In function `_start':
> /usr/lib/crt1.o(.text+0x36): undefined reference to `main'
> collect2: ld returned 1 exit status
> make: *** [teatotal] Error 1
> [alex@hyperspace tea-total]#
>

You are right. I misunderstood you. what I meant is that the library
binaries linked Ok can be striped. That won't affect the running a.out. Is
this right?

> --
> ------- Alex Holden -------
> http://www.linuxhacker.org/
>  http://www.robogeeks.org/
>
>