CONFIG_DESKTOP=n gives "request_module: runaway loop modprobe binfmt-464c"
arno.steffen at googlemail.com
Mon Jan 10 10:44:48 UTC 2011
Thanks for your hints. Indeed, it seems I did something wrong. I
modified the Makefile for using cross compiler instead of using
menuconfig to set it. Maybe that causes the trouble
Now it works with "full blown" or "embedded" - although the difference
in size is not so big anymore 896436 vs 884352 byte.
Indeed file is available not in my embedded system. But it doesn't
matter - I can run it on my PC as well.
setuid ELF 32-bit LSB executable, ARM, version 1 (SYSV), dynamically
linked (uses shared libs), stripped
for my once failed busybox it give
setuid ELF 32-bit LSB executable, Intel 80386, version 1 (SYSV),
dynamically linked (uses shared libs), stripped
So - thanks for help - and sorry. All my fault.
2011/1/10 Gilles Espinasse <g.esp at free.fr>:
> Selon Arno Steffen <arno.steffen at googlemail.com>:
>> Thanks for explaining this. I always thought my kernel supports ELF,
>> as I see this:
>> cat .config | grep ELF
>> # CONFIG_CORE_DUMP_DEFAULT_ELF_HEADERS is not set
>> and in menuconfig
>> BINFMT_ELF [=y]
>> ELF_CORE [=y]
>> CORE_DUMP_DEFAULT_ELF_HEADERS [=n]
>> BINFMT_ELF_FDPIC [=n]
>> BINFMT_ELF [=y]
>> COMPAT_BINFMT_ELF [=n]
>> For everything expect the last "COMPAT_BINFMT_ELF" I have a
>> description in menuconfig.
>> Is this really the problem ?
> Nice explanation from Denys for the 464 part of the error message.
> Looking in make menuconfig (even with ARCH=arm), you could only select
> BINFMT_ELF=y or not select, so that will never be a module.
> I suspect something is wrong in the compilation, probably on busybox side and
> simply with looking with 'file' should show what is wrong
> file <path>linux-<version>/arch/boot/vmlinux
> file <path>/bin/busybox
> I hope to get the right file in the multiples vmlinu* incarnation
More information about the busybox