On Tue, Nov 12, 2002 at 01:22:57PM -0000, atul srivastava wrote:
> Have anybody confirmed stability of binutils for big endian
> mips?
> while little endian gives me no problem
> but i am simply unable to make a working big endian ramdisk.
>
> specific problem is in identifying the e_type and e_machine fields
> in elf header of executables.
> expected value of e_type is 0x2(ET_EXEC) amd e_machine is
> 0x8(EM_MIPS)
> but those are read as 0x200 and 0x800 respectivly ..this is
> obviously the endianness problem.
Possibly a problem in your reader. Are you using od to look at the
header? Od is dumping in the byteorder of the machine it's running on.
> I have tried big endian ramdisk from
> ftp://ftp.ltc.com/pub/linux/mips/ramdisk/ramdisk
> and also tried compiling one using uclibc package , both
> failed because of same reason.
>
> when i run file command on those executables after mounting the
> ramdisk on my host i get o/p like
>
> ELF 32-bit LSB executable, MIPS R3000_BE - invalid byte order,
> version 1, statically linked, stripped
The error message you're seeing simply means byte 5 of the ELF file is
zero that is unset that is not specifying any endianess.
> though i see lot of mails regarding "invalid byteorder" but I
> didn't find specific suggestions.
It's a tool bug but a fairly harmless one.
Ralf
|