This is the same script that worked fine when we were using
egcs-mips-linux-1.0.3a and binutils-mips-linux-2.8.1. Have
there been any changes in the linker that would affect how
we write our linker script?
Thanks,
-John
-----Original Message-----
From: Maciej W. Rozycki [mailto:macro@ds2.pg.gda.pl]
Sent: Thursday, January 04, 2001 1:18 PM
To: John Van Horne
Cc: 'linux-mips@oss.sgi.com'
Subject: Re: objcopy error -- was RE: your mail
On Thu, 4 Jan 2001, John Van Horne wrote:
> [jvhorne@guava-lx linux]$ make orionboot
> make -C arch/mips/orion orionboot
> make[1]: Entering directory
> `/dvlp/jvhorne/jvh_21_lx_mips_cross_test_sv/vobs/gpl/linux/arch/mips/orion'
> mips-linux-objcopy -Obinary --verbose ../../../vmlinux orion.nosym
> copy from ../../../vmlinux(elf32-bigmips) to orion.nosym(binary)
> BFD: Warning: Writing section `.app_header' to huge (ie negative) file
> offset 0x800fec30.
> BFD: Warning: Writing section `.text' to huge (ie negative) file offset
> 0x80100c30.
> BFD: Warning: Writing section `.fixup' to huge (ie negative) file offset
> 0x80236930.
> BFD: Warning: Writing section `.text.exit' to huge (ie negative) file offset
> 0x80237830.
> .
> .
> .
> mips-linux-objcopy: orion.nosym: File truncated
>
>
> Do you think that the reason objcopy fails is because it isn't treating
> the addresses as 32 bit addresses, or could it be something else?
Note that the binary BFD target fills all "holes" with zeroes. I suspect
one or more sections are placed at an offset lower than 0x80000000.
Objcopy would have to make a huge file in this case and it gives up. You
probably need to fix your linker script.
--
+ Maciej W. Rozycki, Technical University of Gdansk, Poland +
+--------------------------------------------------------------+
+ e-mail: macro@ds2.pg.gda.pl, PGP key available +
