Re: MIPS md_apply_fix()(?) problem.

["H . J . Lu" <hjl at lucon dot org>]

On Thu, Nov 15, 2001 at 06:07:43PM -0800, cgd@broadcom.com wrote:
> "H . J . Lu" <hjl@lucon.org> writes:
> > FWIW, I got this with binutils 2.11.92.0.10 and 2.11.92.0.11. It seems
> > ok to me.
> 
> What does that actually mean?  What seems "OK"?  Do you mean that you
> fixed it?  What is the fix?
> 
> (I verified that this is still a problem with the sourceware binutils
> as of 2001-11-15 00:00 UTC -- i.e., this morning.)
> 
> You didn't say what you actually used to build the thing that you
> included the objdump of, either, so, I don't know what it is that you
> think works properly...
> 
> 
> > FYI, I used "-G 0 -mno-abicalls -fno-pic" for kernel.
> 
> Yes, "of course" (that's what the kernel Makefile normally uses, and I
> don't change it).
> 
> But they're not necessary for reproducing the problem, as far as I can
> tell, so i figured it didn't make sense to confuse the matter by
> discussing them.
> 

I don't understand what you meant. I just compiled the test you
provided. Here is typescript. Does the objdump output look ok to
you?


H.J.
--
Script started on Thu Nov 15 18:14:36 2001
sh-2.04$ cat foo.s
# 92 "foo.S"
        .text

        .ent fn # unneeded
        .type fn,@function # unneeded
fn: # unneeded
        lw $5,boot_mem_map+12($17)
        .end fn # unneeded

        .data
        .align 2

spacer:
        .space 4


        .globl boot_mem_map
boot_mem_map:
        .space 4
sh-2.04$ mipsel-linux-gcc -G 0 -mno-abicalls -fno-pic -c foo.s
sh-2.04$ objdump -dr foo.o       

foo.o:     file format elf32-tradlittlemips

Disassembly of section .text:

00000000 <fn>:
   0:	3c050000 	lui	a1,0x0
			0: R_MIPS_HI16	boot_mem_map
   4:	00b12821 	addu	a1,a1,s1
   8:	8ca5000c 	lw	a1,12(a1)
			8: R_MIPS_LO16	boot_mem_map
   c:	00000000 	nop
sh-2.04$ exit

Script done on Thu Nov 15 18:15:07 2001