On Thursday 31 January 2008 04:35:03 am Ingo Molnar wrote:
>
> * Yinghai Lu <[EMAIL PROTECTED]> wrote:
>
> > ENTRY(startup_64)
> > SECTIONS
> > {
> > - /* Be careful parts of head_64.S assume startup_64 is at
> > + /* Be careful parts of head_64.S assume startup_32 is at
> > * address 0.
> > */
> > . = 0;
>
> but this linker script rule is about startup_64, not about startup_32.
>
> startup_32 does exist on 64-bit too (and is 32-bit code), but it's
> linked to a different address that is not 0:
>
> ffffffff806c23c0 r startup_32
>
> startup_64 is at offset 0:
>
> ffffffff80200000 T startup_64
>
> So your patch looks wrong to me.
there are several startup_32 in 64 bit
1. one is for bzImage, it is arch/x86/boot/compressed/head_64.S, and it said
startup_64 will at 0x200. that what i mean
2. one is vmlinux in linux-2.6/ , it is in arch/x86/kernel/trampoline_64.S it
is for AP trampoline and wake up. that is what you mean
actually there was another one in arch/x86_64/kernel/head_64.S, and it was
removed when Eric made linux-2.6/vmlinux to be 64bit elf.
So that patch is needed.
YH
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to [EMAIL PROTECTED]
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
