Re: [crash, bisected] Re: [PATCH 3/4] x86_64: Fold pda into per cpu area

[Eric W. Biederman]

"H. Peter Anvin" <hpa@xxxxxxxxx> writes:

> Eric W. Biederman wrote:
>>
>> For i386 since virtual address space is precious and because there were
>> concerns about putting code in __pa we actually relocate the kernel symbols
>> during load right after decompression.  When we do relocations absolute
>> symbols are a killer.
>>
>
> Well, it means making it clear to the relocator if it should relocate those
> symbols or not.  Since IIRC we're doing "all or nothing" relocation, the
> relative offsets are always the same, even between sections.

Yes the relative offsets stay the same.  I don't remember if ld
generates useable relocations for what it figures are absolute
symbols.  I remember the solution was that anything that we wanted to
relocate we would make section relative, and anything else we would
leave absolute, and that those were easy things to do.

The nasty case is that occasionally ld has a bug where it turns section
relative symbols into global symbols if there is a section without data.
I don't recall us caring in those instances.

While interesting.  All of this is irrelevant until we start talking
unification between x86_63 and x86_64 as only x86_32 has this
restriction.

Eric
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/