Re: [git pull] kmemcheck

["Vegard Nossum" <vegard.nossum@gmail.com>]

Hi,

On Wed, Apr 30, 2008 at 6:29 PM, Roland Dreier <rdreier@cisco.com> wrote:
>  >       kmemcheck: enable in the x86 Kconfig
>
>  This patch does:
>
>
>   > +config KMEMCHECK
>   > +    bool "kmemcheck: trap use of uninitialized memory"
>   > +    depends on X86_32
>
>  So it's only possible to select kmemcheck on 32-bit builds, But you also
>  seem to have 64-bit stuff, eg
>
>
>   >       kmemcheck: support for 64-bit
>   >       kmemcheck: add missing pte_hidden() for x86-64
>
>  so is the Kconfig patch slightly out-of-date?

We don't officially support 64-bit yet, that is why it is disabled in
the Kconfig. The 64-bit parts are preliminaries, and as with 32-bit,
should have no impact when CONFIG_KMEMCHECK=n (as is the case for all
64-bit builds as you pointed out). It is largely untested, but
according to Pekka it at least boots in qemu with kmemcheck enabled
even for 64-bit.

So the 64-bit parts we have is basically to make it _compile_ even if
you force KMEMCHECK=y on a 64-bit build.

(The main obstacle standing in the way of fully supporting 64-bit is
the opcode decoding. I don't really know how the 64-bit registers
works, so I will have to read about it first :-))

Thank you for looking over!

Vegard

-- 
"The animistic metaphor of the bug that maliciously sneaked in while
the programmer was not looking is intellectually dishonest as it
disguises that the error is the programmer's own creation."
	-- E. W. Dijkstra, EWD1036