On 13/01/17 14:46, Mark Rutland wrote:
> On Fri, Jan 13, 2017 at 01:36:12PM +0100, Christoffer Dall wrote:
>> On Fri, Jan 13, 2017 at 11:31:32AM +0000, Marc Zyngier wrote:
>> Further, are we guaranteed that the static branch gets compiled into
>> something that doesn't actually look at cpu_hwcap_keys, which is not
>> mapped in hyp mode?
> The fact that this might happen silently seems to be a larger problem.
> Can we do something like the EFI stub, and ensure that (unintentional)
> references to symbols outside of the hyp-stub will fail to link? That's
> ensrue by some symbol mangling in drivers/firmware/efi/libstub/Makefile.
> I think this may have come up before; I can't recall if there was some
> reason that was problematic.
>From what I remember, this was a gigantic mess... We could revisit it
though (after all, it's been a whole year since we did turn the whole
thing upside down -- time for a rewrite!).
Jazz is not dead. It just smells funny...
kvmarm mailing list