Andi Kleen wrote:
>> so it would be
>> nice to use one of the other serializing instructions in this case.
>>
>
> You would first need to find one that works in ring 3. On x86-64 it is
> used in the gettimeoday vsyscall in ring 3 to synchronize the TSC and
> afaik John was about to implement that for i386 too.
>
Well, that's really usermode code, so I don't think we'd necessarily
touch it at all. It's not the same problem as the (single, at the
moment) ring 0 use.
> BTW another issue that I haven't checked but we will need to make
> this also an alternative() for another case - it is faily important
> to patch it out on Intel systems with a synchronized TSC where it is
> fairly expensive. That is also not done yet on i386, but will be
> likely once vsyscall gettimeofday is implemented.
>
> So basically you would need double patching. Ugly.
>
Yeah. I guess the cleanest way to do that is do the paravirt
substitution, and then nop it out later if it isn't needed in the vsyscall.
> I would recommend to keep it out of para ops.
It's hardly a big deal either way. There's only one in-kernel use of it.
J
