On 31/05/18 10:49, Russell King - ARM Linux wrote:

On Tue, May 29, 2018 at 06:02:28PM +0100, Marc Zyngier wrote:

quoted

On Tue, 29 May 2018 15:55:01 +0100,
Russell King wrote:

quoted

Warn at error level if the context switching function is not what we
are expecting.  This can happen with big.Little systems, which we
currently do not support.

Signed-off-by: Russell King <redacted>
Boot-tested-by: Tony Lindgren [off-list ref]
Reviewed-by: Tony Lindgren <tony@atomide.com>

I assume this is a temporary situation until the ARM port grows the
necessary infrastructure to support this mitigation on heterogeneous
systems.

As I've said, I think that is going to be a very difficult problem to
resolve.  I detailed why in previous emails, but it seems each time I
do that, no one bothers to respond (presumably because no one has any
ideas how to sort that problem either.)

I believe that it's better to get some of the mitigations in the kernel
and warn about non-supported setups than it is to hold it back.

I agree with you that it is better to get this merged quickly, and then
address exotic configurations (which are unfortunately quite common
these days) separately. I haven't had the time to try and understand how
to fix it though.

I notice that you haven't replied to some of the patches (7 and 8),
which makes me think that you have an issue with them - and as tonight
is likely the last linux-next before the merge window, we're basically
out of time to do another respin if there's something you don't like
there and if we want to get them in during the next merge window.

Patches 7 and 8 didn't compile for me, and I had to fix-up things
manually in order to test it. Not that it should hold you back from
merging it.

Thanks,

	M.
-- 
Jazz is not dead. It just smells funny...