Re: [PATCH 16/39] irqdomain: Make normal and nomap irqdomains exclusive

On Mon, 15 Nov 2021 19:05:17 +0000,
Cédric Le Goater [off-list ref] wrote:

Hello Mark,

s/k/c/

On 5/20/21 18:37, Marc Zyngier wrote:

quoted

Direct mappings are completely exclusive of normal mappings, meaning
that we can refactor the code slightly so that we can get rid of
the revmap_direct_max_irq field and use the revmap_size field
instead, reducing the size of the irqdomain structure.

Signed-off-by: Marc Zyngier <maz@kernel.org>

This patch is breaking the POWER9/POWER10 XIVE driver (these are not
old PPC systems :) on machines sharing the same LSI HW IRQ. For instance,
a linux KVM guest with a virtio-rng and a virtio-balloon device. In that
case, Linux creates two distinct IRQ mappings which can lead to some
unexpected behavior.

Either the irq domain translates, or it doesn't. If the driver creates
a nomap domain, and yet expects some sort of translation to happen,
then the driver is fundamentally broken. And even without that: how do
you end-up with a single HW interrupt having two mappings?

A fix to go forward would be to change the XIVE IRQ domain to use a
'Tree' domain for reverse mapping and not the 'No Map' domain mapping.
I will keep you updated for XIVE.

I bet there is a bit more to it. From what you are saying above,
something rather ungodly is happening in the XIVE code.

	M.

-- 
Without deviation from the norm, progress is not possible.