On Tue, 20 Sep 2022 22:46:21 +0100,
Alexey Kardashevskiy <aik@xxxxxxxxx> wrote:
>
>
>
> On 21/09/2022 02:08, Marc Zyngier wr
> ote:
> > On Tue, 20 Sep 2022 13:51:43 +0100,
> > Alexey Kardashevskiy <aik@xxxxxxxxx> wrote:
> >>
> >> When introduced, IRQFD resampling worked on POWER8 with XICS. However
> >> KVM on POWER9 has never implemented it - the compatibility mode code
> >> ("XICS-on-XIVE") misses the kvm_notify_acked_irq() call and the native
> >> XIVE mode does not handle INTx in KVM at all.
> >>
> >> This moved the capability support advertising to platforms and stops
> >> advertising it on XIVE, i.e. POWER9 and later.
> >>
> >> Signed-off-by: Alexey Kardashevskiy <aik@xxxxxxxxx>
> >> Acked-by: Nicholas Piggin <npiggin@xxxxxxxxx>
> >> [For KVM RISC-V]
> >> Acked-by: Anup Patel <anup@xxxxxxxxxxxxxx>
> >> ---
> >> Changes:
> >> v2:
> >> * removed ifdef for ARM64.
> >
> > The same argument applies to both x86 and s390, which do select
> > HAVE_KVM_IRQFD from the KVM config option. Only power allows this
> > option to be selected depending on the underlying interrupt controller
> > emulation.
> >
> > As for riscv and mips, they don't select HAVE_KVM_IRQFD, and this
> > isn't a user-selectable option. So why do they get patched at all?
>
> Before the patch, the capability was advertised on those, with your
> proposal it will stop.
No, they were never advertised, since none of these architectures
select CONFIG_HAVE_KVM_IRQFD, and this option is not selectable by the
user.
> Which is a change in behavior which some may
> say requires a separate patch (like, one per platform). I am
> definitely overthinking it though and you are right.
Well, either there is no change in behaviour (and therefore I am
right), or there is one (and I am wrong).
M.
--
Without deviation from the norm, progress is not possible.
