RE: [PATCH v3 1/1] PCI: hv: Spread load across CPUs when selecting effective affinity CPU

[Michael Kelley]

From: Bjorn Helgaas <helgaas@xxxxxxxxxx> Sent: Wednesday, October 28, 2026 10:14 AM
> 
> On Wed, Oct 28, 2026 at 08:55:59AM -0700, Michael Kelley wrote:
> > In a system with multiple devices with the same configuration (such as
> > NVMe controllers), the affinity masks are the same for all such devices.
> > Linux typicaly must select a single CPU from each mask to be the effective
> 
> s/typicaly/typically/
> 
> > @@ -881,25 +1016,48 @@ static int hv_pci_vec_irq_domain_alloc(struct irq_domain  *domain,
> >  }
> >
> >  /*
> > - * Pick the first cpu as the irq affinity that can be temporarily used for
> > + * Pick the first CPU as the irq affinity that can be temporarily used for
> 
> s/irq/IRQ/ if you want to follow the majority usage in this file, also
> below
> 
> >   * composing MSI from the hypervisor. GIC will eventually set the right
> >   * affinity for the irq and the 'unmask' will retarget the interrupt to that
> > - * cpu.
> > + * CPU.

Thanks. I'll fix these. My own preference is for "IRQ" anyway.

Michael