Re: [PATCH ath-next] wifi: ath10k: spread the PCI copy engine interrupts over the CPUs
From: Julius Bairaktaris
Date: Sat Nov 21 2026 - 12:54:45 EST
On Mon, Nov 16, 2026 at 4:25 PM Loic Poulain wrote:
> So, from what you say I understand pci_alloc_irq_vectors_affinity() is
> the wrong API here because it sizes the allocation around CPUs,
> whereas your goal is to size the allocation around device
> CE/capability (8 vectors).
Yes.
> Can you mention cfe9011a05a8 ("ath10k: remove MSI range support").
> And why we re-introduce MIS support and how is it different from
> previous implementation?
Agreed. The v2 2/2 message names it, gives its reasons, and says that
its per copy engine handlers and tasklets do not come back: every
vector gets ath10k_pci_interrupt_handler(), which schedules the one
NAPI, so only the CPU of the HTT copy engine's interrupt changes.
> Also, you should update the comment/kdoc for
> ath10k_pci_per_engine_handler() which says MSI interrupts are not
> handled there.
Agreed. That function went with cfe9011a05a8; the stale comment sits
above ath10k_pci_interrupt_handler(), and v2 rewrites it.
v2: https://lore.kernel.org/all/20261121175402.644803-1-julius@xxxxxxxxxxxxxx/
Julius