IRQ inbalance on ath79

[Koen Vandeputte]

On Fri, Jun 19, 2026 at 7:42 PM Rosen Penev <rosenp at gmail.com> wrote:
>
> On Fri, Jun 19, 2026 at 7:48 AM Koen Vandeputte
> <koen.vandeputte at citymesh.com> wrote:
> >
> > Hi Rosen,
> >
> > I tested the latest OpenWRT master state including your patch series
> > for ath79 on various ath79 targets:
> >
> > faf8f3760d irq-ath79-intc: add missing \n
> > 00841615f1 irq-ath79-intc: avoid negative values
> > dbb4ac9342 irq-ath79-intc: add irq_dispose_mapping
> > 4ae9c7bdc1 irq-ath79-intc: statically allocate irq_chip
> > 3c2e7b3eee irq-ath79-intc: add chained_irq_enter/exit
> > 5ed10923f9 irq-ath79-intc: rename pending_mask to enable_mask
> > 10d7a164ad irq-ath79-intc: don't use hwirq_max
> > 20275da9cd irq-ath79-intc: use generic_handle_domain_irq
> > e46cc54e9d irq-ath79-intc: add SPDX license
> > 122ea336de irq-ath79-intc: switch from add to create
> > 0354345493 irq-ath79-intc: remove panic and add kfree
> > 86cc4d84bb ath79: move intc driver out of patch
> >
> > They result in IRQ issues:
> > https://pastebin.com/raw/rppfDd6p
> Are those pcie enable_irq warnings also a result of this?

Yes.
They are.

Thanks

> >
> > Removing the series from the build fixes all issues.
> > Do you have an idea where the bug is? :-)
> >
> > Ps.
> > Did you test it on hardware upfront? :-)
> >
> >
> > Thanks!
> >
> > Koen