Re: [PATCHv2] PCI: xilinx: use fwnode_irq_get() for INTx IRQ lookup

Manivannan Sadhasivam <[email protected]> Mon, 3 Aug 2026 17:28:30 +0200
Newsgroups dev.linux.lists.llvm,org.infradead.lists.linux-arm-kernel,org.kernel.vger.linux-kernel,org.kernel.vger.linux-pci
Message-ID <ducc6fhoe6t3bivbpdanmojgtnntwykzwfog7vasps7tj6mp5t@vifnuyn4fjtn>
On Thu, Jul 30, 2026 at 11:54:29AM -0700, Rosen Penev wrote:
> On Wed, Jul 29, 2026 at 8:48 PM Manivannan Sadhasivam <[email protected]> wrote:
> >
> > On Wed, Jul 29, 2026 at 11:18:15AM -0700, Rosen Penev wrote:
> > > On Wed, Jul 29, 2026 at 9:00 AM Manivannan Sadhasivam <[email protected]> wrote:
> > > >
> > > > On Thu, Jul 16, 2026 at 01:58:11PM -0700, Rosen Penev wrote:
> > > > > Replace irq_of_parse_and_map() with fwnode_irq_get(dev_fwnode(dev), 0)
> > > > > in xilinx_pcie_parse_dt(). For an OF-backed device this is equivalent to
> > > > > the previous call, but uses the generic firmware-node API.
> > > > >
> > > >
> > > > What is the benefit of using fwnode API for OF only driver?
> > > I get conflicting information. Apparently of_irq_get needs a call to
> > > irq_dispose_mapping but fwnode_irq_get does not.
> >
> > That's not true. fwnode_irq_get() becomes a wrapper to of_irq_get() if OF is
> > used. So both needs a call to irq_dispose_mapping() in the error path.
> I don't see  irq_dispose_mapping() for fwnode_irq_get() in the codebase.

That doesn't matter. You need to check with irqchip maintainers and see if those
drivers need fix or not.

- Mani

-- 
மணிவண்ணன் சதாசிவம்