On Wed, Mar 12, 2025 at 05:04:24PM -0500, Bjorn Helgaas wrote:
> On Mon, Mar 10, 2025 at 04:16:46PM -0400, Frank Li wrote:
> > If the parent 'ranges' property in DT correctly describes the address
> > translation, the cpu_addr_fixup() callback is not needed. Print a warning
> > message to inform users to correct their DT files.
>
> > +++ b/drivers/pci/controller/dwc/pcie-designware.c
> > @@ -1125,6 +1125,8 @@ int dw_pcie_init_parent_bus_offset(struct dw_pcie *pci, const char *reg_name,
> >
> > fixup = pci->ops->cpu_addr_fixup;
> > if (fixup) {
> > + dev_warn_once(pci->dev, "cpu_addr_fixup() usage detected. Please fix DT!\n");
>
> I don't think we need this. The mere presence of .cpu_addr_fixup()
> doesn't tell us the DT is broken. When we have .cpu_addr_fixup(), the
> DT is only broken if DT tells us something different than
> .cpu_addr_fixup() tells us. And we already warn about that in the
> "reg_addr != fixup_addr" case.
It is encourage people to fix dts first and remove .cpu_addr_fixup().
Most case below "...redundant" is not printed at all at most case, even
there are .cpu_addr_fixup().
>
> > +
> > fixup_addr = fixup(pci, cpu_phy_addr);
> > if (reg_addr == fixup_addr) {
> > dev_info(dev, "%#010llx %s reg[%d] == %#010llx; %ps is redundant\n",
>
> This message is really just a hint that DT is fine and
> .cpu_addr_fixup() is redundant but harmless. If you want a dev_warn()
> here to encourage people to remove .cpu_addr_fixup(), I'm fine with
> that.
It is encourage people remove .cpu_addr_fixup() because dts already fixed.
>
> Seems like "dev_warn()" would be enough, probably no need for
> "dev_warn_once()" since we should only run this once per controller
> anyway, so I don't think we'll get spammed with messages.
