On Thu, Sep 20, 2018 at 11:58:28AM +0100, Lorenzo Pieralisi wrote:
> [+Paul]
>
> On Wed, Sep 19, 2018 at 02:08:08PM -0700, Nick Desaulniers wrote:
> > On Wed, Sep 19, 2018 at 11:55 AM Nathan Chancellor
> > <natechancellor@xxxxxxxxx> wrote:
> > >
> > > WARNING: vmlinux.o(.text+0x4758cc): Section mismatch in reference from
> > > the function kirin_pcie_probe() to the function
> > > .init.text:kirin_add_pcie_port()
> > > The function kirin_pcie_probe() references
> > > the function __init kirin_add_pcie_port().
> > > This is often because kirin_pcie_probe lacks a __init
> > > annotation or the annotation of kirin_add_pcie_port is wrong.
> > >
> > > Add the __init annotation to both kirin_pcie_probe and
> > > kirin_pcie_add_msi then use builtin_platform_driver_probe
> > > instead of builtin_platform_driver + .probe to avoid a section
> > > mismatch warning with kirin_pcie_driver.
> > >
> > > Fixes: fc5165db245a ("PCI: kirin: Add HiSilicon Kirin SoC PCIe controller driver")
> > > Reported-by: Nick Desaulniers <ndesaulniers@xxxxxxxxxx>
> > > Suggested-by: Nick Desaulniers <ndesaulniers@xxxxxxxxxx>
> > > Signed-off-by: Nathan Chancellor <natechancellor@xxxxxxxxx>
> > > ---
> > >
> > > v1 -> v2:
> > >
> > > Instead of removing the annotation on kirin_add_pcie_port, add it to
> > > kirin_pcie_add_msi and kirin_pcie_probe. To avoid a warning with this
> > > configuration, use builtin_platform_driver_probe.
> > >
> > > drivers/pci/controller/dwc/pcie-kirin.c | 9 ++++-----
> > > 1 file changed, 4 insertions(+), 5 deletions(-)
> > >
> > > diff --git a/drivers/pci/controller/dwc/pcie-kirin.c b/drivers/pci/controller/dwc/pcie-kirin.c
> > > index 5352e0c3be82..f64fed12de51 100644
> > > --- a/drivers/pci/controller/dwc/pcie-kirin.c
> > > +++ b/drivers/pci/controller/dwc/pcie-kirin.c
> > > @@ -448,8 +448,8 @@ static const struct dw_pcie_host_ops kirin_pcie_host_ops = {
> > > .host_init = kirin_pcie_host_init,
> > > };
> > >
> > > -static int kirin_pcie_add_msi(struct dw_pcie *pci,
> > > - struct platform_device *pdev)
> > > +static int __init kirin_pcie_add_msi(struct dw_pcie *pci,
> > > + struct platform_device *pdev)
> > > {
> > > int irq;
> > >
> > > @@ -481,7 +481,7 @@ static int __init kirin_add_pcie_port(struct dw_pcie *pci,
> > > return dw_pcie_host_init(&pci->pp);
> > > }
> > >
> > > -static int kirin_pcie_probe(struct platform_device *pdev)
> > > +static int __init kirin_pcie_probe(struct platform_device *pdev)
> > > {
> > > struct device *dev = &pdev->dev;
> > > struct kirin_pcie *kirin_pcie;
> > > @@ -533,11 +533,10 @@ static const struct of_device_id kirin_pcie_match[] = {
> > > };
> > >
> > > static struct platform_driver kirin_pcie_driver = {
> > > - .probe = kirin_pcie_probe,
> > > .driver = {
> > > .name = "kirin-pcie",
> > > .of_match_table = kirin_pcie_match,
> > > .suppress_bind_attrs = true,
> > > },
> > > };
> > > -builtin_platform_driver(kirin_pcie_driver);
> > > +builtin_platform_driver_probe(kirin_pcie_driver, kirin_pcie_probe);
> >
> > It would be good to get additional review from someone who knows more
> > about driver callback lifecycles, but I think this is the correct fix.
> > Thanks Nathan.
> > Reviewed-by: Nick Desaulniers <ndesaulniers@xxxxxxxxxx>
>
> AFAICS we can't use builtin_platform_driver_probe() if the probe()
> itself may need deferral (and I *reckon* it may need that, see eg
> clk_get()) so I suspect we have to go back to v1 for the patch,
> apologies if so.
I think that the best course of action consists in merging v1 since
we have not reached a definitive conclusion on v2, please let me
know.
Thanks,
Lorenzo
