On Tue, Jan 08, 2019 at 11:12:05PM -0800, Andrey Smirnov wrote: > Port of a Linux commit bde4a5a00e761f55be92f62378cf5024ced79ee3 > > Some designs implement reset GPIO via a GPIO expander connected to a > peripheral bus. One such example would be i.MX7 Sabre board where said > GPIO is provided by SPI shift register connected to a bitbanged SPI bus. > To support such designs, allow reset GPIO request to defer probing of the > driver. > > Signed-off-by: Andrey Smirnov <andrew.smirnov@xxxxxxxxx> > Signed-off-by: Bjorn Helgaas <bhelgaas@xxxxxxxxxx> > Reviewed-by: Lucas Stach <l.stach@xxxxxxxxxxxxxx> > Cc: yurovsky@xxxxxxxxx > Cc: Fabio Estevam <fabio.estevam@xxxxxxx> > Cc: Dong Aisheng <dongas86@xxxxxxxxx> > Cc: linux-arm-kernel@xxxxxxxxxxxxxxxxxxx > > Signed-off-by: Andrey Smirnov <andrew.smirnov@xxxxxxxxx> > --- > drivers/pci/pci-imx6.c | 9 ++++++--- > 1 file changed, 6 insertions(+), 3 deletions(-) > > diff --git a/drivers/pci/pci-imx6.c b/drivers/pci/pci-imx6.c > index 6471f95b6..517c6bef8 100644 > --- a/drivers/pci/pci-imx6.c > +++ b/drivers/pci/pci-imx6.c > @@ -562,8 +562,8 @@ static const struct dw_pcie_host_ops imx6_pcie_host_ops = { > .host_init = imx6_pcie_host_init, > }; > > -static int __init imx6_add_pcie_port(struct imx6_pcie *imx6_pcie, > - struct device_d *dev) > +static int imx6_add_pcie_port(struct imx6_pcie *imx6_pcie, > + struct device_d *dev) > { > struct dw_pcie *pci = imx6_pcie->pci; > struct pcie_port *pp = &pci->pp; > @@ -581,7 +581,7 @@ static int __init imx6_add_pcie_port(struct imx6_pcie *imx6_pcie, > return 0; > } > > -static int __init imx6_pcie_probe(struct device_d *dev) > +static int imx6_pcie_probe(struct device_d *dev) > { > struct resource *iores; > struct dw_pcie *pci; > @@ -606,6 +606,9 @@ static int __init imx6_pcie_probe(struct device_d *dev) > > /* Fetch GPIOs */ > imx6_pcie->reset_gpio = of_get_named_gpio(np, "reset-gpio", 0); > + if (imx6_pcie->reset_gpio == -EPROBE_DEFER) > + return imx6_pcie->reset_gpio; Normally probe function run only once in barebox and in that case I do not care about loosing memory. When doing probe deferral here we loose a private data struct with each deferred probe here. Before returning -EPROBE_DEFER Here we already claimed a iomem resource using dev_request_mem_resource() which will fail during next probe, so I believe this patch doesn't work. Probe deferral is something I introduced to barebox out of despair, but it really doesn't work very well without all this devm_* stuff the kernel has. Sascha -- Pengutronix e.K. | | Industrial Linux Solutions | http://www.pengutronix.de/ | Peiner Str. 6-8, 31137 Hildesheim, Germany | Phone: +49-5121-206917-0 | Amtsgericht Hildesheim, HRA 2686 | Fax: +49-5121-206917-5555 | _______________________________________________ barebox mailing list barebox@xxxxxxxxxxxxxxxxxxx http://lists.infradead.org/mailman/listinfo/barebox