Skip to contentSkip to conversation

PCI: kirin: Return -EPROBE_DEFER in case the gpio isn't ready

PATCH Applied Bean Huo 2 participants 5 messages Sep 18, 2020 – Sep 28, 2020

Subsystems (3)
lore
Share & tools
lore

Bean Huo

author
#0 lore raw

PATCH 1/1 PCI: kirin: Return -EPROBE_DEFER in case the gpio isn't ready

From: Bean Huo <[email protected]>

PCI driver might be probed before the gpiochip, so, of_get_named_gpio() can return -EPROBE_DEFER. And let kirin_pcie_probe() directly return -ENODEV, which will result in the PCIe probe failure and the PCIe will not be probed again after the gpiochip driver is loaded.

Fix the above issue by letting kirin_pcie_probe() return -EPROBE_DEFER in such a case.

Fixes 6e0832fa432e ("PCI: Collect all native drivers under drivers/pci/controller")
---
 drivers/pci/controller/dwc/pcie-kirin.c | 6 +++++-
 1 file changed, 5 insertions(+), 1 deletion(-)
drivers/pci/controller/dwc/pcie-kirin.c +5 -1
index e496f51e0152..74b88d158072 100644
@@ -507,8 +507,12 @@ static int kirin_pcie_probe(struct platform_device *pdev)
 
 	kirin_pcie->gpio_id_reset = of_get_named_gpio(dev->of_node,
 						      "reset-gpios", 0);
-	if (kirin_pcie->gpio_id_reset < 0)
+	if (kirin_pcie->gpio_id_reset == -EPROBE_DEFER) {
+		return -EPROBE_DEFER;
+	} else if (!gpio_is_valid(kirin_pcie->gpio_id_reset)) {
+		dev_err(dev, "unable to get a valid gpio pin\n");
 		return -ENODEV;
+	}
 
 	ret = kirin_pcie_power_on(kirin_pcie);
 	if (ret)
Signature · 1 line
2.17.1
#1 lore raw
Quoted 11 lines · Bean Huo

From: Bean Huo <[email protected]>

PCI driver might be probed before the gpiochip, so, of_get_named_gpio() can return -EPROBE_DEFER. And let kirin_pcie_probe() directly return -ENODEV, which will result in the PCIe probe failure and the PCIe will not be probed again after the gpiochip driver is loaded.

Fix the above issue by letting kirin_pcie_probe() return -EPROBE_DEFER in such a case.

Fixes 6e0832fa432e ("PCI: Collect all native drivers under drivers/pci/controller")

This is certainly not the commit that triggered the issue so I would remove it. Kirin maintainers are CC'ed, waiting for their ACK.

Lorenzo

Quoted 26 lines
---
 drivers/pci/controller/dwc/pcie-kirin.c | 6 +++++-
 1 file changed, 5 insertions(+), 1 deletion(-)
drivers/pci/controller/dwc/pcie-kirin.c +5 -1
index e496f51e0152..74b88d158072 100644
@@ -507,8 +507,12 @@ static int kirin_pcie_probe(struct platform_device *pdev)
 
 	kirin_pcie->gpio_id_reset = of_get_named_gpio(dev->of_node,
 						      "reset-gpios", 0);
-	if (kirin_pcie->gpio_id_reset < 0)
+	if (kirin_pcie->gpio_id_reset == -EPROBE_DEFER) {
+		return -EPROBE_DEFER;
+	} else if (!gpio_is_valid(kirin_pcie->gpio_id_reset)) {
+		dev_err(dev, "unable to get a valid gpio pin\n");
 		return -ENODEV;
+	}
 
 	ret = kirin_pcie_power_on(kirin_pcie);
 	if (ret)
Signature · 1 line
2.17.1
author
#2 lore raw
Quoted 11 lines · Lorenzo Pieralisi
Quoted 6 lines

Fix the above issue by letting kirin_pcie_probe() return -EPROBE_DEFER in
such a case.

Fixes 6e0832fa432e ("PCI: Collect all native drivers under

drivers/pci/controller")

This is certainly not the commit that triggered the issue so I would remove it. Kirin maintainers are CC'ed, waiting for their ACK.

Lorenzo

Hi Lorenzo

Thanks very much for your reply. It is true this is a bug which was introduced by the origin commit. It is ok for me to remove this fix tag.

Thanks,
Bean

author
#3 lore raw

seems the Hisilicon PCI driver maintainers are absent, however, we are still using their old platform based on Kirin.

hi, Lorenzo
is it possible to take this patch without Hisilicon maintainter's ACK?

Thanks,
Bean

#4 lore raw
Quoted 5 lines · Bean Huo

seems the Hisilicon PCI driver maintainers are absent, however, we are still using their old platform based on Kirin.

hi, Lorenzo
is it possible to take this patch without Hisilicon maintainter's ACK?

I applied it to pci/kirin, tentatively for v5.10.

Lorenzo

Quoted 3 lines

Thanks,
Bean