It was Jon Hunter (jonathanh@xxxxxxxxxx), who identified the issue. Credits to him. I'm merely upstreaming the fix. Thanks, Vidya Sagar On 13-04-2024 00:44, Bjorn Helgaas wrote:
External email: Use caution opening links or attachments On Mon, Apr 08, 2024 at 03:00:53PM +0530, Vidya Sagar wrote:Tegra194 PCIe probe path is taking failure path in success case for Endpoint mode. Return success from the switch case instead of going into the failure path. Fixes: c57247f940e8 ("PCI: tegra: Add support for PCIe endpoint mode in Tegra194") Signed-off-by: Vidya Sagar <vidyas@xxxxxxxxxx> Reviewed-by: Jon Hunter <jonathanh@xxxxxxxxxx> --- v2: * Added 'Fixes' and 'Reviewed-by' from Jon Hunter drivers/pci/controller/dwc/pcie-tegra194.c | 3 +++ 1 file changed, 3 insertions(+) diff --git a/drivers/pci/controller/dwc/pcie-tegra194.c b/drivers/pci/controller/dwc/pcie-tegra194.c index 4bba31502ce1..1a8178dc899a 100644 --- a/drivers/pci/controller/dwc/pcie-tegra194.c +++ b/drivers/pci/controller/dwc/pcie-tegra194.c @@ -2273,11 +2273,14 @@ static int tegra_pcie_dw_probe(struct platform_device *pdev) ret = tegra_pcie_config_ep(pcie, pdev); if (ret < 0) goto fail; + else + return 0;Wow, how did you ever notice this? It looks like this path would previously have returned "ret" (which was most likely 0 for success) but with an extra tegra_bpmp_put() that we shouldn't have done. Eagle eyes!break; default: dev_err(dev, "Invalid PCIe device type %d\n", pcie->of_data->mode); + ret = -EINVAL; } fail: -- 2.25.1
