From: Bjorn Helgaas <bhelgaas@xxxxxxxxxx> All callers of platform_get_irq() and related functions interpret a negative return value as an error. A few also interpret zero as an error. platform_get_irq() should return either a negative error number or a valid non-zero IRQ, so there's no need to check for zero. This series: - Extends the platform_get_irq() function comment to say it returns a non-zero IRQ number or a negative error number. - Adds a WARN() if platform_get_irq() ever *does* return zero (this would be a bug in the underlying arch code, and most callers are not prepared for this). - Updates drivers/pci/ to check consistently using "irq < 0". This is based on Aman's series [1]. I propose to merge this via the PCI tree, given acks from Greg and Thomas. [1] https://lore.kernel.org/r/cover.1583952275.git.amanharitsh123@xxxxxxxxx Aman Sharma (1): PCI: Check for platform_get_irq() failure consistently Bjorn Helgaas (1): driver core: platform: Clarify that IRQ 0 is invalid drivers/base/platform.c | 40 ++++++++++++------- drivers/pci/controller/dwc/pci-imx6.c | 4 +- drivers/pci/controller/dwc/pcie-tegra194.c | 4 +- .../controller/mobiveil/pcie-mobiveil-host.c | 4 +- drivers/pci/controller/pci-aardvark.c | 3 ++ drivers/pci/controller/pci-v3-semi.c | 4 +- drivers/pci/controller/pcie-mediatek.c | 3 ++ drivers/pci/controller/pcie-tango.c | 4 +- 8 files changed, 41 insertions(+), 25 deletions(-) base-commit: 8f3d9f354286745c751374f5f1fcafee6b3f3136 -- 2.25.1