If the presence detect state (PDS) becomes active after the link comes up (DLLLA), the hotplug code removes the device and then re-loads the driver. For most devices, this is a mere inconvenience, and for PCIe storage devices that are part of a RAID set which started rebuilding, it can get fun. Ideally, we wouldn't remove perfectly good devices. According to the old PCIe spec PDS would always have to come up at or before DLLLA. Since not everyone followed this (looking at you Dell and HPE!!!!), this now got standardized in PCIe 4.0. (*). This series has two solutions for this problem, and is intended to serve as a bikeshedding point for which is better: 1. Try to wait for PDS _before_ loading the driver 2. Load as usual, and recognize the delayed PDS event as such I think (2) is more generic and elegant, but a lot of people seem to like something similar to (1). (*) ECN was approved in Nov 2018, and is normative spec text. A lot of the leaked PCIe 4.0 specs do not have this change. Alexandru Gagniuc (4): PCI: hotplug: Add support for disabling in-band presence PCI: pciehp: Do not turn off slot if presence comes up after link PCI: hotplug: Wait for PDS when in-band presence is disabled PCI: hotplug: Add quirk For Dell nvme pcie switches drivers/pci/hotplug/pciehp.h | 1 + drivers/pci/hotplug/pciehp_ctrl.c | 24 ++++++++++++++ drivers/pci/hotplug/pciehp_hpc.c | 54 ++++++++++++++++++++++++++++++- include/linux/pci.h | 1 + include/uapi/linux/pci_regs.h | 2 ++ 5 files changed, 81 insertions(+), 1 deletion(-) -- 2.19.2
