On Thu, Nov 21, 2024 at 08:43:54AM -0500, Steven Rostedt wrote: > On Thu, 21 Nov 2024 19:34:31 +0800 > Shuai Xue <xueshuai@xxxxxxxxxxxxxxxxx> wrote: > > Sure, I can reorganize the directory. You have two optional proposals: > > > > 1. /sys/kernel/debug/tracing/events/pci_hp_event > > 2. /sys/kernel/debug/tracing/events/pci/pci_hp_event > > > > I personally prefer the second approach. [...] > That said, is all these events going to be related to hotplug? If so, you > probably want "hotplug" or "hp" in the system name. If you make it just > "pci", then it will be expected that all the events will be related to PCI > in general. Grouping them under "pci" probably makes more sense than grouping them under "hotplug". A lot can be hotplugged (cpu, memory, PCI, USB, ...) but it's mostly unrelated hardware and enabling all their events at once doesn't seem to make much sense. Whereas I can imagine people want to enable all PCI trace events. We currently only have tracepoints for AER in the PCI core, but perhaps people will want to add new ones for power faults, DPC recovery, bandwidth control etc. Would something like /sys/kernel/debug/tracing/events/pci/hotplug_event be possible? It would seem more elegant than /sys/kernel/debug/tracing/events/pci/pci_hp_event because it avoids the duplication of "pci" in the path. Thanks, Lukas
