On 31.08.2022 16:11, Volodymyr Babchuk wrote: > There are number of cases where pcidevs_lock() is used to protect > something that is not related to PCI devices per se. > > Probably pcidev_lock in these places should be replaced with some > other lock. > > This patch is not intended to be merged and is present only to discuss > this use of pcidevs_lock()
For all such instances it needs to be understood what (if anything) is being protected and how the same guarding can be achieved in the new model. Since I'm afraid that's simply stating the obvious, I guess I don't really understand what needs discussing here. Jan
