On Tue, Sep 27, 2022 at 05:27:42PM +0200, Lorenzo Pieralisi wrote: > On Mon, Jul 25, 2022 at 06:35:27PM +0100, Marc Zyngier wrote: > > [...] > > > > That is precisely the way I've been testing it and everything appears > > > to be tore down as it should. > > > > > > And a PCI driver that has been unbound should have released its > > > resources, or that's a driver bug. Right? > > > > But that's the thing: you can easily remove part of the infrastructure > > without the endpoint driver even noticing. It may not happen in your > > particular case if removing the RC driver will also nuke the endpoints > > in the process, but I can't see this is an absolute guarantee. The > > crash pointed to by an earlier email is symptomatic of it. > > > > > And for the OF INTx case you mentioned earlier, aren't those mapped by > > > PCI core and could in theory be released by core as well? > > > > Potentially, though I haven't tried to follow the life cycle of those. > > The whole thing is pretty fragile, and this sort of resource is rarely > > expected to be removed... > > This made me notice that we don't undo the actions (ie bridge->map_irq()) > executed in pci_assign_irq() in pci_device_remove(); I don't think this > can be right and that's already a candidate for a fix. There's an inherent asymmetry here as a legacy interrupt can be used by more than one device. It is mapped on first use as each user calls ->map_irq() but can only be disposed when the final user is gone as I mentioned here: https://lore.kernel.org/all/Yt+6azfwd%2FLuMzoG@xxxxxxxxxxxxxxxxxxxx/ > It is not necessarily related to this thread topic, though I believe, > in an _ideal_ world, removing a bridge should guarantee that all > the downstream devices (ie drivers) had a chance of freeing/disposing > the resources they allocated. This in theory; I totally understand > Marc's point of view here and we should make up our mind about what > we want to do on host bridge removal policy - this will take me more > time to get to the bottom of it. Johan
