[Intel-gfx] [RFC PATCH] iommu/vt-d: Fix IOMMU field not populated on device hot re-plug

[Janusz Krzysztofik]

Hi Baolu,

On Thursday, August 29, 2019 3:43:31 AM CEST Lu Baolu wrote:
> Hi Janusz,
> 
> On 8/28/19 10:17 PM, Janusz Krzysztofik wrote:
> >> We should avoid kernel panic when a intel_unmap() is called against
> >> a non-existent domain.
> > Does that mean you suggest to replace
> > 	BUG_ON(!domain);
> > with something like
> > 	if (WARN_ON(!domain))
> > 		return;
> > and to not care of orphaned mappings left allocated?  Is there a way to 
inform
> > users that their active DMA mappings are no longer valid and they 
shouldn't
> > call dma_unmap_*()?
> > 
> >> But we shouldn't expect the IOMMU driver not
> >> cleaning up the domain info when a device remove notification comes and
> >> wait until all file descriptors being closed, right?
> > Shouldn't then the IOMMU driver take care of cleaning up resources still
> > allocated on device remove before it invalidates and forgets their 
pointers?
> > 
> 
> You are right. We need to wait until all allocated resources (iova and
> mappings) to be released.
> 
> How about registering a callback for BUS_NOTIFY_UNBOUND_DRIVER, and
> removing the domain info when the driver detachment completes?

Device core calls BUS_NOTIFY_UNBOUND_DRIVER on each driver unbind, regardless 
of a device being removed or not.  As long as the device is not unplugged and 
the BUS_NOTIFY_REMOVED_DEVICE notification not generated, an unbound driver is 
not a problem here.
Morever, BUS_NOTIFY_UNBOUND_DRIVER  is called even before 
BUS_NOTIFY_REMOVED_DEVICE so that wouldn't help anyway.
Last but not least, bus events are independent of the IOMMU driver use via 
DMA-API it exposes.

If keeping data for unplugged devices and reusing it on device re-plug is not 
acceptable then maybe the IOMMU driver should perform reference counting of 
its internal resources occupied by DMA-API users and perform cleanups on last 
release?

Thanks,
Janusz


> > Thanks,
> > Janusz
> 
> Best regards,
> Baolu
>