Re: [PATCH v1 07/10] iommu/vt-d: Use an xarray for global device_domain_info

[Lu Baolu]

On 2/7/22 3:14 PM, Christoph Hellwig wrote:
> > + #define DEVI_IDX(seg, bus, devfn) ((((u16)(seg)) << 16) | PCI_DEVID(bus, devfn))
>
> Please turn this into an real function.

Sure.

> >   /*
> > - * Iterate over elements in device_domain_list and call the specified
> > + * Iterate over elements in device_domain_array and call the specified
> >    * callback @fn against each element.
> >    */
> >   int for_each_device_domain(int (*fn)(struct device_domain_info *info,
> >   				     void *data), void *data)
> >   {
> >   	struct device_domain_info *info;
> > +	unsigned long index;
> > +	int ret = 0;
> >   
> > +	rcu_read_lock();
> > +	xa_for_each(&device_domain_array, index, info) {
> >   		ret = fn(info, data);
> > +		if (ret)
> > +			break;
> >   	}
> > +	rcu_read_unlock();
>
> Can't we just open code this in the caller now?

That's better. I will remove this helper and make iteration in the only
caller.

> >   const struct iommu_ops intel_iommu_ops;
> > @@ -900,7 +898,8 @@ static void pgtable_walk(struct intel_iommu *iommu, unsigned long pfn, u8 bus, u
> >   	struct dmar_domain *domain;
> >   	int offset, level;
> >   
> > -	info = dmar_search_domain_by_dev_info(iommu->segment, bus, devfn);
> > +	info = xa_load(&device_domain_array,
> > +		       DEVI_IDX(iommu->segment, bus, devfn));
> >   	if (!info || !info->domain) {
> >   		pr_info("device [%02x:%02x.%d] not probed\n",
> >   			bus, PCI_SLOT(devfn), PCI_FUNC(devfn));
>
> Is there any refcounting or other life time protection for the info
> structures?

The info structure's life is managed by iommu_probe/release_device(). It
is created in probe() and freed in release().

Best regards,
baolu