>>> On 31.08.16 at 05:56, <feng...@intel.com> wrote:
> We don't set the affinity for posted format IRTE, since the
> destination of these interrupts is vCPU and the vCPU affinity
> is set during vCPU scheduling.
So is this based on the assumption that after initial setup the function
would only ever get called for affinity changes? I'm not even sure
this is the case today, but I don't think we should build in such a
dependency.
Assuming that the main motivation is to avoid ...
> @@ -637,9 +640,12 @@ static int msi_msg_to_remap_entry(
> remap_rte->address_hi = 0;
> remap_rte->data = index - i;
>
> - memcpy(iremap_entry, &new_ire, sizeof(struct iremap_entry));
> - iommu_flush_cache_entry(iremap_entry, sizeof(struct iremap_entry));
> - iommu_flush_iec_index(iommu, 0, index);
> + if ( !iremap_entry->remap.p || !iremap_entry->remap.im )
> + {
> + memcpy(iremap_entry, &new_ire, sizeof(struct iremap_entry));
> + iommu_flush_cache_entry(iremap_entry, sizeof(struct iremap_entry));
> + iommu_flush_iec_index(iommu, 0, index);
> + }
... the actual updating here, may I suggest that you keep the
construction of new_ire and modify the if() here to check whether
nothing except affinity related bits changed? That would also take
care of certain (older) compiler versions likely warning about new_ire
potentially being used uninitialized.
Jan
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xen.org
https://lists.xen.org/xen-devel
