Message ID | 7671478f-e7c0-c43e-6395-1adc70333e04@suse.com (mailing list archive) |
---|---|
State | New, archived |
Headers | show |
Series | VT-d: address fallout from XSA-400 | expand |
On Thu, Apr 07, 2022 at 11:27:53AM +0200, Jan Beulich wrote: > If get_iommu_domid() in domain_context_unmap_one() fails, we better > wouldn't clear the context entry in the first place, as we're then unable > to issue the corresponding flush. However, we have no need to look up the > DID in the first place: What needs flushing is very specifically the DID > that was in the context entry before our clearing of it. > > Signed-off-by: Jan Beulich <jbeulich@suse.com> Reviewed-by: Roger Pau Monné <roger.pau@citrix.com> Thanks, Roger.
> From: Jan Beulich <jbeulich@suse.com> > Sent: Thursday, April 7, 2022 5:28 PM > > If get_iommu_domid() in domain_context_unmap_one() fails, we better > wouldn't clear the context entry in the first place, as we're then unable > to issue the corresponding flush. However, we have no need to look up the > DID in the first place: What needs flushing is very specifically the DID > that was in the context entry before our clearing of it. > > Signed-off-by: Jan Beulich <jbeulich@suse.com> Reviewed-by: Kevin Tian <kevin.tian@intel.com> > --- > This (an intended follow-up to XSA-399) is actually a prereq to what was > called patch 1 so far in this series. > > --- a/xen/drivers/passthrough/vtd/iommu.c > +++ b/xen/drivers/passthrough/vtd/iommu.c > @@ -1830,18 +1830,12 @@ int domain_context_unmap_one( > return 0; > } > > + iommu_domid = context_domain_id(*context); > + > context_clear_present(*context); > context_clear_entry(*context); > iommu_sync_cache(context, sizeof(struct context_entry)); > > - iommu_domid = get_iommu_did(domid, iommu, !domain->is_dying); > - if ( iommu_domid == -1 ) > - { > - spin_unlock(&iommu->lock); > - unmap_vtd_domain_page(context_entries); > - return -EINVAL; > - } > - > rc = iommu_flush_context_device(iommu, iommu_domid, > PCI_BDF2(bus, devfn), > DMA_CCMD_MASK_NOBIT, 0);
--- a/xen/drivers/passthrough/vtd/iommu.c +++ b/xen/drivers/passthrough/vtd/iommu.c @@ -1830,18 +1830,12 @@ int domain_context_unmap_one( return 0; } + iommu_domid = context_domain_id(*context); + context_clear_present(*context); context_clear_entry(*context); iommu_sync_cache(context, sizeof(struct context_entry)); - iommu_domid = get_iommu_did(domid, iommu, !domain->is_dying); - if ( iommu_domid == -1 ) - { - spin_unlock(&iommu->lock); - unmap_vtd_domain_page(context_entries); - return -EINVAL; - } - rc = iommu_flush_context_device(iommu, iommu_domid, PCI_BDF2(bus, devfn), DMA_CCMD_MASK_NOBIT, 0);
If get_iommu_domid() in domain_context_unmap_one() fails, we better wouldn't clear the context entry in the first place, as we're then unable to issue the corresponding flush. However, we have no need to look up the DID in the first place: What needs flushing is very specifically the DID that was in the context entry before our clearing of it. Signed-off-by: Jan Beulich <jbeulich@suse.com> --- This (an intended follow-up to XSA-399) is actually a prereq to what was called patch 1 so far in this series.