Re: [PATCH v8 04/10] iommu/vt-d: Support flushing more translation cache types

[Jacob Pan]

On Thu, 19 Dec 2019 10:46:51 +0800
Lu Baolu <baolu.lu@xxxxxxxxxxxxxxx> wrote:

> Hi,
> 
> On 12/17/19 3:24 AM, Jacob Pan wrote:
> > When Shared Virtual Memory is exposed to a guest via vIOMMU,
> > scalable IOTLB invalidation may be passed down from outside IOMMU
> > subsystems. This patch adds invalidation functions that can be used
> > for additional translation cache types.
> > 
> > Signed-off-by: Jacob Pan<jacob.jun.pan@xxxxxxxxxxxxxxx>
> > ---
> >   drivers/iommu/dmar.c        | 46
> > +++++++++++++++++++++++++++++++++++++++++++++
> > drivers/iommu/intel-pasid.c |  3 ++- include/linux/intel-iommu.h |
> > 21 +++++++++++++++++---- 3 files changed, 65 insertions(+), 5
> > deletions(-)
> > 
> > diff --git a/drivers/iommu/dmar.c b/drivers/iommu/dmar.c
> > index 3acfa6a25fa2..f2f5d75da94a 100644
> > --- a/drivers/iommu/dmar.c
> > +++ b/drivers/iommu/dmar.c
> > @@ -1348,6 +1348,20 @@ void qi_flush_iotlb(struct intel_iommu
> > *iommu, u16 did, u64 addr, qi_submit_sync(&desc, iommu);
> >   }
> >   
> > +/* PASID-based IOTLB Invalidate */
> > +void qi_flush_iotlb_pasid(struct intel_iommu *iommu, u16 did, u64
> > addr, u32 pasid,
> > +		unsigned int size_order, u64 granu, int ih)
> > +{
> > +	struct qi_desc desc = {.qw2 = 0, .qw3 = 0};
> > +
> > +	desc.qw0 = QI_EIOTLB_PASID(pasid) | QI_EIOTLB_DID(did) |
> > +		QI_EIOTLB_GRAN(granu) | QI_EIOTLB_TYPE;
> > +	desc.qw1 = QI_EIOTLB_ADDR(addr) | QI_EIOTLB_IH(ih) |
> > +		QI_EIOTLB_AM(size_order);
> > +
> > +	qi_submit_sync(&desc, iommu);
> > +}  
> 
> There's another version of pasid-based iotlb invalidation.
> 
> https://lkml.org/lkml/2019/12/10/2128
> 
> Let's consider merging them.
> 
Absolutely, the difference i see is that the granularity is explicit
here. Here we do invalidation request from the guest. Perhaps, we can
look at consolidation once this use case is supported?

> Best regards,
> baolu

[Jacob Pan]