2020-01-29 05:58:51

by Jacob Pan

Subject: [PATCH V9 02/10] iommu/uapi: Define a mask for bind data

Memory type related guest PASID bind data can be grouped together for
one simple check.
Link: https://lore.kernel.org/linux-iommu/20200109095123.17ed5e6b@jacob-builder/

Signed-off-by: Jacob Pan <[email protected]>
---
include/uapi/linux/iommu.h | 5 ++++-
1 file changed, 4 insertions(+), 1 deletion(-)

diff --git a/include/uapi/linux/iommu.h b/include/uapi/linux/iommu.h
index 4ad3496e5c43..fcafb6401430 100644
--- a/include/uapi/linux/iommu.h
+++ b/include/uapi/linux/iommu.h
@@ -284,7 +284,10 @@ struct iommu_gpasid_bind_data_vtd {
__u32 pat;
__u32 emt;
};
-
+#define IOMMU_SVA_VTD_GPASID_EMT_MASK (IOMMU_SVA_VTD_GPASID_CD | \
+ IOMMU_SVA_VTD_GPASID_EMTE | \
+ IOMMU_SVA_VTD_GPASID_PCD | \
+ IOMMU_SVA_VTD_GPASID_PWT)
/**
* struct iommu_gpasid_bind_data - Information about device and guest PASID binding
* @version: Version of this data structure
--
2.7.4

2020-02-12 12:45:57

by Eric Auger

Subject: Re: [PATCH V9 02/10] iommu/uapi: Define a mask for bind data

Hi Jacob,

On 1/29/20 7:01 AM, Jacob Pan wrote:
> Memory type related guest PASID bind data can be grouped together for
> one simple check.
Those are flags related to memory type.
> Link: https://lore.kernel.org/linux-iommu/20200109095123.17ed5e6b@jacob-builder/
not sure the link is really helpful.
>
> Signed-off-by: Jacob Pan <[email protected]>
> ---
> include/uapi/linux/iommu.h | 5 ++++-
> 1 file changed, 4 insertions(+), 1 deletion(-)
>
> diff --git a/include/uapi/linux/iommu.h b/include/uapi/linux/iommu.h
> index 4ad3496e5c43..fcafb6401430 100644
> --- a/include/uapi/linux/iommu.h
> +++ b/include/uapi/linux/iommu.h
> @@ -284,7 +284,10 @@ struct iommu_gpasid_bind_data_vtd {
> __u32 pat;
> __u32 emt;
> };
> -
> +#define IOMMU_SVA_VTD_GPASID_EMT_MASK (IOMMU_SVA_VTD_GPASID_CD | \
> + IOMMU_SVA_VTD_GPASID_EMTE | \
> + IOMMU_SVA_VTD_GPASID_PCD | \
> + IOMMU_SVA_VTD_GPASID_PWT)
Why EMT rather than MT or MTS?
the spec says:
Those fields are treated as Reserved(0) for implementations not
supporting Memory Type (MTS=0 in Extended Capability Register).

> /**
> * struct iommu_gpasid_bind_data - Information about device and guest PASID binding
> * @version: Version of this data structure
>

Thanks

Eric

2020-03-20 21:31:18

by Jacob Pan

Subject: Re: [PATCH V9 02/10] iommu/uapi: Define a mask for bind data

On Wed, 12 Feb 2020 13:43:43 +0100
Auger Eric <[email protected]> wrote:

> Hi Jacob,
>
> On 1/29/20 7:01 AM, Jacob Pan wrote:
> > Memory type related guest PASID bind data can be grouped together
> > for one simple check.
> Those are flags related to memory type.
right, will rephrase.
> > Link:
> > https://lore.kernel.org/linux-iommu/20200109095123.17ed5e6b@jacob-builder/
> not sure the link is really helpful.
> >
will delete. the patch is very simple.

> > Signed-off-by: Jacob Pan <[email protected]>
> > ---
> > include/uapi/linux/iommu.h | 5 ++++-
> > 1 file changed, 4 insertions(+), 1 deletion(-)
> >
> > diff --git a/include/uapi/linux/iommu.h b/include/uapi/linux/iommu.h
> > index 4ad3496e5c43..fcafb6401430 100644
> > --- a/include/uapi/linux/iommu.h
> > +++ b/include/uapi/linux/iommu.h
> > @@ -284,7 +284,10 @@ struct iommu_gpasid_bind_data_vtd {
> > __u32 pat;
> > __u32 emt;
> > };
> > -
> > +#define IOMMU_SVA_VTD_GPASID_EMT_MASK
> > (IOMMU_SVA_VTD_GPASID_CD | \
> > + IOMMU_SVA_VTD_GPASID_EMTE
> > | \
> > + IOMMU_SVA_VTD_GPASID_PCD
> > | \
> > +
> > IOMMU_SVA_VTD_GPASID_PWT)
> Why EMT rather than MT or MTS?
> the spec says:
> Those fields are treated as Reserved(0) for implementations not
> supporting Memory Type (MTS=0 in Extended Capability Register).
>
MTS makes more sense, will change.
It was from hygiene p.o.v. checking the flag to avoid touching these
fields.

Thanks,

Jacob
> > /**
> > * struct iommu_gpasid_bind_data - Information about device and
> > guest PASID binding
> > * @version: Version of this data structure
> >
>
> Thanks
>
> Eric
>

[Jacob Pan]