RE: [RFC PATCH 6/6] iommu/arm-smmu: Update bindings document for smmu-inst-as-data DT option

[Anup Patel]

> -----Original Message-----
> From: Mark Rutland [mailto:mark.rutland@xxxxxxx]
> Sent: 27 January 2016 17:59
> To: Anup Patel
> Cc: Catalin Marinas; Joerg Roedel; Will Deacon; Robin Murphy; Sricharan R;
> Linux IOMMU; Linux ARM Kernel; Rob Herring; Pawel Moll; Ian Campbell; Kumar
> Gala; Device Tree; Ray Jui; Scott Branden; Vikram Prakash; Linux Kernel; bcm-
> kernel-feedback-list
> Subject: Re: [RFC PATCH 6/6] iommu/arm-smmu: Update bindings document for
> smmu-inst-as-data DT option
> 
> On Wed, Jan 27, 2016 at 10:51:19AM +0530, Anup Patel wrote:
> > This patch adds info about 'smmu-inst-as-data' DT option in ARM
> > SMMUv1/SMMUv2 driver bindings document.
> >
> > Signed-off-by: Anup Patel <anup.patel@xxxxxxxxxxxx>
> > Reviewed-by: Ray Jui <rjui@xxxxxxxxxxxx>
> > Reviewed-by: Vikram Prakash <vikramp@xxxxxxxxxxxx>
> > Reviewed-by: Scott Branden <sbranden@xxxxxxxxxxxx>
> > ---
> >  Documentation/devicetree/bindings/iommu/arm,smmu.txt | 8 ++++++++
> >  1 file changed, 8 insertions(+)
> >
> > diff --git a/Documentation/devicetree/bindings/iommu/arm,smmu.txt
> > b/Documentation/devicetree/bindings/iommu/arm,smmu.txt
> > index 7180745..4c4d03e 100644
> > --- a/Documentation/devicetree/bindings/iommu/arm,smmu.txt
> > +++ b/Documentation/devicetree/bindings/iommu/arm,smmu.txt
> > @@ -49,6 +49,14 @@ conditions.
> >                    NOTE: this only applies to the SMMU itself, not
> >                    masters connected upstream of the SMMU.
> >
> > +- smmu-inst-as-data : Treat privilege/unprivilege instruction fetch as
> > +                  data read for SMMUv2. The SMMU driver by default provides
> > +                  unprivilege read-write permission in page table entries.
> > +                  For SMMUv2, privilege instruction fetch from MMU masters
> > +                  will cause a context fault for unprivilege read-write
> > +                  pages. To allow both privilege and unprivilege instruction
> > +                  fetch, we have to forcefully treat it as data read.
> 
> What is this needed for? Which masters do instruction fetches through the
> SMMU, and when?
> 
> Surely this should only need to aplly to a subset of transactions?

The boot_manager of PL330 does privileged instruction fetches which
cause privilege-level context fault in SMMU because current SMMU
driver provides unprivileged read-write permissions.

Regards,
Anup