[PATCHv7 01/26] iommu/arm-smmu: change IOMMU_EXEC to IOMMU_NOEXEC

[Antonios Motakis]

Exposing the XN flag of the SMMU driver as IOMMU_NOEXEC instead of
IOMMU_EXEC makes it enforceable, since for IOMMUs that don't support
the XN flag pages will always be executable.

Signed-off-by: Antonios Motakis 
---
 drivers/iommu/arm-smmu.c | 9 +
 include/linux/iommu.h| 2 +-
 2 files changed, 6 insertions(+), 5 deletions(-)

diff --git a/drivers/iommu/arm-smmu.c b/drivers/iommu/arm-smmu.c
index a83cc2a..c7cbdda 100644
--- a/drivers/iommu/arm-smmu.c
+++ b/drivers/iommu/arm-smmu.c
@@ -1252,7 +1252,7 @@ static int arm_smmu_alloc_init_pte(struct arm_smmu_device 
*smmu, pmd_t *pmd,
   unsigned long pfn, int prot, int stage)
 {
pte_t *pte, *start;
-   pteval_t pteval = ARM_SMMU_PTE_PAGE | ARM_SMMU_PTE_AF | ARM_SMMU_PTE_XN;
+   pteval_t pteval = ARM_SMMU_PTE_PAGE | ARM_SMMU_PTE_AF;
 
if (pmd_none(*pmd)) {
/* Allocate a new set of tables */
@@ -1286,10 +1286,11 @@ static int arm_smmu_alloc_init_pte(struct 
arm_smmu_device *smmu, pmd_t *pmd,
pteval |= ARM_SMMU_PTE_MEMATTR_NC;
}
 
+   if (prot & IOMMU_NOEXEC)
+   pteval |= ARM_SMMU_PTE_XN;
+
/* If no access, create a faulting entry to avoid TLB fills */
-   if (prot & IOMMU_EXEC)
-   pteval &= ~ARM_SMMU_PTE_XN;
-   else if (!(prot & (IOMMU_READ | IOMMU_WRITE)))
+   if (!(prot & (IOMMU_READ | IOMMU_WRITE)))
pteval &= ~ARM_SMMU_PTE_PAGE;
 
pteval |= ARM_SMMU_PTE_SH_IS;
diff --git a/include/linux/iommu.h b/include/linux/iommu.h
index 20f9a52..e1a644c 100644
--- a/include/linux/iommu.h
+++ b/include/linux/iommu.h
@@ -27,7 +27,7 @@
 #define IOMMU_READ (1 << 0)
 #define IOMMU_WRITE(1 << 1)
 #define IOMMU_CACHE(1 << 2) /* DMA cache coherency */
-#define IOMMU_EXEC (1 << 3)
+#define IOMMU_NOEXEC   (1 << 3)
 
 struct iommu_ops;
 struct iommu_group;
-- 
1.8.3.2

--
To unsubscribe from this list: send the line "unsubscribe kvm" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCHv7 01/26] iommu/arm-smmu: change IOMMU_EXEC to IOMMU_NOEXEC

[Will Deacon]

Hi Antonios,

On Tue, Sep 23, 2014 at 03:46:00PM +0100, Antonios Motakis wrote:
> Exposing the XN flag of the SMMU driver as IOMMU_NOEXEC instead of
> IOMMU_EXEC makes it enforceable, since for IOMMUs that don't support
> the XN flag pages will always be executable.
> 
> Signed-off-by: Antonios Motakis 
> ---
>  drivers/iommu/arm-smmu.c | 9 +
>  include/linux/iommu.h| 2 +-
>  2 files changed, 6 insertions(+), 5 deletions(-)

[...]

> diff --git a/include/linux/iommu.h b/include/linux/iommu.h
> index 20f9a52..e1a644c 100644
> --- a/include/linux/iommu.h
> +++ b/include/linux/iommu.h
> @@ -27,7 +27,7 @@
>  #define IOMMU_READ   (1 << 0)
>  #define IOMMU_WRITE  (1 << 1)
>  #define IOMMU_CACHE  (1 << 2) /* DMA cache coherency */
> -#define IOMMU_EXEC   (1 << 3)
> +#define IOMMU_NOEXEC (1 << 3)

This hunk needs to be a separate patch merged by Joerg before I can take the
arm-smmu part (which looks fine).

Will
--
To unsubscribe from this list: send the line "unsubscribe kvm" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCHv7 01/26] iommu/arm-smmu: change IOMMU_EXEC to IOMMU_NOEXEC

[Alex Williamson]

On Tue, 2014-09-23 at 15:58 +0100, Will Deacon wrote:
> Hi Antonios,
> 
> On Tue, Sep 23, 2014 at 03:46:00PM +0100, Antonios Motakis wrote:
> > Exposing the XN flag of the SMMU driver as IOMMU_NOEXEC instead of
> > IOMMU_EXEC makes it enforceable, since for IOMMUs that don't support
> > the XN flag pages will always be executable.
> > 
> > Signed-off-by: Antonios Motakis 
> > ---
> >  drivers/iommu/arm-smmu.c | 9 +
> >  include/linux/iommu.h| 2 +-
> >  2 files changed, 6 insertions(+), 5 deletions(-)
> 
> [...]
> 
> > diff --git a/include/linux/iommu.h b/include/linux/iommu.h
> > index 20f9a52..e1a644c 100644
> > --- a/include/linux/iommu.h
> > +++ b/include/linux/iommu.h
> > @@ -27,7 +27,7 @@
> >  #define IOMMU_READ (1 << 0)
> >  #define IOMMU_WRITE(1 << 1)
> >  #define IOMMU_CACHE(1 << 2) /* DMA cache coherency */
> > -#define IOMMU_EXEC (1 << 3)
> > +#define IOMMU_NOEXEC   (1 << 3)
> 
> This hunk needs to be a separate patch merged by Joerg before I can take the
> arm-smmu part (which looks fine).

That separate hunk would be unbuildable since arm-smmu depends on the
IOMMU_EXEC define.  Patch 2/ is also in iommu code and gates patch 3/ in
arm-smmu.  The IOMMU-core changes are pretty trivial, so perhaps Joerg
would be willing to ACK 1&2 and let Will include the first 3 patches
through his tree.

These first 3 patches should have been sent on their own since they're
small an obvious so they don't get hung up on the reset of the series.
Thanks,

Alex

--
To unsubscribe from this list: send the line "unsubscribe kvm" in
the body of a message to majord...@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html