intel_iommu: Use correct shift for 256 bits qi descriptor

[PATCH] intel_iommu: Use correct shift for 256 bits qi descriptor

Posted by Liu Yi L 3 years, 9 months ago

In chapter 10.4.23 of VT-d spec 3.0, Descriptor Width bit was introduced
in VTD_IQA_REG. Sfotware could set this bit to tell VT-d the QI descriptor
from software would be 256 bits. Accordingly, the VTD_IQH_QH_SHIFT should
be 5 when descriptor size is 256 bits.

This patch adds the DW bit check when deciding the shift used to update
VTD_IQH_REG.

Signed-off-by: Liu Yi L <yi.l.liu@intel.com>
---
 hw/i386/intel_iommu.c          | 7 ++++++-
 hw/i386/intel_iommu_internal.h | 3 ++-
 2 files changed, 8 insertions(+), 2 deletions(-)

diff --git a/hw/i386/intel_iommu.c b/hw/i386/intel_iommu.c
index df7ad25..8703a2d 100644
--- a/hw/i386/intel_iommu.c
+++ b/hw/i386/intel_iommu.c
@@ -2549,6 +2549,11 @@ static bool vtd_process_inv_desc(IntelIOMMUState *s)
 /* Try to fetch and process more Invalidation Descriptors */
 static void vtd_fetch_inv_desc(IntelIOMMUState *s)
 {
+    int qi_shift;
+
+    /* Refer to 10.4.23 of VT-d spec 3.0 */
+    qi_shift = s->iq_dw ? VTD_IQH_QH_SHIFT_5 : VTD_IQH_QH_SHIFT_4;
+
     trace_vtd_inv_qi_fetch();
 
     if (s->iq_tail >= s->iq_size) {
@@ -2567,7 +2572,7 @@ static void vtd_fetch_inv_desc(IntelIOMMUState *s)
         }
         /* Must update the IQH_REG in time */
         vtd_set_quad_raw(s, DMAR_IQH_REG,
-                         (((uint64_t)(s->iq_head)) << VTD_IQH_QH_SHIFT) &
+                         (((uint64_t)(s->iq_head)) << qi_shift) &
                          VTD_IQH_QH_MASK);
     }
 }
diff --git a/hw/i386/intel_iommu_internal.h b/hw/i386/intel_iommu_internal.h
index 862033e..3d5487f 100644
--- a/hw/i386/intel_iommu_internal.h
+++ b/hw/i386/intel_iommu_internal.h
@@ -230,7 +230,8 @@
 #define VTD_IQA_DW_MASK             0x800
 
 /* IQH_REG */
-#define VTD_IQH_QH_SHIFT            4
+#define VTD_IQH_QH_SHIFT_4          4
+#define VTD_IQH_QH_SHIFT_5          5
 #define VTD_IQH_QH_MASK             0x7fff0ULL
 
 /* ICS_REG */
-- 
2.7.4

Re: [PATCH] intel_iommu: Use correct shift for 256 bits qi descriptor

Posted by Peter Xu 3 years, 9 months ago

On Sat, Jul 04, 2020 at 01:07:15AM -0700, Liu Yi L wrote:
> In chapter 10.4.23 of VT-d spec 3.0, Descriptor Width bit was introduced
> in VTD_IQA_REG. Sfotware could set this bit to tell VT-d the QI descriptor
> from software would be 256 bits. Accordingly, the VTD_IQH_QH_SHIFT should
> be 5 when descriptor size is 256 bits.
> 
> This patch adds the DW bit check when deciding the shift used to update
> VTD_IQH_REG.
> 
> Signed-off-by: Liu Yi L <yi.l.liu@intel.com>

Reviewed-by: Peter Xu <peterx@redhat.com>

-- 
Peter Xu

RE: [PATCH] intel_iommu: Use correct shift for 256 bits qi descriptor

Posted by Liu, Yi L 3 years, 9 months ago

> From: Peter Xu <peterx@redhat.com>
> Sent: Tuesday, July 7, 2020 4:58 AM
> 
> On Sat, Jul 04, 2020 at 01:07:15AM -0700, Liu Yi L wrote:
> > In chapter 10.4.23 of VT-d spec 3.0, Descriptor Width bit was
> > introduced in VTD_IQA_REG. Sfotware could set this bit to tell VT-d
> > the QI descriptor from software would be 256 bits. Accordingly, the
> > VTD_IQH_QH_SHIFT should be 5 when descriptor size is 256 bits.
> >
> > This patch adds the DW bit check when deciding the shift used to
> > update VTD_IQH_REG.
> >
> > Signed-off-by: Liu Yi L <yi.l.liu@intel.com>
> 
> Reviewed-by: Peter Xu <peterx@redhat.com>

thanks.

Regards,
Yi Liu

Re: [PATCH] intel_iommu: Use correct shift for 256 bits qi descriptor

Posted by Jason Wang 3 years, 9 months ago

On 2020/7/4 下午4:07, Liu Yi L wrote:
> In chapter 10.4.23 of VT-d spec 3.0, Descriptor Width bit was introduced
> in VTD_IQA_REG. Sfotware could set this bit to tell VT-d the QI descriptor


Typo.


> from software would be 256 bits. Accordingly, the VTD_IQH_QH_SHIFT should
> be 5 when descriptor size is 256 bits.
>
> This patch adds the DW bit check when deciding the shift used to update
> VTD_IQH_REG.
>
> Signed-off-by: Liu Yi L <yi.l.liu@intel.com>


Acked-by: Jason Wang <jasowang@redhat.com>


> ---
>   hw/i386/intel_iommu.c          | 7 ++++++-
>   hw/i386/intel_iommu_internal.h | 3 ++-
>   2 files changed, 8 insertions(+), 2 deletions(-)
>
> diff --git a/hw/i386/intel_iommu.c b/hw/i386/intel_iommu.c
> index df7ad25..8703a2d 100644
> --- a/hw/i386/intel_iommu.c
> +++ b/hw/i386/intel_iommu.c
> @@ -2549,6 +2549,11 @@ static bool vtd_process_inv_desc(IntelIOMMUState *s)
>   /* Try to fetch and process more Invalidation Descriptors */
>   static void vtd_fetch_inv_desc(IntelIOMMUState *s)
>   {
> +    int qi_shift;
> +
> +    /* Refer to 10.4.23 of VT-d spec 3.0 */
> +    qi_shift = s->iq_dw ? VTD_IQH_QH_SHIFT_5 : VTD_IQH_QH_SHIFT_4;
> +
>       trace_vtd_inv_qi_fetch();
>   
>       if (s->iq_tail >= s->iq_size) {
> @@ -2567,7 +2572,7 @@ static void vtd_fetch_inv_desc(IntelIOMMUState *s)
>           }
>           /* Must update the IQH_REG in time */
>           vtd_set_quad_raw(s, DMAR_IQH_REG,
> -                         (((uint64_t)(s->iq_head)) << VTD_IQH_QH_SHIFT) &
> +                         (((uint64_t)(s->iq_head)) << qi_shift) &
>                            VTD_IQH_QH_MASK);
>       }
>   }
> diff --git a/hw/i386/intel_iommu_internal.h b/hw/i386/intel_iommu_internal.h
> index 862033e..3d5487f 100644
> --- a/hw/i386/intel_iommu_internal.h
> +++ b/hw/i386/intel_iommu_internal.h
> @@ -230,7 +230,8 @@
>   #define VTD_IQA_DW_MASK             0x800
>   
>   /* IQH_REG */
> -#define VTD_IQH_QH_SHIFT            4
> +#define VTD_IQH_QH_SHIFT_4          4
> +#define VTD_IQH_QH_SHIFT_5          5
>   #define VTD_IQH_QH_MASK             0x7fff0ULL
>   
>   /* ICS_REG */