iommu/vt-d: Fix intel_pasid_max_id
PATCH 1/1 iommu/vt-d: Fix intel_pasid_max_id
Extended Capability Register PSS field (PASID Size Supported) corresponds to the PASID bit size -1.
"A value of N in this field indicates hardware supports PASID field of N+1 bits (For example, value of 7 in this field, indicates 8-bit PASIDs are supported)".
Fix the computation of intel_pasid_max_id accordingly.
--- drivers/iommu/intel-iommu.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/iommu/intel-iommu.c +1 -1
@@ -3331,7 +3331,7 @@ static int __init init_dmars(void)* than the smallest supported. */ if (pasid_supported (iommu)) { -u32 temp =2 <<ecap_pss (iommu->ecap); +u32 temp =2 << (ecap_pss (iommu->ecap) +1 ); intel_pasid_max_id =min_t (u32, temp, intel_pasid_max_id);
Signature · 1 line
2.20.1
Quoted 21 lines · On Tue, 30 Apr 2019 09:29:40 +0200 Eric Auger <[email protected]> wrote:
Extended Capability Register PSS field (PASID Size Supported) corresponds to the PASID bit size -1.
"A value of N in this field indicates hardware supports PASID field of N+1 bits (For example, value of 7 in this field, indicates 8-bit PASIDs are supported)".
Fix the computation of
intel_pasid_max_idaccordingly.Fixes 562831747f62 ("iommu/vt-d: Global PASID name space")Signed-off-by Eric Auger <[email protected]>--- drivers/iommu/intel-iommu.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)drivers/iommu/intel-iommu.c +0
-0
drivers/iommu/intel-iommu.c static int __init init_dmars(void) · line 3334
Quoted 5 lines
drivers/iommu/intel-iommu.c +1
-1@@ -3331,7 +3331,7 @@ static int __init init_dmars(void)* than the smallest supported. */ if (pasid_supported (iommu)) { -u32 temp =2 <<ecap_pss (iommu->ecap); +u32 temp =2 << (ecap_pss (iommu->ecap) +1 );
here it is "2 << bits" not "1 << bits", so the original code is correct.
But I agree it would be more clear to the spec. if we do:
1 << (ecap_pss(iommu->ecap) + 1);
Quoted 3 lines
drivers/iommu/intel-iommu.c +0
-0intel_pasid_max_id =min_t (u32, temp, intel_pasid_max_id);
Hi Jacob,
Quoted 31 lines · Jacob Pan
Quoted 21 lines · On Tue, 30 Apr 2019 09:29:40 +0200 Eric Auger <[email protected]> wrote:
Extended Capability Register PSS field (PASID Size Supported) corresponds to the PASID bit size -1.
"A value of N in this field indicates hardware supports PASID field of N+1 bits (For example, value of 7 in this field, indicates 8-bit PASIDs are supported)".
Fix the computation of
intel_pasid_max_idaccordingly.Fixes 562831747f62 ("iommu/vt-d: Global PASID name space")Signed-off-by Eric Auger <[email protected]>--- drivers/iommu/intel-iommu.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)drivers/iommu/intel-iommu.c +0
-0drivers/iommu/intel-iommu.c static int __init init_dmars(void) · line 3334
Quoted 5 lines
drivers/iommu/intel-iommu.c +1
-1@@ -3331,7 +3331,7 @@ static int __init init_dmars(void)* than the smallest supported. */ if (pasid_supported (iommu)) { -u32 temp =2 <<ecap_pss (iommu->ecap); +u32 temp =2 << (ecap_pss (iommu->ecap) +1 );here it is "2 << bits" not "1 << bits", so the original code is correct.
Quoted 2 lines
But I agree it would be more clear to the spec. if we do:
1 << (ecap_pss(iommu->ecap)+ 1);
Oups OK, my eyes. Forget it :-(
Thanks
Eric
Quoted 4 lines
Quoted 3 lines
drivers/iommu/intel-iommu.c +0
-0intel_pasid_max_id =min_t (u32, temp, intel_pasid_max_id);