(+ Paul)
Hello,
On 23/01/2019 10:12, Andrii Anisov wrote:
From: Andrii Anisov <andrii_ani...@epam.com>
Taking decission by `need_iommu_pt_sync()` make us never kicking
s/decission/decision/
`iommu_iotlb_flush()` for IOMMUs which do share TLB with CPU.
I am not aware of platform where we share the TLB with the CPU. Do you mean
sharing the P2M?
So check `has_iommu_pt()` instead.
Signed-off-by: Andrii Anisov <andrii_ani...@epam.com>
---
Julien,
Could you please look at this, IMO there is a mistake here.
x86 uses `need_iommu_pt_sync()` to make decission if iommu's map/unmap should
be additionally called.
But ARM has no non-shared pt support in the mainline, so using
`need_iommu_pt_sync()` seems to be odd.
xen/arch/arm/p2m.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/xen/arch/arm/p2m.c b/xen/arch/arm/p2m.c
index 2394f97..059a391 100644
--- a/xen/arch/arm/p2m.c
+++ b/xen/arch/arm/p2m.c
@@ -1019,7 +1019,7 @@ static int __p2m_set_entry(struct p2m_domain *p2m,
!mfn_eq(lpae_get_mfn(*entry), lpae_get_mfn(orig_pte)) )
p2m_free_entry(p2m, orig_pte, level);
- if ( need_iommu_pt_sync(p2m->domain) &&
+ if ( has_iommu_pt(p2m->domain) &&
I think this makes sense because we want to flush the TLB when the P2M is
shared. Although, I would like to hear Paul's opinion here.
(lpae_is_valid(orig_pte) || lpae_is_valid(*entry)) )
{
unsigned int flush_flags = 0; >
Cheers,
--
Julien Grall
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel
