This function can only be reached from EPT-related code which is inherently
HAP. Thus it is not useful to check for shadow_paging (or lack of HAP) there.

Moreover, it is an error to call this function in the non-EPT cases.

Not a functional change.

Signed-off-by: Teddy Astie <teddy.astie@vates.tech>
---
This function is only called through EPT code and by vmx_domain_update_eptp()
called by EPT log-dirty logic, and doesn't look reachable from shadow paging
code.

I think the original reason of this check was for eventually allowing guests to
use both shadow paging and HAP and switch between the 2 dynamically.

 xen/arch/x86/mm/p2m-ept.c | 5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)

diff --git a/xen/arch/x86/mm/p2m-ept.c b/xen/arch/x86/mm/p2m-ept.c
index ce4ef632ae..dfdbfa0afe 100644
--- a/xen/arch/x86/mm/p2m-ept.c
+++ b/xen/arch/x86/mm/p2m-ept.c
@@ -1268,9 +1268,10 @@ static void ept_sync_domain_mask(struct p2m_domain *p2m, const cpumask_t *mask)
 void ept_sync_domain(struct p2m_domain *p2m)
 {
     struct domain *d = p2m->domain;
+    ASSERT(hap_enabled(d));
 
-    /* Only if using EPT and this domain has some VCPUs to dirty. */
-    if ( paging_mode_shadow(d) || !d->vcpu || !d->vcpu[0] )
+    /* Only if this domain has some VCPUs to dirty. */
+    if ( !d->vcpu || !d->vcpu[0] )
         return;
 
     ept_sync_domain_prepare(p2m);
-- 
2.51.2



--
Teddy Astie | Vates XCP-ng Developer

XCP-ng & Xen Orchestra - Vates solutions

web: https://vates.tech

On 27.11.2025 14:39, Teddy Astie wrote:
> This function can only be reached from EPT-related code which is inherently
> HAP. Thus it is not useful to check for shadow_paging (or lack of HAP) there.
> 
> Moreover, it is an error to call this function in the non-EPT cases.
> 
> Not a functional change.
> 
> Signed-off-by: Teddy Astie <teddy.astie@vates.tech>

Reviewed-by: Jan Beulich <jbeulich@suse.com>

> ---
> This function is only called through EPT code and by vmx_domain_update_eptp()
> called by EPT log-dirty logic, and doesn't look reachable from shadow paging
> code.
> 
> I think the original reason of this check was for eventually allowing guests to
> use both shadow paging and HAP and switch between the 2 dynamically.

I don't think there ever was such a plan. The function originally lived in vmx.c,
and there - even if just as a safeguard - having the extra check may have made
sense.

Jan