1. Patchew
  2. Xen
  3. View series

[XEN][PATCH] x86/hvm: revise "cpu_has_vmx" usage for !CONFIG_INTEL_VMX case

Grygorii Strashko posted 1 patch 4 months, 2 weeks ago
Patches applied successfully (tree, apply log)
git fetch https://gitlab.com/xen-project/patchew/xen tags/patchew/20250924101417.229108-1-grygorii._5Fstrashko@epam.com
xen/arch/x86/hvm/dom0_build.c | 2 +-
xen/arch/x86/hvm/hvm.c        | 2 +-
2 files changed, 2 insertions(+), 2 deletions(-)
[XEN][PATCH] x86/hvm: revise "cpu_has_vmx" usage for !CONFIG_INTEL_VMX case
Posted by Grygorii Strashko 4 months, 2 weeks ago 
From: Grygorii Strashko <grygorii_strashko@epam.com>

Since commit b99227347230 ("x86: Fix AMD_SVM and INTEL_VMX dependency") the
HVM Intel VT-x support can be disabled, but it still keeps VMX
code partially built-in. Particularly in HVM code there are two places:

1) hvm/dom0_build.c
 dom0_construct_pvh()->pvh_populate_p2m()
    ...
    if ( cpu_has_vmx && paging_mode_hap(d) && !vmx_unrestricted_guest(v) )
    {
        ...
        [unreachable for !cpu_has_vmx case]
        rc = pvh_setup_vmx_realmode_helpers(d);

pvh_setup_vmx_realmode_helpers() allocates memory and configures
 HVM_PARAM_VM86_TSS_SIZED
 HVM_PARAM_IDENT_PT

2) hvm/hvm.c
 hvm_set_param()
    ...
    case HVM_PARAM_IDENT_PT:

        if ( !paging_mode_hap(d) || !cpu_has_vmx )
        {
            d->arch.hvm.params[index] = value;
            break;
        }
        [unreachable for !cpu_has_vmx case]
        ...

Hence HVM_PARAM_IDENT_PT/HVM_PARAM_VM86_TSS_SIZED are used only by VMX code
above code become unreachable in !cpu_has_vmx case and can be optimazed
when !CONFIG_INTEL_VMX.

Replace "cpu_has_vmx" with using_vmx() to account !CONFIG_INTEL_VMX and allow
compiler DCE to optimize code.

Signed-off-by: Grygorii Strashko <grygorii_strashko@epam.com>
---
add/remove: 0/0 grow/shrink: 0/2 up/down: 0/-604 (-604)
Function                                     old     new   delta
hvm_set_param                               1602    1473    -129
dom0_construct_pvh                          4438    3963    -475
Total: Before=3567191, After=3566587, chg -0.02%

 xen/arch/x86/hvm/dom0_build.c | 2 +-
 xen/arch/x86/hvm/hvm.c        | 2 +-
 2 files changed, 2 insertions(+), 2 deletions(-)

diff --git a/xen/arch/x86/hvm/dom0_build.c b/xen/arch/x86/hvm/dom0_build.c
index 5551f9044836..5ac2cf8394d8 100644
--- a/xen/arch/x86/hvm/dom0_build.c
+++ b/xen/arch/x86/hvm/dom0_build.c
@@ -473,7 +473,7 @@ static int __init pvh_populate_p2m(struct domain *d)
         }
     }
 
-    if ( cpu_has_vmx && paging_mode_hap(d) && !vmx_unrestricted_guest(v) )
+    if ( using_vmx() && paging_mode_hap(d) && !vmx_unrestricted_guest(v) )
     {
         /*
          * Since Dom0 cannot be migrated, we will only setup the
diff --git a/xen/arch/x86/hvm/hvm.c b/xen/arch/x86/hvm/hvm.c
index 95a80369b9b8..70331aeec9a0 100644
--- a/xen/arch/x86/hvm/hvm.c
+++ b/xen/arch/x86/hvm/hvm.c
@@ -4302,7 +4302,7 @@ static int hvm_set_param(struct domain *d, uint32_t index, uint64_t value)
          * Only actually required for VT-x lacking unrestricted_guest
          * capabilities.  Short circuit the pause if possible.
          */
-        if ( !paging_mode_hap(d) || !cpu_has_vmx )
+        if ( !paging_mode_hap(d) || !using_vmx() )
         {
             d->arch.hvm.params[index] = value;
             break;
-- 
2.34.1
Re: [XEN][PATCH] x86/hvm: revise "cpu_has_vmx" usage for !CONFIG_INTEL_VMX case
Posted by Grygorii Strashko 2 months, 4 weeks ago 
Hi

Could it be merged if no other comments, please?

On 24.09.25 13:14, Grygorii Strashko wrote:
> From: Grygorii Strashko <grygorii_strashko@epam.com>
> 
> Since commit b99227347230 ("x86: Fix AMD_SVM and INTEL_VMX dependency") the
> HVM Intel VT-x support can be disabled, but it still keeps VMX
> code partially built-in. Particularly in HVM code there are two places:
> 
> 1) hvm/dom0_build.c
>   dom0_construct_pvh()->pvh_populate_p2m()
>      ...
>      if ( cpu_has_vmx && paging_mode_hap(d) && !vmx_unrestricted_guest(v) )
>      {
>          ...
>          [unreachable for !cpu_has_vmx case]
>          rc = pvh_setup_vmx_realmode_helpers(d);
> 
> pvh_setup_vmx_realmode_helpers() allocates memory and configures
>   HVM_PARAM_VM86_TSS_SIZED
>   HVM_PARAM_IDENT_PT
> 
> 2) hvm/hvm.c
>   hvm_set_param()
>      ...
>      case HVM_PARAM_IDENT_PT:
> 
>          if ( !paging_mode_hap(d) || !cpu_has_vmx )
>          {
>              d->arch.hvm.params[index] = value;
>              break;
>          }
>          [unreachable for !cpu_has_vmx case]
>          ...
> 
> Hence HVM_PARAM_IDENT_PT/HVM_PARAM_VM86_TSS_SIZED are used only by VMX code
> above code become unreachable in !cpu_has_vmx case and can be optimazed
> when !CONFIG_INTEL_VMX.
> 
> Replace "cpu_has_vmx" with using_vmx() to account !CONFIG_INTEL_VMX and allow
> compiler DCE to optimize code.
> 
> Signed-off-by: Grygorii Strashko <grygorii_strashko@epam.com>
> ---
> add/remove: 0/0 grow/shrink: 0/2 up/down: 0/-604 (-604)
> Function                                     old     new   delta
> hvm_set_param                               1602    1473    -129
> dom0_construct_pvh                          4438    3963    -475
> Total: Before=3567191, After=3566587, chg -0.02%
> 
>   xen/arch/x86/hvm/dom0_build.c | 2 +-
>   xen/arch/x86/hvm/hvm.c        | 2 +-
>   2 files changed, 2 insertions(+), 2 deletions(-)
> 
> diff --git a/xen/arch/x86/hvm/dom0_build.c b/xen/arch/x86/hvm/dom0_build.c
> index 5551f9044836..5ac2cf8394d8 100644
> --- a/xen/arch/x86/hvm/dom0_build.c
> +++ b/xen/arch/x86/hvm/dom0_build.c
> @@ -473,7 +473,7 @@ static int __init pvh_populate_p2m(struct domain *d)
>           }
>       }
>   
> -    if ( cpu_has_vmx && paging_mode_hap(d) && !vmx_unrestricted_guest(v) )
> +    if ( using_vmx() && paging_mode_hap(d) && !vmx_unrestricted_guest(v) )
>       {
>           /*
>            * Since Dom0 cannot be migrated, we will only setup the
> diff --git a/xen/arch/x86/hvm/hvm.c b/xen/arch/x86/hvm/hvm.c
> index 95a80369b9b8..70331aeec9a0 100644
> --- a/xen/arch/x86/hvm/hvm.c
> +++ b/xen/arch/x86/hvm/hvm.c
> @@ -4302,7 +4302,7 @@ static int hvm_set_param(struct domain *d, uint32_t index, uint64_t value)
>            * Only actually required for VT-x lacking unrestricted_guest
>            * capabilities.  Short circuit the pause if possible.
>            */
> -        if ( !paging_mode_hap(d) || !cpu_has_vmx )
> +        if ( !paging_mode_hap(d) || !using_vmx() )
>           {
>               d->arch.hvm.params[index] = value;
>               break;

-- 
Best regards,
-grygorii
Re: [XEN][PATCH] x86/hvm: revise "cpu_has_vmx" usage for !CONFIG_INTEL_VMX case
Posted by Jan Beulich 2 months, 3 weeks ago 
On 12.11.2025 21:25, Grygorii Strashko wrote:
> Could it be merged if no other comments, please?

It's on my list of things to put in. However, since you ask: You do realize that
all your recent postings were while the tree was frozen? That's okay-ish, but
when preparing a release people would preferably focus on the release, not on
what is to come afterwards. This change of yours, like quite a few from others,
will need to wait until the tree is fully open again. It having been progressed
to the point where it can in principle go in was already a courtesy, considering
the timing.

Jan

> On 24.09.25 13:14, Grygorii Strashko wrote:
>> From: Grygorii Strashko <grygorii_strashko@epam.com>
>>
>> Since commit b99227347230 ("x86: Fix AMD_SVM and INTEL_VMX dependency") the
>> HVM Intel VT-x support can be disabled, but it still keeps VMX
>> code partially built-in. Particularly in HVM code there are two places:
>>
>> 1) hvm/dom0_build.c
>>   dom0_construct_pvh()->pvh_populate_p2m()
>>      ...
>>      if ( cpu_has_vmx && paging_mode_hap(d) && !vmx_unrestricted_guest(v) )
>>      {
>>          ...
>>          [unreachable for !cpu_has_vmx case]
>>          rc = pvh_setup_vmx_realmode_helpers(d);
>>
>> pvh_setup_vmx_realmode_helpers() allocates memory and configures
>>   HVM_PARAM_VM86_TSS_SIZED
>>   HVM_PARAM_IDENT_PT
>>
>> 2) hvm/hvm.c
>>   hvm_set_param()
>>      ...
>>      case HVM_PARAM_IDENT_PT:
>>
>>          if ( !paging_mode_hap(d) || !cpu_has_vmx )
>>          {
>>              d->arch.hvm.params[index] = value;
>>              break;
>>          }
>>          [unreachable for !cpu_has_vmx case]
>>          ...
>>
>> Hence HVM_PARAM_IDENT_PT/HVM_PARAM_VM86_TSS_SIZED are used only by VMX code
>> above code become unreachable in !cpu_has_vmx case and can be optimazed
>> when !CONFIG_INTEL_VMX.
>>
>> Replace "cpu_has_vmx" with using_vmx() to account !CONFIG_INTEL_VMX and allow
>> compiler DCE to optimize code.
>>
>> Signed-off-by: Grygorii Strashko <grygorii_strashko@epam.com>
>> ---
>> add/remove: 0/0 grow/shrink: 0/2 up/down: 0/-604 (-604)
>> Function                                     old     new   delta
>> hvm_set_param                               1602    1473    -129
>> dom0_construct_pvh                          4438    3963    -475
>> Total: Before=3567191, After=3566587, chg -0.02%
>>
>>   xen/arch/x86/hvm/dom0_build.c | 2 +-
>>   xen/arch/x86/hvm/hvm.c        | 2 +-
>>   2 files changed, 2 insertions(+), 2 deletions(-)
>>
>> diff --git a/xen/arch/x86/hvm/dom0_build.c b/xen/arch/x86/hvm/dom0_build.c
>> index 5551f9044836..5ac2cf8394d8 100644
>> --- a/xen/arch/x86/hvm/dom0_build.c
>> +++ b/xen/arch/x86/hvm/dom0_build.c
>> @@ -473,7 +473,7 @@ static int __init pvh_populate_p2m(struct domain *d)
>>           }
>>       }
>>   
>> -    if ( cpu_has_vmx && paging_mode_hap(d) && !vmx_unrestricted_guest(v) )
>> +    if ( using_vmx() && paging_mode_hap(d) && !vmx_unrestricted_guest(v) )
>>       {
>>           /*
>>            * Since Dom0 cannot be migrated, we will only setup the
>> diff --git a/xen/arch/x86/hvm/hvm.c b/xen/arch/x86/hvm/hvm.c
>> index 95a80369b9b8..70331aeec9a0 100644
>> --- a/xen/arch/x86/hvm/hvm.c
>> +++ b/xen/arch/x86/hvm/hvm.c
>> @@ -4302,7 +4302,7 @@ static int hvm_set_param(struct domain *d, uint32_t index, uint64_t value)
>>            * Only actually required for VT-x lacking unrestricted_guest
>>            * capabilities.  Short circuit the pause if possible.
>>            */
>> -        if ( !paging_mode_hap(d) || !cpu_has_vmx )
>> +        if ( !paging_mode_hap(d) || !using_vmx() )
>>           {
>>               d->arch.hvm.params[index] = value;
>>               break;
>
Re: [XEN][PATCH] x86/hvm: revise "cpu_has_vmx" usage for !CONFIG_INTEL_VMX case
Posted by Grygorii Strashko 2 months, 3 weeks ago 
Hi Jan,

On 13.11.25 09:31, Jan Beulich wrote:
> On 12.11.2025 21:25, Grygorii Strashko wrote:
>> Could it be merged if no other comments, please?
> 
> It's on my list of things to put in. However, since you ask: You do realize that
> all your recent postings were while the tree was frozen? That's okay-ish, but
> when preparing a release people would preferably focus on the release, not on
> what is to come afterwards. This change of yours, like quite a few from others,
> will need to wait until the tree is fully open again. It having been progressed
> to the point where it can in principle go in was already a courtesy, considering
> the timing.
> 

I'm very sorry for disturbing you and being annoying in this regards.

> 
>> On 24.09.25 13:14, Grygorii Strashko wrote:
>>> From: Grygorii Strashko <grygorii_strashko@epam.com>
>>>
>>> Since commit b99227347230 ("x86: Fix AMD_SVM and INTEL_VMX dependency") the
>>> HVM Intel VT-x support can be disabled, but it still keeps VMX
>>> code partially built-in. Particularly in HVM code there are two places:
>>>
>>> 1) hvm/dom0_build.c
>>>    dom0_construct_pvh()->pvh_populate_p2m()
>>>       ...
>>>       if ( cpu_has_vmx && paging_mode_hap(d) && !vmx_unrestricted_guest(v) )
>>>       {
>>>           ...
>>>           [unreachable for !cpu_has_vmx case]
>>>           rc = pvh_setup_vmx_realmode_helpers(d);
>>>
>>> pvh_setup_vmx_realmode_helpers() allocates memory and configures
>>>    HVM_PARAM_VM86_TSS_SIZED
>>>    HVM_PARAM_IDENT_PT
>>>
>>> 2) hvm/hvm.c
>>>    hvm_set_param()
>>>       ...
>>>       case HVM_PARAM_IDENT_PT:
>>>
>>>           if ( !paging_mode_hap(d) || !cpu_has_vmx )
>>>           {
>>>               d->arch.hvm.params[index] = value;
>>>               break;
>>>           }
>>>           [unreachable for !cpu_has_vmx case]
>>>           ...
>>>
>>> Hence HVM_PARAM_IDENT_PT/HVM_PARAM_VM86_TSS_SIZED are used only by VMX code
>>> above code become unreachable in !cpu_has_vmx case and can be optimazed
>>> when !CONFIG_INTEL_VMX.
>>>
>>> Replace "cpu_has_vmx" with using_vmx() to account !CONFIG_INTEL_VMX and allow
>>> compiler DCE to optimize code.
>>>
>>> Signed-off-by: Grygorii Strashko <grygorii_strashko@epam.com>
>>> ---
>>> add/remove: 0/0 grow/shrink: 0/2 up/down: 0/-604 (-604)
>>> Function                                     old     new   delta
>>> hvm_set_param                               1602    1473    -129
>>> dom0_construct_pvh                          4438    3963    -475
>>> Total: Before=3567191, After=3566587, chg -0.02%
>>>
>>>    xen/arch/x86/hvm/dom0_build.c | 2 +-
>>>    xen/arch/x86/hvm/hvm.c        | 2 +-
>>>    2 files changed, 2 insertions(+), 2 deletions(-)
>>>
>>> diff --git a/xen/arch/x86/hvm/dom0_build.c b/xen/arch/x86/hvm/dom0_build.c
>>> index 5551f9044836..5ac2cf8394d8 100644
>>> --- a/xen/arch/x86/hvm/dom0_build.c
>>> +++ b/xen/arch/x86/hvm/dom0_build.c
>>> @@ -473,7 +473,7 @@ static int __init pvh_populate_p2m(struct domain *d)
>>>            }
>>>        }
>>>    
>>> -    if ( cpu_has_vmx && paging_mode_hap(d) && !vmx_unrestricted_guest(v) )
>>> +    if ( using_vmx() && paging_mode_hap(d) && !vmx_unrestricted_guest(v) )
>>>        {
>>>            /*
>>>             * Since Dom0 cannot be migrated, we will only setup the
>>> diff --git a/xen/arch/x86/hvm/hvm.c b/xen/arch/x86/hvm/hvm.c
>>> index 95a80369b9b8..70331aeec9a0 100644
>>> --- a/xen/arch/x86/hvm/hvm.c
>>> +++ b/xen/arch/x86/hvm/hvm.c
>>> @@ -4302,7 +4302,7 @@ static int hvm_set_param(struct domain *d, uint32_t index, uint64_t value)
>>>             * Only actually required for VT-x lacking unrestricted_guest
>>>             * capabilities.  Short circuit the pause if possible.
>>>             */
>>> -        if ( !paging_mode_hap(d) || !cpu_has_vmx )
>>> +        if ( !paging_mode_hap(d) || !using_vmx() )
>>>            {
>>>                d->arch.hvm.params[index] = value;
>>>                break;
>>
> 

-- 
Best regards,
-grygorii
Re: [XEN][PATCH] x86/hvm: revise "cpu_has_vmx" usage for !CONFIG_INTEL_VMX case
Posted by Jan Beulich 4 months ago 
On 24.09.2025 12:14, Grygorii Strashko wrote:
> From: Grygorii Strashko <grygorii_strashko@epam.com>
> 
> Since commit b99227347230 ("x86: Fix AMD_SVM and INTEL_VMX dependency") the
> HVM Intel VT-x support can be disabled, but it still keeps VMX
> code partially built-in. Particularly in HVM code there are two places:
> 
> 1) hvm/dom0_build.c
>  dom0_construct_pvh()->pvh_populate_p2m()
>     ...
>     if ( cpu_has_vmx && paging_mode_hap(d) && !vmx_unrestricted_guest(v) )
>     {
>         ...
>         [unreachable for !cpu_has_vmx case]
>         rc = pvh_setup_vmx_realmode_helpers(d);
> 
> pvh_setup_vmx_realmode_helpers() allocates memory and configures
>  HVM_PARAM_VM86_TSS_SIZED
>  HVM_PARAM_IDENT_PT
> 
> 2) hvm/hvm.c
>  hvm_set_param()
>     ...
>     case HVM_PARAM_IDENT_PT:
> 
>         if ( !paging_mode_hap(d) || !cpu_has_vmx )
>         {
>             d->arch.hvm.params[index] = value;
>             break;
>         }
>         [unreachable for !cpu_has_vmx case]
>         ...
> 
> Hence HVM_PARAM_IDENT_PT/HVM_PARAM_VM86_TSS_SIZED are used only by VMX code
> above code become unreachable in !cpu_has_vmx case and can be optimazed
> when !CONFIG_INTEL_VMX.
> 
> Replace "cpu_has_vmx" with using_vmx() to account !CONFIG_INTEL_VMX and allow
> compiler DCE to optimize code.
> 
> Signed-off-by: Grygorii Strashko <grygorii_strashko@epam.com>

Reviewed-by: Jan Beulich <jbeulich@suse.com>
Re: [XEN][PATCH] x86/hvm: revise "cpu_has_vmx" usage for !CONFIG_INTEL_VMX case
Posted by Alejandro Vallejo 4 months, 1 week ago 
On Wed Sep 24, 2025 at 12:14 PM CEST, Grygorii Strashko wrote:
> From: Grygorii Strashko <grygorii_strashko@epam.com>
>
> Since commit b99227347230 ("x86: Fix AMD_SVM and INTEL_VMX dependency") the
> HVM Intel VT-x support can be disabled, but it still keeps VMX
> code partially built-in. Particularly in HVM code there are two places:
>
> 1) hvm/dom0_build.c
>  dom0_construct_pvh()->pvh_populate_p2m()
>     ...
>     if ( cpu_has_vmx && paging_mode_hap(d) && !vmx_unrestricted_guest(v) )
>     {
>         ...
>         [unreachable for !cpu_has_vmx case]
>         rc = pvh_setup_vmx_realmode_helpers(d);
>
> pvh_setup_vmx_realmode_helpers() allocates memory and configures
>  HVM_PARAM_VM86_TSS_SIZED
>  HVM_PARAM_IDENT_PT
>
> 2) hvm/hvm.c
>  hvm_set_param()
>     ...
>     case HVM_PARAM_IDENT_PT:
>
>         if ( !paging_mode_hap(d) || !cpu_has_vmx )
>         {
>             d->arch.hvm.params[index] = value;
>             break;
>         }
>         [unreachable for !cpu_has_vmx case]
>         ...

nit: These (1) and (2) are rather large for a commit message. I wouldn't mind
if they went away and the rest of the commit message was adjusted to make it
a bit leaner.

Either way, with or without this change...

>
> Hence HVM_PARAM_IDENT_PT/HVM_PARAM_VM86_TSS_SIZED are used only by VMX code
> above code become unreachable in !cpu_has_vmx case and can be optimazed
> when !CONFIG_INTEL_VMX.
>
> Replace "cpu_has_vmx" with using_vmx() to account !CONFIG_INTEL_VMX and allow
> compiler DCE to optimize code.
>
> Signed-off-by: Grygorii Strashko <grygorii_strashko@epam.com>

... Reviewed-by: Alejandro Vallejo <alejandro.garciavallejo@amd.com>

Cheers,
Alejandro

Patchew 2.1-devel-b67e4c2

© 2016 - 2026 Red Hat, Inc.