LLC coloring can be used only on MMU system, move the code
that selects it from ARM_64 to MMU and add the ARM_64
dependency.

Signed-off-by: Luca Fancellu <luca.fancellu@arm.com>
---
v3 changes:
 - Dropped comment on top of boot_fdt_info()

v2 changes:
 - dropped part of the v1 code, now this one is simpler, I will
   discuss better how to design a common boot flow for MPU and
   implement on another patch.

---
---
 xen/arch/arm/Kconfig | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/xen/arch/arm/Kconfig b/xen/arch/arm/Kconfig
index a26d3e11827c..ffdff1f0a36c 100644
--- a/xen/arch/arm/Kconfig
+++ b/xen/arch/arm/Kconfig
@@ -8,7 +8,6 @@ config ARM_64
 	depends on !ARM_32
 	select 64BIT
 	select HAS_FAST_MULTIPLY
-	select HAS_LLC_COLORING if !NUMA
 
 config ARM
 	def_bool y
@@ -76,6 +75,7 @@ choice
 
 config MMU
 	bool "MMU"
+	select HAS_LLC_COLORING if !NUMA && ARM_64
 	select HAS_PMAP
 	select HAS_VMAP
 	select HAS_PASSTHROUGH
-- 
2.34.1

On 18.02.2025 10:51, Luca Fancellu wrote:
> LLC coloring can be used only on MMU system,

Just for my own education: Why is this?

Jan

On 18/02/2025 14:44, Jan Beulich wrote:
> 
> 
> On 18.02.2025 10:51, Luca Fancellu wrote:
>> LLC coloring can be used only on MMU system,
> 
> Just for my own education: Why is this?
I read this as it refers to the feature we have in Xen, not the generic concept.
You could in theory achieve cache coloring without MMU using static allocation
or some custom allocator but in general we benefit from being able to map VA
to PA aligning with cache coloring scheme.

P.S. Personally, I've never seen a solution without MMU.

~Michal

On 18.02.2025 16:01, Orzel, Michal wrote:
> On 18/02/2025 14:44, Jan Beulich wrote:
>> On 18.02.2025 10:51, Luca Fancellu wrote:
>>> LLC coloring can be used only on MMU system,
>>
>> Just for my own education: Why is this?
> I read this as it refers to the feature we have in Xen, not the generic concept.
> You could in theory achieve cache coloring without MMU using static allocation
> or some custom allocator but in general we benefit from being able to map VA
> to PA aligning with cache coloring scheme.

Oh, okay. Then maybe the sentence would better be worded to express that
it's our present choice, rather than an inherent limitation?

Jan

Hi,

> On 18 Feb 2025, at 15:06, Jan Beulich <jbeulich@suse.com> wrote:
> 
> On 18.02.2025 16:01, Orzel, Michal wrote:
>> On 18/02/2025 14:44, Jan Beulich wrote:
>>> On 18.02.2025 10:51, Luca Fancellu wrote:
>>>> LLC coloring can be used only on MMU system,
>>> 
>>> Just for my own education: Why is this?
>> I read this as it refers to the feature we have in Xen, not the generic concept.
>> You could in theory achieve cache coloring without MMU using static allocation
>> or some custom allocator but in general we benefit from being able to map VA
>> to PA aligning with cache coloring scheme.
> 
> Oh, okay. Then maybe the sentence would better be worded to express that
> it's our present choice, rather than an inherent limitation?

Is this better? Can be addressed on commit?

Xen LLC coloring feature can be used only on MMU system, move the code
that selects it from ARM_64 to MMU and add the ARM_64 dependency.


> 
> Jan
> 

On Tue, 4 Mar 2025, Luca Fancellu wrote:
> Hi,
> 
> > On 18 Feb 2025, at 15:06, Jan Beulich <jbeulich@suse.com> wrote:
> > 
> > On 18.02.2025 16:01, Orzel, Michal wrote:
> >> On 18/02/2025 14:44, Jan Beulich wrote:
> >>> On 18.02.2025 10:51, Luca Fancellu wrote:
> >>>> LLC coloring can be used only on MMU system,
> >>> 
> >>> Just for my own education: Why is this?
> >> I read this as it refers to the feature we have in Xen, not the generic concept.
> >> You could in theory achieve cache coloring without MMU using static allocation
> >> or some custom allocator but in general we benefit from being able to map VA
> >> to PA aligning with cache coloring scheme.
> > 
> > Oh, okay. Then maybe the sentence would better be worded to express that
> > it's our present choice, rather than an inherent limitation?
> 
> Is this better? Can be addressed on commit?
> 
> Xen LLC coloring feature can be used only on MMU system, move the code
> that selects it from ARM_64 to MMU and add the ARM_64 dependency.

Hi Luca, I think that's OK. Can this patch be committed without the
previous patch in the series?

Hi Stefano,

> On 4 Mar 2025, at 19:20, Stefano Stabellini <sstabellini@kernel.org> wrote:
> 
> On Tue, 4 Mar 2025, Luca Fancellu wrote:
>> Hi,
>> 
>>> On 18 Feb 2025, at 15:06, Jan Beulich <jbeulich@suse.com> wrote:
>>> 
>>> On 18.02.2025 16:01, Orzel, Michal wrote:
>>>> On 18/02/2025 14:44, Jan Beulich wrote:
>>>>> On 18.02.2025 10:51, Luca Fancellu wrote:
>>>>>> LLC coloring can be used only on MMU system,
>>>>> 
>>>>> Just for my own education: Why is this?
>>>> I read this as it refers to the feature we have in Xen, not the generic concept.
>>>> You could in theory achieve cache coloring without MMU using static allocation
>>>> or some custom allocator but in general we benefit from being able to map VA
>>>> to PA aligning with cache coloring scheme.
>>> 
>>> Oh, okay. Then maybe the sentence would better be worded to express that
>>> it's our present choice, rather than an inherent limitation?
>> 
>> Is this better? Can be addressed on commit?
>> 
>> Xen LLC coloring feature can be used only on MMU system, move the code
>> that selects it from ARM_64 to MMU and add the ARM_64 dependency.
> 
> Hi Luca, I think that's OK. Can this patch be committed without the
> previous patch in the series?

Yes it can, I think we need more time to get an agreement for the first patch,
probably Jan would like to see that together with the rest in order to understand
if it goes to the right path or not.

Cheers,
Luca

On 18/02/2025 10:51, Luca Fancellu wrote:
> 
> 
> LLC coloring can be used only on MMU system, move the code
> that selects it from ARM_64 to MMU and add the ARM_64
> dependency.
> 
> Signed-off-by: Luca Fancellu <luca.fancellu@arm.com>
Reviewed-by: Michal Orzel <michal.orzel@amd.com>

~Michal