In case of ARM_32, all of the RAM will be covered by a permanent contiguous
mapping (where VA == PA) and there will be a single heap. Thus, the memory
allocated from Xen heap uses PGC_xen_heap.
This is similar to the scenario described for
"CONFIG_SEPARATE_XENHEAP=n W/ DIRECT MAP OF ALL RAM" in common/page_alloc.c.

Signed-off-by: Ayan Kumar Halder <ayan.kumar.halder@amd.com>
---
Changes from :-

v1 - 1. Enable is_xen_heap_page() and is_xen_heap_mfn() definitions for ARM_64
or MPU (fixed a build break).

 xen/arch/arm/include/asm/mm.h     | 2 +-
 xen/arch/arm/include/asm/mpu/mm.h | 5 -----
 2 files changed, 1 insertion(+), 6 deletions(-)

diff --git a/xen/arch/arm/include/asm/mm.h b/xen/arch/arm/include/asm/mm.h
index fb79aeb088..7a93dad2ed 100644
--- a/xen/arch/arm/include/asm/mm.h
+++ b/xen/arch/arm/include/asm/mm.h
@@ -170,7 +170,7 @@ struct page_info
 #define _PGC_need_scrub   _PGC_allocated
 #define PGC_need_scrub    PGC_allocated
 
-#ifdef CONFIG_ARM_64
+#if defined(CONFIG_ARM_64) || defined(CONFIG_MPU)
 #define is_xen_heap_page(page) ((page)->count_info & PGC_xen_heap)
 #define is_xen_heap_mfn(mfn) \
     (mfn_valid(mfn) && is_xen_heap_page(mfn_to_page(mfn)))
diff --git a/xen/arch/arm/include/asm/mpu/mm.h b/xen/arch/arm/include/asm/mpu/mm.h
index c32fac8905..e1ded6521d 100644
--- a/xen/arch/arm/include/asm/mpu/mm.h
+++ b/xen/arch/arm/include/asm/mpu/mm.h
@@ -27,11 +27,6 @@ extern pr_t xen_mpumap[MAX_MPU_REGION_NR];
 
 #define virt_to_maddr(va) ((paddr_t)((vaddr_t)(va) & PADDR_MASK))
 
-#ifdef CONFIG_ARM_32
-#define is_xen_heap_page(page) ({ BUG_ON("unimplemented"); false; })
-#define is_xen_heap_mfn(mfn) ({ BUG_ON("unimplemented"); false; })
-#endif
-
 /* On MPU systems there is no translation, ma == va. */
 static inline void *maddr_to_virt(paddr_t ma)
 {
-- 
2.25.1

On 14/08/2025 10:19, Ayan Kumar Halder wrote:
> In case of ARM_32, all of the RAM will be covered by a permanent contiguous
> mapping (where VA == PA) and there will be a single heap. Thus, the memory
> allocated from Xen heap uses PGC_xen_heap.
> This is similar to the scenario described for
> "CONFIG_SEPARATE_XENHEAP=n W/ DIRECT MAP OF ALL RAM" in common/page_alloc.c.
> 
> Signed-off-by: Ayan Kumar Halder <ayan.kumar.halder@amd.com>
Reviewed-by: Michal Orzel <michal.orzel@amd.com>

~Michal

Hi Ayan,

On Thu, Aug 14, 2025 at 09:19:58AM +0000, Ayan Kumar Halder wrote:
> In case of ARM_32, all of the RAM will be covered by a permanent contiguous
> mapping (where VA == PA) and there will be a single heap. Thus, the memory
> allocated from Xen heap uses PGC_xen_heap.
> This is similar to the scenario described for
> "CONFIG_SEPARATE_XENHEAP=n W/ DIRECT MAP OF ALL RAM" in common/page_alloc.c.
> 
> Signed-off-by: Ayan Kumar Halder <ayan.kumar.halder@amd.com>
Reviewed-by: Hari Limaye <hari.limaye@arm.com>
Tested-by: Hari Limaye <hari.limaye@arm.com>

LGTM now, thanks for updating. Tested (build-only) on AArch32 and
AArch64 MMU and MPU systems.

Cheers,
Hari