* [RFC v1 1/2] drivers/vfio_pci_core: Change PXD_ORDER check from switch case to if/else block
@ 2026-02-27 6:16 Ritesh Harjani (IBM)
2026-02-27 6:16 ` [RFC v1 2/2] powerpc/64s: Add support for huge pfnmaps Ritesh Harjani (IBM)
2026-02-27 6:42 ` [RFC v1 1/2] drivers/vfio_pci_core: Change PXD_ORDER check from switch case to if/else block Christophe Leroy (CS GROUP)
0 siblings, 2 replies; 6+ messages in thread
From: Ritesh Harjani (IBM) @ 2026-02-27 6:16 UTC (permalink / raw)
To: linuxppc-dev
Cc: linux-mm, kvm, Alex Williamson, Peter Xu, Ritesh Harjani (IBM)
Architectures like PowerPC uses runtime defined values for
PMD_ORDER/PUD_ORDER. This is because it can use either RADIX or HASH MMU
at runtime using kernel cmdline. So the pXd_index_size is not known at
compile time. Without this fix, when we add huge pfn support on powerpc
in the next patch, vfio_pci_core driver compilation can fail with the
following errors.
CC [M] drivers/vfio/vfio_main.o
CC [M] drivers/vfio/group.o
CC [M] drivers/vfio/container.o
CC [M] drivers/vfio/virqfd.o
CC [M] drivers/vfio/vfio_iommu_spapr_tce.o
CC [M] drivers/vfio/pci/vfio_pci_core.o
CC [M] drivers/vfio/pci/vfio_pci_intrs.o
CC [M] drivers/vfio/pci/vfio_pci_rdwr.o
CC [M] drivers/vfio/pci/vfio_pci_config.o
CC [M] drivers/vfio/pci/vfio_pci.o
AR kernel/built-in.a
../drivers/vfio/pci/vfio_pci_core.c: In function ‘vfio_pci_vmf_insert_pfn’:
../drivers/vfio/pci/vfio_pci_core.c:1678:9: error: case label does not reduce to an integer constant
1678 | case PMD_ORDER:
| ^~~~
../drivers/vfio/pci/vfio_pci_core.c:1682:9: error: case label does not reduce to an integer constant
1682 | case PUD_ORDER:
| ^~~~
make[6]: *** [../scripts/Makefile.build:289: drivers/vfio/pci/vfio_pci_core.o] Error 1
make[6]: *** Waiting for unfinished jobs....
make[5]: *** [../scripts/Makefile.build:546: drivers/vfio/pci] Error 2
make[5]: *** Waiting for unfinished jobs....
make[4]: *** [../scripts/Makefile.build:546: drivers/vfio] Error 2
make[3]: *** [../scripts/Makefile.build:546: drivers] Error 2
Fixes: f9e54c3a2f5b7 ("vfio/pci: implement huge_fault support")
Signed-off-by: Ritesh Harjani (IBM) <ritesh.list@gmail.com>
---
drivers/vfio/pci/vfio_pci_core.c | 15 +++++++--------
1 file changed, 7 insertions(+), 8 deletions(-)
diff --git a/drivers/vfio/pci/vfio_pci_core.c b/drivers/vfio/pci/vfio_pci_core.c
index d43745fe4c84..5395a6f30904 100644
--- a/drivers/vfio/pci/vfio_pci_core.c
+++ b/drivers/vfio/pci/vfio_pci_core.c
@@ -1670,21 +1670,20 @@ vm_fault_t vfio_pci_vmf_insert_pfn(struct vfio_pci_core_device *vdev,
if (vdev->pm_runtime_engaged || !__vfio_pci_memory_enabled(vdev))
return VM_FAULT_SIGBUS;
- switch (order) {
- case 0:
+ if (order == 0) {
return vmf_insert_pfn(vmf->vma, vmf->address, pfn);
+ }
#ifdef CONFIG_ARCH_SUPPORTS_PMD_PFNMAP
- case PMD_ORDER:
+ else if (order == PMD_ORDER) {
return vmf_insert_pfn_pmd(vmf, pfn, false);
+ }
#endif
#ifdef CONFIG_ARCH_SUPPORTS_PUD_PFNMAP
- case PUD_ORDER:
+ else if (order == PUD_ORDER) {
return vmf_insert_pfn_pud(vmf, pfn, false);
- break;
+ }
#endif
- default:
- return VM_FAULT_FALLBACK;
- }
+ return VM_FAULT_FALLBACK;
}
EXPORT_SYMBOL_GPL(vfio_pci_vmf_insert_pfn);
--
2.53.0
^ permalink raw reply [flat|nested] 6+ messages in thread* [RFC v1 2/2] powerpc/64s: Add support for huge pfnmaps
2026-02-27 6:16 [RFC v1 1/2] drivers/vfio_pci_core: Change PXD_ORDER check from switch case to if/else block Ritesh Harjani (IBM)
@ 2026-02-27 6:16 ` Ritesh Harjani (IBM)
2026-02-27 6:47 ` Christophe Leroy (CS GROUP)
2026-02-27 6:42 ` [RFC v1 1/2] drivers/vfio_pci_core: Change PXD_ORDER check from switch case to if/else block Christophe Leroy (CS GROUP)
1 sibling, 1 reply; 6+ messages in thread
From: Ritesh Harjani (IBM) @ 2026-02-27 6:16 UTC (permalink / raw)
To: linuxppc-dev
Cc: linux-mm, kvm, Alex Williamson, Peter Xu, Ritesh Harjani (IBM)
This uses _RPAGE_SW2 bit for the PMD and PUDs similar to PTEs.
This also adds support for {pte,pmd,pud}_pgprot helpers needed for
follow_pfnmap APIs.
This allows us to extend the PFN mappings, e.g. PCI MMIO bars where
it can grow as large as 8GB or even bigger, to map at PMD / PUD level.
VFIO PCI core driver already supports fault handling at PMD / PUD level
for more efficient BAR mappings.
Signed-off-by: Ritesh Harjani (IBM) <ritesh.list@gmail.com>
---
@linux-mm:
Is there any official test which I could use to verify this functionality.
For now I used basic ivshmem setup + vfio using Qemu and validated using some
basic test to see that we are seeing these prints.
[ 4351.435050] vfio_pci_mmap_huge_fault: 3 callbacks suppressed
[ 4351.435234] vfio-pci 0001:00:00.0: vfio_pci_mmap_huge_fault(,order = 5) BAR 2 page offset 0x0: 0x100
[ 4351.457005] vfio-pci 0001:00:00.0: vfio_pci_mmap_huge_fault(,order = 5) BAR 2 page offset 0x40: 0x100
[ 4351.463684] vfio-pci 0001:00:00.0: vfio_pci_mmap_huge_fault(,order = 5) BAR 2 page offset 0x20: 0x100
arch/powerpc/Kconfig | 1 +
arch/powerpc/include/asm/book3s/64/pgtable.h | 23 ++++++++++++++++++++
arch/powerpc/include/asm/pgtable.h | 12 ++++++++++
3 files changed, 36 insertions(+)
diff --git a/arch/powerpc/Kconfig b/arch/powerpc/Kconfig
index ad7a2fe63a2a..cf9283757e5d 100644
--- a/arch/powerpc/Kconfig
+++ b/arch/powerpc/Kconfig
@@ -172,6 +172,7 @@ config PPC
select ARCH_STACKWALK
select ARCH_SUPPORTS_ATOMIC_RMW
select ARCH_SUPPORTS_DEBUG_PAGEALLOC if PPC_BOOK3S || PPC_8xx
+ select ARCH_SUPPORTS_HUGE_PFNMAP if PPC_BOOK3S_64 && TRANSPARENT_HUGEPAGE
select ARCH_SUPPORTS_PAGE_TABLE_CHECK if !HUGETLB_PAGE
select ARCH_SUPPORTS_SCHED_MC if SMP
select ARCH_SUPPORTS_SCHED_SMT if PPC64 && SMP
diff --git a/arch/powerpc/include/asm/book3s/64/pgtable.h b/arch/powerpc/include/asm/book3s/64/pgtable.h
index 1a91762b455d..639cbf34f752 100644
--- a/arch/powerpc/include/asm/book3s/64/pgtable.h
+++ b/arch/powerpc/include/asm/book3s/64/pgtable.h
@@ -1289,6 +1289,29 @@ static inline pud_t pud_mkhuge(pud_t pud)
return pud;
}
+#ifdef CONFIG_ARCH_SUPPORTS_PMD_PFNMAP
+static inline bool pmd_special(pmd_t pmd)
+{
+ return pte_special(pmd_pte(pmd));
+}
+
+static inline pmd_t pmd_mkspecial(pmd_t pmd)
+{
+ return pte_pmd(pte_mkspecial(pmd_pte(pmd)));
+}
+#endif
+
+#ifdef CONFIG_ARCH_SUPPORTS_PUD_PFNMAP
+static inline bool pud_special(pud_t pud)
+{
+ return pte_special(pud_pte(pud));
+}
+
+static inline pud_t pud_mkspecial(pud_t pud)
+{
+ return pte_pud(pte_mkspecial(pud_pte(pud)));
+}
+#endif
#define __HAVE_ARCH_PMDP_SET_ACCESS_FLAGS
extern int pmdp_set_access_flags(struct vm_area_struct *vma,
diff --git a/arch/powerpc/include/asm/pgtable.h b/arch/powerpc/include/asm/pgtable.h
index dcd3a88caaf6..2d27cb1c2334 100644
--- a/arch/powerpc/include/asm/pgtable.h
+++ b/arch/powerpc/include/asm/pgtable.h
@@ -63,6 +63,18 @@ static inline pgprot_t pte_pgprot(pte_t pte)
return __pgprot(pte_flags);
}
+#define pmd_pgprot pmd_pgprot
+static inline pgprot_t pmd_pgprot(pmd_t pmd)
+{
+ return pte_pgprot(pmd_pte(pmd));
+}
+
+#define pud_pgprot pud_pgprot
+static inline pgprot_t pud_pgprot(pud_t pud)
+{
+ return pte_pgprot(pud_pte(pud));
+}
+
static inline pgprot_t pgprot_nx(pgprot_t prot)
{
return pte_pgprot(pte_exprotect(__pte(pgprot_val(prot))));
--
2.53.0
^ permalink raw reply [flat|nested] 6+ messages in thread* Re: [RFC v1 2/2] powerpc/64s: Add support for huge pfnmaps
2026-02-27 6:16 ` [RFC v1 2/2] powerpc/64s: Add support for huge pfnmaps Ritesh Harjani (IBM)
@ 2026-02-27 6:47 ` Christophe Leroy (CS GROUP)
2026-02-27 10:32 ` Ritesh Harjani
0 siblings, 1 reply; 6+ messages in thread
From: Christophe Leroy (CS GROUP) @ 2026-02-27 6:47 UTC (permalink / raw)
To: Ritesh Harjani (IBM), linuxppc-dev
Cc: linux-mm, kvm, Alex Williamson, Peter Xu
Le 27/02/2026 à 07:16, Ritesh Harjani (IBM) a écrit :
> This uses _RPAGE_SW2 bit for the PMD and PUDs similar to PTEs.
> This also adds support for {pte,pmd,pud}_pgprot helpers needed for
> follow_pfnmap APIs.
>
> This allows us to extend the PFN mappings, e.g. PCI MMIO bars where
> it can grow as large as 8GB or even bigger, to map at PMD / PUD level.
> VFIO PCI core driver already supports fault handling at PMD / PUD level
> for more efficient BAR mappings.
>
> Signed-off-by: Ritesh Harjani (IBM) <ritesh.list@gmail.com>
Reviewed-by: Christophe Leroy (CS GROUP) <chleroy@kernel.org>
> ---
>
> @linux-mm:
> Is there any official test which I could use to verify this functionality.
>
> For now I used basic ivshmem setup + vfio using Qemu and validated using some
> basic test to see that we are seeing these prints.
>
> [ 4351.435050] vfio_pci_mmap_huge_fault: 3 callbacks suppressed
> [ 4351.435234] vfio-pci 0001:00:00.0: vfio_pci_mmap_huge_fault(,order = 5) BAR 2 page offset 0x0: 0x100
> [ 4351.457005] vfio-pci 0001:00:00.0: vfio_pci_mmap_huge_fault(,order = 5) BAR 2 page offset 0x40: 0x100
> [ 4351.463684] vfio-pci 0001:00:00.0: vfio_pci_mmap_huge_fault(,order = 5) BAR 2 page offset 0x20: 0x100
>
> arch/powerpc/Kconfig | 1 +
> arch/powerpc/include/asm/book3s/64/pgtable.h | 23 ++++++++++++++++++++
> arch/powerpc/include/asm/pgtable.h | 12 ++++++++++
> 3 files changed, 36 insertions(+)
>
> diff --git a/arch/powerpc/Kconfig b/arch/powerpc/Kconfig
> index ad7a2fe63a2a..cf9283757e5d 100644
> --- a/arch/powerpc/Kconfig
> +++ b/arch/powerpc/Kconfig
> @@ -172,6 +172,7 @@ config PPC
> select ARCH_STACKWALK
> select ARCH_SUPPORTS_ATOMIC_RMW
> select ARCH_SUPPORTS_DEBUG_PAGEALLOC if PPC_BOOK3S || PPC_8xx
> + select ARCH_SUPPORTS_HUGE_PFNMAP if PPC_BOOK3S_64 && TRANSPARENT_HUGEPAGE
> select ARCH_SUPPORTS_PAGE_TABLE_CHECK if !HUGETLB_PAGE
> select ARCH_SUPPORTS_SCHED_MC if SMP
> select ARCH_SUPPORTS_SCHED_SMT if PPC64 && SMP
> diff --git a/arch/powerpc/include/asm/book3s/64/pgtable.h b/arch/powerpc/include/asm/book3s/64/pgtable.h
> index 1a91762b455d..639cbf34f752 100644
> --- a/arch/powerpc/include/asm/book3s/64/pgtable.h
> +++ b/arch/powerpc/include/asm/book3s/64/pgtable.h
> @@ -1289,6 +1289,29 @@ static inline pud_t pud_mkhuge(pud_t pud)
> return pud;
> }
>
> +#ifdef CONFIG_ARCH_SUPPORTS_PMD_PFNMAP
> +static inline bool pmd_special(pmd_t pmd)
> +{
> + return pte_special(pmd_pte(pmd));
> +}
> +
> +static inline pmd_t pmd_mkspecial(pmd_t pmd)
> +{
> + return pte_pmd(pte_mkspecial(pmd_pte(pmd)));
> +}
> +#endif
> +
> +#ifdef CONFIG_ARCH_SUPPORTS_PUD_PFNMAP
> +static inline bool pud_special(pud_t pud)
> +{
> + return pte_special(pud_pte(pud));
> +}
> +
> +static inline pud_t pud_mkspecial(pud_t pud)
> +{
> + return pte_pud(pte_mkspecial(pud_pte(pud)));
> +}
> +#endif
>
> #define __HAVE_ARCH_PMDP_SET_ACCESS_FLAGS
> extern int pmdp_set_access_flags(struct vm_area_struct *vma,
> diff --git a/arch/powerpc/include/asm/pgtable.h b/arch/powerpc/include/asm/pgtable.h
> index dcd3a88caaf6..2d27cb1c2334 100644
> --- a/arch/powerpc/include/asm/pgtable.h
> +++ b/arch/powerpc/include/asm/pgtable.h
> @@ -63,6 +63,18 @@ static inline pgprot_t pte_pgprot(pte_t pte)
> return __pgprot(pte_flags);
> }
>
> +#define pmd_pgprot pmd_pgprot
> +static inline pgprot_t pmd_pgprot(pmd_t pmd)
> +{
> + return pte_pgprot(pmd_pte(pmd));
> +}
> +
> +#define pud_pgprot pud_pgprot
> +static inline pgprot_t pud_pgprot(pud_t pud)
> +{
> + return pte_pgprot(pud_pte(pud));
> +}
> +
> static inline pgprot_t pgprot_nx(pgprot_t prot)
> {
> return pte_pgprot(pte_exprotect(__pte(pgprot_val(prot))));
> --
> 2.53.0
>
>
^ permalink raw reply [flat|nested] 6+ messages in thread* Re: [RFC v1 2/2] powerpc/64s: Add support for huge pfnmaps
2026-02-27 6:47 ` Christophe Leroy (CS GROUP)
@ 2026-02-27 10:32 ` Ritesh Harjani
0 siblings, 0 replies; 6+ messages in thread
From: Ritesh Harjani @ 2026-02-27 10:32 UTC (permalink / raw)
To: Christophe Leroy (CS GROUP), linuxppc-dev
Cc: linux-mm, kvm, Alex Williamson, Peter Xu
"Christophe Leroy (CS GROUP)" <chleroy@kernel.org> writes:
> Le 27/02/2026 à 07:16, Ritesh Harjani (IBM) a écrit :
>> This uses _RPAGE_SW2 bit for the PMD and PUDs similar to PTEs.
>> This also adds support for {pte,pmd,pud}_pgprot helpers needed for
>> follow_pfnmap APIs.
>>
>> This allows us to extend the PFN mappings, e.g. PCI MMIO bars where
>> it can grow as large as 8GB or even bigger, to map at PMD / PUD level.
>> VFIO PCI core driver already supports fault handling at PMD / PUD level
>> for more efficient BAR mappings.
>>
>> Signed-off-by: Ritesh Harjani (IBM) <ritesh.list@gmail.com>
>
> Reviewed-by: Christophe Leroy (CS GROUP) <chleroy@kernel.org>
>
>
Thanks for the review!
>> #define __HAVE_ARCH_PMDP_SET_ACCESS_FLAGS
>> extern int pmdp_set_access_flags(struct vm_area_struct *vma,
>> diff --git a/arch/powerpc/include/asm/pgtable.h b/arch/powerpc/include/asm/pgtable.h
>> index dcd3a88caaf6..2d27cb1c2334 100644
>> --- a/arch/powerpc/include/asm/pgtable.h
>> +++ b/arch/powerpc/include/asm/pgtable.h
>> @@ -63,6 +63,18 @@ static inline pgprot_t pte_pgprot(pte_t pte)
>> return __pgprot(pte_flags);
>> }
>>
>> +#define pmd_pgprot pmd_pgprot
>> +static inline pgprot_t pmd_pgprot(pmd_t pmd)
>> +{
>> + return pte_pgprot(pmd_pte(pmd));
>> +}
>> +
>> +#define pud_pgprot pud_pgprot
>> +static inline pgprot_t pud_pgprot(pud_t pud)
>> +{
>> + return pte_pgprot(pud_pte(pud));
>> +}
>> +
In v2 - I will add above under #ifdef CONFIG_PPC_BOOK3S_64
to avoid build issues with 32-bit PPC.
-ritesh
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [RFC v1 1/2] drivers/vfio_pci_core: Change PXD_ORDER check from switch case to if/else block
2026-02-27 6:16 [RFC v1 1/2] drivers/vfio_pci_core: Change PXD_ORDER check from switch case to if/else block Ritesh Harjani (IBM)
2026-02-27 6:16 ` [RFC v1 2/2] powerpc/64s: Add support for huge pfnmaps Ritesh Harjani (IBM)
@ 2026-02-27 6:42 ` Christophe Leroy (CS GROUP)
2026-02-27 10:30 ` Ritesh Harjani
1 sibling, 1 reply; 6+ messages in thread
From: Christophe Leroy (CS GROUP) @ 2026-02-27 6:42 UTC (permalink / raw)
To: Ritesh Harjani (IBM), linuxppc-dev
Cc: linux-mm, kvm, Alex Williamson, Peter Xu
Le 27/02/2026 à 07:16, Ritesh Harjani (IBM) a écrit :
> Architectures like PowerPC uses runtime defined values for
> PMD_ORDER/PUD_ORDER. This is because it can use either RADIX or HASH MMU
> at runtime using kernel cmdline. So the pXd_index_size is not known at
> compile time. Without this fix, when we add huge pfn support on powerpc
> in the next patch, vfio_pci_core driver compilation can fail with the
> following errors.
>
> CC [M] drivers/vfio/vfio_main.o
> CC [M] drivers/vfio/group.o
> CC [M] drivers/vfio/container.o
> CC [M] drivers/vfio/virqfd.o
> CC [M] drivers/vfio/vfio_iommu_spapr_tce.o
> CC [M] drivers/vfio/pci/vfio_pci_core.o
> CC [M] drivers/vfio/pci/vfio_pci_intrs.o
> CC [M] drivers/vfio/pci/vfio_pci_rdwr.o
> CC [M] drivers/vfio/pci/vfio_pci_config.o
> CC [M] drivers/vfio/pci/vfio_pci.o
> AR kernel/built-in.a
> ../drivers/vfio/pci/vfio_pci_core.c: In function ‘vfio_pci_vmf_insert_pfn’:
> ../drivers/vfio/pci/vfio_pci_core.c:1678:9: error: case label does not reduce to an integer constant
> 1678 | case PMD_ORDER:
> | ^~~~
> ../drivers/vfio/pci/vfio_pci_core.c:1682:9: error: case label does not reduce to an integer constant
> 1682 | case PUD_ORDER:
> | ^~~~
> make[6]: *** [../scripts/Makefile.build:289: drivers/vfio/pci/vfio_pci_core.o] Error 1
> make[6]: *** Waiting for unfinished jobs....
> make[5]: *** [../scripts/Makefile.build:546: drivers/vfio/pci] Error 2
> make[5]: *** Waiting for unfinished jobs....
> make[4]: *** [../scripts/Makefile.build:546: drivers/vfio] Error 2
> make[3]: *** [../scripts/Makefile.build:546: drivers] Error 2
>
> Fixes: f9e54c3a2f5b7 ("vfio/pci: implement huge_fault support")
> Signed-off-by: Ritesh Harjani (IBM) <ritesh.list@gmail.com>
> ---
> drivers/vfio/pci/vfio_pci_core.c | 15 +++++++--------
> 1 file changed, 7 insertions(+), 8 deletions(-)
>
> diff --git a/drivers/vfio/pci/vfio_pci_core.c b/drivers/vfio/pci/vfio_pci_core.c
> index d43745fe4c84..5395a6f30904 100644
> --- a/drivers/vfio/pci/vfio_pci_core.c
> +++ b/drivers/vfio/pci/vfio_pci_core.c
> @@ -1670,21 +1670,20 @@ vm_fault_t vfio_pci_vmf_insert_pfn(struct vfio_pci_core_device *vdev,
> if (vdev->pm_runtime_engaged || !__vfio_pci_memory_enabled(vdev))
> return VM_FAULT_SIGBUS;
>
> - switch (order) {
> - case 0:
> + if (order == 0) {
> return vmf_insert_pfn(vmf->vma, vmf->address, pfn);
> + }
Those braces are unneeded as all legs of the if/else are single lines
> #ifdef CONFIG_ARCH_SUPPORTS_PMD_PFNMAP
ifdef could be replaced by IS_ENABLED() because PxD_ORDER and
vmf_insert_pfn_xxx() are declared all the time
> - case PMD_ORDER:
> + else if (order == PMD_ORDER) {
'else' is not needed because every 'if' leads to a return statement
> return vmf_insert_pfn_pmd(vmf, pfn, false);
> + }
> #endif
> #ifdef CONFIG_ARCH_SUPPORTS_PUD_PFNMAP
> - case PUD_ORDER:
> + else if (order == PUD_ORDER) {
> return vmf_insert_pfn_pud(vmf, pfn, false);
> - break;
> + }
> #endif
> - default:
> - return VM_FAULT_FALLBACK;
> - }
> + return VM_FAULT_FALLBACK;
So at the end we should get something like:
if (!order)
return vmf_insert_pfn(vmf->vma, vmf->address, pfn);
if (IS_ENABLED(CONFIG_ARCH_SUPPORTS_PMD_PFNMAP) && order == PMD_ORDER)
return vmf_insert_pfn_pmd(vmf, pfn, false);
if (IS_ENABLED(CONFIG_ARCH_SUPPORTS_PUD_PFNMAP) && order == PMD_ORDER)
return vmf_insert_pfn_pud(vmf, pfn, false);
return VM_FAULT_FALLBACK;
> }
> EXPORT_SYMBOL_GPL(vfio_pci_vmf_insert_pfn);
>
> --
> 2.53.0
>
>
^ permalink raw reply [flat|nested] 6+ messages in thread* Re: [RFC v1 1/2] drivers/vfio_pci_core: Change PXD_ORDER check from switch case to if/else block
2026-02-27 6:42 ` [RFC v1 1/2] drivers/vfio_pci_core: Change PXD_ORDER check from switch case to if/else block Christophe Leroy (CS GROUP)
@ 2026-02-27 10:30 ` Ritesh Harjani
0 siblings, 0 replies; 6+ messages in thread
From: Ritesh Harjani @ 2026-02-27 10:30 UTC (permalink / raw)
To: Christophe Leroy (CS GROUP), linuxppc-dev
Cc: linux-mm, kvm, Alex Williamson, Peter Xu
"Christophe Leroy (CS GROUP)" <chleroy@kernel.org> writes:
> Le 27/02/2026 à 07:16, Ritesh Harjani (IBM) a écrit :
>> Architectures like PowerPC uses runtime defined values for
>> PMD_ORDER/PUD_ORDER. This is because it can use either RADIX or HASH MMU
>> at runtime using kernel cmdline. So the pXd_index_size is not known at
>> compile time. Without this fix, when we add huge pfn support on powerpc
>> in the next patch, vfio_pci_core driver compilation can fail with the
>> following errors.
>>
>> CC [M] drivers/vfio/vfio_main.o
>> CC [M] drivers/vfio/group.o
>> CC [M] drivers/vfio/container.o
>> CC [M] drivers/vfio/virqfd.o
>> CC [M] drivers/vfio/vfio_iommu_spapr_tce.o
>> CC [M] drivers/vfio/pci/vfio_pci_core.o
>> CC [M] drivers/vfio/pci/vfio_pci_intrs.o
>> CC [M] drivers/vfio/pci/vfio_pci_rdwr.o
>> CC [M] drivers/vfio/pci/vfio_pci_config.o
>> CC [M] drivers/vfio/pci/vfio_pci.o
>> AR kernel/built-in.a
>> ../drivers/vfio/pci/vfio_pci_core.c: In function ‘vfio_pci_vmf_insert_pfn’:
>> ../drivers/vfio/pci/vfio_pci_core.c:1678:9: error: case label does not reduce to an integer constant
>> 1678 | case PMD_ORDER:
>> | ^~~~
>> ../drivers/vfio/pci/vfio_pci_core.c:1682:9: error: case label does not reduce to an integer constant
>> 1682 | case PUD_ORDER:
>> | ^~~~
>> make[6]: *** [../scripts/Makefile.build:289: drivers/vfio/pci/vfio_pci_core.o] Error 1
>> make[6]: *** Waiting for unfinished jobs....
>> make[5]: *** [../scripts/Makefile.build:546: drivers/vfio/pci] Error 2
>> make[5]: *** Waiting for unfinished jobs....
>> make[4]: *** [../scripts/Makefile.build:546: drivers/vfio] Error 2
>> make[3]: *** [../scripts/Makefile.build:546: drivers] Error 2
>>
>> Fixes: f9e54c3a2f5b7 ("vfio/pci: implement huge_fault support")
>> Signed-off-by: Ritesh Harjani (IBM) <ritesh.list@gmail.com>
>> ---
>> drivers/vfio/pci/vfio_pci_core.c | 15 +++++++--------
>> 1 file changed, 7 insertions(+), 8 deletions(-)
>>
>> diff --git a/drivers/vfio/pci/vfio_pci_core.c b/drivers/vfio/pci/vfio_pci_core.c
>> index d43745fe4c84..5395a6f30904 100644
>> --- a/drivers/vfio/pci/vfio_pci_core.c
>> +++ b/drivers/vfio/pci/vfio_pci_core.c
>> @@ -1670,21 +1670,20 @@ vm_fault_t vfio_pci_vmf_insert_pfn(struct vfio_pci_core_device *vdev,
>> if (vdev->pm_runtime_engaged || !__vfio_pci_memory_enabled(vdev))
>> return VM_FAULT_SIGBUS;
>>
>> - switch (order) {
>> - case 0:
>> + if (order == 0) {
>> return vmf_insert_pfn(vmf->vma, vmf->address, pfn);
>> + }
>
> Those braces are unneeded as all legs of the if/else are single lines
>
>> #ifdef CONFIG_ARCH_SUPPORTS_PMD_PFNMAP
>
> ifdef could be replaced by IS_ENABLED() because PxD_ORDER and
> vmf_insert_pfn_xxx() are declared all the time
>
>> - case PMD_ORDER:
>> + else if (order == PMD_ORDER) {
>
> 'else' is not needed because every 'if' leads to a return statement
>
>> return vmf_insert_pfn_pmd(vmf, pfn, false);
>> + }
>> #endif
>> #ifdef CONFIG_ARCH_SUPPORTS_PUD_PFNMAP
>> - case PUD_ORDER:
>> + else if (order == PUD_ORDER) {
>> return vmf_insert_pfn_pud(vmf, pfn, false);
>> - break;
>> + }
>> #endif
>> - default:
>> - return VM_FAULT_FALLBACK;
>> - }
>> + return VM_FAULT_FALLBACK;
>
> So at the end we should get something like:
>
> if (!order)
> return vmf_insert_pfn(vmf->vma, vmf->address, pfn);
>
> if (IS_ENABLED(CONFIG_ARCH_SUPPORTS_PMD_PFNMAP) && order == PMD_ORDER)
> return vmf_insert_pfn_pmd(vmf, pfn, false);
>
> if (IS_ENABLED(CONFIG_ARCH_SUPPORTS_PUD_PFNMAP) && order == PMD_ORDER)
^^^ PUD_ORDER
> return vmf_insert_pfn_pud(vmf, pfn, false);
>
> return VM_FAULT_FALLBACK;
>
>
Looks a lot cleaner. Thanks!
I will make that change in v2.
-ritesh
^ permalink raw reply [flat|nested] 6+ messages in thread
end of thread, other threads:[~2026-02-27 12:27 UTC | newest]
Thread overview: 6+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2026-02-27 6:16 [RFC v1 1/2] drivers/vfio_pci_core: Change PXD_ORDER check from switch case to if/else block Ritesh Harjani (IBM)
2026-02-27 6:16 ` [RFC v1 2/2] powerpc/64s: Add support for huge pfnmaps Ritesh Harjani (IBM)
2026-02-27 6:47 ` Christophe Leroy (CS GROUP)
2026-02-27 10:32 ` Ritesh Harjani
2026-02-27 6:42 ` [RFC v1 1/2] drivers/vfio_pci_core: Change PXD_ORDER check from switch case to if/else block Christophe Leroy (CS GROUP)
2026-02-27 10:30 ` Ritesh Harjani
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox