1. Patchew
  2. linux
  3. View series

[PATCH V2] arm64/hugetlb: Consistently use pud_sect_supported()

Anshuman Khandual posted 1 patch 10 months ago
There is a newer version of this series
arch/arm64/mm/hugetlbpage.c | 9 +++++++--
1 file changed, 7 insertions(+), 2 deletions(-)
[PATCH V2] arm64/hugetlb: Consistently use pud_sect_supported()
Posted by Anshuman Khandual 10 months ago 
Let's be consistent in using pud_sect_supported() for PUD_SIZE sized pages.
Hence change hugetlb_mask_last_page() and arch_make_huge_pte() as required.

Cc: Catalin Marinas <catalin.marinas@arm.com>
Cc: Will Deacon <will@kernel.org>
Cc: linux-arm-kernel@lists.infradead.org
Cc: linux-kernel@vger.kernel.org
Signed-off-by: Anshuman Khandual <anshuman.khandual@arm.com>
---
This patch applies on v6.14-rc3

Changes in V2:

- Added an warning when PUD_SIZE is requested but not supported per Ryan

Changes in V1:

https://lore.kernel.org/all/20250217065414.49489-1-anshuman.khandual@arm.com/

 arch/arm64/mm/hugetlbpage.c | 9 +++++++--
 1 file changed, 7 insertions(+), 2 deletions(-)

diff --git a/arch/arm64/mm/hugetlbpage.c b/arch/arm64/mm/hugetlbpage.c
index 98a2a0e64e25..1d89599a20d7 100644
--- a/arch/arm64/mm/hugetlbpage.c
+++ b/arch/arm64/mm/hugetlbpage.c
@@ -342,7 +342,9 @@ unsigned long hugetlb_mask_last_page(struct hstate *h)
 	switch (hp_size) {
 #ifndef __PAGETABLE_PMD_FOLDED
 	case PUD_SIZE:
-		return PGDIR_SIZE - PUD_SIZE;
+		if (pud_sect_supported())
+			return PGDIR_SIZE - PUD_SIZE;
+		break;
 #endif
 	case CONT_PMD_SIZE:
 		return PUD_SIZE - CONT_PMD_SIZE;
@@ -364,7 +366,10 @@ pte_t arch_make_huge_pte(pte_t entry, unsigned int shift, vm_flags_t flags)
 	switch (pagesize) {
 #ifndef __PAGETABLE_PMD_FOLDED
 	case PUD_SIZE:
-		entry = pud_pte(pud_mkhuge(pte_pud(entry)));
+		if (pud_sect_supported())
+			entry = pud_pte(pud_mkhuge(pte_pud(entry)));
+		else
+			pr_warn("%s: pud huge page not supported\n", __func__);
 		break;
 #endif
 	case CONT_PMD_SIZE:
-- 
2.30.2
Re: [PATCH V2] arm64/hugetlb: Consistently use pud_sect_supported()
Posted by Ryan Roberts 10 months ago 
On 18/02/2025 11:36, Anshuman Khandual wrote:
> Let's be consistent in using pud_sect_supported() for PUD_SIZE sized pages.
> Hence change hugetlb_mask_last_page() and arch_make_huge_pte() as required.
> 
> Cc: Catalin Marinas <catalin.marinas@arm.com>
> Cc: Will Deacon <will@kernel.org>
> Cc: linux-arm-kernel@lists.infradead.org
> Cc: linux-kernel@vger.kernel.org
> Signed-off-by: Anshuman Khandual <anshuman.khandual@arm.com>
> ---
> This patch applies on v6.14-rc3
> 
> Changes in V2:
> 
> - Added an warning when PUD_SIZE is requested but not supported per Ryan
> 
> Changes in V1:
> 
> https://lore.kernel.org/all/20250217065414.49489-1-anshuman.khandual@arm.com/
> 
>  arch/arm64/mm/hugetlbpage.c | 9 +++++++--
>  1 file changed, 7 insertions(+), 2 deletions(-)
> 
> diff --git a/arch/arm64/mm/hugetlbpage.c b/arch/arm64/mm/hugetlbpage.c
> index 98a2a0e64e25..1d89599a20d7 100644
> --- a/arch/arm64/mm/hugetlbpage.c
> +++ b/arch/arm64/mm/hugetlbpage.c
> @@ -342,7 +342,9 @@ unsigned long hugetlb_mask_last_page(struct hstate *h)
>  	switch (hp_size) {
>  #ifndef __PAGETABLE_PMD_FOLDED
>  	case PUD_SIZE:
> -		return PGDIR_SIZE - PUD_SIZE;
> +		if (pud_sect_supported())
> +			return PGDIR_SIZE - PUD_SIZE;
> +		break;
>  #endif
>  	case CONT_PMD_SIZE:
>  		return PUD_SIZE - CONT_PMD_SIZE;
> @@ -364,7 +366,10 @@ pte_t arch_make_huge_pte(pte_t entry, unsigned int shift, vm_flags_t flags)
>  	switch (pagesize) {
>  #ifndef __PAGETABLE_PMD_FOLDED
>  	case PUD_SIZE:
> -		entry = pud_pte(pud_mkhuge(pte_pud(entry)));
> +		if (pud_sect_supported())
> +			entry = pud_pte(pud_mkhuge(pte_pud(entry)));
> +		else
> +			pr_warn("%s: pud huge page not supported\n", __func__);
>  		break;
>  #endif
>  	case CONT_PMD_SIZE:

Personally, I think something like this is cleaner than having 2 warnings:

pte_t arch_make_huge_pte(pte_t entry, unsigned int shift, vm_flags_t flags)
{
	size_t pagesize = 1UL << shift;

	switch (pagesize) {
#ifndef __PAGETABLE_PMD_FOLDED
	case PUD_SIZE:
		if (pud_sect_supported())
			return pud_pte(pud_mkhuge(pte_pud(entry)));
		break;
#endif
	case CONT_PMD_SIZE:
		return pmd_pte(pmd_mkhuge(pmd_mkcont(pte_pmd(entry))));
	case PMD_SIZE:
		return pmd_pte(pmd_mkhuge(pte_pmd(entry)));
	case CONT_PTE_SIZE:
		return pte_mkcont(entry);
	}

	pr_warn("%s: unrecognized huge page size 0x%lx\n",
		__func__, pagesize);
	return entry;
}

Thanks,
Ryan
Re: [PATCH V2] arm64/hugetlb: Consistently use pud_sect_supported()
Posted by Anshuman Khandual 10 months ago 

On 2/18/25 19:31, Ryan Roberts wrote:
> On 18/02/2025 11:36, Anshuman Khandual wrote:
>> Let's be consistent in using pud_sect_supported() for PUD_SIZE sized pages.
>> Hence change hugetlb_mask_last_page() and arch_make_huge_pte() as required.
>>
>> Cc: Catalin Marinas <catalin.marinas@arm.com>
>> Cc: Will Deacon <will@kernel.org>
>> Cc: linux-arm-kernel@lists.infradead.org
>> Cc: linux-kernel@vger.kernel.org
>> Signed-off-by: Anshuman Khandual <anshuman.khandual@arm.com>
>> ---
>> This patch applies on v6.14-rc3
>>
>> Changes in V2:
>>
>> - Added an warning when PUD_SIZE is requested but not supported per Ryan
>>
>> Changes in V1:
>>
>> https://lore.kernel.org/all/20250217065414.49489-1-anshuman.khandual@arm.com/
>>
>>  arch/arm64/mm/hugetlbpage.c | 9 +++++++--
>>  1 file changed, 7 insertions(+), 2 deletions(-)
>>
>> diff --git a/arch/arm64/mm/hugetlbpage.c b/arch/arm64/mm/hugetlbpage.c
>> index 98a2a0e64e25..1d89599a20d7 100644
>> --- a/arch/arm64/mm/hugetlbpage.c
>> +++ b/arch/arm64/mm/hugetlbpage.c
>> @@ -342,7 +342,9 @@ unsigned long hugetlb_mask_last_page(struct hstate *h)
>>  	switch (hp_size) {
>>  #ifndef __PAGETABLE_PMD_FOLDED
>>  	case PUD_SIZE:
>> -		return PGDIR_SIZE - PUD_SIZE;
>> +		if (pud_sect_supported())
>> +			return PGDIR_SIZE - PUD_SIZE;
>> +		break;
>>  #endif
>>  	case CONT_PMD_SIZE:
>>  		return PUD_SIZE - CONT_PMD_SIZE;
>> @@ -364,7 +366,10 @@ pte_t arch_make_huge_pte(pte_t entry, unsigned int shift, vm_flags_t flags)
>>  	switch (pagesize) {
>>  #ifndef __PAGETABLE_PMD_FOLDED
>>  	case PUD_SIZE:
>> -		entry = pud_pte(pud_mkhuge(pte_pud(entry)));
>> +		if (pud_sect_supported())
>> +			entry = pud_pte(pud_mkhuge(pte_pud(entry)));
>> +		else
>> +			pr_warn("%s: pud huge page not supported\n", __func__);
>>  		break;
>>  #endif
>>  	case CONT_PMD_SIZE:
> 
> Personally, I think something like this is cleaner than having 2 warnings:
> 
> pte_t arch_make_huge_pte(pte_t entry, unsigned int shift, vm_flags_t flags)
> {
> 	size_t pagesize = 1UL << shift;
> 
> 	switch (pagesize) {
> #ifndef __PAGETABLE_PMD_FOLDED
> 	case PUD_SIZE:
> 		if (pud_sect_supported())
> 			return pud_pte(pud_mkhuge(pte_pud(entry)));
> 		break;
> #endif
> 	case CONT_PMD_SIZE:
> 		return pmd_pte(pmd_mkhuge(pmd_mkcont(pte_pmd(entry))));
> 	case PMD_SIZE:
> 		return pmd_pte(pmd_mkhuge(pte_pmd(entry)));
> 	case CONT_PTE_SIZE:
> 		return pte_mkcont(entry);

A "default" entry is needed here for the switch statement above. But that
could be just a simple break statement.

> 	}
> 
> 	pr_warn("%s: unrecognized huge page size 0x%lx\n",
> 		__func__, pagesize);


> 	return entry;
> }

Initially thought about that, but wondered if "unsupported" should be called
out as "unrecognized" instead. But that's a trivial detail. So I will do the
changes as suggested.

Patchew 2.1-devel-b67e4c2

© 2016 - 2025 Red Hat, Inc.