From: Alexandre Ghiti alexghiti@rivosinc.com
huge_ptep_get must be reimplemented in order to go through all the PTEs of a NAPOT region: this is needed because the HW can update the A/D bits of any of the PTE that constitutes the NAPOT region.
Fixes: 82a1a1f3bfb6 ("riscv: mm: support Svnapot in hugetlb page") Signed-off-by: Alexandre Ghiti alexghiti@rivosinc.com Reviewed-by: Andrew Jones ajones@ventanamicro.com Link: https://lore.kernel.org/r/20230428120120.21620-2-alexghiti@rivosinc.com Cc: stable@vger.kernel.org Signed-off-by: Palmer Dabbelt palmer@rivosinc.com Signed-off-by: Petko Manolov petko.manolov@konsulko.com --- arch/riscv/include/asm/hugetlb.h | 3 +++ arch/riscv/mm/hugetlbpage.c | 24 ++++++++++++++++++++++++ 2 files changed, 27 insertions(+)
diff --git a/arch/riscv/include/asm/hugetlb.h b/arch/riscv/include/asm/hugetlb.h index fe6f23006641..ce1ebda1a49a 100644 --- a/arch/riscv/include/asm/hugetlb.h +++ b/arch/riscv/include/asm/hugetlb.h @@ -36,6 +36,9 @@ int huge_ptep_set_access_flags(struct vm_area_struct *vma, unsigned long addr, pte_t *ptep, pte_t pte, int dirty);
+#define __HAVE_ARCH_HUGE_PTEP_GET +pte_t huge_ptep_get(pte_t *ptep); + pte_t arch_make_huge_pte(pte_t entry, unsigned int shift, vm_flags_t flags); #define arch_make_huge_pte arch_make_huge_pte
diff --git a/arch/riscv/mm/hugetlbpage.c b/arch/riscv/mm/hugetlbpage.c index 238d00bdac14..e0ef56dc57b9 100644 --- a/arch/riscv/mm/hugetlbpage.c +++ b/arch/riscv/mm/hugetlbpage.c @@ -3,6 +3,30 @@ #include <linux/err.h>
#ifdef CONFIG_RISCV_ISA_SVNAPOT +pte_t huge_ptep_get(pte_t *ptep) +{ + unsigned long pte_num; + int i; + pte_t orig_pte = ptep_get(ptep); + + if (!pte_present(orig_pte) || !pte_napot(orig_pte)) + return orig_pte; + + pte_num = napot_pte_num(napot_cont_order(orig_pte)); + + for (i = 0; i < pte_num; i++, ptep++) { + pte_t pte = ptep_get(ptep); + + if (pte_dirty(pte)) + orig_pte = pte_mkdirty(orig_pte); + + if (pte_young(pte)) + orig_pte = pte_mkyoung(orig_pte); + } + + return orig_pte; +} + pte_t *huge_pte_alloc(struct mm_struct *mm, struct vm_area_struct *vma, unsigned long addr,
Hi Petko,
On Sun, Jul 23, 2023 at 3:48 PM Petko Manolov petko.manolov@konsulko.com wrote:
From: Alexandre Ghiti alexghiti@rivosinc.com
huge_ptep_get must be reimplemented in order to go through all the PTEs of a NAPOT region: this is needed because the HW can update the A/D bits of any of the PTE that constitutes the NAPOT region.
Fixes: 82a1a1f3bfb6 ("riscv: mm: support Svnapot in hugetlb page") Signed-off-by: Alexandre Ghiti alexghiti@rivosinc.com Reviewed-by: Andrew Jones ajones@ventanamicro.com Link: https://lore.kernel.org/r/20230428120120.21620-2-alexghiti@rivosinc.com Cc: stable@vger.kernel.org Signed-off-by: Palmer Dabbelt palmer@rivosinc.com Signed-off-by: Petko Manolov petko.manolov@konsulko.com
arch/riscv/include/asm/hugetlb.h | 3 +++ arch/riscv/mm/hugetlbpage.c | 24 ++++++++++++++++++++++++ 2 files changed, 27 insertions(+)
diff --git a/arch/riscv/include/asm/hugetlb.h b/arch/riscv/include/asm/hugetlb.h index fe6f23006641..ce1ebda1a49a 100644 --- a/arch/riscv/include/asm/hugetlb.h +++ b/arch/riscv/include/asm/hugetlb.h @@ -36,6 +36,9 @@ int huge_ptep_set_access_flags(struct vm_area_struct *vma, unsigned long addr, pte_t *ptep, pte_t pte, int dirty);
+#define __HAVE_ARCH_HUGE_PTEP_GET +pte_t huge_ptep_get(pte_t *ptep);
pte_t arch_make_huge_pte(pte_t entry, unsigned int shift, vm_flags_t flags); #define arch_make_huge_pte arch_make_huge_pte
diff --git a/arch/riscv/mm/hugetlbpage.c b/arch/riscv/mm/hugetlbpage.c index 238d00bdac14..e0ef56dc57b9 100644 --- a/arch/riscv/mm/hugetlbpage.c +++ b/arch/riscv/mm/hugetlbpage.c @@ -3,6 +3,30 @@ #include <linux/err.h>
#ifdef CONFIG_RISCV_ISA_SVNAPOT +pte_t huge_ptep_get(pte_t *ptep) +{
unsigned long pte_num;
int i;
pte_t orig_pte = ptep_get(ptep);
if (!pte_present(orig_pte) || !pte_napot(orig_pte))
return orig_pte;
pte_num = napot_pte_num(napot_cont_order(orig_pte));
for (i = 0; i < pte_num; i++, ptep++) {
pte_t pte = ptep_get(ptep);
if (pte_dirty(pte))
orig_pte = pte_mkdirty(orig_pte);
if (pte_young(pte))
orig_pte = pte_mkyoung(orig_pte);
}
return orig_pte;
+}
pte_t *huge_pte_alloc(struct mm_struct *mm, struct vm_area_struct *vma, unsigned long addr, -- 2.39.2
Not sure I understand why you propose this patch to stable since svnapot support was introduced in 6.4 and this fix (and another one) was merged in 6.4 too. Am I missing something?
Thanks,
Alex
On 23-07-24 10:30:20, Alexandre Ghiti wrote:
Hi Petko,
On Sun, Jul 23, 2023 at 3:48 PM Petko Manolov petko.manolov@konsulko.com wrote:
From: Alexandre Ghiti alexghiti@rivosinc.com
huge_ptep_get must be reimplemented in order to go through all the PTEs of a NAPOT region: this is needed because the HW can update the A/D bits of any of the PTE that constitutes the NAPOT region.
Fixes: 82a1a1f3bfb6 ("riscv: mm: support Svnapot in hugetlb page") Signed-off-by: Alexandre Ghiti alexghiti@rivosinc.com Reviewed-by: Andrew Jones ajones@ventanamicro.com Link: https://lore.kernel.org/r/20230428120120.21620-2-alexghiti@rivosinc.com Cc: stable@vger.kernel.org Signed-off-by: Palmer Dabbelt palmer@rivosinc.com Signed-off-by: Petko Manolov petko.manolov@konsulko.com
arch/riscv/include/asm/hugetlb.h | 3 +++ arch/riscv/mm/hugetlbpage.c | 24 ++++++++++++++++++++++++ 2 files changed, 27 insertions(+)
diff --git a/arch/riscv/include/asm/hugetlb.h b/arch/riscv/include/asm/hugetlb.h index fe6f23006641..ce1ebda1a49a 100644 --- a/arch/riscv/include/asm/hugetlb.h +++ b/arch/riscv/include/asm/hugetlb.h @@ -36,6 +36,9 @@ int huge_ptep_set_access_flags(struct vm_area_struct *vma, unsigned long addr, pte_t *ptep, pte_t pte, int dirty);
+#define __HAVE_ARCH_HUGE_PTEP_GET +pte_t huge_ptep_get(pte_t *ptep);
pte_t arch_make_huge_pte(pte_t entry, unsigned int shift, vm_flags_t flags); #define arch_make_huge_pte arch_make_huge_pte
diff --git a/arch/riscv/mm/hugetlbpage.c b/arch/riscv/mm/hugetlbpage.c index 238d00bdac14..e0ef56dc57b9 100644 --- a/arch/riscv/mm/hugetlbpage.c +++ b/arch/riscv/mm/hugetlbpage.c @@ -3,6 +3,30 @@ #include <linux/err.h>
#ifdef CONFIG_RISCV_ISA_SVNAPOT +pte_t huge_ptep_get(pte_t *ptep) +{
unsigned long pte_num;
int i;
pte_t orig_pte = ptep_get(ptep);
if (!pte_present(orig_pte) || !pte_napot(orig_pte))
return orig_pte;
pte_num = napot_pte_num(napot_cont_order(orig_pte));
for (i = 0; i < pte_num; i++, ptep++) {
pte_t pte = ptep_get(ptep);
if (pte_dirty(pte))
orig_pte = pte_mkdirty(orig_pte);
if (pte_young(pte))
orig_pte = pte_mkyoung(orig_pte);
}
return orig_pte;
+}
pte_t *huge_pte_alloc(struct mm_struct *mm, struct vm_area_struct *vma, unsigned long addr, -- 2.39.2
Not sure I understand why you propose this patch to stable since svnapot support was introduced in 6.4 and this fix (and another one) was merged in 6.4 too. Am I missing something?
Hey there. I've no idea how "Signed-off" with my email address appeared in this patch. I most certainly have not done so, nor i am in any way related to this code. Weird.
Petko
linux-stable-mirror@lists.linaro.org