Thread (31 messages) 31 messages, 8 authors, 2023-09-22

Re: [PATCH v1 3/8] riscv: hugetlb: Convert set_huge_pte_at() to take vma

From: Alexandre Ghiti <alex@ghiti.fr>
Date: 2023-09-22 07:54:43
Also in: linux-arm-kernel, linux-mm, linux-riscv, linux-s390, lkml, sparclinux, stable

Hi Ryan,

On 21/09/2023 18:20, Ryan Roberts wrote:
quoted hunk ↗ jump to hunk
In order to fix a bug, arm64 needs access to the vma inside it's
implementation of set_huge_pte_at(). Provide for this by converting the
mm parameter to be a vma. Any implementations that require the mm can
access it via vma->vm_mm.

This commit makes the required riscv modifications. Separate commits
update the other arches and core code, before the actual bug is fixed in
arm64.

No behavioral changes intended.

Signed-off-by: Ryan Roberts <ryan.roberts@arm.com>
---
  arch/riscv/include/asm/hugetlb.h | 2 +-
  arch/riscv/mm/hugetlbpage.c      | 3 ++-
  2 files changed, 3 insertions(+), 2 deletions(-)
diff --git a/arch/riscv/include/asm/hugetlb.h b/arch/riscv/include/asm/hugetlb.h
index 34e24f078cc1..be1ac8582bc2 100644
--- a/arch/riscv/include/asm/hugetlb.h
+++ b/arch/riscv/include/asm/hugetlb.h
@@ -17,7 +17,7 @@ void huge_pte_clear(struct mm_struct *mm, unsigned long addr,
  		    pte_t *ptep, unsigned long sz);
  
  #define __HAVE_ARCH_HUGE_SET_HUGE_PTE_AT
-void set_huge_pte_at(struct mm_struct *mm,
+void set_huge_pte_at(struct vm_area_struct *vma,
  		     unsigned long addr, pte_t *ptep, pte_t pte);
  
  #define __HAVE_ARCH_HUGE_PTEP_GET_AND_CLEAR
diff --git a/arch/riscv/mm/hugetlbpage.c b/arch/riscv/mm/hugetlbpage.c
index 96225a8533ad..7cdbf0960772 100644
--- a/arch/riscv/mm/hugetlbpage.c
+++ b/arch/riscv/mm/hugetlbpage.c
@@ -177,11 +177,12 @@ pte_t arch_make_huge_pte(pte_t entry, unsigned int shift, vm_flags_t flags)
  	return entry;
  }
  
-void set_huge_pte_at(struct mm_struct *mm,
+void set_huge_pte_at(struct vm_area_struct *vma,
  		     unsigned long addr,
  		     pte_t *ptep,
  		     pte_t pte)
  {
+	struct mm_struct *mm = vma->vm_mm;
  	int i, pte_num;
  
  	if (!pte_napot(pte)) {

You can add:

Reviewed-by: Alexandre Ghiti <redacted>

I realize that we may have the same issue with our contig pte 
implementation (called napot in riscv) as we don't handle swap/migration 
entries at all. So I guess we need something similar, and I'll implement 
it (unless you want to do it of course, but I guess it's easier for me 
to test). One (maybe stupid) question though: wouldn't it be possible to 
extract the contig pte size from the value of ptep instead of using a vma?

Thanks,

Alex
Keyboard shortcuts
hback out one level
jnext message in thread
kprevious message in thread
ldrill in
Escclose help / fold thread tree
?toggle this help