diff mbox series

[v6,16/18] arm64/mm: Implement pte_batch_hint()

Message ID 20240215103205.2607016-17-ryan.roberts@arm.com (mailing list archive)
State New, archived
Headers show
Series Transparent Contiguous PTEs for User Mappings | expand

Commit Message

Ryan Roberts Feb. 15, 2024, 10:32 a.m. UTC 
When core code iterates over a range of ptes and calls ptep_get() for
each of them, if the range happens to cover contpte mappings, the number
of pte reads becomes amplified by a factor of the number of PTEs in a
contpte block. This is because for each call to ptep_get(), the
implementation must read all of the ptes in the contpte block to which
it belongs to gather the access and dirty bits.

This causes a hotspot for fork(), as well as operations that unmap
memory such as munmap(), exit and madvise(MADV_DONTNEED). Fortunately we
can fix this by implementing pte_batch_hint() which allows their
iterators to skip getting the contpte tail ptes when gathering the batch
of ptes to operate on. This results in the number of PTE reads returning
to 1 per pte.

Acked-by: Mark Rutland <mark.rutland@arm.com>
Reviewed-by: David Hildenbrand <david@redhat.com>
Tested-by: John Hubbard <jhubbard@nvidia.com>
Signed-off-by: Ryan Roberts <ryan.roberts@arm.com>
---
 arch/arm64/include/asm/pgtable.h | 9 +++++++++
 1 file changed, 9 insertions(+)

Comments

Catalin Marinas Feb. 16, 2024, 12:34 p.m. UTC | #1 
On Thu, Feb 15, 2024 at 10:32:03AM +0000, Ryan Roberts wrote:
> When core code iterates over a range of ptes and calls ptep_get() for
> each of them, if the range happens to cover contpte mappings, the number
> of pte reads becomes amplified by a factor of the number of PTEs in a
> contpte block. This is because for each call to ptep_get(), the
> implementation must read all of the ptes in the contpte block to which
> it belongs to gather the access and dirty bits.
> 
> This causes a hotspot for fork(), as well as operations that unmap
> memory such as munmap(), exit and madvise(MADV_DONTNEED). Fortunately we
> can fix this by implementing pte_batch_hint() which allows their
> iterators to skip getting the contpte tail ptes when gathering the batch
> of ptes to operate on. This results in the number of PTE reads returning
> to 1 per pte.
> 
> Acked-by: Mark Rutland <mark.rutland@arm.com>
> Reviewed-by: David Hildenbrand <david@redhat.com>
> Tested-by: John Hubbard <jhubbard@nvidia.com>
> Signed-off-by: Ryan Roberts <ryan.roberts@arm.com>

Acked-by: Catalin Marinas <catalin.marinas@arm.com>
diff mbox series

Patch

diff --git a/arch/arm64/include/asm/pgtable.h b/arch/arm64/include/asm/pgtable.h
index a8f1a35e3086..d759a20d2929 100644
--- a/arch/arm64/include/asm/pgtable.h
+++ b/arch/arm64/include/asm/pgtable.h
@@ -1213,6 +1213,15 @@  static inline void contpte_try_unfold(struct mm_struct *mm, unsigned long addr,
 		__contpte_try_unfold(mm, addr, ptep, pte);
 }
 
+#define pte_batch_hint pte_batch_hint
+static inline unsigned int pte_batch_hint(pte_t *ptep, pte_t pte)
+{
+	if (!pte_valid_cont(pte))
+		return 1;
+
+	return CONT_PTES - (((unsigned long)ptep >> 3) & (CONT_PTES - 1));
+}
+
 /*
  * The below functions constitute the public API that arm64 presents to the
  * core-mm to manipulate PTE entries within their page tables (or at least this