Re: [PATCH v4 2/4] riscv: Improve flush_tlb_range() for hugetlb pages

[Samuel Holland]

Hi Alex,

On 2023-09-11 8:12 AM, Alexandre Ghiti wrote:
> flush_tlb_range() uses a fixed stride of PAGE_SIZE and in its current form,
> when a hugetlb mapping needs to be flushed, flush_tlb_range() flushes the
> whole tlb: so set a stride of the size of the hugetlb mapping in order to
> only flush the hugetlb mapping. However, if the hugepage is a NAPOT region,
> all PTEs that constitute this mapping must be invalidated, so the stride
> size must actually be the size of the PTE.
> 
> Note that THPs are directly handled by flush_pmd_tlb_range().
> 
> Signed-off-by: Alexandre Ghiti <alexghiti@xxxxxxxxxxxx>
> Reviewed-by: Andrew Jones <ajones@xxxxxxxxxxxxxxxx>
> ---
>  arch/riscv/mm/tlbflush.c | 39 ++++++++++++++++++++++++++++++++++++++-
>  1 file changed, 38 insertions(+), 1 deletion(-)
> 
> diff --git a/arch/riscv/mm/tlbflush.c b/arch/riscv/mm/tlbflush.c
> index fa03289853d8..5bda6d4fed90 100644
> --- a/arch/riscv/mm/tlbflush.c
> +++ b/arch/riscv/mm/tlbflush.c
> @@ -3,6 +3,7 @@
>  #include <linux/mm.h>
>  #include <linux/smp.h>
>  #include <linux/sched.h>
> +#include <linux/hugetlb.h>
>  #include <asm/sbi.h>
>  #include <asm/mmu_context.h>
>  
> @@ -147,7 +148,43 @@ void flush_tlb_page(struct vm_area_struct *vma, unsigned long addr)
>  void flush_tlb_range(struct vm_area_struct *vma, unsigned long start,
>  		     unsigned long end)
>  {
> -	__flush_tlb_range(vma->vm_mm, start, end - start, PAGE_SIZE);
> +	unsigned long stride_size;
> +
> +	stride_size = is_vm_hugetlb_page(vma) ?
> +				huge_page_size(hstate_vma(vma)) :
> +				PAGE_SIZE;
> +
> +#ifdef CONFIG_RISCV_ISA_SVNAPOT
> +	/*
> +	 * As stated in the privileged specification, every PTE in a NAPOT
> +	 * region must be invalidated, so reset the stride in that case.
> +	 */
> +	if (has_svnapot()) {

This whole block should probably go inside the is_vm_hugetlb_page(vma) check,
since we have to perform that check anyway.

> +		unsigned long order, napot_size;
> +
> +		for_each_napot_order(order) {
> +			napot_size = napot_cont_size(order);
> +
> +			if (stride_size != napot_size)
> +				continue;
> +
> +			if (napot_size >= PGDIR_SIZE)

Can you check stride_size here directly, and drop the loop? We should be able to
assume that the huge page size is valid. Non-NAPOT hugepages will hit one of the
equal-to cases below, which is fine.

Regards,
Samuel

> +				stride_size = PGDIR_SIZE;
> +			else if (napot_size >= P4D_SIZE)
> +				stride_size = P4D_SIZE;
> +			else if (napot_size >= PUD_SIZE)
> +				stride_size = PUD_SIZE;
> +			else if (napot_size >= PMD_SIZE)
> +				stride_size = PMD_SIZE;
> +			else
> +				stride_size = PAGE_SIZE;
> +
> +			break;
> +		}
> +	}
> +#endif
> +
> +	__flush_tlb_range(vma->vm_mm, start, end - start, stride_size);
>  }
>  #ifdef CONFIG_TRANSPARENT_HUGEPAGE
>  void flush_pmd_tlb_range(struct vm_area_struct *vma, unsigned long start,