[PATCH v7 2/3] riscv: mm: support Svnapot in hugetlb page
Qinglin Pan
panqinglin2020 at iscas.ac.cn
Fri Nov 25 19:04:40 PST 2022
Hi Alex,
On 11/25/22 10:23 PM, Alexandre Ghiti wrote:
> On 11/19/22 12:22, panqinglin2020 at iscas.ac.cn wrote:
>> From: Qinglin Pan <panqinglin2020 at iscas.ac.cn>
>>
>> Svnapot can be used to support 64KB hugetlb page, so it can become a new
>> option when using hugetlbfs. Add a basic implementation of hugetlb page,
>> and support 64KB as a size in it by using Svnapot.
>>
>> For test, boot kernel with command line contains "default_hugepagesz=64K
>> hugepagesz=64K hugepages=20" and run a simple test like this:
>>
>> tools/testing/selftests/vm/map_hugetlb 1 16
>>
>> And it should be passed.
>>
>> Signed-off-by: Qinglin Pan <panqinglin2020 at iscas.ac.cn>
>>
>> diff --git a/arch/riscv/Kconfig b/arch/riscv/Kconfig
>> index 31f9a5a160a0..c0307b942ed2 100644
>> --- a/arch/riscv/Kconfig
>> +++ b/arch/riscv/Kconfig
>> @@ -43,7 +43,7 @@ config RISCV
>> select ARCH_USE_QUEUED_RWLOCKS
>> select ARCH_WANT_DEFAULT_TOPDOWN_MMAP_LAYOUT if MMU
>> select ARCH_WANT_FRAME_POINTERS
>> - select ARCH_WANT_GENERAL_HUGETLB
>> + select ARCH_WANT_GENERAL_HUGETLB if !RISCV_ISA_SVNAPOT
>> select ARCH_WANT_HUGE_PMD_SHARE if 64BIT
>> select ARCH_WANTS_THP_SWAP if HAVE_ARCH_TRANSPARENT_HUGEPAGE
>> select BINFMT_FLAT_NO_DATA_START_OFFSET if !MMU
>> diff --git a/arch/riscv/include/asm/hugetlb.h
>> b/arch/riscv/include/asm/hugetlb.h
>> index a5c2ca1d1cd8..854f5db2f45d 100644
>> --- a/arch/riscv/include/asm/hugetlb.h
>> +++ b/arch/riscv/include/asm/hugetlb.h
>> @@ -2,7 +2,39 @@
>> #ifndef _ASM_RISCV_HUGETLB_H
>> #define _ASM_RISCV_HUGETLB_H
>> -#include <asm-generic/hugetlb.h>
>> #include <asm/page.h>
>> +#ifdef CONFIG_RISCV_ISA_SVNAPOT
>> +#define __HAVE_ARCH_HUGE_PTE_CLEAR
>> +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,
>> + unsigned long addr, pte_t *ptep, pte_t pte);
>> +
>> +#define __HAVE_ARCH_HUGE_PTEP_GET_AND_CLEAR
>> +pte_t huge_ptep_get_and_clear(struct mm_struct *mm,
>> + unsigned long addr, pte_t *ptep);
>> +
>> +#define __HAVE_ARCH_HUGE_PTEP_CLEAR_FLUSH
>> +pte_t huge_ptep_clear_flush(struct vm_area_struct *vma,
>> + unsigned long addr, pte_t *ptep);
>> +
>> +#define __HAVE_ARCH_HUGE_PTEP_SET_WRPROTECT
>> +void huge_ptep_set_wrprotect(struct mm_struct *mm,
>> + unsigned long addr, pte_t *ptep);
>> +
>> +#define __HAVE_ARCH_HUGE_PTEP_SET_ACCESS_FLAGS
>> +int huge_ptep_set_access_flags(struct vm_area_struct *vma,
>> + unsigned long addr, pte_t *ptep,
>> + pte_t pte, int dirty);
>> +
>> +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
>> +
>> +#endif /*CONFIG_RISCV_ISA_SVNAPOT*/
>> +
>> +#include <asm-generic/hugetlb.h>
>> +
>> #endif /* _ASM_RISCV_HUGETLB_H */
>> diff --git a/arch/riscv/mm/hugetlbpage.c b/arch/riscv/mm/hugetlbpage.c
>> index 932dadfdca54..130a91b30588 100644
>> --- a/arch/riscv/mm/hugetlbpage.c
>> +++ b/arch/riscv/mm/hugetlbpage.c
>> @@ -2,6 +2,273 @@
>> #include <linux/hugetlb.h>
>> #include <linux/err.h>
>> +#ifdef CONFIG_RISCV_ISA_SVNAPOT
>> +pte_t *huge_pte_alloc(struct mm_struct *mm,
>> + struct vm_area_struct *vma,
>> + unsigned long addr,
>> + unsigned long sz)
>> +{
>> + pgd_t *pgd;
>> + p4d_t *p4d;
>> + pud_t *pud;
>> + pmd_t *pmd;
>> + pte_t *pte = NULL;
>> + unsigned long order;
>> +
>> + pgd = pgd_offset(mm, addr);
>> + p4d = p4d_alloc(mm, pgd, addr);
>> + if (!p4d)
>> + return NULL;
>> +
>> + pud = pud_alloc(mm, p4d, addr);
>> + if (!pud)
>> + return NULL;
>> + if (sz == PUD_SIZE) {
>> + pte = (pte_t *)pud;
>> + goto out;
>> + }
>> +
>> + if (sz == PMD_SIZE) {
>> + if (want_pmd_share(vma, addr) && pud_none(*pud))
>> + pte = huge_pmd_share(mm, vma, addr, pud);
>> + else
>> + pte = (pte_t *)pmd_alloc(mm, pud, addr);
>> + goto out;
>> + }
>> + pmd = pmd_alloc(mm, pud, addr);
>> + if (!pmd)
>> + return NULL;
>> +
>> + for (order = NAPOT_CONT_ORDER_BASE; order < NAPOT_ORDER_MAX;
>> order++) {
>> + if (napot_cont_size(order) == sz) {
>> + pte = pte_alloc_map(mm, pmd, (addr &
>> ~napot_cont_mask(order)));
>> + break;
>> + }
>> + }
>> +out:
>> + WARN_ON_ONCE(pte && pte_present(*pte) && !pte_huge(*pte));
>> + return pte;
>> +}
>> +
>> +pte_t *huge_pte_offset(struct mm_struct *mm,
>> + unsigned long addr,
>> + unsigned long sz)
>> +{
>> + pgd_t *pgd;
>> + p4d_t *p4d;
>> + pud_t *pud;
>> + pmd_t *pmd;
>> + pte_t *pte = NULL;
>> + unsigned long order;
>> +
>> + pgd = pgd_offset(mm, addr);
>> + if (!pgd_present(*pgd))
>> + return NULL;
>> + p4d = p4d_offset(pgd, addr);
>> + if (!p4d_present(*p4d))
>> + return NULL;
>> +
>> + pud = pud_offset(p4d, addr);
>> + if (sz == PUD_SIZE)
>> + /* must be pud huge, non-present or none */
>> + return (pte_t *)pud;
>> + if (!pud_present(*pud))
>> + return NULL;
>> +
>> + pmd = pmd_offset(pud, addr);
>> + if (sz == PMD_SIZE)
>> + /* must be pmd huge, non-present or none */
>> + return (pte_t *)pmd;
>> + if (!pmd_present(*pmd))
>> + return NULL;
>> +
>> + for (order = NAPOT_CONT_ORDER_BASE; order < NAPOT_ORDER_MAX;
>> order++) {
>> + if (napot_cont_size(order) == sz) {
>> + pte = pte_offset_kernel(pmd, (addr &
>> ~napot_cont_mask(order)));
>> + break;
>> + }
>> + }
>> + return pte;
>> +}
>
>
> It's really too bad we can't use generic code for that + an arch
> specific callback here, have you tried to do so?
>
>
>> +
>> +static pte_t get_clear_contig(struct mm_struct *mm,
>> + unsigned long addr,
>> + pte_t *ptep,
>> + unsigned long pte_num)
>> +{
>> + pte_t orig_pte = ptep_get(ptep);
>> + unsigned long i;
>> +
>> + for (i = 0; i < pte_num; i++, addr += PAGE_SIZE, ptep++) {
>> + pte_t pte = ptep_get_and_clear(mm, addr, ptep);
>> +
>> + if (pte_dirty(pte))
>> + orig_pte = pte_mkdirty(orig_pte);
>> +
>> + if (pte_young(pte))
>> + orig_pte = pte_mkyoung(orig_pte);
>> + }
>> + return orig_pte;
>> +}
>
>
> Same here: can't we share code with arm64?
>
>
>> +
>> +static pte_t get_clear_contig_flush(struct mm_struct *mm,
>> + unsigned long addr,
>> + pte_t *ptep,
>> + unsigned long pte_num)
>> +{
>> + pte_t orig_pte = get_clear_contig(mm, addr, ptep, pte_num);
>> + bool valid = !pte_none(orig_pte);
>> + struct vm_area_struct vma = TLB_FLUSH_VMA(mm, 0);
>> +
>> + if (valid)
>> + flush_tlb_range(&vma, addr, addr + (PAGE_SIZE * pte_num));
>> +
>> + return orig_pte;
>> +}
>
>
> Here again: arm64 does something very similar .
>
>
>> +
>> +pte_t arch_make_huge_pte(pte_t entry, unsigned int shift, vm_flags_t
>> flags)
>> +{
>> + unsigned long order;
>> +
>> + for_each_napot_order(order) {
>> + if (shift == napot_cont_shift(order)) {
>> + entry = pte_mknapot(entry, order);
>> + break;
>> + }
>> + }
>> + if (order == NAPOT_ORDER_MAX)
>> + entry = pte_mkhuge(entry);
>> +
>> + return entry;
>> +}
>
>
> Isn't it possible to do the following instead?
>
> if (shift >= napot_cont_shift(NAPOT_ORDER_MAX))
>
> entry = pte_mkhuge(entry);
>
> else
>
> entry = pte_mknapot(entry, order);
>
>
>> +
>> +void set_huge_pte_at(struct mm_struct *mm,
>> + unsigned long addr,
>> + pte_t *ptep,
>> + pte_t pte)
>> +{
>> + int i;
>> + int pte_num;
>> +
>> + if (!pte_napot(pte)) {
>> + set_pte_at(mm, addr, ptep, pte);
>> + return;
>> + }
>> +
>> + pte_num = napot_pte_num(napot_cont_order(pte));
>> + for (i = 0; i < pte_num; i++, ptep++, addr += PAGE_SIZE)
>> + set_pte_at(mm, addr, ptep, pte);
>> +}
>> +
>> +int huge_ptep_set_access_flags(struct vm_area_struct *vma,
>> + unsigned long addr,
>> + pte_t *ptep,
>> + pte_t pte,
>> + int dirty)
>> +{
>> + pte_t orig_pte;
>> + int i;
>> + int pte_num;
>> + struct mm_struct *mm = vma->vm_mm;
>> +
>> + if (!pte_napot(pte))
>> + return ptep_set_access_flags(vma, addr, ptep, pte, dirty);
>> +
>> + pte_num = napot_pte_num(napot_cont_order(pte));
>> + ptep = huge_pte_offset(mm, addr,
>> + napot_cont_size(napot_cont_order(pte)));
>> + orig_pte = get_clear_contig_flush(mm, addr, ptep, pte_num);
>> +
>> + if (pte_dirty(orig_pte))
>> + pte = pte_mkdirty(pte);
>> +
>> + if (pte_young(orig_pte))
>> + pte = pte_mkyoung(pte);
>> +
>> + for (i = 0; i < pte_num; i++, addr += PAGE_SIZE, ptep++)
>> + set_pte_at(mm, addr, ptep, pte);
>> +
>> + return true;
>> +}
>> +
>> +pte_t huge_ptep_get_and_clear(struct mm_struct *mm,
>> + unsigned long addr,
>> + pte_t *ptep)
>> +{
>> + int pte_num;
>> + pte_t orig_pte = ptep_get(ptep);
>> +
>> + if (!pte_napot(orig_pte))
>> + return ptep_get_and_clear(mm, addr, ptep);
>> +
>> + pte_num = napot_pte_num(napot_cont_order(orig_pte));
>> +
>> + return get_clear_contig(mm, addr, ptep, pte_num);
>> +}
>> +
>> +void huge_ptep_set_wrprotect(struct mm_struct *mm,
>> + unsigned long addr,
>> + pte_t *ptep)
>> +{
>> + int i;
>> + int pte_num;
>> + pte_t pte = ptep_get(ptep);
>> +
>> + if (!pte_napot(pte)) {
>> + ptep_set_wrprotect(mm, addr, ptep);
>> + return;
>> + }
>> +
>> + pte_num = napot_pte_num(napot_cont_order(pte));
>> + ptep = huge_pte_offset(mm, addr, napot_cont_size(4UL));
>> +
>> + for (i = 0; i < pte_num; i++, addr += PAGE_SIZE, ptep++)
>> + ptep_set_wrprotect(mm, addr, ptep);
>> +}
>> +
>> +pte_t huge_ptep_clear_flush(struct vm_area_struct *vma,
>> + unsigned long addr,
>> + pte_t *ptep)
>> +{
>> + int pte_num;
>> + pte_t pte = ptep_get(ptep);
>> +
>> + if (!pte_napot(pte))
>> + return ptep_clear_flush(vma, addr, ptep);
>> +
>> + pte_num = napot_pte_num(napot_cont_order(pte));
>> +
>> + return get_clear_contig_flush(vma->vm_mm, addr, ptep, pte_num);
>> +}
>> +
>> +void huge_pte_clear(struct mm_struct *mm,
>> + unsigned long addr,
>> + pte_t *ptep,
>> + unsigned long sz)
>> +{
>> + int i, pte_num = 1;
>> + pte_t pte = READ_ONCE(*ptep);
>> +
>> + if (pte_napot(pte))
>> + pte_num = napot_pte_num(napot_cont_order(pte));
>> +
>> + for (i = 0; i < pte_num; i++, addr += PAGE_SIZE, ptep++)
>> + pte_clear(mm, addr, ptep);
>> +}
>> +
>> +static __init int napot_hugetlbpages_init(void)
>> +{
>> + if (has_svnapot()) {
>> + unsigned long order;
>> +
>> + for_each_napot_order(order)
>> + hugetlb_add_hstate(order);
>> + }
>> + return 0;
>> +}
>> +arch_initcall(napot_hugetlbpages_init);
>> +#endif /*CONFIG_RISCV_ISA_SVNAPOT*/
>> +
>> int pud_huge(pud_t pud)
>> {
>> return pud_leaf(pud);
>> @@ -18,8 +285,17 @@ bool __init arch_hugetlb_valid_size(unsigned long
>> size)
>> return true;
>> else if (IS_ENABLED(CONFIG_64BIT) && size == PUD_SIZE)
>> return true;
>> - else
>> + else if (!has_svnapot())
>> + return false;
>> + else {
>> + unsigned long order;
>> +
>> + for_each_napot_order(order) {
>> + if (size == napot_cont_size(order))
>> + return true;
>> + }
>> return false;
>> + }
>> }
>> #ifdef CONFIG_CONTIG_ALLOC
>
>
> IMO for most of the functions above, we don't need a duplicate of
> arm64/generic functions, otherwise we should really argue why we need
> them. I'm sorry my comment intervenes this late, but I really think it's
> worth giving it a try merging with existing functions, it is quite hard
> to review this way.
Many functions above are really developed with reference to that of
arm64. I agree with you and I will give it a try to add more arch
callbacks in a separate patchset in the future. I think we can merge
these functions now. And we can refactor them and that of arm64 after
we have added new arch callbacks for them in a new patchset. What do
you think of this plan?
Thanks,
Qinglin
More information about the linux-riscv
mailing list