Re: [PATCH v2] arm64: Add support for PTE contiguous bit.

From: Steve Capper
Date: Tue Oct 20 2015 - 08:16:39 EST

On Mon, Oct 19, 2015 at 04:09:09PM -0400, David Woods wrote:
> The arm64 MMU supports a Contiguous bit which is a hint that the TTE
> is one of a set of contiguous entries which can be cached in a single
> TLB entry. Supporting this bit adds new intermediate huge page sizes.
>
> The set of huge page sizes available depends on the base page size.
> Without using contiguous pages the huge page sizes are as follows.
>
> 4KB: 2MB 1GB
> 64KB: 512MB
>
> With a 4KB granule, the contiguous bit groups together sets of 16 pages
> and with a 64KB granule it groups sets of 32 pages. This enables two new
> huge page sizes in each case, so that the full set of available sizes
> is as follows.
>
> 4KB: 64KB 2MB 32MB 1GB
> 64KB: 2MB 512MB 16GB
>
> If a 16KB granule is used then the contiguous bit groups 128 pages
> at the PTE level and 32 pages at the PMD level.
>
> If the base page size is set to 64KB then 2MB pages are enabled by
> default. It is possible in the future to make 2MB the default huge
> page size for both 4KB and 64KB granules.

Thank you for the V2 David,
I have some comments below.

I would recommend running the next version of this series through
the libhugetlbfs test suite, as that may pick up a few things too.

Cheers,
--
Steve

>
> Signed-off-by: David Woods <dwoods@xxxxxxxxxx>
> Reviewed-by: Chris Metcalf <cmetcalf@xxxxxxxxxx>
> ---
> arch/arm64/Kconfig | 3 -
> arch/arm64/include/asm/hugetlb.h | 30 ++---
> arch/arm64/include/asm/pgtable-hwdef.h | 20 ++++
> arch/arm64/include/asm/pgtable.h | 33 +++++-
> arch/arm64/mm/hugetlbpage.c | 211 ++++++++++++++++++++++++++++++++-
> 5 files changed, 272 insertions(+), 25 deletions(-)
>
> diff --git a/arch/arm64/Kconfig b/arch/arm64/Kconfig
> index 07d1811..3aa151d 100644
> --- a/arch/arm64/Kconfig
> +++ b/arch/arm64/Kconfig
> @@ -464,9 +464,6 @@ config HW_PERF_EVENTS
> config SYS_SUPPORTS_HUGETLBFS
> def_bool y
>
> -config ARCH_WANT_GENERAL_HUGETLB
> - def_bool y
> -
> config ARCH_WANT_HUGE_PMD_SHARE
> def_bool y if !ARM64_64K_PAGES
>
> diff --git a/arch/arm64/include/asm/hugetlb.h b/arch/arm64/include/asm/hugetlb.h
> index bb4052e..2b153a9 100644
> --- a/arch/arm64/include/asm/hugetlb.h
> +++ b/arch/arm64/include/asm/hugetlb.h
> @@ -26,12 +26,6 @@ static inline pte_t huge_ptep_get(pte_t *ptep)
> return *ptep;
> }
>
> -static inline void set_huge_pte_at(struct mm_struct *mm, unsigned long addr,
> - pte_t *ptep, pte_t pte)
> -{
> - set_pte_at(mm, addr, ptep, pte);
> -}
> -
> static inline void huge_ptep_clear_flush(struct vm_area_struct *vma,
> unsigned long addr, pte_t *ptep)
> {
> @@ -44,19 +38,6 @@ static inline void huge_ptep_set_wrprotect(struct mm_struct *mm,
> ptep_set_wrprotect(mm, addr, ptep);
> }
>
> -static inline pte_t huge_ptep_get_and_clear(struct mm_struct *mm,
> - unsigned long addr, pte_t *ptep)
> -{
> - return ptep_get_and_clear(mm, addr, ptep);
> -}
> -
> -static inline int huge_ptep_set_access_flags(struct vm_area_struct *vma,
> - unsigned long addr, pte_t *ptep,
> - pte_t pte, int dirty)
> -{
> - return ptep_set_access_flags(vma, addr, ptep, pte, dirty);
> -}
> -
> static inline void hugetlb_free_pgd_range(struct mmu_gather *tlb,
> unsigned long addr, unsigned long end,
> unsigned long floor,
> @@ -97,4 +78,15 @@ static inline void arch_clear_hugepage_flags(struct page *page)
> clear_bit(PG_dcache_clean, &page->flags);
> }
>
> +extern pte_t arch_make_huge_pte(pte_t entry, struct vm_area_struct *vma,
> + struct page *page, int writable);
> +#define arch_make_huge_pte arch_make_huge_pte
> +extern void set_huge_pte_at(struct mm_struct *mm, unsigned long addr,
> + pte_t *ptep, pte_t pte);
> +extern int huge_ptep_set_access_flags(struct vm_area_struct *vma,
> + unsigned long addr, pte_t *ptep,
> + pte_t pte, int dirty);
> +extern pte_t huge_ptep_get_and_clear(struct mm_struct *mm,
> + unsigned long addr, pte_t *ptep);
> +
> #endif /* __ASM_HUGETLB_H */
> diff --git a/arch/arm64/include/asm/pgtable-hwdef.h b/arch/arm64/include/asm/pgtable-hwdef.h
> index 24154b0..1b921a5 100644
> --- a/arch/arm64/include/asm/pgtable-hwdef.h
> +++ b/arch/arm64/include/asm/pgtable-hwdef.h
> @@ -55,6 +55,24 @@
> #define SECTION_MASK (~(SECTION_SIZE-1))
>
> /*
> + * Contiguous page definitions.
> + */
> +#ifdef CONFIG_ARM64_64K_PAGES
> +#define CONT_PTE_SHIFT 5
> +#define CONT_PMD_SHIFT 5
> +#else
> +#define CONT_PTE_SHIFT 4
> +#define CONT_PMD_SHIFT 4
> +#endif
> +
> +#define CONT_PTES (1 << CONT_PTE_SHIFT)
> +#define CONT_PTE_SIZE (CONT_PTES * PAGE_SIZE)
> +#define CONT_PTE_MASK (~(CONT_PTE_SIZE - 1))
> +#define CONT_PMDS (1 << CONT_PMD_SHIFT)
> +#define CONT_PMD_SIZE (CONT_PMDS * PMD_SIZE)
> +#define CONT_PMD_MASK (~(CONT_PMD_SIZE - 1))
> +
> +/*
> * Hardware page table definitions.
> *
> * Level 1 descriptor (PUD).
> @@ -83,6 +101,7 @@
> #define PMD_SECT_S (_AT(pmdval_t, 3) << 8)
> #define PMD_SECT_AF (_AT(pmdval_t, 1) << 10)
> #define PMD_SECT_NG (_AT(pmdval_t, 1) << 11)
> +#define PMD_SECT_CONT (_AT(pmdval_t, 1) << 52)
> #define PMD_SECT_PXN (_AT(pmdval_t, 1) << 53)
> #define PMD_SECT_UXN (_AT(pmdval_t, 1) << 54)
>
> @@ -105,6 +124,7 @@
> #define PTE_AF (_AT(pteval_t, 1) << 10) /* Access Flag */
> #define PTE_NG (_AT(pteval_t, 1) << 11) /* nG */
> #define PTE_DBM (_AT(pteval_t, 1) << 51) /* Dirty Bit Management */
> +#define PTE_CONT (_AT(pteval_t, 1) << 52) /* Contiguous */
> #define PTE_PXN (_AT(pteval_t, 1) << 53) /* Privileged XN */
> #define PTE_UXN (_AT(pteval_t, 1) << 54) /* User XN */
>
> diff --git a/arch/arm64/include/asm/pgtable.h b/arch/arm64/include/asm/pgtable.h
> index 26b0666..cf079a1 100644
> --- a/arch/arm64/include/asm/pgtable.h
> +++ b/arch/arm64/include/asm/pgtable.h
> @@ -140,6 +140,7 @@ extern struct page *empty_zero_page;
> #define pte_special(pte) (!!(pte_val(pte) & PTE_SPECIAL))
> #define pte_write(pte) (!!(pte_val(pte) & PTE_WRITE))
> #define pte_exec(pte) (!(pte_val(pte) & PTE_UXN))
> +#define pte_cont(pte) (!!(pte_val(pte) & PTE_CONT))
>
> #ifdef CONFIG_ARM64_HW_AFDBM
> #define pte_hw_dirty(pte) (pte_write(pte) && !(pte_val(pte) & PTE_RDONLY))
> @@ -202,6 +203,18 @@ static inline pte_t pte_mkspecial(pte_t pte)
> return set_pte_bit(pte, __pgprot(PTE_SPECIAL));
> }
>
> +static inline pte_t pte_mkcont(pte_t pte)
> +{
> + pte = set_pte_bit(pte, __pgprot(PTE_CONT));
> + return set_pte_bit(pte, __pgprot(PTE_TYPE_PAGE));
> + return pte;

The second return should be removed.

> +}
> +
> +static inline pmd_t pmd_mkcont(pmd_t pmd)
> +{
> + return __pmd(pmd_val(pmd) | PMD_SECT_CONT);
> +}
> +
> static inline void set_pte(pte_t *ptep, pte_t pte)
> {
> *ptep = pte;
> @@ -271,7 +284,7 @@ static inline void set_pte_at(struct mm_struct *mm, unsigned long addr,
> /*
> * Hugetlb definitions.
> */
> -#define HUGE_MAX_HSTATE 2
> +#define HUGE_MAX_HSTATE ((2 * CONFIG_PGTABLE_LEVELS) - 1)

Not sure about this definition. I would just go with the maximum possible
which is for a 4KB granule:
1 x 1GB pud
1 x 2MB pmd
16 x 2MB pmds
16 x 4KB ptes

So 4 for now?


> #define HPAGE_SHIFT PMD_SHIFT
> #define HPAGE_SIZE (_AC(1, UL) << HPAGE_SHIFT)
> #define HPAGE_MASK (~(HPAGE_SIZE - 1))
> @@ -496,7 +509,7 @@ static inline pud_t *pud_offset(pgd_t *pgd, unsigned long addr)
> static inline pte_t pte_modify(pte_t pte, pgprot_t newprot)
> {
> const pteval_t mask = PTE_USER | PTE_PXN | PTE_UXN | PTE_RDONLY |
> - PTE_PROT_NONE | PTE_VALID | PTE_WRITE;
> + PTE_PROT_NONE | PTE_VALID | PTE_WRITE | PTE_CONT;

Why has PTE_CONT been added to the pte_modify mask? This will allow
functions such as mprotect to remove the PTE_CONT bit.

> /* preserve the hardware dirty information */
> if (pte_hw_dirty(pte))
> pte = pte_mkdirty(pte);
> @@ -509,6 +522,22 @@ static inline pmd_t pmd_modify(pmd_t pmd, pgprot_t newprot)
> return pte_pmd(pte_modify(pmd_pte(pmd), newprot));
> }
>
> +static inline pte_t pte_modify_pfn(pte_t pte, unsigned long newpfn)
> +{
> + const pteval_t mask = PHYS_MASK & PAGE_MASK;
> +
> + pte_val(pte) = pfn_pte(newpfn, (pte_val(pte) & ~mask));
> + return pte;
> +}
> +
> +static inline pmd_t pmd_modify_pfn(pmd_t pmd, unsigned long newpfn)
> +{
> + const pmdval_t mask = PHYS_MASK & PAGE_MASK;
> +
> + pmd = pfn_pmd(newpfn, (pmd_val(pmd) & ~mask));
> + return pmd;
> +}

pte_modify_pfn and pmd_modify_pfn aren't referenced anywhere in the
patch so should be removed.

> +
> #ifdef CONFIG_ARM64_HW_AFDBM
> /*
> * Atomic pte/pmd modifications.
> diff --git a/arch/arm64/mm/hugetlbpage.c b/arch/arm64/mm/hugetlbpage.c
> index 383b03f..20fd34c 100644
> --- a/arch/arm64/mm/hugetlbpage.c
> +++ b/arch/arm64/mm/hugetlbpage.c
> @@ -41,6 +41,201 @@ int pud_huge(pud_t pud)
> #endif
> }
>
> +static int find_num_contig(struct mm_struct *mm, unsigned long addr,
> + pte_t *ptep, pte_t pte, size_t *pgsize)
> +{
> + pgd_t *pgd = pgd_offset(mm, addr);
> + pud_t *pud;
> + pmd_t *pmd;
> +
> + if (!pte_cont(pte))
> + return 1;
> +
> + pud = pud_offset(pgd, addr);
> + pmd = pmd_offset(pud, addr);

We need to check for pgd_present and pud_present as we walk.
I would be tempted to VM_BUG_ON if they are in an unexpected state.

> + if ((pte_t *)pmd == ptep) {
> + *pgsize = PMD_SIZE;
> + return CONT_PMDS;
> + }

I would check for pmd_present and VM_BUG_ON if it wasn't in an expected
state.

> + *pgsize = PAGE_SIZE;
> + return CONT_PTES;
> +}

Another approach would be something like:

struct vm_area_struct *vma = find_vma(mm, addr);
struct hstate *h = hstate_vma(vma);
size_t size = hpage_size(h);

But I think looking at the page table entries like you've done (with
some checking) may be a little better as it can supply some more robust
debugging with DEBUG_VM selected (and it doesn't need to find_vma).


> +
> +extern void set_huge_pte_at(struct mm_struct *mm, unsigned long addr,
> + pte_t *ptep, pte_t pte)

We don't need this extern.

> +{
> + size_t pgsize;
> + int ncontig = find_num_contig(mm, addr, ptep, pte, &pgsize);
> +
> + if (ncontig == 1) {
> + set_pte_at(mm, addr, ptep, pte);

We can return early here and avoid a level of indentation below.

> + } else {
> + int i;
> + unsigned long pfn = pte_pfn(pte);
> + pgprot_t hugeprot =
> + __pgprot(pte_val(pfn_pte(pfn, 0) ^ pte_val(pte)));
> + for (i = 0; i < ncontig; i++) {
> + pr_debug("%s: set pte %p to 0x%llx\n", __func__, ptep,
> + pfn_pte(pfn, hugeprot));
> + set_pte_at(mm, addr, ptep, pfn_pte(pfn, hugeprot));
> + ptep++;
> + pfn += pgsize / PAGE_SIZE;

nit: pgsize >> PAGE_SHIFT

> + addr += pgsize;
> + }
> + }
> +}

I see... so the contiguous pte and pmd cases are folded together.
The pgsize variable name could be changed, perhaps something like blocksize?
(I am terrible at picking names though :-)).

> +
> +pte_t *huge_pte_alloc(struct mm_struct *mm,
> + unsigned long addr, unsigned long sz)
> +{
> + pgd_t *pgd;
> + pud_t *pud;
> + pte_t *pte = NULL;
> +
> + pr_debug("%s: addr:0x%lx sz:0x%lx\n", __func__, addr, sz);
> + pgd = pgd_offset(mm, addr);
> + pud = pud_alloc(mm, pgd, addr);

Probably better to simplify the levels of indentation with:
if (!pud)
return NULL;
(or goto out before your pr_debug)

> + if (pud) {

Perhaps better to do something with switch(sz) below?

> + if (sz == PUD_SIZE) {
> + pte = (pte_t *)pud;
> + } else if (sz == (PAGE_SIZE * CONT_PTES)) {
> + pmd_t *pmd = pmd_alloc(mm, pud, addr);
> +
> + WARN_ON(addr & (sz - 1));
> + pte = pte_alloc_map(mm, NULL, pmd, addr);
> + } else if (sz == PMD_SIZE) {
> +#ifdef CONFIG_ARCH_WANT_HUGE_PMD_SHARE
> + if (pud_none(*pud))
> + pte = huge_pmd_share(mm, addr, pud);
> + else
> +#endif

This can be simplified to something like:

if (IS_ENABLED(CONFIG_ARCH_WANT_HUGE_PMD_SHARE)
&& pud_none(*pud))
else

So we can remove the preprocessor macros.

> + pte = (pte_t *)pmd_alloc(mm, pud, addr);
> + } else if (sz == (PMD_SIZE * CONT_PMDS)) {
> + pmd_t *pmd;
> +
> + pmd = pmd_alloc(mm, pud, addr);
> + WARN_ON(addr & (sz - 1));
> + return (pte_t *)pmd;
> + }
> + }
> +
> + pr_debug("%s: addr:0x%lx sz:0x%lx ret pte=%p/0x%llx\n", __func__, addr,
> + sz, pte, pte_val(*pte));
> + return pte;
> +}
> +
> +pte_t *huge_pte_offset(struct mm_struct *mm, unsigned long addr)
> +{
> + pgd_t *pgd;
> + pud_t *pud;
> + pmd_t *pmd = NULL;
> + pte_t *pte = NULL;
> +
> + pgd = pgd_offset(mm, addr);
> + pr_debug("%s: addr:0x%lx pgd:%p\n", __func__, addr, pgd);
> + if (pgd_present(*pgd)) {

Again drop a level of indentation with:
if (!pgd_present(*pgd))
return NULL;

Similarly for pud_present and pmd_present.

> + pud = pud_offset(pgd, addr);
> + if (pud_present(*pud)) {
> + if (pud_huge(*pud))
> + return (pte_t *)pud;
> + pmd = pmd_offset(pud, addr);
> + if (pmd_present(*pmd)) {
> + if (pte_cont(pmd_pte(*pmd))) {
> + pmd = pmd_offset(
> + pud, (addr & CONT_PMD_MASK));
> + return (pte_t *)pmd;
> + }
> + if (pmd_huge(*pmd))
> + return (pte_t *)pmd;
> + pte = pte_offset_kernel(pmd, addr);
> + if (pte_present(*pte) && pte_cont(*pte)) {
> + pte = pte_offset_kernel(
> + pmd, (addr & CONT_PTE_MASK));
> + }
> + return pte;
> + }
> + }
> + }
> + return (pte_t *) NULL;
> +}
> +
> +pte_t arch_make_huge_pte(pte_t entry, struct vm_area_struct *vma,
> + struct page *page, int writable)
> +{
> + size_t pagesize = huge_page_size(hstate_vma(vma));
> +

I would go for switch(pagesize) here.

> + if (pagesize == CONT_PTE_SIZE) {
> + entry = pte_mkcont(entry);
> + } else if (pagesize == CONT_PMD_SIZE) {
> + entry = pmd_pte(pmd_mkcont(pte_pmd(entry)));
> + } else if (pagesize != PUD_SIZE && pagesize != PMD_SIZE) {
> + pr_warn("%s: unrecognized huge page size 0x%lx\n",
> + __func__, pagesize);
> + }
> + return entry;
> +}
> +
> +extern pte_t huge_ptep_get_and_clear(struct mm_struct *mm,
> + unsigned long addr, pte_t *ptep)
> +{
> + pte_t pte = {0};

nit: Do we need an initial value for pte?

> +
> + if (pte_cont(*ptep)) {
> + int ncontig, i;
> + size_t pgsize;
> + pte_t *cpte;
> + bool is_dirty = false;
> +
> + cpte = huge_pte_offset(mm, addr);
> + ncontig = find_num_contig(mm, addr, cpte,
> + pte_val(*cpte), &pgsize);
> + /* save the 1st pte to return */
> + pte = ptep_get_and_clear(mm, addr, cpte);
> + for (i = 1; i < ncontig; ++i) {
> + if (pte_dirty(ptep_get_and_clear(mm, addr, ++cpte)))
> + is_dirty = true;
> + }

Nice, we are keeping track of the dirty state. This looks to me like
it *should* work well with the dirty bit management patch that Catalin
introduced:
2f4b829 arm64: Add support for hardware updates of the access and dirty pte bits

Because ptep_get_and_clear will atomically get and clear the pte with
respect to the hardware dirty bit management thus we don't lose any
dirty information. huge_pte_dirty is then called on the extracted pte
by core code.

For a contiguous set of ptes/pmds the individual entry will be dirtied
by DBM rather than the complete set so it's good to check them all for
dirty when going through a get and clear.

Technically we don't need to track dirty if CONFIG_ARM64_HW_AFDBM is
not defined as the core code will fault and modify the entire set of
ptes otherwise.

I would be tempted to keep this code as is, but add a comment that
tracking the dirty variable here helps for when we switch on
CONFIG_ARM64_HW_AFDBM.

> + if (is_dirty)
> + return pte_mkdirty(pte);
> + else
> + return pte;
> + } else {
> + return ptep_get_and_clear(mm, addr, ptep);
> + }
> +}
> +
> +int huge_ptep_set_access_flags(struct vm_area_struct *vma,
> + unsigned long addr, pte_t *ptep,
> + pte_t pte, int dirty)
> +{
> + pte_t *cpte;
> +
> + if (pte_cont(pte)) {
> + int ncontig, i, changed = 0;
> + size_t pgsize = 0;
> + unsigned long pfn = pte_pfn(pte);
> + /* Select all bits except the pfn */
> + pgprot_t hugeprot =
> + __pgprot(pte_val(pfn_pte(pfn, 0) ^ pte_val(pte)));
> +
> + cpte = huge_pte_offset(vma->vm_mm, addr);
> + pfn = pte_pfn(*cpte);
> + ncontig = find_num_contig(vma->vm_mm, addr, cpte,
> + pte_val(*cpte), &pgsize);
> + for (i = 0; i < ncontig; ++i, ++cpte) {
> + changed = ptep_set_access_flags(vma, addr, cpte,
> + pfn_pte(pfn,
> + hugeprot),
> + dirty);
> + pfn += pgsize / PAGE_SIZE;

nit: pgsize >> PAGE_SHIFT

> + }

Similar thing here, we are folding the cont pte and pmd logic together,
probably best go for something like blocksize instead of pgsize.

> + return changed;
> + } else {
> + return ptep_set_access_flags(vma, addr, ptep, pte, dirty);
> + }
> +}
> +
> +
> static __init int setup_hugepagesz(char *opt)
> {
> unsigned long ps = memparse(opt, &opt);
> @@ -48,10 +243,24 @@ static __init int setup_hugepagesz(char *opt)
> hugetlb_add_hstate(PMD_SHIFT - PAGE_SHIFT);

Again switch statement here I think would be clearer.

> } else if (ps == PUD_SIZE) {
> hugetlb_add_hstate(PUD_SHIFT - PAGE_SHIFT);
> + } else if (ps == (PAGE_SIZE * CONT_PTES)) {
> + hugetlb_add_hstate(CONT_PTE_SHIFT);
> + } else if (ps == (PMD_SIZE * CONT_PMDS)) {
> + hugetlb_add_hstate((PMD_SHIFT + CONT_PMD_SHIFT) - PAGE_SHIFT);
> } else {
> - pr_err("hugepagesz: Unsupported page size %lu M\n", ps >> 20);
> + pr_err("hugepagesz: Unsupported page size %lu K\n", ps >> 10);
> return 0;
> }
> return 1;
> }
> __setup("hugepagesz=", setup_hugepagesz);
> +
> +#ifdef CONFIG_ARM64_64K_PAGES
> +static __init int add_default_hugepagesz(void)
> +{
> + if (size_to_hstate(CONT_PTES * PAGE_SIZE) == NULL)
> + hugetlb_add_hstate(CONT_PMD_SHIFT);
> + return 0;
> +}
> +arch_initcall(add_default_hugepagesz);
> +#endif
Why is this initcall defined? Was it for testing?

I think we are missing a few functions:
huge_ptep_set_access_flags
huge_ptep_set_wrprotect
huge_ptep_clear_flush

These functions need to loop through the contiguous set of ptes
or pmds. They should call into the ptep_ equivalents as they will
then work with the DBM patch.

> --
> 2.1.2
>
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/