| Message ID | 1591348266-28392-1-git-send-email-maobibo@loongson.cn (mailing list archive) |
|---|---|
| State | Superseded |
| Headers | show |
| Series | [1/2] MIPS: set page access bit with pgprot on some MIPS platform | expand |
On Fri, 5 Jun 2020 17:11:05 +0800 Bibo Mao <maobibo@loongson.cn> wrote: > On MIPS system which has rixi hardware bit, page access bit is not > set in pgrot. For memory reading, there will be one page fault to > allocate physical page; however valid bit is not set, there will > be the second fast tlb-miss fault handling to set valid/access bit. > > This patch set page access/valid bit with pgrot if there is reading > access privilege. It will reduce one tlb-miss handling for memory > reading access. > > The valid/access bit will be cleared in order to track memory > accessing activity. If the page is accessed, tlb-miss fast handling > will set valid/access bit, pte_sw_mkyoung is not necessary in slow > page fault path. This patch removes pte_sw_mkyoung function which > is defined as empty function except MIPS system. > > Signed-off-by: Bibo Mao <maobibo@loongson.cn> > --- Thanks for tracking it down. Could you please make the patch tittle more clear? "Some" looks confuse to me, "systems with RIXI" would be better. - Jiaxun
On 06/05/2020 05:39 PM, Jiaxun Yang wrote: > On Fri, 5 Jun 2020 17:11:05 +0800 > Bibo Mao <maobibo@loongson.cn> wrote: > >> On MIPS system which has rixi hardware bit, page access bit is not >> set in pgrot. For memory reading, there will be one page fault to >> allocate physical page; however valid bit is not set, there will >> be the second fast tlb-miss fault handling to set valid/access bit. >> >> This patch set page access/valid bit with pgrot if there is reading >> access privilege. It will reduce one tlb-miss handling for memory >> reading access. >> >> The valid/access bit will be cleared in order to track memory >> accessing activity. If the page is accessed, tlb-miss fast handling >> will set valid/access bit, pte_sw_mkyoung is not necessary in slow >> page fault path. This patch removes pte_sw_mkyoung function which >> is defined as empty function except MIPS system. >> >> Signed-off-by: Bibo Mao <maobibo@loongson.cn> >> --- > > Thanks for tracking it down. > > Could you please make the patch tittle more clear? > "Some" looks confuse to me, "systems with RIXI" would be better. Sure, will add it. > > - Jiaxun >
On Fri, 5 Jun 2020 17:11:05 +0800 Bibo Mao <maobibo@loongson.cn> wrote: > On MIPS system which has rixi hardware bit, page access bit is not > set in pgrot. For memory reading, there will be one page fault to > allocate physical page; however valid bit is not set, there will > be the second fast tlb-miss fault handling to set valid/access bit. > > This patch set page access/valid bit with pgrot if there is reading > access privilege. It will reduce one tlb-miss handling for memory > reading access. > > The valid/access bit will be cleared in order to track memory > accessing activity. If the page is accessed, tlb-miss fast handling > will set valid/access bit, pte_sw_mkyoung is not necessary in slow > page fault path. This patch removes pte_sw_mkyoung function which > is defined as empty function except MIPS system. > > ... > > --- a/mm/memory.c > +++ b/mm/memory.c > @@ -2704,7 +2704,6 @@ static vm_fault_t wp_page_copy(struct vm_fault *vmf) > } > flush_cache_page(vma, vmf->address, pte_pfn(vmf->orig_pte)); > entry = mk_pte(new_page, vma->vm_page_prot); > - entry = pte_sw_mkyoung(entry); > entry = maybe_mkwrite(pte_mkdirty(entry), vma); > /* > * Clear the pte entry and flush it first, before updating the > @@ -3379,7 +3378,6 @@ static vm_fault_t do_anonymous_page(struct vm_fault *vmf) > __SetPageUptodate(page); > > entry = mk_pte(page, vma->vm_page_prot); > - entry = pte_sw_mkyoung(entry); > if (vma->vm_flags & VM_WRITE) > entry = pte_mkwrite(pte_mkdirty(entry)); > > @@ -3662,7 +3660,6 @@ vm_fault_t alloc_set_pte(struct vm_fault *vmf, struct mem_cgroup *memcg, > > flush_icache_page(vma, page); > entry = mk_pte(page, vma->vm_page_prot); > - entry = pte_sw_mkyoung(entry); > if (write) > entry = maybe_mkwrite(pte_mkdirty(entry), vma); > /* copy-on-write page */ Only affects mips, so cheerily Acked-by: Andrew Morton <akpm@linux-foundation.org>
diff --git a/arch/mips/include/asm/pgtable.h b/arch/mips/include/asm/pgtable.h index 85b39c9..e2452ab 100644 --- a/arch/mips/include/asm/pgtable.h +++ b/arch/mips/include/asm/pgtable.h @@ -25,6 +25,15 @@ struct mm_struct; struct vm_area_struct; +#define __PP _PAGE_PRESENT +#define __NX _PAGE_NO_EXEC +#define __NR _PAGE_NO_READ +#define ___W _PAGE_WRITE +#define ___A _PAGE_ACCESSED +#define ___R (_PAGE_SILENT_READ | _PAGE_ACCESSED) +#define __PC _page_cachable_default + + #define PAGE_NONE __pgprot(_PAGE_PRESENT | _PAGE_NO_READ | \ _page_cachable_default) #define PAGE_SHARED __pgprot(_PAGE_PRESENT | _PAGE_WRITE | \ @@ -414,8 +423,6 @@ static inline pte_t pte_mkyoung(pte_t pte) return pte; } -#define pte_sw_mkyoung pte_mkyoung - #ifdef CONFIG_MIPS_HUGE_TLB_SUPPORT static inline int pte_huge(pte_t pte) { return pte_val(pte) & _PAGE_HUGE; } diff --git a/arch/mips/mm/cache.c b/arch/mips/mm/cache.c index ad6df1c..f814e43 100644 --- a/arch/mips/mm/cache.c +++ b/arch/mips/mm/cache.c @@ -158,23 +158,23 @@ void __update_cache(unsigned long address, pte_t pte) static inline void setup_protection_map(void) { if (cpu_has_rixi) { - protection_map[0] = __pgprot(_page_cachable_default | _PAGE_PRESENT | _PAGE_NO_EXEC | _PAGE_NO_READ); - protection_map[1] = __pgprot(_page_cachable_default | _PAGE_PRESENT | _PAGE_NO_EXEC); - protection_map[2] = __pgprot(_page_cachable_default | _PAGE_PRESENT | _PAGE_NO_EXEC | _PAGE_NO_READ); - protection_map[3] = __pgprot(_page_cachable_default | _PAGE_PRESENT | _PAGE_NO_EXEC); - protection_map[4] = __pgprot(_page_cachable_default | _PAGE_PRESENT); - protection_map[5] = __pgprot(_page_cachable_default | _PAGE_PRESENT); - protection_map[6] = __pgprot(_page_cachable_default | _PAGE_PRESENT); - protection_map[7] = __pgprot(_page_cachable_default | _PAGE_PRESENT); - - protection_map[8] = __pgprot(_page_cachable_default | _PAGE_PRESENT | _PAGE_NO_EXEC | _PAGE_NO_READ); - protection_map[9] = __pgprot(_page_cachable_default | _PAGE_PRESENT | _PAGE_NO_EXEC); - protection_map[10] = __pgprot(_page_cachable_default | _PAGE_PRESENT | _PAGE_NO_EXEC | _PAGE_WRITE | _PAGE_NO_READ); - protection_map[11] = __pgprot(_page_cachable_default | _PAGE_PRESENT | _PAGE_NO_EXEC | _PAGE_WRITE); - protection_map[12] = __pgprot(_page_cachable_default | _PAGE_PRESENT); - protection_map[13] = __pgprot(_page_cachable_default | _PAGE_PRESENT); - protection_map[14] = __pgprot(_page_cachable_default | _PAGE_PRESENT | _PAGE_WRITE); - protection_map[15] = __pgprot(_page_cachable_default | _PAGE_PRESENT | _PAGE_WRITE); + protection_map[0] = __pgprot(__PC | __PP | __NX | __NR); + protection_map[1] = __pgprot(__PC | __PP | __NX | ___R); + protection_map[2] = __pgprot(__PC | __PP | __NX | __NR); + protection_map[3] = __pgprot(__PC | __PP | __NX | ___R); + protection_map[4] = __pgprot(__PC | __PP | ___R); + protection_map[5] = __pgprot(__PC | __PP | ___R); + protection_map[6] = __pgprot(__PC | __PP | ___R); + protection_map[7] = __pgprot(__PC | __PP | ___R); + + protection_map[8] = __pgprot(__PC | __PP | __NX | __NR); + protection_map[9] = __pgprot(__PC | __PP | __NX | ___R); + protection_map[10] = __pgprot(__PC | __PP | __NX | ___W | __NR); + protection_map[11] = __pgprot(__PC | __PP | __NX | ___W | ___R); + protection_map[12] = __pgprot(__PC | __PP | ___R); + protection_map[13] = __pgprot(__PC | __PP | ___R); + protection_map[14] = __pgprot(__PC | __PP | ___W | ___R); + protection_map[15] = __pgprot(__PC | __PP | ___W | ___R); } else { protection_map[0] = PAGE_NONE; diff --git a/include/asm-generic/pgtable.h b/include/asm-generic/pgtable.h index b5278ec..fa5c73f 100644 --- a/include/asm-generic/pgtable.h +++ b/include/asm-generic/pgtable.h @@ -244,22 +244,6 @@ static inline void ptep_set_wrprotect(struct mm_struct *mm, unsigned long addres } #endif -/* - * On some architectures hardware does not set page access bit when accessing - * memory page, it is responsibilty of software setting this bit. It brings - * out extra page fault penalty to track page access bit. For optimization page - * access bit can be set during all page fault flow on these arches. - * To be differentiate with macro pte_mkyoung, this macro is used on platforms - * where software maintains page access bit. - */ -#ifndef pte_sw_mkyoung -static inline pte_t pte_sw_mkyoung(pte_t pte) -{ - return pte; -} -#define pte_sw_mkyoung pte_sw_mkyoung -#endif - #ifndef pte_savedwrite #define pte_savedwrite pte_write #endif diff --git a/mm/memory.c b/mm/memory.c index c7c8960..8bb31c4 100644 --- a/mm/memory.c +++ b/mm/memory.c @@ -2704,7 +2704,6 @@ static vm_fault_t wp_page_copy(struct vm_fault *vmf) } flush_cache_page(vma, vmf->address, pte_pfn(vmf->orig_pte)); entry = mk_pte(new_page, vma->vm_page_prot); - entry = pte_sw_mkyoung(entry); entry = maybe_mkwrite(pte_mkdirty(entry), vma); /* * Clear the pte entry and flush it first, before updating the @@ -3379,7 +3378,6 @@ static vm_fault_t do_anonymous_page(struct vm_fault *vmf) __SetPageUptodate(page); entry = mk_pte(page, vma->vm_page_prot); - entry = pte_sw_mkyoung(entry); if (vma->vm_flags & VM_WRITE) entry = pte_mkwrite(pte_mkdirty(entry)); @@ -3662,7 +3660,6 @@ vm_fault_t alloc_set_pte(struct vm_fault *vmf, struct mem_cgroup *memcg, flush_icache_page(vma, page); entry = mk_pte(page, vma->vm_page_prot); - entry = pte_sw_mkyoung(entry); if (write) entry = maybe_mkwrite(pte_mkdirty(entry), vma); /* copy-on-write page */
On MIPS system which has rixi hardware bit, page access bit is not set in pgrot. For memory reading, there will be one page fault to allocate physical page; however valid bit is not set, there will be the second fast tlb-miss fault handling to set valid/access bit. This patch set page access/valid bit with pgrot if there is reading access privilege. It will reduce one tlb-miss handling for memory reading access. The valid/access bit will be cleared in order to track memory accessing activity. If the page is accessed, tlb-miss fast handling will set valid/access bit, pte_sw_mkyoung is not necessary in slow page fault path. This patch removes pte_sw_mkyoung function which is defined as empty function except MIPS system. Signed-off-by: Bibo Mao <maobibo@loongson.cn> --- arch/mips/include/asm/pgtable.h | 11 +++++++++-- arch/mips/mm/cache.c | 34 +++++++++++++++++----------------- include/asm-generic/pgtable.h | 16 ---------------- mm/memory.c | 3 --- 4 files changed, 26 insertions(+), 38 deletions(-)