From 35149e2129fe34fc8cb5917e1ecf5156b0fa3415 Mon Sep 17 00:00:00 2001 From: Anthony Liguori Date: Wed, 2 Apr 2008 14:46:56 -0500 Subject: KVM: MMU: Don't assume struct page for x86 This patch introduces a gfn_to_pfn() function and corresponding functions like kvm_release_pfn_dirty(). Using these new functions, we can modify the x86 MMU to no longer assume that it can always get a struct page for any given gfn. We don't want to eliminate gfn_to_page() entirely because a number of places assume they can do gfn_to_page() and then kmap() the results. When we support IO memory, gfn_to_page() will fail for IO pages although gfn_to_pfn() will succeed. This does not implement support for avoiding reference counting for reserved RAM or for IO memory. However, it should make those things pretty straight forward. Since we're only introducing new common symbols, I don't think it will break the non-x86 architectures but I haven't tested those. I've tested Intel, AMD, NPT, and hugetlbfs with Windows and Linux guests. [avi: fix overflow when shifting left pfns by adding casts] Signed-off-by: Anthony Liguori Signed-off-by: Avi Kivity --- arch/x86/kvm/mmu.c | 89 ++++++++++++++++++++++------------------------ arch/x86/kvm/paging_tmpl.h | 26 +++++++------- 2 files changed, 56 insertions(+), 59 deletions(-) (limited to 'arch/x86') diff --git a/arch/x86/kvm/mmu.c b/arch/x86/kvm/mmu.c index c89bf230af6..078a7f1ac34 100644 --- a/arch/x86/kvm/mmu.c +++ b/arch/x86/kvm/mmu.c @@ -240,11 +240,9 @@ static int is_rmap_pte(u64 pte) return is_shadow_present_pte(pte); } -static struct page *spte_to_page(u64 pte) +static pfn_t spte_to_pfn(u64 pte) { - hfn_t hfn = (pte & PT64_BASE_ADDR_MASK) >> PAGE_SHIFT; - - return pfn_to_page(hfn); + return (pte & PT64_BASE_ADDR_MASK) >> PAGE_SHIFT; } static gfn_t pse36_gfn_delta(u32 gpte) @@ -541,20 +539,20 @@ static void rmap_remove(struct kvm *kvm, u64 *spte) struct kvm_rmap_desc *desc; struct kvm_rmap_desc *prev_desc; struct kvm_mmu_page *sp; - struct page *page; + pfn_t pfn; unsigned long *rmapp; int i; if (!is_rmap_pte(*spte)) return; sp = page_header(__pa(spte)); - page = spte_to_page(*spte); + pfn = spte_to_pfn(*spte); if (*spte & PT_ACCESSED_MASK) - mark_page_accessed(page); + kvm_set_pfn_accessed(pfn); if (is_writeble_pte(*spte)) - kvm_release_page_dirty(page); + kvm_release_pfn_dirty(pfn); else - kvm_release_page_clean(page); + kvm_release_pfn_clean(pfn); rmapp = gfn_to_rmap(kvm, sp->gfns[spte - sp->spt], is_large_pte(*spte)); if (!*rmapp) { printk(KERN_ERR "rmap_remove: %p %llx 0->BUG\n", spte, *spte); @@ -635,11 +633,11 @@ static void rmap_write_protect(struct kvm *kvm, u64 gfn) spte = rmap_next(kvm, rmapp, spte); } if (write_protected) { - struct page *page; + pfn_t pfn; spte = rmap_next(kvm, rmapp, NULL); - page = spte_to_page(*spte); - SetPageDirty(page); + pfn = spte_to_pfn(*spte); + kvm_set_pfn_dirty(pfn); } /* check for huge page mappings */ @@ -1036,7 +1034,7 @@ static void mmu_set_spte(struct kvm_vcpu *vcpu, u64 *shadow_pte, unsigned pt_access, unsigned pte_access, int user_fault, int write_fault, int dirty, int *ptwrite, int largepage, gfn_t gfn, - struct page *page, bool speculative) + pfn_t pfn, bool speculative) { u64 spte; int was_rmapped = 0; @@ -1058,10 +1056,9 @@ static void mmu_set_spte(struct kvm_vcpu *vcpu, u64 *shadow_pte, child = page_header(pte & PT64_BASE_ADDR_MASK); mmu_page_remove_parent_pte(child, shadow_pte); - } else if (page != spte_to_page(*shadow_pte)) { + } else if (pfn != spte_to_pfn(*shadow_pte)) { pgprintk("hfn old %lx new %lx\n", - page_to_pfn(spte_to_page(*shadow_pte)), - page_to_pfn(page)); + spte_to_pfn(*shadow_pte), pfn); rmap_remove(vcpu->kvm, shadow_pte); } else { if (largepage) @@ -1090,7 +1087,7 @@ static void mmu_set_spte(struct kvm_vcpu *vcpu, u64 *shadow_pte, if (largepage) spte |= PT_PAGE_SIZE_MASK; - spte |= page_to_phys(page); + spte |= (u64)pfn << PAGE_SHIFT; if ((pte_access & ACC_WRITE_MASK) || (write_fault && !is_write_protection(vcpu) && !user_fault)) { @@ -1135,12 +1132,12 @@ unshadowed: if (!was_rmapped) { rmap_add(vcpu, shadow_pte, gfn, largepage); if (!is_rmap_pte(*shadow_pte)) - kvm_release_page_clean(page); + kvm_release_pfn_clean(pfn); } else { if (was_writeble) - kvm_release_page_dirty(page); + kvm_release_pfn_dirty(pfn); else - kvm_release_page_clean(page); + kvm_release_pfn_clean(pfn); } if (!ptwrite || !*ptwrite) vcpu->arch.last_pte_updated = shadow_pte; @@ -1151,7 +1148,7 @@ static void nonpaging_new_cr3(struct kvm_vcpu *vcpu) } static int __direct_map(struct kvm_vcpu *vcpu, gpa_t v, int write, - int largepage, gfn_t gfn, struct page *page, + int largepage, gfn_t gfn, pfn_t pfn, int level) { hpa_t table_addr = vcpu->arch.mmu.root_hpa; @@ -1166,13 +1163,13 @@ static int __direct_map(struct kvm_vcpu *vcpu, gpa_t v, int write, if (level == 1) { mmu_set_spte(vcpu, &table[index], ACC_ALL, ACC_ALL, - 0, write, 1, &pt_write, 0, gfn, page, false); + 0, write, 1, &pt_write, 0, gfn, pfn, false); return pt_write; } if (largepage && level == 2) { mmu_set_spte(vcpu, &table[index], ACC_ALL, ACC_ALL, - 0, write, 1, &pt_write, 1, gfn, page, false); + 0, write, 1, &pt_write, 1, gfn, pfn, false); return pt_write; } @@ -1187,7 +1184,7 @@ static int __direct_map(struct kvm_vcpu *vcpu, gpa_t v, int write, 1, ACC_ALL, &table[index]); if (!new_table) { pgprintk("nonpaging_map: ENOMEM\n"); - kvm_release_page_clean(page); + kvm_release_pfn_clean(pfn); return -ENOMEM; } @@ -1202,8 +1199,7 @@ static int nonpaging_map(struct kvm_vcpu *vcpu, gva_t v, int write, gfn_t gfn) { int r; int largepage = 0; - - struct page *page; + pfn_t pfn; down_read(¤t->mm->mmap_sem); if (is_largepage_backed(vcpu, gfn & ~(KVM_PAGES_PER_HPAGE-1))) { @@ -1211,18 +1207,18 @@ static int nonpaging_map(struct kvm_vcpu *vcpu, gva_t v, int write, gfn_t gfn) largepage = 1; } - page = gfn_to_page(vcpu->kvm, gfn); + pfn = gfn_to_pfn(vcpu->kvm, gfn); up_read(¤t->mm->mmap_sem); /* mmio */ - if (is_error_page(page)) { - kvm_release_page_clean(page); + if (is_error_pfn(pfn)) { + kvm_release_pfn_clean(pfn); return 1; } spin_lock(&vcpu->kvm->mmu_lock); kvm_mmu_free_some_pages(vcpu); - r = __direct_map(vcpu, v, write, largepage, gfn, page, + r = __direct_map(vcpu, v, write, largepage, gfn, pfn, PT32E_ROOT_LEVEL); spin_unlock(&vcpu->kvm->mmu_lock); @@ -1355,7 +1351,7 @@ static int nonpaging_page_fault(struct kvm_vcpu *vcpu, gva_t gva, static int tdp_page_fault(struct kvm_vcpu *vcpu, gva_t gpa, u32 error_code) { - struct page *page; + pfn_t pfn; int r; int largepage = 0; gfn_t gfn = gpa >> PAGE_SHIFT; @@ -1372,16 +1368,16 @@ static int tdp_page_fault(struct kvm_vcpu *vcpu, gva_t gpa, gfn &= ~(KVM_PAGES_PER_HPAGE-1); largepage = 1; } - page = gfn_to_page(vcpu->kvm, gfn); + pfn = gfn_to_pfn(vcpu->kvm, gfn); up_read(¤t->mm->mmap_sem); - if (is_error_page(page)) { - kvm_release_page_clean(page); + if (is_error_pfn(pfn)) { + kvm_release_pfn_clean(pfn); return 1; } spin_lock(&vcpu->kvm->mmu_lock); kvm_mmu_free_some_pages(vcpu); r = __direct_map(vcpu, gpa, error_code & PFERR_WRITE_MASK, - largepage, gfn, page, TDP_ROOT_LEVEL); + largepage, gfn, pfn, TDP_ROOT_LEVEL); spin_unlock(&vcpu->kvm->mmu_lock); return r; @@ -1525,6 +1521,8 @@ static int init_kvm_softmmu(struct kvm_vcpu *vcpu) static int init_kvm_mmu(struct kvm_vcpu *vcpu) { + vcpu->arch.update_pte.pfn = bad_pfn; + if (tdp_enabled) return init_kvm_tdp_mmu(vcpu); else @@ -1644,7 +1642,7 @@ static void mmu_guess_page_from_pte_write(struct kvm_vcpu *vcpu, gpa_t gpa, gfn_t gfn; int r; u64 gpte = 0; - struct page *page; + pfn_t pfn; vcpu->arch.update_pte.largepage = 0; @@ -1680,15 +1678,15 @@ static void mmu_guess_page_from_pte_write(struct kvm_vcpu *vcpu, gpa_t gpa, gfn &= ~(KVM_PAGES_PER_HPAGE-1); vcpu->arch.update_pte.largepage = 1; } - page = gfn_to_page(vcpu->kvm, gfn); + pfn = gfn_to_pfn(vcpu->kvm, gfn); up_read(¤t->mm->mmap_sem); - if (is_error_page(page)) { - kvm_release_page_clean(page); + if (is_error_pfn(pfn)) { + kvm_release_pfn_clean(pfn); return; } vcpu->arch.update_pte.gfn = gfn; - vcpu->arch.update_pte.page = page; + vcpu->arch.update_pte.pfn = pfn; } void kvm_mmu_pte_write(struct kvm_vcpu *vcpu, gpa_t gpa, @@ -1793,9 +1791,9 @@ void kvm_mmu_pte_write(struct kvm_vcpu *vcpu, gpa_t gpa, } kvm_mmu_audit(vcpu, "post pte write"); spin_unlock(&vcpu->kvm->mmu_lock); - if (vcpu->arch.update_pte.page) { - kvm_release_page_clean(vcpu->arch.update_pte.page); - vcpu->arch.update_pte.page = NULL; + if (!is_error_pfn(vcpu->arch.update_pte.pfn)) { + kvm_release_pfn_clean(vcpu->arch.update_pte.pfn); + vcpu->arch.update_pte.pfn = bad_pfn; } } @@ -2236,8 +2234,7 @@ static void audit_mappings_page(struct kvm_vcpu *vcpu, u64 page_pte, audit_mappings_page(vcpu, ent, va, level - 1); } else { gpa_t gpa = vcpu->arch.mmu.gva_to_gpa(vcpu, va); - struct page *page = gpa_to_page(vcpu, gpa); - hpa_t hpa = page_to_phys(page); + hpa_t hpa = (hpa_t)gpa_to_pfn(vcpu, gpa) << PAGE_SHIFT; if (is_shadow_present_pte(ent) && (ent & PT64_BASE_ADDR_MASK) != hpa) @@ -2250,7 +2247,7 @@ static void audit_mappings_page(struct kvm_vcpu *vcpu, u64 page_pte, && !is_error_hpa(hpa)) printk(KERN_ERR "audit: (%s) notrap shadow," " valid guest gva %lx\n", audit_msg, va); - kvm_release_page_clean(page); + kvm_release_pfn_clean(pfn); } } diff --git a/arch/x86/kvm/paging_tmpl.h b/arch/x86/kvm/paging_tmpl.h index 57d872aec66..156fe10288a 100644 --- a/arch/x86/kvm/paging_tmpl.h +++ b/arch/x86/kvm/paging_tmpl.h @@ -247,7 +247,7 @@ static void FNAME(update_pte)(struct kvm_vcpu *vcpu, struct kvm_mmu_page *page, { pt_element_t gpte; unsigned pte_access; - struct page *npage; + pfn_t pfn; int largepage = vcpu->arch.update_pte.largepage; gpte = *(const pt_element_t *)pte; @@ -260,13 +260,13 @@ static void FNAME(update_pte)(struct kvm_vcpu *vcpu, struct kvm_mmu_page *page, pte_access = page->role.access & FNAME(gpte_access)(vcpu, gpte); if (gpte_to_gfn(gpte) != vcpu->arch.update_pte.gfn) return; - npage = vcpu->arch.update_pte.page; - if (!npage) + pfn = vcpu->arch.update_pte.pfn; + if (is_error_pfn(pfn)) return; - get_page(npage); + kvm_get_pfn(pfn); mmu_set_spte(vcpu, spte, page->role.access, pte_access, 0, 0, gpte & PT_DIRTY_MASK, NULL, largepage, gpte_to_gfn(gpte), - npage, true); + pfn, true); } /* @@ -275,7 +275,7 @@ static void FNAME(update_pte)(struct kvm_vcpu *vcpu, struct kvm_mmu_page *page, static u64 *FNAME(fetch)(struct kvm_vcpu *vcpu, gva_t addr, struct guest_walker *walker, int user_fault, int write_fault, int largepage, - int *ptwrite, struct page *page) + int *ptwrite, pfn_t pfn) { hpa_t shadow_addr; int level; @@ -336,7 +336,7 @@ static u64 *FNAME(fetch)(struct kvm_vcpu *vcpu, gva_t addr, walker->pte_gpa[level - 2], &curr_pte, sizeof(curr_pte)); if (r || curr_pte != walker->ptes[level - 2]) { - kvm_release_page_clean(page); + kvm_release_pfn_clean(pfn); return NULL; } } @@ -349,7 +349,7 @@ static u64 *FNAME(fetch)(struct kvm_vcpu *vcpu, gva_t addr, mmu_set_spte(vcpu, shadow_ent, access, walker->pte_access & access, user_fault, write_fault, walker->ptes[walker->level-1] & PT_DIRTY_MASK, - ptwrite, largepage, walker->gfn, page, false); + ptwrite, largepage, walker->gfn, pfn, false); return shadow_ent; } @@ -378,7 +378,7 @@ static int FNAME(page_fault)(struct kvm_vcpu *vcpu, gva_t addr, u64 *shadow_pte; int write_pt = 0; int r; - struct page *page; + pfn_t pfn; int largepage = 0; pgprintk("%s: addr %lx err %x\n", __func__, addr, error_code); @@ -413,20 +413,20 @@ static int FNAME(page_fault)(struct kvm_vcpu *vcpu, gva_t addr, largepage = 1; } } - page = gfn_to_page(vcpu->kvm, walker.gfn); + pfn = gfn_to_pfn(vcpu->kvm, walker.gfn); up_read(¤t->mm->mmap_sem); /* mmio */ - if (is_error_page(page)) { + if (is_error_pfn(pfn)) { pgprintk("gfn %x is mmio\n", walker.gfn); - kvm_release_page_clean(page); + kvm_release_pfn_clean(pfn); return 1; } spin_lock(&vcpu->kvm->mmu_lock); kvm_mmu_free_some_pages(vcpu); shadow_pte = FNAME(fetch)(vcpu, addr, &walker, user_fault, write_fault, - largepage, &write_pt, page); + largepage, &write_pt, pfn); pgprintk("%s: shadow pte %p %llx ptwrite %d\n", __func__, shadow_pte, *shadow_pte, write_pt); -- cgit v1.2.3