The patch titled
KVM: MMU: oom handling
has been added to the -mm tree. Its filename is
kvm-mmu-oom-handling.patch
See http://www.zip.com.au/~akpm/linux/patches/stuff/added-to-mm.txt to find
out what to do about this
------------------------------------------------------
Subject: KVM: MMU: oom handling
From: Avi Kivity <avi@xxxxxxxxxxxx>
When beginning to process a page fault, make sure we have enough shadow pages
available to service the fault. If not, free some pages.
Signed-off-by: Avi Kivity <avi@xxxxxxxxxxxx>
Acked-by: Ingo Molnar <mingo@xxxxxxx>
Signed-off-by: Andrew Morton <akpm@xxxxxxxx>
---
drivers/kvm/kvm.h | 12 ++++++++
drivers/kvm/mmu.c | 49 +++++++++++++++---------------------
drivers/kvm/paging_tmpl.h | 15 +----------
drivers/kvm/svm.c | 2 -
drivers/kvm/vmx.c | 2 -
5 files changed, 37 insertions(+), 43 deletions(-)
diff -puN drivers/kvm/kvm.h~kvm-mmu-oom-handling drivers/kvm/kvm.h
--- a/drivers/kvm/kvm.h~kvm-mmu-oom-handling
+++ a/drivers/kvm/kvm.h
@@ -52,6 +52,8 @@
#define KVM_MAX_VCPUS 1
#define KVM_MEMORY_SLOTS 4
#define KVM_NUM_MMU_PAGES 256
+#define KVM_MIN_FREE_MMU_PAGES 5
+#define KVM_REFILL_PAGES 25
#define FX_IMAGE_SIZE 512
#define FX_IMAGE_ALIGN 16
@@ -278,6 +280,7 @@ struct kvm {
* Hash table of struct kvm_mmu_page.
*/
struct list_head active_mmu_pages;
+ int n_free_mmu_pages;
struct hlist_head mmu_page_hash[KVM_NUM_MMU_PAGES];
struct kvm_vcpu vcpus[KVM_MAX_VCPUS];
int memory_config_version;
@@ -451,6 +454,15 @@ unsigned long segment_base(u16 selector)
void kvm_mmu_pre_write(struct kvm_vcpu *vcpu, gpa_t gpa, int bytes);
void kvm_mmu_post_write(struct kvm_vcpu *vcpu, gpa_t gpa, int bytes);
int kvm_mmu_unprotect_page_virt(struct kvm_vcpu *vcpu, gva_t gva);
+void kvm_mmu_free_some_pages(struct kvm_vcpu *vcpu);
+
+static inline int kvm_mmu_page_fault(struct kvm_vcpu *vcpu, gva_t gva,
+ u32 error_code)
+{
+ if (unlikely(vcpu->kvm->n_free_mmu_pages < KVM_MIN_FREE_MMU_PAGES))
+ kvm_mmu_free_some_pages(vcpu);
+ return vcpu->mmu.page_fault(vcpu, gva, error_code);
+}
static inline struct page *_gfn_to_page(struct kvm *kvm, gfn_t gfn)
{
diff -puN drivers/kvm/mmu.c~kvm-mmu-oom-handling drivers/kvm/mmu.c
--- a/drivers/kvm/mmu.c~kvm-mmu-oom-handling
+++ a/drivers/kvm/mmu.c
@@ -310,6 +310,7 @@ static void kvm_mmu_free_page(struct kvm
list_del(&page_head->link);
page_head->page_hpa = page_hpa;
list_add(&page_head->link, &vcpu->free_pages);
+ ++vcpu->kvm->n_free_mmu_pages;
}
static int is_empty_shadow_page(hpa_t page_hpa)
@@ -344,6 +345,7 @@ static struct kvm_mmu_page *kvm_mmu_allo
page->global = 1;
page->multimapped = 0;
page->parent_pte = parent_pte;
+ --vcpu->kvm->n_free_mmu_pages;
return page;
}
@@ -544,8 +546,7 @@ static void kvm_mmu_zap_page(struct kvm_
}
kvm_mmu_page_unlink_children(vcpu, page);
hlist_del(&page->hash_link);
- list_del(&page->link);
- list_add(&page->link, &vcpu->free_pages);
+ kvm_mmu_free_page(vcpu, page->page_hpa);
}
static int kvm_mmu_unprotect_page(struct kvm_vcpu *vcpu, gfn_t gfn)
@@ -743,18 +744,6 @@ static void mmu_alloc_roots(struct kvm_v
vcpu->mmu.root_hpa = __pa(vcpu->mmu.pae_root);
}
-static void nonpaging_flush(struct kvm_vcpu *vcpu)
-{
- hpa_t root = vcpu->mmu.root_hpa;
-
- ++kvm_stat.tlb_flush;
- pgprintk("nonpaging_flush\n");
- mmu_free_roots(vcpu);
- mmu_alloc_roots(vcpu);
- kvm_arch_ops->set_cr3(vcpu, root);
- kvm_arch_ops->tlb_flush(vcpu);
-}
-
static gpa_t nonpaging_gva_to_gpa(struct kvm_vcpu *vcpu, gva_t vaddr)
{
return vaddr;
@@ -763,28 +752,19 @@ static gpa_t nonpaging_gva_to_gpa(struct
static int nonpaging_page_fault(struct kvm_vcpu *vcpu, gva_t gva,
u32 error_code)
{
- int ret;
gpa_t addr = gva;
+ hpa_t paddr;
ASSERT(vcpu);
ASSERT(VALID_PAGE(vcpu->mmu.root_hpa));
- for (;;) {
- hpa_t paddr;
- paddr = gpa_to_hpa(vcpu , addr & PT64_BASE_ADDR_MASK);
+ paddr = gpa_to_hpa(vcpu , addr & PT64_BASE_ADDR_MASK);
- if (is_error_hpa(paddr))
- return 1;
+ if (is_error_hpa(paddr))
+ return 1;
- ret = nonpaging_map(vcpu, addr & PAGE_MASK, paddr);
- if (ret) {
- nonpaging_flush(vcpu);
- continue;
- }
- break;
- }
- return ret;
+ return nonpaging_map(vcpu, addr & PAGE_MASK, paddr);
}
static void nonpaging_inval_page(struct kvm_vcpu *vcpu, gva_t addr)
@@ -1093,6 +1073,18 @@ int kvm_mmu_unprotect_page_virt(struct k
return kvm_mmu_unprotect_page(vcpu, gpa >> PAGE_SHIFT);
}
+void kvm_mmu_free_some_pages(struct kvm_vcpu *vcpu)
+{
+ while (vcpu->kvm->n_free_mmu_pages < KVM_REFILL_PAGES) {
+ struct kvm_mmu_page *page;
+
+ page = container_of(vcpu->kvm->active_mmu_pages.prev,
+ struct kvm_mmu_page, link);
+ kvm_mmu_zap_page(vcpu, page);
+ }
+}
+EXPORT_SYMBOL_GPL(kvm_mmu_free_some_pages);
+
static void free_mmu_pages(struct kvm_vcpu *vcpu)
{
while (!list_empty(&vcpu->free_pages)) {
@@ -1124,6 +1116,7 @@ static int alloc_mmu_pages(struct kvm_vc
page_header->page_hpa = (hpa_t)page_to_pfn(page) << PAGE_SHIFT;
memset(__va(page_header->page_hpa), 0, PAGE_SIZE);
list_add(&page_header->link, &vcpu->free_pages);
+ ++vcpu->kvm->n_free_mmu_pages;
}
/*
diff -puN drivers/kvm/paging_tmpl.h~kvm-mmu-oom-handling drivers/kvm/paging_tmpl.h
--- a/drivers/kvm/paging_tmpl.h~kvm-mmu-oom-handling
+++ a/drivers/kvm/paging_tmpl.h
@@ -246,8 +246,6 @@ static u64 *FNAME(fetch)(struct kvm_vcpu
}
shadow_page = kvm_mmu_get_page(vcpu, table_gfn, addr, level-1,
metaphysical, shadow_ent);
- if (!shadow_page)
- return ERR_PTR(-ENOMEM);
shadow_addr = shadow_page->page_hpa;
shadow_pte = shadow_addr | PT_PRESENT_MASK | PT_ACCESSED_MASK
| PT_WRITABLE_MASK | PT_USER_MASK;
@@ -347,17 +345,8 @@ static int FNAME(page_fault)(struct kvm_
/*
* Look up the shadow pte for the faulting address.
*/
- for (;;) {
- FNAME(walk_addr)(&walker, vcpu, addr);
- shadow_pte = FNAME(fetch)(vcpu, addr, &walker);
- if (IS_ERR(shadow_pte)) { /* must be -ENOMEM */
- printk("%s: oom\n", __FUNCTION__);
- nonpaging_flush(vcpu);
- FNAME(release_walker)(&walker);
- continue;
- }
- break;
- }
+ FNAME(walk_addr)(&walker, vcpu, addr);
+ shadow_pte = FNAME(fetch)(vcpu, addr, &walker);
/*
* The page is not mapped by the guest. Let the guest handle it.
diff -puN drivers/kvm/svm.c~kvm-mmu-oom-handling drivers/kvm/svm.c
--- a/drivers/kvm/svm.c~kvm-mmu-oom-handling
+++ a/drivers/kvm/svm.c
@@ -861,7 +861,7 @@ static int pf_interception(struct kvm_vc
fault_address = vcpu->svm->vmcb->control.exit_info_2;
error_code = vcpu->svm->vmcb->control.exit_info_1;
- if (!vcpu->mmu.page_fault(vcpu, fault_address, error_code)) {
+ if (!kvm_mmu_page_fault(vcpu, fault_address, error_code)) {
spin_unlock(&vcpu->kvm->lock);
return 1;
}
diff -puN drivers/kvm/vmx.c~kvm-mmu-oom-handling drivers/kvm/vmx.c
--- a/drivers/kvm/vmx.c~kvm-mmu-oom-handling
+++ a/drivers/kvm/vmx.c
@@ -1318,7 +1318,7 @@ static int handle_exception(struct kvm_v
cr2 = vmcs_readl(EXIT_QUALIFICATION);
spin_lock(&vcpu->kvm->lock);
- if (!vcpu->mmu.page_fault(vcpu, cr2, error_code)) {
+ if (!kvm_mmu_page_fault(vcpu, cr2, error_code)) {
spin_unlock(&vcpu->kvm->lock);
return 1;
}
_
Patches currently in -mm which might be from avi@xxxxxxxxxxxx are
kvm-fix-gfp_kernel-alloc-in-atomic-section-bug.patch
kvm-use-raw_smp_processor_id-instead-of-smp_processor_id-where-applicable.patch
kvm-recover-after-an-arch-module-load-failure.patch
kvm-improve-interrupt-response.patch
kvm-prevent-stale-bits-in-cr0-and-cr4.patch
kvm-mmu-implement-simple-reverse-mapping.patch
kvm-mmu-teach-the-page-table-walker-to-track-guest-page-table-gfns.patch
kvm-mmu-load-the-pae-pdptrs-on-cr3-change-like-the-processor-does.patch
kvm-mmu-fold-fetch_guest-into-init_walker.patch
kvm-mu-special-treatment-for-shadow-pae-root-pages.patch
kvm-mmu-use-the-guest-pdptrs-instead-of-mapping-cr3-in-pae-mode.patch
kvm-mmu-make-the-shadow-page-tables-also-special-case-pae.patch
kvm-mmu-make-kvm_mmu_alloc_page-return-a-kvm_mmu_page-pointer.patch
kvm-mmu-shadow-page-table-caching.patch
kvm-mmu-write-protect-guest-pages-when-a-shadow-is-created-for-them.patch
kvm-mmu-let-the-walker-extract-the-target-page-gfn-from-the-pte.patch
kvm-mmu-support-emulated-writes-into-ram.patch
kvm-mmu-zap-shadow-page-table-entries-on-writes-to-guest-page-tables.patch
kvm-mmu-if-emulating-an-instruction-fails-try-unprotecting-the-page.patch
kvm-mmu-implement-child-shadow-unlinking.patch
kvm-mmu-kvm_mmu_put_page-only-removes-one-link-to-the-page.patch
kvm-mmu-oom-handling.patch
kvm-mmu-remove-invlpg-interception.patch
kvm-mmu-remove-release_pt_page_64.patch
kvm-mmu-handle-misaligned-accesses-to-write-protected-guest-page-tables.patch
kvm-mmu-ove-is_empty_shadow_page-above-kvm_mmu_free_page.patch
kvm-mmu-ensure-freed-shadow-pages-are-clean.patch
kvm-mmu-if-an-empty-shadow-page-is-not-empty-report-more-info.patch
kvm-mmu-page-table-write-flood-protection.patch
kvm-mmu-never-free-a-shadow-page-actively-serving-as-a-root.patch
kvm-mmu-fix-cmpxchg8b-emulation.patch
kvm-mmu-treat-user-mode-faults-as-a-hint-that-a-page-is-no-longer-a-page-table.patch
kvm-mmu-free-pages-on-kvm-destruction.patch
kvm-mmu-replace-atomic-allocations-by-preallocated-objects.patch
kvm-mmu-detect-oom-conditions-and-propagate-error-to-userspace.patch
kvm-mmu-flush-guest-tlb-when-reducing-permissions-on-a-pte.patch
kvm-mmu-destroy-mmu-while-we-still-have-a-vcpu-left.patch
kvm-mmu-add-audit-code-to-check-mappings-etc-are-correct.patch
-
To unsubscribe from this list: send the line "unsubscribe mm-commits" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
