xen64: add extra pv_mmu_ops

We need extra pv_mmu_ops for 64-bit, to deal with the extra level of
pagetable.

Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
Cc: Stephen Tweedie <sct@redhat.com>
Cc: Eduardo Habkost <ehabkost@redhat.com>
Cc: Mark McLoughlin <markmc@redhat.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>

3 files changed, 95 insertions, 4 deletions

diff --git a/arch/x86/xen/enlighten.c b/arch/x86/xen/enlighten.c
index c5f0b40aa39d..afb047e30bdc 100644
--- a/arch/x86/xen/enlighten.c
+++ b/arch/x86/xen/enlighten.c
@@ -803,6 +803,18 @@ static void xen_release_pmd(u32 pfn)
         xen_release_ptpage(pfn, PT_PMD);
 }
 
+#if PAGETABLE_LEVELS == 4
+static void xen_alloc_pud(struct mm_struct *mm, u32 pfn)
+{
+        xen_alloc_ptpage(mm, pfn, PT_PUD);
+}
+
+static void xen_release_pud(u32 pfn)
+{
+        xen_release_ptpage(pfn, PT_PUD);
+}
+#endif
+
 #ifdef CONFIG_HIGHPTE
 static void *xen_kmap_atomic_pte(struct page *page, enum km_type type)
 {
@@ -922,6 +934,11 @@ static __init void xen_pagetable_setup_done(pgd_t *base)
         pv_mmu_ops.alloc_pmd = xen_alloc_pmd;
         pv_mmu_ops.release_pte = xen_release_pte;
         pv_mmu_ops.release_pmd = xen_release_pmd;
+#if PAGETABLE_LEVELS == 4
+        pv_mmu_ops.alloc_pud = xen_alloc_pud;
+        pv_mmu_ops.release_pud = xen_release_pud;
+#endif
+
         pv_mmu_ops.set_pte = xen_set_pte;
 
         xen_setup_shared_info();
@@ -937,6 +954,9 @@ static __init void xen_post_allocator_init(void)
 {
         pv_mmu_ops.set_pmd = xen_set_pmd;
         pv_mmu_ops.set_pud = xen_set_pud;
+#if PAGETABLE_LEVELS == 4
+        pv_mmu_ops.set_pgd = xen_set_pgd;
+#endif
 
         xen_mark_init_mm_pinned();
 }
@@ -1185,15 +1205,26 @@ static const struct pv_mmu_ops xen_mmu_ops __initdata = {
         .make_pte = xen_make_pte,
         .make_pgd = xen_make_pgd,
 
+#ifdef CONFIG_X86_PAE
         .set_pte_atomic = xen_set_pte_atomic,
         .set_pte_present = xen_set_pte_at,
-        .set_pud = xen_set_pud_hyper,
         .pte_clear = xen_pte_clear,
         .pmd_clear = xen_pmd_clear,
+#endif  /* CONFIG_X86_PAE */
+        .set_pud = xen_set_pud_hyper,
 
         .make_pmd = xen_make_pmd,
         .pmd_val = xen_pmd_val,
 
+#if PAGETABLE_LEVELS == 4
+        .pud_val = xen_pud_val,
+        .make_pud = xen_make_pud,
+        .set_pgd = xen_set_pgd_hyper,
+
+        .alloc_pud = xen_alloc_pte_init,
+        .release_pud = xen_release_pte_init,
+#endif  /* PAGETABLE_LEVELS == 4 */
+
         .activate_mm = xen_activate_mm,
         .dup_mmap = xen_dup_mmap,
         .exit_mmap = xen_exit_mmap,
diff --git a/arch/x86/xen/mmu.c b/arch/x86/xen/mmu.c
index 4fca9d88bef0..d0976b87cd2c 100644
--- a/arch/x86/xen/mmu.c
+++ b/arch/x86/xen/mmu.c
@@ -438,14 +438,19 @@ void xen_set_pud(pud_t *ptr, pud_t val)
 
 void xen_set_pte(pte_t *ptep, pte_t pte)
 {
+#ifdef CONFIG_X86_PAE
         ptep->pte_high = pte.pte_high;
         smp_wmb();
         ptep->pte_low = pte.pte_low;
+#else
+        *ptep = pte;
+#endif
 }
 
+#ifdef CONFIG_X86_PAE
 void xen_set_pte_atomic(pte_t *ptep, pte_t pte)
 {
-        set_64bit((u64 *)ptep, pte_val_ma(pte));
+        set_64bit((u64 *)ptep, native_pte_val(pte));
 }
 
 void xen_pte_clear(struct mm_struct *mm, unsigned long addr, pte_t *ptep)
@@ -459,6 +464,7 @@ void xen_pmd_clear(pmd_t *pmdp)
 {
         set_pmd(pmdp, __pmd(0));
 }
+#endif  /* CONFIG_X86_PAE */
 
 pmd_t xen_make_pmd(pmdval_t pmd)
 {
@@ -466,6 +472,49 @@ pmd_t xen_make_pmd(pmdval_t pmd)
         return native_make_pmd(pmd);
 }
 
+#if PAGETABLE_LEVELS == 4
+pudval_t xen_pud_val(pud_t pud)
+{
+        return pte_mfn_to_pfn(pud.pud);
+}
+
+pud_t xen_make_pud(pudval_t pud)
+{
+        pud = pte_pfn_to_mfn(pud);
+
+        return native_make_pud(pud);
+}
+
+void xen_set_pgd_hyper(pgd_t *ptr, pgd_t val)
+{
+        struct mmu_update u;
+
+        preempt_disable();
+
+        xen_mc_batch();
+
+        u.ptr = virt_to_machine(ptr).maddr;
+        u.val = pgd_val_ma(val);
+        extend_mmu_update(&u);
+
+        xen_mc_issue(PARAVIRT_LAZY_MMU);
+
+        preempt_enable();
+}
+
+void xen_set_pgd(pgd_t *ptr, pgd_t val)
+{
+        /* If page is not pinned, we can just update the entry
+           directly */
+        if (!page_pinned(ptr)) {
+                *ptr = val;
+                return;
+        }
+
+        xen_set_pgd_hyper(ptr, val);
+}
+#endif  /* PAGETABLE_LEVELS == 4 */
+
 /*
   (Yet another) pagetable walker.  This one is intended for pinning a
   pagetable.  This means that it walks a pagetable and calls the
diff --git a/arch/x86/xen/mmu.h b/arch/x86/xen/mmu.h
index 7856e37f6044..19d544b0b6c6 100644
--- a/arch/x86/xen/mmu.h
+++ b/arch/x86/xen/mmu.h
@@ -32,13 +32,24 @@ pgd_t xen_make_pgd(pgdval_t);
 void xen_set_pte(pte_t *ptep, pte_t pteval);
 void xen_set_pte_at(struct mm_struct *mm, unsigned long addr,
                     pte_t *ptep, pte_t pteval);
+
+#ifdef CONFIG_X86_PAE
 void xen_set_pte_atomic(pte_t *ptep, pte_t pte);
+void xen_pte_clear(struct mm_struct *mm, unsigned long addr, pte_t *ptep);
+void xen_pmd_clear(pmd_t *pmdp);
+#endif  /* CONFIG_X86_PAE */
+
 void xen_set_pmd(pmd_t *pmdp, pmd_t pmdval);
 void xen_set_pud(pud_t *ptr, pud_t val);
 void xen_set_pmd_hyper(pmd_t *pmdp, pmd_t pmdval);
 void xen_set_pud_hyper(pud_t *ptr, pud_t val);
-void xen_pte_clear(struct mm_struct *mm, unsigned long addr, pte_t *ptep);
-void xen_pmd_clear(pmd_t *pmdp);
+
+#if PAGETABLE_LEVELS == 4
+pudval_t xen_pud_val(pud_t pud);
+pud_t xen_make_pud(pudval_t pudval);
+void xen_set_pgd(pgd_t *pgdp, pgd_t pgd);
+void xen_set_pgd_hyper(pgd_t *pgdp, pgd_t pgd);
+#endif
 
 pte_t xen_ptep_modify_prot_start(struct mm_struct *mm, unsigned long addr, pte_t *ptep);
 void  xen_ptep_modify_prot_commit(struct mm_struct *mm, unsigned long addr,