[PATCH] i386: PARAVIRT: Hooks to set up initial pagetable

This patch introduces paravirt_ops hooks to control how the kernel's
initial pagetable is set up.

In the case of a native boot, the very early bootstrap code creates a
simple non-PAE pagetable to map the kernel and physical memory.  When
the VM subsystem is initialized, it creates a proper pagetable which
respects the PAE mode, large pages, etc.

When booting under a hypervisor, there are many possibilities for what
paging environment the hypervisor establishes for the guest kernel, so
the constructon of the kernel's pagetable depends on the hypervisor.

In the case of Xen, the hypervisor boots the kernel with a fully
constructed pagetable, which is already using PAE if necessary.  Also,
Xen requires particular care when constructing pagetables to make sure
all pagetables are always mapped read-only.

In order to make this easier, kernel's initial pagetable construction
has been changed to only allocate and initialize a pagetable page if
there's no page already present in the pagetable.  This allows the Xen
paravirt backend to make a copy of the hypervisor-provided pagetable,
allowing the kernel to establish any more mappings it needs while
keeping the existing ones.

A slightly subtle point which is worth highlighting here is that Xen
requires all kernel mappings to share the same pte_t pages between all
pagetables, so that updating a kernel page's mapping in one pagetable
is reflected in all other pagetables.  This makes it possible to
allocate a page and attach it to a pagetable without having to
explicitly enumerate that page's mapping in all pagetables.

And:

+From: "Eric W. Biederman" <ebiederm@xmission.com>

If we don't set the leaf page table entries it is quite possible that
will inherit and incorrect page table entry from the initial boot
page table setup in head.S.  So we need to redo the effort here,
so we pick up PSE, PGE and the like.

Hypervisors like Xen require that their page tables be read-only,
which is slightly incompatible with our low identity mappings, however
I discussed this with Jeremy he has modified the Xen early set_pte
function to avoid problems in this area.

Signed-off-by: Eric W. Biederman <ebiederm@xmission.com>
Signed-off-by: Jeremy Fitzhardinge <jeremy@xensource.com>
Signed-off-by: Andi Kleen <ak@suse.de>
Acked-by: William Irwin <bill.irwin@oracle.com>
Cc: Ingo Molnar <mingo@elte.hu>

2 files changed, 94 insertions, 47 deletions

diff --git a/arch/i386/kernel/paravirt.c b/arch/i386/kernel/paravirt.c
index cba7a15..47d075b 100644
--- a/arch/i386/kernel/paravirt.c
+++ b/arch/i386/kernel/paravirt.c
@@ -193,6 +193,9 @@ struct paravirt_ops paravirt_ops = {
 #endif
 	.set_lazy_mode = paravirt_nop,
 
+	.pagetable_setup_start = native_pagetable_setup_start,
+	.pagetable_setup_done = native_pagetable_setup_done,
+
 	.flush_tlb_user = native_flush_tlb,
 	.flush_tlb_kernel = native_flush_tlb_global,
 	.flush_tlb_single = native_flush_tlb_single,
diff --git a/arch/i386/mm/init.c b/arch/i386/mm/init.c
index bd5ef37..e8545dc 100644
--- a/arch/i386/mm/init.c
+++ b/arch/i386/mm/init.c
@@ -43,6 +43,7 @@
 #include <asm/tlb.h>
 #include <asm/tlbflush.h>
 #include <asm/sections.h>
+#include <asm/paravirt.h>
 
 unsigned int __VMALLOC_RESERVE = 128 << 20;
 
@@ -62,17 +63,18 @@ static pmd_t * __init one_md_table_init(pgd_t *pgd)
 	pmd_t *pmd_table;
 		
 #ifdef CONFIG_X86_PAE
-	pmd_table = (pmd_t *) alloc_bootmem_low_pages(PAGE_SIZE);
-	paravirt_alloc_pd(__pa(pmd_table) >> PAGE_SHIFT);
-	set_pgd(pgd, __pgd(__pa(pmd_table) | _PAGE_PRESENT));
-	pud = pud_offset(pgd, 0);
-	if (pmd_table != pmd_offset(pud, 0)) 
-		BUG();
-#else
+	if (!(pgd_val(*pgd) & _PAGE_PRESENT)) {
+		pmd_table = (pmd_t *) alloc_bootmem_low_pages(PAGE_SIZE);
+
+		paravirt_alloc_pd(__pa(pmd_table) >> PAGE_SHIFT);
+		set_pgd(pgd, __pgd(__pa(pmd_table) | _PAGE_PRESENT));
+		pud = pud_offset(pgd, 0);
+		if (pmd_table != pmd_offset(pud, 0))
+			BUG();
+	}
+#endif
 	pud = pud_offset(pgd, 0);
 	pmd_table = pmd_offset(pud, 0);
-#endif
-
 	return pmd_table;
 }
 
@@ -82,14 +84,12 @@ static pmd_t * __init one_md_table_init(pgd_t *pgd)
  */
 static pte_t * __init one_page_table_init(pmd_t *pmd)
 {
-	if (pmd_none(*pmd)) {
+	if (!(pmd_val(*pmd) & _PAGE_PRESENT)) {
 		pte_t *page_table = (pte_t *) alloc_bootmem_low_pages(PAGE_SIZE);
+
 		paravirt_alloc_pt(__pa(page_table) >> PAGE_SHIFT);
 		set_pmd(pmd, __pmd(__pa(page_table) | _PAGE_TABLE));
-		if (page_table != pte_offset_kernel(pmd, 0))
-			BUG();	
-
-		return page_table;
+		BUG_ON(page_table != pte_offset_kernel(pmd, 0));
 	}
 	
 	return pte_offset_kernel(pmd, 0);
@@ -109,7 +109,6 @@ static pte_t * __init one_page_table_init(pmd_t *pmd)
 static void __init page_table_range_init (unsigned long start, unsigned long end, pgd_t *pgd_base)
 {
 	pgd_t *pgd;
-	pud_t *pud;
 	pmd_t *pmd;
 	int pgd_idx, pmd_idx;
 	unsigned long vaddr;
@@ -120,13 +119,10 @@ static void __init page_table_range_init (unsigned long start, unsigned long end
 	pgd = pgd_base + pgd_idx;
 
 	for ( ; (pgd_idx < PTRS_PER_PGD) && (vaddr != end); pgd++, pgd_idx++) {
-		if (pgd_none(*pgd)) 
-			one_md_table_init(pgd);
-		pud = pud_offset(pgd, vaddr);
-		pmd = pmd_offset(pud, vaddr);
+		pmd = one_md_table_init(pgd);
+		pmd = pmd + pmd_index(vaddr);
 		for (; (pmd_idx < PTRS_PER_PMD) && (vaddr != end); pmd++, pmd_idx++) {
-			if (pmd_none(*pmd)) 
-				one_page_table_init(pmd);
+			one_page_table_init(pmd);
 
 			vaddr += PMD_SIZE;
 		}
@@ -168,20 +164,22 @@ static void __init kernel_physical_mapping_init(pgd_t *pgd_base)
 			/* Map with big pages if possible, otherwise create normal page tables. */
 			if (cpu_has_pse) {
 				unsigned int address2 = (pfn + PTRS_PER_PTE - 1) * PAGE_SIZE + PAGE_OFFSET + PAGE_SIZE-1;
-
 				if (is_kernel_text(address) || is_kernel_text(address2))
 					set_pmd(pmd, pfn_pmd(pfn, PAGE_KERNEL_LARGE_EXEC));
 				else
 					set_pmd(pmd, pfn_pmd(pfn, PAGE_KERNEL_LARGE));
+
 				pfn += PTRS_PER_PTE;
 			} else {
 				pte = one_page_table_init(pmd);
 
-				for (pte_ofs = 0; pte_ofs < PTRS_PER_PTE && pfn < max_low_pfn; pte++, pfn++, pte_ofs++) {
-						if (is_kernel_text(address))
-							set_pte(pte, pfn_pte(pfn, PAGE_KERNEL_EXEC));
-						else
-							set_pte(pte, pfn_pte(pfn, PAGE_KERNEL));
+				for (pte_ofs = 0;
+				     pte_ofs < PTRS_PER_PTE && pfn < max_low_pfn;
+				     pte++, pfn++, pte_ofs++, address += PAGE_SIZE) {
+					if (is_kernel_text(address))
+						set_pte(pte, pfn_pte(pfn, PAGE_KERNEL_EXEC));
+					else
+						set_pte(pte, pfn_pte(pfn, PAGE_KERNEL));
 				}
 			}
 		}
@@ -338,24 +336,78 @@ extern void __init remap_numa_kva(void);
 #define remap_numa_kva() do {} while (0)
 #endif
 
-static void __init pagetable_init (void)
+void __init native_pagetable_setup_start(pgd_t *base)
 {
-	unsigned long vaddr;
-	pgd_t *pgd_base = swapper_pg_dir;
-
 #ifdef CONFIG_X86_PAE
 	int i;
-	/* Init entries of the first-level page table to the zero page */
-	for (i = 0; i < PTRS_PER_PGD; i++)
-		set_pgd(pgd_base + i, __pgd(__pa(empty_zero_page) | _PAGE_PRESENT));
+
+	/*
+	 * Init entries of the first-level page table to the
+	 * zero page, if they haven't already been set up.
+	 *
+	 * In a normal native boot, we'll be running on a
+	 * pagetable rooted in swapper_pg_dir, but not in PAE
+	 * mode, so this will end up clobbering the mappings
+	 * for the lower 24Mbytes of the address space,
+	 * without affecting the kernel address space.
+	 */
+	for (i = 0; i < USER_PTRS_PER_PGD; i++)
+		set_pgd(&base[i],
+			__pgd(__pa(empty_zero_page) | _PAGE_PRESENT));
+
+	/* Make sure kernel address space is empty so that a pagetable
+	   will be allocated for it. */
+	memset(&base[USER_PTRS_PER_PGD], 0,
+	       KERNEL_PGD_PTRS * sizeof(pgd_t));
 #else
 	paravirt_alloc_pd(__pa(swapper_pg_dir) >> PAGE_SHIFT);
 #endif
+}
+
+void __init native_pagetable_setup_done(pgd_t *base)
+{
+#ifdef CONFIG_X86_PAE
+	/*
+	 * Add low memory identity-mappings - SMP needs it when
+	 * starting up on an AP from real-mode. In the non-PAE
+	 * case we already have these mappings through head.S.
+	 * All user-space mappings are explicitly cleared after
+	 * SMP startup.
+	 */
+	set_pgd(&base[0], base[USER_PTRS_PER_PGD]);
+#endif
+}
+
+/*
+ * Build a proper pagetable for the kernel mappings.  Up until this
+ * point, we've been running on some set of pagetables constructed by
+ * the boot process.
+ *
+ * If we're booting on native hardware, this will be a pagetable
+ * constructed in arch/i386/kernel/head.S, and not running in PAE mode
+ * (even if we'll end up running in PAE).  The root of the pagetable
+ * will be swapper_pg_dir.
+ *
+ * If we're booting paravirtualized under a hypervisor, then there are
+ * more options: we may already be running PAE, and the pagetable may
+ * or may not be based in swapper_pg_dir.  In any case,
+ * paravirt_pagetable_setup_start() will set up swapper_pg_dir
+ * appropriately for the rest of the initialization to work.
+ *
+ * In general, pagetable_init() assumes that the pagetable may already
+ * be partially populated, and so it avoids stomping on any existing
+ * mappings.
+ */
+static void __init pagetable_init (void)
+{
+	unsigned long vaddr, end;
+	pgd_t *pgd_base = swapper_pg_dir;
+
+	paravirt_pagetable_setup_start(pgd_base);
 
 	/* Enable PSE if available */
-	if (cpu_has_pse) {
+	if (cpu_has_pse)
 		set_in_cr4(X86_CR4_PSE);
-	}
 
 	/* Enable PGE if available */
 	if (cpu_has_pge) {
@@ -372,20 +424,12 @@ static void __init pagetable_init (void)
 	 * created - mappings will be set by set_fixmap():
 	 */
 	vaddr = __fix_to_virt(__end_of_fixed_addresses - 1) & PMD_MASK;
-	page_table_range_init(vaddr, 0, pgd_base);
+	end = (FIXADDR_TOP + PMD_SIZE - 1) & PMD_MASK;
+	page_table_range_init(vaddr, end, pgd_base);
 
 	permanent_kmaps_init(pgd_base);
 
-#ifdef CONFIG_X86_PAE
-	/*
-	 * Add low memory identity-mappings - SMP needs it when
-	 * starting up on an AP from real-mode. In the non-PAE
-	 * case we already have these mappings through head.S.
-	 * All user-space mappings are explicitly cleared after
-	 * SMP startup.
-	 */
-	set_pgd(&pgd_base[0], pgd_base[USER_PTRS_PER_PGD]);
-#endif
+	paravirt_pagetable_setup_done(pgd_base);
 }
 
 #if defined(CONFIG_SOFTWARE_SUSPEND) || defined(CONFIG_ACPI_SLEEP)