Re: [PATCH 6/7] x86, mm: setup page table from top-down

[Yinghai Lu <yinghai@kernel.org>]

[-- Attachment #1: Type: text/plain, Size: 704 bytes --]

On Wed, Oct 10, 2012 at 10:26 AM, Stefano Stabellini
<stefano.stabellini@eu.citrix.com> wrote:
> On Wed, 10 Oct 2012, Yinghai Lu wrote:
>
> It doesn't matter whether they come from BRK or other memory: Xen
> assumes that all the pagetable pages come from
> pgt_buf_start-pgt_buf_top, so if you are going to use another range you
> need to tell Xen about it.
>
> Alternatively, you can follow Peter's suggestion and replace the current
> hooks with a new one with a more precise and well defined semantic.
> Something along the lines of "this pagetable page is about to be hooked
> into the live pagetable". Xen would use the hook to mark it RO.

attached patch on top of this patch will fix the problem?

[-- Attachment #2: fix_xen.patch --]
[-- Type: application/octet-stream, Size: 3358 bytes --]

---
 arch/x86/include/asm/x86_init.h |    2 +-
 arch/x86/kernel/x86_init.c      |    3 ++-
 arch/x86/mm/init_32.c           |    1 +
 arch/x86/mm/init_64.c           |    1 +
 arch/x86/xen/mmu.c              |   15 +++------------
 5 files changed, 8 insertions(+), 14 deletions(-)

Index: linux-2.6/arch/x86/include/asm/x86_init.h
===================================================================
--- linux-2.6.orig/arch/x86/include/asm/x86_init.h
+++ linux-2.6/arch/x86/include/asm/x86_init.h
@@ -76,7 +76,7 @@ struct x86_init_oem {
  * init_memory_mapping and the commit that added it.
  */
 struct x86_init_mapping {
-	void (*pagetable_reserve)(u64 start, u64 end);
+	void (*mark_page_ro)(u64 addr);
 };
 
 /**
Index: linux-2.6/arch/x86/kernel/x86_init.c
===================================================================
--- linux-2.6.orig/arch/x86/kernel/x86_init.c
+++ linux-2.6/arch/x86/kernel/x86_init.c
@@ -28,6 +28,7 @@ void __cpuinit x86_init_noop(void) { }
 void __init x86_init_uint_noop(unsigned int unused) { }
 int __init iommu_init_noop(void) { return 0; }
 void iommu_shutdown_noop(void) { }
+static void mark_page_ro_noop(u64 addr) { }
 
 /*
  * The platform setup functions are preset with the default functions
@@ -63,7 +64,7 @@ struct x86_init_ops x86_init __initdata
 	},
 
 	.mapping = {
-		.pagetable_reserve		= native_pagetable_reserve,
+		.mark_page_ro			= mark_page_ro_noop;
 	},
 
 	.paging = {
Index: linux-2.6/arch/x86/xen/mmu.c
===================================================================
--- linux-2.6.orig/arch/x86/xen/mmu.c
+++ linux-2.6/arch/x86/xen/mmu.c
@@ -1177,18 +1177,9 @@ static void xen_exit_mmap(struct mm_stru
 
 static void xen_post_allocator_init(void);
 
-static __init void xen_mapping_pagetable_reserve(u64 start, u64 end)
+static __init void xen_mapping_mark_page_ro(u64 addr)
 {
-	/* reserve the range used */
-	native_pagetable_reserve(start, end);
-
-	/* set as RW the rest */
-	printk(KERN_DEBUG "xen: setting RW the range %llx - %llx\n", end,
-			PFN_PHYS(pgt_buf_top));
-	while (end < PFN_PHYS(pgt_buf_top)) {
-		make_lowmem_page_readwrite(__va(end));
-		end += PAGE_SIZE;
-	}
+	make_lowmem_page_readonly(__va(addr));
 }
 
 #ifdef CONFIG_X86_64
@@ -2177,7 +2168,7 @@ static const struct pv_mmu_ops xen_mmu_o
 
 void __init xen_init_mmu_ops(void)
 {
-	x86_init.mapping.pagetable_reserve = xen_mapping_pagetable_reserve;
+	x86_init.mapping.mark_page_ro = xen_mapping_mark_page_ro;
 	x86_init.paging.pagetable_init = xen_pagetable_init;
 	pv_mmu_ops = xen_mmu_ops;
 
Index: linux-2.6/arch/x86/mm/init_32.c
===================================================================
--- linux-2.6.orig/arch/x86/mm/init_32.c
+++ linux-2.6/arch/x86/mm/init_32.c
@@ -78,6 +78,7 @@ static __init void *alloc_low_page(void)
 	} else
 		pfn = pgt_buf_end++;
 
+	x86_init.mapping.mark_page_ro(pfn << PAGE_SHIFT);
 	adr = __va(pfn * PAGE_SIZE);
 	clear_page(adr);
 	return adr;
Index: linux-2.6/arch/x86/mm/init_64.c
===================================================================
--- linux-2.6.orig/arch/x86/mm/init_64.c
+++ linux-2.6/arch/x86/mm/init_64.c
@@ -340,6 +340,7 @@ static __ref void *alloc_low_page(unsign
 	} else
 		pfn = pgt_buf_end++;
 
+	x86_init.mapping.mark_page_ro(pfn << PAGE_SHIFT);
 	adr = early_memremap(pfn * PAGE_SIZE, PAGE_SIZE);
 	clear_page(adr);
 	*phys  = pfn * PAGE_SIZE;