Re: bisected boot regression post 2.6.25-rc3.. please revert

[Ingo Molnar <mingo@elte.hu>]

* Ingo Molnar <mingo@elte.hu> wrote:

> Could you try to hack up a debug patch perhaps? Uninline the 
> fucntion(s) then add two versions (one is that breaks on your box and 
> one is that works on your box) of this same pmd_bad()/pud_bad() 
> functions and do something like this (pseudocode):

i.e. something like the (tested) patch below. It is not clear whether 
your testcase is on 32-bit or 64-bit - so i went for the more likely 
32-bit case first.

	Ingo

------------------>
Subject: x86: patches/x86-debug-bad-page.patch
From: Ingo Molnar <mingo@elte.hu>
Date: Mon Mar 03 09:53:17 CET 2008

Signed-off-by: Ingo Molnar <mingo@elte.hu>
---
 arch/x86/mm/pgtable_32.c     |    7 +++++++
 include/asm-x86/pgtable_32.h |    6 +++++-
 2 files changed, 12 insertions(+), 1 deletion(-)

Index: linux-x86.q/arch/x86/mm/pgtable_32.c
===================================================================
--- linux-x86.q.orig/arch/x86/mm/pgtable_32.c
+++ linux-x86.q/arch/x86/mm/pgtable_32.c
@@ -381,3 +381,10 @@ void __pmd_free_tlb(struct mmu_gather *t
 }
 
 #endif
+
+int pmd_bad(pmd_t pmd)
+{
+	WARN_ON_ONCE(pmd_bad_v1(pmd) != pmd_bad_v2(pmd));
+
+	return pmd_bad_v1(pmd);
+}
Index: linux-x86.q/include/asm-x86/pgtable_32.h
===================================================================
--- linux-x86.q.orig/include/asm-x86/pgtable_32.h
+++ linux-x86.q/include/asm-x86/pgtable_32.h
@@ -92,7 +92,11 @@ extern unsigned long pg0[];
 /* To avoid harmful races, pmd_none(x) should check only the lower when PAE */
 #define pmd_none(x)	(!(unsigned long)pmd_val(x))
 #define pmd_present(x)	(pmd_val(x) & _PAGE_PRESENT)
-#define	pmd_bad(x)	((pmd_val(x) \
+
+extern int pmd_bad(pmd_t pmd);
+
+#define pmd_bad_v1(x)	((pmd_val(x) & (~PAGE_MASK & ~_PAGE_USER)) != _KERNPG_TABLE)
+#define	pmd_bad_v2(x)	((pmd_val(x) \
 			  & ~(PAGE_MASK | _PAGE_USER | _PAGE_PSE | _PAGE_NX)) \
 			 != _KERNPG_TABLE)