x86-decouple-pagefault-logic-in-highmem-kmap.patch

From aba1a788bf04417b1abfbc31c2e90c223a46a726 Mon Sep 17 00:00:00 2001
From: Peter Zijlstra <a.p.zijlstra@chello.nl>
Date: Fri, 3 Jul 2009 08:30:37 -0500
Subject: [PATCH] x86: decouple pagefault-logic in highmem/kmap

commit 4beedc411ebd0cd5d294acfe9b33e8c2ad7dc598 in tip.

With the separation of pagefault_{disable,enable}() from the
preempt_count a previously overlooked dependancy became painfully
clear.

kmap_atomic() is per cpu and relies not only on disabling the
pagefault handler, but really needs preemption disabled too.

Make this explicit now - so that we can change pagefault_disable().

Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>

diff --git a/arch/x86/mm/highmem_32.c b/arch/x86/mm/highmem_32.c
index c838ac8..76e0544 100644
--- a/arch/x86/mm/highmem_32.c
+++ b/arch/x86/mm/highmem_32.c
@@ -33,6 +33,7 @@ void *kmap_atomic_prot(struct page *page, enum km_type type, pgprot_t prot)
 	unsigned long vaddr;
 
 	/* even !CONFIG_PREEMPT needs this, for in_atomic in do_page_fault */
+	preempt_disable();
 	pagefault_disable();
 
 	if (!PageHighMem(page))
@@ -74,6 +75,7 @@ void kunmap_atomic(void *kvaddr, enum km_type type)
 	}
 
 	pagefault_enable();
+	preempt_enable();
 }
 
 /*
-- 
1.7.1.1