[PATCH 1/2] x86: add is_f00f_bug helper to fault_32|64.c

[PATCH 1/2] x86: add is_f00f_bug helper to fault_32|64.c

[Harvey Harrison]

Further towards unifying these files, add another helper
in same spirit as is_errata93.

Signed-off-by: Harvey Harrison <harvey.harrison@gmail.com>
---
Incorporated HPA's suggestion of unconditional prototype for
do_invalid_op.

 arch/x86/mm/fault_32.c |   39 ++++++++++++++++++++++-----------------
 arch/x86/mm/fault_64.c |   24 ++++++++++++++++++++++++
 2 files changed, 46 insertions(+), 17 deletions(-)

diff --git a/arch/x86/mm/fault_32.c b/arch/x86/mm/fault_32.c
index 936bb0c..dae4f69 100644
--- a/arch/x86/mm/fault_32.c
+++ b/arch/x86/mm/fault_32.c
@@ -211,8 +211,6 @@ void dump_pagetable(unsigned long address)
 	printk("\n");
 }
 
-void do_invalid_op(struct pt_regs *, unsigned long);
-
 static inline pmd_t *vmalloc_sync_one(pgd_t *pgd, unsigned long address)
 {
 	unsigned index = pgd_index(address);
@@ -288,6 +286,26 @@ static int is_errata93(struct pt_regs *regs, unsigned long address)
 	return 0;
 }
 
+void do_invalid_op(struct pt_regs *, unsigned long);
+
+static int is_f00f_bug(struct pt_regs *regs, unsigned long address)
+{
+#ifdef CONFIG_X86_F00F_BUG
+	unsigned long nr;
+	/*
+	 * Pentium F0 0F C7 C8 bug workaround.
+	 */
+	if (boot_cpu_data.f00f_bug) {
+		nr = (address - idt_descr.address) >> 3;
+
+		if (nr == 6) {
+			do_invalid_op(regs, 0);
+			return 1;
+		}
+	}
+#endif
+	return 0;
+}
 
 /*
  * Handle a fault on the vmalloc or module mapping area
@@ -570,21 +588,8 @@ bad_area_nosemaphore:
 		return;
 	}
 
-#ifdef CONFIG_X86_F00F_BUG
-	/*
-	 * Pentium F0 0F C7 C8 bug workaround.
-	 */
-	if (boot_cpu_data.f00f_bug) {
-		unsigned long nr;
-
-		nr = (address - idt_descr.address) >> 3;
-
-		if (nr == 6) {
-			do_invalid_op(regs, 0);
-			return;
-		}
-	}
-#endif
+	if (is_f00f_bug(regs, address))
+		return;
 
 no_context:
 	/* Are we prepared to handle this kernel fault?  */
diff --git a/arch/x86/mm/fault_64.c b/arch/x86/mm/fault_64.c
index cde110c..ce1a870 100644
--- a/arch/x86/mm/fault_64.c
+++ b/arch/x86/mm/fault_64.c
@@ -256,6 +256,27 @@ static int is_errata93(struct pt_regs *regs, unsigned long address)
 	return 0;
 }
 
+void do_invalid_op(struct pt_regs *, unsigned long);
+
+static int is_f00f_bug(struct pt_regs *regs, unsigned long address)
+{
+#ifdef CONFIG_X86_F00F_BUG
+	unsigned long nr;
+	/*
+	 * Pentium F0 0F C7 C8 bug workaround.
+	 */
+	if (boot_cpu_data.f00f_bug) {
+		nr = (address - idt_descr.address) >> 3;
+
+		if (nr == 6) {
+			do_invalid_op(regs, 0);
+			return 1;
+		}
+	}
+#endif
+	return 0;
+}
+
 static noinline void pgtable_bad(unsigned long address, struct pt_regs *regs,
 				 unsigned long error_code)
 {
@@ -572,6 +593,9 @@ bad_area_nosemaphore:
 		return;
 	}
 
+	if (is_f00f_bug(regs, address))
+		return;
+
 no_context:
 	/* Are we prepared to handle this kernel fault?  */
 	if (fixup_exception(regs))
-- 
1.5.4.rc2.1164.g6451

Re: [PATCH 1/2] x86: add is_f00f_bug helper to fault_32|64.c

[Andi Kleen]

Harvey Harrison <harvey.harrison@gmail.com> writes:

> Further towards unifying these files, add another helper
> in same spirit as is_errata93.

The better way to handle this would be to move all these workarounds
into notifiers that only get registered on the CPUs that actually have
the bugs.

There is right now no die notifier in the right place for this,
but you could just add one there. This is no performance critical
place.

-Andi

Re: [PATCH 1/2] x86: add is_f00f_bug helper to fault_32|64.c

[Ingo Molnar]

* Andi Kleen <andi@firstfloor.org> wrote:

> Harvey Harrison <harvey.harrison@gmail.com> writes:
> 
> > Further towards unifying these files, add another helper in same 
> > spirit as is_errata93.
> 
> The better way to handle this would be to move all these workarounds 
> into notifiers that only get registered on the CPUs that actually have 
> the bugs.
> 
> There is right now no die notifier in the right place for this, but 
> you could just add one there. This is no performance critical place.

agreed in principle, but i think it's perhaps a bit more maintainable if 
we first aimed for unification, then did such cleanups ontop of the 
unified code. Almost everything we do prior unification is double the 
work.

	Ingo

Re: [PATCH 1/2] x86: add is_f00f_bug helper to fault_32|64.c

[Andi Kleen]

> Almost everything we do prior unification is double the work.

That's not correct in this case as there are no common workarounds for i386 and 
x86-64. 

-Andi