Skip to contentSkip to conversation

patches to settle i8042 with Intel MID platforms

PATCH In review [email protected] 5 participants 18 messages Jul 5, 2010 – Jul 8, 2010

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

v2 cover patches to settle i8042 with Intel MID platforms

From: Feng Tang <[email protected]>

Hi All,

Some x86 platforms like intel MID platforms don't have i8042 controllers, and i8042 driver's probe to some non-exist legacy IO ports may hang the MID processor.

Current solution for this is add dependency of SERIO_I8042 over !X86_MRST, then if we build a x86 kernel image with X86_MRST=y, it won't have i8042 driver support over common x86 platforms.

So we add a i8042 pre-detect hook into x86_platforms ops, so that i8042 driver can use it to run-time judge whether to go on with probe process.

Please review these patches.

Thanks,
Feng

Changelog:

	v2:
	   * only call x86_platforms.i8042_detect() for x86 platforms

----------------
Feng Tang (5):
x86: add i8042 pre-detection hook to x86_platform_ops x86, mrst: add i8042_detect API for Moorestwon platform Revert "Input: do not force selecting i8042 on Moorestown" Revert "Input: fixup X86_MRST selects"
input: i8042 - add runtime check in x86's i8042_platform_init

 arch/x86/include/asm/x86_init.h       |    2 ++
 arch/x86/kernel/mrst.c                |    7 +++++++
 arch/x86/kernel/x86_init.c            |    4 +++-
 drivers/input/keyboard/Kconfig        |    4 ++--
 drivers/input/mouse/Kconfig           |    2 +-
 drivers/input/serio/Kconfig           |    2 +-
 drivers/input/serio/i8042-x86ia64io.h |    6 ++++++
 7 files changed, 22 insertions(+), 5 deletions(-)
#1 lore raw

PATCH v2 1/5 x86: add i8042 pre-detection hook to x86_platform_ops

From: Feng Tang <[email protected]>

Some x86 platforms like intel MID platforms don't have i8042 controllers, and i8042 driver's probe to some legacy IO ports may hang the MID processor. With this hook, i8042 driver can runtime check and skip the probe when the pretection fail which also saves some probe time

---
 arch/x86/include/asm/x86_init.h |    2 ++
 arch/x86/kernel/x86_init.c      |    4 +++-
 2 files changed, 5 insertions(+), 1 deletions(-)
arch/x86/include/asm/x86_init.h +2 -0
index 519b543..baa579c 100644
@@ -142,6 +142,7 @@ struct x86_cpuinit_ops {
  * @set_wallclock:		set time back to HW clock
  * @is_untracked_pat_range	exclude from PAT logic
  * @nmi_init			enable NMI on cpus
+ * @i8042_detect		pre-detect if i8042 controller exists
  */
 struct x86_platform_ops {
 	unsigned long (*calibrate_tsc)(void);
@@ -150,6 +151,7 @@ struct x86_platform_ops {
 	void (*iommu_shutdown)(void);
 	bool (*is_untracked_pat_range)(u64 start, u64 end);
 	void (*nmi_init)(void);
+	int (*i8042_detect)(void);
 };
 
 extern struct x86_init_ops x86_init;
arch/x86/kernel/x86_init.c +3 -1
index 61a1e8c..1e0c863 100644
@@ -85,6 +85,7 @@ struct x86_cpuinit_ops x86_cpuinit __cpuinitdata = {
 };
 
 static void default_nmi_init(void) { };
+static int default_i8042_detect(void) { return 1; };
 
 struct x86_platform_ops x86_platform = {
 	.calibrate_tsc			= native_calibrate_tsc,
@@ -92,5 +93,6 @@ struct x86_platform_ops x86_platform = {
 	.set_wallclock			= mach_set_rtc_mmss,
 	.iommu_shutdown			= iommu_shutdown_noop,
 	.is_untracked_pat_range		= is_ISA_range,
-	.nmi_init			= default_nmi_init
+	.nmi_init			= default_nmi_init,
+	.i8042_detect			= default_i8042_detect
 };
Signature · 1 line
1.7.0.4
#2 lore raw

[tip:x86/urgent] x86: Add i8042 pre-detection hook to x86_platform_ops

Commit-ID:  c516ac583973196162b1ba7e4d597d6f6892dac0
Gitweb:     http://git.kernel.org/tip/c516ac583973196162b1ba7e4d597d6f6892dac0
Author:     Feng Tang <[email protected]>
AuthorDate: Mon, 5 Jul 2010 23:03:18 +0800
Committer:  H. Peter Anvin <[email protected]>
CommitDate: Wed, 7 Jul 2010 17:05:06 -0700

x86: Add i8042 pre-detection hook to x86_platform_ops

Some x86 platforms like Intel MID platforms don't have i8042 controllers, and i8042 driver's probe to some legacy IO ports may hang the MID processor. With this hook, i8042 driver can runtime check and skip the probe when the pretection fail which also saves some probe time

[ hpa note: this is currently a compile-time check, which breaks the
i386 allyesconfig build. This patch series thus does fix a regression. ]

LKML-Reference: <[email protected]>

---
 arch/x86/include/asm/x86_init.h |    2 ++
 arch/x86/kernel/x86_init.c      |    4 +++-
 2 files changed, 5 insertions(+), 1 deletions(-)
arch/x86/include/asm/x86_init.h +2 -0
index 519b543..baa579c 100644
@@ -142,6 +142,7 @@ struct x86_cpuinit_ops {
  * @set_wallclock:		set time back to HW clock
  * @is_untracked_pat_range	exclude from PAT logic
  * @nmi_init			enable NMI on cpus
+ * @i8042_detect		pre-detect if i8042 controller exists
  */
 struct x86_platform_ops {
 	unsigned long (*calibrate_tsc)(void);
@@ -150,6 +151,7 @@ struct x86_platform_ops {
 	void (*iommu_shutdown)(void);
 	bool (*is_untracked_pat_range)(u64 start, u64 end);
 	void (*nmi_init)(void);
+	int (*i8042_detect)(void);
 };
 
 extern struct x86_init_ops x86_init;
arch/x86/kernel/x86_init.c +3 -1
index ebfb8e4..cd6da6b 100644
@@ -86,6 +86,7 @@ struct x86_cpuinit_ops x86_cpuinit __cpuinitdata = {
 };
 
 static void default_nmi_init(void) { };
+static int default_i8042_detect(void) { return 1; };
 
 struct x86_platform_ops x86_platform = {
 	.calibrate_tsc			= native_calibrate_tsc,
@@ -93,7 +94,8 @@ struct x86_platform_ops x86_platform = {
 	.set_wallclock			= mach_set_rtc_mmss,
 	.iommu_shutdown			= iommu_shutdown_noop,
 	.is_untracked_pat_range		= is_ISA_range,
-	.nmi_init			= default_nmi_init
+	.nmi_init			= default_nmi_init,
+	.i8042_detect			= default_i8042_detect
 };
 
 EXPORT_SYMBOL_GPL(x86_platform);
#3 lore raw

PATCH v2 2/5 x86, mrst: add i8042_detect API for Moorestwon platform

In reply to [email protected]

From: Feng Tang <[email protected]>

It will just return 0 as there is no i8042 controller

---
 arch/x86/kernel/mrst.c |    7 +++++++
 1 files changed, 7 insertions(+), 0 deletions(-)
arch/x86/kernel/mrst.c +7 -0
index e796448..5915e0b 100644
@@ -216,6 +216,12 @@ static void __init mrst_setup_boot_clock(void)
 		setup_boot_APIC_clock();
 };
 
+/* MID systems don't have i8042 controller */
+static int mrst_i8042_detect(void)
+{
+	return 0;
+}
+
 /*
  * Moorestown specific x86_init function overrides and early setup
  * calls.
@@ -233,6 +239,7 @@ void __init x86_mrst_early_setup(void)
 	x86_cpuinit.setup_percpu_clockev = mrst_setup_secondary_clock;
 
 	x86_platform.calibrate_tsc = mrst_calibrate_tsc;
+	x86_platform.i8042_detect = mrst_i8042_detect;
 	x86_init.pci.init = pci_mrst_init;
 	x86_init.pci.fixup_irqs = x86_init_noop;
 
Signature · 1 line
1.7.0.4
#4 lore raw

[tip:x86/urgent] x86, mrst: Add i8042_detect API for Moorestwon platform

Commit-ID:  6d2cce62017efe957e34cfcbba23861b7671980b
Gitweb:     http://git.kernel.org/tip/6d2cce62017efe957e34cfcbba23861b7671980b
Author:     Feng Tang <[email protected]>
AuthorDate: Mon, 5 Jul 2010 23:03:19 +0800
Committer:  H. Peter Anvin <[email protected]>
CommitDate: Wed, 7 Jul 2010 17:05:06 -0700

x86, mrst: Add i8042_detect API for Moorestwon platform

It will just return 0 as there is no i8042 controller

LKML-Reference: <[email protected]>

---
 arch/x86/kernel/mrst.c |    7 +++++++
 1 files changed, 7 insertions(+), 0 deletions(-)
arch/x86/kernel/mrst.c +7 -0
index e796448..5915e0b 100644
@@ -216,6 +216,12 @@ static void __init mrst_setup_boot_clock(void)
 		setup_boot_APIC_clock();
 };
 
+/* MID systems don't have i8042 controller */
+static int mrst_i8042_detect(void)
+{
+	return 0;
+}
+
 /*
  * Moorestown specific x86_init function overrides and early setup
  * calls.
@@ -233,6 +239,7 @@ void __init x86_mrst_early_setup(void)
 	x86_cpuinit.setup_percpu_clockev = mrst_setup_secondary_clock;
 
 	x86_platform.calibrate_tsc = mrst_calibrate_tsc;
+	x86_platform.i8042_detect = mrst_i8042_detect;
 	x86_init.pci.init = pci_mrst_init;
 	x86_init.pci.fixup_irqs = x86_init_noop;
 
#5 lore raw

PATCH v2 3/5 Revert "Input: do not force selecting i8042 on Moorestown"

In reply to [email protected]

From: Feng Tang <[email protected]>

This reverts commit 685afae02557a178185a4be36f58332976e79f63.

After adding x86_platform's detection for i8042 controller, we don't need the force dependency on !X86_MRST any more

---
 drivers/input/keyboard/Kconfig |    2 +-
 drivers/input/mouse/Kconfig    |    2 +-
 2 files changed, 2 insertions(+), 2 deletions(-)
drivers/input/keyboard/Kconfig +1 -1
index 0f9a478..917eec3 100644
@@ -73,7 +73,7 @@ config KEYBOARD_ATKBD
 	default y
 	select SERIO
 	select SERIO_LIBPS2
-	select SERIO_I8042 if X86 && !X86_MRST
+	select SERIO_I8042 if X86
 	select SERIO_GSCPS2 if GSC
 	help
 	  Say Y here if you want to use a standard AT or PS/2 keyboard. Usually
drivers/input/mouse/Kconfig +1 -1
index eeb58c1..c714ca2 100644
@@ -17,7 +17,7 @@ config MOUSE_PS2
 	default y
 	select SERIO
 	select SERIO_LIBPS2
-	select SERIO_I8042 if X86 && !X86_MRST
+	select SERIO_I8042 if X86
 	select SERIO_GSCPS2 if GSC
 	help
 	  Say Y here if you have a PS/2 mouse connected to your system. This
Signature · 1 line
1.7.0.4
#6 lore raw

[tip:x86/urgent] Revert "Input: do not force selecting i8042 on Moorestown"

Commit-ID:  44631ac64d06d2f7ce006c2a6f2c8e003a9c6ace
Gitweb:     http://git.kernel.org/tip/44631ac64d06d2f7ce006c2a6f2c8e003a9c6ace
Author:     Feng Tang <[email protected]>
AuthorDate: Mon, 5 Jul 2010 23:03:20 +0800
Committer:  H. Peter Anvin <[email protected]>
CommitDate: Wed, 7 Jul 2010 17:05:06 -0700

Revert "Input: do not force selecting i8042 on Moorestown"

This reverts commit 685afae02557a178185a4be36f58332976e79f63.

After adding x86_platform's detection for i8042 controller, we don't need the force dependency on !X86_MRST any more

LKML-Reference: <[email protected]>

---
 drivers/input/keyboard/Kconfig |    2 +-
 drivers/input/mouse/Kconfig    |    2 +-
 2 files changed, 2 insertions(+), 2 deletions(-)
drivers/input/keyboard/Kconfig +1 -1
index 0f9a478..917eec3 100644
@@ -73,7 +73,7 @@ config KEYBOARD_ATKBD
 	default y
 	select SERIO
 	select SERIO_LIBPS2
-	select SERIO_I8042 if X86 && !X86_MRST
+	select SERIO_I8042 if X86
 	select SERIO_GSCPS2 if GSC
 	help
 	  Say Y here if you want to use a standard AT or PS/2 keyboard. Usually
drivers/input/mouse/Kconfig +1 -1
index eeb58c1..c714ca2 100644
@@ -17,7 +17,7 @@ config MOUSE_PS2
 	default y
 	select SERIO
 	select SERIO_LIBPS2
-	select SERIO_I8042 if X86 && !X86_MRST
+	select SERIO_I8042 if X86
 	select SERIO_GSCPS2 if GSC
 	help
 	  Say Y here if you have a PS/2 mouse connected to your system. This
#7 lore raw

PATCH v2 4/5 Revert "Input: fixup X86_MRST selects"

In reply to [email protected]

From: Feng Tang <[email protected]>

This reverts commit 0b28bac5aef7bd1ab213723df031e61db9ff151a.

After adding x86_platform's detection for i8042 controller, we don't need the force dependency on !X86_MRST any more

Cc Randy Dunlap <[email protected]>
---
 drivers/input/keyboard/Kconfig |    2 +-
 drivers/input/serio/Kconfig    |    2 +-
 2 files changed, 2 insertions(+), 2 deletions(-)
drivers/input/keyboard/Kconfig +1 -1
index 917eec3..3525f53 100644
@@ -69,7 +69,7 @@ config KEYBOARD_ATARI
 	  module will be called atakbd.
 
 config KEYBOARD_ATKBD
-	tristate "AT keyboard" if EMBEDDED || !X86 || X86_MRST
+	tristate "AT keyboard" if EMBEDDED || !X86
 	default y
 	select SERIO
 	select SERIO_LIBPS2
drivers/input/serio/Kconfig +1 -1
index 256b9e9..3bfe8fa 100644
@@ -22,7 +22,7 @@ config SERIO_I8042
 	tristate "i8042 PC Keyboard controller" if EMBEDDED || !X86
 	default y
 	depends on !PARISC && (!ARM || ARCH_SHARK || FOOTBRIDGE_HOST) && \
-		   (!SUPERH || SH_CAYMAN) && !M68K && !BLACKFIN && !X86_MRST
+		   (!SUPERH || SH_CAYMAN) && !M68K && !BLACKFIN
 	help
 	  i8042 is the chip over which the standard AT keyboard and PS/2
 	  mouse are connected to the computer. If you use these devices,
Signature · 1 line
1.7.0.4
#8 lore raw

[tip:x86/urgent] Revert "Input: fixup X86_MRST selects"

Commit-ID:  c9d46f63f8e89fd70f97b83fdc4e5d2e37d92aeb
Gitweb:     http://git.kernel.org/tip/c9d46f63f8e89fd70f97b83fdc4e5d2e37d92aeb
Author:     Feng Tang <[email protected]>
AuthorDate: Mon, 5 Jul 2010 23:03:21 +0800
Committer:  H. Peter Anvin <[email protected]>
CommitDate: Wed, 7 Jul 2010 17:05:07 -0700

Revert "Input: fixup X86_MRST selects"

This reverts commit 0b28bac5aef7bd1ab213723df031e61db9ff151a.

After adding x86_platform's detection for i8042 controller, we don't need the force dependency on !X86_MRST any more

Cc Randy Dunlap <[email protected]>

LKML-Reference: <[email protected]>

---
 drivers/input/keyboard/Kconfig |    2 +-
 drivers/input/serio/Kconfig    |    2 +-
 2 files changed, 2 insertions(+), 2 deletions(-)
drivers/input/keyboard/Kconfig +1 -1
index 917eec3..3525f53 100644
@@ -69,7 +69,7 @@ config KEYBOARD_ATARI
 	  module will be called atakbd.
 
 config KEYBOARD_ATKBD
-	tristate "AT keyboard" if EMBEDDED || !X86 || X86_MRST
+	tristate "AT keyboard" if EMBEDDED || !X86
 	default y
 	select SERIO
 	select SERIO_LIBPS2
drivers/input/serio/Kconfig +1 -1
index 256b9e9..3bfe8fa 100644
@@ -22,7 +22,7 @@ config SERIO_I8042
 	tristate "i8042 PC Keyboard controller" if EMBEDDED || !X86
 	default y
 	depends on !PARISC && (!ARM || ARCH_SHARK || FOOTBRIDGE_HOST) && \
-		   (!SUPERH || SH_CAYMAN) && !M68K && !BLACKFIN && !X86_MRST
+		   (!SUPERH || SH_CAYMAN) && !M68K && !BLACKFIN
 	help
 	  i8042 is the chip over which the standard AT keyboard and PS/2
 	  mouse are connected to the computer. If you use these devices,
#9 lore raw

PATCH v2 5/5 input: i8042 - add runtime check in x86's i8042_platform_init

In reply to [email protected]

From: Feng Tang <[email protected]>

Then it will first check x86_platforms's i8042 detection result, then go on with normal probe.

Cc Dmitry Torokhov <[email protected]>
---
 drivers/input/serio/i8042-x86ia64io.h |    8 ++++++++
 1 files changed, 8 insertions(+), 0 deletions(-)
drivers/input/serio/i8042-x86ia64io.h +8 -0
index 6168469..6228526 100644
@@ -7,6 +7,8 @@
  * the Free Software Foundation.
  */
 
+#include <asm/x86_init.h>
+
 /*
  * Names.
  */
@@ -840,6 +842,12 @@ static int __init i8042_platform_init(void)
 {
 	int retval;
 
+#ifdef CONFIG_X86
+	/* Just return if pre-detection shows no i8042 controller exist */
+	if (!x86_platform.i8042_detect())
+		return -ENODEV;
+#endif
+
 /*
  * On ix86 platforms touching the i8042 data register region can do really
  * bad things. Because of this the region is always reserved on ix86 boxes.
Signature · 1 line
1.7.0.4
#10 lore raw

Re: [PATCH v2 5/5] input: i8042 - add runtime check in x86's i8042_platform_init

From: Feng Tang <[email protected]>

Then it will first check x86_platforms's i8042 detection result, then go on with normal probe.

---

Added #ifdef around #include <asm/x86_init.h> -- Tony

drivers/input/serio/i8042-x86ia64io.h +10 -0
index 6168469..81003c4 100644
@@ -7,6 +7,10 @@
  * the Free Software Foundation.
  */
 
+#ifdef CONFIG_X86
+#include <asm/x86_init.h>
+#endif
+
 /*
  * Names.
  */
@@ -840,6 +844,12 @@ static int __init i8042_platform_init(void)
 {
 	int retval;
 
+#ifdef CONFIG_X86
+	/* Just return if pre-detection shows no i8042 controller exist */
+	if (!x86_platform.i8042_detect())
+		return -ENODEV;
+#endif
+
 /*
  * On ix86 platforms touching the i8042 data register region can do really
  * bad things. Because of this the region is always reserved on ix86 boxes.
#11 lore raw

[tip:x86/urgent] input: i8042 - add runtime check in x86's i8042_platform_init

Commit-ID:  5cdfa1c3bbabb809ef3134f741a63e13373a8cad
Gitweb:     http://git.kernel.org/tip/5cdfa1c3bbabb809ef3134f741a63e13373a8cad
Author:     Feng Tang <[email protected]>
AuthorDate: Wed, 7 Jul 2010 13:02:16 -0700
Committer:  H. Peter Anvin <[email protected]>
CommitDate: Wed, 7 Jul 2010 17:05:07 -0700

input: i8042 - add runtime check in x86's i8042_platform_init

Then it will first check x86_platforms's i8042 detection result, then go on with normal probe.

LKML-Reference: <[email protected]>

---
 drivers/input/serio/i8042-x86ia64io.h |   10 ++++++++++
 1 files changed, 10 insertions(+), 0 deletions(-)
drivers/input/serio/i8042-x86ia64io.h +10 -0
index 6168469..81003c4 100644
@@ -7,6 +7,10 @@
  * the Free Software Foundation.
  */
 
+#ifdef CONFIG_X86
+#include <asm/x86_init.h>
+#endif
+
 /*
  * Names.
  */
@@ -840,6 +844,12 @@ static int __init i8042_platform_init(void)
 {
 	int retval;
 
+#ifdef CONFIG_X86
+	/* Just return if pre-detection shows no i8042 controller exist */
+	if (!x86_platform.i8042_detect())
+		return -ENODEV;
+#endif
+
 /*
  * On ix86 platforms touching the i8042 data register region can do really
  * bad things. Because of this the region is always reserved on ix86 boxes.
#12 lore raw
In reply to [email protected]
Quoted 12 lines · On 07/05/2010 08:03 AM, [email protected] wrote:

From: Feng Tang <[email protected]>

Hi All,

Some x86 platforms like intel MID platforms don't have i8042 controllers, and i8042 driver's probe to some non-exist legacy IO ports may hang the MID processor.

Current solution for this is add dependency of SERIO_I8042 over !X86_MRST, then if we build a x86 kernel image with X86_MRST=y, it won't have i8042 driver support over common x86 platforms.

Quoted 6 lines

So we add a i8042 pre-detect hook into x86_platforms ops, so that i8042 driver can use it to run-time judge whether to go on with probe process.

Please review these patches.

Looks good to me. Dmitry, what do you think?

-hpa

Signature · 2 lines
H. Peter Anvin, Intel Open Source Technology Center
I work for Intel.  I don't speak on their behalf.
maintainer
#13 lore raw
Quoted 13 lines · H. Peter Anvin
Quoted 12 lines · On 07/05/2010 08:03 AM, [email protected] wrote:

From: Feng Tang <[email protected]>

Hi All,

Some x86 platforms like intel MID platforms don't have i8042 controllers, and i8042 driver's probe to some non-exist legacy IO ports may hang the MID processor.

Current solution for this is add dependency of SERIO_I8042 over !X86_MRST, then if we build a x86 kernel image with X86_MRST=y, it won't have i8042 driver support over common x86 platforms.

Quoted 9 lines
Quoted 6 lines

So we add a i8042 pre-detect hook into x86_platforms ops, so that i8042 driver can use it to run-time judge whether to go on with probe process.

Please review these patches.

Looks good to me. Dmitry, what do you think?

Looks hood to me as well. We can either merge it through my tree or x86 tree, I don't have a preference.

Signature · 1 line
Dmitry
#14 lore raw
Quoted 4 lines · Dmitry Torokhov

Looks hood to me as well. We can either merge it through my tree or x86 tree, I don't have a preference.

OK if I merge it then with your Acked-by:?

-hpa

Signature · 2 lines
H. Peter Anvin, Intel Open Source Technology Center
I work for Intel.  I don't speak on their behalf.
maintainer
#15 lore raw
Quoted 8 lines · H. Peter Anvin
Quoted 4 lines · Dmitry Torokhov

Looks hood to me as well. We can either merge it through my tree or x86 tree, I don't have a preference.

OK if I merge it then with your Acked-by:?

Certainly.

Signature · 1 line
Dmitry
#16 lore raw
In reply to Dmitry Torokhov
Quoted 4 lines · Dmitry Torokhov

Looks hood to me as well. We can either merge it through my tree or x86 tree, I don't have a preference.

OK, I found that this series breaks allmodconfig, because i8042 can be a module. The obvious solution is to export x86_platform to modules, I presume that is okay with everyone (especially tglx)? If so I'll push this tree out tomorrow morning.

-hpa

#17 lore raw

[tip:x86/urgent] x86, platform: Export x86_platform to modules

In reply to [email protected]
Commit-ID:  72550b3ae545c75897c769d43d62d4be3f3d48fe
Gitweb:     http://git.kernel.org/tip/72550b3ae545c75897c769d43d62d4be3f3d48fe
Author:     H. Peter Anvin <[email protected]>
AuthorDate: Wed, 7 Jul 2010 16:57:46 -0700
Committer:  H. Peter Anvin <[email protected]>
CommitDate: Wed, 7 Jul 2010 17:05:05 -0700

x86, platform: Export x86_platform to modules

Export x86_platform to modules in preparation of using it for i8042 discovery control.

LKML-Reference: <[email protected]>
Cc: Thomas Gleixner <[email protected]>
Cc: Feng Tang <[email protected]>
Cc: Dmitry Torokhov <[email protected]>

---
 arch/x86/kernel/x86_init.c |    3 +++
 1 files changed, 3 insertions(+), 0 deletions(-)
arch/x86/kernel/x86_init.c +3 -0
index 61a1e8c..ebfb8e4 100644
@@ -5,6 +5,7 @@
  */
 #include <linux/init.h>
 #include <linux/ioport.h>
+#include <linux/module.h>
 
 #include <asm/bios_ebda.h>
 #include <asm/paravirt.h>
@@ -94,3 +95,5 @@ struct x86_platform_ops x86_platform = {
 	.is_untracked_pat_range		= is_ISA_range,
 	.nmi_init			= default_nmi_init
 };
+
+EXPORT_SYMBOL_GPL(x86_platform);