Platform: add Samsung Laptop platform driver

From: Greg Kroah-Hartman <[email protected]>

This adds the samsung-laptop driver to the kernel. It now supports all known Samsung laptops that use the SABI interface.

---
 Documentation/ABI/testing/sysfs-driver-samsung-laptop |   19 
 drivers/platform/x86/Kconfig                          |   13 
 drivers/platform/x86/Makefile                         |    1 
 drivers/platform/x86/samsung-laptop.c                 |  832 ++++++++++++++++++
 4 files changed, 865 insertions(+)

Matthew, this is a copy of the latest version in the linux-next tree of the driver. Once it has been accepted into your tree, I'll delete the staging driver. Or, if you want to give your ACK, I can just move the driver from staging into this directory in my tree, which will properly preserve the history of the code.

v2 - I have also addressed all of the review comments you gave me
previously, and fixed a bug in the previous version that broke the driver on the N128 laptop, as well as added support for a wide range of other laptops thanks to many testers who sent in patches.

v3 - Added a few more device ids, const changes, and now supports
variable brightness levels as not all laptops are the same with this.

new file mode 100644
index 0000000..0a81023
@@ -0,0 +1,19 @@
+What:		/sys/devices/platform/samsung/performance_level
+Date:		January 1, 2010
+KernelVersion:	2.6.33
+Contact:	Greg Kroah-Hartman <[email protected]>
+Description:	Some Samsung laptops have different "performance levels"
+		that are can be modified by a function key, and by this
+		sysfs file.  These values don't always make a whole lot
+		of sense, but some users like to modify them to keep
+		their fans quiet at all costs.  Reading from this file
+		will show the current performance level.  Writing to the
+		file can change this value.
+			Valid options:
+				"silent"
+				"normal"
+				"overclock"
+		Note that not all laptops support all of these options.
+		Specifically, not all support the "overclock" option,
+		and it's still unknown if this value even changes
+		anything, other than making the user feel a bit better.

index a59af5b..5624267 100644
@@ -654,4 +654,17 @@ config XO1_RFKILL
 	  Support for enabling/disabling the WLAN interface on the OLPC XO-1
 	  laptop.
 
+config SAMSUNG_LAPTOP
+	tristate "Samsung Laptop driver"
+	depends on RFKILL && BACKLIGHT_CLASS_DEVICE && X86
+	---help---
+	  This module implements a driver for a wide range of different
+	  Samsung laptops.  It offers control over the different
+	  function keys, wireless LED, LCD backlight level, and
+	  sometimes provides a "performance_control" sysfs file to allow
+	  the performance level of the laptop to be changed.
+
+	  To compile this driver as a module, choose M here: the module
+	  will be called samsung-laptop.
+
 endif # X86_PLATFORM_DEVICES

index 4ec4ff8..ad2fb97 100644
@@ -34,3 +34,4 @@ obj-$(CONFIG_INTEL_IPS)		+= intel_ips.o
 obj-$(CONFIG_GPIO_INTEL_PMIC)	+= intel_pmic_gpio.o
 obj-$(CONFIG_XO1_RFKILL)	+= xo1-rfkill.o
 obj-$(CONFIG_IBM_RTL)		+= ibm_rtl.o
+obj-$(CONFIG_SAMSUNG_LAPTOP)	+= samsung-laptop.o

Applied, thanks.

Matthew Garrett | [email protected]

Great, any pointers to where it was applied? I think I need to send you a few more 'add a new device id' patches and I want to make sure I get the right version to patch against.

thanks,

greg k-h

http://git.kernel.org/?p=linux/kernel/git/mjg59/platform-drivers-x86.git;a=shortlog;h=refs/heads/linux-next
- ideally we'd just have it id against all Samsung laptops and check for SABI? Or are there machines where that might break?

Matthew Garrett | [email protected]

Thanks.

I thought all samsung laptops had SABI, but as the other samsung laptop driver shows, that's not true.

I don't know what happens if we try to check for SABI on those machines. We are just rooting around in the BIOS rom and reading values, _hopefully_ that wouldn't break any other machines, but I don't really know.

I wouldn't mind just doing a DMI match on all Samsung laptops, it would save me having to constantly send patch updates, so I'll let this be your call :)

thanks,

greg k-h

I think it'd be worth a go, at least. dell-laptop has example alias lines for that, there's a few different DMI classes that can be used for laptops so you'd need to check which one Samsung are using.

Matthew Garrett | [email protected]

Hi Matthew,

concering the hp-wmi driver I would like to answer a few questions.

My GNU/Linux Debian system with a self compiled custom kernel 2.6.37.3 is running on a HP Pavilion DV7-4151sg notebook (dmidecode output, DSDT, and kernel config attached).
I activated the driver on my HP notebook in order to achieve several goals:
(1) make the rfkill hardware button work
(seems to work according to the output of rfkill list and acpi_listen)
(2) make the buttons for lowering and increasing display brightness work (do not work, no output with acpi_listen)
(3) the display back side has an illuminated "hp" logo which I would like to be able to switch off (alone or together with the display) to save a little energy
How can I found out if these functionality is achievable via ACPI WMI functions or somehow else?
What would I have to change in hp-wmi.c?

Any help is appreciated.
Best regards,
Jörg-Volker.

Hi Matthew,

concering the hp-wmi driver I would like to answer a few questions.

My GNU/Linux Debian system with a self compiled custom kernel 2.6.37.3 is running on a HP Pavilion DV7-4151sg notebook (dmidecode output, DSDT, and kernel config attached).
I activated the driver on my HP notebook in order to achieve several goals:
(1) make the rfkill hardware button work
(seems to work according to the output of rfkill list and acpi_listen)
(2) make the buttons for lowering and increasing display brightness work (do not work, no output with acpi_listen)
(3) the display back side has an illuminated "hp" logo which I would like to be able to switch off (alone or together with the display) to save a little energy
How can I found out if these functionality is achievable via ACPI WMI functions or somehow else?
What would I have to change in hp-wmi.c?

By the way, I see the following messages in dmesg concerning WMI (these messages did not appear before 2.6.37.4):

ACPI Error: Field [D128] at 1040 exceeds Buffer [NULL] size 160 (bits) (20101013/dsopcode-597)
ACPI Error: Method parse/execution failed [\_SB_.WMID.HWMC] (Node ffff88011be42320), AE_AML_BUFFER_LIMIT (20101013/psparse-537) ACPI Error: Method parse/execution failed [\_SB_.WMID.WMAD] (Node ffff88011be42500), AE_AML_BUFFER_LIMIT (20101013/psparse-537) ACPI Error: Field [D128] at 1040 exceeds Buffer [NULL] size 160 (bits) (20101013/dsopcode-597)
ACPI Error: Method parse/execution failed [\_SB_.WMID.HWMC] (Node ffff88011be42320), AE_AML_BUFFER_LIMIT (20101013/psparse-537) ACPI Error: Method parse/execution failed [\_SB_.WMID.WMAD] (Node ffff88011be42500), AE_AML_BUFFER_LIMIT (20101013/psparse-537) ACPI Error: Field [D128] at 1040 exceeds Buffer [NULL] size 160 (bits) (20101013/dsopcode-597)
ACPI Error: Method parse/execution failed [\_SB_.WMID.HWMC] (Node ffff88011be42320), AE_AML_BUFFER_LIMIT (20101013/psparse-537) ACPI Error: Method parse/execution failed [\_SB_.WMID.WMAD] (Node ffff88011be42500), AE_AML_BUFFER_LIMIT (20101013/psparse-537)
hp-wmi: probe of hp-wmi failed with error -22

Any help is appreciated.
Best regards,
Jörg-Volker.

P.S.: Please excuse the previous unready e-mail with the wrong subject.