[Lucid SRU] UBUNTU: SAUCE: dell-laptop: fire SMI when toggling hardware killswitch

[Eric Miao]

On Wed, Jun 9, 2010 at 3:44 PM, Keng-Yu Lin <keng-yu.lin at canonical.com> wrote:
> SRU Justification:
>
> Impact:
> A type of Dell Inspiron Mini does not turn off bluetooth physically when
> pressing F2 (rfkill hotkey).
>
> Fix:
> Current kernel code expects hardware to handle the hard-rfkill switching
> spontaneously and only notify rfkill subsystem with the status change.
> This patch makes kernel to explicitly fire a SMI to switch on/off
> rfkill devices if BIOS reports that the hardware switch is not supported.
>
> This patch depends on the following sauce patch in lucid kernel so is not going upstream:
>        UBUNTU: SAUCE: dell-laptop: Store the HW switch status internally rather than requerying every time
>
> Testcase:
> I tested the patch on a Dell Mini 10. It works smoothly.
>
> --
> From b17359e833677297d4d5daf78e4d9dc460e51285 Mon Sep 17 00:00:00 2001
> From: Keng-Yu Lin <keng-yu.lin at canonical.com>
> Date: Thu, 3 Jun 2010 10:31:56 +0800
> Subject: [PATCH] UBUNTU: SAUCE: dell-laptop: fire SMI when toggling hardware killswitch
>
> When BIOS reports hardware switch is not supported, it does not turn on/off all
> devices spontaneously. So it is neccessary to fire an SMI to switch the rfkill
> status explicitly when toggling the hardware killswitch.
>
> BugLink: http://bugs.launchpad.net/bugs/590607
>
> Signed-off-by: Keng-Yu Lin <keng-yu.lin at canonical.com>
> ---
>  drivers/platform/x86/dell-laptop.c |    4 +++-
>  1 files changed, 3 insertions(+), 1 deletions(-)
>
> diff --git a/drivers/platform/x86/dell-laptop.c b/drivers/platform/x86/dell-laptop.c
> index 15d96a0..e54ab12 100644
> --- a/drivers/platform/x86/dell-laptop.c
> +++ b/drivers/platform/x86/dell-laptop.c
> @@ -29,6 +29,7 @@
>  #define WLAN_SWITCH_MASK 0
>  #define BT_SWITCH_MASK 1
>  #define WWAN_SWITCH_MASK 2
> +#define HW_SWITCH_SUPPORT 3
>  #define HW_SWITCH_MASK 16
>
>  /* This structure will be modified by the firmware when we enter
> @@ -240,7 +241,7 @@ static int dell_rfkill_set(void *data, bool blocked)
>        int disable = blocked ? 1 : 0;
>        unsigned long radio = (unsigned long)data;
>
> -       if (!(hw_switch_status & BIT(radio-1)) || !(hw_switch_status & BIT(HW_SWITCH_MASK))) {
> +       if (!(hw_switch_status & BIT(radio-1)) || !(hw_switch_status & BIT(HW_SWITCH_MASK)) || !(hw_switch_status & BIT(HW_SWITCH_SUPPORT))) {

Coding style perspective, this is too long. You can either wrap it
or introduce a static inline function, e.g. hw_support_switch() to
clean up the conditions here.

>                memset(&buffer, 0, sizeof(struct calling_interface_buffer));
>                buffer.input[0] = (1 | (radio<<8) | (disable << 16));
>                dell_send_request(&buffer, 17, 11);
> @@ -258,6 +259,7 @@ static void dell_rfkill_query(struct rfkill *rfkill, void *data)
>        dell_send_request(&buffer, 17, 11);
>        status = buffer.output[1];
>
> +       hw_switch_status |= (status & BIT(0)) << BIT(HW_SWITCH_SUPPORT);
>        hw_switch_status |= (status & BIT(HW_SWITCH_MASK)) ^ BIT(HW_SWITCH_MASK);
>

Patch looks good to me, and we should get this upstreamed as well.