Re: [PATCH v2 1/2] toshiba_acpi: Add support for WWAN devices

[Azael Avalos]

Hi Darren,

2015-11-20 15:49 GMT-07:00 Darren Hart <dvhart@xxxxxxxxxxxxx>:
> On Thu, Nov 19, 2015 at 08:49:24AM -0700, Azael Avalos wrote:
>> Toshiba laptops with WWAN devices installed cannot use the device unless
>> it is attached and powered, similar to how Toshiba Bluetooth devices
>> work.
>>
>> This patch adds support to WWAN devices, introducing three functions,
>> one to query the overall status of the wireless devices (RFKill, WLAN,
>> BT, WWAN), the second queries WWAN support, and finally the third
>> (de)activates the device.
>>
>> Signed-off-by: Fabian Koester <fabian.koester@xxxxxxxxxxxx>
>> Signed-off-by: Azael Avalos <coproscefalo@xxxxxxxxx>
>
> Thanks Azael,
>
> A few comments on code flow and one bug I think below.
>
>> ---
>>  drivers/platform/x86/toshiba_acpi.c | 92 +++++++++++++++++++++++++++++++++++++
>>  1 file changed, 92 insertions(+)
>>
>> diff --git a/drivers/platform/x86/toshiba_acpi.c b/drivers/platform/x86/toshiba_acpi.c
>> index c013029..60d1ad9 100644
>> --- a/drivers/platform/x86/toshiba_acpi.c
>> +++ b/drivers/platform/x86/toshiba_acpi.c
>> @@ -114,6 +114,7 @@ MODULE_LICENSE("GPL");
>>  #define HCI_VIDEO_OUT                        0x001c
>>  #define HCI_HOTKEY_EVENT             0x001e
>>  #define HCI_LCD_BRIGHTNESS           0x002a
>> +#define HCI_WIRELESS                 0x0056
>>  #define HCI_ACCELEROMETER            0x006d
>>  #define HCI_KBD_ILLUMINATION         0x0095
>>  #define HCI_ECO_MODE                 0x0097
>> @@ -148,6 +149,10 @@ MODULE_LICENSE("GPL");
>>  #define SCI_KBD_MODE_ON                      0x8
>>  #define SCI_KBD_MODE_OFF             0x10
>>  #define SCI_KBD_TIME_MAX             0x3c001a
>> +#define HCI_WIRELESS_STATUS          0x1
>> +#define HCI_WIRELESS_WWAN            0x3
>> +#define HCI_WIRELESS_WWAN_STATUS     0x2000
>> +#define HCI_WIRELESS_WWAN_POWER              0x4000
>>  #define SCI_USB_CHARGE_MODE_MASK     0xff
>>  #define SCI_USB_CHARGE_DISABLED              0x00
>>  #define SCI_USB_CHARGE_ALTERNATE     0x09
>> @@ -197,12 +202,14 @@ struct toshiba_acpi_dev {
>>       unsigned int kbd_function_keys_supported:1;
>>       unsigned int panel_power_on_supported:1;
>>       unsigned int usb_three_supported:1;
>> +     unsigned int wwan_supported:1;
>>       unsigned int sysfs_created:1;
>>       unsigned int special_functions;
>>
>>       bool kbd_led_registered;
>>       bool illumination_led_registered;
>>       bool eco_led_registered;
>> +     bool killswitch;
>>  };
>>
>>  static struct toshiba_acpi_dev *toshiba_acpi;
>> @@ -1085,6 +1092,87 @@ static int toshiba_hotkey_event_type_get(struct toshiba_acpi_dev *dev,
>>       return -EIO;
>>  }
>>
>> +/* Wireless status (RFKill, WLAN, BT, WWAN) */
>> +static int toshiba_wireless_status(struct toshiba_acpi_dev *dev)
>> +{
>> +     u32 in[TCI_WORDS] = { HCI_GET, HCI_WIRELESS, 0, 0, 0, 0 };
>> +     u32 out[TCI_WORDS];
>> +     acpi_status status;
>> +
>> +     in[3] = HCI_WIRELESS_STATUS;
>> +     status = tci_raw(dev, in, out);
>> +     if (ACPI_FAILURE(status)) {
>> +             pr_err("ACPI call to get Wireless status failed\n");
>> +     } else if (out[0] == TOS_NOT_SUPPORTED) {
>> +             return -ENODEV;
>> +     } else if (out[0] == TOS_SUCCESS) {
>> +             dev->killswitch =
>> +                         (out[2] & HCI_WIRELESS_STATUS) ? true : false;
>
> This should assign successfully without the need for the ternary operator. You
> can also then drop the extra newline. You can always use:
>
> !!(out[2] & HCI_WIRELESS_STATUS)
>
> To ensure a 1 or 0 assignment.

OK, will change on v3.

>
>> +             return 0;
>> +     }
>> +
>> +     return -EIO;
>
> Also, we should be testing for error and do the expected path outside the if
> blocks.
>
>
>         if (ACPI_FAILURE(status) {
>                 pr_err("ACPI call to get Wireless status failed\n");
>                 return -EIO;
>         }
>
>         if (out[0] == TOS_NOT_SUPPORTED)
>                 return -ENODEV;
>
>         if (out[0] != TOS_SUCCESS)
>                 return -EIO;
>
>         dev->killswitch = !!(out[2] & HCI_WIRELESS_STATUS);
>
>         return 0;
>

OK, will change the functions to this style on v3.

>> +}
>> +
>> +/* WWAN */
>> +static void toshiba_wwan_available(struct toshiba_acpi_dev *dev)
>> +{
>> +     u32 in[TCI_WORDS] = { HCI_GET, HCI_WIRELESS, 0, 0, 0, 0 };
>> +     u32 out[TCI_WORDS];
>> +     acpi_status status;
>> +
>> +     dev->wwan_supported = 0;
>> +
>> +     /*
>> +      * WWAN support can be queried by setting the in[3] value to
>> +      * HCI_WIRELESS_WWAN (0x03).
>> +      *
>> +      * If supported, out[0] contains TOS_SUCCESS and out[2] contains
>> +      * HCI_WIRELESS_WWAN_STATUS (0x2000).
>> +      *
>> +      * If not supported, out[0] contains TOS_INPUT_DATA_ERROR (0x8300)
>> +      * or TOS_NOT_SUPPORTED (0x8000).
>> +      */
>> +     in[3] = HCI_WIRELESS_WWAN;
>> +     status = tci_raw(dev, in, out);
>> +     if (ACPI_FAILURE(status))
>> +             pr_err("ACPI call to get WWAN status failed\n");
>> +     else if (out[0] == TOS_SUCCESS && out[2] == HCI_WIRELESS_WWAN_STATUS)
>> +             dev->wwan_supported = 1;
>
> This block similarly intermixes error checking with the primary functional
> logic, making it less legible in my opinion. Consider:
>
>
>         in[3] = HCI_WIRELESS_WWAN;
>         status = tci_raw(dev, in, out);
>
>         if (ACPI_FAILURE(status) || (out[0] != TOS_SUCCESS)) {
>                 pr_err("ACPI call to get WWAN status failed\n");
>                 return;
>         }
>
>         dev->wwan_supported = (out[2] == HCI_WIRELESS_WWAN_STATUS);
>
>> +}
>> +
>> +static int toshiba_wwan_set(struct toshiba_acpi_dev *dev, u32 state)
>> +{
>> +     u32 in[TCI_WORDS] = { HCI_SET, HCI_WIRELESS, state, 0, 0, 0 };
>> +     u32 out[TCI_WORDS];
>> +     acpi_status status;
>> +
>> +     in[3] = HCI_WIRELESS_WWAN_STATUS;
>> +     status = tci_raw(dev, in, out);
>> +     if (ACPI_FAILURE(status)) {
>> +             pr_err("ACPI call to set WWAN status failed\n");
>> +             return -EIO;
>> +     } else if (out[0] == TOS_NOT_SUPPORTED) {
>> +             return -ENODEV;
>> +     } else if (out[0] != TOS_SUCCESS) {
>> +             return -EIO;
>> +     }
>> +
>> +     /*
>> +      * Some devices only need to call HCI_WIRELESS_WWAN_STATUS to
>> +      * (de)activate the device, but some others need the
>> +      * HCI_WIRELESS_WWAN_POWER call as well.
>> +      */
>> +     in[3] = HCI_WIRELESS_WWAN_POWER;
>> +     status = tci_raw(dev, in, out);
>> +     if (ACPI_FAILURE(status))
>> +             pr_err("ACPI call to set WWAN power failed\n");
>
> I believe you want a return -EIO here?
>

Yes, and actually, I think the whole driver functions are like this,
I'll check once
I get back home and send a separate patch for this issue.

>> +     else if (out[0] == TOS_NOT_SUPPORTED)
>> +             return -ENODEV;
>> +
>> +     return out[0] == TOS_SUCCESS ? 0 : -EIO;
>
> So much ternary! :-) I suppose this one is OK.


Alright, once I get back home I'll update these patches according
to your comments and send a v3.


Cheers
Azael


-- 
-- El mundo apesta y vosotros apestais tambien --
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/