Re: [PATCH 6/6] phy: twl4030-usb: add extcon to report cable connections.

From: Chanwoo Choi
Date: Mon May 11 2015 - 11:58:39 EST


Hi,

On Mon, May 11, 2015 at 10:38 PM, Kishon Vijay Abraham I <kishon@xxxxxx> wrote:
> +extcon MAINTAINERS
>
> Hi,
>
> On Thursday 16 April 2015 01:33 PM, NeilBrown wrote:
>>
>> From: NeilBrown <neilb@xxxxxxx>
>
>
> commit msg pls.
>
>>
>> Signed-off-by: NeilBrown <neilb@xxxxxxx>
>> ---
>> drivers/phy/phy-twl4030-usb.c | 67
>> +++++++++++++++++++++++++++++++++++++++++
>> 1 file changed, 67 insertions(+)
>>
>> diff --git a/drivers/phy/phy-twl4030-usb.c b/drivers/phy/phy-twl4030-usb.c
>> index 1d6f3e70193e..c42153d43ec2 100644
>> --- a/drivers/phy/phy-twl4030-usb.c
>> +++ b/drivers/phy/phy-twl4030-usb.c
>> @@ -40,6 +40,7 @@
>> #include <linux/regulator/consumer.h>
>> #include <linux/err.h>
>> #include <linux/slab.h>
>> +#include <linux/extcon.h>
>>
>> /* Register defines */
>>
>> @@ -173,6 +174,9 @@ struct twl4030_usb {
>> enum omap_musb_vbus_id_status linkstat;
>> bool vbus_supplied;
>>
>> + /* cable connection */
>> + struct extcon_dev edev;
>> +
>> struct delayed_work id_workaround_work;
>> };
>>
>> @@ -592,6 +596,54 @@ static ssize_t twl4030_usb_id_show(struct device
>> *dev,
>> }
>> static DEVICE_ATTR(id, 0444, twl4030_usb_id_show, NULL);
>>
>> +static const char *usb_cables[] = {
>> + "USB", /* id is floating */
>> + "Charger-downstream", /* id is floating and D+ shorted to D- */
>> + "USB-Host", /* id is ground */
>> + "USB-ACA", /* "Accessory Charger Adapter" id is something else */
>> + NULL

Current extcon core use the string for external connector when
registering the extcon device.
But, It is not clear. So, I'm working to implement it with the unique
id for each external connectors
instead of previous string name. You can refer to ongoing patch[1].

[1] https://git.kernel.org/cgit/linux/kernel/git/chanwoo/extcon.git/commit/?h=v4.2/extcon-next&id=ab73c3cb76816f904d91191b13b9140f3684fa35

I recommend that you stop the implementation of this patch until
finishing the update[2] of extcon core.
After complete the update[2], you'd better to implement this path
again with new method
to register the extcon device.
[2] https://git.kernel.org/cgit/linux/kernel/git/chanwoo/extcon.git/log/?h=v4.2/extcon-next


>> +};
>> +enum {
>> + TWL_CABLE_USB,
>> + TWL_CABLE_CHARGER, /* Not used - twl4030 can detect charger,
>> + * but driver cannot yet */
>> + TWL_CABLE_OTG,
>> + TWL_CABLE_ACA,
>> +};
>> +static u32 all_exclusive[] = {0xFFFFFFFF, 0};
>> +
>> +static void twl4030_usb_report_cable(struct twl4030_usb *twl)
>> +{
>> + enum twl4030_id_status sts;
>> +
>> + if (!cable_present(twl->linkstat)) {
>> + extcon_set_state(&twl->edev, 0);
>> + return;
>> + }
>> +
>> + sts = twl4030_get_id(twl);
>> +
>> + switch (sts) {
>> + case TWL4030_FLOATING: /* USB downstream */
>> + extcon_update_state(&twl->edev,
>> + 1<<TWL_CABLE_USB,
>> + 1<<TWL_CABLE_USB);
>> + break;
>> + case TWL4030_GROUND: /* USB host */
>> + extcon_update_state(&twl->edev,
>> + 1<<TWL_CABLE_OTG,
>> + 1<<TWL_CABLE_OTG);
>> + break;
>> + default: /* Some resistor */
>> + extcon_update_state(&twl->edev,
>> + 1<<TWL_CABLE_ACA,
>> + 1<<TWL_CABLE_ACA);
>> + /* An ACA should set port to 'host' mode */
>> + twl->linkstat = OMAP_MUSB_ID_GROUND;
>> + break;
>> + }

I don't prefer to use the extcon_update_state() function because
the extcon_update_state() function make the difficult code to handle
the extcon device driver. So, I'm planning to handle the extcon_update_state()
in only extcon core. Instead, extcon driver can use the
extcon_set_cable_{state|state_}
to update the state of cable.

>> +}
>> +
>> static irqreturn_t twl4030_usb_irq(int irq, void *_twl)
>> {
>> struct twl4030_usb *twl = _twl;
>> @@ -628,6 +680,7 @@ static irqreturn_t twl4030_usb_irq(int irq, void
>> *_twl)
>> pm_runtime_put_autosuspend(twl->dev);
>> }
>> omap_musb_mailbox(status);
>> + twl4030_usb_report_cable(twl);
>> }
>>
>> /* don't schedule during sleep - irq works right then */
>> @@ -766,6 +819,20 @@ static int twl4030_usb_probe(struct platform_device
>> *pdev)
>> }
>> usb_add_phy_dev(&twl->phy);
>>
>> + twl->edev.name = devm_kasprintf(twl->dev, GFP_KERNEL, "%s-usb",
>> + dev_name(twl->dev->parent));

The name of extcon device should be set by extcon core. you can refer
to patch[3]
about the extcon device's name.
[3] https://lkml.org/lkml/2015/4/27/258
- extcon: Modify the device name as extcon[X] for sysfs

>> + twl->edev.supported_cable = usb_cables;
>> + twl->edev.mutually_exclusive = all_exclusive;
>> + twl->edev.print_name = NULL; /* why would you change this? */
>> + twl->edev.print_state = NULL; /* probably want to change this */

I don't agree that you store some variable to extcon_dev structure directly.

There are both extcon provider driver and extcon consumer driver. So,
the extcon_dev structure should be handled on extcon provider driver
which included in drivers/extcon. Namely, current extcon subsystem
has the problem about this. Also, I'm working to resolve this problem.

I don't prefer to implement the new extcon provider driver on
non-'drivers/extcon' directory.

>
>
> Not sure about those two callbacks. Chanwoo?
>
> Thanks
> Kishon


Best Regards,
Chanwoo Choi
--
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/