Hi Igor,
On Wed, Mar 20, 2013 at 11:14 AM, Igor Grinberg <grinberg@xxxxxxxxxxxxxx> wrote:
> On 03/20/13 15:35, Fabio Estevam wrote:
>> From: Fabio Estevam <fabio.estevam@xxxxxxxxxxxxx>
>>
>> Building a kernel for imx_v4_v5_defconfig with CONFIG_USB_ULPI disabled, results
>> in the following error:
>>
>> arch/arm/mach-imx/built-in.o: In function 'pca100_init':
>> platform-mx2-emma.c:(.init.text+0x6788): undefined reference to 'otg_ulpi_create'
>> platform-mx2-emma.c:(.init.text+0x682c): undefined reference to 'mxc_ulpi_access_ops'
>
> Is this patch fixes both undefined references?
Yes, it fixes both.
> It does not look so..
> If it fixes only the first one, I would suggest adjusting the commit message.
Currently 'mxc_ulpi_access_ops' is defined in
arch/arm/mach-imx/ulpi.c, so if CONFIG_USB_ULPI is unselected the
unreference to 'mxc_ulpi_access_ops' will happen as reported in the
commit log.
If it makes sense I can move 'mxc_ulpi_access_ops' into
arch/arm/mach-imx/ulpi.h or even try to get rid of it and use the
generic otg_ulpi_create instead.
Either way, this should be handled by a different patch.
>> +#if IS_ENABLED(CONFIG_USB_ULPI)
>> struct usb_phy *otg_ulpi_create(struct usb_phy_io_ops *ops,
>> unsigned int flags);
>> +#else
>> +static inline struct usb_phy *otg_ulpi_create(struct usb_phy_io_ops *ops,
>> + unsigned int flags)
>> +{
>> + return NULL;
>
> Would ERR_PTR(-ENXIO) do better here?
The real *otg_ulpi_create function returns NULL on errors. Better to
keep the real and no-op version with the same return type, right?
Regards,
Fabio Estevam
--
To unsubscribe from this list: send the line "unsubscribe linux-usb" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
