Re: [next:master 1495/1772] drivers/staging/iio/iio_simple_dummy_buffer.c:86:2: error: implicit declaration of function 'iio_push_to_buffers_with_timestamp'

[Lars-Peter Clausen <lars@metafoo.de>]

On 09/25/2013 06:22 PM, Jonathan Cameron wrote:
> 
> 
> Jonathan Cameron <jic23@kernel.org> wrote:
>>
>>
>> Lars-Peter Clausen <lars@metafoo.de> wrote:
>>> On 09/24/2013 05:28 AM, kbuild test robot wrote:
>>>> tree:  
>>> git://git.kernel.org/pub/scm/linux/kernel/git/next/linux-next.git
>>> master
>>>> head:   1fa8f71dfa6e28c89afad7ac71dcb19b8c8da8b7
>>>> commit: 09a799d04eb17e19bbc80cb0e350a3ec89552b01 [1495/1772]
>>> staging:iio:dummy: Use iio_push_to_buffers_with_timestamp()
>>>> config: i386-randconfig-j6-09240827 (attached as .config)
>>>>
>>>> All error/warnings:
>>>>
>>>>     drivers/staging/iio/iio_simple_dummy_buffer.c: In function
>>> 'iio_simple_dummy_trigger_h':
>>>>>> drivers/staging/iio/iio_simple_dummy_buffer.c:86:2: error:
>> implicit
>>> declaration of function 'iio_push_to_buffers_with_timestamp'
>>> [-Werror=implicit-function-declaration]
>>>>       iio_push_to_buffers_with_timestamp(indio_dev, data,
>>> iio_get_time_ns());
>>>>       ^
>>>>     drivers/staging/iio/iio_simple_dummy_buffer.c: At top level:
>>>>     drivers/staging/iio/iio_simple_dummy_buffer.c:108:16: error:
>>> 'iio_sw_buffer_preenable' undeclared here (not in a function)
>>>>       .preenable = &iio_sw_buffer_preenable,
>>>>                     ^
>>>>     drivers/staging/iio/iio_simple_dummy_buffer.c: In function
>>> 'iio_simple_dummy_configure_buffer':
>>>>     drivers/staging/iio/iio_simple_dummy_buffer.c:141:8: error:
>>> dereferencing pointer to incomplete type
>>>>       buffer->scan_timestamp = true;
>>>>             ^
>>>>     cc1: some warnings being treated as errors
>>>>
>>>> vim +/iio_push_to_buffers_with_timestamp +86
>>> drivers/staging/iio/iio_simple_dummy_buffer.c
>>>>
>>>>      80				/* random access read from the 'device' */
>>>>      81				data[i] = fakedata[j];
>>>>      82				len += 2;
>>>>      83			}
>>>>      84		}
>>>>      85	
>>>>    > 86		iio_push_to_buffers_with_timestamp(indio_dev, data,
>>> iio_get_time_ns());
>>>>      87	
>>>>      88		kfree(data);
>>>>      89	
>>>
>>>
>>> Hm, the config that generated this has:
>>>
>>> CONFIG_IIO=m
>>> # CONFIG_IIO_BUFFER is not set
>>> CONFIG_IIO_KFIFO_BUF=m
>>>
>>> How is it possible that IIO_KFIFO_BUF is selected but not IIO_BUFFER?
>>>
>> How does a SELECT interact with an IF in kconfig?  I am not sure..
>>
>> We select the kfifo_buf and that is under IF IIO_BUFFER.
> 
> The documentation for select makes it clear that it does not enforce dependencies so I guess any select of Kfifo_buf must also select iio_buffer?  Sometimes there seems to be no right way of doing this stuff...

Just wanted to say that we never do this, but then looked at the latest
iio/togreg and the commit that causes it seems to be "ca9a5638 iio:
ti_am335x_adc: Add continuous sampling support". Which selects the KFIFO
buffer without either selecting IIO_BUFFER as well or depending on IIO_BUFFER.

- Lars