On 01/30/2016 03:18 PM, Jonathan Cameron wrote:
> On 25/01/16 15:50, Arnd Bergmann wrote:
>> The ad5933_i2c_read function returns an error code to indicate
>> whether it could read data or not. However ad5933_work() ignores
>> this return code and just accesses the data unconditionally,
>> which gets detected by gcc as a possible bug:
>>
>> drivers/staging/iio/impedance-analyzer/ad5933.c: In function 'ad5933_work':
>> drivers/staging/iio/impedance-analyzer/ad5933.c:649:16: warning: 'status' may be used uninitialized in this function [-Wmaybe-uninitialized]
>>
>> This adds minimal error handling so we only evaluate the
>> data if it was correctly read.
>>
>> Signed-off-by: Arnd Bergmann <arnd@xxxxxxxx>
> Hi Arnd,
>
> Thanks for the patch. The handling in here is a little fiddly
> by the look of things. Lars can you take a look at this when
> you have a minute?
>
> At a very high level, it doesn't make sense to fix this instance and
> not the one in the context of the patch below.
> See below...
>> ---
>> drivers/staging/iio/impedance-analyzer/ad5933.c | 5 +++--
>> 1 file changed, 3 insertions(+), 2 deletions(-)
>>
>> diff --git a/drivers/staging/iio/impedance-analyzer/ad5933.c b/drivers/staging/iio/impedance-analyzer/ad5933.c
>> index 10c43dda0f5a..304bb464e478 100644
>> --- a/drivers/staging/iio/impedance-analyzer/ad5933.c
>> +++ b/drivers/staging/iio/impedance-analyzer/ad5933.c
>> @@ -647,6 +647,7 @@ static void ad5933_work(struct work_struct *work)
>> __be16 buf[2];
>> int val[2];
>> unsigned char status;
>> + int ret;
>>
>> mutex_lock(&indio_dev->mlock);
>> if (st->state == AD5933_CTRL_INIT_START_FREQ) {
>> @@ -658,9 +659,9 @@ static void ad5933_work(struct work_struct *work)
>> return;
>> }
>>
>> - ad5933_i2c_read(st->client, AD5933_REG_STATUS, 1, &status);
>> + ret = ad5933_i2c_read(st->client, AD5933_REG_STATUS, 1, &status);
>>
>> - if (status & AD5933_STAT_DATA_VALID) {
>> + if (!ret && (status & AD5933_STAT_DATA_VALID)) {
> The else is non trivial here as it assumes we will get the data later. If we
> get such a failure, we probably want to drop out completely rather than paper
> over the gaps..
I agree. Although we could argue that Arnd's approach allows to recover from
temporary failure. But then again we don't want to keep polling forever if
it's a permanent failure. I'd say the best thing for a quick fix is to just
error out and assume the error is permanent.
>> int scan_count = bitmap_weight(indio_dev->active_scan_mask,
>> indio_dev->masklength);
> Same issue on the next line - this results in known garbage data being spooled
> out.
Also agreed.
_______________________________________________
devel mailing list
devel@xxxxxxxxxxxxxxxxxxxxxx
http://driverdev.linuxdriverproject.org/mailman/listinfo/driverdev-devel
