On 5/18/20 12:06 PM, Jonathan Cameron wrote:
>
>
> On 18 May 2020 01:09:47 BST, "Andrew F. Davis" <afd@xxxxxx> wrote:
>> On 5/17/20 1:29 PM, jic23@xxxxxxxxxx wrote:
>>> From: Jonathan Cameron <Jonathan.Cameron@xxxxxxxxxx>
>>>
>>> One of a class of bugs pointed out by Lars in a recent review.
>>> iio_push_to_buffers_with_timestamp assumes the buffer used is aligned
>>> to the size of the timestamp (8 bytes). This is not guaranteed in
>>> this driver which uses a 32 byte array of smaller elements on the
>> stack.
>>> As Lars also noted this anti pattern can involve a leak of data to
>>> userspace and that indeed can happen here. We close both issues by
>>> moving to a suitable structure in the iio_priv() data with alignment
>>> explicitly requested. This data is allocated with kzalloc so no
>>> data can leak appart from previous readings.
>>>
>>> Fixes: eec96d1e2d31 ("iio: health: Add driver for the TI AFE4403
>> heart monitor")
>>> Reported-by: Lars-Peter Clausen <lars@xxxxxxxxxx>
>>> Signed-off-by: Jonathan Cameron <Jonathan.Cameron@xxxxxxxxxx>
>>> Cc: Andrew F. Davis <afd@xxxxxx>
>>> ---
>>> drivers/iio/health/afe4403.c | 9 ++++++---
>>> 1 file changed, 6 insertions(+), 3 deletions(-)
>>>
>>> diff --git a/drivers/iio/health/afe4403.c
>> b/drivers/iio/health/afe4403.c
>>> index e9f87e42ff4f..a3507624b30f 100644
>>> --- a/drivers/iio/health/afe4403.c
>>> +++ b/drivers/iio/health/afe4403.c
>>> @@ -65,6 +65,7 @@ static const struct reg_field afe4403_reg_fields[]
>> = {
>>> * @regulator: Pointer to the regulator for the IC
>>> * @trig: IIO trigger for this device
>>> * @irq: ADC_RDY line interrupt number
>>> + * @buffer: Used to construct data layout to push into IIO buffer.
>>> */
>>> struct afe4403_data {
>>> struct device *dev;
>>> @@ -74,6 +75,8 @@ struct afe4403_data {
>>> struct regulator *regulator;
>>> struct iio_trigger *trig;
>>> int irq;
>>> + /* Ensure suitable alignment for timestamp */
>>> + s32 buffer[8] __aligned(8);
>>
>>
>> One of those fancy structs with the timestamp specified would be nice
>> here like the other patches. IIRC we have 6 s32 channels, plus a s64
>> ts.
>
> I think we may only have some of those channels enabled. So ts may be it several
> locations in the buffer.
>
Might have been better to have the ts at the beginning, could have also
helped with alignment for when an odd number of channels are enabled.
> Hence we could use the structure approach but it might give a false sense
> of what is going on.
>
That's true, it can always be cleaned later then.
Andrew
> J
>>
>> Other than that everything looks good.
>>
>> Andrew
>>
>>
>>> };
>>>
>>> enum afe4403_chan_id {
>>> @@ -309,7 +312,6 @@ static irqreturn_t afe4403_trigger_handler(int
>> irq, void *private)
>>> struct iio_dev *indio_dev = pf->indio_dev;
>>> struct afe4403_data *afe = iio_priv(indio_dev);
>>> int ret, bit, i = 0;
>>> - s32 buffer[8];
>>> u8 tx[4] = {AFE440X_CONTROL0, 0x0, 0x0, AFE440X_CONTROL0_READ};
>>> u8 rx[3];
>>>
>>> @@ -326,7 +328,7 @@ static irqreturn_t afe4403_trigger_handler(int
>> irq, void *private)
>>> if (ret)
>>> goto err;
>>>
>>> - buffer[i++] = get_unaligned_be24(&rx[0]);
>>> + afe->buffer[i++] = get_unaligned_be24(&rx[0]);
>>> }
>>>
>>> /* Disable reading from the device */
>>> @@ -335,7 +337,8 @@ static irqreturn_t afe4403_trigger_handler(int
>> irq, void *private)
>>> if (ret)
>>> goto err;
>>>
>>> - iio_push_to_buffers_with_timestamp(indio_dev, buffer,
>> pf->timestamp);
>>> + iio_push_to_buffers_with_timestamp(indio_dev, afe->buffer,
>>> + pf->timestamp);
>>> err:
>>> iio_trigger_notify_done(indio_dev->trig);
>>>
>>>
>
