On 2013-08-30 01:47, H Hartley Sweeten wrote:
The subdevice 'io_bits' is a bit mask of the i/o configuration for digital subdevices. '0' values indicate that a channel is configured as an input and '1' values that the channel is an output. Since the subdevice data is kzalloc()'d, all channels default as inputs. Modify __comedi_device_postconfig() so that the 'io_bits' are correctly initialized for Digital Output subdevices. Remove all the unnecessary initializations of 's->io_bits' from the drivers. Also, remove the unnecessary initialization of the 's->state'. Signed-off-by: H Hartley Sweeten <hsweeten@xxxxxxxxxxxxxxxxxxx> Cc: Ian Abbott <abbotti@xxxxxxxxx> Cc: Greg Kroah-Hartman <gregkh@xxxxxxxxxxxxxxxxxxx>
[snip]
diff --git a/drivers/staging/comedi/drivers.c b/drivers/staging/comedi/drivers.c index 317a821..1c81d65 100644 --- a/drivers/staging/comedi/drivers.c +++ b/drivers/staging/comedi/drivers.c @@ -285,6 +285,9 @@ static int __comedi_device_postconfig(struct comedi_device *dev) if (s->type == COMEDI_SUBD_UNUSED) continue; + if (s->type == COMEDI_SUBD_DO) + s->io_bits = (1 << s->n_chan) - 1; +
As pointed out by Dan Carpenter, this won't work for s->n_chan >= 32. We definitely care about it working for s->n_chan == 32, so like in your original patch, s->io_bits should be set to 0xffffffff when s->n_chan >= 32. It's my fault for leading you astray.
[snip] -- -=( Ian Abbott @ MEV Ltd. E-mail: <abbotti@xxxxxxxxx> )=- -=( Tel: +44 (0)161 477 1898 FAX: +44 (0)161 718 3587 )=- _______________________________________________ devel mailing list devel@xxxxxxxxxxxxxxxxxxxxxx http://driverdev.linuxdriverproject.org/mailman/listinfo/driverdev-devel
