On Friday, November 08, 2013 1:07 PM, Dave Jones wrote:
> On Thu, Nov 07, 2013 at 06:27:54AM +0000, Linux Kernel wrote:
>> Gitweb:
>> http://git.kernel.org/linus/;a=commit;h=b3ff824a81e8978deb56f6d163479c1a0a606037
>> Commit: b3ff824a81e8978deb56f6d163479c1a0a606037
>> Parent: 97f4289ad08cffe55de06d4ac4f89ac540450aee
>> Author: H Hartley Sweeten <hswee...@visionengravers.com>
>> AuthorDate: Fri Aug 30 11:06:17 2013 -0700
>> Committer: Greg Kroah-Hartman <gre...@linuxfoundation.org>
>> CommitDate: Tue Sep 17 07:47:40 2013 -0700
>>
>> staging: comedi: drivers: use comedi_dio_update_state() for complex cases
>>
>> Use comedi_dio_update_state() to handle the boilerplate code to update
>> the subdevice s->state for more complex cases where the hardware is only
>> updated based on the 'mask' of the channels that are modified.
>>
>> Signed-off-by: H Hartley Sweeten <hswee...@visionengravers.com>
>> Reviewed-by: Ian Abbott <abbo...@mev.co.uk>
>> Signed-off-by: Greg Kroah-Hartman <gre...@linuxfoundation.org>
>
> This bit looks suspect :-
>
>> if (mask) {
>> - s->state &= ~mask;
>> - s->state |= (bits & mask);
>> -
>> if (mask & 0x00ff)
>> outb(s->state & 0xff, dev->iobase + reg);
>> - if ((mask & 0xff00) && (s->n_chan > 8))
>> + if ((mask & 0xff00) & (s->n_chan > 8))
>> outb((s->state >> 8) & 0xff, dev->iobase + reg + 1);
>> - if ((mask & 0xff0000) && (s->n_chan > 16))
>> + if ((mask & 0xff0000) & (s->n_chan > 16))
>> outb((s->state >> 16) & 0xff, dev->iobase + reg + 2);
>> - if ((mask & 0xff000000) && (s->n_chan > 24))
>> + if ((mask & 0xff000000) & (s->n_chan > 24))
>> outb((s->state >> 24) & 0xff, dev->iobase + reg + 3);
>> }
>
>
> should those be shifts ? Like so ?
>
>
> diff --git a/drivers/staging/comedi/drivers/pcl730.c
> b/drivers/staging/comedi/drivers/pcl730.c
> index d041b714db29..29a54e5d73d6 100644
> --- a/drivers/staging/comedi/drivers/pcl730.c
> +++ b/drivers/staging/comedi/drivers/pcl730.c
> @@ -173,11 +173,11 @@ static int pcl730_do_insn_bits(struct comedi_device
> *dev,
> if (mask) {
> if (mask & 0x00ff)
> outb(s->state & 0xff, dev->iobase + reg);
> - if ((mask & 0xff00) & (s->n_chan > 8))
> + if ((mask & 0xff00) & (s->n_chan >> 8))
> outb((s->state >> 8) & 0xff, dev->iobase + reg + 1);
> - if ((mask & 0xff0000) & (s->n_chan > 16))
> + if ((mask & 0xff0000) & (s->n_chan >> 16))
> outb((s->state >> 16) & 0xff, dev->iobase + reg + 2);
> - if ((mask & 0xff000000) & (s->n_chan > 24))
> + if ((mask & 0xff000000) & (s->n_chan >> 24))
> outb((s->state >> 24) & 0xff, dev->iobase + reg + 3);
> }
No, they should not be bit-shifts.
But, there is a bug here. The if tests should be like this:
if ((mask & 0xff00) && (s->n_chan > 8))
Sorry about that.
Do you want to post a patch? If I don't see one in the next couple days I'll
fix it.
Regards,
Hartley
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
