On Sat, Feb 10, 2024 at 5:09 PM Jonathan Cameron <jic23@xxxxxxxxxx> wrote: > On Mon, 5 Feb 2024 18:39:02 +0200 > Petre Rodan <petre.rodan@xxxxxxxxxxxxxxx> wrote: > > On Sun, Feb 04, 2024 at 06:03:28PM +0200, andy.shevchenko@xxxxxxxxx wrote: > > > > + memcpy(&data->scan.chan[1], &data->buffer[2], 2); > > > > > > Hmm... We don't have fixed-size memcpy() :-( > > > > __be16 *ptr; > > > > ptr = (__be16 *) data->buffer; > > data->scan.chan[0] = *ptr; > > data->scan.chan[1] = *++ptr; > > > > is this an acceptable replacement? I do not understand that your concern was, my > > intent was to copy exactly 2 bytes over. > > Andy? > > I'm not sure what you meant here either. It was just a rhetorical remark, no AR implied. I.o.w. the current code is okay. > There is an existing oddity that the read_raw deals with this as a be32 and > masking out the right sections for each channel rather than perhaps more logical > be16 pair here. -- With Best Regards, Andy Shevchenko
