Messages in this thread | | | Date | Mon, 15 Oct 2012 22:42:41 +0200 | Subject | Re: [PATCH v1] i2c-hid: introduce HID over i2c specification implementation | From | Benjamin Tissoires <> |
| |
On Sun, Oct 7, 2012 at 9:03 PM, Jean Delvare <khali@linux-fr.org> wrote: > On Sun, 7 Oct 2012 18:57:36 +0200, Benjamin Tissoires wrote: >> On Sun, Oct 7, 2012 at 4:28 PM, Jean Delvare <khali@linux-fr.org> wrote: >> > On Fri, 14 Sep 2012 15:41:43 +0200, benjamin.tissoires wrote: >> >> + u16 readRegister = ihid->hdesc.wDataRegister; >> > >> > This is missing le16_to_cpu(). >> >> I agree this is awful, but not putting it allows me to not have to >> check the endianness when I'm using it. >> But I may be totally wrong on this. > > I'm afraid I don't follow you. I want to see: > > u16 readRegister = le16_to_cpu(ihid->hdesc.wDataRegister); > > If you don't do that, your driver is broken on bigendian systems. And > there's no need to "check the endianness when you're using it", the > above should be enough for things to work just fine. >
a little bit late, but yes, you are entirely right. I was confused by the fact that I only wanted to use the number coming from the device to communicate with it, but as I made bit shifting within the CPU, I need to convert it. So forget my previous words here, it is fixed in v2.
Thanks Benjamin
> -- > Jean Delvare
| |