On Sun, Aug 13, 2017 at 7:15 AM, Andy Shevchenko <andriy.shevchenko@xxxxxxxxxxxxxxx> wrote: > On Sun, 2017-08-06 at 08:43 -0700, Andy Lutomirski wrote: >> On Fri, Aug 4, 2017 at 8:27 AM, Andy Shevchenko >> <andy.shevchenko@xxxxxxxxx> wrote: >> > On Fri, Aug 4, 2017 at 6:01 PM, Andy Lutomirski <luto@xxxxxxxxxx> >> > wrote: >> > > > On Aug 2, 2017, at 9:28 AM, Andy Shevchenko <andriy.shevchenko@l >> > > > inux.intel.com> wrote: > >> > > NAK. guid_block is a firmware interface, so opaque kernel types >> > > don't >> > > belong in it. >> > >> > I f we leave this, what do you think about everything else? >> >> Assuming it works, it's fine with me. I'd be happy to test. > > Just sent v2. > >> Keep in mind that this beast is a *little-endian* GUID abomination, >> and I don't see generic conversion helpers. Something might need to >> be added. > > Do you mean something like char16_to_guid() / char16_to_uuid() ? No, I mean something like uuid_le_to_guid(). There's an existing helper uuid_le_to_bin() which, for all that the implementation and signature is ridiculous, does the right thing.
