On Thu, Apr 27, 2017 at 4:50 PM, Eric Sandeen <sandeen@xxxxxxxxxxx> wrote: > > > On 4/27/17 8:22 AM, David Howells wrote: >> Amir Goldstein <amir73il@xxxxxxxxx> wrote: >> >>> You mean if it is NULL_UUID_LE? Yes, that is what I do now. >>> Christoph suggested that this is not robust enough. >>> I'm rather indifferent about the need of a flag. >> >> I was thinking more that s_uuid would be a NULL pointer if the fs doesn't have >> a uuid. > > but s_uuid in the vfs superblock isn't a pointer: > > u8 s_uuid[16]; /* UUID */ > > (or am I missing your point?) > > -Eric > IMO, its just too simple to compare to NULL_UUID_LE and/or the set/check a flag. I don't think we should bother with any other option. Amir.
