Re: [PATCH 1/2] afs: Move UUID struct to linux/uuid.h

[Arnd Bergmann]

On Thursday, January 12, 2017 2:12:56 PM CET David Howells wrote:
> Arnd Bergmann <arnd@xxxxxxxx> wrote:
> 
> > Looks good to me, but I wonder if this part:
> > 
> >                 r = call->request;
> > -               r->time_low                     = ntohl(b[0]);
> > -               r->time_mid                     = ntohl(b[1]);
> > -               r->time_hi_and_version          = ntohl(b[2]);
> > +               r->time_low                     = b[0];
> > +               r->time_mid                     = htons(ntohl(b[1]));
> > +               r->time_hi_and_version          = htons(ntohl(b[2]));
> >                 r->clock_seq_hi_and_reserved    = ntohl(b[3]);
> >                 r->clock_seq_low                = ntohl(b[4]);
> >  
> > should be considered a bugfix and split out into a
> > separate patch.
> 
> I changed the definitions in the struct from u16/u32 to __be16/__be32 so it's
> not a bugfix.

Ok.

> > From what I understand about the mess in UUID formats, the time fields can
> > either be big-endian (as defined) or little-endian (for all things
> > Microsoft),
> 
> RFC 4122 specified that the multi-octet fields are stored MSB-first.
> 
> > and you are changing the representation from CPU-specific to big-endian,
> > which makes it different for x86 and most ARM at least.
> 
> In-kernel, not in the protocol.

Ok, I assumed that the uuid was later sent out over the wire again
in the in-memory format, but you are right, it does get sent out in
the AFS specific format as a series of 32-bit big-endian values
rather than the RFC4122 format.
 
> The problem is that you can't do what you put in your suggested patch and just
> copy the UUID produced by the generate_random_uuid() over the afs_uuid struct
> since that puts the version in the wrong place.

Got it. One more thing then:

> @@ -569,9 +569,9 @@ static void SRXAFSCB_TellMeAboutYourself(struct
> work_struct *work)> 
>         memset(&reply, 0, sizeof(reply));
>         reply.ia.nifs = htonl(nifs);
> 
> -       reply.ia.uuid[0] = htonl(afs_uuid.time_low);
> -       reply.ia.uuid[1] = htonl(afs_uuid.time_mid);
> -       reply.ia.uuid[2] = htonl(afs_uuid.time_hi_and_version);
> +       reply.ia.uuid[0] = afs_uuid.time_low;
> +       reply.ia.uuid[1] = htonl(ntohl(afs_uuid.time_mid));
> +       reply.ia.uuid[2] = htonl(ntohl(afs_uuid.time_hi_and_version));
> 
>         reply.ia.uuid[3] = htonl((s8) afs_uuid.clock_seq_hi_and_reserved);
>         reply.ia.uuid[4] = htonl((s8) afs_uuid.clock_seq_low);
>         for (loop = 0; loop < 6; loop++)

Shouldn't this be ntohs() instead of ntohl(), like this:

       reply.ia.uuid[1] = htonl(ntohl(afs_uuid.time_mid));
       reply.ia.uuid[2] = htonl(ntohl(afs_uuid.time_hi_and_version));

My head is spinning a little from all the byteswapping, but it
looks to me like the data here ends up in the wrong half of the
on-wire data. Can you double-check this?

	Arnd