Re: [PATCH] Time-based RFC 4122 UUID generator

[Helge Deller <deller@gmx.de>]

On Tuesday 20 November 2007, Matt Mackall wrote:
> On Tue, Nov 20, 2007 at 10:59:58PM +0100, Helge Deller wrote:
> > > > Current implemenations use userspace-libraries. In userspace you e.g. can't 
> > > > easily protect the uniquness of a UUID against other running _processes_.
> > > > If you try do, you'll need to do locking e.g. with shared memory, which can 
> > > > get very expensive.
> > > 
> > > Even with a futex? Or userspace atomics? 
> > 
> > Yes, you'll need a futex or similiar.
> > The problem is then more, where will you put that futex to be able to protect against other processes ? 
> > Best solution is probably shared memory, but then the question will be, who is allowed to access this memory/futex ?
> > Will any process (shared library) be allowed to read/write/delete it ?
> > At this stage you then suddenly run from a locking-problem into a security problem, which is probably equally hard to solve.
> > Btw, this is how Novell tried to solve the time-based UUID generator problem in SLES and it's still not 100% fixed.
> > 
> > > I think something as simple 
> > > as a server stuffing a bunch of clock sequence numbers into a pipe
> > > for clients to pop into their generated UUIDs should be plenty fast
> > > enough.
> > 
> > Sounds simple and is probably fast enough.
> > But do you really want to add then another daemon to the Linux system, just in case "some" application needs somewhen a UUID ?
> 
> This really is the crux of the problem. I really don't want to add 1K
> of unpageable memory to every kernel in the world for a feature that
> can be implemented in userspace, just in case "some" application needs
> a UUID.

Again, it could be made a config option in which case you could disable 
it if you don't want it.

> 
> > True, but let's look at the facts.
> > 
> > Current libuuid.so (from e2fsprogs) library on Fedora 7 (i386):
> >    text    data     bss     dec     hex filename
> >    8101     368      40    8509    213d /lib/libuuid.so.1
> > 
> > And the kernel implementation:
> >    text    data     bss     dec     hex filename
> >    4877     604    2080    7561    1d89 drivers/char/random.o.without_uuid
> >    5976     752    2080    8808    2268 drivers/char/random.o.withuuid
> 
> I don't think that's a very good comparison. Here's a trivial (but untested)
> implementation of RFC 4122 (variant 4) that's collision-safe and very tiny:
> 
> /* RFC4122-compliant UUID containing 128 - 4 - 2 - 1 = 121 bits of entropy */
> void genrfc4122(char *buf)
> {
>         int f;
>         f = open("/dev/urandom", O_RDONLY);
>         read(f, buf, 16); /* fill our buffer */
> 	close(f);
>         /* sec4.4: set clock_seq_hi_and_reserved bits 6 and 7 to 0 and 1 */
>         buf[8] = (buf[8] & ~0x3f) | 0x80;
>         /* sec4.4: and high nibble of time_hi_and_version to 4 = "random" */
>         buf[6] = (buf[6] & 0xf) | 0x40;
>         /* sec4.5: set multicast bit to indicate random node (lsb of node[0])*/
>         buf[10] |= 1;
> }
> 
> $ size rfc4122.o
>    text    data     bss     dec     hex filename
>      95       0       0      95      5f rfc4122.o

Nice example, but it's not comparable since it's not what this thread is about.
As you mentioned, you showed here a variant 4 (fully random) version.
You could have shown this even more easily, without any additional code:

[deller@halden linux-2.6]$ cat /proc/sys/kernel/random/uuid
607e598a-b0f2-4d60-9ca7-22838d2120ba

This is already in the kernel and it's allocating some non-swapable memory already.

My patch is about variant 1 (time-based), which is not that easy to make unique!
 
> Modern kernels guarantee that simultaneous readers don't see the same
> pool state, so collisions should be exceedingly rare. While collisions
> are still possible here, frankly I think they are much less likely
> than with schemes that involve persistent state, hardware ids, or
> time. The odds of the persistent state or hardware ids being
> mismanaged or the clock being off are quite terrestrial rather than
> astronomical.

That's only relevant to variant 4, not 1.

Helge