> From: David Gwynne <da...@gwynne.id.au>
> Date: Tue, 25 Apr 2017 11:06:52 +1000
> 
> > On 25 Apr 2017, at 6:18 am, Mark Kettenis <mark.kette...@xs4all.nl> wrote:
> > 
> >> Date: Mon, 24 Apr 2017 16:01:39 -0400
> >> From: Dale Rahn <dr...@dalerahn.com>
> >> 
> >> The below code is set up to attach and identify processors for arm64
> >> currently it only identifies model, not cache size, but that can be added
> >> later. It is set up to attach secondary processors later (for when SMP
> >> is present).
> > 
> > I don't think this cpubus(4) device makes much sense.  We don't have
> > something like that on other FDT/OpenFirmware architectures.
> 
> there's this:
> 
> mainbus0 at root: SPARC Enterprise M4000 Server
> cmp0 at mainbus0
> core0 at cmp0
> cpu0 at core0: FJSV,SPARC64-VI (rev 9.3) @ 2150 MHz
> cpu0: physical 128K instruction (64 b/l), 128K data (64 b/l), 5120K external 
> (256 b/l)
> cpu1 at core0: FJSV,SPARC64-VI (rev 9.3) @ 2150 MHz
> cpu1: physical 128K instruction (64 b/l), 128K data (64 b/l), 5120K external 
> (256 b/l)
> core1 at cmp0
> cpu2 at core1: FJSV,SPARC64-VI (rev 9.3) @ 2150 MHz
> cpu2: physical 128K instruction (64 b/l), 128K data (64 b/l), 5120K external 
> (256 b/l)
> cpu3 at core1: FJSV,SPARC64-VI (rev 9.3) @ 2150 MHz
> cpu3: physical 128K instruction (64 b/l), 128K data (64 b/l), 5120K external 
> (256 b/l)
> cmp1 at mainbus0
> core2 at cmp1
> cpu4 at core2: FJSV,SPARC64-VI (rev 9.3) @ 2150 MHz
> cpu4: physical 128K instruction (64 b/l), 128K data (64 b/l), 5120K external 
> (256 b/l)
> cpu5 at core2: FJSV,SPARC64-VI (rev 9.3) @ 2150 MHz
> cpu5: physical 128K instruction (64 b/l), 128K data (64 b/l), 5120K external 
> (256 b/l)
> core3 at cmp1
> cpu6 at core3: FJSV,SPARC64-VI (rev 9.3) @ 2150 MHz
> cpu6: physical 128K instruction (64 b/l), 128K data (64 b/l), 5120K external 
> (256 b/l)
> cpu7 at core3: FJSV,SPARC64-VI (rev 9.3) @ 2150 MHz
> cpu7: physical 128K instruction (64 b/l), 128K data (64 b/l), 5120K external 
> (256 b/l)

That's a bit of a special case though.  For one thing Sun never put
the CPUs under /cpus.  And on these machines the hierarchy expresses
the topology of sockets, cores and threads, where the nodes all have
interesting proprties.

Cheers,

Mark

Reply via email to