Hey, Tim

Tim Kelly wrote:
> 
> At 5:47 PM +0100 12/7/04, Riccardo Mottola wrote:
> >The bad news, is that my own kernel exhibits the same "hang after scsi
> >devise settle" and then disk activity progresses only if I repeat
> >pressing keys on the keyboard.
> 
> What is the code in arch/powerpc/powerpc/locore_subr.c -> Idle (it's the
> second or third function)? Are you running mesh sync or non-sync?
O have only an assembler file there... .S

there If ound this
...
ASENTRY(Idle)
        lis     %r8,_C_LABEL(sched_whichqs)@ha
        ldint   %r9,_C_LABEL(sched_whichqs)@l(%r8)

        or.     %r9,%r9,%r9
        bne+    .Lsw1                   /* at least one queue non-empty
*/

#if defined(MULTIPROCESSOR) || defined(LOCKDEBUG)
        bl      _C_LABEL(sched_unlock_idle)
#endif

....

I don't know which mesh I am running. I should use sync with Michael's
patches since otherwise my disk causes problems.
Howeverm those problems are not like those I am experiencing now.
 
> >Could this be caused by Tim's CPU detect patch? were lately other IRQ
> >affecting patches commited?
> 
> The CPU detection patch does not affect any IRQs, and once it's reported
> the CPU speed has no further interaction with the boot process. :-)
ok. I hope so :) Maybe you set something that is important for later
use.
In any case the cpu is correctly detected.

-R