John Hay wrote:
> > As of a few minutes ago, a minimal set of changes to bring the so-called
> > 'new-bus' functionality to the i386 kernel in -current.
> >
>
> It looks like the stat clock isn't started after this. I have tried a SMP
> and UNI kernel and both behave the same. Looking with vmstat -i and
> systat -vmstat, there is no sign of the stat clock. Sysctl kern.clockrate
> also have profhz and stathz as 100. A kernel of about a week ago doesn't
> have this problem.
Hmm...
[7:38pm]~-103# sysctl kern.clockrate
kern.clockrate: { hz = 100, tick = 10000, tickadj = 5, profhz = 1024,
stathz = 128 }
And from systat:
zfod Interrupts
Proc:r p d s w Csw Trp Sys Int Sof Flt cow 230 total
2 2 23 59 2 141 230 30 6 16016 wire pci irq11
15636 act 2 pci irq5
1.1%Sys 1.1%Intr 0.8%User 97.0%Nice 0.0%Idl 24480 inact pci irq6
| | | | | | | | | | 3736 cache pci irq4
=>---------------------------------------------- 1400 free pci irq7
daefr 100 clk0 irq2
Namei Name-cache Dir-cache prcfr 128 rtc0 irq8
The labelling of the source column is known to be broken at the moment, there
are a number of key places where the fake isa unit number is getting passed
down to the lower level code. This will be fixed up shortly.
Cheers,
-Peter
To Unsubscribe: send mail to [email protected]
with "unsubscribe freebsd-current" in the body of the message
