Re: dentries rule. [was: Re: Pros and cons of newer kernels]

Steven N. Hirsch (
Thu, 20 Nov 1997 08:36:13 -0500 (EST)

On Wed, 19 Nov 1997, Mike Jagdis wrote:

> On Tue, 18 Nov 1997, Steven N. Hirsch wrote:
> > 2.0.31: 0.72user 5.11system 0:06.19elapsed 94%CPU (486/133)
> >
> > 2.1.64: 0.32user 0.57system 0:17.04elapsed 5%CPU (6x86 P150+)
> >
> >
> > Just a bit of difference in the CPU usage, no? Any suggestions as to why
> > the 6x86 shows so little CPU use? Top shows the system as being 97%
> > idle! Could this be some odd misbehavior with the Cyrix idle patch?
> It looks like the 6x86 machine spent around 10 seconds waiting
> for I/O. Almost as if your hard disk interrupts aren't bringing
> the processor out of hlt (but others must be). What do you get
> without suspend on halt enabled or booting with the no-hlt boot
> option?

Mike, et al,

I rebuilt the kernel w/o suspend-on-halt, and there was no change in the
behavior. The odd part about this is that the machine functions normally
in all other respects. Not that this is a scientific benchmark, but a
full kernel build of 2.1..65 runs a bit under 10-minutes - certainly not
indicative of long dead periods.

I'll re-test again under 2.0.32 this evening, though I believe I've
already tried this without seeing the problem..