Carrier Grade Linux/Gaps Alpha1

[greg@xxxxxxxxx (Greg KH)]

On Tue, May 27, 2008 at 06:43:29PM -0400, Theodore Tso wrote:
> On Tue, May 27, 2008 at 10:23:56AM -0700, Greg KH wrote:
> > See previous message (1% speed increase was the best possible
> > measurement that I got, but the variance in booting was about 5%
> > overall, so it was lost in the noise pretty much all the time.)
> 
> It should be noted that on a typical system there's typically only one
> probe which takes a loooooong time and that's the SCSI/FC bus
> enumeration.  So even if you parallelize everything else, it's not at
> all surprising that the reduction would only be 2 seconds.

You do know we can do parallel scsi/fc probes now right?  That's been in
the kernel for a while now and does provide a huge decrease in boot time
on larger boxes.

And yes, other than disks, PCI probe time is pretty much instant (USB
devices are discovered asynchronously, the boot process does not wait
for them), so doing multi-threaded driver probing saves no real time at
all (and slowed down single processor machines).

thanks,

greg k-h