On Monday 04 August 2008, Al Johnson wrote: > On Saturday 02 August 2008, Michael Shiloh wrote: > > I'm part of Openmoko and I'm not clear on the fix either. > > > > I'm trying to come up with two side-by-side tests we can run, one with > > the SW fix, one without, that is easy to document and requires the same > > distro, then test w/ two different kernel images (one with, one without > > the SW fix). > > Can anyone see a problem with the methodology in the script below for > testing the effect of the sd_drive and sd_idleclk parameters on TTFF? Am I > doing something stupid like setting the wrong parameters or values?
The stupid mistake was not forcing some transaction on the SD card so that the clock settings get applied. I was seeing a similar spread of ttff for all settings before. After adding a touch and sync after the clock settings I am seeing a clear pattern emerging when there's no SD activity. _______________________________________________ support mailing list [email protected] https://lists.openmoko.org/mailman/listinfo/support
