On Wed, 2009-10-21 at 09:49 +0200, Christof Schmitt wrote: > >> Added timestamp comparison checks are straightforward without any > >> additional conversion on each check since added timestamps and > >> queue_ramp_up_period both are jiffies, so works better this way > with > >> stored queue_ramp_up_period in jiffies. > >> > >> I do see your point as well but jiffies unit is also common in > Linux. > >> However if you or some other feels strongly to change this to ms or > >> seconds unit then I could change store or show sysfs functions to > handle > >> this value in ms or second unit while storing it in jiffies in > >> queue_ramp_up_period. > >> > >> Vasu > > > > I agree with Christof on this, HZ may be 100 on some machines and > 1000 > > on others. Management tools (if they ever adjust this) would like > > standard units, so it should be in ms or seconds, and I prefer ms. > > I would just change the sysfs functions as you say, and use jiffies > internally. > > I would also vote for ms in sysfs and using jiffies internally. This > should be fairly easy using the functions jiffies_to_msecs and > msecs_to_jiffies. Yeap it will be easy change and I'll do this change to have ms unit for only sysfs i/f as mostly preferred ms as unit, though Mike Reed said in other response that ms unit is too fine grained but that should be okay as oppose to not giving choice of fine granularity and I also don't see any risk to large value limitation by use of fine ms unit here. I wasn't against the Christof's suggestion to use ms or sec in first place since as I said "I do see your point as well". And now all these responses gives solid confirmation to use ms unit here. Thanks all for your inputs. Vasu -- To unsubscribe from this list: send the line "unsubscribe linux-scsi" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html