On Thu, Jun 11, 2009 at 11:16:31AM +0800, Alek Du wrote: > On Thu, 11 Jun 2009 10:52:53 +0800 > Greg KH <greg@xxxxxxxxx> wrote: > > > On Wed, Jun 10, 2009 at 01:31:13PM +0800, Alek Du wrote: > > > >From b076459721690a5a03bc6490d0a98237ada74628 Mon Sep 17 00:00:00 2001 > > > From: Alek Du <alek.du@xxxxxxxxx> > > > Date: Wed, 10 Jun 2009 12:29:57 +0800 > > > Subject: [PATCH] EHCI: Add module parameter to specify EHCI_IO_JIFFIES > > > > > > The default EHCI_IO_JIFFIES is just waking up system too much (~10 times per > > > second) and it won't benefit those no interrupt quirk EHCI controllers. This > > > patch adds an io_jiffies parameter to ehci-hcd, user either could specify the > > > parameter during modprobe time or provide ehci-hcd.io_jiffies=xxxx in kernel > > > cmdline for ehci-hcd built-in case. > > > > > > Signed-off-by: Alek Du <alek.du@xxxxxxxxx> > > > > What happens if you have two different EHCI controllers, one needing > > this and one that doesn't? Shouldn't this be a 'per device' type > > attribute? > > > > And I agree with Alan, is there anyway we can dynamically determine > > this? That would make it much better, adding new config options just > > increases the complexity that a system integrator needs to handle. > > > > hmmm... > > Then how about adding a need_io_watchdog flag to ehci_hcd structure to > identify if someone need io watchdog? Those quirk drivers need to fill > the item and the default good device drivers just do nothing. That sounds good to me as long as you determine the drivers that need the flag at the same time you introduce it so that you don't break anything :) How about adding it to all drivers, and then we can turn it off on drivers that don't need it after it has been tested? thanks, greg k-h -- To unsubscribe from this list: send the line "unsubscribe linux-usb" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
