On Fri, 1 Dec 2006, Greg KH wrote: > From: Greg Kroah-Hartman <[EMAIL PROTECTED]> > > Might speed up some systems. If nothing else, a bad driver should not > take the whole USB subsystem down with it. > > Signed-off-by: Greg Kroah-Hartman <[EMAIL PROTECTED]> > --- > drivers/usb/core/Kconfig | 15 ++++++++ > drivers/usb/core/hub.c | 81 +++++++++++++++++++++++++++++++++------------ > 2 files changed, 74 insertions(+), 22 deletions(-) > > diff --git a/drivers/usb/core/Kconfig b/drivers/usb/core/Kconfig > index 6e3b535..f8324d8 100644 > --- a/drivers/usb/core/Kconfig > +++ b/drivers/usb/core/Kconfig > @@ -72,6 +72,21 @@ config USB_SUSPEND > > If you are unsure about this, say N here. > > +config USB_MULTITHREAD_PROBE > + bool "USB Multi-threaded probe (EXPERIMENTAL)" > + depends on USB && EXPERIMENTAL > + default n > + help > + Say Y here if you want the USB core to spawn a new thread for > + every USB device that is probed. This can cause a small speedup > + in boot times on systems with a lot of different USB devices. > + > + This option should be safe to enable, but if any odd probing > + problems are found, please disable it, or dynamically turn it > + off in the /sys/module/usbcore/parameters/multithread_probe > + file > + > + When in doubt, say N.
You sent your original patch off to Linus instead of the variant I sent in, even though your patch violates the locking rules (see the kerneldoc for usb_new_device)! By the way, have you observed any significant improvements in peformance from this patch? I don't have any experience with systems having lots of USB devices attached. For that matter, does speeding up khubd make any significant change to the overall boot time for a system? One wouldn't think the other boot-up steps would have to wait for khubd... This patch will help when many devices are plugged into the same hub (or same root hub), but it won't probe multiple hubs in parallel. It's not clear which approach would be better in general, or whether some combination of the two would be best. It's not even clear whether the overhead from spawning all these new threads is worth the savings from doing I/O in parallel. Also, I would expect the largest effects to show up at boot time. At other times people tend to turn USB devices on or off (or plug and unplug them) one at a time, not many all at once. But there still is all the overhead from creating new threads to handle these events. Alan Stern ------------------------------------------------------------------------- Take Surveys. Earn Cash. Influence the Future of IT Join SourceForge.net's Techsay panel and you'll get the chance to share your opinions on IT & business topics through brief surveys - and earn cash http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV _______________________________________________ linux-usb-devel@lists.sourceforge.net To unsubscribe, use the last form field at: https://lists.sourceforge.net/lists/listinfo/linux-usb-devel
