On Thu, Sep 26, 2013 at 2:13 AM, Greg KH <gregkh@xxxxxxxxxxxxxxxxxxx> wrote: > On Tue, Sep 24, 2013 at 09:12:39PM +0200, Markus Rechberger wrote: >> This patch adds memory mapping support to USBFS for isochronous and bulk >> data transfers, it allows to pre-allocate usb transfer buffers. > > Does libusb support this? > not yet, we're directly accessing usbfs with our userspace usb driver. It should be easy for libusb devs to pick up that extension. The devices are USB video devices. http://sundtek.de/support/toggo2.mpg (captured from our usb device) http://sundtek.de/images/mc_st_640.jpg >> The CPU usage decreases 1-2% on my 1.3ghz U7300 notebook when >> transferring 20mbyte/sec, it should be more interesting to see those >> statistics on embedded systems where copying data is more expensive. > > Any chance you could test this on a system with a "smaller" processor to > see if that really is the case or not? > I'm also curious about that, unfortunately I don't have any time for that at the moment. We support some embedded systems where copying data is very expensive, that's what that patch has been written for. If this patch would enter the kernel now it would probably take one year or longer to get picked up by those systems. http://sundtek.de/support/devio_wo_mmap.mts http://sundtek.de/support/devio_mmap.mts both tests are using the same kernel, the wo_mmap file is using the legacy mode, it sticks between 8-10% (even 12%) CPU usage, while the other one sticks between 6-8% (max 10%). At the time writing this email the files are still uploading, so you might wait another hour before downloading them. Markus > 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
