On Fri, Mar 16, 2012 at 09:51:51AM -0700, Greg KH wrote: > On Fri, Mar 16, 2012 at 05:37:42PM +0100, zouid abdelhamid wrote: > > > > Hi greg, > > I was trying to add this patch http://permalink.gmane.org/gmane.linux.usb.general/17150 to my 2.6.32.16 , that was posted by sarah sharp, in fact I am trying to make the same experience that did sarah in 2008 demo when the USB 3.0 was newly introtuced with Linux, > > the CONFIG_USB_HCD_STAT flag was removed from all the new releases. Yes, because the statistics code doesn't actually work properly, as the replies from Alan Stern point out. Don't use that code to measure the xHCI driver performance. Instead, use dd to measure a USB 3.0 disk performance when you're reading or writing directly to the block device (like /dev/sdb). If you don't set up a filesystem on the disk, then you avoid measuring any overhead included in that filesystem. You'll need to run this command to drop any caches before each performance test, so you're measuring the xHCI host controller performance, not the disk cache performance: # echo 3 > /proc/sys/vm/drop_caches dd will give you some performance numbers. I would also suggest running iostat and mpstat to get a report of the CPU and memory utilization when you're running dd. I'm not going to tell you how to use those tools. I know this is for a college project, because you sent me private mail, and you need to figure that out on your own. Sarah Sharp -- 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
