>-----Original Message----- >Here's a quick analysis of a typical READ command from the log. I >have >removed some extraneous information and added line numbers and decimal >points to the timestamps (second column) so they will be easier to >read.
Thanks for the quick work. >In other words, the time-consuming parts of this data transfer >consisted >entirely of the computer waiting for the drive to send data. >Transferring >the command and the status took practically no time by comparison. > >This does make it look like the bottleneck is in the drive, not the >computer. However I wouldn't put it past the VIA controller to be >slowing >things down. Those timestamps indicate when the computer learns that >the >transfers have completed, which obviously doesn't happen until the >controller tells it. VIA hardware is known to have a number of funky >problems. My reason for not suspecting the hardware (Given the cheapness of the USB HDD adapter, I would be inclined to blame it) is that I get the same performance with my iRiver H340 which I have seen working faster (With a different kernel on different hardware all of which is in another country, making it hard to make comparisons). One interesting thing that I have found. The following line from lspci -vv for the USB2 device on the card talks about routing interrupt A to IRQ 11. This is the same as is claimed for the two USB1.1 (UHCI) devices that also appear on the chipset. "Interrupt: pin A routed to IRQ 11" However, the diagram at the bottom of: http://www.xbitlabs.com/articles/storage/display/vt6212.html Shows lines A,B and C used for the different chips. I'm just in the process of trying to work out if I can see whether or not the interrupts are making it through, but given that the diagram suggests that the USB2 device is on line C and that the kernel has routed line A three times, this might be a possibility. And I know it's a VT6212L because I took the top off the card. ;-) Jon ------------------------------------------------------------------------- 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-users@lists.sourceforge.net To unsubscribe, use the last form field at: https://lists.sourceforge.net/lists/listinfo/linux-usb-users