Featured Articles
Announcing Pi
Technical Specifications
Raspberry Pi FAQs
Win a Pi
Raspberry Pi Wishlist
Actions
Forum Thread Details
  • Replies 2 replies
  • Subscribers 664 subscribers
  • Views 506 views
  • Users 0 members are here
Related

FTDI USB-to-serial issue with NAK packets

Former Member

Drew has a blog post about the FTDI USB-to-serial converter:

http://www.element14.com/community/groups/raspberry-pi/blog/2012/07/18/look-ma-no-display-using-the-raspberry-pi-serial-console

 

Apparently there is an issue with excessive interrups from NAK packets:

 

http://www.raspberrypi.org/phpBB3/viewtopic.php?f=28&t=7866&start=111

http://www.raspberrypi.org/phpBB3/viewtopic.php?f=28&t=7866&start=114

  • Interesting thread.  I've not had any trouble using FTDI cable for console access on the Pi, but it is acting as a USB device connected to my laptop.  I think the person reporting the problem must have been using FTDI cable connected to a Pi USB host port.

     

    I found the commands in the thread interesting:

    pi@raspberrypi ~ $ cat /boot/cmdline.txt

    dwc_otg.lpm_enable=0 console=ttyAMA0,115200 kgdboc=ttyAMA0,115200 console=tty1 root=/dev/mmcblk0p2 rootfstype=ext4 elevator=deadline rootwait

    pi@raspberrypi ~ $ cat /proc/interrupts

               CPU0      

      3:     842516   ARMCTRL  BCM2708 Timer Tick

    52:          0   ARMCTRL  BCM2708 GPIO catchall handler

    65:          2   ARMCTRL  ARM Mailbox IRQ

    66:          1   ARMCTRL  VCHIQ doorbell

    75:   65568338   ARMCTRL  dwc_otg, dwc_otg_hcd:usb1

    77:       8312   ARMCTRL  bcm2708_sdhci (dma)

    79:        338   ARMCTRL  bcm2708_i2c.0, bcm2708_i2c.1

    80:          0   ARMCTRL  bcm2708_spi.0

    83:      47024   ARMCTRL  uart-pl011

    84:      18082   ARMCTRL  mmc0

    Err:          0

    pi@raspberrypi ~ $  sh -c "cat /proc/interrupts | grep dwc && sleep 10 && cat /proc/interrupts | grep dwc" | awk '{ if (s==0) s=$2 } END { print ($2-s)/10 }'

    8035.9

    So this would seem to confirm that USB controller causes 8,000 interrupts a sec.

  • in reply to fustini

    I am unsure how old previous report is, but the cmdline.txt isn't having the latest patches enabled that reduce the interrupt rate.

    The patch isn't really reducing the actual interrupt rate, it's just handling some interrupts that have no interesting usb data in a short assembly routine that just acknowledges the interrupt if I understand things well.

    So, if you have a device like an ftdi connected, interrupt range will go up again, but probably only when you start to use the device.