Re: Advanced USB snooping

From: Bernd Walter (ticso_at_cicely12.cicely.de)
Date: 02/22/05

  • Next message: John Baldwin: "Re: Kernel monitor, the return"
    Date: Tue, 22 Feb 2005 22:53:56 +0100
    To: Peter B <pb@ludd.ltu.se>
    
    

    On Tue, Feb 22, 2005 at 09:00:58PM +0100, Peter B wrote:
    > >>
    > >> Is it possible to program an "ordinary" (like Intel 82801CA/CAM (ICH3)) USB
    > >> host controller into acting as an usb device instead? (just like scsi can).
    > >>
    > >> Idea:
    > >> M$Win-Usb -> FreeBSD-USB#1 .. software .. FreeBSD-USB#2 -> Device
    > >
    > >No - this wonn't work.
    > >Host Controllers are dedicated host controllers.
    >
    > >If you use a microcontroller you can use normale device controllers,
    > >such as the PDIUSBD12 or the ISP1581 - there are many others as well.
    > >But the they are all limited to work as a single device and not as a
    > >path through device.
    >
    > If I can connect it to the computer from the "other" side it solves the problem
    > I think.

    But it does behave a a sinlge device only - and even that not in
    every aspect.
    They are limited to the number of channels they can do and so on.

    > >All USB sniffers I know run with some kind of ASIC, which is not very
    > >hard to do if you are familar with such devices.
    >
    > Fpga is doable, asic requires a large factory run.
    > Unless there is a ready to buy asic.

    FPGA are ASIC - just a special kind of.
    Take an ISP1106 and an 48MHz oscillator with an 48MHz capable FPGA
    and you should get it runninng.
    48MHz is the usuall clock used for oversampling the 12MHz line signal.
    The ISP1106 does the electrical part of the bus for you.

    > >> The advantage would be then to possible use scripts to debug protocol
    > >> in order to port drivers to freebsd.
    >
    > >Sniffer Software exist at driver level at least for Windows and NetBSD
    > >derived stacks and I would be surprised if there isn't anything
    > >available for Linux as well.
    > >Hardware sniffers are very usefull if you are into debugging host
    > >controllers and sometimes device controllers on your own, but for
    > >debugging device drivers it's rarely usefull.
    >
    > The catch with normal setup is that you need to poke around in the m$-win
    > machine which can cause other problems. So the idea would be to just insert
    > some device with one male + female connector and a connection to the snoop
    > computer. So that the m$-win computer don't need any special software.
    > Also m$-win host software snoop lack the nice programming enviroment of unix.
    > In order to filter out the stuff that one really wants.

    Public available Windows software can output ascii.

    -- 
    B.Walter                   BWCT                http://www.bwct.de
    bernd@bwct.de                                  info@bwct.de
    _______________________________________________
    freebsd-hackers@freebsd.org mailing list
    http://lists.freebsd.org/mailman/listinfo/freebsd-hackers
    To unsubscribe, send any mail to "freebsd-hackers-unsubscribe@freebsd.org"
    

  • Next message: John Baldwin: "Re: Kernel monitor, the return"

    Relevant Pages

    • Re: USB mass storage - more trouble
      ... I had to keep lockcdr.flt or RSJ couldn't mount a ... check that RSJ still worked - it did; ... USB package which tells you what kind of USB controllers you have and ... each one, thus if you have two open host controllers, then you need ...
      (comp.os.os2.setup.misc)
    • Re: SATA Drive firmware
      ... Prior to the introduction of SATA, the vast majority of ATA host controllers ...
      (comp.periphs.scsi)
    • Re: DDR/DDR2 controller - core
      ... I have chosen DDR because is avaiable and cheaper than SDR. ... I have no experience with memory controllers and not to big with FPGA, ... When I compiled a DDR2 controler from MegaCore it take 3000LE. ...
      (comp.arch.fpga)
    • DDR/DDR2 controller - core
      ... I have chosen DDR because is avaiable and cheaper than SDR. ... I have no experience with memory controllers and not to big with FPGA, ... When I compiled a DDR2 controler from MegaCore it take 3000LE. ...
      (comp.arch.fpga)
    • Re: DDR/DDR2 controller - core
      ... I have chosen DDR because is avaiable and cheaper than SDR. ... I have no experience with memory controllers and not to big with FPGA, ... that's why I don't know which design solution to choose. ...
      (comp.arch.fpga)