SUCCESS! ..and SuSE (was Re: Fedora Core 3 USB testing)

  • Previous message: Fedora Core 3 USB testing (100% failures so far)
  • Next message: SUCCESS! ..and SuSE (was Re: Fedora Core 3 USB testing)
  • David A. Desrosiers hacker at gnu-designs.com
    Sun Nov 21 21:57:08 EST 2004

     

    > 	I'm soliciting for testers to give me feedback about their 
    > experiences with Palm synchronization under Fedora Core 3 (FC3) and 
    > their usb-connected Palm device.
    
     	I'm replying to my own message here:
    
     	I spent a few more hours and tested the 2.6.9-1.667, 668, and 
    now 681 stock FC3 kernels, all with failures of the same type.. kernel 
    crashes in the uhci_hcd layer of the kernel. With the help of some 
    people on irc on #fedora and #fedora-devel, I was able to piece 
    together some clues.
    
     	I downloaded pristine 2.6.9 kernel source from kernel.org, and 
    put in the .681 .config file and built myself a clean 2.6.9 kernel, 
    free of any FC patches or fixes.
    
     	After rebooting, EVERYTHING worked perfectly, including sync 
    with every tool that was previously failing with the 3 prior "stock" 
    kernels.
    
     	At this point, none of us (including #fedora-devel) know what 
    is causing the crashes with the three prior kernel revisions, but I 
    now have something I can use to test and debug udev issues with.
    
     	The other tricky thing I had to do was to apply some changes 
    to the udev-specific portions of FC3, namely permissions and some 
    symlinking of the default "convenience" device, /dev/pilot.
    
     	To do that, follow these steps:
    
     	1.) Create /etc/udev/rules.d/10-visor.rules and put this in
     	    it:
    
     		KERNEL="ttyUSB1" SYMLINK="pilot"
    
     	2.) Create /etc/udev/permissions.d/10-visor.permissions and
     	    put this in it:
    
     		ttyUSB1:$local:uucp:0660
    
     	Thats it. Once I did that, everything hooked up and started 
    working nicely on FC3. There are some underlying issues with the way 
    gnome-pilot's conduits work, but I'm sure I'll figure those out in 
    time too.
    
     	SuSE is another nest of bees I have to figure out. I spent a 
    good portion of this weekend debugging SuSE issues with an i500 that 
    Jeremy Wise sent me (his personal unit) to get it working for him, 
    after we had exhausted all possibilities of making it work "remotely".
    
     	After some tussle with the way SuSE handles USB, I managed to 
    get that working as well. He wanted to sync to Evolution, but was 
    unable to. With the stock SuSE 9.1 Professional build, which ships 
    with gnome-pilot 2.0.10, I was unable to sync, because the daemon 
    would consistently crash. Compiling gnome-pilot 2.0.12 from source 
    onto the machine worked like a champ, and I was able to sync right 
    away.
    
     	Once he gets the unit back, we'll try to apply the techniques 
    I used locally to his own personal machine, remotely, so he can sync 
    his (not-so-inexpensive) PDA/Phone to his Linux machine.
    
     	What a weekend. Whew. Now back to real paying work. ;)
    
    
    David A. Desrosiers
    desrod at gnu-designs.com
    http://gnu-designs.com
    
    
    

     

     

    More information about the pilot-link-general mailing list