Forums

First impressions with new RidgeRun SDK and Leopardboard 365

  • Hhhmmm I'm getting the following error (during boot) when I want to save to NAND (similar to what someone posted here: http://pastebin.com/eLWFGAvs )

    Root-NFS: No NFS server available, giving up.                                  

    VFS: Unable to mount root fs via NFS, trying floppy.                            

    List of all partitions:                                                        

    No filesystem could mount root, tried:  jffs2                                  

    Kernel panic - not syncing: VFS: Unable to mount root fs on unknown-block(2,0)  

    Backtrace:                                                                      

    [<c0031700>] (dump_backtrace+0x0/0x110) from [<c0031844>] (dump_stack+0x18/0x1c)

    r6:c0026850 r5:c048a464 r4:c2029f58

  • It would be helpful if you provided all the boot output.

    It looks like your uboot bootargs are set up for NFS root file system, not NAND root file system.

    cd $DEVDIR

    make config

    in the file system menu select NAND for the target hardware and JFFS2 for the file system.

    make

    make install

    Todd

  • Sorry to come back with this problem but I don't see any mention of NAND in the make config menu.

    Here is what I have in the "File System Configuration menu":

      *** File system target image configuration  --- ***
      │ │        File system image target (JFFS2 disk image)  --->
      │ │    (64KiB) Erase block size of JFFS2 (NEW)

      │ │   [*] Do not include the cleanmaker block (say yes if you are using
      │ │    [ ] Change JFFS2 endianness (NEW) 
      │ │    [ ] Strip the target file system binaries (may render debug usele
      │ │        *** File system contents --- ***         
      │ │        Select target's file system software  --->
      │ │        *** Toolchain file system components ***    
      │ │    -*- Install C++ (libstdc++) support on target file system  
      │ │        *** Toolchain provided applications ***  
      │ │    [ ] Install GDB server on target file system   
      │ │    [ ] Install ldd on target file system

     

    I have also gone through the Architecture configurations,  Installer Configuration, Bootloader Configuration,  File System Configuration but no trace of a hardware file system menu (with NAND or NFS) :-(

  • This is the patch required by 'make' after enabling JFFS2 as a filesystem via 'make config':

    =============================== BEGIN PATCH  ==================================

    Index: /home/vleo/00_HW/LeopardBoard/RRSDK.latest/trunk/bsp/oscheck/Fedora.common

    ===================================================================

    --- /home/vleo/00_HW/LeopardBoard/RRSDK.latest/trunk/bsp/oscheck/Fedora.common (revision 56)

    +++ /home/vleo/00_HW/LeopardBoard/RRSDK.latest/trunk/bsp/oscheck/Fedora.common (working copy)

    @@ -43,6 +43,6 @@

    curl:curl

    # Not available on fedora?

    -#mtd-tools:mtd-tools

    +mtd-tools:mtd-utils

    #cramfsswap:cramfsswap

    #genext2fs:genext2fs

    =============================== END PATCH ==================================

    after that 'make' did not go through though since there were some files owned by root on the $DEVDIR/fsfs tree,

    owned by root and not having world read/write permissions:

    sudo find -uid 0 -exec chmod go+rw {} \;

    after that 'make' completed w/o errors.

    After that trying to install the fs image over serial line (/dev/ttyS0) results in a failure. Maybe we need to enable another load method (over the ethernet)

  • Hi Alex,

    Did you use the bootargs listed here: www.ridgerun.com/.../DM365_Leopard_Pipelines when you booted the kernel for H.264 network streaming?

    When I follow the steps you listed, ipipe_init -d 1 (or just -d) generates the following error:

    davinci_previewer davinci_previewer.2: Operation mode doesn't match with current hw mode

    Can't get operation mode

    : Invalid argument

    Error in initializing previewer

    Has anyone else encountered this problem?

  • We find that (without changing the V4L2_CID_VFLIP default and with flipping disabled in the MT9P031 register setup) the image from a LI-5M03 5MP camera (MT9P031) is right-way-up, and the image from a LI-LBCM5M1 5MP camera (MT9P001) is upside-down.  No doubt this is to do with the lens that comes with the LI-5M03.  These two sensors are software-compatible, the MT9P031 is larger though (and possibly has extra settings that the MT9P001 doesn't have).  Not sure about setting the flipping in the kernel by default.  I mean, who's to say you're not standing on your head ;)