From: Robert C. <rc...@ae...> - 2011-11-10 13:41:50
|
I've had to modify some kernel reboot parameters before. The .h file you're referencing is changed by a patch file, so changing it won't probably work. I suggest taking a look at one of my postings about modifying the kernel command line. You might see something that'll help http://old.nabble.com/FYI---steps-to-modify-kernel-boot-command-line-to32625614.html#a32625614 -cy On Wed, Nov 9, 2011 at 8:16 PM, chr...@gm... <chr...@gm...>wrote: > > A bit of followup: > > 1) I've tried looking at the u-boot.bin generated in the u-boot tmp > directory; it does not contain the string "defaultdisplay=lcd43". It does, > however, have long strings of nandarguments and mmcarguments referring to > ${defaultdisplay}, which is earlier declared to be equal to "dvi". > Therefore > the problem is in compilation somewhere. > > 2) Modifying the u-boot/tmp directory's copy of u-boot.bin to have the > desired string value of "defaultdisplay=lcd43" makes a boot with that > u-boot.bin binary fail, as might be reasonably expected. Whups. > 2b) However, copying that binary onto my uSD card and then checking for > presence of string "defaultdisplay=lc343" in the uSD card's copy of > u-boot.bin confirms that at least the file copying operation is working as > expected. > > My current source of mystification, then, is why > $bitbake -c clean u-boot > $bitbake -c clean virtual/bootloader > $bitbake virtual/bootloader > > does not create a binary image with the string as defined by omap3_overo.h > . > My best speculation on a cause for this is that an intermediate product of > the compilation process is not getting removed by the -c clean directive > and > is acting as a cache for the string's old value. My next question, then - > if > -c clean isn't sufficient to rebuild u-boot and all its dependencies from > scratch, what is? My not-best speculation on the cause is that other > options > set in omap3_overo.h, like dvimode, are getting caught by a preprocessing > script which then undoes the change I made to defaultdisplay to preserve > sanity. > > Can anyone provide any feedback for these theories, or a fix for the issue > I'm seeing? > thanks in advance, > Chris > > > chr...@gm... wrote: > > > > Hi, all. > > > > Having confirmed that my build environment appears to be operating > > correctly, I would now like to configure my Tide+Chestnut43 to boot using > > the touchscreen without my having to intervene each time during the boot > > process. http://www.gumstix.org/getting-started.html tells me to modify > > omap3_overo.h to do so, but does not provide specifics; I assumed that > > changing 'defaultdisplay=dvi' to 'defaultdisplay=lcd43' would be > > sufficient, but when I check the output of printenv after building with > > the modified header, defaultdisplay is still 'dvi'. The system uses the > > display happily when I manually set the variable via u-boot, so I've got > > the right system image... but why are my changes not being reflected in > > system behavior? Where are they being overridden? > > > > Thanks in advance, > > Chris > > > > -- > View this message in context: > http://old.nabble.com/Modifying-omap3_overo.h-does-not-modify-output-of-overo-printenv---why--tp32788700p32815682.html > Sent from the Gumstix mailing list archive at Nabble.com. > > > > ------------------------------------------------------------------------------ > RSA(R) Conference 2012 > Save $700 by Nov 18 > Register now > http://p.sf.net/sfu/rsa-sfdev2dev1 > _______________________________________________ > gumstix-users mailing list > gum...@li... > https://lists.sourceforge.net/lists/listinfo/gumstix-users > -- Robert Corley |