Interesting... I find that the machine ID is checked in 2 steps in kernel/setup.c function setup_arch()
First it tries to use the device tree by calling setup_machine_fdt() in devtree.c But if this function return a NULL pointer, it will call setup_machine_tags() in kernel/atags_parse.c. This is the case with a large initramfs is used... It should not be different... that means maybe that there is something wrong with kernel loading that overwrites the device tree... Le mardi 1 octobre 2013 23:06:23 UTC+2, Slyde a écrit : > > It seems that my issue is due to the initramfs size. > > But I checked the address used to load the kernel and device tree and > their sizes and it should not overlap. > > What is strange is that the bad machine code I have (0x00000e05) is the > same as other post that I found on the web. But they were related to old > u-boot... > > I tried to change my optargs to very minimal, only console=/dev/ttyO0 but > same issue. > > Now I will put some debug messages in the code to try to compare the > values between small and large initramfs > > > Le mardi 1 octobre 2013 21:50:46 UTC+2, RobertCNelson a écrit : >> >> On Tue, Oct 1, 2013 at 2:49 PM, Slyde <[email protected]> wrote: >> > Hi rh_ , >> > >> > Thanks I will try to look around the optargs and atags parsing... >> > >> > I don't want to but I need this to work :-( >> >> "too large initramfs" your not by chance over writing the memory used >> for the device tree binary? >> >> Regards, >> >> -- >> Robert Nelson >> http://www.rcn-ee.com/ >> > -- For more options, visit http://beagleboard.org/discuss --- You received this message because you are subscribed to the Google Groups "BeagleBoard" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. For more options, visit https://groups.google.com/groups/opt_out.
