+U-Boot again Hi Andre,
On 30 March 2015 at 22:41, Andre Wolokita <andre.wolok...@analog.com> wrote: > > > On 31/03/15 13:42, Stephen Warren wrote: >> On 03/30/2015 05:42 PM, Andre Wolokita wrote: >> ... >>> I am able to boot a uImage produced by buildroot successfully, although I >>> still haven't tried a zImage. One point of clarification: the uImage (and >>> presumably zImage) produced by buildroot has the initrd linked in to the >>> image. Could you think of any reason that this would affect a fitImage from >>> loading? >> Presumably this combined kernel+initrd is very large, so the amount of >> space taken by the decompressed kernel (roughly located at the start of >> RAM) is much larger than typical. Perhaps the decompression is >> over-writing the DTB or something like that? Is the entire kernel_initrd >> located within the first 256MB of RAM, if you're uses AUTO_ZRELADDR in >> the kernel config? > Well, I figured it out. It turns out that using objcopy to create a binary of > vmlinux and using the vmlinux.bin as the "kernel" as well as creating a cpio > of the rootfs for the "ramdisk" allows Linux to boot perfectly. > > Thanks for your help, Simon and Stephen. :) > OK good! I wonder if the Linux build process still creates a file called 'Image' which does the objcopy for you? Regards, Simon _______________________________________________ U-Boot mailing list U-Boot@lists.denx.de http://lists.denx.de/mailman/listinfo/u-boot
