Hi,
On kernel 5.4 (yocto hardknott), bootm hangs after "starting kernel" if the kernel config RANDOMIZE_BASE is disabled.
The issue is that with RANDOMIZE_BASE enabled, even with the nokaslr kernel option, there is a offset of 0x188000 between the vmlinux symbols and the actual running kernel. This makes debugging / dump analyzing complicated and not future proof, especially that I didn't find where this 0x188000 comes from.
Thanks in advance for your help !
I am using the default yocto fitimage layout:
/dts-v1/;
/ {
description = "FIT Image";
#address-cells = <1>;
images {
kernel {
description = "Kernel";
data = /incbin/("vmlinux.bin");
type = "kernel";
arch = "arm64";
os = "linux";
compression = "none";
load = <0x80008000>;
entry = <0x80008000>;
hash {
algo = "sha1";
};
};
fdt {
description = "DTB";
data = /incbin/("fsl-ls1048a.dtb");
type = "flat_dt";
arch = "arm64";
compression = "none";
load = <0x90000000>;
hash {
algo = "sha1";
};
};
};
configurations {
default = "standard";
standard {
description = "Standard Boot";
kernel = "kernel";
fdt = "fdt";
hash {
algo = "sha1";
};
};
};
};
The issue is that we want to use bootm, because we need fitimage support (we want to enable secure boot)
I have escalated this case to the SE team, please refer to the following update from them.
Since maybe it's uboot's issue introduced by customer's custom needs, we have no plan to fix it.
Please refer to the following suggestion from the SE team.
Suggest customer use booti instead of bootm when disabling config RANDOMIZE_BASE. Refer the following link for details.
https://www.mail-archive.com/u-boot@lists.denx.de/msg411672.html