On 09.04.20 10:44, Ye Li wrote:
The FIT config node has reversed ATF and u-boot: ATF is set to firmware but u-boot is set to loadable. This script can work previously because spl fit driver wrongly appends fdt to all loadable images. With the issue fixed, the u-boot in loadable does not have fdt appended and fails to work. So correct script by moving u-boot to firmware and ATF to loadable.
I know this has been applied and the change itself is probably correct. But when I apply this to my 2020.01-based tree, TFA/U-Boot isn't loaded anymore.
Please note, that I have disabled CONFIG_SPL_FIT_IMAGE_TINY, so 9d15d1d1c24f ("Revert "common: spl_fit: Default to IH_OS_U_BOOT if FIT_IMAGE_TINY enabled"") won't help.
I haven't tested with master so far, so I'm not sure if I miss some other patch or if there is an actual issue.
I just wondered why I switched the order in the first place in fa99af41e0da ("imx: mkimage_fit_atf: Fix FIT image for correct boot order").
It would also have been nice if I would have been cc-ed for this patch as I was the one introducing the "wrong" order in fa99af41e0da.
Signed-off-by: Ye Li ye.li@nxp.com
arch/arm/mach-imx/mkimage_fit_atf.sh | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-)
diff --git a/arch/arm/mach-imx/mkimage_fit_atf.sh b/arch/arm/mach-imx/mkimage_fit_atf.sh index ad81d5e..dd1ca5a 100755 --- a/arch/arm/mach-imx/mkimage_fit_atf.sh +++ b/arch/arm/mach-imx/mkimage_fit_atf.sh @@ -116,8 +116,8 @@ if [ -f $BL32 ]; then cat << __CONF_SECTION_EOF config@$cnt { description = "$(basename $dtname .dtb)";
firmware = "atf@1";loadables = "uboot@1", "tee@1";
firmware = "uboot@1";loadables = "atf@1", "tee@1"; fdt = "fdt@$cnt";@@ -125,8 +125,8 @@ else cat << __CONF_SECTION1_EOF config@$cnt { description = "$(basename $dtname .dtb)";
firmware = "atf@1";loadables = "uboot@1";
firmware = "uboot@1";loadables = "atf@1"; fdt = "fdt@$cnt";