[Armbian build PR] - more of rpardini's dirty laundry, very late April/'23

more of rpardini's dirty laundry, very late April/'23

Remove legacy branches after trying, and failing, to ressurect them

• rockchip: completely remove legacy branch; fixes #5118
  • there's absolutely no point in trying to keep this alive, current/edge are good enough, legacy's abandoned for years and broken
• jetson-nano legacy: give up on this legacy and remove it; fixes #5117
• jetson-nano legacy: try, and fail, to fix this legacy; (does NOT) fixes #5117
  • jetson-nano legacy: kernel config: CONFIG_SECTION_MISMATCH_WARN_ONLY=y
  • jetson-nano legacy: add -mno-outline-atomics to fix linking
  • jetson-nano legacy: add xxd host-dependency
  • jetson-nano legacy: downgrade a lot of gcc errors to allow this legacy to build

Fix (sucessfully) broken stuff

• rtl8822bs: fix: remove binaries (firmware?) being added to kernel driver source and breaking patching; related to #5117
• kernel-make: introduce KERNEL_EXTRA_CFLAGS; fill it to add to what's passed to KCFLAGS; related to #5117
• macchiatobin-doubleshot: fix u-boot/atf build: export SCP_BL2 had been wrongly converted to declare -g SCP_BL2 when it is in fact required; fixes #5116
• rk3399-4.4: remove mbox-confusing long explanation from ancient patch in ancient kernel; possibly fixes #5119
• driver_uwe5622_allwinner: only apply "pre-6.1" park-link patch if kernel >= 5 (eg: not for rockchip64-legacy); possibly fixes #5120
• rockchip64_common: add explicit make target u-boot.dtb before u-boot.itb in the target map for SPI; fixes #5114
  • when BOOT_SUPPORT_SPI=yes (and not using the rkspi_loader)
  • it seems the makefile target u-boot.itb does not correctly depend on the u-boot.dtb target, and works "by accident" on multicore machines, which are faster and build it fast enough in parallel. this adds u-boot.dtb to the target map, so make's hand is forced to do the right thing.
• artifact-uboot: fix: include UBOOT_TARGET_MAP in variables hash of version
• kernel-debs: postinst/etc: fix: unify is_boot_dev_vfat() function; make work during image build too via env var
  • determinining if /boot is vfat is more complex than it looks;
    • use a common function to unify across all scripts
  • during image build, the kernel install is always done with a non-fat, non-mounted /boot
    • use a special variable passed through the env so unified function can know ahead of time
    • introduce extra_apt_envs param to chroot_sdcard_apt_get(); array with extra environment vars
  • if vfat /boot:
    • linux-update-symlinks is not called (Debian would just choke on this)
    • don't even try to symlink Armbian stuff, move instead
  • more/better logging board-side (unfortunately mixed in all the set -x debugs)

View the full article