In my case, a csb_patcher.sh script [1] - among other things - delivers the example config files for the opensource AGESA boards. Although these configs haven't been updated for a while (i.e. a last update of AMD Lenovo G505S config [2] is 1 year ago), I got away with this: they are still working, just get refreshed when a user does "menuconfig" & "make" and result in a working coreboot build. However, with a commit 0eab62b [3] they are rejected [see 4].
Of course I'm going to manually "refresh" the configs to temporarily fix this problem, but I see this could increase the maintenance burden and reduce the usefulness of coreboot config files shared online... Are there any advantages of KCONFIG_STRICT / KCONFIG_WERROR that outweigh these potential issues?
[1] https://review.coreboot.org/c/coreboot/+/64873 [2] https://review.coreboot.org/c/coreboot/+/64829 [3] https://review.coreboot.org/c/coreboot/+/79259 [4]
./coreboot/$ make menuconfig src/soc/intel/meteorlake/Kconfig:457:warning: config symbol defined without type ./coreboot/.config:26:warning: unknown symbol: COMPRESS_RAMSTAGE ./coreboot/.config:230:warning: unknown symbol: ARCH_ALL_STAGES_X86 ./coreboot/.config:249:warning: unknown symbol: VBT_DATA_SIZE_KB ./coreboot/.config:256:warning: unknown symbol: UART_PCI_ADDR ./coreboot/.config:264:warning: unknown symbol: S3_DATA_POS ./coreboot/.config:265:warning: unknown symbol: S3_DATA_SIZE ./coreboot/.config:274:warning: unknown symbol: LOGICAL_CPUS ./coreboot/.config:459:warning: unknown symbol: INTEL_GMA_OPREGION_2_0 ./coreboot/.config:572:warning: unknown symbol: PAYLOAD_YABITS ./coreboot/.config:574:warning: unknown symbol: PAYLOAD_TIANOCORE
ERROR: 10 warnings encountered, and warnings are errors.
make: *** [build/util/kconfig/Makefile.real:47: menuconfig] Error 1