On 3/10/20 10:14 AM, John Garry wrote: > + > > On 10/03/2020 14:37, Randy Dunlap wrote: >> On 3/10/20 2:19 AM, Stephen Rothwell wrote: >>> Hi all, >>> >>> Changes since 20200306: >>> >>> Removed tree: nfc-next (abanboned) >> >> doned >> >> >> on i386: >> >> WARNING: unmet direct dependencies detected for MTD_SPI_NOR > > Uh, so kbuild honors a "select" of a config option which has unmet dependencies: > > https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/Documentation/kbuild/kconfig-language.rst?h=v5.6-rc5#n143 Yes. > So this looks introduced by: > > https://git.kernel.org/pub/scm/linux/kernel/git/next/linux-next.git/commit/drivers/spi/Kconfig?h=next-20200310&id=e14572c52546c16e159c4c1814984843a119e823 > >> Depends on [m]: MTD [=m] && SPI_MASTER [=y] > Selected by [y]: >> - SPI_HISI_SFC_V3XX [=y] && SPI [=y] && SPI_MASTER [=y] && (ARM64 && ACPI [=y] || COMPILE_TEST [=y]) && HAS_IOMEM [=y] Would it be acceptable to move the Kconfig entry for this: config SPI_HISI_SFC_V3XX tristate "HiSilicon SPI-NOR Flash Controller for Hi16XX chipsets" depends on (ARM64 && ACPI) || COMPILE_TEST depends on HAS_IOMEM select MTD_SPI_NOR into drivers/mtd/spi-nor/Kconfig, say after this one: config SPI_HISI_SFC tristate "Hisilicon FMC SPI-NOR Flash Controller(SFC)" depends on ARCH_HISI || COMPILE_TEST depends on HAS_IOMEM and drop the "select MTD_SPI_NOR" because when it's in drivers/mtd/spi-nor/Kconfig, it will depend on MTD_SPI_NOR. >> ERROR: "spi_nor_scan" [drivers/mtd/spi-nor/intel-spi.ko] undefined! >> ERROR: "spi_nor_scan" [drivers/mtd/spi-nor/mtk-quadspi.ko] undefined! >> ERROR: "spi_nor_scan" [drivers/mtd/spi-nor/hisi-sfc.ko] undefined! >> ERROR: "spi_nor_scan" [drivers/mtd/spi-nor/aspeed-smc.ko] undefined! >> >> >> Full randconfig file is attached. >> -- ~Randy
