On Sat, Dec 22, 2018 at 6:47 PM Sinan Kaya <okaya@xxxxxxxxxx> wrote: > > Collect reviewed-by for 3/11 and 6/11 from Lukas > Add a few words about 8/11 why PCI dependency is being added > > Sinan Kaya (11): > ACPI / LPSS: Add guards against CONFIG_PCI > ata: make PCI dependency explicit for PATA_ACPI > vga-switcheroo: make PCI dependency explicit > platform/x86: make PCI dependency explicit > platform/x86: intel_pmc: Hide PCI specific pieces behind CONFIG_PCI > apple-gmux: Make PCI dependency explicit > drivers: thermal: Hide PCI driver when CONFIG_PCI is unset > ASoC: Intel: Make PCI dependency explicit > mmc: add PCI dependency into IOSF_MBI > x86: select IOSF_MBI only when CONFIG_PCI is set > drivers: thermal: Make PCI dependency explicit > > arch/x86/Kconfig | 2 +- > drivers/acpi/acpi_lpss.c | 5 ++++- > drivers/ata/Kconfig | 2 +- > drivers/gpu/vga/Kconfig | 1 + > drivers/mmc/host/Kconfig | 2 +- > drivers/platform/x86/Kconfig | 3 ++- > drivers/platform/x86/intel_pmc_ipc.c | 6 ++++++ > drivers/thermal/intel/int340x_thermal/Kconfig | 2 +- > .../intel/int340x_thermal/processor_thermal_device.c | 8 ++++++-- > sound/soc/intel/Kconfig | 2 +- > 10 files changed, 24 insertions(+), 9 deletions(-) Why exactly do you think that adding #ifdefs around stuff in random places just because then don't build without CONFIG_PCI makes any sense at all? Please don't do that. If something requires CONFIG_PCI to build, make it depend on PCI, unless you *know* it for a fact that it *will* work with your new #ifdefs and without PCI. In which case the changelog must say that and specify the platform you have tested in on.
