On Thu, Apr 8, 2021 at 4:36 PM Jonas Gorski <jonas.gorski@xxxxxxxxx> wrote: > On Thu, 8 Apr 2021 at 14:57, Linus Walleij <linus.walleij@xxxxxxxxxx> wrote: > > On Thu, Apr 8, 2021 at 2:10 PM kernel test robot <lkp@xxxxxxxxx> wrote: > > > > > hppa-linux-ld: drivers/pinctrl/bcm/pinctrl-bcm63268.o: in function `pinconf_generic_dt_node_to_map_pin': > > > >> (.text+0x174): undefined reference to `pinconf_generic_dt_node_to_map' > > > > Well that is interesting, as the Kconfig looks like this in the latest > > -next tree: > > > > config PINCTRL_BCM63268 > > bool "Broadcom BCM63268 GPIO driver" > > depends on (BMIPS_GENERIC || COMPILE_TEST) > > depends on OF > > select PINCTRL_BCM63XX > > default BMIPS_GENERIC > > help > > Say Y here to enable the Broadcom BCM63268 GPIO driver. > > > > depends on OF. But this config has: > > # CONFIG_OF is not set > > So how is this happening? > > Not sure if this is a serious question, but if you check the robot > output, it is explicitly building the commit that adds the driver (it > even links the commit ;P), which is then obviously still missing the > OF dependency fixes you and Randy added later. Aha that explains it. No I didn't see the commit ID, I blame it on information stress, when the robot sends me something I assume it is building/testing HEAD. I think the robot used to say something like "HEAD works, this only hurts bisectability", that was helpful. Yours, Linus Walleij
