On 14:38-20190115, Tero Kristo wrote: [...] > > The mentioned failure only happens with a specific .config. Not sure what is > the actual Kconfig that masks the failure though, but anyway it looks like > interrupt.h gets included via some indirect path with arm64 defconfig making > it pass. I am able to reproduce the build fail as well: .config: -> Notice that all v8 arch other than K3 is disabled, expert mode is enabled and few ancillary "optimization" of config https://pastebin.ubuntu.com/p/w5t9Wvp2jJ/ Here is what I did: git checkout next-20190115 wget -O inta.mbox https://lore.kernel.org/patchwork/series/377815/mbox/ git am inta.mbox I put the above .config as my .config make oldconfig make drivers/irqchip/irq-ti-sci-inta.o I got the exact failure as well.. https://pastebin.ubuntu.com/p/JzFXGz7Gh9/ I agree with peter than this should be fixed. randconfig would have caught this anyways.. I did'nt bother looking further in the config to figure out why the side effect, but glad it got caught early enough. Thanks Peter. Lokesh, I know you are out this week, will appreciate if you could post a v5 once you are back? -- Regards, Nishanth Menon
