On Thu, 30 Jan 2025 at 19:30, Nathan Chancellor <nathan@xxxxxxxxxx> wrote: > It looks like it is in Josh's tree but that tree does not flow into > -next; IIRC, they have to be merged into -tip to show up there. > > https://git.kernel.org/pub/scm/linux/kernel/git/jpoimboe/linux.git/log/?h=objtool/core Ah nice - thanks for the info. > For the record, this will be disruptive for clang users because a number > of warnings have crept up in recent releases and this option will get > enabled for allmodconfig. [snip] > I think Josh already mentioned it but exposing -Werror for objtool is a > big committment. OK yeah, I hadn't really taken the implications on board, i.e. I hadn't really internalised the fact that this affects builds where the user didn't explicitly opt-in to strictness. Do you have a mental picture of how sources of objtool regressions are distributed in the kernel? I'm wondering if it would be alleviated if we enabled it for stuff like defconfig and tinyconfig, while disabling it for allmodconfig/allyesconfig. Looks like LTO_CLANG_FULL does the latter (forcefully) by depending on !COMPILE_TEST, maybe there's another way. But I can also envisage a world where that creates exactly as much work for you, just introducing Kconfig hackery for no reason! > If exposing this to the world feels too premature, maybe the flag could > be added then have a make variable like OBJTOOL_FLAGS to allow a > developer to pass it through if they wish? Yeah, that would definitely be a reasonable start. I'll wait and see if Josh has any additional thoughts.
