On Fri, Jan 28, 2022 at 3:26 PM Elliot Berman <quic_eberman@xxxxxxxxxxx> wrote: > > > > On 1/28/2022 2:49 PM, Nick Desaulniers wrote: > > On Fri, Jan 28, 2022 at 2:45 PM Fangrui Song <maskray@xxxxxxxxxx> wrote: > >> > >> On 2022-01-28, Elliot Berman wrote: > >>> Allow additional arguments be passed to userprogs compilation. > >>> Reproducible clang builds need to provide a sysroot and gcc path to > >>> ensure same toolchain is used across hosts. KCFLAGS is not currently > >>> used for any user programs compilation, so add new USERCFLAGS and > >>> USERLDFLAGS which serves similar purpose as HOSTCFLAGS/HOSTLDFLAGS. > >>> > >>> Specifically, I'm trying to force CC_CAN_LINK to consistently fail in > >>> an environment where a user sysroot is not specifically available. > >>> Currently, Clang might automatically detect GCC installation on hosts > >>> which have it installed to a default location in /. With addition of > >>> these environment variables, you can specify flags such as: > >>> > >>> $ make USERCFLAGS=--sysroot=/dev/null USERLDFLAGS=-Wl,--sysroot=/dev/null > >>> > >>> to force sysroot detection to fail. > >> > >> -Wl,--sysroot=/dev/null => --sysroot > >> > >> As I mentioned in > >> https://lore.kernel.org/all/20220128031549.w5a4bilxbkppagfu@xxxxxxxxxx/ > >> -Wl,--sysroot=/dev/null does not suppress search paths like -L/lib . > > > > In that case, Elliot, can you please test whether USERLDFLAGS is > > necessary to be specified AT ALL? Maybe we can drop that addition from > > this patch if so? > > USERLDFLAGS as --sysroot=/dev/null isn't necessary and not needed for > the Android kernel usecase. I've included here since we have HOSTLDFLAGS > and LDFLAGS for host programs and kernel respectively. > Ok, for symmetry it's ok to keep HOSTLDFLAGS, consider dropping it from the "example" in the commit message. The note about -Wl, prefixes exists in the docs now which is good enough for me. -- Thanks, ~Nick Desaulniers
