On Tue, Nov 5, 2013 at 2:01 AM, Gleb Smirnoff <glebius_at_freebsd.org> wrote: > On Mon, Nov 04, 2013 at 10:09:54PM -0700, Ian Lepore wrote: > I> > > /usr/src/sys/dev/aic7xxx/aic7xxx.seq > I> > > > make[2]: exec(aicasm) failed (No such file or directory) > I> > > > *** Error code 1 > I> > > > > I> > > > Stop. > I> > > > make[2]: stopped in /usr/obj/usr/src/sys/GENERIC > I> > > > *** Error code 1 > I> > > > > I> > > > Stop. > I> > > > make[1]: stopped in /usr/src > I> > > > I> > > Did you update your source and then "make buildkernel" without > I> > > buildworld? If so, a "make kernel-toolchain" should create the > aicasm > I> > > tool and get you back on track. > I> > > > I> > > I> > really odd as i built a kernel this morning no problem, then updated > the > I> > tree and went to build another kernel and got that. > I> > working through it, thanks for the input.......... > I> > > I> > I> You were just unlucky that your updates bracketed my checkin that > I> changed the build process for the aicasm tool so that it gets built as > I> part of the toolchain rather than as part of the kernel now. > > Before this change, the toolchain was not required for kernel build if you > aren't cross building. And now it is. This breaks the kernel build > procedure > documented in handbook for years, and brings a lot of discomfort to > developers. > > Now to test a trivial change to kernel I need first to compile clang. > > I concur, I feel this is a bit invasive and unneccesary, as custom build scripts now also fail, what has worked for years without issue is now completely broken, it makes no sense to me to have to build a toolchain just to build a kernel everytime. > -- > Totus tuus, Glebius. >Received on Tue Nov 05 2013 - 12:00:23 UTC
This archive was generated by hypermail 2.4.0 : Wed May 19 2021 - 11:40:44 UTC