On Wed, 17 Mar 2021 05:08:50 -0700 David Wolfskill <david_at_catwhisker.org> wrote: > My laptop is currently running main-n245489-15b82e00a164; after updating > sources to n245498-096a84721670, I am performing a source-based update. > > A simialr update on my build machine (which is headless, and thus does > not use anything related to X11) was successful. > > The laptop is set up to rebuild x11/nvidia-driver when the kernel > is updated. > > The buildkernel step on it fails with: > > ... > awk -f /usr/src/sys/conf/kmod_syms.awk nvidia-modeset.ko export_syms | xargs -J% objcopy % nvidia-modeset.ko > ===> lib (all) > ===> lib/libglvnd (all) > ... > ===> x11/driver (all) > ===> x11/extension (all) > ===> doc (all) > make[6]: "/usr/share/mk/bsd.man.mk" line 53: Malformed conditional (${MK_MANSPLITPKG} == "no") > make[6]: Fatal errors encountered -- cannot continue > make[6]: stopped in /common/S4/obj/usr/src/amd64.amd64/sys/CANARY/common/ports/x11/nvidia-driver/work/NVIDIA-FreeBSD-x86_64-460.56/doc > *** Error code 1 > > Stop. > > > On reviewing the list of files changed in 15b82e00a164..096a84721670, I > note a couple of promising-looking candidates: > > share/mk/bsd.opts.mk | 1 + > share/mk/src.opts.mk | 1 - > > Reviewing the commit log for share/mk/bsd.opts.mk, I see that the most > recent entry is: > > | commit 6827435548d257c672f934db5c6ff01012d96995 > | Author: Jung-uk Kim <jkim_at_FreeBSD.org> > | Date: Tue Mar 16 14:16:10 2021 -0400 > | > | pkgbase: Fix building out-of-tree manual pages > | > | c7e6cb9e08d6 introduced MK_MANSPLITPKG but it was not available for > | building out-of-tree manual pages. For example, x11/nvidia-driver fails > | with the following error: > | > | ===> doc (all) > | make[3]: "/usr/share/mk/bsd.man.mk" line 53: Malformed conditional (${MK_MANSPLITPKG} == "no") > | make[3]: Fatal errors encountered -- cannot continue > | > | Move the definition from src.opts.mk to bsd.opts.mk to make it visible. > > which looks ... apropos. > > Indeed, it appears that the n245494-6827435548d2 change was intended to > fix the issue that I am now just seeing. > > But I readily confess that I have neither familairity nor expertise > with share/mk/* (and that delving into it reminds me of "You are > in a mazy twist of passages, all different") > > So... help? What do I need to do to be able to build the kernel now? > > (E.g., if I need to just skip building x11/nvidia-driver once, get > everything installed, then build "normally" (with x11/nvidia-driver) > -- that's fine; I just need a clue.) > > Thanks. > > Peace, > david > -- > David H. Wolfskill david_at_catwhisker.org > That broadly-popular "American Rescue Plan" (stimulus/COVID relrief)? > Ever Republican vote in Congress was against it. > > See https://www.catwhisker.org/~david/publickey.gpg for my public key. Hi. Once world whichever between git c7e6cb9e08d6 and 0b373f26bea1 is installed, nvidia-driver SHOULD ALWAYS FAIL until world after git 6827435548d2 is installed. This is the primary reason why I don't like/don't use current PORTS_MODULES approach. (Not all modules are mandatory to boot.) Building PORTS_MODULES would be better having separate target, something like buildportsmodules/installportsmodules and do NOT block buildkernel/installkernel. With the approach, buildportsmodules/installportsmodules may depend on buildkernel/installkernel. Regards. -- Tomoaki AOKI <junchoon_at_dec.sakura.ne.jp>Received on Wed Mar 17 2021 - 13:27:02 UTC
This archive was generated by hypermail 2.4.0 : Wed May 19 2021 - 11:41:27 UTC