[PATCH] kbuild: unexport sub_make_done to fix child make invocations
Tom Rini
trini at konsulko.com
Wed Mar 4 16:45:49 CET 2026
On Wed, Mar 04, 2026 at 03:24:45PM +0000, Conor Dooley wrote:
> On Wed, Mar 04, 2026 at 09:11:54AM -0600, Tom Rini wrote:
> > On Wed, Mar 04, 2026 at 03:09:40PM +0000, Conor Dooley wrote:
> > > On Wed, Mar 04, 2026 at 09:01:45AM -0600, Tom Rini wrote:
> > > > On Wed, Mar 04, 2026 at 09:54:22AM +0000, Conor Dooley wrote:
> > > > > On Tue, Mar 03, 2026 at 12:50:55PM -0600, Tom Rini wrote:
> > > > > > On Fri, 13 Feb 2026 19:47:34 -0700, Simon Glass wrote:
> > > > > >
> > > > > > > The exported sub_make_done variable leaks into the environment of all
> > > > > > > child processes. When make targets like tcheck spawn independent make
> > > > > > > invocations with O=, those child makes inherit sub_make_done=1, skip
> > > > > > > the KBUILD_OUTPUT setup and try to build in the source tree.
> > > > > > >
> > > > > > > There is a workaround that resets sub_make_done to 0 for specific test
> > > > > > > targets, but this isn't great since it has tolist every target that
> > > > > > > spawns independent make invocations.
> > > > > > >
> > > > > > > [...]
> > > > > >
> > > > > > Applied to u-boot/master, thanks!
> > > > > >
> > > > > > [1/1] kbuild: unexport sub_make_done to fix child make invocations
> > > > > > commit: 4284306d22c5b6d64ecd62b462551d9d313c8104
> > > > >
> > > > >
> > > > > This breaks the build for me:
> > > > > make ARCH=riscv CROSS_COMPILE=riscv64-unknown-linux-gnu- O=build microchip_mpfs_generic_defconfig
> > > > > GEN Makefile
> > > > > #
> > > > > # configuration written to .config
> > > > > #
> > > > > make ARCH=riscv CROSS_COMPILE=riscv64-unknown-linux-gnu- O=build
> > > > > HOSTCC scripts/basic/fixdep
> > > > > GEN Makefile
> > > > > HOSTCC scripts/kconfig/conf.o
> > > > > YACC scripts/kconfig/zconf.tab.[ch]
> > > > > LEX scripts/kconfig/zconf.lex.c
> > > > > HOSTCC scripts/kconfig/zconf.tab.o
> > > > > HOSTLD scripts/kconfig/conf
> > > > > scripts/kconfig/conf --syncconfig Kconfig
> > > > > ***
> > > > > *** Configuration file ".config" not found!
> > > > > ***
> > > > > *** Please run some configurator (e.g. "make oldconfig" or
> > > > > *** "make menuconfig" or "make xconfig").
> > > > > ***
> > > > > make[4]: *** [/stuff/u-boot/scripts/kconfig/Makefile:75: syncconfig] Error 1
> > > > > make[3]: *** [/stuff/u-boot/Makefile:702: syncconfig] Error 2
> > > > > make[2]: *** [../Makefile:189: __sub-make] Error 2
> > > > > make[1]: *** No rule to make target 'include/config/auto.conf', needed by 'include/config/uboot.release'. Stop.
> > > > > make: *** [Makefile:189: __sub-make] Error 2
> > > >
> > > > That's odd. What host distro and make? I don't see that on Ubuntu 24.04.
> > >
> > > Originally my CI found it in a docker container with "from: ubuntu:noble",
> > > so 24.04. I reproduced it locally on debian unstable with make 4.4.1
> >
> > That's even more puzzling (noble is 24.04) since that sounds very close
> > to our CI too. Can you please look more in to what's going on?
>
> Can I assume your CI also sets O=<something> then? Is the CI config
> stuff in tree, or somewhere else?
We use buildman, which in turn does separate output directories. We
also, in Azure, have one job that uses O= specifically. Things are
defiend in .gitlab-ci.yml and .azure-pipelines.yml
--
Tom
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 228 bytes
Desc: not available
URL: <https://lists.denx.de/pipermail/u-boot/attachments/20260304/3935bf0c/attachment.sig>
More information about the U-Boot
mailing list