Re: [U-Boot] [PATCH 1/2] kbuild: Fix a false error of generic board support
Hello Masahiro Yamada,
Even the with this patch, we encountered a false error of generic board support. The problem was very interrmittent for us, but we were able to debug the problem to performing builds on EXT3 file systems which have a time stamp resolution of one second. To reproduce the problem, touch ./include/config/auto.conf and .config on a configured uboot tree on an EXT3 file system.
The patch below fixes the problem for us.
Thanks,
Matthew Gerlach
--- Makefile | 8 ++++++-- 1 file changed, 6 insertions(+), 2 deletions(-)
diff --git a/Makefile b/Makefile index 99097e1..84de2f9 100644 --- a/Makefile +++ b/Makefile @@ -493,8 +493,12 @@ include/config/%.conf: $(KCONFIG_CONFIG) include/config/auto.conf.cmd # is up-to-date. When we switch to a different board configuration, old CONFIG # macros are still remaining in include/config/auto.conf. Without the following # gimmick, wrong config.mk would be included leading nasty warnings/errors. -autoconf_is_current := $(if $(wildcard $(KCONFIG_CONFIG)),$(shell find . \ - -path ./include/config/auto.conf -newer $(KCONFIG_CONFIG))) +# We use if not (not -newer) so that we include config.mk in the event that the +# file timestamps are exacty equal which can happen on EXT3 filesystems. +autoconf_is_current := $(if $(wildcard $(KCONFIG_CONFIG)),\ + $(if $(shell find . -path ./include/config/auto.conf \ + ! -newer $(KCONFIG_CONFIG)),,./include/config/auto.conf)) + ifneq ($(autoconf_is_current),) include $(srctree)/config.mk endif
On Tue, Dec 16, 2014 at 02:56:44PM -0600, mgerlach wrote:
Hello Masahiro Yamada,
Even the with this patch, we encountered a false error of generic board support. The problem was very interrmittent for us, but we were able to debug the problem to performing builds on EXT3 file systems which have a time stamp resolution of one second. To reproduce the problem, touch ./include/config/auto.conf and .config on a configured uboot tree on an EXT3 file system.
The patch below fixes the problem for us.
This makes an odd race condition problem I run into when doing massively paralell builds worse :( With MAKEALL no ARM boards build (ARCH wasn't set in time so it tried arch//Makefile for something) and buildman was also broken in a bunch of places.
On Wednesday, December 17, 2014 at 12:56:46 AM, Tom Rini wrote:
On Tue, Dec 16, 2014 at 02:56:44PM -0600, mgerlach wrote:
Hello Masahiro Yamada,
Even the with this patch, we encountered a false error of generic board support. The problem was very interrmittent for us, but we were able to debug the problem to performing builds on EXT3 file systems which have a time stamp resolution of one second. To reproduce the problem, touch ./include/config/auto.conf and .config on a configured uboot tree on an EXT3 file system.
The patch below fixes the problem for us.
This makes an odd race condition problem I run into when doing massively paralell builds worse :( With MAKEALL no ARM boards build (ARCH wasn't set in time so it tried arch//Makefile for something) and buildman was also broken in a bunch of places.
Let me just chime in, I recall I did report something possibly similar some time ago [1]. I did not manage to put a finger on this issue though and I can no longer trigger it.
[1] http://lists.denx.de/pipermail/u-boot/2013-June/157481.html
Best regards, Marek Vasut
On Wed, Dec 17, 2014 at 02:04:32AM +0100, Marek Vasut wrote:
On Wednesday, December 17, 2014 at 12:56:46 AM, Tom Rini wrote:
On Tue, Dec 16, 2014 at 02:56:44PM -0600, mgerlach wrote:
Hello Masahiro Yamada,
Even the with this patch, we encountered a false error of generic board support. The problem was very interrmittent for us, but we were able to debug the problem to performing builds on EXT3 file systems which have a time stamp resolution of one second. To reproduce the problem, touch ./include/config/auto.conf and .config on a configured uboot tree on an EXT3 file system.
The patch below fixes the problem for us.
This makes an odd race condition problem I run into when doing massively paralell builds worse :( With MAKEALL no ARM boards build (ARCH wasn't set in time so it tried arch//Makefile for something) and buildman was also broken in a bunch of places.
Let me just chime in, I recall I did report something possibly similar some time ago [1]. I did not manage to put a finger on this issue though and I can no longer trigger it.
[1] http://lists.denx.de/pipermail/u-boot/2013-June/157481.html
No, my problem was introduced with the switch to Kbuild itself.
Hi Marek, Tom,
On Wed, 17 Dec 2014 10:52:32 -0500 Tom Rini trini@ti.com wrote:
On Wed, Dec 17, 2014 at 02:04:32AM +0100, Marek Vasut wrote:
On Wednesday, December 17, 2014 at 12:56:46 AM, Tom Rini wrote:
On Tue, Dec 16, 2014 at 02:56:44PM -0600, mgerlach wrote:
Hello Masahiro Yamada,
Even the with this patch, we encountered a false error of generic board support. The problem was very interrmittent for us, but we were able to debug the problem to performing builds on EXT3 file systems which have a time stamp resolution of one second. To reproduce the problem, touch ./include/config/auto.conf and .config on a configured uboot tree on an EXT3 file system.
The patch below fixes the problem for us.
This makes an odd race condition problem I run into when doing massively paralell builds worse :( With MAKEALL no ARM boards build (ARCH wasn't set in time so it tried arch//Makefile for something) and buildman was also broken in a bunch of places.
Let me just chime in, I recall I did report something possibly similar some time ago [1]. I did not manage to put a finger on this issue though and I can no longer trigger it.
[1] http://lists.denx.de/pipermail/u-boot/2013-June/157481.html
No, my problem was introduced with the switch to Kbuild itself.
Marek's problem is unrelated to this issue. (2013-June is before Kbuild)
York Sun also reported the similar build error in this thread. http://thread.gmane.org/gmane.comp.boot-loaders.u-boot/203331/focus=203468
This is really related to what Mathew saw, I think.
Some bug reports imply "autoconf_is_current" is not working on some cases although I have never been able to reproduce the problem on my box.
I guess it is almost time for further build-system refactoring.
Best Regards Masahiro Yamada
Hi Matthew, Tom,
On Tue, 16 Dec 2014 18:56:46 -0500 Tom Rini trini@ti.com wrote:
On Tue, Dec 16, 2014 at 02:56:44PM -0600, mgerlach wrote:
Hello Masahiro Yamada,
Even the with this patch, we encountered a false error of generic board support. The problem was very interrmittent for us, but we were able to debug the problem to performing builds on EXT3 file systems which have a time stamp resolution of one second. To reproduce the problem, touch ./include/config/auto.conf and .config on a configured uboot tree on an EXT3 file system.
The patch below fixes the problem for us.
This makes an odd race condition problem I run into when doing massively paralell builds worse :( With MAKEALL no ARM boards build (ARCH wasn't set in time so it tried arch//Makefile for something) and buildman was also broken in a bunch of places.
Me too.
Although I do not know why, Methew's patch makes it worse. Without MAKEALL or buildman, I hit the same error every time.
$ LANG=C make -j8 CROSS_COMPILE=arm-linux-gnueabi- ph1_ld4_defconfig all HOSTCC scripts/basic/fixdep HOSTCC scripts/kconfig/conf.o SHIPPED scripts/kconfig/zconf.tab.c SHIPPED scripts/kconfig/zconf.lex.c SHIPPED scripts/kconfig/zconf.hash.c HOSTCC scripts/kconfig/zconf.tab.o HOSTLD scripts/kconfig/conf # # configuration written to .config # # # configuration written to spl/.config # ./Makefile:511: arch//Makefile: No such file or directory make[1]: *** No rule to make target `arch//Makefile'. Stop. make: *** [__build_one_by_one] Error 2
Best Regards Masahiro Yamada
Hi Tom and Masahiro,
Neither my collegue nor I could reproduce the problem you encountered with our patch and the build line, "LANG=C make -j8 CROSS_COMPILE=arm-linux-gnueabi- ph1_ld4_defconfig all". Any suggestions on other ways to fix the false error and not break paralell builds would be extremely appreciated.
Thanks,
Matthew Gerlach
On Thu, 18 Dec 2014, Masahiro Yamada wrote:
Hi Matthew, Tom,
On Tue, 16 Dec 2014 18:56:46 -0500 Tom Rini trini@ti.com wrote:
On Tue, Dec 16, 2014 at 02:56:44PM -0600, mgerlach wrote:
Hello Masahiro Yamada,
Even the with this patch, we encountered a false error of generic board support. The problem was very interrmittent for us, but we were able to debug the problem to performing builds on EXT3 file systems which have a time stamp resolution of one second. To reproduce the problem, touch ./include/config/auto.conf and .config on a configured uboot tree on an EXT3 file system.
The patch below fixes the problem for us.
This makes an odd race condition problem I run into when doing massively paralell builds worse :( With MAKEALL no ARM boards build (ARCH wasn't set in time so it tried arch//Makefile for something) and buildman was also broken in a bunch of places.
Me too.
Although I do not know why, Methew's patch makes it worse. Without MAKEALL or buildman, I hit the same error every time.
$ LANG=C make -j8 CROSS_COMPILE=arm-linux-gnueabi- ph1_ld4_defconfig all HOSTCC scripts/basic/fixdep HOSTCC scripts/kconfig/conf.o SHIPPED scripts/kconfig/zconf.tab.c SHIPPED scripts/kconfig/zconf.lex.c SHIPPED scripts/kconfig/zconf.hash.c HOSTCC scripts/kconfig/zconf.tab.o HOSTLD scripts/kconfig/conf # # configuration written to .config # # # configuration written to spl/.config # ./Makefile:511: arch//Makefile: No such file or directory make[1]: *** No rule to make target `arch//Makefile'. Stop. make: *** [__build_one_by_one] Error 2
Best Regards Masahiro Yamada
participants (4)
-
Marek Vasut -
Masahiro Yamada -
mgerlach -
Tom Rini