| Message ID | 20220601165909.46588-2-anthony.perard@citrix.com (mailing list archive) |
|---|---|
| State | Superseded |
| Headers | show |
| Series | xen: rework compat headers generation | expand |
On 01.06.2022 18:59, Anthony PERARD wrote: > Use "define" for the headers*_chk commands as otherwise the "#" > is interpreted as a comment and make can't find the end of > $(foreach,). In cmd_xlat_lst you use $(pound) - any reason this doesn't work in these rules? Note that I don't mind the use of "define", just that I'm puzzled by the justification. Jan
On Thu, Jun 02, 2022 at 11:11:15AM +0200, Jan Beulich wrote: > On 01.06.2022 18:59, Anthony PERARD wrote: > > Use "define" for the headers*_chk commands as otherwise the "#" > > is interpreted as a comment and make can't find the end of > > $(foreach,). > > In cmd_xlat_lst you use $(pound) - any reason this doesn't work in > these rules? Note that I don't mind the use of "define", just that I'm > puzzled by the justification. I think I just forgot about $(pound) when I tried to debug why the command didn't work in some environment (that is when not using define). I think the second thing that make me not replacing '#' by "$(pound)" is that reading "#include" in the Makefile is probably better that reading "$(pound)include". I guess we should add something to the justification like: That allow us to keep writing "#include" in the Makefile without having to replace that by "$(pound)include" which would be a bit less obvious about the command line purpose. Thanks,
On 06.06.2022 15:39, Anthony PERARD wrote: > On Thu, Jun 02, 2022 at 11:11:15AM +0200, Jan Beulich wrote: >> On 01.06.2022 18:59, Anthony PERARD wrote: >>> Use "define" for the headers*_chk commands as otherwise the "#" >>> is interpreted as a comment and make can't find the end of >>> $(foreach,). >> >> In cmd_xlat_lst you use $(pound) - any reason this doesn't work in >> these rules? Note that I don't mind the use of "define", just that I'm >> puzzled by the justification. > > I think I just forgot about $(pound) when I tried to debug why the > command didn't work in some environment (that is when not using define). > > I think the second thing that make me not replacing '#' by "$(pound)" is > that reading "#include" in the Makefile is probably better that > reading "$(pound)include". > > I guess we should add something to the justification like: > That allow us to keep writing "#include" in the Makefile without > having to replace that by "$(pound)include" which would be a bit > less obvious about the command line purpose. I'd be okay with this, and I'd also be okay with adding this while committing. With it added Acked-by: Jan Beulich <jbeulich@suse.com> Jan
Hi Anthony, > On 1 Jun 2022, at 17:59, Anthony PERARD <anthony.perard@citrix.com> wrote: > > Use "define" for the headers*_chk commands as otherwise the "#" > is interpreted as a comment and make can't find the end of > $(foreach,). > > Adding several .PRECIOUS as without them `make` deletes the > intermediate targets. This is an issue because the macro $(if_changed,) > check if the target exist in order to decide whether to recreate the > target. > > Removing the call to `mkdir` from the commands. Those aren't needed > anymore because a rune in Rules.mk creates the directory for each > $(targets). > > Remove "export PYTHON" as it is already exported. With this change, compiling for x86 is now ending up in: CHK include/headers99.chk make[9]: execvp: /bin/sh: Argument list too long make[9]: *** [include/Makefile:181: include/headers++.chk] Error 127 Not quite sure yet why but I wanted to signal it early as other might be impacted. Arm and arm64 builds are not impacted. Cheers Bertrand > > Signed-off-by: Anthony PERARD <anthony.perard@citrix.com> > --- > xen/include/Makefile | 108 ++++++++++++++++++++++++++++++------------- > 1 file changed, 76 insertions(+), 32 deletions(-) > > diff --git a/xen/include/Makefile b/xen/include/Makefile > index 03baf10efb..6d9bcc19b0 100644 > --- a/xen/include/Makefile > +++ b/xen/include/Makefile > @@ -45,38 +45,65 @@ public-$(CONFIG_ARM) := $(wildcard $(srcdir)/public/arch-arm/*.h $(srcdir)/publi > .PHONY: all > all: $(addprefix $(obj)/,$(headers-y)) > > -$(obj)/compat/%.h: $(obj)/compat/%.i $(srcdir)/Makefile $(srctree)/tools/compat-build-header.py > - $(PYTHON) $(srctree)/tools/compat-build-header.py <$< $(patsubst $(obj)/%,%,$@) >>$@.new; \ > - mv -f $@.new $@ > - > -$(obj)/compat/%.i: $(obj)/compat/%.c $(srcdir)/Makefile > - $(CPP) $(filter-out -Wa$(comma)% -include %/include/xen/config.h,$(XEN_CFLAGS)) $(cppflags-y) -o $@ $< > - > -$(obj)/compat/%.c: $(src)/public/%.h $(srcdir)/xlat.lst $(srcdir)/Makefile $(srctree)/tools/compat-build-source.py > - mkdir -p $(@D) > - $(PYTHON) $(srctree)/tools/compat-build-source.py $(srcdir)/xlat.lst <$< >$@.new > - mv -f $@.new $@ > - > -$(obj)/compat/.xlat/%.h: $(obj)/compat/%.h $(obj)/compat/.xlat/%.lst $(srctree)/tools/get-fields.sh $(srcdir)/Makefile > - export PYTHON=$(PYTHON); \ > - while read what name; do \ > - $(SHELL) $(srctree)/tools/get-fields.sh "$$what" compat_$$name $< || exit $$?; \ > - done <$(patsubst $(obj)/compat/%,$(obj)/compat/.xlat/%,$(basename $<)).lst >$@.new > - mv -f $@.new $@ > +quiet_cmd_compat_h = GEN $@ > +cmd_compat_h = \ > + $(PYTHON) $(srctree)/tools/compat-build-header.py <$< $(patsubst $(obj)/%,%,$@) >>$@.new; \ > + mv -f $@.new $@ > + > +quiet_cmd_compat_i = CPP $@ > +cmd_compat_i = $(CPP) $(filter-out -Wa$(comma)% -include %/include/xen/config.h,$(XEN_CFLAGS)) $(cppflags-y) -o $@ $< > + > +quiet_cmd_compat_c = GEN $@ > +cmd_compat_c = \ > + $(PYTHON) $(srctree)/tools/compat-build-source.py $(srcdir)/xlat.lst <$< >$@.new; \ > + mv -f $@.new $@ > + > +quiet_cmd_xlat_headers = GEN $@ > +cmd_xlat_headers = \ > + while read what name; do \ > + $(SHELL) $(srctree)/tools/get-fields.sh "$$what" compat_$$name $< || exit $$?; \ > + done <$(patsubst $(obj)/compat/%,$(obj)/compat/.xlat/%,$(basename $<)).lst >$@.new; \ > + mv -f $@.new $@ > + > +targets += $(headers-y) > +$(obj)/compat/%.h: $(obj)/compat/%.i $(srctree)/tools/compat-build-header.py FORCE > + $(call if_changed,compat_h) > + > +.PRECIOUS: $(obj)/compat/%.i > +targets += $(patsubst %.h, %.i, $(headers-y)) > +$(obj)/compat/%.i: $(obj)/compat/%.c FORCE > + $(call if_changed,compat_i) > + > +.PRECIOUS: $(obj)/compat/%.c > +targets += $(patsubst %.h, %.c, $(headers-y)) > +$(obj)/compat/%.c: $(src)/public/%.h $(srcdir)/xlat.lst $(srctree)/tools/compat-build-source.py FORCE > + $(call if_changed,compat_c) > + > +targets += $(patsubst compat/%, compat/.xlat/%, $(headers-y)) > +$(obj)/compat/.xlat/%.h: $(obj)/compat/%.h $(obj)/compat/.xlat/%.lst $(srctree)/tools/get-fields.sh FORCE > + $(call if_changed,xlat_headers) > + > +quiet_cmd_xlat_lst = GEN $@ > +cmd_xlat_lst = \ > + grep -v '^[[:blank:]]*$(pound)' $< | sed -ne 's,@arch@,$(compat-arch-y),g' -re 's,[[:blank:]]+$*\.h[[:blank:]]*$$,,p' >$@.new; \ > + $(call move-if-changed,$@.new,$@) > > .PRECIOUS: $(obj)/compat/.xlat/%.lst > -$(obj)/compat/.xlat/%.lst: $(srcdir)/xlat.lst $(srcdir)/Makefile > - mkdir -p $(@D) > - grep -v '^[[:blank:]]*#' $< | sed -ne 's,@arch@,$(compat-arch-y),g' -re 's,[[:blank:]]+$*\.h[[:blank:]]*$$,,p' >$@.new > - $(call move-if-changed,$@.new,$@) > +targets += $(patsubst compat/%.h, compat/.xlat/%.lst, $(headers-y)) > +$(obj)/compat/.xlat/%.lst: $(srcdir)/xlat.lst FORCE > + $(call if_changed,xlat_lst) > > xlat-y := $(shell sed -ne 's,@arch@,$(compat-arch-y),g' -re 's,^[?!][[:blank:]]+[^[:blank:]]+[[:blank:]]+,,p' $(srcdir)/xlat.lst | uniq) > xlat-y := $(filter $(patsubst compat/%,%,$(headers-y)),$(xlat-y)) > > -$(obj)/compat/xlat.h: $(addprefix $(obj)/compat/.xlat/,$(xlat-y)) $(obj)/config/auto.conf $(srcdir)/Makefile > - cat $(filter %.h,$^) >$@.new > +quiet_cmd_xlat_h = GEN $@ > +cmd_xlat_h = \ > + cat $(filter %.h,$^) >$@.new; \ > mv -f $@.new $@ > > +$(obj)/compat/xlat.h: $(addprefix $(obj)/compat/.xlat/,$(xlat-y)) $(obj)/config/auto.conf FORCE > + $(call if_changed,xlat_h) > + > ifeq ($(XEN_TARGET_ARCH),$(XEN_COMPILE_ARCH)) > > all: $(obj)/headers.chk $(obj)/headers99.chk $(obj)/headers++.chk > @@ -102,27 +129,31 @@ PUBLIC_C99_HEADERS := $(call public-filter-headers,public-c99-headers) > $(src)/public/io/9pfs.h-prereq := string > $(src)/public/io/pvcalls.h-prereq := string > > -$(obj)/headers.chk: $(PUBLIC_ANSI_HEADERS) $(srcdir)/Makefile > +quiet_cmd_header_chk = CHK $@ > +cmd_header_chk = \ > for i in $(filter %.h,$^); do \ > $(CC) -x c -ansi -Wall -Werror -include stdint.h \ > -S -o /dev/null $$i || exit 1; \ > echo $$i; \ > - done >$@.new > + done >$@.new; \ > mv $@.new $@ > > -$(obj)/headers99.chk: $(PUBLIC_C99_HEADERS) $(srcdir)/Makefile > - rm -f $@.new > +quiet_cmd_headers99_chk = CHK $@ > +define cmd_headers99_chk > + rm -f $@.new; \ > $(foreach i, $(filter %.h,$^), \ > echo "#include "\"$(i)\" \ > | $(CC) -x c -std=c99 -Wall -Werror \ > -include stdint.h \ > $(foreach j, $($(patsubst $(srctree)/%,%,$i)-prereq), -include $(j).h) \ > -S -o /dev/null - \ > - || exit $$?; echo $(i) >> $@.new;) > + || exit $$?; echo $(i) >> $@.new;) \ > mv $@.new $@ > +endef > > -$(obj)/headers++.chk: $(PUBLIC_HEADERS) $(srcdir)/Makefile > - rm -f $@.new > +quiet_cmd_headerscxx_chk = CHK $@ > +define cmd_headerscxx_chk > + rm -f $@.new; \ > if ! $(CXX) -v >/dev/null 2>&1; then \ > touch $@.new; \ > exit 0; \ > @@ -133,8 +164,21 @@ $(obj)/headers++.chk: $(PUBLIC_HEADERS) $(srcdir)/Makefile > -include stdint.h -include $(srcdir)/public/xen.h \ > $(foreach j, $($(patsubst $(srctree)/%,%,$i)-prereq), -include c$(j)) \ > -S -o /dev/null - \ > - || exit $$?; echo $(i) >> $@.new;) > + || exit $$?; echo $(i) >> $@.new;) \ > mv $@.new $@ > +endef > + > +targets += headers.chk > +$(obj)/headers.chk: $(PUBLIC_ANSI_HEADERS) FORCE > + $(call if_changed,header_chk) > + > +targets += headers99.chk > +$(obj)/headers99.chk: $(PUBLIC_C99_HEADERS) FORCE > + $(call if_changed,headers99_chk) > + > +targets += headers++.chk > +$(obj)/headers++.chk: $(PUBLIC_HEADERS) FORCE > + $(call if_changed,headerscxx_chk) > > endif > > -- > Anthony PERARD > >
On 09.06.2022 12:16, Bertrand Marquis wrote: >> On 1 Jun 2022, at 17:59, Anthony PERARD <anthony.perard@citrix.com> wrote: >> >> Use "define" for the headers*_chk commands as otherwise the "#" >> is interpreted as a comment and make can't find the end of >> $(foreach,). >> >> Adding several .PRECIOUS as without them `make` deletes the >> intermediate targets. This is an issue because the macro $(if_changed,) >> check if the target exist in order to decide whether to recreate the >> target. >> >> Removing the call to `mkdir` from the commands. Those aren't needed >> anymore because a rune in Rules.mk creates the directory for each >> $(targets). >> >> Remove "export PYTHON" as it is already exported. > > With this change, compiling for x86 is now ending up in: > CHK include/headers99.chk > make[9]: execvp: /bin/sh: Argument list too long > make[9]: *** [include/Makefile:181: include/headers++.chk] Error 127 > > Not quite sure yet why but I wanted to signal it early as other might be impacted. > > Arm and arm64 builds are not impacted. Hmm, that patch has passed the smoke push gate already, so there likely is more to it than there being an unconditional issue. I did build-test this before pushing, and I've just re-tested on a 2nd system without seeing an issue. Also please remember to trim your replies. Jan
Hi, > On 9 Jun 2022, at 11:26, Jan Beulich <jbeulich@suse.com> wrote: > > On 09.06.2022 12:16, Bertrand Marquis wrote: >>> On 1 Jun 2022, at 17:59, Anthony PERARD <anthony.perard@citrix.com> wrote: >>> >>> Use "define" for the headers*_chk commands as otherwise the "#" >>> is interpreted as a comment and make can't find the end of >>> $(foreach,). >>> >>> Adding several .PRECIOUS as without them `make` deletes the >>> intermediate targets. This is an issue because the macro $(if_changed,) >>> check if the target exist in order to decide whether to recreate the >>> target. >>> >>> Removing the call to `mkdir` from the commands. Those aren't needed >>> anymore because a rune in Rules.mk creates the directory for each >>> $(targets). >>> >>> Remove "export PYTHON" as it is already exported. >> >> With this change, compiling for x86 is now ending up in: >> CHK include/headers99.chk >> make[9]: execvp: /bin/sh: Argument list too long >> make[9]: *** [include/Makefile:181: include/headers++.chk] Error 127 >> >> Not quite sure yet why but I wanted to signal it early as other might be impacted. >> >> Arm and arm64 builds are not impacted. > > Hmm, that patch has passed the smoke push gate already, so there likely is > more to it than there being an unconditional issue. I did build-test this > before pushing, and I've just re-tested on a 2nd system without seeing an > issue. I have the problem only when building using Yocto, I did a normal build and the issue is not coming. Doing a verbose compilation I have this (sorry for the long lines): for i in include/public/vcpu.h include/public/errno.h include/public/kexec.h include/public/argo.h include/public/xen.h include/public/nmi.h include/public/xencomm.h include/public/xenoprof.h include/public/device_tree_defs.h include/public/version.h include/public/memory.h include/public/features.h include/public/sched.h include/public/xen-compat.h include/public/callback.h include/public/vm_event.h include/public/grant_table.h include/public/physdev.h include/public/tmem.h include/public/hypfs.h include/public/platform.h include/public/pmu.h include/public/elfnote.h include/public/trace.h include/public/event_channel.h include/public/io/vscsiif.h include/public/io/kbdif.h include/public/io/protocols.h include/public/io/ring.h include/public/io/displif.h include/public/io/fsif.h include/public/io/blkif.h include/public/io/console.h include/public/io/sndif.h include/public/io/fbif.h include/public/io/libxenvchan.h include/public/io/netif.h include/public/io/usbif.h include/public/io/pciif.h include/public/io/tpmif.h include/public/io/xs_wire.h include/public/io/xenbus.h include/public/io/cameraif.h include/public/hvm/pvdrivers.h include/public/hvm/e820.h include/public/hvm/hvm_xs_strings.h include/public/hvm/dm_op.h include/public/hvm/ioreq.h include/public/hvm/hvm_info_table.h include/public/hvm/hvm_vcpu.h include/public/hvm/hvm_op.h include/public/hvm/params.h; do x86_64-poky-linux-gcc --sysroot=/home/bermar01/Development/xen-dev/build/profile-qemu-x86_64.prj/tmp/work/core2-64-poky-linux/xen/4.17+git1-r0/recipe-sysroot -x c -ansi -Wall -Werror -include stdint.h -S -o /dev/null $i || exit 1; echo $i; done >include/headers.chk.new; mv include/headers.chk.new include/headers.chk | rm -f include/headers99.chk.new; echo "#include "\"include/public/io/9pfs.h\" | x86_64-poky-linux-gcc --sysroot=/home/bermar01/Development/xen-dev/build/profile-qemu-x86_64.prj/tmp/work/core2-64-poky-linux/xen/4.17+git1-r0/recipe-sysroot -x c -std=c99 -Wall -Werror -include stdint.h -include string.h -S -o /dev/null - || exit $?; echo include/public/io/9pfs.h >> include/headers99.chk.new; echo "#include "\"include/public/io/pvcalls.h\" | x86_64-poky-linux-gcc --sysroot=/home/bermar01/Development/xen-dev/build/profile-qemu-x86_64.prj/tmp/work/core2-64-poky-linux/xen/4.17+git1-r0/recipe-sysroot -x c -std=c99 -Wall -Werror -include stdint.h -include string.h -S -o /dev/null - || exit $?; echo include/public/io/pvcalls.h >> include/headers99.chk.new; mv include/headers99.chk.new include/headers99.chk | make[9]: execvp: /bin/sh: Argument list too long | make[9]: *** [include/Makefile:181: include/headers++.chk] Error 127 | make[9]: *** Waiting for unfinished jobs.... So the command passed to the sub shell by make is quite long. No idea why this comes out only when building in Yocto but I will dig a bit. > > Also please remember to trim your replies. > Will do. Bertrand
On 09.06.2022 13:51, Bertrand Marquis wrote: >> On 9 Jun 2022, at 11:26, Jan Beulich <jbeulich@suse.com> wrote: >> On 09.06.2022 12:16, Bertrand Marquis wrote: >>>> On 1 Jun 2022, at 17:59, Anthony PERARD <anthony.perard@citrix.com> wrote: >>>> >>>> Use "define" for the headers*_chk commands as otherwise the "#" >>>> is interpreted as a comment and make can't find the end of >>>> $(foreach,). >>>> >>>> Adding several .PRECIOUS as without them `make` deletes the >>>> intermediate targets. This is an issue because the macro $(if_changed,) >>>> check if the target exist in order to decide whether to recreate the >>>> target. >>>> >>>> Removing the call to `mkdir` from the commands. Those aren't needed >>>> anymore because a rune in Rules.mk creates the directory for each >>>> $(targets). >>>> >>>> Remove "export PYTHON" as it is already exported. >>> >>> With this change, compiling for x86 is now ending up in: >>> CHK include/headers99.chk >>> make[9]: execvp: /bin/sh: Argument list too long >>> make[9]: *** [include/Makefile:181: include/headers++.chk] Error 127 >>> >>> Not quite sure yet why but I wanted to signal it early as other might be impacted. >>> >>> Arm and arm64 builds are not impacted. >> >> Hmm, that patch has passed the smoke push gate already, so there likely is >> more to it than there being an unconditional issue. I did build-test this >> before pushing, and I've just re-tested on a 2nd system without seeing an >> issue. > > I have the problem only when building using Yocto, I did a normal build and the > issue is not coming. > > Doing a verbose compilation I have this (sorry for the long lines): > > for i in include/public/vcpu.h include/public/errno.h include/public/kexec.h include/public/argo.h include/public/xen.h include/public/nmi.h include/public/xencomm.h include/public/xenoprof.h include/public/device_tree_defs.h include/public/version.h include/public/memory.h include/public/features.h include/public/sched.h include/public/xen-compat.h include/public/callback.h include/public/vm_event.h include/public/grant_table.h include/public/physdev.h include/public/tmem.h include/public/hypfs.h include/public/platform.h include/public/pmu.h include/public/elfnote.h include/public/trace.h include/public/event_channel.h include/public/io/vscsiif.h include/public/io/kbdif.h include/public/io/protocols.h include/public/io/ring.h include/public/io/displif.h include/public/io/fsif.h include/public/io/blkif.h include/public/io/console.h include/public/io/sndif.h include/public/io/fbif.h include/public/io/libxenvchan.h include/public/io/netif.h include/public/io/usbif.h include/public/io/pciif.h include/public/io/tpmif.h include/public/io/xs_wire.h include/public/io/xenbus.h include/public/io/cameraif.h include/public/hvm/pvdrivers.h include/public/hvm/e820.h include/public/hvm/hvm_xs_strings.h include/public/hvm/dm_op.h include/public/hvm/ioreq.h include/public/hvm/hvm_info_table.h include/public/hvm/hvm_vcpu.h include/public/hvm/hvm_op.h include/public/hvm/params.h; do x86_64-poky-linux-gcc --sysroot=/home/bermar01/Development/xen-dev/build/profile-qemu-x86_64.prj/tmp/work/core2-64-poky-linux/xen/4.17+git1-r0/recipe-sysroot -x c -ansi -Wall -Werror -include stdint.h -S -o /dev/null $i || exit 1; echo $i; done >include/headers.chk.new; mv include/headers.chk.new include/headers.chk > | rm -f include/headers99.chk.new; echo "#include "\"include/public/io/9pfs.h\" | x86_64-poky-linux-gcc --sysroot=/home/bermar01/Development/xen-dev/build/profile-qemu-x86_64.prj/tmp/work/core2-64-poky-linux/xen/4.17+git1-r0/recipe-sysroot -x c -std=c99 -Wall -Werror -include stdint.h -include string.h -S -o /dev/null - || exit $?; echo include/public/io/9pfs.h >> include/headers99.chk.new; echo "#include "\"include/public/io/pvcalls.h\" | x86_64-poky-linux-gcc --sysroot=/home/bermar01/Development/xen-dev/build/profile-qemu-x86_64.prj/tmp/work/core2-64-poky-linux/xen/4.17+git1-r0/recipe-sysroot -x c -std=c99 -Wall -Werror -include stdint.h -include string.h -S -o /dev/null - || exit $?; echo include/public/io/pvcalls.h >> include/headers99.chk.new; mv include/headers99.chk.new include/headers99.chk > | make[9]: execvp: /bin/sh: Argument list too long > | make[9]: *** [include/Makefile:181: include/headers++.chk] Error 127 > | make[9]: *** Waiting for unfinished jobs.... > > So the command passed to the sub shell by make is quite long. > > No idea why this comes out only when building in Yocto but I will dig a bit. Maybe Yocto has an unusually low limit on command arguments' total size? The whole thing is just over 2500 chars, which doesn't look to be unusually long for Unix-like environments. Jan
Hi, > On 9 Jun 2022, at 13:16, Jan Beulich <jbeulich@suse.com> wrote: > > On 09.06.2022 13:51, Bertrand Marquis wrote: >>> On 9 Jun 2022, at 11:26, Jan Beulich <jbeulich@suse.com> wrote: >>> On 09.06.2022 12:16, Bertrand Marquis wrote: >>>>> On 1 Jun 2022, at 17:59, Anthony PERARD <anthony.perard@citrix.com> wrote: >>>>> >>>>> Use "define" for the headers*_chk commands as otherwise the "#" >>>>> is interpreted as a comment and make can't find the end of >>>>> $(foreach,). >>>>> >>>>> Adding several .PRECIOUS as without them `make` deletes the >>>>> intermediate targets. This is an issue because the macro $(if_changed,) >>>>> check if the target exist in order to decide whether to recreate the >>>>> target. >>>>> >>>>> Removing the call to `mkdir` from the commands. Those aren't needed >>>>> anymore because a rune in Rules.mk creates the directory for each >>>>> $(targets). >>>>> >>>>> Remove "export PYTHON" as it is already exported. >>>> >>>> With this change, compiling for x86 is now ending up in: >>>> CHK include/headers99.chk >>>> make[9]: execvp: /bin/sh: Argument list too long >>>> make[9]: *** [include/Makefile:181: include/headers++.chk] Error 127 >>>> >>>> Not quite sure yet why but I wanted to signal it early as other might be impacted. >>>> >>>> Arm and arm64 builds are not impacted. >>> >>> Hmm, that patch has passed the smoke push gate already, so there likely is >>> more to it than there being an unconditional issue. I did build-test this >>> before pushing, and I've just re-tested on a 2nd system without seeing an >>> issue. >> >> I have the problem only when building using Yocto, I did a normal build and the >> issue is not coming. >> >> Doing a verbose compilation I have this (sorry for the long lines): >> >> for i in include/public/vcpu.h include/public/errno.h include/public/kexec.h include/public/argo.h include/public/xen.h include/public/nmi.h include/public/xencomm.h include/public/xenoprof.h include/public/device_tree_defs.h include/public/version.h include/public/memory.h include/public/features.h include/public/sched.h include/public/xen-compat.h include/public/callback.h include/public/vm_event.h include/public/grant_table.h include/public/physdev.h include/public/tmem.h include/public/hypfs.h include/public/platform.h include/public/pmu.h include/public/elfnote.h include/public/trace.h include/public/event_channel.h include/public/io/vscsiif.h include/public/io/kbdif.h include/public/io/protocols.h include/public/io/ring.h include/public/io/displif.h include/public/io/fsif.h include/public/io/blkif.h include/public/io/console.h include/public/io/sndif.h include/public/io/fbif.h include/public/io/libxenvchan.h include/public/io/netif.h include/public/io/usbif.h include/public/io/pciif.h include/public/io/tpmif.h include/public/io/xs_wire.h include/public/io/xenbus.h include/public/io/cameraif.h include/public/hvm/pvdrivers.h include/public/hvm/e820.h include/public/hvm/hvm_xs_strings.h include/public/hvm/dm_op.h include/public/hvm/ioreq.h include/public/hvm/hvm_info_table.h include/public/hvm/hvm_vcpu.h include/public/hvm/hvm_op.h include/public/hvm/params.h; do x86_64-poky-linux-gcc --sysroot=/home/bermar01/Development/xen-dev/build/profile-qemu-x86_64.prj/tmp/work/core2-64-poky-linux/xen/4.17+git1-r0/recipe-sysroot -x c -ansi -Wall -Werror -include stdint.h -S -o /dev/null $i || exit 1; echo $i; done >include/headers.chk.new; mv include/headers.chk.new include/headers.chk >> | rm -f include/headers99.chk.new; echo "#include "\"include/public/io/9pfs.h\" | x86_64-poky-linux-gcc --sysroot=/home/bermar01/Development/xen-dev/build/profile-qemu-x86_64.prj/tmp/work/core2-64-poky-linux/xen/4.17+git1-r0/recipe-sysroot -x c -std=c99 -Wall -Werror -include stdint.h -include string.h -S -o /dev/null - || exit $?; echo include/public/io/9pfs.h >> include/headers99.chk.new; echo "#include "\"include/public/io/pvcalls.h\" | x86_64-poky-linux-gcc --sysroot=/home/bermar01/Development/xen-dev/build/profile-qemu-x86_64.prj/tmp/work/core2-64-poky-linux/xen/4.17+git1-r0/recipe-sysroot -x c -std=c99 -Wall -Werror -include stdint.h -include string.h -S -o /dev/null - || exit $?; echo include/public/io/pvcalls.h >> include/headers99.chk.new; mv include/headers99.chk.new include/headers99.chk >> | make[9]: execvp: /bin/sh: Argument list too long >> | make[9]: *** [include/Makefile:181: include/headers++.chk] Error 127 >> | make[9]: *** Waiting for unfinished jobs.... >> >> So the command passed to the sub shell by make is quite long. >> >> No idea why this comes out only when building in Yocto but I will dig a bit. > > Maybe Yocto has an unusually low limit on command arguments' total size? > The whole thing is just over 2500 chars, which doesn't look to be unusually > long for Unix-like environments. > Actually the command to generate headers++.chk is 15294 characters when I build xen normally. In Yocto it becomes a lot bigger as CC includes a sysroot parameter with a path. Bertrand
On Thu, Jun 09, 2022 at 11:51:20AM +0000, Bertrand Marquis wrote: > Hi, > > > On 9 Jun 2022, at 11:26, Jan Beulich <jbeulich@suse.com> wrote: > > > > On 09.06.2022 12:16, Bertrand Marquis wrote: > >> With this change, compiling for x86 is now ending up in: > >> CHK include/headers99.chk > >> make[9]: execvp: /bin/sh: Argument list too long > >> make[9]: *** [include/Makefile:181: include/headers++.chk] Error 127 > >> > >> Not quite sure yet why but I wanted to signal it early as other might be impacted. > >> > >> Arm and arm64 builds are not impacted. > > > > Hmm, that patch has passed the smoke push gate already, so there likely is > > more to it than there being an unconditional issue. I did build-test this > > before pushing, and I've just re-tested on a 2nd system without seeing an > > issue. > > I have the problem only when building using Yocto, I did a normal build and the > issue is not coming. > Will the following patch help? From 0f32f749304b233c0d5574dc6b14f66e8709feba Mon Sep 17 00:00:00 2001 From: Anthony PERARD <anthony.perard@citrix.com> Date: Thu, 9 Jun 2022 13:42:52 +0100 Subject: [XEN PATCH] build,include: rework shell script for headers++.chk The command line generated for headers++.chk by make is quite long, and in some environment it is too long. This issue have been seen in Yocto build environment. Error messages: make[9]: execvp: /bin/sh: Argument list too long make[9]: *** [include/Makefile:181: include/headers++.chk] Error 127 Rework to that we do the foreach loop in shell rather that make to reduce the command line size by a lot. We also need a way to get headers prerequisite for some public headers so we use a switch "case" in shell to be able to do some simple pattern matching. Variables alone in POSIX shell don't allow to work with associative array or variables with "/". Reported-by: Bertrand Marquis <Bertrand.Marquis@arm.com> Fixes: 28e13c7f43 ("build: xen/include: use if_changed") Signed-off-by: Anthony PERARD <anthony.perard@citrix.com> --- xen/include/Makefile | 17 +++++++++++++---- 1 file changed, 13 insertions(+), 4 deletions(-) diff --git a/xen/include/Makefile b/xen/include/Makefile index 6d9bcc19b0..ca5e868f38 100644 --- a/xen/include/Makefile +++ b/xen/include/Makefile @@ -158,13 +158,22 @@ define cmd_headerscxx_chk touch $@.new; \ exit 0; \ fi; \ - $(foreach i, $(filter %.h,$^), \ - echo "#include "\"$(i)\" \ + get_prereq() { \ + case $$1 in \ + $(foreach i, $(filter %.h,$^), \ + $(if $($(patsubst $(srctree)/%,%,$i)-prereq), \ + $(patsubst $(srctree)/%,%,$i)$(close) \ + echo "$(foreach j, $($(patsubst $(srctree)/%,%,$i)-prereq), \ + -include c$(j))";;)) \ + esac; \ + }; \ + for i in $(filter %.h,$^); do \ + echo "#include "\"$$i\" \ | $(CXX) -x c++ -std=gnu++98 -Wall -Werror -D__XEN_TOOLS__ \ -include stdint.h -include $(srcdir)/public/xen.h \ - $(foreach j, $($(patsubst $(srctree)/%,%,$i)-prereq), -include c$(j)) \ + `get_prereq $$i` \ -S -o /dev/null - \ - || exit $$?; echo $(i) >> $@.new;) \ + || exit $$?; echo $$i >> $@.new; done; \ mv $@.new $@ endef
Hi Antony, > On 9 Jun 2022, at 13:55, Anthony PERARD <anthony.perard@citrix.com> wrote: > > On Thu, Jun 09, 2022 at 11:51:20AM +0000, Bertrand Marquis wrote: >> Hi, >> >>> On 9 Jun 2022, at 11:26, Jan Beulich <jbeulich@suse.com> wrote: >>> >>> On 09.06.2022 12:16, Bertrand Marquis wrote: >>>> With this change, compiling for x86 is now ending up in: >>>> CHK include/headers99.chk >>>> make[9]: execvp: /bin/sh: Argument list too long >>>> make[9]: *** [include/Makefile:181: include/headers++.chk] Error 127 >>>> >>>> Not quite sure yet why but I wanted to signal it early as other might be impacted. >>>> >>>> Arm and arm64 builds are not impacted. >>> >>> Hmm, that patch has passed the smoke push gate already, so there likely is >>> more to it than there being an unconditional issue. I did build-test this >>> before pushing, and I've just re-tested on a 2nd system without seeing an >>> issue. >> >> I have the problem only when building using Yocto, I did a normal build and the >> issue is not coming. >> > > Will the following patch help? Yes it does, thanks a lot. You can add my: Reviewed-by: Bertrand Marquis <bertrand.marquis@arm.com> Cheers Bertrand > > > From 0f32f749304b233c0d5574dc6b14f66e8709feba Mon Sep 17 00:00:00 2001 > From: Anthony PERARD <anthony.perard@citrix.com> > Date: Thu, 9 Jun 2022 13:42:52 +0100 > Subject: [XEN PATCH] build,include: rework shell script for headers++.chk > > The command line generated for headers++.chk by make is quite long, > and in some environment it is too long. This issue have been seen in > Yocto build environment. > > Error messages: > make[9]: execvp: /bin/sh: Argument list too long > make[9]: *** [include/Makefile:181: include/headers++.chk] Error 127 > > Rework to that we do the foreach loop in shell rather that make to > reduce the command line size by a lot. We also need a way to get > headers prerequisite for some public headers so we use a switch "case" > in shell to be able to do some simple pattern matching. Variables > alone in POSIX shell don't allow to work with associative array or > variables with "/". > > Reported-by: Bertrand Marquis <Bertrand.Marquis@arm.com> > Fixes: 28e13c7f43 ("build: xen/include: use if_changed") > Signed-off-by: Anthony PERARD <anthony.perard@citrix.com> > --- > xen/include/Makefile | 17 +++++++++++++---- > 1 file changed, 13 insertions(+), 4 deletions(-) > > diff --git a/xen/include/Makefile b/xen/include/Makefile > index 6d9bcc19b0..ca5e868f38 100644 > --- a/xen/include/Makefile > +++ b/xen/include/Makefile > @@ -158,13 +158,22 @@ define cmd_headerscxx_chk > touch $@.new; \ > exit 0; \ > fi; \ > - $(foreach i, $(filter %.h,$^), \ > - echo "#include "\"$(i)\" \ > + get_prereq() { \ > + case $$1 in \ > + $(foreach i, $(filter %.h,$^), \ > + $(if $($(patsubst $(srctree)/%,%,$i)-prereq), \ > + $(patsubst $(srctree)/%,%,$i)$(close) \ > + echo "$(foreach j, $($(patsubst $(srctree)/%,%,$i)-prereq), \ > + -include c$(j))";;)) \ > + esac; \ > + }; \ > + for i in $(filter %.h,$^); do \ > + echo "#include "\"$$i\" \ > | $(CXX) -x c++ -std=gnu++98 -Wall -Werror -D__XEN_TOOLS__ \ > -include stdint.h -include $(srcdir)/public/xen.h \ > - $(foreach j, $($(patsubst $(srctree)/%,%,$i)-prereq), -include c$(j)) \ > + `get_prereq $$i` \ > -S -o /dev/null - \ > - || exit $$?; echo $(i) >> $@.new;) \ > + || exit $$?; echo $$i >> $@.new; done; \ > mv $@.new $@ > endef > > > > > -- > Anthony PERARD
Hi Anthony, > On 9 Jun 2022, at 13:55, Anthony PERARD <anthony.perard@citrix.com> wrote: > > On Thu, Jun 09, 2022 at 11:51:20AM +0000, Bertrand Marquis wrote: >> Hi, >> >>> On 9 Jun 2022, at 11:26, Jan Beulich <jbeulich@suse.com> wrote: >>> >>> On 09.06.2022 12:16, Bertrand Marquis wrote: >>>> With this change, compiling for x86 is now ending up in: >>>> CHK include/headers99.chk >>>> make[9]: execvp: /bin/sh: Argument list too long >>>> make[9]: *** [include/Makefile:181: include/headers++.chk] Error 127 >>>> >>>> Not quite sure yet why but I wanted to signal it early as other might be impacted. >>>> >>>> Arm and arm64 builds are not impacted. >>> >>> Hmm, that patch has passed the smoke push gate already, so there likely is >>> more to it than there being an unconditional issue. I did build-test this >>> before pushing, and I've just re-tested on a 2nd system without seeing an >>> issue. >> >> I have the problem only when building using Yocto, I did a normal build and the >> issue is not coming. >> > > Will the following patch help? > > > From 0f32f749304b233c0d5574dc6b14f66e8709feba Mon Sep 17 00:00:00 2001 > From: Anthony PERARD <anthony.perard@citrix.com> > Date: Thu, 9 Jun 2022 13:42:52 +0100 > Subject: [XEN PATCH] build,include: rework shell script for headers++.chk > > The command line generated for headers++.chk by make is quite long, > and in some environment it is too long. This issue have been seen in > Yocto build environment. > > Error messages: > make[9]: execvp: /bin/sh: Argument list too long > make[9]: *** [include/Makefile:181: include/headers++.chk] Error 127 > > Rework to that we do the foreach loop in shell rather that make to > reduce the command line size by a lot. We also need a way to get > headers prerequisite for some public headers so we use a switch "case" > in shell to be able to do some simple pattern matching. Variables > alone in POSIX shell don't allow to work with associative array or > variables with "/". > > Reported-by: Bertrand Marquis <Bertrand.Marquis@arm.com> > Fixes: 28e13c7f43 ("build: xen/include: use if_changed") > Signed-off-by: Anthony PERARD <anthony.perard@citrix.com> > --- > xen/include/Makefile | 17 +++++++++++++---- > 1 file changed, 13 insertions(+), 4 deletions(-) > > diff --git a/xen/include/Makefile b/xen/include/Makefile > index 6d9bcc19b0..ca5e868f38 100644 > --- a/xen/include/Makefile > +++ b/xen/include/Makefile > @@ -158,13 +158,22 @@ define cmd_headerscxx_chk > touch $@.new; \ > exit 0; \ > fi; \ > - $(foreach i, $(filter %.h,$^), \ > - echo "#include "\"$(i)\" \ > + get_prereq() { \ > + case $$1 in \ > + $(foreach i, $(filter %.h,$^), \ > + $(if $($(patsubst $(srctree)/%,%,$i)-prereq), \ > + $(patsubst $(srctree)/%,%,$i)$(close) \ > + echo "$(foreach j, $($(patsubst $(srctree)/%,%,$i)-prereq), \ > + -include c$(j))";;)) \ > + esac; \ > + }; \ > + for i in $(filter %.h,$^); do \ > + echo "#include "\"$$i\" \ > | $(CXX) -x c++ -std=gnu++98 -Wall -Werror -D__XEN_TOOLS__ \ > -include stdint.h -include $(srcdir)/public/xen.h \ > - $(foreach j, $($(patsubst $(srctree)/%,%,$i)-prereq), -include c$(j)) \ > + `get_prereq $$i` \ > -S -o /dev/null - \ > - || exit $$?; echo $(i) >> $@.new;) \ > + || exit $$?; echo $$i >> $@.new; done; \ > mv $@.new $@ > endef > Just a small reminder that you need to push this patch officially :-) Cheers Bertrand
diff --git a/xen/include/Makefile b/xen/include/Makefile index 03baf10efb..6d9bcc19b0 100644 --- a/xen/include/Makefile +++ b/xen/include/Makefile @@ -45,38 +45,65 @@ public-$(CONFIG_ARM) := $(wildcard $(srcdir)/public/arch-arm/*.h $(srcdir)/publi .PHONY: all all: $(addprefix $(obj)/,$(headers-y)) -$(obj)/compat/%.h: $(obj)/compat/%.i $(srcdir)/Makefile $(srctree)/tools/compat-build-header.py - $(PYTHON) $(srctree)/tools/compat-build-header.py <$< $(patsubst $(obj)/%,%,$@) >>$@.new; \ - mv -f $@.new $@ - -$(obj)/compat/%.i: $(obj)/compat/%.c $(srcdir)/Makefile - $(CPP) $(filter-out -Wa$(comma)% -include %/include/xen/config.h,$(XEN_CFLAGS)) $(cppflags-y) -o $@ $< - -$(obj)/compat/%.c: $(src)/public/%.h $(srcdir)/xlat.lst $(srcdir)/Makefile $(srctree)/tools/compat-build-source.py - mkdir -p $(@D) - $(PYTHON) $(srctree)/tools/compat-build-source.py $(srcdir)/xlat.lst <$< >$@.new - mv -f $@.new $@ - -$(obj)/compat/.xlat/%.h: $(obj)/compat/%.h $(obj)/compat/.xlat/%.lst $(srctree)/tools/get-fields.sh $(srcdir)/Makefile - export PYTHON=$(PYTHON); \ - while read what name; do \ - $(SHELL) $(srctree)/tools/get-fields.sh "$$what" compat_$$name $< || exit $$?; \ - done <$(patsubst $(obj)/compat/%,$(obj)/compat/.xlat/%,$(basename $<)).lst >$@.new - mv -f $@.new $@ +quiet_cmd_compat_h = GEN $@ +cmd_compat_h = \ + $(PYTHON) $(srctree)/tools/compat-build-header.py <$< $(patsubst $(obj)/%,%,$@) >>$@.new; \ + mv -f $@.new $@ + +quiet_cmd_compat_i = CPP $@ +cmd_compat_i = $(CPP) $(filter-out -Wa$(comma)% -include %/include/xen/config.h,$(XEN_CFLAGS)) $(cppflags-y) -o $@ $< + +quiet_cmd_compat_c = GEN $@ +cmd_compat_c = \ + $(PYTHON) $(srctree)/tools/compat-build-source.py $(srcdir)/xlat.lst <$< >$@.new; \ + mv -f $@.new $@ + +quiet_cmd_xlat_headers = GEN $@ +cmd_xlat_headers = \ + while read what name; do \ + $(SHELL) $(srctree)/tools/get-fields.sh "$$what" compat_$$name $< || exit $$?; \ + done <$(patsubst $(obj)/compat/%,$(obj)/compat/.xlat/%,$(basename $<)).lst >$@.new; \ + mv -f $@.new $@ + +targets += $(headers-y) +$(obj)/compat/%.h: $(obj)/compat/%.i $(srctree)/tools/compat-build-header.py FORCE + $(call if_changed,compat_h) + +.PRECIOUS: $(obj)/compat/%.i +targets += $(patsubst %.h, %.i, $(headers-y)) +$(obj)/compat/%.i: $(obj)/compat/%.c FORCE + $(call if_changed,compat_i) + +.PRECIOUS: $(obj)/compat/%.c +targets += $(patsubst %.h, %.c, $(headers-y)) +$(obj)/compat/%.c: $(src)/public/%.h $(srcdir)/xlat.lst $(srctree)/tools/compat-build-source.py FORCE + $(call if_changed,compat_c) + +targets += $(patsubst compat/%, compat/.xlat/%, $(headers-y)) +$(obj)/compat/.xlat/%.h: $(obj)/compat/%.h $(obj)/compat/.xlat/%.lst $(srctree)/tools/get-fields.sh FORCE + $(call if_changed,xlat_headers) + +quiet_cmd_xlat_lst = GEN $@ +cmd_xlat_lst = \ + grep -v '^[[:blank:]]*$(pound)' $< | sed -ne 's,@arch@,$(compat-arch-y),g' -re 's,[[:blank:]]+$*\.h[[:blank:]]*$$,,p' >$@.new; \ + $(call move-if-changed,$@.new,$@) .PRECIOUS: $(obj)/compat/.xlat/%.lst -$(obj)/compat/.xlat/%.lst: $(srcdir)/xlat.lst $(srcdir)/Makefile - mkdir -p $(@D) - grep -v '^[[:blank:]]*#' $< | sed -ne 's,@arch@,$(compat-arch-y),g' -re 's,[[:blank:]]+$*\.h[[:blank:]]*$$,,p' >$@.new - $(call move-if-changed,$@.new,$@) +targets += $(patsubst compat/%.h, compat/.xlat/%.lst, $(headers-y)) +$(obj)/compat/.xlat/%.lst: $(srcdir)/xlat.lst FORCE + $(call if_changed,xlat_lst) xlat-y := $(shell sed -ne 's,@arch@,$(compat-arch-y),g' -re 's,^[?!][[:blank:]]+[^[:blank:]]+[[:blank:]]+,,p' $(srcdir)/xlat.lst | uniq) xlat-y := $(filter $(patsubst compat/%,%,$(headers-y)),$(xlat-y)) -$(obj)/compat/xlat.h: $(addprefix $(obj)/compat/.xlat/,$(xlat-y)) $(obj)/config/auto.conf $(srcdir)/Makefile - cat $(filter %.h,$^) >$@.new +quiet_cmd_xlat_h = GEN $@ +cmd_xlat_h = \ + cat $(filter %.h,$^) >$@.new; \ mv -f $@.new $@ +$(obj)/compat/xlat.h: $(addprefix $(obj)/compat/.xlat/,$(xlat-y)) $(obj)/config/auto.conf FORCE + $(call if_changed,xlat_h) + ifeq ($(XEN_TARGET_ARCH),$(XEN_COMPILE_ARCH)) all: $(obj)/headers.chk $(obj)/headers99.chk $(obj)/headers++.chk @@ -102,27 +129,31 @@ PUBLIC_C99_HEADERS := $(call public-filter-headers,public-c99-headers) $(src)/public/io/9pfs.h-prereq := string $(src)/public/io/pvcalls.h-prereq := string -$(obj)/headers.chk: $(PUBLIC_ANSI_HEADERS) $(srcdir)/Makefile +quiet_cmd_header_chk = CHK $@ +cmd_header_chk = \ for i in $(filter %.h,$^); do \ $(CC) -x c -ansi -Wall -Werror -include stdint.h \ -S -o /dev/null $$i || exit 1; \ echo $$i; \ - done >$@.new + done >$@.new; \ mv $@.new $@ -$(obj)/headers99.chk: $(PUBLIC_C99_HEADERS) $(srcdir)/Makefile - rm -f $@.new +quiet_cmd_headers99_chk = CHK $@ +define cmd_headers99_chk + rm -f $@.new; \ $(foreach i, $(filter %.h,$^), \ echo "#include "\"$(i)\" \ | $(CC) -x c -std=c99 -Wall -Werror \ -include stdint.h \ $(foreach j, $($(patsubst $(srctree)/%,%,$i)-prereq), -include $(j).h) \ -S -o /dev/null - \ - || exit $$?; echo $(i) >> $@.new;) + || exit $$?; echo $(i) >> $@.new;) \ mv $@.new $@ +endef -$(obj)/headers++.chk: $(PUBLIC_HEADERS) $(srcdir)/Makefile - rm -f $@.new +quiet_cmd_headerscxx_chk = CHK $@ +define cmd_headerscxx_chk + rm -f $@.new; \ if ! $(CXX) -v >/dev/null 2>&1; then \ touch $@.new; \ exit 0; \ @@ -133,8 +164,21 @@ $(obj)/headers++.chk: $(PUBLIC_HEADERS) $(srcdir)/Makefile -include stdint.h -include $(srcdir)/public/xen.h \ $(foreach j, $($(patsubst $(srctree)/%,%,$i)-prereq), -include c$(j)) \ -S -o /dev/null - \ - || exit $$?; echo $(i) >> $@.new;) + || exit $$?; echo $(i) >> $@.new;) \ mv $@.new $@ +endef + +targets += headers.chk +$(obj)/headers.chk: $(PUBLIC_ANSI_HEADERS) FORCE + $(call if_changed,header_chk) + +targets += headers99.chk +$(obj)/headers99.chk: $(PUBLIC_C99_HEADERS) FORCE + $(call if_changed,headers99_chk) + +targets += headers++.chk +$(obj)/headers++.chk: $(PUBLIC_HEADERS) FORCE + $(call if_changed,headerscxx_chk) endif
Use "define" for the headers*_chk commands as otherwise the "#" is interpreted as a comment and make can't find the end of $(foreach,). Adding several .PRECIOUS as without them `make` deletes the intermediate targets. This is an issue because the macro $(if_changed,) check if the target exist in order to decide whether to recreate the target. Removing the call to `mkdir` from the commands. Those aren't needed anymore because a rune in Rules.mk creates the directory for each $(targets). Remove "export PYTHON" as it is already exported. Signed-off-by: Anthony PERARD <anthony.perard@citrix.com> --- xen/include/Makefile | 108 ++++++++++++++++++++++++++++++------------- 1 file changed, 76 insertions(+), 32 deletions(-)