Re: linux-next: build failure after merge of the kbuild tree

Re: linux-next: build failure after merge of the kbuild tree

[Michal Marek]

On Wed, Apr 20, 2011 at 08:25:14AM +1000, Stephen Rothwell wrote:
> Hi Michal,
> 
> After merging the kbuild tree, today's linux-next build (s390 allnoconfig)
> failed like this:
> 
> /opt/crosstool/gcc-4.3.2-nolibc/bin/s390x-linux-ar: illegal option -- D
> 
> Caused by commit 09ff9fecc039 ("kbuild: Use the deterministic mode of ar").
> 
> $ /opt/crosstool/gcc-4.3.2-nolibc/bin/s390x-linux-ar --version
> GNU ar (GNU Binutils) 2.19
> 
> I presume that this version of ar is now too old to build the kernel?
> Documentation/Changes only mentions Binutils verion 2.12.

Sorry about that, the following patch should fix it.

Michal


Subject: [PATCH] kbuild: Fix build with binutils <= 2.19

The D option of ar is only available in newer versions.

Signed-off-by: Michal Marek <mmarek@suse.cz>

diff --git a/Documentation/kbuild/makefiles.txt b/Documentation/kbuild/makefiles.txt
index 5d145bb..40e082b 100644
--- a/Documentation/kbuild/makefiles.txt
+++ b/Documentation/kbuild/makefiles.txt
@@ -955,6 +955,11 @@ When kbuild executes, the following steps are followed (roughly):
 	used when linking modules. This is often a linker script.
 	From commandline LDFLAGS_MODULE shall be used (see kbuild.txt).
 
+    KBUILD_ARFLAGS   Options for $(AR) when creating archives
+
+	$(KBUILD_ARFLAGS) set by the top level Makefile to "D" (deterministic
+	mode) if this option is supported by $(AR).
+
 --- 6.2 Add prerequisites to archprepare:
 
 	The archprepare: rule is used to list prerequisites that need to be
diff --git a/Makefile b/Makefile
index c3bd316..d7c4231 100644
--- a/Makefile
+++ b/Makefile
@@ -382,6 +382,7 @@ export KBUILD_CFLAGS CFLAGS_KERNEL CFLAGS_MODULE CFLAGS_GCOV
 export KBUILD_AFLAGS AFLAGS_KERNEL AFLAGS_MODULE
 export KBUILD_AFLAGS_MODULE KBUILD_CFLAGS_MODULE KBUILD_LDFLAGS_MODULE
 export KBUILD_AFLAGS_KERNEL KBUILD_CFLAGS_KERNEL
+export KBUILD_ARFLAGS
 
 # When compiling out-of-tree modules, put MODVERDIR in the module
 # tree rather than in the kernel tree. The kernel tree might
@@ -612,6 +613,9 @@ KBUILD_CFLAGS	+= $(call cc-option,-fno-strict-overflow)
 # conserve stack if available
 KBUILD_CFLAGS   += $(call cc-option,-fconserve-stack)
 
+# use the deterministic mode of AR if available
+KBUILD_ARFLAGS := $(call ar-option,D)
+
 # check for 'asm goto'
 ifeq ($(shell $(CONFIG_SHELL) $(srctree)/scripts/gcc-goto.sh $(CC)), y)
 	KBUILD_CFLAGS += -DCC_HAVE_ASM_GOTO
diff --git a/scripts/Kbuild.include b/scripts/Kbuild.include
index ed2773e..fae2d8d 100644
--- a/scripts/Kbuild.include
+++ b/scripts/Kbuild.include
@@ -141,6 +141,11 @@ cc-ldoption = $(call try-run,\
 ld-option = $(call try-run,\
 	$(CC) /dev/null -c -o "$$TMPO" ; $(LD) $(1) "$$TMPO" -o "$$TMP",$(1),$(2))
 
+# ar-option
+# Usage: KBUILD_ARFLAGS := $(call ar-option,D)
+# Important: no spaces around options
+ar-option = $(call try-run, $(AR) rc$(1) "$$TMP",$(1),$(2))
+
 ######
 
 ###
diff --git a/scripts/Makefile.build b/scripts/Makefile.build
index 5f87d37..e1244ef 100644
--- a/scripts/Makefile.build
+++ b/scripts/Makefile.build
@@ -345,7 +345,7 @@ quiet_cmd_link_o_target = LD      $@
 cmd_link_o_target = $(if $(strip $(obj-y)),\
 		      $(LD) $(ld_flags) -r -o $@ $(filter $(obj-y), $^) \
 		      $(cmd_secanalysis),\
-		      rm -f $@; $(AR) rcsD $@)
+		      rm -f $@; $(AR) rcs$(KBUILD_ARFLAGS) $@)
 
 $(builtin-target): $(obj-y) FORCE
 	$(call if_changed,link_o_target)
@@ -371,7 +371,7 @@ $(modorder-target): $(subdir-ym) FORCE
 #
 ifdef lib-target
 quiet_cmd_link_l_target = AR      $@
-cmd_link_l_target = rm -f $@; $(AR) rcsD $@ $(lib-y)
+cmd_link_l_target = rm -f $@; $(AR) rcs$(KBUILD_ARFLAGS) $@ $(lib-y)
 
 $(lib-target): $(lib-y) FORCE
 	$(call if_changed,link_l_target)

linux-next: build failure after merge of the kbuild tree

[Stephen Rothwell]

Hi all,

After merging the dax-misc tree, today's linux-next build (powerpc
ppc44x_defconfig) failed like this:

arch/powerpc/kernel/built-in.o: In function `__ksymtab_clear_pages':
(___ksymtab+clear_pages+0x0): undefined reference to `clear_pages'

Caused by commit

  9445aa1a3062 ("ppc: move exports to definitions")

I applied the following fix patch:

From: Stephen Rothwell <sfr@canb.auug.org.au>
Date: Thu, 15 Sep 2016 16:04:12 +1000
Subject: [PATCH] ppc: there is no clear_pages to export

Fixes: 9445aa1a3062 ("ppc: move exports to definitions")
Signed-off-by: Stephen Rothwell <sfr@canb.auug.org.au>
---
 arch/powerpc/kernel/misc_32.S | 1 -
 1 file changed, 1 deletion(-)

diff --git a/arch/powerpc/kernel/misc_32.S b/arch/powerpc/kernel/misc_32.S
index ea26a9181586..f5156105c5f3 100644
--- a/arch/powerpc/kernel/misc_32.S
+++ b/arch/powerpc/kernel/misc_32.S
@@ -453,7 +453,6 @@ END_FTR_SECTION_IFSET(CPU_FTR_COHERENT_ICACHE)
 	stw	r7,8(r3);	\
 	stw	r8,12(r3);	\
 	stwu	r9,16(r3)
-EXPORT_SYMBOL(clear_pages)
 
 _GLOBAL(copy_page)
 	addi	r3,r3,-4
-- 
2.8.1

-- 
Cheers,
Stephen Rothwell

Re: linux-next: build failure after merge of the kbuild tree

[Michal Marek]

On 2016-09-15 08:08, Stephen Rothwell wrote:
> Hi all,
> 
> After merging the dax-misc tree, today's linux-next build (powerpc
> ppc44x_defconfig) failed like this:
> 
> arch/powerpc/kernel/built-in.o: In function `__ksymtab_clear_pages':
> (___ksymtab+clear_pages+0x0): undefined reference to `clear_pages'
> 
> Caused by commit
> 
>   9445aa1a3062 ("ppc: move exports to definitions")
> 
> I applied the following fix patch:
> 
> From: Stephen Rothwell <sfr@canb.auug.org.au>
> Date: Thu, 15 Sep 2016 16:04:12 +1000
> Subject: [PATCH] ppc: there is no clear_pages to export
> 
> Fixes: 9445aa1a3062 ("ppc: move exports to definitions")
> Signed-off-by: Stephen Rothwell <sfr@canb.auug.org.au>

Applied to kbuild.git#kbuild now.

Thanks,
Michal

linux-next: build failure after merge of the kbuild tree

[Stephen Rothwell]

[-- Attachment #1: Type: text/plain, Size: 991 bytes --]

Hi all,

After merging the kbuild tree, today's linux-next build (powerpc
allyesconfig) failed like this:

/usr/bin/ld: /usr/lib/gcc/powerpc64le-linux-gnu/14/../../../powerpc64le-linux-gnu/libgcc_s.so.1: error adding symbols: file in wrong format
collect2: error: ld returned 1 exit status
make[5]: *** [scripts/Makefile.userprogs:29: samples/pfsm/pfsm-wakeup] Error 1
make[5]: Target 'samples/pfsm/' not remade because of errors.

an so on for lots more ...

Caused by commit

  478494044bb4 ("kbuild: userprogs: also inherit byte order and ABI from kernel")

I have reverted that commit for today.

Build was done on ppc64le, POWER9

$ gcc --version
gcc (Debian 14.2.0-19) 14.2.0
$ ld --version
GNU ld (GNU Binutils for Debian) 2.44

CONFIG_VDSO32=y
CONFIG_CPU_BIG_ENDIAN=y
# CONFIG_CPU_LITTLE_ENDIAN is not set
CONFIG_PPC64_ELF_ABI_V2=y
CONFIG_CC_HAS_ELFV2=y
CONFIG_CC_HAS_PREFIXED=y
CONFIG_CC_HAS_PCREL=y
CONFIG_64BIT=y

-- 
Cheers,
Stephen Rothwell

[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 488 bytes --]

Re: linux-next: build failure after merge of the kbuild tree

[Nathan Chancellor]

On Mon, Aug 18, 2025 at 02:01:43PM +1000, Stephen Rothwell wrote:
> Hi all,
> 
> After merging the kbuild tree, today's linux-next build (powerpc
> allyesconfig) failed like this:
> 
> /usr/bin/ld: /usr/lib/gcc/powerpc64le-linux-gnu/14/../../../powerpc64le-linux-gnu/libgcc_s.so.1: error adding symbols: file in wrong format
> collect2: error: ld returned 1 exit status
> make[5]: *** [scripts/Makefile.userprogs:29: samples/pfsm/pfsm-wakeup] Error 1
> make[5]: Target 'samples/pfsm/' not remade because of errors.
> 
> an so on for lots more ...
> 
> Caused by commit
> 
>   478494044bb4 ("kbuild: userprogs: also inherit byte order and ABI from kernel")
> 
> I have reverted that commit for today.
> 
> Build was done on ppc64le, POWER9
> 
> $ gcc --version
> gcc (Debian 14.2.0-19) 14.2.0
> $ ld --version
> GNU ld (GNU Binutils for Debian) 2.44
> 
> CONFIG_VDSO32=y
> CONFIG_CPU_BIG_ENDIAN=y
> # CONFIG_CPU_LITTLE_ENDIAN is not set
> CONFIG_PPC64_ELF_ABI_V2=y
> CONFIG_CC_HAS_ELFV2=y
> CONFIG_CC_HAS_PREFIXED=y
> CONFIG_CC_HAS_PCREL=y
> CONFIG_64BIT=y

Thanks a lot for the report. Now that I am looking at the test for
CONFIG_CC_CAN_LINK, I am not sure that change is fully correct because
those additional flags are not accounted for in that test. We would need
something like this but if I understand correctly, the filtering of the
flags would need to happen much earlier that it currently does. I have
dropped that change from kbuild-next for now. Thomas, please investigate
and send me a new patch.

Cheers,
Nathan

diff --git a/init/Kconfig b/init/Kconfig
index 836320251219..ab8aa6ab8aa4 100644
--- a/init/Kconfig
+++ b/init/Kconfig
@@ -84,8 +84,8 @@ config RUSTC_LLVM_VERSION
 
 config CC_CAN_LINK
 	bool
-	default $(success,$(srctree)/scripts/cc-can-link.sh $(CC) $(CLANG_FLAGS) $(USERCFLAGS) $(USERLDFLAGS) $(m64-flag)) if 64BIT
-	default $(success,$(srctree)/scripts/cc-can-link.sh $(CC) $(CLANG_FLAGS) $(USERCFLAGS) $(USERLDFLAGS) $(m32-flag))
+	default $(success,$(srctree)/scripts/cc-can-link.sh $(CC) $(CLANG_FLAGS) $(KBUILD_USERCFLAGS) $(KBUILD_USERLDFLAGS) $(m64-flag)) if 64BIT
+	default $(success,$(srctree)/scripts/cc-can-link.sh $(CC) $(CLANG_FLAGS) $(KBUILD_USERCFLAGS) $(KBUILD_USERLDFLAGS) $(m32-flag))
 
 # Fixed in GCC 14, 13.3, 12.4 and 11.5
 # https://gcc.gnu.org/bugzilla/show_bug.cgi?id=113921