Kbuild: move to -std=gnu11
Subsystems (12)
PATCH v3 1/3 Kbuild: move to -std=gnu11
From: Arnd Bergmann <[email protected]>
During a patch discussion, Linus brought up the option of changing the C standard version from gnu89 to gnu99, which allows using variable declaration inside of a for() loop. While the C99, C11 and later standards introduce many other features, most of these are already available in gnu89 as GNU extensions as well.
An earlier attempt to do this when gcc-5 started defaulting to -std=gnu11 failed because at the time that caused warnings about designated initializers with older compilers. Now that gcc-5.1 is the minimum compiler version used for building kernels, that is no longer a concern. Similarly, the behavior of 'inline' functions changes between gnu89 using gnu_inline behavior and gnu11 using standard c99+ behavior, but this was taken care of by defining 'inline' to include __attribute__((gnu_inline)) in order to allow building with clang a while ago.
One minor issue that remains is an added gcc warning for shifts of negative integers when building with -Wextra, which happens with the 'make W=1' option, as well as for three drivers in the kernel that always enable -Wextra, but it was only observed with the i915 driver so far. To be on the safe side, add -Wno-shift-negative-value to any -Wextra in a Makefile.
Nathan Chancellor reported an additional -Wdeclaration-after-statement warning that appears in a system header on arm, this still needs a workaround.
The differences between gnu99, gnu11, gnu1x and gnu17 are fairly minimal and mainly impact warnings at the -Wpedantic level that the kernel never enables. Between these, gnu11 is the newest version that is supported by all supported compiler versions, though it is only the default on gcc-5, while all other supported versions of gcc or clang default to gnu1x/gnu17.
---
[v3]
- split out USERCFLAGS to a separate patch
- add -Wdeclaration-after-statement patch from Mark Rutland
- leave out C17 reference
- more rewording the descrption
[v2]
- added -std=gnu11 back, rather than just relying on the default
- minor changes to changelog text
--- Documentation/process/programming-language.rst | 6 +++--- .../translations/it_IT/process/programming-language.rst | 4 ++-- .../translations/zh_CN/process/programming-language.rst | 3 +-- .../translations/zh_TW/process/programming-language.rst | 3 +-- Makefile | 4 ++-- arch/arm64/kernel/vdso32/Makefile | 2 +- drivers/gpu/drm/i915/Makefile | 1 + drivers/staging/greybus/tools/Makefile | 3 ++- fs/btrfs/Makefile | 1 + scripts/Makefile.extrawarn | 1 + 10 files changed, 15 insertions(+), 13 deletions(-)
Documentation/process/programming-language.rst +3 -3
@@ -5,9 +5,9 @@ Programming Language The kernel is written in the C programming language [c-language]_. More precisely, the kernel is typically compiled with ``gcc`` [gcc]_ -under ``-std=gnu89`` [gcc-c-dialect-options]_: the GNU dialect of ISO C90 -(including some C99 features). ``clang`` [clang]_ is also supported, see -docs on :ref:`Building Linux with Clang/LLVM <kbuild_llvm>`. +under ``-std=gnu11`` [gcc-c-dialect-options]_: the GNU dialect of ISO C11. +``clang`` [clang]_ is also supported, see docs on +:ref:`Building Linux with Clang/LLVM <kbuild_llvm>`. This dialect contains many extensions to the language [gnu-extensions]_, and many of them are used within the kernel as a matter of course.
Documentation/translations/it_IT/process/programming-language.rst +2 -2
@@ -10,8 +10,8 @@ Linguaggio di programmazione Il kernel è scritto nel linguaggio di programmazione C [it-c-language]_. Più precisamente, il kernel viene compilato con ``gcc`` [it-gcc]_ usando -l'opzione ``-std=gnu89`` [it-gcc-c-dialect-options]_: il dialetto GNU -dello standard ISO C90 (con l'aggiunta di alcune funzionalità da C99). +l'opzione ``-std=gnu11`` [it-gcc-c-dialect-options]_: il dialetto GNU +dello standard ISO C11. Linux supporta anche ``clang`` [it-clang]_, leggete la documentazione :ref:`Building Linux with Clang/LLVM <kbuild_llvm>`.
Documentation/translations/zh_CN/process/programming-language.rst +1 -2
@@ -9,8 +9,7 @@ ============ 内核是用C语言 :ref:`c-language <cn_c-language>` 编写的。更准确地说,内核通常是用 :ref:`gcc <cn_gcc>` -在 ``-std=gnu89`` :ref:`gcc-c-dialect-options <cn_gcc-c-dialect-options>` 下编译的:ISO C90的 GNU 方言( -包括一些C99特性) +在 ``-std=gnu11`` :ref:`gcc-c-dialect-options <cn_gcc-c-dialect-options>` 下编译的:ISO C11的 GNU 方言 这种方言包含对语言 :ref:`gnu-extensions <cn_gnu-extensions>` 的许多扩展,当然,它们许多都在内核中使用。
Documentation/translations/zh_TW/process/programming-language.rst +1 -2
@@ -12,8 +12,7 @@ ============ 內核是用C語言 :ref:`c-language <tw_c-language>` 編寫的。更準確地說,內核通常是用 :ref:`gcc <tw_gcc>` -在 ``-std=gnu89`` :ref:`gcc-c-dialect-options <tw_gcc-c-dialect-options>` 下編譯的:ISO C90的 GNU 方言( -包括一些C99特性) +在 ``-std=gnu11`` :ref:`gcc-c-dialect-options <tw_gcc-c-dialect-options>` 下編譯的:ISO C11的 GNU 方言 這種方言包含對語言 :ref:`gnu-extensions <tw_gnu-extensions>` 的許多擴展,當然,它們許多都在內核中使用。
Makefile +2 -2
@@ -515,7 +515,7 @@ KBUILD_CFLAGS := -Wall -Wundef -Werror=strict-prototypes -Wno-trigraphs \ -fno-strict-aliasing -fno-common -fshort-wchar -fno-PIE \ -Werror=implicit-function-declaration -Werror=implicit-int \ -Werror=return-type -Wno-format-security \ - -std=gnu89 + -std=gnu11 KBUILD_CPPFLAGS := -D__KERNEL__ KBUILD_AFLAGS_KERNEL := KBUILD_CFLAGS_KERNEL :=
@@ -782,7 +782,7 @@ KBUILD_CFLAGS += $(KBUILD_CFLAGS-y) $(CONFIG_CC_IMPLICIT_FALLTHROUGH) ifdef CONFIG_CC_IS_CLANG KBUILD_CPPFLAGS += -Qunused-arguments -# The kernel builds with '-std=gnu89' so use of GNU extensions is acceptable. +# The kernel builds with '-std=gnu11' so use of GNU extensions is acceptable. KBUILD_CFLAGS += -Wno-gnu # CLANG uses a _MergedGlobals as optimization, but this breaks modpost, as the # source of a reference will be _MergedGlobals and not on of the whitelisted names.
arch/arm64/kernel/vdso32/Makefile +1 -1
@@ -68,7 +68,7 @@ VDSO_CFLAGS += -Wall -Wundef -Wstrict-prototypes -Wno-trigraphs \ -fno-strict-aliasing -fno-common \ -Werror-implicit-function-declaration \ -Wno-format-security \ - -std=gnu89 + -std=gnu11 VDSO_CFLAGS += -O2 # Some useful compiler-dependent flags from top-level Makefile VDSO_CFLAGS += $(call cc32-option,-Wdeclaration-after-statement,)
drivers/gpu/drm/i915/Makefile +1 -0
@@ -17,6 +17,7 @@ subdir-ccflags-y += -Wno-unused-parameter subdir-ccflags-y += -Wno-type-limits subdir-ccflags-y += -Wno-missing-field-initializers subdir-ccflags-y += -Wno-sign-compare +subdir-ccflags-y += -Wno-shift-negative-value subdir-ccflags-y += $(call cc-disable-warning, unused-but-set-variable) subdir-ccflags-y += $(call cc-disable-warning, frame-address) subdir-ccflags-$(CONFIG_DRM_I915_WERROR) += -Werror
drivers/staging/greybus/tools/Makefile +2 -1
@@ -12,7 +12,8 @@ CFLAGS += -std=gnu99 -Wall -Wextra -g \ -Wredundant-decls \ -Wcast-align \ -Wsign-compare \ - -Wno-missing-field-initializers + -Wno-missing-field-initializers \ + -Wno-shift-negative-value CC := $(CROSS_COMPILE)gcc
fs/btrfs/Makefile +1 -0
@@ -17,6 +17,7 @@ subdir-ccflags-y += $(condflags) subdir-ccflags-y += -Wno-missing-field-initializers subdir-ccflags-y += -Wno-sign-compare subdir-ccflags-y += -Wno-type-limits +subdir-ccflags-y += -Wno-shift-negative-value obj-$(CONFIG_BTRFS_FS) := btrfs.o
scripts/Makefile.extrawarn +1 -0
@@ -36,6 +36,7 @@ KBUILD_CFLAGS += $(call cc-option, -Wstringop-truncation) KBUILD_CFLAGS += -Wno-missing-field-initializers KBUILD_CFLAGS += -Wno-sign-compare KBUILD_CFLAGS += -Wno-type-limits +KBUILD_CFLAGS += -Wno-shift-negative-value KBUILD_CPPFLAGS += -DKBUILD_EXTRA_WARN1
Signature · 1 line
2.29.2
PATCH 2/3 treewide: use -Wdeclaration-after-statement
From: Mark Rutland <[email protected]>
In a subsequent patch we'll move the kernel from using -std=gnu89 to -std=gnu11, permitting the use of additional C11 features such as for-loop initial declarations.
One contentious aspect of C99 is that it permits mixed declarations and code, and for now at least, it seems preferable to enforce that declarations must come first.
These warnings were already disabled in the kernel itself, but not for KBUILD_USERCFLAGS or the compat VDSO on arch/arm64, which uses a separate set of CFLAGS.
This patch fixes an existing violation in modpost.c, which is not reported because of the missing flag in KBUILD_USERCFLAGS:
| scripts/mod/modpost.c: In function ‘match’: | scripts/mod/modpost.c:837:3: warning: ISO C90 forbids mixed declarations and code [-Wdeclaration-after-statement] | 837 | const char *endp = p + strlen(p) - 1; | | ^~~~~
--- Makefile | 3 ++- arch/arm64/kernel/vdso32/Makefile | 1 + scripts/mod/modpost.c | 4 +++- 3 files changed, 6 insertions(+), 2 deletions(-)
Makefile +2 -1
@@ -432,7 +432,8 @@ HOSTCXX = g++ endif export KBUILD_USERCFLAGS := -Wall -Wmissing-prototypes -Wstrict-prototypes \ - -O2 -fomit-frame-pointer -std=gnu89 + -O2 -fomit-frame-pointer -std=gnu89 \ + -Wdeclaration-after-statement export KBUILD_USERLDFLAGS := KBUILD_HOSTCFLAGS := $(KBUILD_USERCFLAGS) $(HOST_LFS_CFLAGS) $(HOSTCFLAGS)
arch/arm64/kernel/vdso32/Makefile +1 -0
@@ -68,6 +68,7 @@ VDSO_CFLAGS += -Wall -Wundef -Wstrict-prototypes -Wno-trigraphs \ -fno-strict-aliasing -fno-common \ -Werror-implicit-function-declaration \ -Wno-format-security \ + -Wdeclaration-after-statement \ -std=gnu11 VDSO_CFLAGS += -O2 # Some useful compiler-dependent flags from top-level Makefile
scripts/mod/modpost.c +3 -1
@@ -833,8 +833,10 @@ static int match(const char *sym, const char * const pat[]) {const char *p;while (*pat) { +const char *endp; + p = *pat++; -const char *endp = p +strlen (p) -1 ; + endp = p +strlen (p) -1 ;/* "*foo*" */ if (*p =='*' && *endp =='*' ) {
Signature · 1 line
2.29.2
PATCH 3/3 Kbuild: use -std=gnu11 for KBUILD_USERCFLAGS
In reply to Arnd BergmannFrom: Arnd Bergmann <[email protected]>
As we change the C language standard for the kernel from gnu89 to gnu11, it makes sense to also update the version for user space compilation.
Some users have older native compilers than what they use for kernel builds, so I considered using gnu99 as the default version for wider compatibility with gcc-4.6 and earlier.
However, testing with older compilers showed that we already require HOSTCC version 5.1 as well because a lot of host tools include linux/compiler.h that uses __has_attribute():
CC tools/objtool/exec-cmd.o
In file included from tools/include/linux/compiler_types.h:36:0,
from tools/include/linux/compiler.h:5,
from exec-cmd.c:2:tools/include/linux/compiler-gcc.h:19:5: error: "__has_attribute" is not defined [-Werror=undef]
--- Makefile | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
Makefile +1 -1
@@ -432,7 +432,7 @@ HOSTCXX = g++ endif export KBUILD_USERCFLAGS := -Wall -Wmissing-prototypes -Wstrict-prototypes \ - -O2 -fomit-frame-pointer -std=gnu89 \ + -O2 -fomit-frame-pointer -std=gnu11 \ -Wdeclaration-after-statement export KBUILD_USERLDFLAGS :=
Signature · 1 line
2.29.2
Quoted 36 lines · Arnd Bergmann
From: Arnd Bergmann <[email protected]>
During a patch discussion, Linus brought up the option of changing the C standard version from gnu89 to gnu99, which allows using variable declaration inside of a
for()loop. While the C99, C11 and later standards introduce many other features, most of these are already available in gnu89 as GNU extensions as well.An earlier attempt to do this when gcc-5 started defaulting to -std=gnu11 failed because at the time that caused warnings about designated initializers with older compilers. Now that gcc-5.1 is the minimum compiler version used for building kernels, that is no longer a concern. Similarly, the behavior of 'inline' functions changes between gnu89 using
gnu_inlinebehavior and gnu11 using standard c99+ behavior, but this was taken care of by defining 'inline' to include __attribute__((gnu_inline)) in order to allow building with clang a while ago.One minor issue that remains is an added gcc warning for shifts of negative integers when building with -Wextra, which happens with the 'make W=1' option, as well as for three drivers in the kernel that always enable -Wextra, but it was only observed with the i915 driver so far. To be on the safe side, add -Wno-shift-negative-value to any -Wextra in a
Makefile.Nathan Chancellor reported an additional -Wdeclaration-after-statement warning that appears in a system header on arm, this still needs a workaround.
The differences between gnu99, gnu11, gnu1x and gnu17 are fairly minimal and mainly impact warnings at the -Wpedantic level that the kernel never enables. Between these, gnu11 is the newest version that is supported by all supported compiler versions, though it is only the default on gcc-5, while all other supported versions of gcc or clang default to gnu1x/gnu17.
Quoted 16 lines
Link https://lore.kernel.org/lkml/CAHk-=wiyCH7xeHcmiFJ-YgXUy2Jaj7pnkdKpcovt8fYbVFW3TA@mail.gmail.com/ open hereSuggested-by Linus Torvalds <[email protected]>Acked-by Marco Elver <[email protected]>Acked-by Jani Nikula <[email protected]>Acked-by David Sterba <[email protected]>Reviewed-by Alex Shi <[email protected]>Reviewed-by Nick Desaulniers <[email protected]>Reviewed-by Miguel Ojeda <[email protected]>Cc Masahiro Yamada <[email protected]>Cc Mark Rutland <[email protected]>Signed-off-by Arnd Bergmann <[email protected]>
One comment below.
Quoted 135 lines
---
[v3]
- split out USERCFLAGS to a separate patch
- add -Wdeclaration-after-statement patch from Mark Rutland
- leave out C17 reference
- more rewording the descrption[v2]
- added -std=gnu11 back, rather than just relying on the default
- minor changes to changelog text--- Documentation/process/programming-language.rst | 6 +++--- .../translations/it_IT/process/programming-language.rst | 4 ++-- .../translations/zh_CN/process/programming-language.rst | 3 +-- .../translations/zh_TW/process/programming-language.rst | 3 +-- Makefile | 4 ++-- arch/arm64/kernel/vdso32/Makefile | 2 +- drivers/gpu/drm/i915/Makefile | 1 + drivers/staging/greybus/tools/Makefile | 3 ++- fs/btrfs/Makefile | 1 + scripts/Makefile.extrawarn | 1 + 10 files changed, 15 insertions(+), 13 deletions(-)Documentation/process/programming-language.rst +3
-3@@ -5,9 +5,9 @@ Programming Language The kernel is written in the C programming language [c-language]_. More precisely, the kernel is typically compiled with ``gcc`` [gcc]_ -under ``-std=gnu89`` [gcc-c-dialect-options]_: the GNU dialect of ISO C90 -(including some C99 features). ``clang`` [clang]_ is also supported, see -docs on :ref:`Building Linux with Clang/LLVM <kbuild_llvm>`. +under ``-std=gnu11`` [gcc-c-dialect-options]_: the GNU dialect of ISO C11. +``clang`` [clang]_ is also supported, see docs on +:ref:`Building Linux with Clang/LLVM <kbuild_llvm>`. This dialect contains many extensions to the language [gnu-extensions]_, and many of them are used within the kernel as a matter of course.Documentation/translations/it_IT/process/programming-language.rst +2
-2@@ -10,8 +10,8 @@ Linguaggio di programmazione Il kernel è scritto nel linguaggio di programmazione C [it-c-language]_. Più precisamente, il kernel viene compilato con ``gcc`` [it-gcc]_ usando -l'opzione ``-std=gnu89`` [it-gcc-c-dialect-options]_: il dialetto GNU -dello standard ISO C90 (con l'aggiunta di alcune funzionalità da C99). +l'opzione ``-std=gnu11`` [it-gcc-c-dialect-options]_: il dialetto GNU +dello standard ISO C11. Linux supporta anche ``clang`` [it-clang]_, leggete la documentazione :ref:`Building Linux with Clang/LLVM <kbuild_llvm>`.Documentation/translations/zh_CN/process/programming-language.rst +1
-2@@ -9,8 +9,7 @@ ============ 内核是用C语言 :ref:`c-language <cn_c-language>` 编写的。更准确地说,内核通常是用 :ref:`gcc <cn_gcc>` -在 ``-std=gnu89`` :ref:`gcc-c-dialect-options <cn_gcc-c-dialect-options>` 下编译的:ISO C90的 GNU 方言( -包括一些C99特性) +在 ``-std=gnu11`` :ref:`gcc-c-dialect-options <cn_gcc-c-dialect-options>` 下编译的:ISO C11的 GNU 方言 这种方言包含对语言 :ref:`gnu-extensions <cn_gnu-extensions>` 的许多扩展,当然,它们许多都在内核中使用。Documentation/translations/zh_TW/process/programming-language.rst +1
-2@@ -12,8 +12,7 @@ ============ 內核是用C語言 :ref:`c-language <tw_c-language>` 編寫的。更準確地說,內核通常是用 :ref:`gcc <tw_gcc>` -在 ``-std=gnu89`` :ref:`gcc-c-dialect-options <tw_gcc-c-dialect-options>` 下編譯的:ISO C90的 GNU 方言( -包括一些C99特性) +在 ``-std=gnu11`` :ref:`gcc-c-dialect-options <tw_gcc-c-dialect-options>` 下編譯的:ISO C11的 GNU 方言 這種方言包含對語言 :ref:`gnu-extensions <tw_gnu-extensions>` 的許多擴展,當然,它們許多都在內核中使用。Makefile +2
-2@@ -515,7 +515,7 @@ KBUILD_CFLAGS := -Wall -Wundef -Werror=strict-prototypes -Wno-trigraphs \ -fno-strict-aliasing -fno-common -fshort-wchar -fno-PIE \ -Werror=implicit-function-declaration -Werror=implicit-int \ -Werror=return-type -Wno-format-security \ - -std=gnu89 + -std=gnu11 KBUILD_CPPFLAGS := -D__KERNEL__ KBUILD_AFLAGS_KERNEL := KBUILD_CFLAGS_KERNEL :=@@ -782,7 +782,7 @@ KBUILD_CFLAGS += $(KBUILD_CFLAGS-y) $(CONFIG_CC_IMPLICIT_FALLTHROUGH) ifdef CONFIG_CC_IS_CLANG KBUILD_CPPFLAGS += -Qunused-arguments -# The kernel builds with '-std=gnu89' so use of GNU extensions is acceptable. +# The kernel builds with '-std=gnu11' so use of GNU extensions is acceptable. KBUILD_CFLAGS += -Wno-gnu # CLANG uses a _MergedGlobals as optimization, but this breaks modpost, as the # source of a reference will be _MergedGlobals and not on of the whitelisted names.arch/arm64/kernel/vdso32/Makefile +1
-1@@ -68,7 +68,7 @@ VDSO_CFLAGS += -Wall -Wundef -Wstrict-prototypes -Wno-trigraphs \ -fno-strict-aliasing -fno-common \ -Werror-implicit-function-declaration \ -Wno-format-security \ - -std=gnu89 + -std=gnu11 VDSO_CFLAGS += -O2 # Some useful compiler-dependent flags from top-level Makefile VDSO_CFLAGS += $(call cc32-option,-Wdeclaration-after-statement,)drivers/gpu/drm/i915/Makefile +1
-0@@ -17,6 +17,7 @@ subdir-ccflags-y += -Wno-unused-parameter subdir-ccflags-y += -Wno-type-limits subdir-ccflags-y += -Wno-missing-field-initializers subdir-ccflags-y += -Wno-sign-compare +subdir-ccflags-y += -Wno-shift-negative-value subdir-ccflags-y += $(call cc-disable-warning, unused-but-set-variable) subdir-ccflags-y += $(call cc-disable-warning, frame-address) subdir-ccflags-$(CONFIG_DRM_I915_WERROR) += -Werrordrivers/staging/greybus/tools/Makefile +0
-0
drivers/staging/greybus/tools/Makefile CFLAGS += -std=gnu99 -Wall -Wextra -g \ · line 16
Quoted 6 lines
drivers/staging/greybus/tools/Makefile +2
-1@@ -12,7 +12,8 @@ CFLAGS += -std=gnu99 -Wall -Wextra -g \ -Wredundant-decls \ -Wcast-align \ -Wsign-compare \ - -Wno-missing-field-initializers + -Wno-missing-field-initializers \ + -Wno-shift-negative-value
This Makefile uses its own -std= option because it is a userland tool. Seems like this hunk could just be dropped or the -std= argument could be bumped here too?
Quoted 30 lines
drivers/staging/greybus/tools/Makefile +0
-0CC := $(CROSS_COMPILE)gccfs/btrfs/Makefile +1
-0@@ -17,6 +17,7 @@ subdir-ccflags-y += $(condflags) subdir-ccflags-y += -Wno-missing-field-initializers subdir-ccflags-y += -Wno-sign-compare subdir-ccflags-y += -Wno-type-limits +subdir-ccflags-y += -Wno-shift-negative-value obj-$(CONFIG_BTRFS_FS) := btrfs.oscripts/Makefile.extrawarn +1
-0@@ -36,6 +36,7 @@ KBUILD_CFLAGS += $(call cc-option, -Wstringop-truncation) KBUILD_CFLAGS += -Wno-missing-field-initializers KBUILD_CFLAGS += -Wno-sign-compare KBUILD_CFLAGS += -Wno-type-limits +KBUILD_CFLAGS += -Wno-shift-negative-value KBUILD_CPPFLAGS += -DKBUILD_EXTRA_WARN1Signature · 1 line
2.29.2
Re: [PATCH 2/3] treewide: use -Wdeclaration-after-statement
In reply to Arnd BergmannQuoted 17 lines · Arnd Bergmann
From: Mark Rutland <[email protected]>
In a subsequent patch we'll move the kernel from using
-std=gnu89to-std=gnu11, permitting the use of additional C11 features such as for-loop initial declarations.One contentious aspect of C99 is that it permits mixed declarations and code, and for now at least, it seems preferable to enforce that declarations must come first.
These warnings were already disabled in the kernel itself, but not for
KBUILD_USERCFLAGSor the compat VDSO onarch/arm64, which uses a separate set of CFLAGS.This patch fixes an existing violation in
modpost.c, which is not reported because of the missing flag inKBUILD_USERCFLAGS:
Quoted 8 lines
| scripts/mod/modpost.c: In function ‘match’: | scripts/mod/modpost.c:837:3: warning: ISO C90 forbids mixed declarations and code [-Wdeclaration-after-statement] | 837 | const char *endp = p + strlen(p) - 1; | | ^~~~~Signed-off-by Mark Rutland <[email protected]>[arnd: don't add a duplicate flag to the default set, update changelog]Signed-off-by Arnd Bergmann <[email protected]>
Quoted 52 lines
--- Makefile | 3 ++- arch/arm64/kernel/vdso32/Makefile | 1 + scripts/mod/modpost.c | 4 +++- 3 files changed, 6 insertions(+), 2 deletions(-)Makefile +2
-1@@ -432,7 +432,8 @@ HOSTCXX = g++ endif export KBUILD_USERCFLAGS := -Wall -Wmissing-prototypes -Wstrict-prototypes \ - -O2 -fomit-frame-pointer -std=gnu89 + -O2 -fomit-frame-pointer -std=gnu89 \ + -Wdeclaration-after-statement export KBUILD_USERLDFLAGS := KBUILD_HOSTCFLAGS := $(KBUILD_USERCFLAGS) $(HOST_LFS_CFLAGS) $(HOSTCFLAGS)arch/arm64/kernel/vdso32/Makefile +1
-0@@ -68,6 +68,7 @@ VDSO_CFLAGS += -Wall -Wundef -Wstrict-prototypes -Wno-trigraphs \ -fno-strict-aliasing -fno-common \ -Werror-implicit-function-declaration \ -Wno-format-security \ + -Wdeclaration-after-statement \ -std=gnu11 VDSO_CFLAGS += -O2 # Some useful compiler-dependent flags from top-level Makefilescripts/mod/modpost.c +3
-1@@ -833,8 +833,10 @@ static int match(const char *sym, const char * const pat[]) {const char *p;while (*pat) { +const char *endp; + p = *pat++; -const char *endp = p +strlen (p) -1 ; + endp = p +strlen (p) -1 ;/* "*foo*" */ if (*p =='*' && *endp =='*' ) {Signature · 1 line
2.29.2
Re: [PATCH 3/3] Kbuild: use -std=gnu11 for KBUILD_USERCFLAGS
In reply to Arnd BergmannQuoted 14 lines · Arnd Bergmann
From: Arnd Bergmann <[email protected]>
As we change the C language standard for the kernel from gnu89 to gnu11, it makes sense to also update the version for user space compilation.
Some users have older native compilers than what they use for kernel builds, so I considered using gnu99 as the default version for wider compatibility with gcc-4.6 and earlier.
However, testing with older compilers showed that we already require HOSTCC version 5.1 as well because a lot of host tools include
linux/compiler.hthat uses__has_attribute():
Quoted 5 lines
CC
tools/objtool/exec-cmd.o
In file included fromtools/include/linux/compiler_types.h:36:0,from tools/include/linux/compiler.h:5, from exec-cmd.c:2:
tools/include/linux/compiler-gcc.h:19:5: error: "__has_attribute" is not defined [-Werror=undef]
For what it's worth, I think this is a bug:
https://lore.kernel.org/r/[email protected]/
https://lore.kernel.org/r/[email protected]/
I never got clear feedback on resolving it from the first thread and nobody bothered to pick up the second patch.
Quoted 1 line
Signed-off-by Arnd Bergmann <[email protected]>
Nobody has yelled that loud though, so this is probably fine.
Quoted 21 lines
--- Makefile | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)Makefile +1
-1@@ -432,7 +432,7 @@ HOSTCXX = g++ endif export KBUILD_USERCFLAGS := -Wall -Wmissing-prototypes -Wstrict-prototypes \ - -O2 -fomit-frame-pointer -std=gnu89 \ + -O2 -fomit-frame-pointer -std=gnu11 \ -Wdeclaration-after-statement export KBUILD_USERLDFLAGS :=Signature · 1 line
2.29.2
Re: [PATCH 3/3] Kbuild: use -std=gnu11 for KBUILD_USERCFLAGS
In reply to Arnd BergmannQuoted 15 lines · Arnd Bergmann
From: Arnd Bergmann <[email protected]>
As we change the C language standard for the kernel from gnu89 to gnu11, it makes sense to also update the version for user space compilation.
Some users have older native compilers than what they use for kernel builds, so I considered using gnu99 as the default version for wider compatibility with gcc-4.6 and earlier.
However, testing with older compilers showed that we already require HOSTCC version 5.1 as well because a lot of host tools include
linux/compiler.hthat uses__has_attribute():
Quoted 7 lines
CC
tools/objtool/exec-cmd.o
In file included fromtools/include/linux/compiler_types.h:36:0,from tools/include/linux/compiler.h:5, from exec-cmd.c:2:
tools/include/linux/compiler-gcc.h:19:5: error: "__has_attribute" is not defined [-Werror=undef]Signed-off-by Arnd Bergmann <[email protected]>
Thanks for the patches!
Quoted 20 lines
--- Makefile | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)Makefile +1
-2@@ -432,7 +432,7 @@ HOSTCXX = g++ endif export KBUILD_USERCFLAGS := -Wall -Wmissing-prototypes -Wstrict-prototypes \ - -O2 -fomit-frame-pointer -std=gnu89 \ + -O2 -fomit-frame-pointer -std=gnu11 \ -Wdeclaration-after-statement export KBUILD_USERLDFLAGS := --2.29.2
Signature · 2 lines
Thanks, ~Nick Desaulniers
Re: [PATCH 2/3] treewide: use -Wdeclaration-after-statement
In reply to Arnd BergmannQuoted 18 lines · Arnd Bergmann
From: Mark Rutland <[email protected]>
In a subsequent patch we'll move the kernel from using
-std=gnu89to-std=gnu11, permitting the use of additional C11 features such as for-loop initial declarations.One contentious aspect of C99 is that it permits mixed declarations and code, and for now at least, it seems preferable to enforce that declarations must come first.
These warnings were already disabled in the kernel itself, but not for
KBUILD_USERCFLAGSor the compat VDSO onarch/arm64, which uses a separate set of CFLAGS.This patch fixes an existing violation in
modpost.c, which is not reported because of the missing flag inKBUILD_USERCFLAGS:
Quoted 8 lines
| scripts/mod/modpost.c: In function ‘match’: | scripts/mod/modpost.c:837:3: warning: ISO C90 forbids mixed declarations and code [-Wdeclaration-after-statement] | 837 | const char *endp = p + strlen(p) - 1; | | ^~~~~Signed-off-by Mark Rutland <[email protected]>[arnd: don't add a duplicate flag to the default set, update changelog]Signed-off-by Arnd Bergmann <[email protected]>
Thanks for the patches!
Quoted 51 lines
--- Makefile | 3 ++- arch/arm64/kernel/vdso32/Makefile | 1 + scripts/mod/modpost.c | 4 +++- 3 files changed, 6 insertions(+), 2 deletions(-)Makefile +2
-1@@ -432,7 +432,8 @@ HOSTCXX = g++ endif export KBUILD_USERCFLAGS := -Wall -Wmissing-prototypes -Wstrict-prototypes \ - -O2 -fomit-frame-pointer -std=gnu89 + -O2 -fomit-frame-pointer -std=gnu89 \ + -Wdeclaration-after-statement export KBUILD_USERLDFLAGS := KBUILD_HOSTCFLAGS := $(KBUILD_USERCFLAGS) $(HOST_LFS_CFLAGS) $(HOSTCFLAGS)arch/arm64/kernel/vdso32/Makefile +1
-0@@ -68,6 +68,7 @@ VDSO_CFLAGS += -Wall -Wundef -Wstrict-prototypes -Wno-trigraphs \ -fno-strict-aliasing -fno-common \ -Werror-implicit-function-declaration \ -Wno-format-security \ + -Wdeclaration-after-statement \ -std=gnu11 VDSO_CFLAGS += -O2 # Some useful compiler-dependent flags from top-level Makefilescripts/mod/modpost.c +3
-2@@ -833,8 +833,10 @@ static int match(const char *sym, const char * const pat[]) {const char *p;while (*pat) { +const char *endp; + p = *pat++; -const char *endp = p +strlen (p) -1 ; + endp = p +strlen (p) -1 ;/* "*foo*" */ if (*p =='*' && *endp =='*' ) { --2.29.2
Signature · 2 lines
Thanks, ~Nick Desaulniers
Re: [PATCH 2/3] treewide: use -Wdeclaration-after-statement
Quoted 25 lines · Nick Desaulniers
Quoted 23 lines · Arnd Bergmann
From: Mark Rutland <[email protected]>
In a subsequent patch we'll move the kernel from using
-std=gnu89to-std=gnu11, permitting the use of additional C11 features such as for-loop initial declarations.One contentious aspect of C99 is that it permits mixed declarations and code, and for now at least, it seems preferable to enforce that declarations must come first.
These warnings were already disabled in the kernel itself, but not for
KBUILD_USERCFLAGSor the compat VDSO onarch/arm64, which uses a separate set of CFLAGS.This patch fixes an existing violation in
modpost.c, which is not reported because of the missing flag inKBUILD_USERCFLAGS:| scripts/mod/modpost.c: In function ‘match’: | scripts/mod/modpost.c:837:3: warning: ISO C90 forbids mixed declarations and code [-Wdeclaration-after-statement] | 837 | const char *endp = p + strlen(p) - 1; | | ^~~~~
Quoted 6 lines
Quoted 3 lines
Signed-off-by Mark Rutland <[email protected]>[arnd: don't add a duplicate flag to the default set, update changelog]Signed-off-by Arnd Bergmann <[email protected]>Thanks for the patches!
Reviewed-by Nick Desaulniers <[email protected]>
That said, there's a few additional places that reset KBUILD_CFLAGS.
$ git grep -rn "KBUILD_CFLAGS :=" | grep -v filter-out | grep -v subst arch/mips/boot/compressed/Makefile:30:KBUILD_CFLAGS := $(KBUILD_CFLAGS) -D__KERNEL__ -D__DISABLE_EXPORTS \ arch/mips/vdso/Makefile:115:$(obj-vdso): KBUILD_CFLAGS := $(cflags-vdso) $(native-abi) arch/mips/vdso/Makefile:144:$(obj-vdso-o32): KBUILD_CFLAGS := $(cflags-vdso) -mabi=32 arch/mips/vdso/Makefile:184:$(obj-vdso-n32): KBUILD_CFLAGS := $(cflags-vdso) -mabi=n32 arch/parisc/boot/compressed/Makefile:17:KBUILD_CFLAGS := -D__KERNEL__ -O2 -DBOOTLOADER arch/s390/boot/Makefile:13:KBUILD_CFLAGS := $(KBUILD_CFLAGS_DECOMPRESSOR) arch/s390/boot/compressed/Makefile:23:KBUILD_CFLAGS := $(KBUILD_CFLAGS_DECOMPRESSOR) arch/s390/purgatory/Makefile:24:KBUILD_CFLAGS := -fno-strict-aliasing -Wall -Wstrict-prototypes arch/x86/boot/compressed/Makefile:35:KBUILD_CFLAGS := -m$(BITS) -O2 $(CLANG_FLAGS)
The parisc, s390, and x86 cases look like true positives to me (mips looks fine FWICT). I didn't want to nack the patch for being incomplete, but it's not necessarily treewide.
Quoted 57 lines
Quoted 50 lines
--- Makefile | 3 ++- arch/arm64/kernel/vdso32/Makefile | 1 + scripts/mod/modpost.c | 4 +++- 3 files changed, 6 insertions(+), 2 deletions(-)Makefile +2
-1@@ -432,7 +432,8 @@ HOSTCXX = g++ endif export KBUILD_USERCFLAGS := -Wall -Wmissing-prototypes -Wstrict-prototypes \ - -O2 -fomit-frame-pointer -std=gnu89 + -O2 -fomit-frame-pointer -std=gnu89 \ + -Wdeclaration-after-statement export KBUILD_USERLDFLAGS := KBUILD_HOSTCFLAGS := $(KBUILD_USERCFLAGS) $(HOST_LFS_CFLAGS) $(HOSTCFLAGS)arch/arm64/kernel/vdso32/Makefile +1
-0@@ -68,6 +68,7 @@ VDSO_CFLAGS += -Wall -Wundef -Wstrict-prototypes -Wno-trigraphs \ -fno-strict-aliasing -fno-common \ -Werror-implicit-function-declaration \ -Wno-format-security \ + -Wdeclaration-after-statement \ -std=gnu11 VDSO_CFLAGS += -O2 # Some useful compiler-dependent flags from top-level Makefilescripts/mod/modpost.c +3
-2@@ -833,8 +833,10 @@ static int match(const char *sym, const char * const pat[]) {const char *p;while (*pat) { +const char *endp; + p = *pat++; -const char *endp = p +strlen (p) -1 ; + endp = p +strlen (p) -1 ;/* "*foo*" */ if (*p =='*' && *endp =='*' ) { --Quoted 2 lines
2.29.2
--
Thanks,
~Nick Desaulniers
Signature · 2 lines
Thanks, ~Nick Desaulniers
Re: [PATCH 2/3] treewide: use -Wdeclaration-after-statement
Quoted 2 lines · Nick Desaulniers
That said, there's a few additional places that reset
KBUILD_CFLAGS.
Yeah, the boot code and vdso code often sets its own private flags, because it's _so_ different.
The generic kernel CFLAGS often don't work at all, because that code may be doing some truly horrendous things.
So I agree that this may not catch all the kernel code, but I don't think it's much of a problem.
Linus
Re: [PATCH 2/3] treewide: use -Wdeclaration-after-statement
Quoted 6 lines · Linus Torvalds
Quoted 2 lines · Nick Desaulniers
That said, there's a few additional places that reset
KBUILD_CFLAGS.
Quoted 8 lines
Yeah, the boot code and vdso code often sets its own private flags, because it's _so_ different.
The generic kernel CFLAGS often don't work at all, because that code may be doing some truly horrendous things.
So I agree that this may not catch all the kernel code, but I don't think it's much of a problem.
Agreed; they can still mirror the style imposed by -Wdeclaration-after-statement like the rest of the codebase though.
Signature · 2 lines
Thanks, ~Nick Desaulniers
Re: [PATCH 3/3] Kbuild: use -std=gnu11 for KBUILD_USERCFLAGS
In reply to Arnd BergmannQuoted 7 lines · Arnd Bergmann
From: Arnd Bergmann <[email protected]>
As we change the C language standard for the kernel from gnu89 to gnu11, it makes sense to also update the version for user space compilation.
Overdue.
Can you point me to a lore link where I can easily fetch the full 3/3 v3 patchset?
I would like to give this a try on x86-64 with my latest kernel-config (based on the latest kernel-config v5.17-rc5 from Debian/unstable).
What Linux source base do I need?
Is Linux v5.17-rc6 good enough (patchset applicable) for testing?
Thanks, Arnd.
Regards,
- Sedat -
Quoted 35 lines
Some users have older native compilers than what they use for kernel builds, so I considered using gnu99 as the default version for wider compatibility with gcc-4.6 and earlier.
However, testing with older compilers showed that we already require HOSTCC version 5.1 as well because a lot of host tools include
linux/compiler.hthat uses__has_attribute():CC
tools/objtool/exec-cmd.o
In file included fromtools/include/linux/compiler_types.h:36:0,from tools/include/linux/compiler.h:5, from exec-cmd.c:2:
tools/include/linux/compiler-gcc.h:19:5: error: "__has_attribute" is not defined [-Werror=undef]Signed-off-by Arnd Bergmann <[email protected]>--- Makefile | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)Makefile +1
-2@@ -432,7 +432,7 @@ HOSTCXX = g++ endif export KBUILD_USERCFLAGS := -Wall -Wmissing-prototypes -Wstrict-prototypes \ - -O2 -fomit-frame-pointer -std=gnu89 \ + -O2 -fomit-frame-pointer -std=gnu11 \ -Wdeclaration-after-statement export KBUILD_USERLDFLAGS := --2.29.2
Re: [PATCH 3/3] Kbuild: use -std=gnu11 for KBUILD_USERCFLAGS
Quoted 10 lines · Sedat Dilek
Quoted 7 lines · Arnd Bergmann
From: Arnd Bergmann <[email protected]>
As we change the C language standard for the kernel from gnu89 to gnu11, it makes sense to also update the version for user space compilation.
Quoted 4 lines
Overdue.
Can you point me to a lore link where I can easily fetch the full 3/3 v3 patchset?
Hey Sedat!
$ b4 am https://lore.kernel.org/lkml/[email protected]/ -o - | git am -3
https://people.kernel.org/monsieuricon/introducing-b4-and-patch-attestation
Quoted 6 lines
I would like to give this a try on x86-64 with my latest kernel-config (based on the latest kernel-config v5.17-rc5 from Debian/unstable).
What Linux source base do I need?
Is Linux v5.17-rc6 good enough (patchset applicable) for testing?
I suspect so.
Signature · 2 lines
Thanks, ~Nick Desaulniers
Re: [PATCH 2/3] treewide: use -Wdeclaration-after-statement
In reply to Arnd BergmannQuoted 4 lines · Arnd Bergmann
From: Mark Rutland <[email protected]>
In a subsequent patch we'll move the kernel from using
-std=gnu89to
Precisely speaking,
"In a subsequent patch we'll move the userspace tools ..."
The kernel part was done by 1/3.
Quoted 8 lines
-std=gnu11, permitting the use of additional C11 features such as for-loop initial declarations.One contentious aspect of C99 is that it permits mixed declarations and code, and for now at least, it seems preferable to enforce that declarations must come first.
These warnings were already disabled in the kernel itself, but not
Is this the opposite?
"These warnings were already _enabled_ in the kernel itself, ..."
If you have a chance to resend this,
can you please rebase on top of kbuild/for-next?
This is not applicable due to the conflict with:
https://patchwork.kernel.org/project/linux-kbuild/patch/[email protected]/
Quoted 65 lines
for
KBUILD_USERCFLAGSor the compat VDSO onarch/arm64, which uses a separate set of CFLAGS.This patch fixes an existing violation in
modpost.c, which is not reported because of the missing flag inKBUILD_USERCFLAGS:| scripts/mod/modpost.c: In function ‘match’: | scripts/mod/modpost.c:837:3: warning: ISO C90 forbids mixed declarations and code [-Wdeclaration-after-statement] | 837 | const char *endp = p + strlen(p) - 1; | | ^~~~~Signed-off-by Mark Rutland <[email protected]>[arnd: don't add a duplicate flag to the default set, update changelog]Signed-off-by Arnd Bergmann <[email protected]>--- Makefile | 3 ++- arch/arm64/kernel/vdso32/Makefile | 1 + scripts/mod/modpost.c | 4 +++- 3 files changed, 6 insertions(+), 2 deletions(-)Makefile +2
-1@@ -432,7 +432,8 @@ HOSTCXX = g++ endif export KBUILD_USERCFLAGS := -Wall -Wmissing-prototypes -Wstrict-prototypes \ - -O2 -fomit-frame-pointer -std=gnu89 + -O2 -fomit-frame-pointer -std=gnu89 \ + -Wdeclaration-after-statement export KBUILD_USERLDFLAGS := KBUILD_HOSTCFLAGS := $(KBUILD_USERCFLAGS) $(HOST_LFS_CFLAGS) $(HOSTCFLAGS)arch/arm64/kernel/vdso32/Makefile +1
-0@@ -68,6 +68,7 @@ VDSO_CFLAGS += -Wall -Wundef -Wstrict-prototypes -Wno-trigraphs \ -fno-strict-aliasing -fno-common \ -Werror-implicit-function-declaration \ -Wno-format-security \ + -Wdeclaration-after-statement \ -std=gnu11 VDSO_CFLAGS += -O2 # Some useful compiler-dependent flags from top-level Makefilescripts/mod/modpost.c +3
-2@@ -833,8 +833,10 @@ static int match(const char *sym, const char * const pat[]) {const char *p;while (*pat) { +const char *endp; + p = *pat++; -const char *endp = p +strlen (p) -1 ; + endp = p +strlen (p) -1 ;/* "*foo*" */ if (*p =='*' && *endp =='*' ) { --2.29.2
Signature · 2 lines
Best Regards Masahiro Yamada
Re: [PATCH 3/3] Kbuild: use -std=gnu11 for KBUILD_USERCFLAGS
In reply to Nick DesaulniersQuoted 14 lines · Nick Desaulniers
Quoted 12 lines · Sedat Dilek
Quoted 7 lines · Arnd Bergmann
From: Arnd Bergmann <[email protected]>
As we change the C language standard for the kernel from gnu89 to gnu11, it makes sense to also update the version for user space compilation.
Overdue.
Quoted 8 lines
Quoted 2 lines
Can you point me to a lore link where I can easily fetch the full 3/3 v3 patchset?
Hey Sedat!
$ b4 am https://lore.kernel.org/lkml/[email protected]/ -o - | git am -3
Hey Nick!
This only applies 1/3.
$ b4 --version 0.8.0
$ b4 am https://lore.kernel.org/lkml/[email protected]/ -o - | git am -3 Analyzing 14 messages in the thread Will use the latest revision: v3 You can pick other revisions using the -vN flag Checking attestation on all messages, may take a moment... --- ✓ [PATCH v3 1/3] Kbuild: move to -std=gnu11 ✓ Signed: DKIM/kernel.org + Reviewed-by: Nathan Chancellor <[email protected]> (✓ DKIM/kernel.org) ERROR: missing [2/3]! ERROR: missing [3/3]! --- NOTE: install patatt for end-to-end signature verification --- Total patches: 1 --- WARNING: Thread incomplete! Link: https://lore.kernel.org/r/[email protected] Base: not specified Wende an: Kbuild: move to -std=gnu11
- Sedat -
Quoted 13 lines
https://people.kernel.org/monsieuricon/introducing-b4-and-patch-attestation
Quoted 6 lines
I would like to give this a try on x86-64 with my latest kernel-config (based on the latest kernel-config v5.17-rc5 from Debian/unstable).
What Linux source base do I need?
Is Linux v5.17-rc6 good enough (patchset applicable) for testing?I suspect so.
--
Thanks,
~Nick Desaulniers
Nathan Chancellor
maintainerRe: [PATCH 3/3] Kbuild: use -std=gnu11 for KBUILD_USERCFLAGS
Hi Sedat,
Quoted 20 lines · Sedat Dilek
Hey Nick!
This only applies 1/3.
$ b4 --version 0.8.0$ b4 am https://lore.kernel.org/lkml/[email protected]/ -o - | git am -3 Analyzing 14 messages in the thread Will use the latest revision: v3 You can pick other revisions using the -vN flag Checking attestation on all messages, may take a moment... --- ✓ [PATCH v3 1/3] Kbuild: move to -std=gnu11 ✓ Signed: DKIM/kernel.org + Reviewed-by: Nathan Chancellor <[email protected]> (✓ DKIM/kernel.org) ERROR: missing [2/3]! ERROR: missing [3/3]! ---
Quoted 8 lines
NOTE: install patatt for end-to-end signature verification --- Total patches: 1 --- WARNING: Thread incomplete! Link: https://lore.kernel.org/r/[email protected] Base: not specified Wende an: Kbuild: move to -std=gnu11
It looks like the threading somehow got broken, likely due to the [v3] on the first patch and not the second or third:
This worked for me on v5.17-rc6:
$ for i in $(seq 1 3); do b4 shazam -P _ 20220301145233.3689119-"$i"[email protected]; done
"b4 shazam" is the equivalent of "b4 am -o - ... | git am" and the "-P _" tells b4 to only fetch that exact message ID, not the whole thread.
Cheers,
Nathan
Quoted 196 lines · Nathan Chancellor
Quoted 36 lines · Arnd Bergmann
From: Arnd Bergmann <[email protected]>
During a patch discussion, Linus brought up the option of changing the C standard version from gnu89 to gnu99, which allows using variable declaration inside of a
for()loop. While the C99, C11 and later standards introduce many other features, most of these are already available in gnu89 as GNU extensions as well.An earlier attempt to do this when gcc-5 started defaulting to -std=gnu11 failed because at the time that caused warnings about designated initializers with older compilers. Now that gcc-5.1 is the minimum compiler version used for building kernels, that is no longer a concern. Similarly, the behavior of 'inline' functions changes between gnu89 using
gnu_inlinebehavior and gnu11 using standard c99+ behavior, but this was taken care of by defining 'inline' to include __attribute__((gnu_inline)) in order to allow building with clang a while ago.One minor issue that remains is an added gcc warning for shifts of negative integers when building with -Wextra, which happens with the 'make W=1' option, as well as for three drivers in the kernel that always enable -Wextra, but it was only observed with the i915 driver so far. To be on the safe side, add -Wno-shift-negative-value to any -Wextra in a
Makefile.Nathan Chancellor reported an additional -Wdeclaration-after-statement warning that appears in a system header on arm, this still needs a workaround.
The differences between gnu99, gnu11, gnu1x and gnu17 are fairly minimal and mainly impact warnings at the -Wpedantic level that the kernel never enables. Between these, gnu11 is the newest version that is supported by all supported compiler versions, though it is only the default on gcc-5, while all other supported versions of gcc or clang default to gnu1x/gnu17.
Quoted 16 lines
Link https://lore.kernel.org/lkml/CAHk-=wiyCH7xeHcmiFJ-YgXUy2Jaj7pnkdKpcovt8fYbVFW3TA@mail.gmail.com/ open hereSuggested-by Linus Torvalds <[email protected]>Acked-by Marco Elver <[email protected]>Acked-by Jani Nikula <[email protected]>Acked-by David Sterba <[email protected]>Reviewed-by Alex Shi <[email protected]>Reviewed-by Nick Desaulniers <[email protected]>Reviewed-by Miguel Ojeda <[email protected]>Cc Masahiro Yamada <[email protected]>Cc Mark Rutland <[email protected]>Signed-off-by Arnd Bergmann <[email protected]>Reviewed-by Nathan Chancellor <[email protected]>One comment below.
Quoted 137 lines
---
[v3]
- split out USERCFLAGS to a separate patch
- add -Wdeclaration-after-statement patch from Mark Rutland
- leave out C17 reference
- more rewording the descrption[v2]
- added -std=gnu11 back, rather than just relying on the default
- minor changes to changelog text--- Documentation/process/programming-language.rst | 6 +++--- .../translations/it_IT/process/programming-language.rst | 4 ++-- .../translations/zh_CN/process/programming-language.rst | 3 +-- .../translations/zh_TW/process/programming-language.rst | 3 +-- Makefile | 4 ++-- arch/arm64/kernel/vdso32/Makefile | 2 +- drivers/gpu/drm/i915/Makefile | 1 + drivers/staging/greybus/tools/Makefile | 3 ++- fs/btrfs/Makefile | 1 + scripts/Makefile.extrawarn | 1 + 10 files changed, 15 insertions(+), 13 deletions(-)Documentation/process/programming-language.rst +3
-3@@ -5,9 +5,9 @@ Programming Language The kernel is written in the C programming language [c-language]_. More precisely, the kernel is typically compiled with ``gcc`` [gcc]_ -under ``-std=gnu89`` [gcc-c-dialect-options]_: the GNU dialect of ISO C90 -(including some C99 features). ``clang`` [clang]_ is also supported, see -docs on :ref:`Building Linux with Clang/LLVM <kbuild_llvm>`. +under ``-std=gnu11`` [gcc-c-dialect-options]_: the GNU dialect of ISO C11. +``clang`` [clang]_ is also supported, see docs on +:ref:`Building Linux with Clang/LLVM <kbuild_llvm>`. This dialect contains many extensions to the language [gnu-extensions]_, and many of them are used within the kernel as a matter of course.Documentation/translations/it_IT/process/programming-language.rst +2
-2@@ -10,8 +10,8 @@ Linguaggio di programmazione Il kernel è scritto nel linguaggio di programmazione C [it-c-language]_. Più precisamente, il kernel viene compilato con ``gcc`` [it-gcc]_ usando -l'opzione ``-std=gnu89`` [it-gcc-c-dialect-options]_: il dialetto GNU -dello standard ISO C90 (con l'aggiunta di alcune funzionalità da C99). +l'opzione ``-std=gnu11`` [it-gcc-c-dialect-options]_: il dialetto GNU +dello standard ISO C11. Linux supporta anche ``clang`` [it-clang]_, leggete la documentazione :ref:`Building Linux with Clang/LLVM <kbuild_llvm>`.Documentation/translations/zh_CN/process/programming-language.rst +1
-2@@ -9,8 +9,7 @@ ============ 内核是用C语言 :ref:`c-language <cn_c-language>` 编写的。更准确地说,内核通常是用 :ref:`gcc <cn_gcc>` -在 ``-std=gnu89`` :ref:`gcc-c-dialect-options <cn_gcc-c-dialect-options>` 下编译的:ISO C90的 GNU 方言( -包括一些C99特性) +在 ``-std=gnu11`` :ref:`gcc-c-dialect-options <cn_gcc-c-dialect-options>` 下编译的:ISO C11的 GNU 方言 这种方言包含对语言 :ref:`gnu-extensions <cn_gnu-extensions>` 的许多扩展,当然,它们许多都在内核中使用。Documentation/translations/zh_TW/process/programming-language.rst +1
-2@@ -12,8 +12,7 @@ ============ 內核是用C語言 :ref:`c-language <tw_c-language>` 編寫的。更準確地說,內核通常是用 :ref:`gcc <tw_gcc>` -在 ``-std=gnu89`` :ref:`gcc-c-dialect-options <tw_gcc-c-dialect-options>` 下編譯的:ISO C90的 GNU 方言( -包括一些C99特性) +在 ``-std=gnu11`` :ref:`gcc-c-dialect-options <tw_gcc-c-dialect-options>` 下編譯的:ISO C11的 GNU 方言 這種方言包含對語言 :ref:`gnu-extensions <tw_gnu-extensions>` 的許多擴展,當然,它們許多都在內核中使用。Makefile +2
-2@@ -515,7 +515,7 @@ KBUILD_CFLAGS := -Wall -Wundef -Werror=strict-prototypes -Wno-trigraphs \ -fno-strict-aliasing -fno-common -fshort-wchar -fno-PIE \ -Werror=implicit-function-declaration -Werror=implicit-int \ -Werror=return-type -Wno-format-security \ - -std=gnu89 + -std=gnu11 KBUILD_CPPFLAGS := -D__KERNEL__ KBUILD_AFLAGS_KERNEL := KBUILD_CFLAGS_KERNEL :=@@ -782,7 +782,7 @@ KBUILD_CFLAGS += $(KBUILD_CFLAGS-y) $(CONFIG_CC_IMPLICIT_FALLTHROUGH) ifdef CONFIG_CC_IS_CLANG KBUILD_CPPFLAGS += -Qunused-arguments -# The kernel builds with '-std=gnu89' so use of GNU extensions is acceptable. +# The kernel builds with '-std=gnu11' so use of GNU extensions is acceptable. KBUILD_CFLAGS += -Wno-gnu # CLANG uses a _MergedGlobals as optimization, but this breaks modpost, as the # source of a reference will be _MergedGlobals and not on of the whitelisted names.arch/arm64/kernel/vdso32/Makefile +1
-1@@ -68,7 +68,7 @@ VDSO_CFLAGS += -Wall -Wundef -Wstrict-prototypes -Wno-trigraphs \ -fno-strict-aliasing -fno-common \ -Werror-implicit-function-declaration \ -Wno-format-security \ - -std=gnu89 + -std=gnu11 VDSO_CFLAGS += -O2 # Some useful compiler-dependent flags from top-level Makefile VDSO_CFLAGS += $(call cc32-option,-Wdeclaration-after-statement,)drivers/gpu/drm/i915/Makefile +1
-0@@ -17,6 +17,7 @@ subdir-ccflags-y += -Wno-unused-parameter subdir-ccflags-y += -Wno-type-limits subdir-ccflags-y += -Wno-missing-field-initializers subdir-ccflags-y += -Wno-sign-compare +subdir-ccflags-y += -Wno-shift-negative-value subdir-ccflags-y += $(call cc-disable-warning, unused-but-set-variable) subdir-ccflags-y += $(call cc-disable-warning, frame-address) subdir-ccflags-$(CONFIG_DRM_I915_WERROR) += -Werrordrivers/staging/greybus/tools/Makefile +0
-0@@ -12,7 +12,8 @@ CFLAGS += -std=gnu99 -Wall -Wextra -g \ -Wredundant-decls \ -Wcast-align \
Quoted 9 lines
drivers/staging/greybus/tools/Makefile CFLAGS += -std=gnu99 -Wall -Wextra -g \ · line 16
Quoted 4 lines
drivers/staging/greybus/tools/Makefile +2
-1-Wsign-compare \ - -Wno-missing-field-initializers + -Wno-missing-field-initializers \ + -Wno-shift-negative-valueThis
Makefileuses its own -std= option because it is a userland tool. Seems like this hunk could just be dropped or the -std= argument could be bumped here too?
Right, this is a userland tool.
I do not mind either way.
BTW, this makefile is somewhat corrupted because ../greybus_protocols.h does not exist. (but it is working, relying on the built-in rules)
If Arnd does not mind sending v4, maybe we
can split this into two paches:
1/4 : add -Wno-shift-negative-value where -Wextra is used 2/4 : Kbuild: move to -std=gnu11
1/4 adds -Wno-shift-negative-value
whether it is the kernel space or user-land.
--
Best Regards
Masahiro Yamada
Re: [PATCH 3/3] Kbuild: use -std=gnu11 for KBUILD_USERCFLAGS
In reply to Nathan ChancellorQuoted 36 lines · Nathan Chancellor
Hi Sedat,
Quoted 20 lines · Sedat Dilek
Hey Nick!
This only applies 1/3.
$ b4 --version 0.8.0$ b4 am https://lore.kernel.org/lkml/[email protected]/ -o - | git am -3 Analyzing 14 messages in the thread Will use the latest revision: v3 You can pick other revisions using the -vN flag Checking attestation on all messages, may take a moment... --- ✓ [PATCH v3 1/3] Kbuild: move to -std=gnu11 ✓ Signed: DKIM/kernel.org + Reviewed-by: Nathan Chancellor <[email protected]> (✓ DKIM/kernel.org) ERROR: missing [2/3]! ERROR: missing [3/3]! ---Quoted 8 lines
NOTE: install patatt for end-to-end signature verification --- Total patches: 1 --- WARNING: Thread incomplete! Link: https://lore.kernel.org/r/[email protected] Base: not specified Wende an: Kbuild: move to -std=gnu11It looks like the threading somehow got broken, likely due to the [v3] on the first patch and not the second or third:
Quoted 8 lines
This worked for me on v5.17-rc6:
$ for i in $(seq 1 3); do b4 shazam -P _ 20220301145233.3689119-"$i"[email protected]; done"b4 shazam" is the equivalent of "b4 am -o - ... | git am" and the "-P _" tells b4 to only fetch that exact message ID, not the whole thread.
Hmm, the universe is not good to me...
$ for i in $(seq 1 3); do b4 shazam -P _ 20220301145233.3689119-"$i"[email protected]; done usage: b4 [-h] [--version] [-d] [-q] {mbox,am,attest,pr,ty,diff,kr} ... b4: error: argument subcmd: invalid choice: 'shazam' (choose from 'mbox', 'am', 'attest', 'pr', 'ty', 'diff', 'kr') usage: b4 [-h] [--version] [-d] [-q] {mbox,am,attest,pr,ty,diff,kr} ... b4: error: argument subcmd: invalid choice: 'shazam' (choose from 'mbox', 'am', 'attest', 'pr', 'ty', 'diff', 'kr') usage: b4 [-h] [--version] [-d] [-q] {mbox,am,attest,pr,ty,diff,kr} ... b4: error: argument subcmd: invalid choice: 'shazam' (choose from 'mbox', 'am', 'attest', 'pr', 'ty', 'diff', 'kr')
Do I need a higher version of b4 (here: v0.8.0)?
Check myself... b4.git:
commit 7c1d044ff1d5235e598d4c777c4abfe60e0a09a8 ("shazam: change default behaviour to be "apply-here"")
...is post-v0.8.0.
Lemme see if I can apply this patch...
# cd /usr/lib/python3/dist-packages
# LC_ALL=C git apply --check --verbose /root/b4-shazam.patch Checking patch b4/command.py...
error: while searching for:
sp.add_argument('-M', '--save-as-maildir', dest='maildir',
action='store_true', default=False,
help='Save as maildir (avoids mbox format ambiguities)')
def cmd_am_common_opts(sp):
sp.add_argument('-v', '--use-version', dest='wantver', type=int,
default=None,
help='Get a specific version of the patch/series')
error: patch failed: b4/command.py:35
error: b4/command.py: patch does not apply Checking patch b4/mbox.py...
error: while searching for:
ifh = io.StringIO() b4.save_git_am_mbox(am_msgs, ifh) ambytes = ifh.getvalue().encode()
if cmdargs.applyhere:
amflags = config.get('git-am-flags', '') sp = shlex.shlex(amflags, posix=True) sp.whitespace_split = True
error: patch failed: b4/mbox.py:262
error: b4/mbox.py: patch does not apply
Nope.
Dunno if I am willing to do that manually or build-from-git.
Anyway, can you add this b4 shazam tipp/trick to our wiki, please?
Last question:
LLVM/Clang-14...
Do I need any patches to Linux v5.17-rc6 or upstream Git? Dependent/Independent of "std-gnu-11"?
I can see on my Debian/unstable AMD64 system:
# LC_ALL=C apt-cache policy clang-14
clang-14:
Installed: (none)
Candidate: 1:14.0.0~++20220301114802+19149538e9a9-1~exp1~20220301234814.85 Version table:
1:14.0.0~++20220301114802+19149538e9a9-1~exp1~20220301234814.85 99
99 https://apt.llvm.org/unstable llvm-toolchain-14/main amd64 Packages
1:14.0.0~+rc1-1 99
99 https://ftp.debian.org/debian unstable/main amd64 Packages 99 https://deb.debian.org/debian unstable/main amd64 Packages
The one from apt.llvm.org I guess is LLVM/Clang v14.0.0-rc2?
Maybe, I wait until Masahiroy has the triple in his kbuild Git tree...
Thanks.
Have a nice Friday,
- sed@ -
[1] https://git.kernel.org/pub/scm/utils/b4/b4.git/tag/?h=v0.8.0
[2] https://git.kernel.org/pub/scm/utils/b4/b4.git/commit/?id=7c1d044ff1d5235e598d4c777c4abfe60e0a09a8
[3] https://github.com/ClangBuiltLinux/linux/wiki/Command-line-tips-and-tricks
Re: [PATCH 3/3] Kbuild: use -std=gnu11 for KBUILD_USERCFLAGS
Quoted 61 lines · Sedat Dilek
Quoted 36 lines · Nathan Chancellor
Hi Sedat,
Quoted 20 lines · Sedat Dilek
Hey Nick!
This only applies 1/3.
$ b4 --version 0.8.0$ b4 am https://lore.kernel.org/lkml/[email protected]/ -o - | git am -3 Analyzing 14 messages in the thread Will use the latest revision: v3 You can pick other revisions using the -vN flag Checking attestation on all messages, may take a moment... --- ✓ [PATCH v3 1/3] Kbuild: move to -std=gnu11 ✓ Signed: DKIM/kernel.org + Reviewed-by: Nathan Chancellor <[email protected]> (✓ DKIM/kernel.org) ERROR: missing [2/3]! ERROR: missing [3/3]! ---Quoted 8 lines
NOTE: install patatt for end-to-end signature verification --- Total patches: 1 --- WARNING: Thread incomplete! Link: https://lore.kernel.org/r/[email protected] Base: not specified Wende an: Kbuild: move to -std=gnu11It looks like the threading somehow got broken, likely due to the [v3] on the first patch and not the second or third:
Quoted 8 lines
This worked for me on v5.17-rc6:
$ for i in $(seq 1 3); do b4 shazam -P _ 20220301145233.3689119-"$i"[email protected]; done"b4 shazam" is the equivalent of "b4 am -o - ... | git am" and the "-P _" tells b4 to only fetch that exact message ID, not the whole thread.
Hmm, the universe is not good to me...
$ for i in $(seq 1 3); do b4 shazam -P _ 20220301145233.3689119-"$i"[email protected]; done usage: b4 [-h] [--version] [-d] [-q] {mbox,am,attest,pr,ty,diff,kr} ... b4: error: argument subcmd: invalid choice: 'shazam' (choose from 'mbox', 'am', 'attest', 'pr', 'ty', 'diff', 'kr') usage: b4 [-h] [--version] [-d] [-q] {mbox,am,attest,pr,ty,diff,kr} ... b4: error: argument subcmd: invalid choice: 'shazam' (choose from 'mbox', 'am', 'attest', 'pr', 'ty', 'diff', 'kr') usage: b4 [-h] [--version] [-d] [-q] {mbox,am,attest,pr,ty,diff,kr} ... b4: error: argument subcmd: invalid choice: 'shazam' (choose from 'mbox', 'am', 'attest', 'pr', 'ty', 'diff', 'kr')
Quoted 9 lines
Do I need a higher version of b4 (here: v0.8.0)?
Check myself... b4.git:
commit
7c1d044ff1d5235e598d4c777c4abfe60e0a09a8("shazam: change default behaviour to be "apply-here"")...is post-v0.8.0.
I brutally applied the post-v0.8.0 patches stolen from b4.git over my local distro b4 files.
And was able to apply the triple:
$ git log --oneline --no-merges 5.17.0-rc6-1-amd64-clang13-lto..5.17.0-rc6-2-amd64-clang13-lto 96a4222bdd4c (for-5.17/kbuild-std_gnu11-arndb-20220301) Kbuild: use -std=gnu11 for KBUILD_USERCFLAGS c4e8cef401a8 treewide: use -Wdeclaration-after-statement 6a7cc105b238 Kbuild: move to -std=gnu11
- sed@ -
Sedat Dilek
Tested-byRe: [PATCH 3/3] Kbuild: use -std=gnu11 for KBUILD_USERCFLAGS
Quoted 76 lines · Sedat Dilek
Quoted 61 lines · Sedat Dilek
Quoted 36 lines · Nathan Chancellor
Hi Sedat,
Quoted 20 lines · Sedat Dilek
Hey Nick!
This only applies 1/3.
$ b4 --version 0.8.0$ b4 am https://lore.kernel.org/lkml/[email protected]/ -o - | git am -3 Analyzing 14 messages in the thread Will use the latest revision: v3 You can pick other revisions using the -vN flag Checking attestation on all messages, may take a moment... --- ✓ [PATCH v3 1/3] Kbuild: move to -std=gnu11 ✓ Signed: DKIM/kernel.org + Reviewed-by: Nathan Chancellor <[email protected]> (✓ DKIM/kernel.org) ERROR: missing [2/3]! ERROR: missing [3/3]! ---Quoted 8 lines
NOTE: install patatt for end-to-end signature verification --- Total patches: 1 --- WARNING: Thread incomplete! Link: https://lore.kernel.org/r/[email protected] Base: not specified Wende an: Kbuild: move to -std=gnu11It looks like the threading somehow got broken, likely due to the [v3] on the first patch and not the second or third:
Quoted 8 lines
This worked for me on v5.17-rc6:
$ for i in $(seq 1 3); do b4 shazam -P _ 20220301145233.3689119-"$i"[email protected]; done"b4 shazam" is the equivalent of "b4 am -o - ... | git am" and the "-P _" tells b4 to only fetch that exact message ID, not the whole thread.
Hmm, the universe is not good to me...
$ for i in $(seq 1 3); do b4 shazam -P _ 20220301145233.3689119-"$i"[email protected]; done usage: b4 [-h] [--version] [-d] [-q] {mbox,am,attest,pr,ty,diff,kr} ... b4: error: argument subcmd: invalid choice: 'shazam' (choose from 'mbox', 'am', 'attest', 'pr', 'ty', 'diff', 'kr') usage: b4 [-h] [--version] [-d] [-q] {mbox,am,attest,pr,ty,diff,kr} ... b4: error: argument subcmd: invalid choice: 'shazam' (choose from 'mbox', 'am', 'attest', 'pr', 'ty', 'diff', 'kr') usage: b4 [-h] [--version] [-d] [-q] {mbox,am,attest,pr,ty,diff,kr} ... b4: error: argument subcmd: invalid choice: 'shazam' (choose from 'mbox', 'am', 'attest', 'pr', 'ty', 'diff', 'kr')Quoted 9 lines
Do I need a higher version of b4 (here: v0.8.0)?
Check myself... b4.git:
commit
7c1d044ff1d5235e598d4c777c4abfe60e0a09a8("shazam: change default behaviour to be "apply-here"")...is post-v0.8.0.
I brutally applied the post-v0.8.0 patches stolen from b4.git over my local distro b4 files.
Quoted 9 lines
And was able to apply the triple:
$ git log --oneline --no-merges 5.17.0-rc6-1-amd64-clang13-lto..5.17.0-rc6-2-amd64-clang13-lto 96a4222bdd4c (for-5.17/kbuild-std_gnu11-arndb-20220301) Kbuild: use -std=gnu11 for KBUILD_USERCFLAGS c4e8cef401a8 treewide: use -Wdeclaration-after-statement 6a7cc105b238 Kbuild: move to -std=gnu11
I was able to build and boot on bare metal.
No new warnings in my build-log here after switching to -std=gnu11.
- sed@ -
Quoted 53 lines · Arnd Bergmann
From: Arnd Bergmann <[email protected]>
During a patch discussion, Linus brought up the option of changing the C standard version from gnu89 to gnu99, which allows using variable declaration inside of a
for()loop. While the C99, C11 and later standards introduce many other features, most of these are already available in gnu89 as GNU extensions as well.An earlier attempt to do this when gcc-5 started defaulting to -std=gnu11 failed because at the time that caused warnings about designated initializers with older compilers. Now that gcc-5.1 is the minimum compiler version used for building kernels, that is no longer a concern. Similarly, the behavior of 'inline' functions changes between gnu89 using
gnu_inlinebehavior and gnu11 using standard c99+ behavior, but this was taken care of by defining 'inline' to include __attribute__((gnu_inline)) in order to allow building with clang a while ago.One minor issue that remains is an added gcc warning for shifts of negative integers when building with -Wextra, which happens with the 'make W=1' option, as well as for three drivers in the kernel that always enable -Wextra, but it was only observed with the i915 driver so far. To be on the safe side, add -Wno-shift-negative-value to any -Wextra in a
Makefile.Nathan Chancellor reported an additional -Wdeclaration-after-statement warning that appears in a system header on arm, this still needs a workaround.
The differences between gnu99, gnu11, gnu1x and gnu17 are fairly minimal and mainly impact warnings at the -Wpedantic level that the kernel never enables. Between these, gnu11 is the newest version that is supported by all supported compiler versions, though it is only the default on gcc-5, while all other supported versions of gcc or clang default to gnu1x/gnu17.
Link https://lore.kernel.org/lkml/CAHk-=wiyCH7xeHcmiFJ-YgXUy2Jaj7pnkdKpcovt8fYbVFW3TA@mail.gmail.com/ open hereSuggested-by Linus Torvalds <[email protected]>Acked-by Marco Elver <[email protected]>Acked-by Jani Nikula <[email protected]>Acked-by David Sterba <[email protected]>Reviewed-by Alex Shi <[email protected]>Reviewed-by Nick Desaulniers <[email protected]>Reviewed-by Miguel Ojeda <[email protected]>Cc Masahiro Yamada <[email protected]>Cc Mark Rutland <[email protected]>Signed-off-by Arnd Bergmann <[email protected]>
Quoted 7 lines
---
[v3]
- split out USERCFLAGS to a separate patch
- add -Wdeclaration-after-statement patch from Mark Rutland
- leave out C17 reference
- more rewording the descrption
I was able to build and boot on bare metal.
No new warnings in my build-log here after switching to -std=gnu11.
- sed@ -
Quoted 164 lines
[v2]
- added -std=gnu11 back, rather than just relying on the default
- minor changes to changelog text--- Documentation/process/programming-language.rst | 6 +++--- .../translations/it_IT/process/programming-language.rst | 4 ++-- .../translations/zh_CN/process/programming-language.rst | 3 +-- .../translations/zh_TW/process/programming-language.rst | 3 +-- Makefile | 4 ++-- arch/arm64/kernel/vdso32/Makefile | 2 +- drivers/gpu/drm/i915/Makefile | 1 + drivers/staging/greybus/tools/Makefile | 3 ++- fs/btrfs/Makefile | 1 + scripts/Makefile.extrawarn | 1 + 10 files changed, 15 insertions(+), 13 deletions(-)Documentation/process/programming-language.rst +3
-3@@ -5,9 +5,9 @@ Programming Language The kernel is written in the C programming language [c-language]_. More precisely, the kernel is typically compiled with ``gcc`` [gcc]_ -under ``-std=gnu89`` [gcc-c-dialect-options]_: the GNU dialect of ISO C90 -(including some C99 features). ``clang`` [clang]_ is also supported, see -docs on :ref:`Building Linux with Clang/LLVM <kbuild_llvm>`. +under ``-std=gnu11`` [gcc-c-dialect-options]_: the GNU dialect of ISO C11. +``clang`` [clang]_ is also supported, see docs on +:ref:`Building Linux with Clang/LLVM <kbuild_llvm>`. This dialect contains many extensions to the language [gnu-extensions]_, and many of them are used within the kernel as a matter of course.Documentation/translations/it_IT/process/programming-language.rst +2
-2@@ -10,8 +10,8 @@ Linguaggio di programmazione Il kernel è scritto nel linguaggio di programmazione C [it-c-language]_. Più precisamente, il kernel viene compilato con ``gcc`` [it-gcc]_ usando -l'opzione ``-std=gnu89`` [it-gcc-c-dialect-options]_: il dialetto GNU -dello standard ISO C90 (con l'aggiunta di alcune funzionalità da C99). +l'opzione ``-std=gnu11`` [it-gcc-c-dialect-options]_: il dialetto GNU +dello standard ISO C11. Linux supporta anche ``clang`` [it-clang]_, leggete la documentazione :ref:`Building Linux with Clang/LLVM <kbuild_llvm>`.Documentation/translations/zh_CN/process/programming-language.rst +1
-2@@ -9,8 +9,7 @@ ============ 内核是用C语言 :ref:`c-language <cn_c-language>` 编写的。更准确地说,内核通常是用 :ref:`gcc <cn_gcc>` -在 ``-std=gnu89`` :ref:`gcc-c-dialect-options <cn_gcc-c-dialect-options>` 下编译的:ISO C90的 GNU 方言( -包括一些C99特性) +在 ``-std=gnu11`` :ref:`gcc-c-dialect-options <cn_gcc-c-dialect-options>` 下编译的:ISO C11的 GNU 方言 这种方言包含对语言 :ref:`gnu-extensions <cn_gnu-extensions>` 的许多扩展,当然,它们许多都在内核中使用。Documentation/translations/zh_TW/process/programming-language.rst +1
-2@@ -12,8 +12,7 @@ ============ 內核是用C語言 :ref:`c-language <tw_c-language>` 編寫的。更準確地說,內核通常是用 :ref:`gcc <tw_gcc>` -在 ``-std=gnu89`` :ref:`gcc-c-dialect-options <tw_gcc-c-dialect-options>` 下編譯的:ISO C90的 GNU 方言( -包括一些C99特性) +在 ``-std=gnu11`` :ref:`gcc-c-dialect-options <tw_gcc-c-dialect-options>` 下編譯的:ISO C11的 GNU 方言 這種方言包含對語言 :ref:`gnu-extensions <tw_gnu-extensions>` 的許多擴展,當然,它們許多都在內核中使用。Makefile +2
-2@@ -515,7 +515,7 @@ KBUILD_CFLAGS := -Wall -Wundef -Werror=strict-prototypes -Wno-trigraphs \ -fno-strict-aliasing -fno-common -fshort-wchar -fno-PIE \ -Werror=implicit-function-declaration -Werror=implicit-int \ -Werror=return-type -Wno-format-security \ - -std=gnu89 + -std=gnu11 KBUILD_CPPFLAGS := -D__KERNEL__ KBUILD_AFLAGS_KERNEL := KBUILD_CFLAGS_KERNEL :=@@ -782,7 +782,7 @@ KBUILD_CFLAGS += $(KBUILD_CFLAGS-y) $(CONFIG_CC_IMPLICIT_FALLTHROUGH) ifdef CONFIG_CC_IS_CLANG KBUILD_CPPFLAGS += -Qunused-arguments -# The kernel builds with '-std=gnu89' so use of GNU extensions is acceptable. +# The kernel builds with '-std=gnu11' so use of GNU extensions is acceptable. KBUILD_CFLAGS += -Wno-gnu # CLANG uses a _MergedGlobals as optimization, but this breaks modpost, as the # source of a reference will be _MergedGlobals and not on of the whitelisted names.arch/arm64/kernel/vdso32/Makefile +1
-1@@ -68,7 +68,7 @@ VDSO_CFLAGS += -Wall -Wundef -Wstrict-prototypes -Wno-trigraphs \ -fno-strict-aliasing -fno-common \ -Werror-implicit-function-declaration \ -Wno-format-security \ - -std=gnu89 + -std=gnu11 VDSO_CFLAGS += -O2 # Some useful compiler-dependent flags from top-level Makefile VDSO_CFLAGS += $(call cc32-option,-Wdeclaration-after-statement,)drivers/gpu/drm/i915/Makefile +1
-0@@ -17,6 +17,7 @@ subdir-ccflags-y += -Wno-unused-parameter subdir-ccflags-y += -Wno-type-limits subdir-ccflags-y += -Wno-missing-field-initializers subdir-ccflags-y += -Wno-sign-compare +subdir-ccflags-y += -Wno-shift-negative-value subdir-ccflags-y += $(call cc-disable-warning, unused-but-set-variable) subdir-ccflags-y += $(call cc-disable-warning, frame-address) subdir-ccflags-$(CONFIG_DRM_I915_WERROR) += -Werrordrivers/staging/greybus/tools/Makefile +2
-1@@ -12,7 +12,8 @@ CFLAGS += -std=gnu99 -Wall -Wextra -g \ -Wredundant-decls \ -Wcast-align \ -Wsign-compare \ - -Wno-missing-field-initializers + -Wno-missing-field-initializers \ + -Wno-shift-negative-value CC := $(CROSS_COMPILE)gccfs/btrfs/Makefile +1
-0@@ -17,6 +17,7 @@ subdir-ccflags-y += $(condflags) subdir-ccflags-y += -Wno-missing-field-initializers subdir-ccflags-y += -Wno-sign-compare subdir-ccflags-y += -Wno-type-limits +subdir-ccflags-y += -Wno-shift-negative-value obj-$(CONFIG_BTRFS_FS) := btrfs.oscripts/Makefile.extrawarn +1
-1@@ -36,6 +36,7 @@ KBUILD_CFLAGS += $(call cc-option, -Wstringop-truncation) KBUILD_CFLAGS += -Wno-missing-field-initializers KBUILD_CFLAGS += -Wno-sign-compare KBUILD_CFLAGS += -Wno-type-limits +KBUILD_CFLAGS += -Wno-shift-negative-value KBUILD_CPPFLAGS += -DKBUILD_EXTRA_WARN1 --2.29.2
Re: [PATCH 2/3] treewide: use -Wdeclaration-after-statement
In reply to Arnd BergmannQuoted 18 lines · Arnd Bergmann
From: Mark Rutland <[email protected]>
In a subsequent patch we'll move the kernel from using
-std=gnu89to-std=gnu11, permitting the use of additional C11 features such as for-loop initial declarations.One contentious aspect of C99 is that it permits mixed declarations and code, and for now at least, it seems preferable to enforce that declarations must come first.
These warnings were already disabled in the kernel itself, but not for
KBUILD_USERCFLAGSor the compat VDSO onarch/arm64, which uses a separate set of CFLAGS.This patch fixes an existing violation in
modpost.c, which is not reported because of the missing flag inKBUILD_USERCFLAGS:
Quoted 8 lines
| scripts/mod/modpost.c: In function ‘match’: | scripts/mod/modpost.c:837:3: warning: ISO C90 forbids mixed declarations and code [-Wdeclaration-after-statement] | 837 | const char *endp = p + strlen(p) - 1; | | ^~~~~Signed-off-by Mark Rutland <[email protected]>[arnd: don't add a duplicate flag to the default set, update changelog]Signed-off-by Arnd Bergmann <[email protected]>
I was able to build and boot on bare metal.
No new warnings in my build-log here after switching to -std=gnu11.
- sed@ -
Quoted 51 lines
--- Makefile | 3 ++- arch/arm64/kernel/vdso32/Makefile | 1 + scripts/mod/modpost.c | 4 +++- 3 files changed, 6 insertions(+), 2 deletions(-)Makefile +2
-1@@ -432,7 +432,8 @@ HOSTCXX = g++ endif export KBUILD_USERCFLAGS := -Wall -Wmissing-prototypes -Wstrict-prototypes \ - -O2 -fomit-frame-pointer -std=gnu89 + -O2 -fomit-frame-pointer -std=gnu89 \ + -Wdeclaration-after-statement export KBUILD_USERLDFLAGS := KBUILD_HOSTCFLAGS := $(KBUILD_USERCFLAGS) $(HOST_LFS_CFLAGS) $(HOSTCFLAGS)arch/arm64/kernel/vdso32/Makefile +1
-0@@ -68,6 +68,7 @@ VDSO_CFLAGS += -Wall -Wundef -Wstrict-prototypes -Wno-trigraphs \ -fno-strict-aliasing -fno-common \ -Werror-implicit-function-declaration \ -Wno-format-security \ + -Wdeclaration-after-statement \ -std=gnu11 VDSO_CFLAGS += -O2 # Some useful compiler-dependent flags from top-level Makefilescripts/mod/modpost.c +3
-2@@ -833,8 +833,10 @@ static int match(const char *sym, const char * const pat[]) {const char *p;while (*pat) { +const char *endp; + p = *pat++; -const char *endp = p +strlen (p) -1 ; + endp = p +strlen (p) -1 ;/* "*foo*" */ if (*p =='*' && *endp =='*' ) { --2.29.2
Quoted 53 lines · Arnd Bergmann
From: Arnd Bergmann <[email protected]>
During a patch discussion, Linus brought up the option of changing the C standard version from gnu89 to gnu99, which allows using variable declaration inside of a
for()loop. While the C99, C11 and later standards introduce many other features, most of these are already available in gnu89 as GNU extensions as well.An earlier attempt to do this when gcc-5 started defaulting to -std=gnu11 failed because at the time that caused warnings about designated initializers with older compilers. Now that gcc-5.1 is the minimum compiler version used for building kernels, that is no longer a concern. Similarly, the behavior of 'inline' functions changes between gnu89 using
gnu_inlinebehavior and gnu11 using standard c99+ behavior, but this was taken care of by defining 'inline' to include __attribute__((gnu_inline)) in order to allow building with clang a while ago.One minor issue that remains is an added gcc warning for shifts of negative integers when building with -Wextra, which happens with the 'make W=1' option, as well as for three drivers in the kernel that always enable -Wextra, but it was only observed with the i915 driver so far. To be on the safe side, add -Wno-shift-negative-value to any -Wextra in a
Makefile.Nathan Chancellor reported an additional -Wdeclaration-after-statement warning that appears in a system header on arm, this still needs a workaround.
The differences between gnu99, gnu11, gnu1x and gnu17 are fairly minimal and mainly impact warnings at the -Wpedantic level that the kernel never enables. Between these, gnu11 is the newest version that is supported by all supported compiler versions, though it is only the default on gcc-5, while all other supported versions of gcc or clang default to gnu1x/gnu17.
Link https://lore.kernel.org/lkml/CAHk-=wiyCH7xeHcmiFJ-YgXUy2Jaj7pnkdKpcovt8fYbVFW3TA@mail.gmail.com/ open hereSuggested-by Linus Torvalds <[email protected]>Acked-by Marco Elver <[email protected]>Acked-by Jani Nikula <[email protected]>Acked-by David Sterba <[email protected]>Reviewed-by Alex Shi <[email protected]>Reviewed-by Nick Desaulniers <[email protected]>Reviewed-by Miguel Ojeda <[email protected]>Cc Masahiro Yamada <[email protected]>Cc Mark Rutland <[email protected]>Signed-off-by Arnd Bergmann <[email protected]>
Quoted 1 line
---
Arnd, please CC me on further patches/patchsets.
Thanks.
- Sedat -
Quoted 170 lines
[v3]
- split out USERCFLAGS to a separate patch
- add -Wdeclaration-after-statement patch from Mark Rutland
- leave out C17 reference
- more rewording the descrption[v2]
- added -std=gnu11 back, rather than just relying on the default
- minor changes to changelog text--- Documentation/process/programming-language.rst | 6 +++--- .../translations/it_IT/process/programming-language.rst | 4 ++-- .../translations/zh_CN/process/programming-language.rst | 3 +-- .../translations/zh_TW/process/programming-language.rst | 3 +-- Makefile | 4 ++-- arch/arm64/kernel/vdso32/Makefile | 2 +- drivers/gpu/drm/i915/Makefile | 1 + drivers/staging/greybus/tools/Makefile | 3 ++- fs/btrfs/Makefile | 1 + scripts/Makefile.extrawarn | 1 + 10 files changed, 15 insertions(+), 13 deletions(-)Documentation/process/programming-language.rst +3
-3@@ -5,9 +5,9 @@ Programming Language The kernel is written in the C programming language [c-language]_. More precisely, the kernel is typically compiled with ``gcc`` [gcc]_ -under ``-std=gnu89`` [gcc-c-dialect-options]_: the GNU dialect of ISO C90 -(including some C99 features). ``clang`` [clang]_ is also supported, see -docs on :ref:`Building Linux with Clang/LLVM <kbuild_llvm>`. +under ``-std=gnu11`` [gcc-c-dialect-options]_: the GNU dialect of ISO C11. +``clang`` [clang]_ is also supported, see docs on +:ref:`Building Linux with Clang/LLVM <kbuild_llvm>`. This dialect contains many extensions to the language [gnu-extensions]_, and many of them are used within the kernel as a matter of course.Documentation/translations/it_IT/process/programming-language.rst +2
-2@@ -10,8 +10,8 @@ Linguaggio di programmazione Il kernel è scritto nel linguaggio di programmazione C [it-c-language]_. Più precisamente, il kernel viene compilato con ``gcc`` [it-gcc]_ usando -l'opzione ``-std=gnu89`` [it-gcc-c-dialect-options]_: il dialetto GNU -dello standard ISO C90 (con l'aggiunta di alcune funzionalità da C99). +l'opzione ``-std=gnu11`` [it-gcc-c-dialect-options]_: il dialetto GNU +dello standard ISO C11. Linux supporta anche ``clang`` [it-clang]_, leggete la documentazione :ref:`Building Linux with Clang/LLVM <kbuild_llvm>`.Documentation/translations/zh_CN/process/programming-language.rst +1
-2@@ -9,8 +9,7 @@ ============ 内核是用C语言 :ref:`c-language <cn_c-language>` 编写的。更准确地说,内核通常是用 :ref:`gcc <cn_gcc>` -在 ``-std=gnu89`` :ref:`gcc-c-dialect-options <cn_gcc-c-dialect-options>` 下编译的:ISO C90的 GNU 方言( -包括一些C99特性) +在 ``-std=gnu11`` :ref:`gcc-c-dialect-options <cn_gcc-c-dialect-options>` 下编译的:ISO C11的 GNU 方言 这种方言包含对语言 :ref:`gnu-extensions <cn_gnu-extensions>` 的许多扩展,当然,它们许多都在内核中使用。Documentation/translations/zh_TW/process/programming-language.rst +1
-2@@ -12,8 +12,7 @@ ============ 內核是用C語言 :ref:`c-language <tw_c-language>` 編寫的。更準確地說,內核通常是用 :ref:`gcc <tw_gcc>` -在 ``-std=gnu89`` :ref:`gcc-c-dialect-options <tw_gcc-c-dialect-options>` 下編譯的:ISO C90的 GNU 方言( -包括一些C99特性) +在 ``-std=gnu11`` :ref:`gcc-c-dialect-options <tw_gcc-c-dialect-options>` 下編譯的:ISO C11的 GNU 方言 這種方言包含對語言 :ref:`gnu-extensions <tw_gnu-extensions>` 的許多擴展,當然,它們許多都在內核中使用。Makefile +2
-2@@ -515,7 +515,7 @@ KBUILD_CFLAGS := -Wall -Wundef -Werror=strict-prototypes -Wno-trigraphs \ -fno-strict-aliasing -fno-common -fshort-wchar -fno-PIE \ -Werror=implicit-function-declaration -Werror=implicit-int \ -Werror=return-type -Wno-format-security \ - -std=gnu89 + -std=gnu11 KBUILD_CPPFLAGS := -D__KERNEL__ KBUILD_AFLAGS_KERNEL := KBUILD_CFLAGS_KERNEL :=@@ -782,7 +782,7 @@ KBUILD_CFLAGS += $(KBUILD_CFLAGS-y) $(CONFIG_CC_IMPLICIT_FALLTHROUGH) ifdef CONFIG_CC_IS_CLANG KBUILD_CPPFLAGS += -Qunused-arguments -# The kernel builds with '-std=gnu89' so use of GNU extensions is acceptable. +# The kernel builds with '-std=gnu11' so use of GNU extensions is acceptable. KBUILD_CFLAGS += -Wno-gnu # CLANG uses a _MergedGlobals as optimization, but this breaks modpost, as the # source of a reference will be _MergedGlobals and not on of the whitelisted names.arch/arm64/kernel/vdso32/Makefile +1
-1@@ -68,7 +68,7 @@ VDSO_CFLAGS += -Wall -Wundef -Wstrict-prototypes -Wno-trigraphs \ -fno-strict-aliasing -fno-common \ -Werror-implicit-function-declaration \ -Wno-format-security \ - -std=gnu89 + -std=gnu11 VDSO_CFLAGS += -O2 # Some useful compiler-dependent flags from top-level Makefile VDSO_CFLAGS += $(call cc32-option,-Wdeclaration-after-statement,)drivers/gpu/drm/i915/Makefile +1
-0@@ -17,6 +17,7 @@ subdir-ccflags-y += -Wno-unused-parameter subdir-ccflags-y += -Wno-type-limits subdir-ccflags-y += -Wno-missing-field-initializers subdir-ccflags-y += -Wno-sign-compare +subdir-ccflags-y += -Wno-shift-negative-value subdir-ccflags-y += $(call cc-disable-warning, unused-but-set-variable) subdir-ccflags-y += $(call cc-disable-warning, frame-address) subdir-ccflags-$(CONFIG_DRM_I915_WERROR) += -Werrordrivers/staging/greybus/tools/Makefile +2
-1@@ -12,7 +12,8 @@ CFLAGS += -std=gnu99 -Wall -Wextra -g \ -Wredundant-decls \ -Wcast-align \ -Wsign-compare \ - -Wno-missing-field-initializers + -Wno-missing-field-initializers \ + -Wno-shift-negative-value CC := $(CROSS_COMPILE)gccfs/btrfs/Makefile +1
-0@@ -17,6 +17,7 @@ subdir-ccflags-y += $(condflags) subdir-ccflags-y += -Wno-missing-field-initializers subdir-ccflags-y += -Wno-sign-compare subdir-ccflags-y += -Wno-type-limits +subdir-ccflags-y += -Wno-shift-negative-value obj-$(CONFIG_BTRFS_FS) := btrfs.oscripts/Makefile.extrawarn +1
-1@@ -36,6 +36,7 @@ KBUILD_CFLAGS += $(call cc-option, -Wstringop-truncation) KBUILD_CFLAGS += -Wno-missing-field-initializers KBUILD_CFLAGS += -Wno-sign-compare KBUILD_CFLAGS += -Wno-type-limits +KBUILD_CFLAGS += -Wno-shift-negative-value KBUILD_CPPFLAGS += -DKBUILD_EXTRA_WARN1 --2.29.2
Quoted 37 lines · Arnd Bergmann
From: Arnd Bergmann <[email protected]>
During a patch discussion, Linus brought up the option of changing the C standard version from gnu89 to gnu99, which allows using variable declaration inside of a
for()loop. While the C99, C11 and later standards introduce many other features, most of these are already available in gnu89 as GNU extensions as well.An earlier attempt to do this when gcc-5 started defaulting to -std=gnu11 failed because at the time that caused warnings about designated initializers with older compilers. Now that gcc-5.1 is the minimum compiler version used for building kernels, that is no longer a concern. Similarly, the behavior of 'inline' functions changes between gnu89 using
gnu_inlinebehavior and gnu11 using standard c99+ behavior, but this was taken care of by defining 'inline' to include __attribute__((gnu_inline)) in order to allow building with clang a while ago.One minor issue that remains is an added gcc warning for shifts of negative integers when building with -Wextra, which happens with the 'make W=1' option, as well as for three drivers in the kernel that always enable -Wextra, but it was only observed with the i915 driver so far. To be on the safe side, add -Wno-shift-negative-value to any -Wextra in a
Makefile.Nathan Chancellor reported an additional -Wdeclaration-after-statement warning that appears in a system header on arm, this still needs a workaround.
The differences between gnu99, gnu11, gnu1x and gnu17 are fairly minimal and mainly impact warnings at the -Wpedantic level that the kernel never enables. Between these, gnu11 is the newest version that is supported by all supported compiler versions, though it is only the default on gcc-5, while all other supported versions of gcc or clang default to gnu1x/gnu17.
Quoted 16 lines
Link https://lore.kernel.org/lkml/CAHk-=wiyCH7xeHcmiFJ-YgXUy2Jaj7pnkdKpcovt8fYbVFW3TA@mail.gmail.com/ open hereSuggested-by Linus Torvalds <[email protected]>Acked-by Marco Elver <[email protected]>Acked-by Jani Nikula <[email protected]>Acked-by David Sterba <[email protected]>Reviewed-by Alex Shi <[email protected]>Reviewed-by Nick Desaulniers <[email protected]>Reviewed-by Miguel Ojeda <[email protected]>Cc Masahiro Yamada <[email protected]>Cc Mark Rutland <[email protected]>Signed-off-by Arnd Bergmann <[email protected]>
When I saw the first patches in this area, I was thinking what about the tools/ build-system?
This ECO system has its own rules.
I use CONFIG_DEBUG_INFO_BTF=y in my kernels which means tools/bpf/resolve_btfids will be built.
This uses hardcoded -std=gnu89.
$ git grep 'std=gnu89' tools/ tools/lib/bpf/Makefile:override CFLAGS += -std=gnu89 tools/testing/selftests/sync/Makefile:CFLAGS += -O2 -g -std=gnu89 -pthread -Wall -Wextra
Furthermore, we have several places using C99 which can be replaced by C11 now.
$ git grep 'std=gnu99' tools/ | wc -l 20
We already have C11 in tools/:
$ git grep 'std=gnu11' tools/ tools/testing/selftests/wireguard/qemu/Makefile: $(MUSL_CC) -o $@ $(CFLAGS) $(LDFLAGS) -std=gnu11 $<
What about having a central place where we define the STD_GNU_C_VERSION (as a suggestion)?
This might be helpful for people dealing with diverse C version standards.
Also with STD_GNU_CXX_VERSION which is used AFAICS only in tools/:
$ git grep 'std=gnu++' tools/
tools/build/feature/Makefile: $(BUILDXX) -std=gnu++11
tools/build/feature/Makefile: $(BUILDXX) -std=gnu++14
\
tools/build/feature/Makefile: $(BUILDXX) -std=gnu++14
\
tools/build/feature/Makefile: $(BUILDXX) -std=gnu++14
\
tools/perf/Makefile.config:CXXFLAGS += -std=gnu++14 -fno-exceptions -fno-rttiLooks like one place should be -std=gnu++14.
( I had no deeper look. No testing. )
Just some thoughts.
- Sedat -
Quoted 171 lines
---
[v3]
- split out USERCFLAGS to a separate patch
- add -Wdeclaration-after-statement patch from Mark Rutland
- leave out C17 reference
- more rewording the descrption[v2]
- added -std=gnu11 back, rather than just relying on the default
- minor changes to changelog text--- Documentation/process/programming-language.rst | 6 +++--- .../translations/it_IT/process/programming-language.rst | 4 ++-- .../translations/zh_CN/process/programming-language.rst | 3 +-- .../translations/zh_TW/process/programming-language.rst | 3 +-- Makefile | 4 ++-- arch/arm64/kernel/vdso32/Makefile | 2 +- drivers/gpu/drm/i915/Makefile | 1 + drivers/staging/greybus/tools/Makefile | 3 ++- fs/btrfs/Makefile | 1 + scripts/Makefile.extrawarn | 1 + 10 files changed, 15 insertions(+), 13 deletions(-)Documentation/process/programming-language.rst +3
-3@@ -5,9 +5,9 @@ Programming Language The kernel is written in the C programming language [c-language]_. More precisely, the kernel is typically compiled with ``gcc`` [gcc]_ -under ``-std=gnu89`` [gcc-c-dialect-options]_: the GNU dialect of ISO C90 -(including some C99 features). ``clang`` [clang]_ is also supported, see -docs on :ref:`Building Linux with Clang/LLVM <kbuild_llvm>`. +under ``-std=gnu11`` [gcc-c-dialect-options]_: the GNU dialect of ISO C11. +``clang`` [clang]_ is also supported, see docs on +:ref:`Building Linux with Clang/LLVM <kbuild_llvm>`. This dialect contains many extensions to the language [gnu-extensions]_, and many of them are used within the kernel as a matter of course.Documentation/translations/it_IT/process/programming-language.rst +2
-2@@ -10,8 +10,8 @@ Linguaggio di programmazione Il kernel è scritto nel linguaggio di programmazione C [it-c-language]_. Più precisamente, il kernel viene compilato con ``gcc`` [it-gcc]_ usando -l'opzione ``-std=gnu89`` [it-gcc-c-dialect-options]_: il dialetto GNU -dello standard ISO C90 (con l'aggiunta di alcune funzionalità da C99). +l'opzione ``-std=gnu11`` [it-gcc-c-dialect-options]_: il dialetto GNU +dello standard ISO C11. Linux supporta anche ``clang`` [it-clang]_, leggete la documentazione :ref:`Building Linux with Clang/LLVM <kbuild_llvm>`.Documentation/translations/zh_CN/process/programming-language.rst +1
-2@@ -9,8 +9,7 @@ ============ 内核是用C语言 :ref:`c-language <cn_c-language>` 编写的。更准确地说,内核通常是用 :ref:`gcc <cn_gcc>` -在 ``-std=gnu89`` :ref:`gcc-c-dialect-options <cn_gcc-c-dialect-options>` 下编译的:ISO C90的 GNU 方言( -包括一些C99特性) +在 ``-std=gnu11`` :ref:`gcc-c-dialect-options <cn_gcc-c-dialect-options>` 下编译的:ISO C11的 GNU 方言 这种方言包含对语言 :ref:`gnu-extensions <cn_gnu-extensions>` 的许多扩展,当然,它们许多都在内核中使用。Documentation/translations/zh_TW/process/programming-language.rst +1
-2@@ -12,8 +12,7 @@ ============ 內核是用C語言 :ref:`c-language <tw_c-language>` 編寫的。更準確地說,內核通常是用 :ref:`gcc <tw_gcc>` -在 ``-std=gnu89`` :ref:`gcc-c-dialect-options <tw_gcc-c-dialect-options>` 下編譯的:ISO C90的 GNU 方言( -包括一些C99特性) +在 ``-std=gnu11`` :ref:`gcc-c-dialect-options <tw_gcc-c-dialect-options>` 下編譯的:ISO C11的 GNU 方言 這種方言包含對語言 :ref:`gnu-extensions <tw_gnu-extensions>` 的許多擴展,當然,它們許多都在內核中使用。Makefile +2
-2@@ -515,7 +515,7 @@ KBUILD_CFLAGS := -Wall -Wundef -Werror=strict-prototypes -Wno-trigraphs \ -fno-strict-aliasing -fno-common -fshort-wchar -fno-PIE \ -Werror=implicit-function-declaration -Werror=implicit-int \ -Werror=return-type -Wno-format-security \ - -std=gnu89 + -std=gnu11 KBUILD_CPPFLAGS := -D__KERNEL__ KBUILD_AFLAGS_KERNEL := KBUILD_CFLAGS_KERNEL :=@@ -782,7 +782,7 @@ KBUILD_CFLAGS += $(KBUILD_CFLAGS-y) $(CONFIG_CC_IMPLICIT_FALLTHROUGH) ifdef CONFIG_CC_IS_CLANG KBUILD_CPPFLAGS += -Qunused-arguments -# The kernel builds with '-std=gnu89' so use of GNU extensions is acceptable. +# The kernel builds with '-std=gnu11' so use of GNU extensions is acceptable. KBUILD_CFLAGS += -Wno-gnu # CLANG uses a _MergedGlobals as optimization, but this breaks modpost, as the # source of a reference will be _MergedGlobals and not on of the whitelisted names.arch/arm64/kernel/vdso32/Makefile +1
-1@@ -68,7 +68,7 @@ VDSO_CFLAGS += -Wall -Wundef -Wstrict-prototypes -Wno-trigraphs \ -fno-strict-aliasing -fno-common \ -Werror-implicit-function-declaration \ -Wno-format-security \ - -std=gnu89 + -std=gnu11 VDSO_CFLAGS += -O2 # Some useful compiler-dependent flags from top-level Makefile VDSO_CFLAGS += $(call cc32-option,-Wdeclaration-after-statement,)drivers/gpu/drm/i915/Makefile +1
-0@@ -17,6 +17,7 @@ subdir-ccflags-y += -Wno-unused-parameter subdir-ccflags-y += -Wno-type-limits subdir-ccflags-y += -Wno-missing-field-initializers subdir-ccflags-y += -Wno-sign-compare +subdir-ccflags-y += -Wno-shift-negative-value subdir-ccflags-y += $(call cc-disable-warning, unused-but-set-variable) subdir-ccflags-y += $(call cc-disable-warning, frame-address) subdir-ccflags-$(CONFIG_DRM_I915_WERROR) += -Werrordrivers/staging/greybus/tools/Makefile +2
-1@@ -12,7 +12,8 @@ CFLAGS += -std=gnu99 -Wall -Wextra -g \ -Wredundant-decls \ -Wcast-align \ -Wsign-compare \ - -Wno-missing-field-initializers + -Wno-missing-field-initializers \ + -Wno-shift-negative-value CC := $(CROSS_COMPILE)gccfs/btrfs/Makefile +1
-0@@ -17,6 +17,7 @@ subdir-ccflags-y += $(condflags) subdir-ccflags-y += -Wno-missing-field-initializers subdir-ccflags-y += -Wno-sign-compare subdir-ccflags-y += -Wno-type-limits +subdir-ccflags-y += -Wno-shift-negative-value obj-$(CONFIG_BTRFS_FS) := btrfs.oscripts/Makefile.extrawarn +1
-1@@ -36,6 +36,7 @@ KBUILD_CFLAGS += $(call cc-option, -Wstringop-truncation) KBUILD_CFLAGS += -Wno-missing-field-initializers KBUILD_CFLAGS += -Wno-sign-compare KBUILD_CFLAGS += -Wno-type-limits +KBUILD_CFLAGS += -Wno-shift-negative-value KBUILD_CPPFLAGS += -DKBUILD_EXTRA_WARN1 --2.29.2
Sedat Dilek
Tested-byQuoted 63 lines · Sedat Dilek
Quoted 53 lines · Arnd Bergmann
From: Arnd Bergmann <[email protected]>
During a patch discussion, Linus brought up the option of changing the C standard version from gnu89 to gnu99, which allows using variable declaration inside of a
for()loop. While the C99, C11 and later standards introduce many other features, most of these are already available in gnu89 as GNU extensions as well.An earlier attempt to do this when gcc-5 started defaulting to -std=gnu11 failed because at the time that caused warnings about designated initializers with older compilers. Now that gcc-5.1 is the minimum compiler version used for building kernels, that is no longer a concern. Similarly, the behavior of 'inline' functions changes between gnu89 using
gnu_inlinebehavior and gnu11 using standard c99+ behavior, but this was taken care of by defining 'inline' to include __attribute__((gnu_inline)) in order to allow building with clang a while ago.One minor issue that remains is an added gcc warning for shifts of negative integers when building with -Wextra, which happens with the 'make W=1' option, as well as for three drivers in the kernel that always enable -Wextra, but it was only observed with the i915 driver so far. To be on the safe side, add -Wno-shift-negative-value to any -Wextra in a
Makefile.Nathan Chancellor reported an additional -Wdeclaration-after-statement warning that appears in a system header on arm, this still needs a workaround.
The differences between gnu99, gnu11, gnu1x and gnu17 are fairly minimal and mainly impact warnings at the -Wpedantic level that the kernel never enables. Between these, gnu11 is the newest version that is supported by all supported compiler versions, though it is only the default on gcc-5, while all other supported versions of gcc or clang default to gnu1x/gnu17.
Link https://lore.kernel.org/lkml/CAHk-=wiyCH7xeHcmiFJ-YgXUy2Jaj7pnkdKpcovt8fYbVFW3TA@mail.gmail.com/ open hereSuggested-by Linus Torvalds <[email protected]>Acked-by Marco Elver <[email protected]>Acked-by Jani Nikula <[email protected]>Acked-by David Sterba <[email protected]>Reviewed-by Alex Shi <[email protected]>Reviewed-by Nick Desaulniers <[email protected]>Reviewed-by Miguel Ojeda <[email protected]>Cc Masahiro Yamada <[email protected]>Cc Mark Rutland <[email protected]>Signed-off-by Arnd Bergmann <[email protected]>Quoted 7 lines
---
[v3]
- split out USERCFLAGS to a separate patch
- add -Wdeclaration-after-statement patch from Mark Rutland
- leave out C17 reference
- more rewording the descrption
Quoted 6 lines
I was able to build and boot on bare metal.
No new warnings in my build-log here after switching to -std=gnu11.
I have re-tested with Debian's LLVM/Clang v14.0.0-rc2.
- sed@ -
Quoted 166 lines
- sed@ -
Quoted 164 lines
[v2]
- added -std=gnu11 back, rather than just relying on the default
- minor changes to changelog text--- Documentation/process/programming-language.rst | 6 +++--- .../translations/it_IT/process/programming-language.rst | 4 ++-- .../translations/zh_CN/process/programming-language.rst | 3 +-- .../translations/zh_TW/process/programming-language.rst | 3 +-- Makefile | 4 ++-- arch/arm64/kernel/vdso32/Makefile | 2 +- drivers/gpu/drm/i915/Makefile | 1 + drivers/staging/greybus/tools/Makefile | 3 ++- fs/btrfs/Makefile | 1 + scripts/Makefile.extrawarn | 1 + 10 files changed, 15 insertions(+), 13 deletions(-)Documentation/process/programming-language.rst +3
-3@@ -5,9 +5,9 @@ Programming Language The kernel is written in the C programming language [c-language]_. More precisely, the kernel is typically compiled with ``gcc`` [gcc]_ -under ``-std=gnu89`` [gcc-c-dialect-options]_: the GNU dialect of ISO C90 -(including some C99 features). ``clang`` [clang]_ is also supported, see -docs on :ref:`Building Linux with Clang/LLVM <kbuild_llvm>`. +under ``-std=gnu11`` [gcc-c-dialect-options]_: the GNU dialect of ISO C11. +``clang`` [clang]_ is also supported, see docs on +:ref:`Building Linux with Clang/LLVM <kbuild_llvm>`. This dialect contains many extensions to the language [gnu-extensions]_, and many of them are used within the kernel as a matter of course.Documentation/translations/it_IT/process/programming-language.rst +2
-2@@ -10,8 +10,8 @@ Linguaggio di programmazione Il kernel è scritto nel linguaggio di programmazione C [it-c-language]_. Più precisamente, il kernel viene compilato con ``gcc`` [it-gcc]_ usando -l'opzione ``-std=gnu89`` [it-gcc-c-dialect-options]_: il dialetto GNU -dello standard ISO C90 (con l'aggiunta di alcune funzionalità da C99). +l'opzione ``-std=gnu11`` [it-gcc-c-dialect-options]_: il dialetto GNU +dello standard ISO C11. Linux supporta anche ``clang`` [it-clang]_, leggete la documentazione :ref:`Building Linux with Clang/LLVM <kbuild_llvm>`.Documentation/translations/zh_CN/process/programming-language.rst +1
-2@@ -9,8 +9,7 @@ ============ 内核是用C语言 :ref:`c-language <cn_c-language>` 编写的。更准确地说,内核通常是用 :ref:`gcc <cn_gcc>` -在 ``-std=gnu89`` :ref:`gcc-c-dialect-options <cn_gcc-c-dialect-options>` 下编译的:ISO C90的 GNU 方言( -包括一些C99特性) +在 ``-std=gnu11`` :ref:`gcc-c-dialect-options <cn_gcc-c-dialect-options>` 下编译的:ISO C11的 GNU 方言 这种方言包含对语言 :ref:`gnu-extensions <cn_gnu-extensions>` 的许多扩展,当然,它们许多都在内核中使用。Documentation/translations/zh_TW/process/programming-language.rst +1
-2@@ -12,8 +12,7 @@ ============ 內核是用C語言 :ref:`c-language <tw_c-language>` 編寫的。更準確地說,內核通常是用 :ref:`gcc <tw_gcc>` -在 ``-std=gnu89`` :ref:`gcc-c-dialect-options <tw_gcc-c-dialect-options>` 下編譯的:ISO C90的 GNU 方言( -包括一些C99特性) +在 ``-std=gnu11`` :ref:`gcc-c-dialect-options <tw_gcc-c-dialect-options>` 下編譯的:ISO C11的 GNU 方言 這種方言包含對語言 :ref:`gnu-extensions <tw_gnu-extensions>` 的許多擴展,當然,它們許多都在內核中使用。Makefile +2
-2@@ -515,7 +515,7 @@ KBUILD_CFLAGS := -Wall -Wundef -Werror=strict-prototypes -Wno-trigraphs \ -fno-strict-aliasing -fno-common -fshort-wchar -fno-PIE \ -Werror=implicit-function-declaration -Werror=implicit-int \ -Werror=return-type -Wno-format-security \ - -std=gnu89 + -std=gnu11 KBUILD_CPPFLAGS := -D__KERNEL__ KBUILD_AFLAGS_KERNEL := KBUILD_CFLAGS_KERNEL :=@@ -782,7 +782,7 @@ KBUILD_CFLAGS += $(KBUILD_CFLAGS-y) $(CONFIG_CC_IMPLICIT_FALLTHROUGH) ifdef CONFIG_CC_IS_CLANG KBUILD_CPPFLAGS += -Qunused-arguments -# The kernel builds with '-std=gnu89' so use of GNU extensions is acceptable. +# The kernel builds with '-std=gnu11' so use of GNU extensions is acceptable. KBUILD_CFLAGS += -Wno-gnu # CLANG uses a _MergedGlobals as optimization, but this breaks modpost, as the # source of a reference will be _MergedGlobals and not on of the whitelisted names.arch/arm64/kernel/vdso32/Makefile +1
-1@@ -68,7 +68,7 @@ VDSO_CFLAGS += -Wall -Wundef -Wstrict-prototypes -Wno-trigraphs \ -fno-strict-aliasing -fno-common \ -Werror-implicit-function-declaration \ -Wno-format-security \ - -std=gnu89 + -std=gnu11 VDSO_CFLAGS += -O2 # Some useful compiler-dependent flags from top-level Makefile VDSO_CFLAGS += $(call cc32-option,-Wdeclaration-after-statement,)drivers/gpu/drm/i915/Makefile +1
-0@@ -17,6 +17,7 @@ subdir-ccflags-y += -Wno-unused-parameter subdir-ccflags-y += -Wno-type-limits subdir-ccflags-y += -Wno-missing-field-initializers subdir-ccflags-y += -Wno-sign-compare +subdir-ccflags-y += -Wno-shift-negative-value subdir-ccflags-y += $(call cc-disable-warning, unused-but-set-variable) subdir-ccflags-y += $(call cc-disable-warning, frame-address) subdir-ccflags-$(CONFIG_DRM_I915_WERROR) += -Werrordrivers/staging/greybus/tools/Makefile +2
-1@@ -12,7 +12,8 @@ CFLAGS += -std=gnu99 -Wall -Wextra -g \ -Wredundant-decls \ -Wcast-align \ -Wsign-compare \ - -Wno-missing-field-initializers + -Wno-missing-field-initializers \ + -Wno-shift-negative-value CC := $(CROSS_COMPILE)gccfs/btrfs/Makefile +1
-0@@ -17,6 +17,7 @@ subdir-ccflags-y += $(condflags) subdir-ccflags-y += -Wno-missing-field-initializers subdir-ccflags-y += -Wno-sign-compare subdir-ccflags-y += -Wno-type-limits +subdir-ccflags-y += -Wno-shift-negative-value obj-$(CONFIG_BTRFS_FS) := btrfs.oscripts/Makefile.extrawarn +1
-1@@ -36,6 +36,7 @@ KBUILD_CFLAGS += $(call cc-option, -Wstringop-truncation) KBUILD_CFLAGS += -Wno-missing-field-initializers KBUILD_CFLAGS += -Wno-sign-compare KBUILD_CFLAGS += -Wno-type-limits +KBUILD_CFLAGS += -Wno-shift-negative-value KBUILD_CPPFLAGS += -DKBUILD_EXTRA_WARN1 --2.29.2
Sedat Dilek
Tested-byRe: [PATCH 2/3] treewide: use -Wdeclaration-after-statement
In reply to Sedat DilekQuoted 29 lines · Sedat Dilek
Quoted 18 lines · Arnd Bergmann
From: Mark Rutland <[email protected]>
In a subsequent patch we'll move the kernel from using
-std=gnu89to-std=gnu11, permitting the use of additional C11 features such as for-loop initial declarations.One contentious aspect of C99 is that it permits mixed declarations and code, and for now at least, it seems preferable to enforce that declarations must come first.
These warnings were already disabled in the kernel itself, but not for
KBUILD_USERCFLAGSor the compat VDSO onarch/arm64, which uses a separate set of CFLAGS.This patch fixes an existing violation in
modpost.c, which is not reported because of the missing flag inKBUILD_USERCFLAGS:Quoted 8 lines
| scripts/mod/modpost.c: In function ‘match’: | scripts/mod/modpost.c:837:3: warning: ISO C90 forbids mixed declarations and code [-Wdeclaration-after-statement] | 837 | const char *endp = p + strlen(p) - 1; | | ^~~~~Signed-off-by Mark Rutland <[email protected]>[arnd: don't add a duplicate flag to the default set, update changelog]Signed-off-by Arnd Bergmann <[email protected]>
Quoted 6 lines
I was able to build and boot on bare metal.
No new warnings in my build-log here after switching to -std=gnu11.
I have re-tested with Debian's LLVM/Clang v14.0.0-rc2.
- sed@ -
Quoted 53 lines
- sed@ -
Quoted 51 lines
--- Makefile | 3 ++- arch/arm64/kernel/vdso32/Makefile | 1 + scripts/mod/modpost.c | 4 +++- 3 files changed, 6 insertions(+), 2 deletions(-)Makefile +2
-1@@ -432,7 +432,8 @@ HOSTCXX = g++ endif export KBUILD_USERCFLAGS := -Wall -Wmissing-prototypes -Wstrict-prototypes \ - -O2 -fomit-frame-pointer -std=gnu89 + -O2 -fomit-frame-pointer -std=gnu89 \ + -Wdeclaration-after-statement export KBUILD_USERLDFLAGS := KBUILD_HOSTCFLAGS := $(KBUILD_USERCFLAGS) $(HOST_LFS_CFLAGS) $(HOSTCFLAGS)arch/arm64/kernel/vdso32/Makefile +1
-0@@ -68,6 +68,7 @@ VDSO_CFLAGS += -Wall -Wundef -Wstrict-prototypes -Wno-trigraphs \ -fno-strict-aliasing -fno-common \ -Werror-implicit-function-declaration \ -Wno-format-security \ + -Wdeclaration-after-statement \ -std=gnu11 VDSO_CFLAGS += -O2 # Some useful compiler-dependent flags from top-level Makefilescripts/mod/modpost.c +3
-2@@ -833,8 +833,10 @@ static int match(const char *sym, const char * const pat[]) {const char *p;while (*pat) { +const char *endp; + p = *pat++; -const char *endp = p +strlen (p) -1 ; + endp = p +strlen (p) -1 ;/* "*foo*" */ if (*p =='*' && *endp =='*' ) { --2.29.2
Sedat Dilek
Tested-byRe: [PATCH 3/3] Kbuild: use -std=gnu11 for KBUILD_USERCFLAGS
In reply to Sedat DilekQuoted 88 lines · Sedat Dilek
Quoted 76 lines · Sedat Dilek
Quoted 61 lines · Sedat Dilek
Quoted 36 lines · Nathan Chancellor
Hi Sedat,
Quoted 20 lines · Sedat Dilek
Hey Nick!
This only applies 1/3.
$ b4 --version 0.8.0$ b4 am https://lore.kernel.org/lkml/[email protected]/ -o - | git am -3 Analyzing 14 messages in the thread Will use the latest revision: v3 You can pick other revisions using the -vN flag Checking attestation on all messages, may take a moment... --- ✓ [PATCH v3 1/3] Kbuild: move to -std=gnu11 ✓ Signed: DKIM/kernel.org + Reviewed-by: Nathan Chancellor <[email protected]> (✓ DKIM/kernel.org) ERROR: missing [2/3]! ERROR: missing [3/3]! ---Quoted 8 lines
NOTE: install patatt for end-to-end signature verification --- Total patches: 1 --- WARNING: Thread incomplete! Link: https://lore.kernel.org/r/[email protected] Base: not specified Wende an: Kbuild: move to -std=gnu11It looks like the threading somehow got broken, likely due to the [v3] on the first patch and not the second or third:
Quoted 8 lines
This worked for me on v5.17-rc6:
$ for i in $(seq 1 3); do b4 shazam -P _ 20220301145233.3689119-"$i"[email protected]; done"b4 shazam" is the equivalent of "b4 am -o - ... | git am" and the "-P _" tells b4 to only fetch that exact message ID, not the whole thread.
Hmm, the universe is not good to me...
$ for i in $(seq 1 3); do b4 shazam -P _ 20220301145233.3689119-"$i"[email protected]; done usage: b4 [-h] [--version] [-d] [-q] {mbox,am,attest,pr,ty,diff,kr} ... b4: error: argument subcmd: invalid choice: 'shazam' (choose from 'mbox', 'am', 'attest', 'pr', 'ty', 'diff', 'kr') usage: b4 [-h] [--version] [-d] [-q] {mbox,am,attest,pr,ty,diff,kr} ... b4: error: argument subcmd: invalid choice: 'shazam' (choose from 'mbox', 'am', 'attest', 'pr', 'ty', 'diff', 'kr') usage: b4 [-h] [--version] [-d] [-q] {mbox,am,attest,pr,ty,diff,kr} ... b4: error: argument subcmd: invalid choice: 'shazam' (choose from 'mbox', 'am', 'attest', 'pr', 'ty', 'diff', 'kr')Quoted 9 lines
Do I need a higher version of b4 (here: v0.8.0)?
Check myself... b4.git:
commit
7c1d044ff1d5235e598d4c777c4abfe60e0a09a8("shazam: change default behaviour to be "apply-here"")...is post-v0.8.0.
I brutally applied the post-v0.8.0 patches stolen from b4.git over my local distro b4 files.
Quoted 9 lines
And was able to apply the triple:
$ git log --oneline --no-merges 5.17.0-rc6-1-amd64-clang13-lto..5.17.0-rc6-2-amd64-clang13-lto 96a4222bdd4c (for-5.17/kbuild-std_gnu11-arndb-20220301) Kbuild: use -std=gnu11 for KBUILD_USERCFLAGS c4e8cef401a8 treewide: use -Wdeclaration-after-statement 6a7cc105b238 Kbuild: move to -std=gnu11
Quoted 6 lines
I was able to build and boot on bare metal.
No new warnings in my build-log here after switching to -std=gnu11.
I have re-tested with Debian's LLVM/Clang v14.0.0-rc2.
- sed@ -
On Fri, Mar 4, 2022 at 5:25 PM Sedat Dilek <[email protected]> wrote:
[ ... ]
Quoted 4 lines
When I saw the first patches in this area, I was thinking what about the
tools/build-system?
This ECO system has its own rules.
My other goal was to build perf with my selfmade or distro LLVM/Clang and -std=gnu11.
I was able to do so:
$ git log --oneline --no-merges v5.17-rc6..for-5.17/tools-std_gnu11-dileks-v2 2d99b1cfa897 (for-5.17/tools-std_gnu11-dileks-v2) perf: Use -std=gnu11 c345a183c4ef tools: libtraceevent: Use -std=gnu11 26f77082f233 tools: libsubcmd: Use -std=gnu11 aaba58ec81c8 tools: libbpf: Use -std=gnu11 ffc800af3be4 tools: libapi: Use -std=gnu11
For perf/libperl-support I needed an extra patch:
$ git log --oneline --no-merges v5.17-rc6..for-5.17/perf-libperl_support-clang-dileks 316a1917ec05 (for-5.17/perf-libperl_support-clang-dileks) perf: Fix libperl support with clang and perl v5.34
For details see [1].
I can send the patches if someone wants them.