* [PATCH 1/2] vdso: remove meaningless undefining CONFIG_OPTIMIZE_INLINING
@ 2020-02-20 7:11 Masahiro Yamada
2020-02-20 7:11 ` [PATCH 2/2] compiler: Remove CONFIG_OPTIMIZE_INLINING entirely Masahiro Yamada
2020-02-20 9:28 ` [PATCH 1/2] vdso: remove meaningless undefining CONFIG_OPTIMIZE_INLINING Miguel Ojeda
0 siblings, 2 replies; 4+ messages in thread
From: Masahiro Yamada @ 2020-02-20 7:11 UTC (permalink / raw)
To: Andrew Morton
Cc: Arnd Bergmann, Ingo Molnar, Thomas Gleixner, x86,
clang-built-linux, Miguel Ojeda, sparclinux, Masahiro Yamada,
Andy Lutomirski, Borislav Petkov, David S. Miller,
H. Peter Anvin, linux-kernel
The code, #undef CONFIG_OPTIMIZE_INLINING, is not working as expected.
It is because <linux/compiler_types.h> is parsed before vclock_gettime.c
since 28128c61e08e ("kconfig.h: Include compiler types to avoid missed
struct attributes").
Since then, <linux/compiler_types.h> is included really early by
using the '-include' option. So, you cannot negate the decision of
<linux/compiler_types.h> in this way.
You can confirm it by checking the pre-processed code, like this:
$ make arch/x86/entry/vdso/vdso32/vclock_gettime.i
There is no difference with/without CONFIG_CC_OPTIMIZE_FOR_SIZE.
It it about two years since 28128c61e08e. Nobody has reported a
problem (or, nobody has even noticed the fact that this code is not
working).
It is ugly and unreliable to attempt to undefine a CONFIG option from
C files, and anyway the inlining heuristic is up to the compiler.
Just remove the broken code.
Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>
---
arch/sparc/vdso/vdso32/vclock_gettime.c | 4 ----
arch/x86/entry/vdso/vdso32/vclock_gettime.c | 4 ----
2 files changed, 8 deletions(-)
diff --git a/arch/sparc/vdso/vdso32/vclock_gettime.c b/arch/sparc/vdso/vdso32/vclock_gettime.c
index 026abb3b826c..d7f99e6745ea 100644
--- a/arch/sparc/vdso/vdso32/vclock_gettime.c
+++ b/arch/sparc/vdso/vdso32/vclock_gettime.c
@@ -4,10 +4,6 @@
#define BUILD_VDSO32
-#ifndef CONFIG_CC_OPTIMIZE_FOR_SIZE
-#undef CONFIG_OPTIMIZE_INLINING
-#endif
-
#ifdef CONFIG_SPARC64
/*
diff --git a/arch/x86/entry/vdso/vdso32/vclock_gettime.c b/arch/x86/entry/vdso/vdso32/vclock_gettime.c
index 9242b28418d5..3c26488db94d 100644
--- a/arch/x86/entry/vdso/vdso32/vclock_gettime.c
+++ b/arch/x86/entry/vdso/vdso32/vclock_gettime.c
@@ -1,10 +1,6 @@
// SPDX-License-Identifier: GPL-2.0
#define BUILD_VDSO32
-#ifndef CONFIG_CC_OPTIMIZE_FOR_SIZE
-#undef CONFIG_OPTIMIZE_INLINING
-#endif
-
#ifdef CONFIG_X86_64
/*
--
2.17.1
^ permalink raw reply related [flat|nested] 4+ messages in thread
* [PATCH 2/2] compiler: Remove CONFIG_OPTIMIZE_INLINING entirely
2020-02-20 7:11 [PATCH 1/2] vdso: remove meaningless undefining CONFIG_OPTIMIZE_INLINING Masahiro Yamada
@ 2020-02-20 7:11 ` Masahiro Yamada
2020-02-20 9:26 ` Miguel Ojeda
2020-02-20 9:28 ` [PATCH 1/2] vdso: remove meaningless undefining CONFIG_OPTIMIZE_INLINING Miguel Ojeda
1 sibling, 1 reply; 4+ messages in thread
From: Masahiro Yamada @ 2020-02-20 7:11 UTC (permalink / raw)
To: Andrew Morton
Cc: Arnd Bergmann, Ingo Molnar, Thomas Gleixner, x86,
clang-built-linux, Miguel Ojeda, sparclinux, Masahiro Yamada,
Borislav Petkov, H. Peter Anvin, linux-kernel
Commit ac7c3e4ff401 ("compiler: enable CONFIG_OPTIMIZE_INLINING
forcibly") made this always-on option. We released v5.4 and v5.5
including that commit.
Remove the CONFIG option and clean up the code now.
Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>
---
arch/x86/configs/i386_defconfig | 1 -
arch/x86/configs/x86_64_defconfig | 1 -
include/linux/compiler_types.h | 11 +----------
kernel/configs/tiny.config | 1 -
lib/Kconfig.debug | 12 ------------
5 files changed, 1 insertion(+), 25 deletions(-)
diff --git a/arch/x86/configs/i386_defconfig b/arch/x86/configs/i386_defconfig
index 59ce9ed58430..d961d831c266 100644
--- a/arch/x86/configs/i386_defconfig
+++ b/arch/x86/configs/i386_defconfig
@@ -288,7 +288,6 @@ CONFIG_EARLY_PRINTK_DBGP=y
CONFIG_DEBUG_STACKOVERFLOW=y
# CONFIG_DEBUG_RODATA_TEST is not set
CONFIG_DEBUG_BOOT_PARAMS=y
-CONFIG_OPTIMIZE_INLINING=y
CONFIG_SECURITY=y
CONFIG_SECURITY_NETWORK=y
CONFIG_SECURITY_SELINUX=y
diff --git a/arch/x86/configs/x86_64_defconfig b/arch/x86/configs/x86_64_defconfig
index 0b9654c7a05c..4826254c6140 100644
--- a/arch/x86/configs/x86_64_defconfig
+++ b/arch/x86/configs/x86_64_defconfig
@@ -285,7 +285,6 @@ CONFIG_EARLY_PRINTK_DBGP=y
CONFIG_DEBUG_STACKOVERFLOW=y
# CONFIG_DEBUG_RODATA_TEST is not set
CONFIG_DEBUG_BOOT_PARAMS=y
-CONFIG_OPTIMIZE_INLINING=y
CONFIG_UNWINDER_ORC=y
CONFIG_SECURITY=y
CONFIG_SECURITY_NETWORK=y
diff --git a/include/linux/compiler_types.h b/include/linux/compiler_types.h
index 72393a8c1a6c..e970f97a7fcb 100644
--- a/include/linux/compiler_types.h
+++ b/include/linux/compiler_types.h
@@ -129,22 +129,13 @@ struct ftrace_likely_data {
#define __compiler_offsetof(a, b) __builtin_offsetof(a, b)
/*
- * Force always-inline if the user requests it so via the .config.
* Prefer gnu_inline, so that extern inline functions do not emit an
* externally visible function. This makes extern inline behave as per gnu89
* semantics rather than c99. This prevents multiple symbol definition errors
* of extern inline functions at link time.
* A lot of inline functions can cause havoc with function tracing.
- * Do not use __always_inline here, since currently it expands to inline again
- * (which would break users of __always_inline).
*/
-#if !defined(CONFIG_OPTIMIZE_INLINING)
-#define inline inline __attribute__((__always_inline__)) __gnu_inline \
- __inline_maybe_unused notrace
-#else
-#define inline inline __gnu_inline \
- __inline_maybe_unused notrace
-#endif
+#define inline inline __gnu_inline __inline_maybe_unused notrace
/*
* gcc provides both __inline__ and __inline as alternate spellings of
diff --git a/kernel/configs/tiny.config b/kernel/configs/tiny.config
index 7fa0c4ae6394..8a44b93da0f3 100644
--- a/kernel/configs/tiny.config
+++ b/kernel/configs/tiny.config
@@ -6,7 +6,6 @@ CONFIG_CC_OPTIMIZE_FOR_SIZE=y
CONFIG_KERNEL_XZ=y
# CONFIG_KERNEL_LZO is not set
# CONFIG_KERNEL_LZ4 is not set
-CONFIG_OPTIMIZE_INLINING=y
# CONFIG_SLAB is not set
# CONFIG_SLUB is not set
CONFIG_SLOB=y
diff --git a/lib/Kconfig.debug b/lib/Kconfig.debug
index 69def4a9df00..5abde39c3c69 100644
--- a/lib/Kconfig.debug
+++ b/lib/Kconfig.debug
@@ -305,18 +305,6 @@ config HEADERS_INSTALL
user-space program samples. It is also needed by some features such
as uapi header sanity checks.
-config OPTIMIZE_INLINING
- def_bool y
- help
- This option determines if the kernel forces gcc to inline the functions
- developers have marked 'inline'. Doing so takes away freedom from gcc to
- do what it thinks is best, which is desirable for the gcc 3.x series of
- compilers. The gcc 4.x series have a rewritten inlining algorithm and
- enabling this option will generate a smaller kernel there. Hopefully
- this algorithm is so good that allowing gcc 4.x and above to make the
- decision will become the default in the future. Until then this option
- is there to test gcc for this.
-
config DEBUG_SECTION_MISMATCH
bool "Enable full Section mismatch analysis"
help
--
2.17.1
^ permalink raw reply related [flat|nested] 4+ messages in thread
* Re: [PATCH 2/2] compiler: Remove CONFIG_OPTIMIZE_INLINING entirely
2020-02-20 7:11 ` [PATCH 2/2] compiler: Remove CONFIG_OPTIMIZE_INLINING entirely Masahiro Yamada
@ 2020-02-20 9:26 ` Miguel Ojeda
0 siblings, 0 replies; 4+ messages in thread
From: Miguel Ojeda @ 2020-02-20 9:26 UTC (permalink / raw)
To: Masahiro Yamada
Cc: Andrew Morton, Arnd Bergmann, Ingo Molnar, Thomas Gleixner,
maintainer:X86 ARCHITECTURE (32-BIT AND 64-BIT),
clang-built-linux, sparclinux, Borislav Petkov, H. Peter Anvin,
linux-kernel
On Thu, Feb 20, 2020 at 8:12 AM Masahiro Yamada <masahiroy@kernel.org> wrote:
>
> Commit ac7c3e4ff401 ("compiler: enable CONFIG_OPTIMIZE_INLINING
> forcibly") made this always-on option. We released v5.4 and v5.5
> including that commit.
>
> Remove the CONFIG option and clean up the code now.
Yes, please!
Reviewed-by: Miguel Ojeda <miguel.ojeda.sandonis@gmail.com>
Cheers,
Miguel
^ permalink raw reply [flat|nested] 4+ messages in thread
* Re: [PATCH 1/2] vdso: remove meaningless undefining CONFIG_OPTIMIZE_INLINING
2020-02-20 7:11 [PATCH 1/2] vdso: remove meaningless undefining CONFIG_OPTIMIZE_INLINING Masahiro Yamada
2020-02-20 7:11 ` [PATCH 2/2] compiler: Remove CONFIG_OPTIMIZE_INLINING entirely Masahiro Yamada
@ 2020-02-20 9:28 ` Miguel Ojeda
1 sibling, 0 replies; 4+ messages in thread
From: Miguel Ojeda @ 2020-02-20 9:28 UTC (permalink / raw)
To: Masahiro Yamada
Cc: Andrew Morton, Arnd Bergmann, Ingo Molnar, Thomas Gleixner,
maintainer:X86 ARCHITECTURE (32-BIT AND 64-BIT),
clang-built-linux, sparclinux, Andy Lutomirski, Borislav Petkov,
David S. Miller, H. Peter Anvin, linux-kernel
On Thu, Feb 20, 2020 at 8:12 AM Masahiro Yamada <masahiroy@kernel.org> wrote:
>
> It it about two years since 28128c61e08e. Nobody has reported a
Nit: "It it" -> "it is"
> It is ugly and unreliable to attempt to undefine a CONFIG option from
> C files, and anyway the inlining heuristic is up to the compiler.
+1
Acked-by: Miguel Ojeda <miguel.ojeda.sandonis@gmail.com>
Cheers,
Miguel
^ permalink raw reply [flat|nested] 4+ messages in thread
end of thread, other threads:[~2020-02-20 9:28 UTC | newest]
Thread overview: 4+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2020-02-20 7:11 [PATCH 1/2] vdso: remove meaningless undefining CONFIG_OPTIMIZE_INLINING Masahiro Yamada
2020-02-20 7:11 ` [PATCH 2/2] compiler: Remove CONFIG_OPTIMIZE_INLINING entirely Masahiro Yamada
2020-02-20 9:26 ` Miguel Ojeda
2020-02-20 9:28 ` [PATCH 1/2] vdso: remove meaningless undefining CONFIG_OPTIMIZE_INLINING Miguel Ojeda
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).