From: Ingo Molnar <mingo@kernel.org>
To: Masahiro Yamada <yamada.masahiro@socionext.com>
Cc: x86@kernel.org, Thomas Gleixner <tglx@linutronix.de>,
Ingo Molnar <mingo@redhat.com>, Borislav Petkov <bp@alien8.de>,
"H . Peter Anvin" <hpa@zytor.com>,
Richard Biener <rguenther@suse.de>,
Segher Boessenkool <segher@kernel.crashing.org>,
Peter Zijlstra <peterz@infradead.org>,
Juergen Gross <jgross@suse.com>,
Josh Poimboeuf <jpoimboe@redhat.com>,
Kees Cook <keescook@chromium.org>,
Linus Torvalds <torvalds@linux-foundation.org>,
Arnd Bergmann <arnd@arndb.de>,
Andrey Ryabinin <aryabinin@virtuozzo.com>,
virtualization@lists.linux-foundation.org,
Luc Van Oostenryck <luc.vanoostenryck@gmail.com>,
Alok Kataria <akataria@vmware.com>,
Ard Biesheuvel <ard.biesheuvel@linaro.org>,
Jann Horn <jannh@google.com>,
linux-arch@vger.kernel.org, Alexey Dobriyan <adobriyan@gmail.com>,
linux-sparse@vger.kernel.org,
Andrew Morton <akpm@linux-foundation.org>,
linux-kbuild@vger.kernel.org, Yonghong Song <yhs@fb.com>,
Michal Marek <michal.lkml@markovi.net>,
Arnaldo Carvalho de Melo <acme@redhat.com>,
Jan Beulich <JBeulich@suse.com>, Nadav Amit <namit@vmware.com>,
David Woodhouse <dwmw@amazon.co.uk>,
Alexei Starovoitov <ast@kernel.org>,
linux-kernel@vger.kernel.org
Subject: Re: [PATCH v3 00/12] x86, kbuild: revert macrofying inline assembly code
Date: Wed, 19 Dec 2018 12:20:28 +0100 [thread overview]
Message-ID: <20181219112028.GA38175@gmail.com> (raw)
In-Reply-To: <1545062607-8599-1-git-send-email-yamada.masahiro@socionext.com>
* Masahiro Yamada <yamada.masahiro@socionext.com> wrote:
> This series reverts the in-kernel workarounds for inlining issues.
>
> The commit description of 77b0bf55bc67 mentioned
> "We also hope that GCC will eventually get fixed,..."
>
> Now, GCC provides a solution.
>
> https://gcc.gnu.org/onlinedocs/gcc/Extended-Asm.html
> explains the new "asm inline" syntax.
>
> The performance issue will be eventually solved.
>
> [About Code cleanups]
>
> I know Nadam Amit is opposed to the full revert.
> He also claims his motivation for macrofying was not only
> performance, but also cleanups.
>
> IIUC, the criticism addresses the code duplication between C and ASM.
>
> If so, I'd like to suggest a different approach for cleanups.
> Please see the last 3 patches.
> IMHO, preprocessor approach is more straight-forward, and readable.
> Basically, this idea should work because it is what we already do for
> __ASM_FORM() etc.
>
> [Quick Test of "asm inline" of GCC 9]
>
> If you want to try "asm inline" feature, the patch is available:
> https://lore.kernel.org/patchwork/patch/1024590/
>
> The number of symbols for arch/x86/configs/x86_64_defconfig:
>
> nr_symbols
> [1] v4.20-rc7 : 96502
> [2] [1]+full revert : 96705 (+203)
> [3] [2]+"asm inline": 96568 (-137)
>
> [3]: apply my patch, then replace "asm" -> "asm_inline"
> for _BUG_FLAGS(), refcount_add(), refcount_inc(), refcount_dec(),
> annotate_reachable(), annotate_unreachable()
>
>
> Changes in v3:
> - Split into per-commit revert (per Nadav Amit)
> - Add some cleanups with preprocessor approach
>
> Changes in v2:
> - Revive clean-ups made by 5bdcd510c2ac (per Peter Zijlstra)
> - Fix commit quoting style (per Peter Zijlstra)
>
> Masahiro Yamada (12):
> Revert "x86/jump-labels: Macrofy inline assembly code to work around
> GCC inlining bugs"
> Revert "x86/cpufeature: Macrofy inline assembly code to work around
> GCC inlining bugs"
> Revert "x86/extable: Macrofy inline assembly code to work around GCC
> inlining bugs"
> Revert "x86/paravirt: Work around GCC inlining bugs when compiling
> paravirt ops"
> Revert "x86/bug: Macrofy the BUG table section handling, to work
> around GCC inlining bugs"
> Revert "x86/alternatives: Macrofy lock prefixes to work around GCC
> inlining bugs"
> Revert "x86/refcount: Work around GCC inlining bug"
> Revert "x86/objtool: Use asm macros to work around GCC inlining bugs"
> Revert "kbuild/Makefile: Prepare for using macros in inline assembly
> code to work around asm() related GCC inlining bugs"
> linux/linkage: add ASM() macro to reduce duplication between C/ASM
> code
> x86/alternatives: consolidate LOCK_PREFIX macro
> x86/asm: consolidate ASM_EXTABLE_* macros
>
> Makefile | 9 +--
> arch/x86/Makefile | 7 ---
> arch/x86/include/asm/alternative-asm.h | 22 +------
> arch/x86/include/asm/alternative-common.h | 47 +++++++++++++++
> arch/x86/include/asm/alternative.h | 30 +---------
> arch/x86/include/asm/asm.h | 46 +++++----------
> arch/x86/include/asm/bug.h | 98 +++++++++++++------------------
> arch/x86/include/asm/cpufeature.h | 82 +++++++++++---------------
> arch/x86/include/asm/jump_label.h | 22 +++++--
> arch/x86/include/asm/paravirt_types.h | 56 +++++++++---------
> arch/x86/include/asm/refcount.h | 81 +++++++++++--------------
> arch/x86/kernel/macros.S | 16 -----
> include/asm-generic/bug.h | 8 +--
> include/linux/compiler.h | 56 ++++--------------
> include/linux/linkage.h | 8 +++
> scripts/Kbuild.include | 4 +-
> scripts/mod/Makefile | 2 -
> 17 files changed, 249 insertions(+), 345 deletions(-)
> create mode 100644 arch/x86/include/asm/alternative-common.h
> delete mode 100644 arch/x86/kernel/macros.S
I absolutely agree that this needs to be resolved in v4.20.
So I did the 1-9 reverts manually myself as well, because I think the
first commit should be reverted fully to get as close to the starting
point as possible (we are late in the cycle) - and came to the attached
interdiff between your series and mine.
Does this approach look OK to you, or did I miss something?
Thanks,
Ingo
=============>
entry/calling.h | 2 -
include/asm/jump_label.h | 50 ++++++++++++++++++++++++++++++++++-------------
2 files changed, 38 insertions(+), 14 deletions(-)
diff --git a/arch/x86/entry/calling.h b/arch/x86/entry/calling.h
index 25e5a6bda8c3..20d0885b00fb 100644
--- a/arch/x86/entry/calling.h
+++ b/arch/x86/entry/calling.h
@@ -352,7 +352,7 @@ For 32-bit we have the following conventions - kernel is built with
.macro CALL_enter_from_user_mode
#ifdef CONFIG_CONTEXT_TRACKING
#ifdef HAVE_JUMP_LABEL
- STATIC_BRANCH_JMP l_yes=.Lafter_call_\@, key=context_tracking_enabled, branch=1
+ STATIC_JUMP_IF_FALSE .Lafter_call_\@, context_tracking_enabled, def=0
#endif
call enter_from_user_mode
.Lafter_call_\@:
diff --git a/arch/x86/include/asm/jump_label.h b/arch/x86/include/asm/jump_label.h
index cf88ebf9a4ca..21efc9d07ed9 100644
--- a/arch/x86/include/asm/jump_label.h
+++ b/arch/x86/include/asm/jump_label.h
@@ -2,6 +2,19 @@
#ifndef _ASM_X86_JUMP_LABEL_H
#define _ASM_X86_JUMP_LABEL_H
+#ifndef HAVE_JUMP_LABEL
+/*
+ * For better or for worse, if jump labels (the gcc extension) are missing,
+ * then the entire static branch patching infrastructure is compiled out.
+ * If that happens, the code in here will malfunction. Raise a compiler
+ * error instead.
+ *
+ * In theory, jump labels and the static branch patching infrastructure
+ * could be decoupled to fix this.
+ */
+#error asm/jump_label.h included on a non-jump-label kernel
+#endif
+
#define JUMP_LABEL_NOP_SIZE 5
#ifdef CONFIG_X86_64
@@ -53,26 +66,37 @@ static __always_inline bool arch_static_branch_jump(struct static_key *key, bool
#else /* __ASSEMBLY__ */
-.macro STATIC_BRANCH_NOP l_yes:req key:req branch:req
-.Lstatic_branch_nop_\@:
- .byte STATIC_KEY_INIT_NOP
-.Lstatic_branch_no_after_\@:
+.macro STATIC_JUMP_IF_TRUE target, key, def
+.Lstatic_jump_\@:
+ .if \def
+ /* Equivalent to "jmp.d32 \target" */
+ .byte 0xe9
+ .long \target - .Lstatic_jump_after_\@
+.Lstatic_jump_after_\@:
+ .else
+ .byte STATIC_KEY_INIT_NOP
+ .endif
.pushsection __jump_table, "aw"
_ASM_ALIGN
- .long .Lstatic_branch_nop_\@ - ., \l_yes - .
- _ASM_PTR \key + \branch - .
+ .long .Lstatic_jump_\@ - ., \target - .
+ _ASM_PTR \key - .
.popsection
.endm
-.macro STATIC_BRANCH_JMP l_yes:req key:req branch:req
-.Lstatic_branch_jmp_\@:
- .byte 0xe9
- .long \l_yes - .Lstatic_branch_jmp_after_\@
-.Lstatic_branch_jmp_after_\@:
+.macro STATIC_JUMP_IF_FALSE target, key, def
+.Lstatic_jump_\@:
+ .if \def
+ .byte STATIC_KEY_INIT_NOP
+ .else
+ /* Equivalent to "jmp.d32 \target" */
+ .byte 0xe9
+ .long \target - .Lstatic_jump_after_\@
+.Lstatic_jump_after_\@:
+ .endif
.pushsection __jump_table, "aw"
_ASM_ALIGN
- .long .Lstatic_branch_jmp_\@ - ., \l_yes - .
- _ASM_PTR \key + \branch - .
+ .long .Lstatic_jump_\@ - ., \target - .
+ _ASM_PTR \key + 1 - .
.popsection
.endm
next prev parent reply other threads:[~2018-12-19 11:20 UTC|newest]
Thread overview: 17+ messages / expand[flat|nested] mbox.gz Atom feed top
2018-12-17 16:03 [PATCH v3 00/12] x86, kbuild: revert macrofying inline assembly code Masahiro Yamada
2018-12-17 16:03 ` [PATCH v3 01/12] Revert "x86/jump-labels: Macrofy inline assembly code to work around GCC inlining bugs" Masahiro Yamada
2018-12-17 16:03 ` [PATCH v3 02/12] Revert "x86/cpufeature: " Masahiro Yamada
2018-12-17 16:03 ` [PATCH v3 03/12] Revert "x86/extable: " Masahiro Yamada
2018-12-17 16:03 ` [PATCH v3 04/12] Revert "x86/paravirt: Work around GCC inlining bugs when compiling paravirt ops" Masahiro Yamada
2018-12-17 16:03 ` [PATCH v3 05/12] Revert "x86/bug: Macrofy the BUG table section handling, to work around GCC inlining bugs" Masahiro Yamada
2018-12-17 16:03 ` [PATCH v3 06/12] Revert "x86/alternatives: Macrofy lock prefixes " Masahiro Yamada
2018-12-17 16:03 ` [PATCH v3 07/12] Revert "x86/refcount: Work around GCC inlining bug" Masahiro Yamada
2018-12-17 16:03 ` [PATCH v3 08/12] Revert "x86/objtool: Use asm macros to work around GCC inlining bugs" Masahiro Yamada
2018-12-17 16:03 ` [PATCH v3 09/12] Revert "kbuild/Makefile: Prepare for using macros in inline assembly code to work around asm() related " Masahiro Yamada
2018-12-17 16:03 ` [PATCH v3 10/12] linux/linkage: add ASM() macro to reduce duplication between C/ASM code Masahiro Yamada
2018-12-17 16:03 ` [PATCH v3 11/12] x86/alternatives: consolidate LOCK_PREFIX macro Masahiro Yamada
2018-12-17 16:03 ` [PATCH v3 12/12] x86/asm: consolidate ASM_EXTABLE_* macros Masahiro Yamada
2018-12-18 19:43 ` [PATCH v3 00/12] x86, kbuild: revert macrofying inline assembly code Nadav Amit
2018-12-19 3:19 ` Masahiro Yamada
2018-12-19 11:20 ` Ingo Molnar [this message]
2018-12-19 14:33 ` Masahiro Yamada
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20181219112028.GA38175@gmail.com \
--to=mingo@kernel.org \
--cc=JBeulich@suse.com \
--cc=acme@redhat.com \
--cc=adobriyan@gmail.com \
--cc=akataria@vmware.com \
--cc=akpm@linux-foundation.org \
--cc=ard.biesheuvel@linaro.org \
--cc=arnd@arndb.de \
--cc=aryabinin@virtuozzo.com \
--cc=ast@kernel.org \
--cc=bp@alien8.de \
--cc=dwmw@amazon.co.uk \
--cc=hpa@zytor.com \
--cc=jannh@google.com \
--cc=jgross@suse.com \
--cc=jpoimboe@redhat.com \
--cc=keescook@chromium.org \
--cc=linux-arch@vger.kernel.org \
--cc=linux-kbuild@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-sparse@vger.kernel.org \
--cc=luc.vanoostenryck@gmail.com \
--cc=michal.lkml@markovi.net \
--cc=mingo@redhat.com \
--cc=namit@vmware.com \
--cc=peterz@infradead.org \
--cc=rguenther@suse.de \
--cc=segher@kernel.crashing.org \
--cc=tglx@linutronix.de \
--cc=torvalds@linux-foundation.org \
--cc=virtualization@lists.linux-foundation.org \
--cc=x86@kernel.org \
--cc=yamada.masahiro@socionext.com \
--cc=yhs@fb.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox