[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20200423182728.GI26021@zn.tnic>
Date: Thu, 23 Apr 2020 20:27:28 +0200
From: Borislav Petkov <bp@...en8.de>
To: Nick Desaulniers <ndesaulniers@...gle.com>
Cc: Michael Matz <matz@...e.de>, Jakub Jelinek <jakub@...hat.com>,
Sergei Trofimovich <slyfox@...too.org>,
LKML <linux-kernel@...r.kernel.org>,
Thomas Gleixner <tglx@...utronix.de>,
Ingo Molnar <mingo@...hat.com>,
"H. Peter Anvin" <hpa@...or.com>,
Andy Lutomirski <luto@...nel.org>,
Peter Zijlstra <peterz@...radead.org>,
"maintainer:X86 ARCHITECTURE (32-BIT AND 64-BIT)" <x86@...nel.org>,
clang-built-linux <clang-built-linux@...glegroups.com>,
Kees Cook <keescook@...omium.org>,
Martin Liška <mliska@...e.cz>,
Masahiro Yamada <masahiroy@...nel.org>
Subject: Re: [PATCH] x86: Fix early boot crash on gcc-10, next try
On Thu, Apr 23, 2020 at 11:02:09AM -0700, Nick Desaulniers wrote:
> Reviewed-by: Nick Desaulniers <ndesaulniers@...gle.com>
> Tested-by: Nick Desaulniers <ndesaulniers@...gle.com>
>
> It's too bad that $(CC) isn't exported yet; IIUC we include the arch
> specific Makefiles then later export $(CC). If that was the case, we
> could just use $(CC) in the shell script, rather than passing it along
> as an argument. Oh well.
Aha, so that's why others pass it. I used
gcc-x86_64-has-stack-protector.sh as an example to slap that one
together.
Below new version with proper commit message this time.
> If I add `echo "hello world"` to the end of
> scripts/x86-check-compiler.sh to verify this stops a build, this is
> the error message I would observe:
> arch/x86/Makefile:6: *** clang check failed. Aborting. Stop.
Right, or you can comment out the asm("") in the script and then it
matches the "jmp" and thus fails the build. As it should be.
Thx.
---
From: Borislav Petkov <bp@...e.de>
Date: Thu, 23 Apr 2020 19:28:28 +0200
Subject: [PATCH] x86: Check whether the compiler is sane
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
Add a check script to verify whether the compiler is sane. This is
x86-only for now and checks one thing only but should be useful for more
checks in the future.
Suggested-by: Martin Liška <mliska@...e.cz>
Signed-off-by: Borislav Petkov <bp@...e.de>
Reviewed-by: Nick Desaulniers <ndesaulniers@...gle.com>
Tested-by: Nick Desaulniers <ndesaulniers@...gle.com>
---
arch/x86/Makefile | 4 ++++
scripts/x86-check-compiler.sh | 9 +++++++++
2 files changed, 13 insertions(+)
create mode 100755 scripts/x86-check-compiler.sh
diff --git a/arch/x86/Makefile b/arch/x86/Makefile
index 00e378de8bc0..38d3eec5062e 100644
--- a/arch/x86/Makefile
+++ b/arch/x86/Makefile
@@ -1,6 +1,10 @@
# SPDX-License-Identifier: GPL-2.0
# Unified Makefile for i386 and x86_64
+# Check the compiler
+sane_compiler := $(shell $(srctree)/scripts/x86-check-compiler.sh $(CC))
+$(if $(sane_compiler),$(error $(CC) check failed. Aborting),)
+
# select defconfig based on actual architecture
ifeq ($(ARCH),x86)
ifeq ($(shell uname -m),x86_64)
diff --git a/scripts/x86-check-compiler.sh b/scripts/x86-check-compiler.sh
new file mode 100755
index 000000000000..b2b5b54b6939
--- /dev/null
+++ b/scripts/x86-check-compiler.sh
@@ -0,0 +1,9 @@
+#!/bin/sh
+# SPDX-License-Identifier: GPL-2.0
+
+# Check whether the compiler tail-call optimizes across an asm() statement.
+# Fail the build if it does.
+
+echo "int foo(int a); int bar(int a) { int r = foo(a); asm(\"\"); return r; }" |\
+ $* -O2 -x c -c -S - -o - 2>/dev/null |\
+ grep -E "^[[:blank:]]+jmp[[:blank:]]+.*"
--
2.21.0
--
Regards/Gruss,
Boris.
https://people.kernel.org/tglx/notes-about-netiquette
Powered by blists - more mailing lists