Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
MIME-Version: 1.0
Message-ID: <87411705-893f-46d3-b899-b09ed9fa8d1b@default>
Date:   Mon, 17 Sep 2018 22:17:30 -0700 (PDT)
From:   Zhenzhong Duan <zhenzhong.duan@oracle.com>
To:     <linux-kernel@vger.kernel.org>
Cc:     <mingo@redhat.com>, <konrad.wilk@oracle.com>, <x86@kernel.org>,
        <dwmw@amazon.co.uk>, <tglx@linutronix.de>,
        Srinivas REDDY Eeda <srinivas.eeda@oracle.com>, <bp@suse.de>,
        <hpa@zytor.com>
Subject: [PATCH] x86/speculation: Use AMD specific retpoline for inline asm on
 AMD
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: quoted-printable
Content-Disposition: inline
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

Lfence is preferred than general retpoline on AMD, add this option
in C / inline asm just as the ASM code does.

For x86_64, it still help to have minimal retpoline for kernel even
if gcc doesn't support it, change the inline asm for x86 so that it
could also be used by x86_64.

Add ANNOTATE_NOSPEC_ALTERNATIVE for i386 to avoid below warning:
"warning: objtool: .altinstr_replacement+0x10: unsupported
intra-function call"
"warning: objtool: If this is a retpoline, please patch it
in with alternatives and annotate it with ANNOTATE_NOSPEC_ALTERNATIVE."

Signed-off-by: Zhenzhong Duan <zhenzhong.duan@oracle.com>
---
 arch/x86/include/asm/nospec-branch.h |   23 ++++++++++++++++-------
 1 files changed, 16 insertions(+), 7 deletions(-)

diff --git a/arch/x86/include/asm/nospec-branch.h b/arch/x86/include/asm/no=
spec-branch.h
index fd2a8c1..2d49eab 100644
--- a/arch/x86/include/asm/nospec-branch.h
+++ b/arch/x86/include/asm/nospec-branch.h
@@ -170,21 +170,26 @@
  */
 # define CALL_NOSPEC=09=09=09=09=09=09\
 =09ANNOTATE_NOSPEC_ALTERNATIVE=09=09=09=09\
-=09ALTERNATIVE(=09=09=09=09=09=09\
+=09ALTERNATIVE_2(=09=09=09=09=09=09\
 =09ANNOTATE_RETPOLINE_SAFE=09=09=09=09=09\
 =09"call *%[thunk_target]\n",=09=09=09=09\
 =09"call __x86_indirect_thunk_%V[thunk_target]\n",=09=09\
-=09X86_FEATURE_RETPOLINE)
+=09X86_FEATURE_RETPOLINE,=09=09=09=09=09\
+=09"lfence;\n"=09=09=09=09=09=09\
+=09ANNOTATE_RETPOLINE_SAFE=09=09=09=09=09\
+=09"call *%[thunk_target]\n",=09=09=09=09\
+=09X86_FEATURE_RETPOLINE_AMD)
 # define THUNK_TARGET(addr) [thunk_target] "r" (addr)
=20
-#elif defined(CONFIG_X86_32) && defined(CONFIG_RETPOLINE)
+#elif defined(CONFIG_RETPOLINE)
 /*
  * For i386 we use the original ret-equivalent retpoline, because
  * otherwise we'll run out of registers. We don't care about CET
  * here, anyway.
  */
 # define CALL_NOSPEC=09=09=09=09=09=09\
-=09ALTERNATIVE(=09=09=09=09=09=09\
+=09ANNOTATE_NOSPEC_ALTERNATIVE=09=09=09=09\
+=09ALTERNATIVE_2(=09=09=09=09=09=09\
 =09ANNOTATE_RETPOLINE_SAFE=09=09=09=09=09\
 =09"call *%[thunk_target]\n",=09=09=09=09\
 =09"       jmp    904f;\n"=09=09=09=09=09\
@@ -194,12 +199,16 @@
 =09"    =09lfence;\n"=09=09=09=09=09\
 =09"       jmp    902b;\n"=09=09=09=09=09\
 =09"       .align 16\n"=09=09=09=09=09\
-=09"903:=09addl   $4, %%esp;\n"=09=09=09=09\
-=09"       pushl  %[thunk_target];\n"=09=09=09\
+=09"903:=09add   $4, %%" _ASM_SP ";\n"=09=09=09\
+=09"       push  %[thunk_target];\n"=09=09=09\
 =09"       ret;\n"=09=09=09=09=09=09\
 =09"       .align 16\n"=09=09=09=09=09\
 =09"904:=09call   901b;\n",=09=09=09=09\
-=09X86_FEATURE_RETPOLINE)
+=09X86_FEATURE_RETPOLINE,=09=09=09=09=09\
+=09"lfence;\n"=09=09=09=09=09=09\
+=09ANNOTATE_RETPOLINE_SAFE=09=09=09=09=09\
+=09"call *%[thunk_target]\n",=09=09=09=09\
+=09X86_FEATURE_RETPOLINE_AMD)
=20
 # define THUNK_TARGET(addr) [thunk_target] "rm" (addr)
 #else /* No retpoline for C / inline asm */
--=20
1.7.3