[PATCH 13/13] jump_label,x86: Allow short NOPs

From: Peter Zijlstra <peterz@infradead.org>
To: x86@kernel.org, jpoimboe@redhat.com, jbaron@akamai.com,
	rostedt@goodmis.org, ardb@kernel.org
Cc: linux-kernel@vger.kernel.org, peterz@infradead.org
Subject: [PATCH 13/13] jump_label,x86: Allow short NOPs
Date: Thu, 06 May 2021 21:34:05 +0200	[thread overview]
Message-ID: <20210506194158.216763632@infradead.org> (raw)
In-Reply-To: 20210506193352.719596001@infradead.org

Now that objtool is able to rewrite jump_label instructions, have the
compiler emit a JMP, such that it can decide on the optimal encoding,
and set jump_entry::key bit1 to indicate that objtool should rewrite
the instruction to a matching NOP.

For x86_64-allyesconfig this gives:

jl\     NOP     JMP
short:  22997   124
long:   30874   90

IOW, we save (22997+124) * 3 bytes of kernel text in hotpaths.

Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
---
 arch/x86/include/asm/jump_label.h |   18 ++++++++++++++++++
 1 file changed, 18 insertions(+)

--- a/arch/x86/include/asm/jump_label.h
+++ b/arch/x86/include/asm/jump_label.h
@@ -28,6 +28,22 @@
 	_ASM_PTR "%c0 + %c1 - .\n\t"			\
 	".popsection \n\t"
 
+#ifdef CONFIG_STACK_VALIDATION
+
+static __always_inline bool arch_static_branch(struct static_key *key, bool branch)
+{
+	asm_volatile_goto("1:"
+		"jmp %l[l_yes] # objtool NOPs this \n\t"
+		JUMP_TABLE_ENTRY
+		: :  "i" (key), "i" (2 | branch) : : l_yes);
+
+	return false;
+l_yes:
+	return true;
+}
+
+#else
+
 static __always_inline bool arch_static_branch(struct static_key * const key, const bool branch)
 {
 	asm_volatile_goto("1:"
@@ -40,6 +56,8 @@ static __always_inline bool arch_static_
 	return true;
 }
 
+#endif /* STACK_VALIDATION */
+
 static __always_inline bool arch_static_branch_jump(struct static_key * const key, const bool branch)
 {
 	asm_volatile_goto("1:"


