[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20221028194453.592512209@infradead.org>
Date: Fri, 28 Oct 2022 21:40:27 +0200
From: Peter Zijlstra <peterz@...radead.org>
To: x86@...nel.org
Cc: linux-kernel@...r.kernel.org, peterz@...radead.org,
djwong@...nel.org, yujie.liu@...el.com, tglx@...utronix.de,
jpoimboe@...nel.org, joao.moreira@...el.com,
samitolvanen@...gle.com
Subject: [PATCH 5/5] x86: Add prefix symbols for function padding
When code is compiled with:
-fpatchable-function-entry=${PADDING_BYTES},${PADDING_BYTES}
functions will have PADDING_BYTES of NOP in front of them. Unwinders
and other things that symbolize code locations will typically
attribute these bytes to the preceding function.
Given that these bytes nominally belong to the following symbol this
mis-attribution is confusing.
Inspired by the fact that CFI_CLANG emits __cfi_##name symbols to
claim these bytes, use objtool to emit __pfx_##name symbols to do
the same when CFI_CLANG is not used.
This then shows the callthunk for symbol 'name' as:
__pfx_##name+0x6/0x10
Signed-off-by: Peter Zijlstra (Intel) <peterz@...radead.org>
---
arch/x86/Kconfig | 4 ++++
scripts/Makefile.lib | 1 +
2 files changed, 5 insertions(+)
--- a/arch/x86/Kconfig
+++ b/arch/x86/Kconfig
@@ -2472,6 +2472,10 @@ config CALL_THUNKS
def_bool n
select FUNCTION_ALIGNMENT_16B
+config PREFIX_SYMBOLS
+ def_bool y
+ depends on CALL_THUNKS && !CFI_CLANG
+
menuconfig SPECULATION_MITIGATIONS
bool "Mitigations for speculative execution vulnerabilities"
default y
--- a/scripts/Makefile.lib
+++ b/scripts/Makefile.lib
@@ -265,6 +265,7 @@ objtool-args-$(CONFIG_STACK_VALIDATION)
objtool-args-$(CONFIG_HAVE_STATIC_CALL_INLINE) += --static-call
objtool-args-$(CONFIG_HAVE_UACCESS_VALIDATION) += --uaccess
objtool-args-$(CONFIG_GCOV_KERNEL) += --no-unreachable
+objtool-args-$(CONFIG_PREFIX_SYMBOLS) += --prefix=$(CONFIG_FUNCTION_PADDING_BYTES)
objtool-args = $(objtool-args-y) \
$(if $(delay-objtool), --link) \
Powered by blists - more mailing lists