lists.openwall.net   lists  /  announce  owl-users  owl-dev  john-users  john-dev  passwdqc-users  yescrypt  popa3d-users  /  oss-security  kernel-hardening  musl  sabotage  tlsify  passwords  /  crypt-dev  xvendor  /  Bugtraq  Full-Disclosure  linux-kernel  linux-netdev  linux-ext4  linux-hardening  linux-cve-announce  PHC 
Open Source and information security mailing list archives
 
Hash Suite: Windows password security audit tool. GUI, reports in PDF.
[<prev] [next>] [thread-next>] [day] [month] [year] [list]
Message-ID: <20251103185237.2284456-1-benniu@meta.com>
Date: Mon,  3 Nov 2025 10:52:35 -0800
From: Ben Niu <niuben003@...il.com>
To: Catalin Marinas <catalin.marinas@....com>
Cc: Mark Rutland <mark.rutland@....com>,
	Will Deacon <will@...nel.org>,
	tytso@....edu,
	Jason@...c4.com,
	Ben Niu <benniu@...a.com>,
	linux-arm-kernel@...ts.infradead.org,
	linux-kernel@...r.kernel.org
Subject: [PATCH v2] tracing: Enable kprobe for selected Arm64 assembly

When ftrace is enabled, a function can only be kprobe'd if
  (1) the function has proper nops inserted at the beginning
  (2) the first inserted nop's address is added in section
      __patchable_function_entries.

See function within_notrace_func in kernel/trace/trace_kprobe.c
for more details.

This patch adds a new asm function macro SYM_FUNC_START_TRACE
that makes an asm funtion satisfy the above two conditions so that
it can be kprobe'd. In addition, the macro is applied to
__arch_copy_to_user and __arch_copy_from_user, which were found
to be hot in certain workloads.

Note: although this patch unblocks kprobe tracing, fentry is still
broken because no BTF info gets generated from assembly files. A
separate patch is needed to fix that.

This patch was built with different combos of the following
features and kprobe tracing for asm functions __arch_copy_to_user
and __arch_copy_from_user worked in all cases:

CONFIG_DYNAMIC_FTRACE_WITH_CALL_OPS
CONFIG_DYNAMIC_FTRACE_WITH_ARGS
CONFIG_ARM64_BTI_KERNEL

Signed-off-by: Ben Niu <benniu@...a.com>
---
v2: only opt-in __arch_copy_from_user and __arch_copy_to_user for
    tracing since tracing asm functions in general may be unsafe

v1: aQOpd14RXnyQLOWR@...a.com
---
 arch/arm64/include/asm/linkage.h | 91 ++++++++++++++++++++++++--------
 arch/arm64/lib/copy_from_user.S  |  2 +-
 arch/arm64/lib/copy_to_user.S    |  2 +-
 3 files changed, 70 insertions(+), 25 deletions(-)

diff --git a/arch/arm64/include/asm/linkage.h b/arch/arm64/include/asm/linkage.h
index d3acd9c87509..b9a7513dc7d9 100644
--- a/arch/arm64/include/asm/linkage.h
+++ b/arch/arm64/include/asm/linkage.h
@@ -5,8 +5,47 @@
 #include <asm/assembler.h>
 #endif
 
-#define __ALIGN		.balign CONFIG_FUNCTION_ALIGNMENT
-#define __ALIGN_STR	".balign " #CONFIG_FUNCTION_ALIGNMENT
+#define __ALIGN .balign CONFIG_FUNCTION_ALIGNMENT
+#define __ALIGN_STR ".balign " #CONFIG_FUNCTION_ALIGNMENT
+
+#ifdef CONFIG_DYNAMIC_FTRACE_WITH_CALL_OPS
+
+#define PRE_FUNCTION_NOPS                                                   \
+	ALIGN;                                                              \
+	nops CONFIG_FUNCTION_ALIGNMENT / 4 - 2;                             \
+	.pushsection __patchable_function_entries, "awo", @progbits, .text; \
+	.p2align 3;                                                         \
+	.8byte 1f;                                                          \
+	.popsection;                                                        \
+	1 :;                                                                \
+	nops 2;
+
+#define PRE_PROLOGUE_NOPS nops 2;
+
+#elif defined(CONFIG_DYNAMIC_FTRACE_WITH_ARGS)
+
+#define PRE_FUNCTION_NOPS
+
+#define PRE_PROLOGUE_NOPS                                                   \
+	.pushsection __patchable_function_entries, "awo", @progbits, .text; \
+	.p2align 3;                                                         \
+	.8byte 1f;                                                          \
+	.popsection;                                                        \
+	1 :;                                                                \
+	nops 2;
+
+#else
+
+#define PRE_FUNCTION_NOPS
+#define PRE_PROLOGUE_NOPS
+
+#endif
+
+#ifdef CONFIG_ARM64_BTI_KERNEL
+#define BTI_C bti c;
+#else
+#define BTI_C
+#endif
 
 /*
  * When using in-kernel BTI we need to ensure that PCS-conformant
@@ -15,32 +54,38 @@
  * everything, the override is done unconditionally so we're more
  * likely to notice any drift from the overridden definitions.
  */
-#define SYM_FUNC_START(name)				\
-	SYM_START(name, SYM_L_GLOBAL, SYM_A_ALIGN)	\
-	bti c ;
+#define SYM_FUNC_START_TRACE(name)                 \
+	PRE_FUNCTION_NOPS                          \
+	SYM_START(name, SYM_L_GLOBAL, SYM_A_ALIGN) \
+	BTI_C                                      \
+	PRE_PROLOGUE_NOPS
+
+#define SYM_FUNC_START(name)                       \
+	SYM_START(name, SYM_L_GLOBAL, SYM_A_ALIGN) \
+	BTI_C
 
-#define SYM_FUNC_START_NOALIGN(name)			\
-	SYM_START(name, SYM_L_GLOBAL, SYM_A_NONE)	\
-	bti c ;
+#define SYM_FUNC_START_NOALIGN(name)              \
+	SYM_START(name, SYM_L_GLOBAL, SYM_A_NONE) \
+	BTI_C                                     \
 
-#define SYM_FUNC_START_LOCAL(name)			\
-	SYM_START(name, SYM_L_LOCAL, SYM_A_ALIGN)	\
-	bti c ;
+#define SYM_FUNC_START_LOCAL(name)                \
+	SYM_START(name, SYM_L_LOCAL, SYM_A_ALIGN) \
+	BTI_C                                     \
 
-#define SYM_FUNC_START_LOCAL_NOALIGN(name)		\
-	SYM_START(name, SYM_L_LOCAL, SYM_A_NONE)	\
-	bti c ;
+#define SYM_FUNC_START_LOCAL_NOALIGN(name)       \
+	SYM_START(name, SYM_L_LOCAL, SYM_A_NONE) \
+	BTI_C                                    \
 
-#define SYM_FUNC_START_WEAK(name)			\
-	SYM_START(name, SYM_L_WEAK, SYM_A_ALIGN)	\
-	bti c ;
+#define SYM_FUNC_START_WEAK(name)                \
+	SYM_START(name, SYM_L_WEAK, SYM_A_ALIGN) \
+	BTI_C                                    \
 
-#define SYM_FUNC_START_WEAK_NOALIGN(name)		\
-	SYM_START(name, SYM_L_WEAK, SYM_A_NONE)		\
-	bti c ;
+#define SYM_FUNC_START_WEAK_NOALIGN(name)       \
+	SYM_START(name, SYM_L_WEAK, SYM_A_NONE) \
+	BTI_C                                   \
 
-#define SYM_TYPED_FUNC_START(name)				\
-	SYM_TYPED_START(name, SYM_L_GLOBAL, SYM_A_ALIGN)	\
-	bti c ;
+#define SYM_TYPED_FUNC_START(name)                       \
+	SYM_TYPED_START(name, SYM_L_GLOBAL, SYM_A_ALIGN) \
+	BTI_C                                            \
 
 #endif
diff --git a/arch/arm64/lib/copy_from_user.S b/arch/arm64/lib/copy_from_user.S
index 400057d607ec..07b6d6c9573f 100644
--- a/arch/arm64/lib/copy_from_user.S
+++ b/arch/arm64/lib/copy_from_user.S
@@ -61,7 +61,7 @@
 
 end	.req	x5
 srcin	.req	x15
-SYM_FUNC_START(__arch_copy_from_user)
+SYM_FUNC_START_TRACE(__arch_copy_from_user)
 	add	end, x0, x2
 	mov	srcin, x1
 #include "copy_template.S"
diff --git a/arch/arm64/lib/copy_to_user.S b/arch/arm64/lib/copy_to_user.S
index 819f2e3fc7a9..e7e663aea8ca 100644
--- a/arch/arm64/lib/copy_to_user.S
+++ b/arch/arm64/lib/copy_to_user.S
@@ -60,7 +60,7 @@
 
 end	.req	x5
 srcin	.req	x15
-SYM_FUNC_START(__arch_copy_to_user)
+SYM_FUNC_START_TRACE(__arch_copy_to_user)
 	add	end, x0, x2
 	mov	srcin, x1
 #include "copy_template.S"
-- 
2.47.3


Powered by blists - more mailing lists

Powered by Openwall GNU/*/Linux Powered by OpenVZ