Fix backtrace support in THUMB2 mode

From: Nikolay Borisov <[email protected]>

Currently there are numerous places where "struct pt_regs" are used to populate "struct stackframe", however all of those location do not consider the situation where the kernel might be compiled in THUMB2 mode, in which case the framepointer member of pt_regs become ARM_r7 instead of ARM_fp (r11). Document this idiosyncracy in the definition of "struct stackframe"

The easiest solution is to introduce a new function (in the spirit of
https://groups.google.com/forum/#!topic/linux.kernel/dA2YuUcSpZ4)
which would hide the complexity of initializing the stackframe struct from pt_regs.

Also implement a macro frame_pointer(regs) that would return the correct register so that we can use it in cases where we just require the frame pointer and not a whole struct stackframe

Signed-off-by Nikolay Borisov <[email protected]>

---
 arch/arm/include/asm/ptrace.h     |  6 ++++++
 arch/arm/include/asm/stacktrace.h | 12 ++++++++++++
 2 files changed, 18 insertions(+)

arch/arm/include/asm/ptrace.h +6 -0

index c877654..601264d 100644
@@ -84,6 +84,12 @@ static inline long regs_return_value(struct pt_regs *regs)
 
 #define instruction_pointer(regs)	(regs)->ARM_pc
 
+#ifdef CONFIG_THUMB2_KERNEL
+#define frame_pointer(regs) (regs)->ARM_r7
+#else
+#define frame_pointer(regs) (regs)->ARM_fp
+#endif
+
 static inline void instruction_pointer_set(struct pt_regs *regs,
 					   unsigned long val)
 {

arch/arm/include/asm/stacktrace.h +0 -0

index 4d0a164..6a5b13e 100644

arch/arm/include/asm/ptrace.h +6 -0

index c877654..601264d 100644
@@ -84,6 +84,12 @@ static inline long regs_return_value(struct pt_regs *regs)
 
 #define instruction_pointer(regs)	(regs)->ARM_pc
 
+#ifdef CONFIG_THUMB2_KERNEL
+#define frame_pointer(regs) (regs)->ARM_r7
+#else
+#define frame_pointer(regs) (regs)->ARM_fp
+#endif
+
 static inline void instruction_pointer_set(struct pt_regs *regs,
 					   unsigned long val)
 {

arch/arm/include/asm/stacktrace.h +3 -0

index 4d0a164..6a5b13e 100644
@@ -2,12 +2,24 @@
 #define __ASM_STACKTRACE_H
 
 struct stackframe {
+	/* FP member should hold R7 when CONFIG_THUMB2_KERNEL is enabled.
+	 * and R11 otherwise
+	 */

arch/arm/include/asm/stacktrace.h +4 -0

+		frame->sp = regs->ARM_sp;
+		frame->lr = regs->ARM_lr;
+		frame->pc = regs->ARM_pc;
+}

From: Nikolay Borisov <[email protected]>

Currently all the code which deals with backtrace support assumes that R11 is the frame-pointer. While this is the case for ARM mode and is explicitly documented in the AAPCS, this is not the case for THUMB2 mode.

There is no official document requiring that R11 has to be the frame pointer and GCC uses R7 as FP and given that R7's usage is so intertwined within GCC's mechanics it is unlikely to change, so fixing backtrace in THUMB2 mode seems in order.

This patch series rectifies the problem by first fixing the thread_save_fp macro to reference the correct register. Furthermore, there a lot of repetetive sequences of code such as :

stackframe.fp = pt_regs->ARM_fp
stackframe.lr = pt_regs->ARM_lr

so introducing a function arm_get_current_stack_frame which both hides this repetition and also utilizes teh frame_pointer(regs) macro to reference the correct register depending on the mode.

Finally, change all the call sites so that they utilize the new routine.

Changes since v1:
* Nothing major, just made arg_get_current_stack frame an inline routine.

From: Nikolay Borisov <[email protected]>

Currently all the code which deals with backtrace support assumes that R11 is the frame-pointer. While this is the case for ARM mode and is explicitly documented in the AAPCS, this is not the case for THUMB2 mode.

There is no official document requiring that R11 has to be the frame pointer and GCC uses R7 as FP and given that R7's usage is so intertwined within GCC's mechanics it is unlikely to change, so fixing backtrace in THUMB2 mode seems in order.

This patch series rectifies the problem by first fixing the thread_save_fp macro to reference the correct register. Furthermore, there a lot of repetetive sequences of code such as :

stackframe.fp = pt_regs->ARM_fp
stackframe.lr = pt_regs->ARM_lr

so introducing a function arm_get_current_stack_frame which both hides this repetition and also utilizes teh frame_pointer(regs) macro to reference the correct register depending on the mode.

Finally, change all the call sites so that they utilize the new routine.

-----Original Message-----
From: Arnd Bergmann [mailto:arnd at arndb.de]
Sent: 31 May 2014 19:32
To: linux-arm-kernel at lists.infradead.org
Cc: Nikolay Borisov; rric at kernel.org; Will Deacon; u.kleine- koenig at pengutronix.de
Subject: Re: [PATCH v2 0/7] Fix backtrace support in THUMB2 mode

Quoted 3 lines · On Friday 30 May 2014 17:59:44 Nikolay Borisov wrote:

From: Nikolay Borisov <[email protected]>

Currently all the code which deals with backtrace support assumes

that R11

Quoted 1 line

is the frame-pointer. While this is the case for ARM mode and is

explicitly

Quoted 3 lines

documented in the AAPCS, this is not the case for THUMB2 mode.

There is no official document requiring that R11 has to be the frame

pointer

Quoted 1 line

and GCC uses R7 as FP and given that R7's usage is so intertwined

within GCC's

Quoted 1 line

mechanics it is unlikely to change, so fixing backtrace in THUMB2

mode seems

Quoted 4 lines

in order.

This patch series rectifies the problem by first fixing the thread_save_fp macro to reference the correct register. Furthermore,

there

Quoted 9 lines

a lot of repetetive sequences of code such as :

stackframe.fp = pt_regs->ARM_fp
stackframe.lr = pt_regs->ARM_lr

so introducing a function arm_get_current_stack_frame which both hides this repetition and also utilizes teh frame_pointer(regs) macro to reference the correct register depending on the mode.

Quoted 1 line

Finally, change all the call sites so that they utilize the new

routine.

Sorry for the stupid question, but does your series relate to the #warning "TODO: return_address should use unwind tables" that we see all the time in arch/arm/kernel/return_address.c?

If I remember the story correctly, the problem is that so far we could never use the frame pointer in thumb2 mode, and nobody has implemented a version of that function using the arm unwinder.

Are we now able to use the frame pointer after all?

Arnd