[RFC PATCH 00/19] Remove STACK_FRAME_OVERHEAD

[RFC PATCH 00/19] Remove STACK_FRAME_OVERHEAD

[Nicholas Piggin]

This is some quick hacking, hardly tested but might have potential.

I think we're not validating the perf kernel stack walker bounds
quite correctly, and not setting up decent stack frames for the child
in copy_thread. So at least those two things we could do. Maybe
patch 1 should go upstream as a fix.

Thanks,
Nick

Nicholas Piggin (19):
  powerpc/perf: callchain validate kernel stack pointer bounds
  powerpc: Rearrange copy_thread child stack creation
  powerpc/64: Remove asm interrupt tracing call helpers
  powerpc/pseries: hvcall stack frame overhead
  powerpc/32: Use load and store multiple in GPR save/restore macros
  powerpc: simplify ppc_save_regs
  powerpc: add definition for pt_regs offset within an interrupt frame
  powerpc: add a definition for the marker offset within the interrupt
    frame
  powerpc: Rename STACK_FRAME_MARKER and derive it from frame offset
  powerpc: add a define for the user interrupt frame size
  powerpc: add a define for the switch frame size and regs offset
  powerpc: copy_thread fill in interrupt frame marker and back chain
  powerpc: copy_thread add a back chain to the switch stack frame
  powerpc: split validate_sp into two functions
  powerpc: allow minimum sized kernel stack frames
  powerpc/64: ELFv2 use minimal stack frames in int and switch frame
    sizes
  powerpc: remove STACK_FRAME_OVERHEAD
  powerpc: change stack marker memory operations to 32-bit
  powerpc/64: ELFv2 use reserved word in the stack frame for the regs
    marker

 arch/powerpc/include/asm/irqflags.h           | 29 -------
 arch/powerpc/include/asm/ppc_asm.h            | 18 +++-
 arch/powerpc/include/asm/processor.h          | 15 +++-
 arch/powerpc/include/asm/ptrace.h             | 41 +++++++---
 arch/powerpc/kernel/asm-offsets.c             |  9 +-
 arch/powerpc/kernel/entry_32.S                | 14 ++--
 arch/powerpc/kernel/exceptions-64e.S          | 44 +++++-----
 arch/powerpc/kernel/exceptions-64s.S          | 82 +++++++++----------
 arch/powerpc/kernel/head_32.h                 |  4 +-
 arch/powerpc/kernel/head_40x.S                |  2 +-
 arch/powerpc/kernel/head_44x.S                |  6 +-
 arch/powerpc/kernel/head_64.S                 |  6 +-
 arch/powerpc/kernel/head_85xx.S               |  8 +-
 arch/powerpc/kernel/head_8xx.S                |  2 +-
 arch/powerpc/kernel/head_book3s_32.S          |  4 +-
 arch/powerpc/kernel/head_booke.h              |  4 +-
 arch/powerpc/kernel/interrupt_64.S            | 32 ++++----
 arch/powerpc/kernel/irq.c                     |  4 +-
 arch/powerpc/kernel/kgdb.c                    |  2 +-
 arch/powerpc/kernel/misc_32.S                 |  2 +-
 arch/powerpc/kernel/misc_64.S                 |  4 +-
 arch/powerpc/kernel/optprobes_head.S          |  4 +-
 arch/powerpc/kernel/ppc_save_regs.S           | 58 ++++---------
 arch/powerpc/kernel/process.c                 | 54 +++++++-----
 arch/powerpc/kernel/smp.c                     |  2 +-
 arch/powerpc/kernel/stacktrace.c              | 10 +--
 arch/powerpc/kernel/tm.S                      |  8 +-
 arch/powerpc/kernel/trace/ftrace_mprofile.S   |  2 +-
 arch/powerpc/kvm/book3s_hv_rmhandlers.S       |  2 +-
 .../lib/test_emulate_step_exec_instr.S        |  2 +-
 arch/powerpc/perf/callchain.c                 |  9 +-
 arch/powerpc/platforms/pseries/hvCall.S       | 38 +++++----
 arch/powerpc/xmon/xmon.c                      | 10 +--
 33 files changed, 263 insertions(+), 268 deletions(-)

-- 
2.37.2

[RFC PATCH 01/19] powerpc/perf: callchain validate kernel stack pointer bounds

[Nicholas Piggin]

The interrupt frame detection and loads from the hypothetical pt_regs
are not bounds-checked. The next-frame validation only bounds-checks
STACK_FRAME_OVERHEAD, which does not include the pt_regs. Add another
test for this.

Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
---

Could the user set r1 to be equal to the address matching the first
interrupt frame - STACK_INT_FRAME_SIZE, which is in the previous page
due to the kernel redzone, and induce the kernel to load the marker from
there? Possibly it could cause a crash at least.

It also seems a bit rude to put a fancy next-frame-validation out in
perf/ rather than with the rest of the frame validation code.

Thanks,
Nick

 arch/powerpc/perf/callchain.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/arch/powerpc/perf/callchain.c b/arch/powerpc/perf/callchain.c
index 082f6d0308a4..8718289c051d 100644
--- a/arch/powerpc/perf/callchain.c
+++ b/arch/powerpc/perf/callchain.c
@@ -61,6 +61,7 @@ perf_callchain_kernel(struct perf_callchain_entry_ctx *entry, struct pt_regs *re
 		next_sp = fp[0];
 
 		if (next_sp == sp + STACK_INT_FRAME_SIZE &&
+		    validate_sp(sp, current, STACK_INT_FRAME_SIZE) &&
 		    fp[STACK_FRAME_MARKER] == STACK_FRAME_REGS_MARKER) {
 			/*
 			 * This looks like an interrupt frame for an
-- 
2.37.2

Re: [RFC PATCH 01/19] powerpc/perf: callchain validate kernel stack pointer bounds

[Segher Boessenkool]

On Mon, Oct 31, 2022 at 03:54:22PM +1000, Nicholas Piggin wrote:
> Could the user set r1 to be equal to the address matching the first
> interrupt frame - STACK_INT_FRAME_SIZE, which is in the previous page
> due to the kernel redzone, and induce the kernel to load the marker from
> there? Possibly it could cause a crash at least.

Yes, the user can set r1 to anything, it is just a general purpose
register.  This isn't a valid thing to do of course, the ABI requires
r1 to point at a valid stack at all times, but it is an obvious attack
point if we do not harden against this :-)


Segher

[RFC PATCH 02/19] powerpc: Rearrange copy_thread child stack creation

[Nicholas Piggin]

This makes it a bit clearer where the stack frame is created, and will
allow easier use of some of the stack offset constants in a later
change.

Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
---
 arch/powerpc/kernel/process.c | 11 ++++++-----
 1 file changed, 6 insertions(+), 5 deletions(-)

diff --git a/arch/powerpc/kernel/process.c b/arch/powerpc/kernel/process.c
index 67da147fe34d..acfa197fb2df 100644
--- a/arch/powerpc/kernel/process.c
+++ b/arch/powerpc/kernel/process.c
@@ -1726,13 +1726,16 @@ int copy_thread(struct task_struct *p, const struct kernel_clone_args *args)
 
 	klp_init_thread_info(p);
 
+	/* Create initial stack frame. */
+	sp -= (sizeof(struct pt_regs) + STACK_FRAME_OVERHEAD);
+	((unsigned long *)sp)[0] = 0;
+
 	/* Copy registers */
-	sp -= sizeof(struct pt_regs);
-	childregs = (struct pt_regs *) sp;
+	childregs = (struct pt_regs *)(sp + STACK_FRAME_OVERHEAD);
 	if (unlikely(args->fn)) {
 		/* kernel thread */
 		memset(childregs, 0, sizeof(struct pt_regs));
-		childregs->gpr[1] = sp + sizeof(struct pt_regs);
+		childregs->gpr[1] = sp + (sizeof(struct pt_regs) + STACK_FRAME_OVERHEAD);
 		/* function */
 		if (args->fn)
 			childregs->gpr[14] = ppc_function_entry((void *)args->fn);
@@ -1767,7 +1770,6 @@ int copy_thread(struct task_struct *p, const struct kernel_clone_args *args)
 			f = ret_from_fork;
 	}
 	childregs->msr &= ~(MSR_FP|MSR_VEC|MSR_VSX);
-	sp -= STACK_FRAME_OVERHEAD;
 
 	/*
 	 * The way this works is that at some point in the future
@@ -1777,7 +1779,6 @@ int copy_thread(struct task_struct *p, const struct kernel_clone_args *args)
 	 * do some house keeping and then return from the fork or clone
 	 * system call, using the stack frame created above.
 	 */
-	((unsigned long *)sp)[0] = 0;
 	sp -= sizeof(struct pt_regs);
 	kregs = (struct pt_regs *) sp;
 	sp -= STACK_FRAME_OVERHEAD;
-- 
2.37.2

[RFC PATCH 03/19] powerpc/64: Remove asm interrupt tracing call helpers

[Nicholas Piggin]

These are unused. Remove.

Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
---
 arch/powerpc/include/asm/irqflags.h | 29 -----------------------------
 1 file changed, 29 deletions(-)

diff --git a/arch/powerpc/include/asm/irqflags.h b/arch/powerpc/include/asm/irqflags.h
index 1a6c1ce17735..81e0a5025be8 100644
--- a/arch/powerpc/include/asm/irqflags.h
+++ b/arch/powerpc/include/asm/irqflags.h
@@ -13,32 +13,6 @@
 
 #else
 #ifdef CONFIG_TRACE_IRQFLAGS
-#ifdef CONFIG_IRQSOFF_TRACER
-/*
- * Since the ftrace irqsoff latency trace checks CALLER_ADDR1,
- * which is the stack frame here, we need to force a stack frame
- * in case we came from user space.
- */
-#define TRACE_WITH_FRAME_BUFFER(func)		\
-	mflr	r0;				\
-	stdu	r1, -STACK_FRAME_OVERHEAD(r1);	\
-	std	r0, 16(r1);			\
-	stdu	r1, -STACK_FRAME_OVERHEAD(r1);	\
-	bl func;				\
-	ld	r1, 0(r1);			\
-	ld	r1, 0(r1);
-#else
-#define TRACE_WITH_FRAME_BUFFER(func)		\
-	bl func;
-#endif
-
-/*
- * These are calls to C code, so the caller must be prepared for volatiles to
- * be clobbered.
- */
-#define TRACE_ENABLE_INTS	TRACE_WITH_FRAME_BUFFER(trace_hardirqs_on)
-#define TRACE_DISABLE_INTS	TRACE_WITH_FRAME_BUFFER(trace_hardirqs_off)
-
 /*
  * This is used by assembly code to soft-disable interrupts first and
  * reconcile irq state.
@@ -59,9 +33,6 @@
 44:
 
 #else
-#define TRACE_ENABLE_INTS
-#define TRACE_DISABLE_INTS
-
 #define RECONCILE_IRQ_STATE(__rA, __rB)		\
 	lbz	__rA,PACAIRQHAPPENED(r13);	\
 	li	__rB,IRQS_DISABLED;		\
-- 
2.37.2

[RFC PATCH 04/19] powerpc/pseries: hvcall stack frame overhead

[Nicholas Piggin]

This call may use the min size stack frame. The scratch space used is
in the caller's parameter area frame, not this function's frame.

Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
---
 arch/powerpc/platforms/pseries/hvCall.S | 38 +++++++++++++------------
 1 file changed, 20 insertions(+), 18 deletions(-)

diff --git a/arch/powerpc/platforms/pseries/hvCall.S b/arch/powerpc/platforms/pseries/hvCall.S
index 762eb15d3bd4..783c16ad648b 100644
--- a/arch/powerpc/platforms/pseries/hvCall.S
+++ b/arch/powerpc/platforms/pseries/hvCall.S
@@ -27,7 +27,9 @@ hcall_tracepoint_refcount:
 
 /*
  * precall must preserve all registers.  use unused STK_PARAM()
- * areas to save snapshots and opcode.
+ * areas to save snapshots and opcode. STK_PARAM() in the caller's
+ * frame will be available even on ELFv2 because these are all
+ * variadic functions.
  */
 #define HCALL_INST_PRECALL(FIRST_REG)				\
 	mflr	r0;						\
@@ -41,29 +43,29 @@ hcall_tracepoint_refcount:
 	std	r10,STK_PARAM(R10)(r1);				\
 	std	r0,16(r1);					\
 	addi	r4,r1,STK_PARAM(FIRST_REG);			\
-	stdu	r1,-STACK_FRAME_OVERHEAD(r1);			\
+	stdu	r1,-STACK_FRAME_MIN_SIZE(r1);			\
 	bl	__trace_hcall_entry;				\
-	ld	r3,STACK_FRAME_OVERHEAD+STK_PARAM(R3)(r1);	\
-	ld	r4,STACK_FRAME_OVERHEAD+STK_PARAM(R4)(r1);	\
-	ld	r5,STACK_FRAME_OVERHEAD+STK_PARAM(R5)(r1);	\
-	ld	r6,STACK_FRAME_OVERHEAD+STK_PARAM(R6)(r1);	\
-	ld	r7,STACK_FRAME_OVERHEAD+STK_PARAM(R7)(r1);	\
-	ld	r8,STACK_FRAME_OVERHEAD+STK_PARAM(R8)(r1);	\
-	ld	r9,STACK_FRAME_OVERHEAD+STK_PARAM(R9)(r1);	\
-	ld	r10,STACK_FRAME_OVERHEAD+STK_PARAM(R10)(r1)
+	ld	r3,STACK_FRAME_MIN_SIZE+STK_PARAM(R3)(r1);	\
+	ld	r4,STACK_FRAME_MIN_SIZE+STK_PARAM(R4)(r1);	\
+	ld	r5,STACK_FRAME_MIN_SIZE+STK_PARAM(R5)(r1);	\
+	ld	r6,STACK_FRAME_MIN_SIZE+STK_PARAM(R6)(r1);	\
+	ld	r7,STACK_FRAME_MIN_SIZE+STK_PARAM(R7)(r1);	\
+	ld	r8,STACK_FRAME_MIN_SIZE+STK_PARAM(R8)(r1);	\
+	ld	r9,STACK_FRAME_MIN_SIZE+STK_PARAM(R9)(r1);	\
+	ld	r10,STACK_FRAME_MIN_SIZE+STK_PARAM(R10)(r1)
 
 /*
  * postcall is performed immediately before function return which
  * allows liberal use of volatile registers.
  */
 #define __HCALL_INST_POSTCALL					\
-	ld	r0,STACK_FRAME_OVERHEAD+STK_PARAM(R3)(r1);	\
-	std	r3,STACK_FRAME_OVERHEAD+STK_PARAM(R3)(r1);	\
+	ld	r0,STACK_FRAME_MIN_SIZE+STK_PARAM(R3)(r1);	\
+	std	r3,STACK_FRAME_MIN_SIZE+STK_PARAM(R3)(r1);	\
 	mr	r4,r3;						\
 	mr	r3,r0;						\
 	bl	__trace_hcall_exit;				\
-	ld	r0,STACK_FRAME_OVERHEAD+16(r1);			\
-	addi	r1,r1,STACK_FRAME_OVERHEAD;			\
+	ld	r0,STACK_FRAME_MIN_SIZE+16(r1);			\
+	addi	r1,r1,STACK_FRAME_MIN_SIZE;			\
 	ld	r3,STK_PARAM(R3)(r1);				\
 	mtlr	r0
 
@@ -303,14 +305,14 @@ plpar_hcall9_trace:
 	mr	r7,r8
 	mr	r8,r9
 	mr	r9,r10
-	ld	r10,STACK_FRAME_OVERHEAD+STK_PARAM(R11)(r1)
-	ld	r11,STACK_FRAME_OVERHEAD+STK_PARAM(R12)(r1)
-	ld	r12,STACK_FRAME_OVERHEAD+STK_PARAM(R13)(r1)
+	ld	r10,STACK_FRAME_MIN_SIZE+STK_PARAM(R11)(r1)
+	ld	r11,STACK_FRAME_MIN_SIZE+STK_PARAM(R12)(r1)
+	ld	r12,STACK_FRAME_MIN_SIZE+STK_PARAM(R13)(r1)
 
 	HVSC
 
 	mr	r0,r12
-	ld	r12,STACK_FRAME_OVERHEAD+STK_PARAM(R4)(r1)
+	ld	r12,STACK_FRAME_MIN_SIZE+STK_PARAM(R4)(r1)
 	std	r4,0(r12)
 	std	r5,8(r12)
 	std	r6,16(r12)
-- 
2.37.2

[RFC PATCH 05/19] powerpc/32: Use load and store multiple in GPR save/restore macros

[Nicholas Piggin]

---
 arch/powerpc/include/asm/ppc_asm.h | 18 ++++++++++++++++--
 1 file changed, 16 insertions(+), 2 deletions(-)

diff --git a/arch/powerpc/include/asm/ppc_asm.h b/arch/powerpc/include/asm/ppc_asm.h
index 753a2757bcd4..ac44383d350a 100644
--- a/arch/powerpc/include/asm/ppc_asm.h
+++ b/arch/powerpc/include/asm/ppc_asm.h
@@ -57,8 +57,22 @@
 #define SAVE_NVGPRS(base)		SAVE_GPRS(14, 31, base)
 #define REST_NVGPRS(base)		REST_GPRS(14, 31, base)
 #else
-#define SAVE_GPRS(start, end, base)	OP_REGS stw, 4, start, end, base, GPR0
-#define REST_GPRS(start, end, base)	OP_REGS lwz, 4, start, end, base, GPR0
+.macro __SAVE_GPRS start, end, base, offset
+	.if \end == 31
+		stmw	\start,\offset(\base)
+	.else
+		OP_REGS stw, 4, \start, \end, \base, \offset
+	.endif
+.endm
+.macro __REST_GPRS start, end, base, offset
+	.if \end == 31
+		lmw	\start,\offset(\base)
+	.else
+		OP_REGS lwz, 4, \start, \end, \base, \offset
+	.endif
+.endm
+#define SAVE_GPRS(start, end, base)	__SAVE_GPRS start, end, base, GPR0
+#define REST_GPRS(start, end, base)	__REST_GPRS start, end, base, GPR0
 #define SAVE_NVGPRS(base)		SAVE_GPRS(13, 31, base)
 #define REST_NVGPRS(base)		REST_GPRS(13, 31, base)
 #endif
-- 
2.37.2

Re: [RFC PATCH 05/19] powerpc/32: Use load and store multiple in GPR save/restore macros

[Christophe Leroy]

It contradicts commit a85c728cb5e1 ("powerpc/32: Don't use lmw/stmw for 
saving/restoring non volatile regs")

Le 31/10/2022 à 06:54, Nicholas Piggin a écrit :
> ---
>   arch/powerpc/include/asm/ppc_asm.h | 18 ++++++++++++++++--
>   1 file changed, 16 insertions(+), 2 deletions(-)
> 
> diff --git a/arch/powerpc/include/asm/ppc_asm.h b/arch/powerpc/include/asm/ppc_asm.h
> index 753a2757bcd4..ac44383d350a 100644
> --- a/arch/powerpc/include/asm/ppc_asm.h
> +++ b/arch/powerpc/include/asm/ppc_asm.h
> @@ -57,8 +57,22 @@
>   #define SAVE_NVGPRS(base)		SAVE_GPRS(14, 31, base)
>   #define REST_NVGPRS(base)		REST_GPRS(14, 31, base)
>   #else
> -#define SAVE_GPRS(start, end, base)	OP_REGS stw, 4, start, end, base, GPR0
> -#define REST_GPRS(start, end, base)	OP_REGS lwz, 4, start, end, base, GPR0
> +.macro __SAVE_GPRS start, end, base, offset
> +	.if \end == 31
> +		stmw	\start,\offset(\base)
> +	.else
> +		OP_REGS stw, 4, \start, \end, \base, \offset
> +	.endif
> +.endm
> +.macro __REST_GPRS start, end, base, offset
> +	.if \end == 31
> +		lmw	\start,\offset(\base)
> +	.else
> +		OP_REGS lwz, 4, \start, \end, \base, \offset
> +	.endif
> +.endm
> +#define SAVE_GPRS(start, end, base)	__SAVE_GPRS start, end, base, GPR0
> +#define REST_GPRS(start, end, base)	__REST_GPRS start, end, base, GPR0
>   #define SAVE_NVGPRS(base)		SAVE_GPRS(13, 31, base)
>   #define REST_NVGPRS(base)		REST_GPRS(13, 31, base)
>   #endif

Re: [RFC PATCH 05/19] powerpc/32: Use load and store multiple in GPR save/restore macros

[Nicholas Piggin]

On Thu Nov 3, 2022 at 6:26 PM AEST, Christophe Leroy wrote:
> It contradicts commit a85c728cb5e1 ("powerpc/32: Don't use lmw/stmw for 
> saving/restoring non volatile regs")

Hmm okay. Do you want to keep the #ifdef case in ppc_save_regs for ppc32
to save ~124 bytes of text? I don't mind either way.

Thanks,
Nick

>
> Le 31/10/2022 à 06:54, Nicholas Piggin a écrit :
> > ---
> >   arch/powerpc/include/asm/ppc_asm.h | 18 ++++++++++++++++--
> >   1 file changed, 16 insertions(+), 2 deletions(-)
> > 
> > diff --git a/arch/powerpc/include/asm/ppc_asm.h b/arch/powerpc/include/asm/ppc_asm.h
> > index 753a2757bcd4..ac44383d350a 100644
> > --- a/arch/powerpc/include/asm/ppc_asm.h
> > +++ b/arch/powerpc/include/asm/ppc_asm.h
> > @@ -57,8 +57,22 @@
> >   #define SAVE_NVGPRS(base)		SAVE_GPRS(14, 31, base)
> >   #define REST_NVGPRS(base)		REST_GPRS(14, 31, base)
> >   #else
> > -#define SAVE_GPRS(start, end, base)	OP_REGS stw, 4, start, end, base, GPR0
> > -#define REST_GPRS(start, end, base)	OP_REGS lwz, 4, start, end, base, GPR0
> > +.macro __SAVE_GPRS start, end, base, offset
> > +	.if \end == 31
> > +		stmw	\start,\offset(\base)
> > +	.else
> > +		OP_REGS stw, 4, \start, \end, \base, \offset
> > +	.endif
> > +.endm
> > +.macro __REST_GPRS start, end, base, offset
> > +	.if \end == 31
> > +		lmw	\start,\offset(\base)
> > +	.else
> > +		OP_REGS lwz, 4, \start, \end, \base, \offset
> > +	.endif
> > +.endm
> > +#define SAVE_GPRS(start, end, base)	__SAVE_GPRS start, end, base, GPR0
> > +#define REST_GPRS(start, end, base)	__REST_GPRS start, end, base, GPR0
> >   #define SAVE_NVGPRS(base)		SAVE_GPRS(13, 31, base)
> >   #define REST_NVGPRS(base)		REST_GPRS(13, 31, base)
> >   #endif

Re: [RFC PATCH 05/19] powerpc/32: Use load and store multiple in GPR save/restore macros

[Christophe Leroy]

Le 07/11/2022 à 13:45, Nicholas Piggin a écrit :
> On Thu Nov 3, 2022 at 6:26 PM AEST, Christophe Leroy wrote:
>> It contradicts commit a85c728cb5e1 ("powerpc/32: Don't use lmw/stmw for
>> saving/restoring non volatile regs")
> 
> Hmm okay. Do you want to keep the #ifdef case in ppc_save_regs for ppc32
> to save ~124 bytes of text? I don't mind either way.

I think you can make ppc_save_regs() generic.

I don't think the lmw/stmw is worth it in ppc_save_regs(), just drop it.

Christophe

> 
> Thanks,
> Nick
> 
>>
>> Le 31/10/2022 à 06:54, Nicholas Piggin a écrit :
>>> ---
>>>    arch/powerpc/include/asm/ppc_asm.h | 18 ++++++++++++++++--
>>>    1 file changed, 16 insertions(+), 2 deletions(-)
>>>
>>> diff --git a/arch/powerpc/include/asm/ppc_asm.h b/arch/powerpc/include/asm/ppc_asm.h
>>> index 753a2757bcd4..ac44383d350a 100644
>>> --- a/arch/powerpc/include/asm/ppc_asm.h
>>> +++ b/arch/powerpc/include/asm/ppc_asm.h
>>> @@ -57,8 +57,22 @@
>>>    #define SAVE_NVGPRS(base)		SAVE_GPRS(14, 31, base)
>>>    #define REST_NVGPRS(base)		REST_GPRS(14, 31, base)
>>>    #else
>>> -#define SAVE_GPRS(start, end, base)	OP_REGS stw, 4, start, end, base, GPR0
>>> -#define REST_GPRS(start, end, base)	OP_REGS lwz, 4, start, end, base, GPR0
>>> +.macro __SAVE_GPRS start, end, base, offset
>>> +	.if \end == 31
>>> +		stmw	\start,\offset(\base)
>>> +	.else
>>> +		OP_REGS stw, 4, \start, \end, \base, \offset
>>> +	.endif
>>> +.endm
>>> +.macro __REST_GPRS start, end, base, offset
>>> +	.if \end == 31
>>> +		lmw	\start,\offset(\base)
>>> +	.else
>>> +		OP_REGS lwz, 4, \start, \end, \base, \offset
>>> +	.endif
>>> +.endm
>>> +#define SAVE_GPRS(start, end, base)	__SAVE_GPRS start, end, base, GPR0
>>> +#define REST_GPRS(start, end, base)	__REST_GPRS start, end, base, GPR0
>>>    #define SAVE_NVGPRS(base)		SAVE_GPRS(13, 31, base)
>>>    #define REST_NVGPRS(base)		REST_GPRS(13, 31, base)
>>>    #endif
> 

[RFC PATCH 06/19] powerpc: simplify ppc_save_regs

[Nicholas Piggin]

Adjust the pt_regs pointer so the interrupt frame offsets can be used
directly to save registers.

Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
---
 arch/powerpc/kernel/ppc_save_regs.S | 58 +++++++----------------------
 1 file changed, 14 insertions(+), 44 deletions(-)

diff --git a/arch/powerpc/kernel/ppc_save_regs.S b/arch/powerpc/kernel/ppc_save_regs.S
index 2d4d21bb46a9..2908a5c2cd2f 100644
--- a/arch/powerpc/kernel/ppc_save_regs.S
+++ b/arch/powerpc/kernel/ppc_save_regs.S
@@ -21,60 +21,30 @@
  * different ABIs, though).
  */
 _GLOBAL(ppc_save_regs)
-	PPC_STL	r0,0*SZL(r3)
-#ifdef CONFIG_PPC32
-	stmw	r2, 2*SZL(r3)
-#else
-	PPC_STL	r2,2*SZL(r3)
-	PPC_STL	r3,3*SZL(r3)
-	PPC_STL	r4,4*SZL(r3)
-	PPC_STL	r5,5*SZL(r3)
-	PPC_STL	r6,6*SZL(r3)
-	PPC_STL	r7,7*SZL(r3)
-	PPC_STL	r8,8*SZL(r3)
-	PPC_STL	r9,9*SZL(r3)
-	PPC_STL	r10,10*SZL(r3)
-	PPC_STL	r11,11*SZL(r3)
-	PPC_STL	r12,12*SZL(r3)
-	PPC_STL	r13,13*SZL(r3)
-	PPC_STL	r14,14*SZL(r3)
-	PPC_STL	r15,15*SZL(r3)
-	PPC_STL	r16,16*SZL(r3)
-	PPC_STL	r17,17*SZL(r3)
-	PPC_STL	r18,18*SZL(r3)
-	PPC_STL	r19,19*SZL(r3)
-	PPC_STL	r20,20*SZL(r3)
-	PPC_STL	r21,21*SZL(r3)
-	PPC_STL	r22,22*SZL(r3)
-	PPC_STL	r23,23*SZL(r3)
-	PPC_STL	r24,24*SZL(r3)
-	PPC_STL	r25,25*SZL(r3)
-	PPC_STL	r26,26*SZL(r3)
-	PPC_STL	r27,27*SZL(r3)
-	PPC_STL	r28,28*SZL(r3)
-	PPC_STL	r29,29*SZL(r3)
-	PPC_STL	r30,30*SZL(r3)
-	PPC_STL	r31,31*SZL(r3)
+	/* This allows stack frame accessor macros and offsets to be used */
+	subi	r3,r3,STACK_FRAME_OVERHEAD
+	SAVE_GPRS(0, 31, r3)
+#ifdef CONFIG_PPC64
 	lbz	r0,PACAIRQSOFTMASK(r13)
-	PPC_STL	r0,SOFTE-STACK_FRAME_OVERHEAD(r3)
+	PPC_STL	r0,SOFTE(r3)
 #endif
 	/* go up one stack frame for SP */
 	PPC_LL	r4,0(r1)
-	PPC_STL	r4,1*SZL(r3)
+	PPC_STL	r4,GPR1(r3)
 	/* get caller's LR */
 	PPC_LL	r0,LRSAVE(r4)
-	PPC_STL	r0,_LINK-STACK_FRAME_OVERHEAD(r3)
+	PPC_STL	r0,_LINK(r3)
 	mflr	r0
-	PPC_STL	r0,_NIP-STACK_FRAME_OVERHEAD(r3)
+	PPC_STL	r0,_NIP(r3)
 	mfmsr	r0
-	PPC_STL	r0,_MSR-STACK_FRAME_OVERHEAD(r3)
+	PPC_STL	r0,_MSR(r3)
 	mfctr	r0
-	PPC_STL	r0,_CTR-STACK_FRAME_OVERHEAD(r3)
+	PPC_STL	r0,_CTR(r3)
 	mfxer	r0
-	PPC_STL	r0,_XER-STACK_FRAME_OVERHEAD(r3)
+	PPC_STL	r0,_XER(r3)
 	mfcr	r0
-	PPC_STL	r0,_CCR-STACK_FRAME_OVERHEAD(r3)
+	PPC_STL	r0,_CCR(r3)
 	li	r0,0
-	PPC_STL	r0,_TRAP-STACK_FRAME_OVERHEAD(r3)
-	PPC_STL	r0,ORIG_GPR3-STACK_FRAME_OVERHEAD(r3)
+	PPC_STL	r0,_TRAP(r3)
+	PPC_STL	r0,ORIG_GPR3(r3)
 	blr
-- 
2.37.2

[RFC PATCH 07/19] powerpc: add definition for pt_regs offset within an interrupt frame

[Nicholas Piggin]

This is a common offset that currently uses the overloaded
STACK_FRAME_OVERHEAD constant. It's easier to read and more
flexible to use a specific regs offset for this.

Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
---
 arch/powerpc/include/asm/ptrace.h             |  2 +
 arch/powerpc/kernel/asm-offsets.c             |  7 +-
 arch/powerpc/kernel/entry_32.S                |  6 +-
 arch/powerpc/kernel/exceptions-64e.S          | 42 +++++-----
 arch/powerpc/kernel/exceptions-64s.S          | 80 +++++++++----------
 arch/powerpc/kernel/head_32.h                 |  2 +-
 arch/powerpc/kernel/head_85xx.S               |  4 +-
 arch/powerpc/kernel/head_booke.h              |  2 +-
 arch/powerpc/kernel/interrupt_64.S            | 22 ++---
 arch/powerpc/kernel/kgdb.c                    |  2 +-
 arch/powerpc/kernel/optprobes_head.S          |  4 +-
 arch/powerpc/kernel/ppc_save_regs.S           |  2 +-
 arch/powerpc/kernel/process.c                 |  4 +-
 arch/powerpc/kernel/tm.S                      |  8 +-
 arch/powerpc/kernel/trace/ftrace_mprofile.S   |  2 +-
 .../lib/test_emulate_step_exec_instr.S        |  2 +-
 arch/powerpc/perf/callchain.c                 |  2 +-
 arch/powerpc/xmon/xmon.c                      |  7 +-
 18 files changed, 100 insertions(+), 100 deletions(-)

diff --git a/arch/powerpc/include/asm/ptrace.h b/arch/powerpc/include/asm/ptrace.h
index 2efec6d87049..a4ae67aa9b76 100644
--- a/arch/powerpc/include/asm/ptrace.h
+++ b/arch/powerpc/include/asm/ptrace.h
@@ -124,6 +124,7 @@ struct pt_regs
 #define STACK_FRAME_LR_SAVE	2	/* Location of LR in stack frame */
 #define STACK_INT_FRAME_SIZE	(sizeof(struct pt_regs) + \
 				 STACK_FRAME_OVERHEAD + KERNEL_REDZONE_SIZE)
+#define STACK_INT_FRAME_REGS	STACK_FRAME_OVERHEAD
 #define STACK_FRAME_MARKER	12
 
 #ifdef CONFIG_PPC64_ELF_ABI_V2
@@ -143,6 +144,7 @@ struct pt_regs
 #define STACK_FRAME_OVERHEAD	16	/* size of minimum stack frame */
 #define STACK_FRAME_LR_SAVE	1	/* Location of LR in stack frame */
 #define STACK_INT_FRAME_SIZE	(sizeof(struct pt_regs) + STACK_FRAME_OVERHEAD)
+#define STACK_INT_FRAME_REGS	STACK_FRAME_OVERHEAD
 #define STACK_FRAME_MARKER	2
 #define STACK_FRAME_MIN_SIZE	STACK_FRAME_OVERHEAD
 
diff --git a/arch/powerpc/kernel/asm-offsets.c b/arch/powerpc/kernel/asm-offsets.c
index 4ce2a4aa3985..db5e66c1d031 100644
--- a/arch/powerpc/kernel/asm-offsets.c
+++ b/arch/powerpc/kernel/asm-offsets.c
@@ -72,7 +72,7 @@
 #endif
 
 #define STACK_PT_REGS_OFFSET(sym, val)	\
-	DEFINE(sym, STACK_FRAME_OVERHEAD + offsetof(struct pt_regs, val))
+	DEFINE(sym, STACK_INT_FRAME_REGS + offsetof(struct pt_regs, val))
 
 int main(void)
 {
@@ -167,9 +167,8 @@ int main(void)
 	OFFSET(THREAD_CKVRSTATE, thread_struct, ckvr_state.vr);
 	OFFSET(THREAD_CKVRSAVE, thread_struct, ckvrsave);
 	OFFSET(THREAD_CKFPSTATE, thread_struct, ckfp_state.fpr);
-	/* Local pt_regs on stack for Transactional Memory funcs. */
-	DEFINE(TM_FRAME_SIZE, STACK_FRAME_OVERHEAD +
-	       sizeof(struct pt_regs) + 16);
+	/* Local pt_regs on stack in int frame form, plus 16 bytes for TM */
+	DEFINE(TM_FRAME_SIZE, STACK_INT_FRAME_SIZE + 16);
 #endif /* CONFIG_PPC_TRANSACTIONAL_MEM */
 
 	OFFSET(TI_LOCAL_FLAGS, thread_info, local_flags);
diff --git a/arch/powerpc/kernel/entry_32.S b/arch/powerpc/kernel/entry_32.S
index 3fc7c9886bb7..24c8d84a56c9 100644
--- a/arch/powerpc/kernel/entry_32.S
+++ b/arch/powerpc/kernel/entry_32.S
@@ -123,12 +123,12 @@ transfer_to_syscall:
 	kuep_lock
 
 	/* Calling convention has r3 = regs, r4 = orig r0 */
-	addi	r3,r1,STACK_FRAME_OVERHEAD
+	addi	r3,r1,STACK_INT_FRAME_REGS
 	mr	r4,r0
 	bl	system_call_exception
 
 ret_from_syscall:
-	addi    r4,r1,STACK_FRAME_OVERHEAD
+	addi    r4,r1,STACK_INT_FRAME_REGS
 	li	r5,0
 	bl	syscall_exit_prepare
 #ifdef CONFIG_PPC_47x
@@ -293,7 +293,7 @@ _ASM_NOKPROBE_SYMBOL(fast_exception_return)
 	.globl interrupt_return
 interrupt_return:
 	lwz	r4,_MSR(r1)
-	addi	r3,r1,STACK_FRAME_OVERHEAD
+	addi	r3,r1,STACK_INT_FRAME_REGS
 	andi.	r0,r4,MSR_PR
 	beq	.Lkernel_interrupt_return
 	bl	interrupt_exit_user_prepare
diff --git a/arch/powerpc/kernel/exceptions-64e.S b/arch/powerpc/kernel/exceptions-64e.S
index 930e36099015..f6b36083a8c2 100644
--- a/arch/powerpc/kernel/exceptions-64e.S
+++ b/arch/powerpc/kernel/exceptions-64e.S
@@ -455,7 +455,7 @@ exc_##n##_bad_stack:							    \
 	EXCEPTION_COMMON(trapnum)					\
 	ack(r8);							\
 	CHECK_NAPPING();						\
-	addi	r3,r1,STACK_FRAME_OVERHEAD;				\
+	addi	r3,r1,STACK_INT_FRAME_REGS;				\
 	bl	hdlr;							\
 	b	interrupt_return
 
@@ -504,7 +504,7 @@ __end_interrupts:
 	EXCEPTION_COMMON_CRIT(0x100)
 	bl	special_reg_save
 	CHECK_NAPPING();
-	addi	r3,r1,STACK_FRAME_OVERHEAD
+	addi	r3,r1,STACK_INT_FRAME_REGS
 	bl	unknown_nmi_exception
 	b	ret_from_crit_except
 
@@ -515,7 +515,7 @@ __end_interrupts:
 	EXCEPTION_COMMON_MC(0x000)
 	bl	special_reg_save
 	CHECK_NAPPING();
-	addi	r3,r1,STACK_FRAME_OVERHEAD
+	addi	r3,r1,STACK_INT_FRAME_REGS
 	bl	machine_check_exception
 	b	ret_from_mc_except
 
@@ -570,7 +570,7 @@ __end_interrupts:
 	std	r14,_ESR(r1)
 	ld	r14,PACA_EXGEN+EX_R14(r13)
 	EXCEPTION_COMMON(0x700)
-	addi	r3,r1,STACK_FRAME_OVERHEAD
+	addi	r3,r1,STACK_INT_FRAME_REGS
 	bl	program_check_exception
 	REST_NVGPRS(r1)
 	b	interrupt_return
@@ -586,7 +586,7 @@ __end_interrupts:
 	beq-	1f
 	bl	load_up_fpu
 	b	fast_interrupt_return
-1:	addi	r3,r1,STACK_FRAME_OVERHEAD
+1:	addi	r3,r1,STACK_INT_FRAME_REGS
 	bl	kernel_fp_unavailable_exception
 	b	interrupt_return
 
@@ -606,7 +606,7 @@ BEGIN_FTR_SECTION
 1:
 END_FTR_SECTION_IFSET(CPU_FTR_ALTIVEC)
 #endif
-	addi	r3,r1,STACK_FRAME_OVERHEAD
+	addi	r3,r1,STACK_INT_FRAME_REGS
 	bl	altivec_unavailable_exception
 	b	interrupt_return
 
@@ -616,7 +616,7 @@ END_FTR_SECTION_IFSET(CPU_FTR_ALTIVEC)
 				BOOKE_INTERRUPT_ALTIVEC_ASSIST,
 				PROLOG_ADDITION_NONE)
 	EXCEPTION_COMMON(0x220)
-	addi	r3,r1,STACK_FRAME_OVERHEAD
+	addi	r3,r1,STACK_INT_FRAME_REGS
 #ifdef CONFIG_ALTIVEC
 BEGIN_FTR_SECTION
 	bl	altivec_assist_exception
@@ -643,7 +643,7 @@ END_FTR_SECTION_IFSET(CPU_FTR_ALTIVEC)
 	EXCEPTION_COMMON_CRIT(0x9f0)
 	bl	special_reg_save
 	CHECK_NAPPING();
-	addi	r3,r1,STACK_FRAME_OVERHEAD
+	addi	r3,r1,STACK_INT_FRAME_REGS
 #ifdef CONFIG_BOOKE_WDT
 	bl	WatchdogException
 #else
@@ -664,7 +664,7 @@ END_FTR_SECTION_IFSET(CPU_FTR_ALTIVEC)
 	NORMAL_EXCEPTION_PROLOG(0xf20, BOOKE_INTERRUPT_AP_UNAVAIL,
 				PROLOG_ADDITION_NONE)
 	EXCEPTION_COMMON(0xf20)
-	addi	r3,r1,STACK_FRAME_OVERHEAD
+	addi	r3,r1,STACK_INT_FRAME_REGS
 	bl	unknown_exception
 	b	interrupt_return
 
@@ -731,7 +731,7 @@ END_FTR_SECTION_IFSET(CPU_FTR_ALTIVEC)
 	ld	r14,PACA_EXCRIT+EX_R14(r13)
 	ld	r15,PACA_EXCRIT+EX_R15(r13)
 	EXCEPTION_COMMON_CRIT(0xd00)
-	addi	r3,r1,STACK_FRAME_OVERHEAD
+	addi	r3,r1,STACK_INT_FRAME_REGS
 	bl	DebugException
 	REST_NVGPRS(r1)
 	b	interrupt_return
@@ -802,7 +802,7 @@ kernel_dbg_exc:
 	ld	r14,PACA_EXDBG+EX_R14(r13)
 	ld	r15,PACA_EXDBG+EX_R15(r13)
 	EXCEPTION_COMMON_DBG(0xd08)
-	addi	r3,r1,STACK_FRAME_OVERHEAD
+	addi	r3,r1,STACK_INT_FRAME_REGS
 	bl	DebugException
 	REST_NVGPRS(r1)
 	b	interrupt_return
@@ -812,7 +812,7 @@ kernel_dbg_exc:
 				PROLOG_ADDITION_NONE)
 	EXCEPTION_COMMON(0x260)
 	CHECK_NAPPING()
-	addi	r3,r1,STACK_FRAME_OVERHEAD
+	addi	r3,r1,STACK_INT_FRAME_REGS
 	bl	performance_monitor_exception
 	b	interrupt_return
 
@@ -827,7 +827,7 @@ kernel_dbg_exc:
 	EXCEPTION_COMMON_CRIT(0x2a0)
 	bl	special_reg_save
 	CHECK_NAPPING();
-	addi	r3,r1,STACK_FRAME_OVERHEAD
+	addi	r3,r1,STACK_INT_FRAME_REGS
 	bl	unknown_nmi_exception
 	b	ret_from_crit_except
 
@@ -839,7 +839,7 @@ kernel_dbg_exc:
 	GDBELL_EXCEPTION_PROLOG(0x2c0, BOOKE_INTERRUPT_GUEST_DBELL,
 			        PROLOG_ADDITION_NONE)
 	EXCEPTION_COMMON(0x2c0)
-	addi	r3,r1,STACK_FRAME_OVERHEAD
+	addi	r3,r1,STACK_INT_FRAME_REGS
 	bl	unknown_exception
 	b	interrupt_return
 
@@ -850,7 +850,7 @@ kernel_dbg_exc:
 	EXCEPTION_COMMON_CRIT(0x2e0)
 	bl	special_reg_save
 	CHECK_NAPPING();
-	addi	r3,r1,STACK_FRAME_OVERHEAD
+	addi	r3,r1,STACK_INT_FRAME_REGS
 	bl	unknown_nmi_exception
 	b	ret_from_crit_except
 
@@ -859,7 +859,7 @@ kernel_dbg_exc:
 	NORMAL_EXCEPTION_PROLOG(0x310, BOOKE_INTERRUPT_HV_SYSCALL,
 			        PROLOG_ADDITION_NONE)
 	EXCEPTION_COMMON(0x310)
-	addi	r3,r1,STACK_FRAME_OVERHEAD
+	addi	r3,r1,STACK_INT_FRAME_REGS
 	bl	unknown_exception
 	b	interrupt_return
 
@@ -868,7 +868,7 @@ kernel_dbg_exc:
 	NORMAL_EXCEPTION_PROLOG(0x320, BOOKE_INTERRUPT_HV_PRIV,
 			        PROLOG_ADDITION_NONE)
 	EXCEPTION_COMMON(0x320)
-	addi	r3,r1,STACK_FRAME_OVERHEAD
+	addi	r3,r1,STACK_INT_FRAME_REGS
 	bl	unknown_exception
 	b	interrupt_return
 
@@ -877,7 +877,7 @@ kernel_dbg_exc:
 	NORMAL_EXCEPTION_PROLOG(0x340, BOOKE_INTERRUPT_LRAT_ERROR,
 			        PROLOG_ADDITION_NONE)
 	EXCEPTION_COMMON(0x340)
-	addi	r3,r1,STACK_FRAME_OVERHEAD
+	addi	r3,r1,STACK_INT_FRAME_REGS
 	bl	unknown_exception
 	b	interrupt_return
 
@@ -972,7 +972,7 @@ masked_interrupt_book3e_0x2c0:
  * original values stashed away in the PACA
  */
 storage_fault_common:
-	addi	r3,r1,STACK_FRAME_OVERHEAD
+	addi	r3,r1,STACK_INT_FRAME_REGS
 	bl	do_page_fault
 	b	interrupt_return
 
@@ -981,7 +981,7 @@ storage_fault_common:
  * continues here.
  */
 alignment_more:
-	addi	r3,r1,STACK_FRAME_OVERHEAD
+	addi	r3,r1,STACK_INT_FRAME_REGS
 	bl	alignment_exception
 	REST_NVGPRS(r1)
 	b	interrupt_return
@@ -1062,7 +1062,7 @@ bad_stack_book3e:
 	ZEROIZE_GPR(12)
 	std	r12,0(r11)
 	LOAD_PACA_TOC()
-1:	addi	r3,r1,STACK_FRAME_OVERHEAD
+1:	addi	r3,r1,STACK_INT_FRAME_REGS
 	bl	kernel_bad_stack
 	b	1b
 
diff --git a/arch/powerpc/kernel/exceptions-64s.S b/arch/powerpc/kernel/exceptions-64s.S
index 5381a43e50fe..907054940a07 100644
--- a/arch/powerpc/kernel/exceptions-64s.S
+++ b/arch/powerpc/kernel/exceptions-64s.S
@@ -1061,7 +1061,7 @@ EXC_COMMON_BEGIN(system_reset_common)
 	subi	r1,r1,INT_FRAME_SIZE
 	__GEN_COMMON_BODY system_reset
 
-	addi	r3,r1,STACK_FRAME_OVERHEAD
+	addi	r3,r1,STACK_INT_FRAME_REGS
 	bl	system_reset_exception
 
 	/* Clear MSR_RI before setting SRR0 and SRR1. */
@@ -1208,7 +1208,7 @@ EXC_COMMON_BEGIN(machine_check_early_common)
 BEGIN_FTR_SECTION
 	bl	enable_machine_check
 END_FTR_SECTION_IFSET(CPU_FTR_HVMODE)
-	addi	r3,r1,STACK_FRAME_OVERHEAD
+	addi	r3,r1,STACK_INT_FRAME_REGS
 BEGIN_FTR_SECTION
 	bl	machine_check_early_boot
 END_FTR_SECTION(0, 1)     // nop out after boot
@@ -1298,7 +1298,7 @@ EXC_COMMON_BEGIN(machine_check_common)
 	 * save area: PACA_EXMC instead of PACA_EXGEN.
 	 */
 	GEN_COMMON machine_check
-	addi	r3,r1,STACK_FRAME_OVERHEAD
+	addi	r3,r1,STACK_INT_FRAME_REGS
 	bl	machine_check_exception_async
 	b	interrupt_return_srr
 
@@ -1364,14 +1364,14 @@ END_FTR_SECTION_IFSET(CPU_FTR_HVMODE)
 	 * This is the NMI version of the handler because we are called from
 	 * the early handler which is a true NMI.
 	 */
-	addi	r3,r1,STACK_FRAME_OVERHEAD
+	addi	r3,r1,STACK_INT_FRAME_REGS
 	bl	machine_check_exception
 
 	/*
 	 * We will not reach here. Even if we did, there is no way out.
 	 * Call unrecoverable_exception and die.
 	 */
-	addi	r3,r1,STACK_FRAME_OVERHEAD
+	addi	r3,r1,STACK_INT_FRAME_REGS
 	bl	unrecoverable_exception
 	b	.
 
@@ -1422,7 +1422,7 @@ EXC_VIRT_END(data_access, 0x4300, 0x80)
 EXC_COMMON_BEGIN(data_access_common)
 	GEN_COMMON data_access
 	ld	r4,_DSISR(r1)
-	addi	r3,r1,STACK_FRAME_OVERHEAD
+	addi	r3,r1,STACK_INT_FRAME_REGS
 	andis.	r0,r4,DSISR_DABRMATCH@h
 	bne-	1f
 #ifdef CONFIG_PPC_64S_HASH_MMU
@@ -1479,7 +1479,7 @@ EXC_COMMON_BEGIN(data_access_slb_common)
 #ifdef CONFIG_PPC_64S_HASH_MMU
 BEGIN_MMU_FTR_SECTION
 	/* HPT case, do SLB fault */
-	addi	r3,r1,STACK_FRAME_OVERHEAD
+	addi	r3,r1,STACK_INT_FRAME_REGS
 	bl	do_slb_fault
 	cmpdi	r3,0
 	bne-	1f
@@ -1493,7 +1493,7 @@ ALT_MMU_FTR_SECTION_END_IFCLR(MMU_FTR_TYPE_RADIX)
 	li	r3,-EFAULT
 #endif
 	std	r3,RESULT(r1)
-	addi	r3,r1,STACK_FRAME_OVERHEAD
+	addi	r3,r1,STACK_INT_FRAME_REGS
 	bl	do_bad_segment_interrupt
 	b	interrupt_return_srr
 
@@ -1525,7 +1525,7 @@ EXC_VIRT_BEGIN(instruction_access, 0x4400, 0x80)
 EXC_VIRT_END(instruction_access, 0x4400, 0x80)
 EXC_COMMON_BEGIN(instruction_access_common)
 	GEN_COMMON instruction_access
-	addi	r3,r1,STACK_FRAME_OVERHEAD
+	addi	r3,r1,STACK_INT_FRAME_REGS
 #ifdef CONFIG_PPC_64S_HASH_MMU
 BEGIN_MMU_FTR_SECTION
 	bl	do_hash_fault
@@ -1567,7 +1567,7 @@ EXC_COMMON_BEGIN(instruction_access_slb_common)
 #ifdef CONFIG_PPC_64S_HASH_MMU
 BEGIN_MMU_FTR_SECTION
 	/* HPT case, do SLB fault */
-	addi	r3,r1,STACK_FRAME_OVERHEAD
+	addi	r3,r1,STACK_INT_FRAME_REGS
 	bl	do_slb_fault
 	cmpdi	r3,0
 	bne-	1f
@@ -1581,7 +1581,7 @@ ALT_MMU_FTR_SECTION_END_IFCLR(MMU_FTR_TYPE_RADIX)
 	li	r3,-EFAULT
 #endif
 	std	r3,RESULT(r1)
-	addi	r3,r1,STACK_FRAME_OVERHEAD
+	addi	r3,r1,STACK_INT_FRAME_REGS
 	bl	do_bad_segment_interrupt
 	b	interrupt_return_srr
 
@@ -1635,7 +1635,7 @@ EXC_VIRT_BEGIN(hardware_interrupt, 0x4500, 0x100)
 EXC_VIRT_END(hardware_interrupt, 0x4500, 0x100)
 EXC_COMMON_BEGIN(hardware_interrupt_common)
 	GEN_COMMON hardware_interrupt
-	addi	r3,r1,STACK_FRAME_OVERHEAD
+	addi	r3,r1,STACK_INT_FRAME_REGS
 	bl	do_IRQ
 	BEGIN_FTR_SECTION
 	b	interrupt_return_hsrr
@@ -1665,7 +1665,7 @@ EXC_VIRT_BEGIN(alignment, 0x4600, 0x100)
 EXC_VIRT_END(alignment, 0x4600, 0x100)
 EXC_COMMON_BEGIN(alignment_common)
 	GEN_COMMON alignment
-	addi	r3,r1,STACK_FRAME_OVERHEAD
+	addi	r3,r1,STACK_INT_FRAME_REGS
 	bl	alignment_exception
 	REST_NVGPRS(r1) /* instruction emulation may change GPRs */
 	b	interrupt_return_srr
@@ -1731,7 +1731,7 @@ EXC_COMMON_BEGIN(program_check_common)
 	__GEN_COMMON_BODY program_check
 
 .Ldo_program_check:
-	addi	r3,r1,STACK_FRAME_OVERHEAD
+	addi	r3,r1,STACK_INT_FRAME_REGS
 	bl	program_check_exception
 	REST_NVGPRS(r1) /* instruction emulation may change GPRs */
 	b	interrupt_return_srr
@@ -1762,7 +1762,7 @@ EXC_VIRT_END(fp_unavailable, 0x4800, 0x100)
 EXC_COMMON_BEGIN(fp_unavailable_common)
 	GEN_COMMON fp_unavailable
 	bne	1f			/* if from user, just load it up */
-	addi	r3,r1,STACK_FRAME_OVERHEAD
+	addi	r3,r1,STACK_INT_FRAME_REGS
 	bl	kernel_fp_unavailable_exception
 0:	trap
 	EMIT_BUG_ENTRY 0b, __FILE__, __LINE__, 0
@@ -1780,7 +1780,7 @@ END_FTR_SECTION_IFSET(CPU_FTR_TM)
 	b	fast_interrupt_return_srr
 #ifdef CONFIG_PPC_TRANSACTIONAL_MEM
 2:	/* User process was in a transaction */
-	addi	r3,r1,STACK_FRAME_OVERHEAD
+	addi	r3,r1,STACK_INT_FRAME_REGS
 	bl	fp_unavailable_tm
 	b	interrupt_return_srr
 #endif
@@ -1824,7 +1824,7 @@ EXC_VIRT_BEGIN(decrementer, 0x4900, 0x80)
 EXC_VIRT_END(decrementer, 0x4900, 0x80)
 EXC_COMMON_BEGIN(decrementer_common)
 	GEN_COMMON decrementer
-	addi	r3,r1,STACK_FRAME_OVERHEAD
+	addi	r3,r1,STACK_INT_FRAME_REGS
 	bl	timer_interrupt
 	b	interrupt_return_srr
 
@@ -1909,7 +1909,7 @@ EXC_VIRT_BEGIN(doorbell_super, 0x4a00, 0x100)
 EXC_VIRT_END(doorbell_super, 0x4a00, 0x100)
 EXC_COMMON_BEGIN(doorbell_super_common)
 	GEN_COMMON doorbell_super
-	addi	r3,r1,STACK_FRAME_OVERHEAD
+	addi	r3,r1,STACK_INT_FRAME_REGS
 #ifdef CONFIG_PPC_DOORBELL
 	bl	doorbell_exception
 #else
@@ -2076,7 +2076,7 @@ EXC_VIRT_BEGIN(single_step, 0x4d00, 0x100)
 EXC_VIRT_END(single_step, 0x4d00, 0x100)
 EXC_COMMON_BEGIN(single_step_common)
 	GEN_COMMON single_step
-	addi	r3,r1,STACK_FRAME_OVERHEAD
+	addi	r3,r1,STACK_INT_FRAME_REGS
 	bl	single_step_exception
 	b	interrupt_return_srr
 
@@ -2110,7 +2110,7 @@ EXC_VIRT_BEGIN(h_data_storage, 0x4e00, 0x20)
 EXC_VIRT_END(h_data_storage, 0x4e00, 0x20)
 EXC_COMMON_BEGIN(h_data_storage_common)
 	GEN_COMMON h_data_storage
-	addi    r3,r1,STACK_FRAME_OVERHEAD
+	addi    r3,r1,STACK_INT_FRAME_REGS
 BEGIN_MMU_FTR_SECTION
 	bl      do_bad_page_fault_segv
 MMU_FTR_SECTION_ELSE
@@ -2139,7 +2139,7 @@ EXC_VIRT_BEGIN(h_instr_storage, 0x4e20, 0x20)
 EXC_VIRT_END(h_instr_storage, 0x4e20, 0x20)
 EXC_COMMON_BEGIN(h_instr_storage_common)
 	GEN_COMMON h_instr_storage
-	addi	r3,r1,STACK_FRAME_OVERHEAD
+	addi	r3,r1,STACK_INT_FRAME_REGS
 	bl	unknown_exception
 	b	interrupt_return_hsrr
 
@@ -2162,7 +2162,7 @@ EXC_VIRT_BEGIN(emulation_assist, 0x4e40, 0x20)
 EXC_VIRT_END(emulation_assist, 0x4e40, 0x20)
 EXC_COMMON_BEGIN(emulation_assist_common)
 	GEN_COMMON emulation_assist
-	addi	r3,r1,STACK_FRAME_OVERHEAD
+	addi	r3,r1,STACK_INT_FRAME_REGS
 	bl	emulation_assist_interrupt
 	REST_NVGPRS(r1) /* instruction emulation may change GPRs */
 	b	interrupt_return_hsrr
@@ -2222,7 +2222,7 @@ EXC_COMMON_BEGIN(hmi_exception_early_common)
 
 	__GEN_COMMON_BODY hmi_exception_early
 
-	addi	r3,r1,STACK_FRAME_OVERHEAD
+	addi	r3,r1,STACK_INT_FRAME_REGS
 	bl	hmi_exception_realmode
 	cmpdi	cr0,r3,0
 	bne	1f
@@ -2240,7 +2240,7 @@ EXC_COMMON_BEGIN(hmi_exception_early_common)
 
 EXC_COMMON_BEGIN(hmi_exception_common)
 	GEN_COMMON hmi_exception
-	addi	r3,r1,STACK_FRAME_OVERHEAD
+	addi	r3,r1,STACK_INT_FRAME_REGS
 	bl	handle_hmi_exception
 	b	interrupt_return_hsrr
 
@@ -2274,7 +2274,7 @@ EXC_VIRT_BEGIN(h_doorbell, 0x4e80, 0x20)
 EXC_VIRT_END(h_doorbell, 0x4e80, 0x20)
 EXC_COMMON_BEGIN(h_doorbell_common)
 	GEN_COMMON h_doorbell
-	addi	r3,r1,STACK_FRAME_OVERHEAD
+	addi	r3,r1,STACK_INT_FRAME_REGS
 #ifdef CONFIG_PPC_DOORBELL
 	bl	doorbell_exception
 #else
@@ -2310,7 +2310,7 @@ EXC_VIRT_BEGIN(h_virt_irq, 0x4ea0, 0x20)
 EXC_VIRT_END(h_virt_irq, 0x4ea0, 0x20)
 EXC_COMMON_BEGIN(h_virt_irq_common)
 	GEN_COMMON h_virt_irq
-	addi	r3,r1,STACK_FRAME_OVERHEAD
+	addi	r3,r1,STACK_INT_FRAME_REGS
 	bl	do_IRQ
 	b	interrupt_return_hsrr
 
@@ -2356,7 +2356,7 @@ EXC_VIRT_BEGIN(performance_monitor, 0x4f00, 0x20)
 EXC_VIRT_END(performance_monitor, 0x4f00, 0x20)
 EXC_COMMON_BEGIN(performance_monitor_common)
 	GEN_COMMON performance_monitor
-	addi	r3,r1,STACK_FRAME_OVERHEAD
+	addi	r3,r1,STACK_INT_FRAME_REGS
 	bl	performance_monitor_exception
 	b	interrupt_return_srr
 
@@ -2398,14 +2398,14 @@ BEGIN_FTR_SECTION
 	b	fast_interrupt_return_srr
 #ifdef CONFIG_PPC_TRANSACTIONAL_MEM
 2:	/* User process was in a transaction */
-	addi	r3,r1,STACK_FRAME_OVERHEAD
+	addi	r3,r1,STACK_INT_FRAME_REGS
 	bl	altivec_unavailable_tm
 	b	interrupt_return_srr
 #endif
 1:
 END_FTR_SECTION_IFSET(CPU_FTR_ALTIVEC)
 #endif
-	addi	r3,r1,STACK_FRAME_OVERHEAD
+	addi	r3,r1,STACK_INT_FRAME_REGS
 	bl	altivec_unavailable_exception
 	b	interrupt_return_srr
 
@@ -2446,14 +2446,14 @@ BEGIN_FTR_SECTION
 	b	load_up_vsx
 #ifdef CONFIG_PPC_TRANSACTIONAL_MEM
 2:	/* User process was in a transaction */
-	addi	r3,r1,STACK_FRAME_OVERHEAD
+	addi	r3,r1,STACK_INT_FRAME_REGS
 	bl	vsx_unavailable_tm
 	b	interrupt_return_srr
 #endif
 1:
 END_FTR_SECTION_IFSET(CPU_FTR_VSX)
 #endif
-	addi	r3,r1,STACK_FRAME_OVERHEAD
+	addi	r3,r1,STACK_INT_FRAME_REGS
 	bl	vsx_unavailable_exception
 	b	interrupt_return_srr
 
@@ -2480,7 +2480,7 @@ EXC_VIRT_BEGIN(facility_unavailable, 0x4f60, 0x20)
 EXC_VIRT_END(facility_unavailable, 0x4f60, 0x20)
 EXC_COMMON_BEGIN(facility_unavailable_common)
 	GEN_COMMON facility_unavailable
-	addi	r3,r1,STACK_FRAME_OVERHEAD
+	addi	r3,r1,STACK_INT_FRAME_REGS
 	bl	facility_unavailable_exception
 	REST_NVGPRS(r1) /* instruction emulation may change GPRs */
 	b	interrupt_return_srr
@@ -2508,7 +2508,7 @@ EXC_VIRT_BEGIN(h_facility_unavailable, 0x4f80, 0x20)
 EXC_VIRT_END(h_facility_unavailable, 0x4f80, 0x20)
 EXC_COMMON_BEGIN(h_facility_unavailable_common)
 	GEN_COMMON h_facility_unavailable
-	addi	r3,r1,STACK_FRAME_OVERHEAD
+	addi	r3,r1,STACK_INT_FRAME_REGS
 	bl	facility_unavailable_exception
 	REST_NVGPRS(r1) /* XXX Shouldn't be necessary in practice */
 	b	interrupt_return_hsrr
@@ -2538,7 +2538,7 @@ EXC_REAL_END(cbe_system_error, 0x1200, 0x100)
 EXC_VIRT_NONE(0x5200, 0x100)
 EXC_COMMON_BEGIN(cbe_system_error_common)
 	GEN_COMMON cbe_system_error
-	addi	r3,r1,STACK_FRAME_OVERHEAD
+	addi	r3,r1,STACK_INT_FRAME_REGS
 	bl	cbe_system_error_exception
 	b	interrupt_return_hsrr
 
@@ -2569,7 +2569,7 @@ EXC_VIRT_BEGIN(instruction_breakpoint, 0x5300, 0x100)
 EXC_VIRT_END(instruction_breakpoint, 0x5300, 0x100)
 EXC_COMMON_BEGIN(instruction_breakpoint_common)
 	GEN_COMMON instruction_breakpoint
-	addi	r3,r1,STACK_FRAME_OVERHEAD
+	addi	r3,r1,STACK_INT_FRAME_REGS
 	bl	instruction_breakpoint_exception
 	b	interrupt_return_srr
 
@@ -2691,7 +2691,7 @@ END_FTR_SECTION_IFSET(CPU_FTR_CFAR)
 
 EXC_COMMON_BEGIN(denorm_exception_common)
 	GEN_COMMON denorm_exception
-	addi	r3,r1,STACK_FRAME_OVERHEAD
+	addi	r3,r1,STACK_INT_FRAME_REGS
 	bl	unknown_exception
 	b	interrupt_return_hsrr
 
@@ -2708,7 +2708,7 @@ EXC_REAL_END(cbe_maintenance, 0x1600, 0x100)
 EXC_VIRT_NONE(0x5600, 0x100)
 EXC_COMMON_BEGIN(cbe_maintenance_common)
 	GEN_COMMON cbe_maintenance
-	addi	r3,r1,STACK_FRAME_OVERHEAD
+	addi	r3,r1,STACK_INT_FRAME_REGS
 	bl	cbe_maintenance_exception
 	b	interrupt_return_hsrr
 
@@ -2733,7 +2733,7 @@ EXC_VIRT_BEGIN(altivec_assist, 0x5700, 0x100)
 EXC_VIRT_END(altivec_assist, 0x5700, 0x100)
 EXC_COMMON_BEGIN(altivec_assist_common)
 	GEN_COMMON altivec_assist
-	addi	r3,r1,STACK_FRAME_OVERHEAD
+	addi	r3,r1,STACK_INT_FRAME_REGS
 #ifdef CONFIG_ALTIVEC
 	bl	altivec_assist_exception
 	REST_NVGPRS(r1) /* instruction emulation may change GPRs */
@@ -2755,7 +2755,7 @@ EXC_REAL_END(cbe_thermal, 0x1800, 0x100)
 EXC_VIRT_NONE(0x5800, 0x100)
 EXC_COMMON_BEGIN(cbe_thermal_common)
 	GEN_COMMON cbe_thermal
-	addi	r3,r1,STACK_FRAME_OVERHEAD
+	addi	r3,r1,STACK_INT_FRAME_REGS
 	bl	cbe_thermal_exception
 	b	interrupt_return_hsrr
 
@@ -2788,7 +2788,7 @@ EXC_COMMON_BEGIN(soft_nmi_common)
 	subi	r1,r1,INT_FRAME_SIZE
 	__GEN_COMMON_BODY soft_nmi
 
-	addi	r3,r1,STACK_FRAME_OVERHEAD
+	addi	r3,r1,STACK_INT_FRAME_REGS
 	bl	soft_nmi_interrupt
 
 	/* Clear MSR_RI before setting SRR0 and SRR1. */
diff --git a/arch/powerpc/kernel/head_32.h b/arch/powerpc/kernel/head_32.h
index c3286260a7d1..117d25330e13 100644
--- a/arch/powerpc/kernel/head_32.h
+++ b/arch/powerpc/kernel/head_32.h
@@ -127,7 +127,7 @@ _ASM_NOKPROBE_SYMBOL(\name\()_virt)
 	mfspr	r10,SPRN_XER
 	addi	r2, r2, -THREAD
 	stw	r10,_XER(r1)
-	addi	r3,r1,STACK_FRAME_OVERHEAD
+	addi	r3,r1,STACK_INT_FRAME_REGS
 .endm
 
 .macro prepare_transfer_to_handler
diff --git a/arch/powerpc/kernel/head_85xx.S b/arch/powerpc/kernel/head_85xx.S
index 52c0ab416326..24f39abf81df 100644
--- a/arch/powerpc/kernel/head_85xx.S
+++ b/arch/powerpc/kernel/head_85xx.S
@@ -972,10 +972,10 @@ _GLOBAL(__giveup_spe)
 	li	r4,THREAD_ACC
 	evstddx	evr6, r4, r3		/* save off accumulator */
 	beq	1f
-	lwz	r4,_MSR-STACK_FRAME_OVERHEAD(r5)
+	lwz	r4,_MSR-STACK_INT_FRAME_REGS(r5)
 	lis	r3,MSR_SPE@h
 	andc	r4,r4,r3		/* disable SPE for previous task */
-	stw	r4,_MSR-STACK_FRAME_OVERHEAD(r5)
+	stw	r4,_MSR-STACK_INT_FRAME_REGS(r5)
 1:
 	blr
 #endif /* CONFIG_SPE */
diff --git a/arch/powerpc/kernel/head_booke.h b/arch/powerpc/kernel/head_booke.h
index 1cb9d0f7cbf2..3149ac20b18e 100644
--- a/arch/powerpc/kernel/head_booke.h
+++ b/arch/powerpc/kernel/head_booke.h
@@ -99,7 +99,7 @@ END_BTB_FLUSH_SECTION
 	mfspr	r10,SPRN_XER
 	addi	r2, r2, -THREAD
 	stw	r10,_XER(r1)
-	addi	r3,r1,STACK_FRAME_OVERHEAD
+	addi	r3,r1,STACK_INT_FRAME_REGS
 .endm
 
 .macro prepare_transfer_to_handler
diff --git a/arch/powerpc/kernel/interrupt_64.S b/arch/powerpc/kernel/interrupt_64.S
index 978a173eb339..12807d5298a3 100644
--- a/arch/powerpc/kernel/interrupt_64.S
+++ b/arch/powerpc/kernel/interrupt_64.S
@@ -78,7 +78,7 @@ _ASM_NOKPROBE_SYMBOL(system_call_vectored_\name)
 	std	r12,_CCR(r1)
 	std	r3,ORIG_GPR3(r1)
 	/* Calling convention has r3 = regs, r4 = orig r0 */
-	addi	r3,r1,STACK_FRAME_OVERHEAD
+	addi	r3,r1,STACK_INT_FRAME_REGS
 	mr	r4,r0
 	LOAD_REG_IMMEDIATE(r11, STACK_FRAME_REGS_MARKER)
 	std	r11,-16(r3)		/* "regshere" marker */
@@ -99,7 +99,7 @@ END_FTR_SECTION_IFSET(CPU_FTR_HAS_PPR)
 	bl	system_call_exception
 
 .Lsyscall_vectored_\name\()_exit:
-	addi	r4,r1,STACK_FRAME_OVERHEAD
+	addi	r4,r1,STACK_INT_FRAME_REGS
 	li	r5,1 /* scv */
 	bl	syscall_exit_prepare
 	std	r1,PACA_EXIT_SAVE_R1(r13) /* save r1 for restart */
@@ -176,7 +176,7 @@ _ASM_NOKPROBE_SYMBOL(syscall_vectored_\name\()_restart)
 	ld	r1,PACA_EXIT_SAVE_R1(r13)
 	LOAD_PACA_TOC()
 	ld	r3,RESULT(r1)
-	addi	r4,r1,STACK_FRAME_OVERHEAD
+	addi	r4,r1,STACK_INT_FRAME_REGS
 	li	r11,IRQS_ALL_DISABLED
 	stb	r11,PACAIRQSOFTMASK(r13)
 	bl	syscall_exit_restart
@@ -251,7 +251,7 @@ END_BTB_FLUSH_SECTION
 	std	r12,_CCR(r1)
 	std	r3,ORIG_GPR3(r1)
 	/* Calling convention has r3 = regs, r4 = orig r0 */
-	addi	r3,r1,STACK_FRAME_OVERHEAD
+	addi	r3,r1,STACK_INT_FRAME_REGS
 	mr	r4,r0
 	LOAD_REG_IMMEDIATE(r11, STACK_FRAME_REGS_MARKER)
 	std	r11,-16(r3)		/* "regshere" marker */
@@ -278,7 +278,7 @@ END_BTB_FLUSH_SECTION
 	bl	system_call_exception
 
 .Lsyscall_exit:
-	addi	r4,r1,STACK_FRAME_OVERHEAD
+	addi	r4,r1,STACK_INT_FRAME_REGS
 	li	r5,0 /* !scv */
 	bl	syscall_exit_prepare
 	std	r1,PACA_EXIT_SAVE_R1(r13) /* save r1 for restart */
@@ -357,7 +357,7 @@ _ASM_NOKPROBE_SYMBOL(syscall_restart)
 	ld	r1,PACA_EXIT_SAVE_R1(r13)
 	LOAD_PACA_TOC()
 	ld	r3,RESULT(r1)
-	addi	r4,r1,STACK_FRAME_OVERHEAD
+	addi	r4,r1,STACK_INT_FRAME_REGS
 	li	r11,IRQS_ALL_DISABLED
 	stb	r11,PACAIRQSOFTMASK(r13)
 	bl	syscall_exit_restart
@@ -388,7 +388,7 @@ _ASM_NOKPROBE_SYMBOL(fast_interrupt_return_srr)
 	andi.	r0,r5,MSR_RI
 	li	r3,0 /* 0 return value, no EMULATE_STACK_STORE */
 	bne+	.Lfast_kernel_interrupt_return_srr
-	addi	r3,r1,STACK_FRAME_OVERHEAD
+	addi	r3,r1,STACK_INT_FRAME_REGS
 	bl	unrecoverable_exception
 	b	. /* should not get here */
 #else
@@ -406,7 +406,7 @@ _ASM_NOKPROBE_SYMBOL(interrupt_return_\srr\())
 	beq	interrupt_return_\srr\()_kernel
 interrupt_return_\srr\()_user: /* make backtraces match the _kernel variant */
 _ASM_NOKPROBE_SYMBOL(interrupt_return_\srr\()_user)
-	addi	r3,r1,STACK_FRAME_OVERHEAD
+	addi	r3,r1,STACK_INT_FRAME_REGS
 	bl	interrupt_exit_user_prepare
 	cmpdi	r3,0
 	bne-	.Lrestore_nvgprs_\srr
@@ -503,7 +503,7 @@ _ASM_NOKPROBE_SYMBOL(interrupt_return_\srr\()_user_restart)
 	GET_PACA(r13)
 	ld	r1,PACA_EXIT_SAVE_R1(r13)
 	LOAD_PACA_TOC()
-	addi	r3,r1,STACK_FRAME_OVERHEAD
+	addi	r3,r1,STACK_INT_FRAME_REGS
 	li	r11,IRQS_ALL_DISABLED
 	stb	r11,PACAIRQSOFTMASK(r13)
 	bl	interrupt_exit_user_restart
@@ -518,7 +518,7 @@ RESTART_TABLE(.Linterrupt_return_\srr\()_user_rst_start, .Linterrupt_return_\srr
 	.balign IFETCH_ALIGN_BYTES
 interrupt_return_\srr\()_kernel:
 _ASM_NOKPROBE_SYMBOL(interrupt_return_\srr\()_kernel)
-	addi	r3,r1,STACK_FRAME_OVERHEAD
+	addi	r3,r1,STACK_INT_FRAME_REGS
 	bl	interrupt_exit_kernel_prepare
 
 	std	r1,PACA_EXIT_SAVE_R1(r13) /* save r1 for restart */
@@ -675,7 +675,7 @@ _ASM_NOKPROBE_SYMBOL(interrupt_return_\srr\()_kernel_restart)
 	GET_PACA(r13)
 	ld	r1,PACA_EXIT_SAVE_R1(r13)
 	LOAD_PACA_TOC()
-	addi	r3,r1,STACK_FRAME_OVERHEAD
+	addi	r3,r1,STACK_INT_FRAME_REGS
 	li	r11,IRQS_ALL_DISABLED
 	stb	r11,PACAIRQSOFTMASK(r13)
 	bl	interrupt_exit_kernel_restart
diff --git a/arch/powerpc/kernel/kgdb.c b/arch/powerpc/kernel/kgdb.c
index 1a1e9995dae3..ebe4d1645ca1 100644
--- a/arch/powerpc/kernel/kgdb.c
+++ b/arch/powerpc/kernel/kgdb.c
@@ -191,7 +191,7 @@ static int kgdb_break_match(struct pt_regs *regs)
 void sleeping_thread_to_gdb_regs(unsigned long *gdb_regs, struct task_struct *p)
 {
 	struct pt_regs *regs = (struct pt_regs *)(p->thread.ksp +
-						  STACK_FRAME_OVERHEAD);
+						  STACK_INT_FRAME_REGS);
 	unsigned long *ptr = gdb_regs;
 	int reg;
 
diff --git a/arch/powerpc/kernel/optprobes_head.S b/arch/powerpc/kernel/optprobes_head.S
index cd4e7bc32609..35932f45fb4e 100644
--- a/arch/powerpc/kernel/optprobes_head.S
+++ b/arch/powerpc/kernel/optprobes_head.S
@@ -85,7 +85,7 @@ optprobe_template_op_address:
 	TEMPLATE_FOR_IMM_LOAD_INSNS
 
 	/* 2. pt_regs pointer in r4 */
-	addi	r4,r1,STACK_FRAME_OVERHEAD
+	addi	r4,r1,STACK_INT_FRAME_REGS
 
 	.global optprobe_template_call_handler
 optprobe_template_call_handler:
@@ -96,7 +96,7 @@ optprobe_template_call_handler:
 	 * Parameters for instruction emulation:
 	 * 1. Pass SP in register r3.
 	 */
-	addi	r3,r1,STACK_FRAME_OVERHEAD
+	addi	r3,r1,STACK_INT_FRAME_REGS
 
 	.global optprobe_template_insn
 optprobe_template_insn:
diff --git a/arch/powerpc/kernel/ppc_save_regs.S b/arch/powerpc/kernel/ppc_save_regs.S
index 2908a5c2cd2f..272b632fe0d5 100644
--- a/arch/powerpc/kernel/ppc_save_regs.S
+++ b/arch/powerpc/kernel/ppc_save_regs.S
@@ -22,7 +22,7 @@
  */
 _GLOBAL(ppc_save_regs)
 	/* This allows stack frame accessor macros and offsets to be used */
-	subi	r3,r3,STACK_FRAME_OVERHEAD
+	subi	r3,r3,STACK_INT_FRAME_REGS
 	SAVE_GPRS(0, 31, r3)
 #ifdef CONFIG_PPC64
 	lbz	r0,PACAIRQSOFTMASK(r13)
diff --git a/arch/powerpc/kernel/process.c b/arch/powerpc/kernel/process.c
index acfa197fb2df..e7010f71de24 100644
--- a/arch/powerpc/kernel/process.c
+++ b/arch/powerpc/kernel/process.c
@@ -2231,12 +2231,12 @@ void __no_sanitize_address show_stack(struct task_struct *tsk,
 
 		/*
 		 * See if this is an exception frame.
-		 * We look for the "regshere" marker in the current frame.
+		 * We look for the "regs" marker in the current frame.
 		 */
 		if (validate_sp(sp, tsk, STACK_FRAME_WITH_PT_REGS)
 		    && stack[STACK_FRAME_MARKER] == STACK_FRAME_REGS_MARKER) {
 			struct pt_regs *regs = (struct pt_regs *)
-				(sp + STACK_FRAME_OVERHEAD);
+				(sp + STACK_INT_FRAME_REGS);
 
 			lr = regs->link;
 			printk("%s--- interrupt: %lx at %pS\n",
diff --git a/arch/powerpc/kernel/tm.S b/arch/powerpc/kernel/tm.S
index 5a0f023a26e9..9feab5e0485b 100644
--- a/arch/powerpc/kernel/tm.S
+++ b/arch/powerpc/kernel/tm.S
@@ -117,7 +117,7 @@ _GLOBAL(tm_reclaim)
 	std	r2, STK_GOT(r1)
 	stdu	r1, -TM_FRAME_SIZE(r1)
 
-	/* We've a struct pt_regs at [r1+STACK_FRAME_OVERHEAD]. */
+	/* We've a struct pt_regs at [r1+STACK_INT_FRAME_REGS]. */
 
 	std	r3, STK_PARAM(R3)(r1)
 	SAVE_NVGPRS(r1)
@@ -222,7 +222,7 @@ _GLOBAL(tm_reclaim)
 	 * Make r7 look like an exception frame so that we can use the neat
 	 * GPRx(n) macros. r7 is NOT a pt_regs ptr!
 	 */
-	subi	r7, r7, STACK_FRAME_OVERHEAD
+	subi	r7, r7, STACK_INT_FRAME_REGS
 
 	/* Sync the userland GPRs 2-12, 14-31 to thread->regs: */
 	SAVE_GPR(0, r7)				/* user r0 */
@@ -359,7 +359,7 @@ _GLOBAL(__tm_recheckpoint)
 	stdu	r1, -TM_FRAME_SIZE(r1)
 
 	/*
-	 * We've a struct pt_regs at [r1+STACK_FRAME_OVERHEAD].
+	 * We've a struct pt_regs at [r1+STACK_INT_FRAME_REGS].
 	 * This is used for backing up the NVGPRs:
 	 */
 	SAVE_NVGPRS(r1)
@@ -379,7 +379,7 @@ _GLOBAL(__tm_recheckpoint)
 	 * Make r7 look like an exception frame so that we can use the neat
 	 * GPRx(n) macros. r7 is now NOT a pt_regs ptr!
 	 */
-	subi	r7, r7, STACK_FRAME_OVERHEAD
+	subi	r7, r7, STACK_INT_FRAME_REGS
 
 	/* We need to setup MSR for FP/VMX/VSX register save instructions. */
 	mfmsr	r6
diff --git a/arch/powerpc/kernel/trace/ftrace_mprofile.S b/arch/powerpc/kernel/trace/ftrace_mprofile.S
index d031093bc436..ffb1db386849 100644
--- a/arch/powerpc/kernel/trace/ftrace_mprofile.S
+++ b/arch/powerpc/kernel/trace/ftrace_mprofile.S
@@ -110,7 +110,7 @@
 	.endif
 
 	/* Load &pt_regs in r6 for call below */
-	addi    r6, r1, STACK_FRAME_OVERHEAD
+	addi    r6, r1, STACK_INT_FRAME_REGS
 .endm
 
 .macro	ftrace_regs_exit allregs
diff --git a/arch/powerpc/lib/test_emulate_step_exec_instr.S b/arch/powerpc/lib/test_emulate_step_exec_instr.S
index 5473f9d03df3..e2b646a4f7fa 100644
--- a/arch/powerpc/lib/test_emulate_step_exec_instr.S
+++ b/arch/powerpc/lib/test_emulate_step_exec_instr.S
@@ -16,7 +16,7 @@ _GLOBAL(exec_instr)
 
 	/*
 	 * Stack frame layout (INT_FRAME_SIZE bytes)
-	 *   In-memory pt_regs	(SP + STACK_FRAME_OVERHEAD)
+	 *   In-memory pt_regs	(SP + STACK_INT_FRAME_REGS)
 	 *   Scratch space	(SP + 8)
 	 *   Back chain		(SP + 0)
 	 */
diff --git a/arch/powerpc/perf/callchain.c b/arch/powerpc/perf/callchain.c
index 8718289c051d..9e254aed1f61 100644
--- a/arch/powerpc/perf/callchain.c
+++ b/arch/powerpc/perf/callchain.c
@@ -67,7 +67,7 @@ perf_callchain_kernel(struct perf_callchain_entry_ctx *entry, struct pt_regs *re
 			 * This looks like an interrupt frame for an
 			 * interrupt that occurred in the kernel
 			 */
-			regs = (struct pt_regs *)(sp + STACK_FRAME_OVERHEAD);
+			regs = (struct pt_regs *)(sp + STACK_INT_FRAME_REGS);
 			next_ip = regs->nip;
 			lr = regs->link;
 			level = 0;
diff --git a/arch/powerpc/xmon/xmon.c b/arch/powerpc/xmon/xmon.c
index f51c882bf902..e403f14eb6eb 100644
--- a/arch/powerpc/xmon/xmon.c
+++ b/arch/powerpc/xmon/xmon.c
@@ -1781,14 +1781,13 @@ static void xmon_show_stack(unsigned long sp, unsigned long lr,
 			xmon_print_symbol(ip, " ", "\n");
 		}
 
-		/* Look for "regshere" marker to see if this is
+		/* Look for "regs" marker to see if this is
 		   an exception frame. */
 		if (mread(sp + MARKER_OFFSET, &marker, sizeof(unsigned long))
 		    && marker == STACK_FRAME_REGS_MARKER) {
-			if (mread(sp + STACK_FRAME_OVERHEAD, &regs, sizeof(regs))
-			    != sizeof(regs)) {
+			if (mread(sp + STACK_INT_FRAME_REGS, &regs, sizeof(regs)) != sizeof(regs)) {
 				printf("Couldn't read registers at %lx\n",
-				       sp + STACK_FRAME_OVERHEAD);
+				       sp + STACK_INT_FRAME_REGS);
 				break;
 			}
 			printf("--- Exception: %lx %s at ", regs.trap,
-- 
2.37.2

[RFC PATCH 08/19] powerpc: add a definition for the marker offset within the interrupt frame

[Nicholas Piggin]

Define a constant rather than open-code the offset for the
"regs" marker.

Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
---
 arch/powerpc/include/asm/ptrace.h       |  2 ++
 arch/powerpc/kernel/entry_32.S          |  2 +-
 arch/powerpc/kernel/exceptions-64e.S    |  2 +-
 arch/powerpc/kernel/exceptions-64s.S    |  2 +-
 arch/powerpc/kernel/head_32.h           |  2 +-
 arch/powerpc/kernel/head_booke.h        |  2 +-
 arch/powerpc/kernel/interrupt_64.S      | 10 +++++-----
 arch/powerpc/kvm/book3s_hv_rmhandlers.S |  2 +-
 8 files changed, 13 insertions(+), 11 deletions(-)

diff --git a/arch/powerpc/include/asm/ptrace.h b/arch/powerpc/include/asm/ptrace.h
index a4ae67aa9b76..8a9f4cf8c4c5 100644
--- a/arch/powerpc/include/asm/ptrace.h
+++ b/arch/powerpc/include/asm/ptrace.h
@@ -125,6 +125,7 @@ struct pt_regs
 #define STACK_INT_FRAME_SIZE	(sizeof(struct pt_regs) + \
 				 STACK_FRAME_OVERHEAD + KERNEL_REDZONE_SIZE)
 #define STACK_INT_FRAME_REGS	STACK_FRAME_OVERHEAD
+#define STACK_INT_FRAME_MARKER	(STACK_FRAME_OVERHEAD - 16)
 #define STACK_FRAME_MARKER	12
 
 #ifdef CONFIG_PPC64_ELF_ABI_V2
@@ -145,6 +146,7 @@ struct pt_regs
 #define STACK_FRAME_LR_SAVE	1	/* Location of LR in stack frame */
 #define STACK_INT_FRAME_SIZE	(sizeof(struct pt_regs) + STACK_FRAME_OVERHEAD)
 #define STACK_INT_FRAME_REGS	STACK_FRAME_OVERHEAD
+#define STACK_INT_FRAME_MARKER	(STACK_FRAME_OVERHEAD - 8)
 #define STACK_FRAME_MARKER	2
 #define STACK_FRAME_MIN_SIZE	STACK_FRAME_OVERHEAD
 
diff --git a/arch/powerpc/kernel/entry_32.S b/arch/powerpc/kernel/entry_32.S
index 24c8d84a56c9..2f61b7d3677c 100644
--- a/arch/powerpc/kernel/entry_32.S
+++ b/arch/powerpc/kernel/entry_32.S
@@ -114,7 +114,7 @@ transfer_to_syscall:
 	addi	r12,r12,STACK_FRAME_REGS_MARKER@l
 	stw	r9,_MSR(r1)
 	li	r2, INTERRUPT_SYSCALL
-	stw	r12,8(r1)
+	stw	r12,STACK_INT_FRAME_MARKER(r1)
 	stw	r2,_TRAP(r1)
 	SAVE_GPR(0, r1)
 	SAVE_GPRS(3, 8, r1)
diff --git a/arch/powerpc/kernel/exceptions-64e.S b/arch/powerpc/kernel/exceptions-64e.S
index f6b36083a8c2..d74c2a53af13 100644
--- a/arch/powerpc/kernel/exceptions-64e.S
+++ b/arch/powerpc/kernel/exceptions-64e.S
@@ -391,7 +391,7 @@ exc_##n##_common:							    \
 	std	r10,_CCR(r1);		/* store orig CR in stackframe */   \
 	std	r9,GPR1(r1);		/* store stack frame back link */   \
 	std	r11,SOFTE(r1);		/* and save it to stackframe */     \
-	std	r12,STACK_FRAME_OVERHEAD-16(r1); /* mark the frame */	    \
+	std	r12,STACK_INT_FRAME_MARKER(r1); /* mark the frame */	    \
 	std	r3,_TRAP(r1);		/* set trap number		*/  \
 	std	r0,RESULT(r1);		/* clear regs->result */	    \
 	SAVE_NVGPRS(r1);
diff --git a/arch/powerpc/kernel/exceptions-64s.S b/arch/powerpc/kernel/exceptions-64s.S
index 907054940a07..39f08ec56126 100644
--- a/arch/powerpc/kernel/exceptions-64s.S
+++ b/arch/powerpc/kernel/exceptions-64s.S
@@ -591,7 +591,7 @@ END_FTR_SECTION_IFSET(CPU_FTR_CFAR)
 	li	r10,0
 	LOAD_REG_IMMEDIATE(r11, STACK_FRAME_REGS_MARKER)
 	std	r10,RESULT(r1)		/* clear regs->result		*/
-	std	r11,STACK_FRAME_OVERHEAD-16(r1) /* mark the frame	*/
+	std	r11,STACK_INT_FRAME_MARKER(r1) /* mark the frame	*/
 .endm
 
 /*
diff --git a/arch/powerpc/kernel/head_32.h b/arch/powerpc/kernel/head_32.h
index 117d25330e13..f8e2911478a7 100644
--- a/arch/powerpc/kernel/head_32.h
+++ b/arch/powerpc/kernel/head_32.h
@@ -112,7 +112,7 @@ _ASM_NOKPROBE_SYMBOL(\name\()_virt)
 	stw	r0,GPR0(r1)
 	lis	r10,STACK_FRAME_REGS_MARKER@ha /* exception frame marker */
 	addi	r10,r10,STACK_FRAME_REGS_MARKER@l
-	stw	r10,8(r1)
+	stw	r10,STACK_INT_FRAME_MARKER(r1)
 	li	r10, \trapno
 	stw	r10,_TRAP(r1)
 	SAVE_GPRS(3, 8, r1)
diff --git a/arch/powerpc/kernel/head_booke.h b/arch/powerpc/kernel/head_booke.h
index 3149ac20b18e..37d43c172676 100644
--- a/arch/powerpc/kernel/head_booke.h
+++ b/arch/powerpc/kernel/head_booke.h
@@ -84,7 +84,7 @@ END_BTB_FLUSH_SECTION
 	stw	r0,GPR0(r1)
 	lis	r10, STACK_FRAME_REGS_MARKER@ha	/* exception frame marker */
 	addi	r10, r10, STACK_FRAME_REGS_MARKER@l
-	stw	r10, 8(r1)
+	stw	r10, STACK_INT_FRAME_MARKER(r1)
 	li	r10, \trapno
 	stw	r10,_TRAP(r1)
 	SAVE_GPRS(3, 8, r1)
diff --git a/arch/powerpc/kernel/interrupt_64.S b/arch/powerpc/kernel/interrupt_64.S
index 12807d5298a3..2306b979e71f 100644
--- a/arch/powerpc/kernel/interrupt_64.S
+++ b/arch/powerpc/kernel/interrupt_64.S
@@ -77,11 +77,11 @@ _ASM_NOKPROBE_SYMBOL(system_call_vectored_\name)
 	std	r11,_TRAP(r1)
 	std	r12,_CCR(r1)
 	std	r3,ORIG_GPR3(r1)
+	LOAD_REG_IMMEDIATE(r11, STACK_FRAME_REGS_MARKER)
+	std	r11,STACK_INT_FRAME_MARKER(r1)		/* "regs" marker */
 	/* Calling convention has r3 = regs, r4 = orig r0 */
 	addi	r3,r1,STACK_INT_FRAME_REGS
 	mr	r4,r0
-	LOAD_REG_IMMEDIATE(r11, STACK_FRAME_REGS_MARKER)
-	std	r11,-16(r3)		/* "regshere" marker */
 
 BEGIN_FTR_SECTION
 	HMT_MEDIUM
@@ -250,11 +250,11 @@ END_BTB_FLUSH_SECTION
 	std	r11,_TRAP(r1)
 	std	r12,_CCR(r1)
 	std	r3,ORIG_GPR3(r1)
+	LOAD_REG_IMMEDIATE(r11, STACK_FRAME_REGS_MARKER)
+	std	r11,STACK_INT_FRAME_MARKER(r1)		/* "regs" marker */
 	/* Calling convention has r3 = regs, r4 = orig r0 */
 	addi	r3,r1,STACK_INT_FRAME_REGS
 	mr	r4,r0
-	LOAD_REG_IMMEDIATE(r11, STACK_FRAME_REGS_MARKER)
-	std	r11,-16(r3)		/* "regshere" marker */
 
 #ifdef CONFIG_PPC_BOOK3S
 	li	r11,1
@@ -628,7 +628,7 @@ ALT_FTR_SECTION_END_IFCLR(CPU_FTR_STCX_CHECKS_ADDRESS)
 	 * Leaving a stale STACK_FRAME_REGS_MARKER on the stack can confuse
 	 * the reliable stack unwinder later on. Clear it.
 	 */
-	std	r0,STACK_FRAME_OVERHEAD-16(r1)
+	std	r0,STACK_INT_FRAME_MARKER(r1)
 
 	REST_GPRS(2, 5, r1)
 
diff --git a/arch/powerpc/kvm/book3s_hv_rmhandlers.S b/arch/powerpc/kvm/book3s_hv_rmhandlers.S
index 37f50861dd98..a9e162a1deec 100644
--- a/arch/powerpc/kvm/book3s_hv_rmhandlers.S
+++ b/arch/powerpc/kvm/book3s_hv_rmhandlers.S
@@ -2729,7 +2729,7 @@ kvmppc_bad_host_intr:
 	std	r6, SOFTE(r1)
 	LOAD_PACA_TOC()
 	LOAD_REG_IMMEDIATE(3, STACK_FRAME_REGS_MARKER)
-	std	r3, STACK_FRAME_OVERHEAD-16(r1)
+	std	r3, STACK_INT_FRAME_MARKER(r1)
 
 	/*
 	 * XXX On POWER7 and POWER8, we just spin here since we don't
-- 
2.37.2

[RFC PATCH 09/19] powerpc: Rename STACK_FRAME_MARKER and derive it from frame offset

[Nicholas Piggin]

This is a count of longs from the stack pointer to the regs marker.
Rename it to make it more distinct from the other byte offsets. It
can be derived from the byte offset definitions just added.

Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
---
 arch/powerpc/include/asm/ptrace.h | 4 ++--
 arch/powerpc/kernel/process.c     | 2 +-
 arch/powerpc/kernel/stacktrace.c  | 2 +-
 arch/powerpc/perf/callchain.c     | 2 +-
 arch/powerpc/xmon/xmon.c          | 3 +--
 5 files changed, 6 insertions(+), 7 deletions(-)

diff --git a/arch/powerpc/include/asm/ptrace.h b/arch/powerpc/include/asm/ptrace.h
index 8a9f4cf8c4c5..fdd50648df56 100644
--- a/arch/powerpc/include/asm/ptrace.h
+++ b/arch/powerpc/include/asm/ptrace.h
@@ -126,7 +126,6 @@ struct pt_regs
 				 STACK_FRAME_OVERHEAD + KERNEL_REDZONE_SIZE)
 #define STACK_INT_FRAME_REGS	STACK_FRAME_OVERHEAD
 #define STACK_INT_FRAME_MARKER	(STACK_FRAME_OVERHEAD - 16)
-#define STACK_FRAME_MARKER	12
 
 #ifdef CONFIG_PPC64_ELF_ABI_V2
 #define STACK_FRAME_MIN_SIZE	32
@@ -147,7 +146,6 @@ struct pt_regs
 #define STACK_INT_FRAME_SIZE	(sizeof(struct pt_regs) + STACK_FRAME_OVERHEAD)
 #define STACK_INT_FRAME_REGS	STACK_FRAME_OVERHEAD
 #define STACK_INT_FRAME_MARKER	(STACK_FRAME_OVERHEAD - 8)
-#define STACK_FRAME_MARKER	2
 #define STACK_FRAME_MIN_SIZE	STACK_FRAME_OVERHEAD
 
 /* Size of stack frame allocated when calling signal handler. */
@@ -155,6 +153,8 @@ struct pt_regs
 
 #endif /* __powerpc64__ */
 
+#define STACK_INT_FRAME_MARKER_LONGS	(STACK_INT_FRAME_MARKER/sizeof(long))
+
 #ifndef __ASSEMBLY__
 #include <asm/paca.h>
 
diff --git a/arch/powerpc/kernel/process.c b/arch/powerpc/kernel/process.c
index e7010f71de24..b0a9e5eeec4c 100644
--- a/arch/powerpc/kernel/process.c
+++ b/arch/powerpc/kernel/process.c
@@ -2234,7 +2234,7 @@ void __no_sanitize_address show_stack(struct task_struct *tsk,
 		 * We look for the "regs" marker in the current frame.
 		 */
 		if (validate_sp(sp, tsk, STACK_FRAME_WITH_PT_REGS)
-		    && stack[STACK_FRAME_MARKER] == STACK_FRAME_REGS_MARKER) {
+		    && stack[STACK_INT_FRAME_MARKER_LONGS] == STACK_FRAME_REGS_MARKER) {
 			struct pt_regs *regs = (struct pt_regs *)
 				(sp + STACK_INT_FRAME_REGS);
 
diff --git a/arch/powerpc/kernel/stacktrace.c b/arch/powerpc/kernel/stacktrace.c
index a2443d61728e..7efa0ec9dd77 100644
--- a/arch/powerpc/kernel/stacktrace.c
+++ b/arch/powerpc/kernel/stacktrace.c
@@ -136,7 +136,7 @@ int __no_sanitize_address arch_stack_walk_reliable(stack_trace_consume_fn consum
 
 		/* Mark stacktraces with exception frames as unreliable. */
 		if (sp <= stack_end - STACK_INT_FRAME_SIZE &&
-		    stack[STACK_FRAME_MARKER] == STACK_FRAME_REGS_MARKER) {
+		    stack[STACK_INT_FRAME_MARKER_LONGS] == STACK_FRAME_REGS_MARKER) {
 			return -EINVAL;
 		}
 
diff --git a/arch/powerpc/perf/callchain.c b/arch/powerpc/perf/callchain.c
index 9e254aed1f61..b01497ed5173 100644
--- a/arch/powerpc/perf/callchain.c
+++ b/arch/powerpc/perf/callchain.c
@@ -62,7 +62,7 @@ perf_callchain_kernel(struct perf_callchain_entry_ctx *entry, struct pt_regs *re
 
 		if (next_sp == sp + STACK_INT_FRAME_SIZE &&
 		    validate_sp(sp, current, STACK_INT_FRAME_SIZE) &&
-		    fp[STACK_FRAME_MARKER] == STACK_FRAME_REGS_MARKER) {
+		    fp[STACK_INT_FRAME_MARKER_LONGS] == STACK_FRAME_REGS_MARKER) {
 			/*
 			 * This looks like an interrupt frame for an
 			 * interrupt that occurred in the kernel
diff --git a/arch/powerpc/xmon/xmon.c b/arch/powerpc/xmon/xmon.c
index e403f14eb6eb..bbdaa42ba4ba 100644
--- a/arch/powerpc/xmon/xmon.c
+++ b/arch/powerpc/xmon/xmon.c
@@ -1720,7 +1720,6 @@ static void get_function_bounds(unsigned long pc, unsigned long *startp,
 }
 
 #define LRSAVE_OFFSET		(STACK_FRAME_LR_SAVE * sizeof(unsigned long))
-#define MARKER_OFFSET		(STACK_FRAME_MARKER * sizeof(unsigned long))
 
 static void xmon_show_stack(unsigned long sp, unsigned long lr,
 			    unsigned long pc)
@@ -1783,7 +1782,7 @@ static void xmon_show_stack(unsigned long sp, unsigned long lr,
 
 		/* Look for "regs" marker to see if this is
 		   an exception frame. */
-		if (mread(sp + MARKER_OFFSET, &marker, sizeof(unsigned long))
+		if (mread(sp + STACK_INT_FRAME_MARKER, &marker, sizeof(unsigned long))
 		    && marker == STACK_FRAME_REGS_MARKER) {
 			if (mread(sp + STACK_INT_FRAME_REGS, &regs, sizeof(regs)) != sizeof(regs)) {
 				printf("Couldn't read registers at %lx\n",
-- 
2.37.2

[RFC PATCH 10/19] powerpc: add a define for the user interrupt frame size

[Nicholas Piggin]

The user interrupt frame is a different size from the kernel frame, so
give it its own name.

Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
---
 arch/powerpc/include/asm/ptrace.h | 6 +++---
 arch/powerpc/kernel/process.c     | 6 +++---
 arch/powerpc/kernel/stacktrace.c  | 4 ++--
 3 files changed, 8 insertions(+), 8 deletions(-)

diff --git a/arch/powerpc/include/asm/ptrace.h b/arch/powerpc/include/asm/ptrace.h
index fdd50648df56..705ce26ae887 100644
--- a/arch/powerpc/include/asm/ptrace.h
+++ b/arch/powerpc/include/asm/ptrace.h
@@ -122,8 +122,7 @@ struct pt_regs
 
 #define STACK_FRAME_OVERHEAD	112	/* size of minimum stack frame */
 #define STACK_FRAME_LR_SAVE	2	/* Location of LR in stack frame */
-#define STACK_INT_FRAME_SIZE	(sizeof(struct pt_regs) + \
-				 STACK_FRAME_OVERHEAD + KERNEL_REDZONE_SIZE)
+#define STACK_USER_INT_FRAME_SIZE	(sizeof(struct pt_regs) + STACK_FRAME_OVERHEAD)
 #define STACK_INT_FRAME_REGS	STACK_FRAME_OVERHEAD
 #define STACK_INT_FRAME_MARKER	(STACK_FRAME_OVERHEAD - 16)
 
@@ -143,7 +142,7 @@ struct pt_regs
 #define KERNEL_REDZONE_SIZE	0
 #define STACK_FRAME_OVERHEAD	16	/* size of minimum stack frame */
 #define STACK_FRAME_LR_SAVE	1	/* Location of LR in stack frame */
-#define STACK_INT_FRAME_SIZE	(sizeof(struct pt_regs) + STACK_FRAME_OVERHEAD)
+#define STACK_USER_INT_FRAME_SIZE	(sizeof(struct pt_regs) + STACK_FRAME_OVERHEAD)
 #define STACK_INT_FRAME_REGS	STACK_FRAME_OVERHEAD
 #define STACK_INT_FRAME_MARKER	(STACK_FRAME_OVERHEAD - 8)
 #define STACK_FRAME_MIN_SIZE	STACK_FRAME_OVERHEAD
@@ -153,6 +152,7 @@ struct pt_regs
 
 #endif /* __powerpc64__ */
 
+#define STACK_INT_FRAME_SIZE	(KERNEL_REDZONE_SIZE + STACK_USER_INT_FRAME_SIZE)
 #define STACK_INT_FRAME_MARKER_LONGS	(STACK_INT_FRAME_MARKER/sizeof(long))
 
 #ifndef __ASSEMBLY__
diff --git a/arch/powerpc/kernel/process.c b/arch/powerpc/kernel/process.c
index b0a9e5eeec4c..d6daf0d073b3 100644
--- a/arch/powerpc/kernel/process.c
+++ b/arch/powerpc/kernel/process.c
@@ -1727,15 +1727,15 @@ int copy_thread(struct task_struct *p, const struct kernel_clone_args *args)
 	klp_init_thread_info(p);
 
 	/* Create initial stack frame. */
-	sp -= (sizeof(struct pt_regs) + STACK_FRAME_OVERHEAD);
+	sp -= STACK_USER_INT_FRAME_SIZE;
 	((unsigned long *)sp)[0] = 0;
 
 	/* Copy registers */
-	childregs = (struct pt_regs *)(sp + STACK_FRAME_OVERHEAD);
+	childregs = (struct pt_regs *)(sp + STACK_INT_FRAME_REGS);
 	if (unlikely(args->fn)) {
 		/* kernel thread */
 		memset(childregs, 0, sizeof(struct pt_regs));
-		childregs->gpr[1] = sp + (sizeof(struct pt_regs) + STACK_FRAME_OVERHEAD);
+		childregs->gpr[1] = sp + STACK_USER_INT_FRAME_SIZE;
 		/* function */
 		if (args->fn)
 			childregs->gpr[14] = ppc_function_entry((void *)args->fn);
diff --git a/arch/powerpc/kernel/stacktrace.c b/arch/powerpc/kernel/stacktrace.c
index 7efa0ec9dd77..453ac317a6cf 100644
--- a/arch/powerpc/kernel/stacktrace.c
+++ b/arch/powerpc/kernel/stacktrace.c
@@ -77,7 +77,7 @@ int __no_sanitize_address arch_stack_walk_reliable(stack_trace_consume_fn consum
 		/*
 		 * For user tasks, this is the SP value loaded on
 		 * kernel entry, see "PACAKSAVE(r13)" in _switch() and
-		 * system_call_common()/EXCEPTION_PROLOG_COMMON().
+		 * system_call_common().
 		 *
 		 * Likewise for non-swapper kernel threads,
 		 * this also happens to be the top of the stack
@@ -88,7 +88,7 @@ int __no_sanitize_address arch_stack_walk_reliable(stack_trace_consume_fn consum
 		 * an unreliable stack trace until it's been
 		 * _switch()'ed to for the first time.
 		 */
-		stack_end -= STACK_FRAME_OVERHEAD + sizeof(struct pt_regs);
+		stack_end -= STACK_USER_INT_FRAME_SIZE;
 	} else {
 		/*
 		 * idle tasks have a custom stack layout,
-- 
2.37.2

[RFC PATCH 11/19] powerpc: add a define for the switch frame size and regs offset

[Nicholas Piggin]

This is open-coded in process.c, ppc32 uses a different define with the
same value, and the C definition is name differently which makes it an
extra indirection to grep for.

Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
---
 arch/powerpc/include/asm/ptrace.h |  6 ++++--
 arch/powerpc/kernel/asm-offsets.c |  2 +-
 arch/powerpc/kernel/entry_32.S    |  6 +++---
 arch/powerpc/kernel/process.c     | 12 ++++++++----
 4 files changed, 16 insertions(+), 10 deletions(-)

diff --git a/arch/powerpc/include/asm/ptrace.h b/arch/powerpc/include/asm/ptrace.h
index 705ce26ae887..412ef0749775 100644
--- a/arch/powerpc/include/asm/ptrace.h
+++ b/arch/powerpc/include/asm/ptrace.h
@@ -97,8 +97,6 @@ struct pt_regs
 #endif
 
 
-#define STACK_FRAME_WITH_PT_REGS (STACK_FRAME_OVERHEAD + sizeof(struct pt_regs))
-
 // Always displays as "REGS" in memory dumps
 #ifdef CONFIG_CPU_BIG_ENDIAN
 #define STACK_FRAME_REGS_MARKER	ASM_CONST(0x52454753)
@@ -125,6 +123,8 @@ struct pt_regs
 #define STACK_USER_INT_FRAME_SIZE	(sizeof(struct pt_regs) + STACK_FRAME_OVERHEAD)
 #define STACK_INT_FRAME_REGS	STACK_FRAME_OVERHEAD
 #define STACK_INT_FRAME_MARKER	(STACK_FRAME_OVERHEAD - 16)
+#define STACK_SWITCH_FRAME_SIZE	(sizeof(struct pt_regs) + STACK_FRAME_OVERHEAD)
+#define STACK_SWITCH_FRAME_REGS	STACK_FRAME_OVERHEAD
 
 #ifdef CONFIG_PPC64_ELF_ABI_V2
 #define STACK_FRAME_MIN_SIZE	32
@@ -146,6 +146,8 @@ struct pt_regs
 #define STACK_INT_FRAME_REGS	STACK_FRAME_OVERHEAD
 #define STACK_INT_FRAME_MARKER	(STACK_FRAME_OVERHEAD - 8)
 #define STACK_FRAME_MIN_SIZE	STACK_FRAME_OVERHEAD
+#define STACK_SWITCH_FRAME_SIZE	(sizeof(struct pt_regs) + STACK_FRAME_OVERHEAD)
+#define STACK_SWITCH_FRAME_REGS	STACK_FRAME_OVERHEAD
 
 /* Size of stack frame allocated when calling signal handler. */
 #define __SIGNAL_FRAMESIZE	64
diff --git a/arch/powerpc/kernel/asm-offsets.c b/arch/powerpc/kernel/asm-offsets.c
index db5e66c1d031..f7dff906c24b 100644
--- a/arch/powerpc/kernel/asm-offsets.c
+++ b/arch/powerpc/kernel/asm-offsets.c
@@ -260,7 +260,7 @@ int main(void)
 
 	/* Interrupt register frame */
 	DEFINE(INT_FRAME_SIZE, STACK_INT_FRAME_SIZE);
-	DEFINE(SWITCH_FRAME_SIZE, STACK_FRAME_WITH_PT_REGS);
+	DEFINE(SWITCH_FRAME_SIZE, STACK_SWITCH_FRAME_SIZE);
 	STACK_PT_REGS_OFFSET(GPR0, gpr[0]);
 	STACK_PT_REGS_OFFSET(GPR1, gpr[1]);
 	STACK_PT_REGS_OFFSET(GPR2, gpr[2]);
diff --git a/arch/powerpc/kernel/entry_32.S b/arch/powerpc/kernel/entry_32.S
index 2f61b7d3677c..6e99ec10be89 100644
--- a/arch/powerpc/kernel/entry_32.S
+++ b/arch/powerpc/kernel/entry_32.S
@@ -215,9 +215,9 @@ ret_from_kernel_thread:
  * in arch/ppc/kernel/process.c
  */
 _GLOBAL(_switch)
-	stwu	r1,-INT_FRAME_SIZE(r1)
+	stwu	r1,-SWITCH_FRAME_SIZE(r1)
 	mflr	r0
-	stw	r0,INT_FRAME_SIZE+4(r1)
+	stw	r0,SWITCH_FRAME_SIZE+4(r1)
 	/* r3-r12 are caller saved -- Cort */
 	SAVE_NVGPRS(r1)
 	stw	r0,_NIP(r1)	/* Return to switch caller */
@@ -248,7 +248,7 @@ _GLOBAL(_switch)
 
 	lwz	r4,_NIP(r1)	/* Return to _switch caller in new task */
 	mtlr	r4
-	addi	r1,r1,INT_FRAME_SIZE
+	addi	r1,r1,SWITCH_FRAME_SIZE
 	blr
 
 	.globl	fast_exception_return
diff --git a/arch/powerpc/kernel/process.c b/arch/powerpc/kernel/process.c
index d6daf0d073b3..a097879b0474 100644
--- a/arch/powerpc/kernel/process.c
+++ b/arch/powerpc/kernel/process.c
@@ -1779,10 +1779,10 @@ int copy_thread(struct task_struct *p, const struct kernel_clone_args *args)
 	 * do some house keeping and then return from the fork or clone
 	 * system call, using the stack frame created above.
 	 */
-	sp -= sizeof(struct pt_regs);
-	kregs = (struct pt_regs *) sp;
-	sp -= STACK_FRAME_OVERHEAD;
+	sp -= STACK_SWITCH_FRAME_SIZE;
+	kregs = (struct pt_regs *)(sp + STACK_SWITCH_FRAME_REGS);
 	p->thread.ksp = sp;
+
 #ifdef CONFIG_HAVE_HW_BREAKPOINT
 	for (i = 0; i < nr_wp_slots(); i++)
 		p->thread.ptrace_bps[i] = NULL;
@@ -2232,8 +2232,12 @@ void __no_sanitize_address show_stack(struct task_struct *tsk,
 		/*
 		 * See if this is an exception frame.
 		 * We look for the "regs" marker in the current frame.
+		 *
+		 * STACK_SWITCH_FRAME_SIZE being the smallest frame that
+		 * could hold a pt_regs, if that does not fit then it can't
+		 * have regs.
 		 */
-		if (validate_sp(sp, tsk, STACK_FRAME_WITH_PT_REGS)
+		if (validate_sp(sp, tsk, STACK_SWITCH_FRAME_SIZE)
 		    && stack[STACK_INT_FRAME_MARKER_LONGS] == STACK_FRAME_REGS_MARKER) {
 			struct pt_regs *regs = (struct pt_regs *)
 				(sp + STACK_INT_FRAME_REGS);
-- 
2.37.2

[RFC PATCH 12/19] powerpc: copy_thread fill in interrupt frame marker and back chain

[Nicholas Piggin]

Backtraces will not recognise the fork system call interrupt without
the regs marker. And at least perf starts unwinding the user stack
from gpr[1] directly, but regular interrupt entry from userspace creates
the back chain to the user stack, so do this too, to be consistent.

Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
---
 arch/powerpc/kernel/process.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

diff --git a/arch/powerpc/kernel/process.c b/arch/powerpc/kernel/process.c
index a097879b0474..27956831fa5d 100644
--- a/arch/powerpc/kernel/process.c
+++ b/arch/powerpc/kernel/process.c
@@ -1728,12 +1728,13 @@ int copy_thread(struct task_struct *p, const struct kernel_clone_args *args)
 
 	/* Create initial stack frame. */
 	sp -= STACK_USER_INT_FRAME_SIZE;
-	((unsigned long *)sp)[0] = 0;
+	*(unsigned long *)(sp + STACK_INT_FRAME_MARKER) = STACK_FRAME_REGS_MARKER;
 
 	/* Copy registers */
 	childregs = (struct pt_regs *)(sp + STACK_INT_FRAME_REGS);
 	if (unlikely(args->fn)) {
 		/* kernel thread */
+		((unsigned long *)sp)[0] = 0;
 		memset(childregs, 0, sizeof(struct pt_regs));
 		childregs->gpr[1] = sp + STACK_USER_INT_FRAME_SIZE;
 		/* function */
@@ -1753,6 +1754,7 @@ int copy_thread(struct task_struct *p, const struct kernel_clone_args *args)
 		*childregs = *regs;
 		if (usp)
 			childregs->gpr[1] = usp;
+		((unsigned long *)sp)[0] = childregs->gpr[1];
 		p->thread.regs = childregs;
 		/* 64s sets this in ret_from_fork */
 		if (!IS_ENABLED(CONFIG_PPC_BOOK3S_64))
-- 
2.37.2

[RFC PATCH 13/19] powerpc: copy_thread add a back chain to the switch stack frame

[Nicholas Piggin]

Stack unwinders need LR and the back chain as a minimum. The switch
stack uses regs->nip for its return pointer rather than lrsave, so
that was not filled in, and neither was the back chain.

With this and the previous change, a stack trace in the switch or
interrupt stack goes from looking like this:

  Oops: Exception in kernel mode, sig: 5 [#1]
  LE PAGE_SIZE=64K MMU=Radix SMP NR_CPUS=2048 NUMA pSeries
  Modules linked in:
  CPU: 3 PID: 90 Comm: systemd Not tainted
  NIP:  c000000000011060 LR: c000000000010f68 CTR: 0000000000007fff
  [ ... regs ... ]
  NIP [c000000000011060] _switch+0x160/0x17c
  LR [c000000000010f68] _switch+0x68/0x17c
  Call Trace:

To this:

  Oops: Exception in kernel mode, sig: 5 [#1]
  LE PAGE_SIZE=64K MMU=Radix SMP NR_CPUS=2048 NUMA pSeries
  CPU: 0 PID: 93 Comm: systemd Not tainted
  NIP:  c000000000011060 LR: c000000000010f68 CTR: 0000000000007fff
  [ ... regs ... ]
  NIP [c000000000011060] _switch+0x160/0x17c
  LR [c000000000010f68] _switch+0x68/0x17c
  Call Trace:
  [c000000005a93e10] [c00000000000cdbc] ret_from_fork_scv+0x0/0x54
  --- interrupt: 3000 at 0x7fffa72f56d8
  NIP:  00007fffa72f56d8 LR: 0000000000000000 CTR: 0000000000000000
  [ ... regs ... ]
  NIP [00007fffa72f56d8] 0x7fffa72f56d8
  LR [0000000000000000] 0x0
  --- interrupt: 3000

Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
---
 arch/powerpc/kernel/process.c | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/arch/powerpc/kernel/process.c b/arch/powerpc/kernel/process.c
index 27956831fa5d..6cb3982a11ef 100644
--- a/arch/powerpc/kernel/process.c
+++ b/arch/powerpc/kernel/process.c
@@ -1781,7 +1781,9 @@ int copy_thread(struct task_struct *p, const struct kernel_clone_args *args)
 	 * do some house keeping and then return from the fork or clone
 	 * system call, using the stack frame created above.
 	 */
+	((unsigned long *)sp)[STACK_FRAME_LR_SAVE] = (unsigned long)f;
 	sp -= STACK_SWITCH_FRAME_SIZE;
+	((unsigned long *)sp)[0] = sp + STACK_SWITCH_FRAME_SIZE;
 	kregs = (struct pt_regs *)(sp + STACK_SWITCH_FRAME_REGS);
 	p->thread.ksp = sp;
 
-- 
2.37.2

[RFC PATCH 14/19] powerpc: split validate_sp into two functions

[Nicholas Piggin]

Most callers just want to validate an arbitrary kernel stack pointer,
some need a particular size. Make the size case the exceptional one
with an extra function.

Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
---
 arch/powerpc/include/asm/processor.h | 15 ++++++++++++---
 arch/powerpc/kernel/process.c        | 23 ++++++++++++++---------
 arch/powerpc/kernel/stacktrace.c     |  2 +-
 arch/powerpc/perf/callchain.c        |  6 +++---
 4 files changed, 30 insertions(+), 16 deletions(-)

diff --git a/arch/powerpc/include/asm/processor.h b/arch/powerpc/include/asm/processor.h
index 631802999d59..e96c9b8c2a60 100644
--- a/arch/powerpc/include/asm/processor.h
+++ b/arch/powerpc/include/asm/processor.h
@@ -374,9 +374,18 @@ static inline unsigned long __pack_fe01(unsigned int fpmode)
 
 #endif
 
-/* Check that a certain kernel stack pointer is valid in task_struct p */
-int validate_sp(unsigned long sp, struct task_struct *p,
-                       unsigned long nbytes);
+/*
+ * Check that a certain kernel stack pointer is a valid (minimum sized)
+ * stack frame in task_struct p.
+ */
+int validate_sp(unsigned long sp, struct task_struct *p);
+
+/*
+ * validate the stack frame of a particular minimum size, used for when we are
+ * looking at a certain object in the stack beyond the minimum.
+ */
+int validate_sp_size(unsigned long sp, struct task_struct *p,
+		     unsigned long nbytes);
 
 /*
  * Prefetch macros.
diff --git a/arch/powerpc/kernel/process.c b/arch/powerpc/kernel/process.c
index 6cb3982a11ef..b5defea32e75 100644
--- a/arch/powerpc/kernel/process.c
+++ b/arch/powerpc/kernel/process.c
@@ -2128,9 +2128,12 @@ static inline int valid_emergency_stack(unsigned long sp, struct task_struct *p,
 	return 0;
 }
 
-
-int validate_sp(unsigned long sp, struct task_struct *p,
-		       unsigned long nbytes)
+/*
+ * validate the stack frame of a particular minimum size, used for when we are
+ * looking at a certain object in the stack beyond the minimum.
+ */
+int validate_sp_size(unsigned long sp, struct task_struct *p,
+		     unsigned long nbytes)
 {
 	unsigned long stack_page = (unsigned long)task_stack_page(p);
 
@@ -2146,7 +2149,10 @@ int validate_sp(unsigned long sp, struct task_struct *p,
 	return valid_emergency_stack(sp, p, nbytes);
 }
 
-EXPORT_SYMBOL(validate_sp);
+int validate_sp(unsigned long sp, struct task_struct *p)
+{
+	return validate_sp(sp, p, STACK_FRAME_OVERHEAD);
+}
 
 static unsigned long ___get_wchan(struct task_struct *p)
 {
@@ -2154,13 +2160,12 @@ static unsigned long ___get_wchan(struct task_struct *p)
 	int count = 0;
 
 	sp = p->thread.ksp;
-	if (!validate_sp(sp, p, STACK_FRAME_OVERHEAD))
+	if (!validate_sp(sp, p))
 		return 0;
 
 	do {
 		sp = READ_ONCE_NOCHECK(*(unsigned long *)sp);
-		if (!validate_sp(sp, p, STACK_FRAME_OVERHEAD) ||
-		    task_is_running(p))
+		if (!validate_sp(sp, p) || task_is_running(p))
 			return 0;
 		if (count > 0) {
 			ip = READ_ONCE_NOCHECK(((unsigned long *)sp)[STACK_FRAME_LR_SAVE]);
@@ -2214,7 +2219,7 @@ void __no_sanitize_address show_stack(struct task_struct *tsk,
 	lr = 0;
 	printk("%sCall Trace:\n", loglvl);
 	do {
-		if (!validate_sp(sp, tsk, STACK_FRAME_OVERHEAD))
+		if (!validate_sp(sp, tsk))
 			break;
 
 		stack = (unsigned long *) sp;
@@ -2241,7 +2246,7 @@ void __no_sanitize_address show_stack(struct task_struct *tsk,
 		 * could hold a pt_regs, if that does not fit then it can't
 		 * have regs.
 		 */
-		if (validate_sp(sp, tsk, STACK_SWITCH_FRAME_SIZE)
+		if (validate_sp_size(sp, tsk, STACK_SWITCH_FRAME_SIZE)
 		    && stack[STACK_INT_FRAME_MARKER_LONGS] == STACK_FRAME_REGS_MARKER) {
 			struct pt_regs *regs = (struct pt_regs *)
 				(sp + STACK_INT_FRAME_REGS);
diff --git a/arch/powerpc/kernel/stacktrace.c b/arch/powerpc/kernel/stacktrace.c
index 453ac317a6cf..1dbbf30f265e 100644
--- a/arch/powerpc/kernel/stacktrace.c
+++ b/arch/powerpc/kernel/stacktrace.c
@@ -43,7 +43,7 @@ void __no_sanitize_address arch_stack_walk(stack_trace_consume_fn consume_entry,
 		unsigned long *stack = (unsigned long *) sp;
 		unsigned long newsp, ip;
 
-		if (!validate_sp(sp, task, STACK_FRAME_OVERHEAD))
+		if (!validate_sp(sp, task))
 			return;
 
 		newsp = stack[0];
diff --git a/arch/powerpc/perf/callchain.c b/arch/powerpc/perf/callchain.c
index b01497ed5173..6b4434dd0ff3 100644
--- a/arch/powerpc/perf/callchain.c
+++ b/arch/powerpc/perf/callchain.c
@@ -27,7 +27,7 @@ static int valid_next_sp(unsigned long sp, unsigned long prev_sp)
 {
 	if (sp & 0xf)
 		return 0;		/* must be 16-byte aligned */
-	if (!validate_sp(sp, current, STACK_FRAME_OVERHEAD))
+	if (!validate_sp(sp, current))
 		return 0;
 	if (sp >= prev_sp + STACK_FRAME_MIN_SIZE)
 		return 1;
@@ -53,7 +53,7 @@ perf_callchain_kernel(struct perf_callchain_entry_ctx *entry, struct pt_regs *re
 	sp = regs->gpr[1];
 	perf_callchain_store(entry, perf_instruction_pointer(regs));
 
-	if (!validate_sp(sp, current, STACK_FRAME_OVERHEAD))
+	if (!validate_sp(sp, current))
 		return;
 
 	for (;;) {
@@ -61,7 +61,7 @@ perf_callchain_kernel(struct perf_callchain_entry_ctx *entry, struct pt_regs *re
 		next_sp = fp[0];
 
 		if (next_sp == sp + STACK_INT_FRAME_SIZE &&
-		    validate_sp(sp, current, STACK_INT_FRAME_SIZE) &&
+		    validate_sp_size(sp, current, STACK_INT_FRAME_SIZE) &&
 		    fp[STACK_INT_FRAME_MARKER_LONGS] == STACK_FRAME_REGS_MARKER) {
 			/*
 			 * This looks like an interrupt frame for an
-- 
2.37.2

Re: [RFC PATCH 14/19] powerpc: split validate_sp into two functions

[Russell Currey]

On Mon, 2022-10-31 at 15:54 +1000, Nicholas Piggin wrote:
> Most callers just want to validate an arbitrary kernel stack pointer,
> some need a particular size. Make the size case the exceptional one
> with an extra function.
> 
> Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
> ---
>  arch/powerpc/include/asm/processor.h | 15 ++++++++++++---
>  arch/powerpc/kernel/process.c        | 23 ++++++++++++++---------
>  arch/powerpc/kernel/stacktrace.c     |  2 +-
>  arch/powerpc/perf/callchain.c        |  6 +++---
>  4 files changed, 30 insertions(+), 16 deletions(-)
> 
> diff --git a/arch/powerpc/include/asm/processor.h
> b/arch/powerpc/include/asm/processor.h
> index 631802999d59..e96c9b8c2a60 100644
> --- a/arch/powerpc/include/asm/processor.h
> +++ b/arch/powerpc/include/asm/processor.h
> @@ -374,9 +374,18 @@ static inline unsigned long __pack_fe01(unsigned
> int fpmode)
>  
>  #endif
>  
> -/* Check that a certain kernel stack pointer is valid in task_struct
> p */
> -int validate_sp(unsigned long sp, struct task_struct *p,
> -                       unsigned long nbytes);
> +/*
> + * Check that a certain kernel stack pointer is a valid (minimum
> sized)
> + * stack frame in task_struct p.
> + */
> +int validate_sp(unsigned long sp, struct task_struct *p);
> +
> +/*
> + * validate the stack frame of a particular minimum size, used for
> when we are
> + * looking at a certain object in the stack beyond the minimum.
> + */
> +int validate_sp_size(unsigned long sp, struct task_struct *p,
> +                    unsigned long nbytes);
>  
>  /*
>   * Prefetch macros.
> diff --git a/arch/powerpc/kernel/process.c
> b/arch/powerpc/kernel/process.c
> index 6cb3982a11ef..b5defea32e75 100644
> --- a/arch/powerpc/kernel/process.c
> +++ b/arch/powerpc/kernel/process.c
> @@ -2128,9 +2128,12 @@ static inline int
> valid_emergency_stack(unsigned long sp, struct task_struct *p,
>         return 0;
>  }
>  
> -
> -int validate_sp(unsigned long sp, struct task_struct *p,
> -                      unsigned long nbytes)
> +/*
> + * validate the stack frame of a particular minimum size, used for
> when we are
> + * looking at a certain object in the stack beyond the minimum.
> + */
> +int validate_sp_size(unsigned long sp, struct task_struct *p,
> +                    unsigned long nbytes)
>  {
>         unsigned long stack_page = (unsigned long)task_stack_page(p);
>  
> @@ -2146,7 +2149,10 @@ int validate_sp(unsigned long sp, struct
> task_struct *p,
>         return valid_emergency_stack(sp, p, nbytes);
>  }
>  
> -EXPORT_SYMBOL(validate_sp);
> +int validate_sp(unsigned long sp, struct task_struct *p)
> +{
> +       return validate_sp(sp, p, STACK_FRAME_OVERHEAD);

Hi Nick, I assume this supposed to be validate_sp_size()?  Did you get
this to compile?

> +}
>  
>  static unsigned long ___get_wchan(struct task_struct *p)
>  {
> @@ -2154,13 +2160,12 @@ static unsigned long ___get_wchan(struct
> task_struct *p)
>         int count = 0;
>  
>         sp = p->thread.ksp;
> -       if (!validate_sp(sp, p, STACK_FRAME_OVERHEAD))
> +       if (!validate_sp(sp, p))
>                 return 0;
>  
>         do {
>                 sp = READ_ONCE_NOCHECK(*(unsigned long *)sp);
> -               if (!validate_sp(sp, p, STACK_FRAME_OVERHEAD) ||
> -                   task_is_running(p))
> +               if (!validate_sp(sp, p) || task_is_running(p))
>                         return 0;
>                 if (count > 0) {
>                         ip = READ_ONCE_NOCHECK(((unsigned long
> *)sp)[STACK_FRAME_LR_SAVE]);
> @@ -2214,7 +2219,7 @@ void __no_sanitize_address show_stack(struct
> task_struct *tsk,
>         lr = 0;
>         printk("%sCall Trace:\n", loglvl);
>         do {
> -               if (!validate_sp(sp, tsk, STACK_FRAME_OVERHEAD))
> +               if (!validate_sp(sp, tsk))
>                         break;
>  
>                 stack = (unsigned long *) sp;
> @@ -2241,7 +2246,7 @@ void __no_sanitize_address show_stack(struct
> task_struct *tsk,
>                  * could hold a pt_regs, if that does not fit then it
> can't
>                  * have regs.
>                  */
> -               if (validate_sp(sp, tsk, STACK_SWITCH_FRAME_SIZE)
> +               if (validate_sp_size(sp, tsk,
> STACK_SWITCH_FRAME_SIZE)
>                     && stack[STACK_INT_FRAME_MARKER_LONGS] ==
> STACK_FRAME_REGS_MARKER) {
>                         struct pt_regs *regs = (struct pt_regs *)
>                                 (sp + STACK_INT_FRAME_REGS);
> diff --git a/arch/powerpc/kernel/stacktrace.c
> b/arch/powerpc/kernel/stacktrace.c
> index 453ac317a6cf..1dbbf30f265e 100644
> --- a/arch/powerpc/kernel/stacktrace.c
> +++ b/arch/powerpc/kernel/stacktrace.c
> @@ -43,7 +43,7 @@ void __no_sanitize_address
> arch_stack_walk(stack_trace_consume_fn consume_entry,
>                 unsigned long *stack = (unsigned long *) sp;
>                 unsigned long newsp, ip;
>  
> -               if (!validate_sp(sp, task, STACK_FRAME_OVERHEAD))
> +               if (!validate_sp(sp, task))
>                         return;
>  
>                 newsp = stack[0];
> diff --git a/arch/powerpc/perf/callchain.c
> b/arch/powerpc/perf/callchain.c
> index b01497ed5173..6b4434dd0ff3 100644
> --- a/arch/powerpc/perf/callchain.c
> +++ b/arch/powerpc/perf/callchain.c
> @@ -27,7 +27,7 @@ static int valid_next_sp(unsigned long sp, unsigned
> long prev_sp)
>  {
>         if (sp & 0xf)
>                 return 0;               /* must be 16-byte aligned */
> -       if (!validate_sp(sp, current, STACK_FRAME_OVERHEAD))
> +       if (!validate_sp(sp, current))
>                 return 0;
>         if (sp >= prev_sp + STACK_FRAME_MIN_SIZE)
>                 return 1;
> @@ -53,7 +53,7 @@ perf_callchain_kernel(struct
> perf_callchain_entry_ctx *entry, struct pt_regs *re
>         sp = regs->gpr[1];
>         perf_callchain_store(entry, perf_instruction_pointer(regs));
>  
> -       if (!validate_sp(sp, current, STACK_FRAME_OVERHEAD))
> +       if (!validate_sp(sp, current))
>                 return;
>  
>         for (;;) {
> @@ -61,7 +61,7 @@ perf_callchain_kernel(struct
> perf_callchain_entry_ctx *entry, struct pt_regs *re
>                 next_sp = fp[0];
>  
>                 if (next_sp == sp + STACK_INT_FRAME_SIZE &&
> -                   validate_sp(sp, current, STACK_INT_FRAME_SIZE) &&
> +                   validate_sp_size(sp, current,
> STACK_INT_FRAME_SIZE) &&
>                     fp[STACK_INT_FRAME_MARKER_LONGS] ==
> STACK_FRAME_REGS_MARKER) {
>                         /*
>                          * This looks like an interrupt frame for an

Re: [RFC PATCH 14/19] powerpc: split validate_sp into two functions

[Nicholas Piggin]

On Mon Nov 7, 2022 at 10:58 AM AEST, Russell Currey wrote:
> On Mon, 2022-10-31 at 15:54 +1000, Nicholas Piggin wrote:
> > Most callers just want to validate an arbitrary kernel stack pointer,
> > some need a particular size. Make the size case the exceptional one
> > with an extra function.
> > 
> > Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
> > ---
> >  arch/powerpc/include/asm/processor.h | 15 ++++++++++++---
> >  arch/powerpc/kernel/process.c        | 23 ++++++++++++++---------
> >  arch/powerpc/kernel/stacktrace.c     |  2 +-
> >  arch/powerpc/perf/callchain.c        |  6 +++---
> >  4 files changed, 30 insertions(+), 16 deletions(-)
> > 
> > diff --git a/arch/powerpc/include/asm/processor.h
> > b/arch/powerpc/include/asm/processor.h
> > index 631802999d59..e96c9b8c2a60 100644
> > --- a/arch/powerpc/include/asm/processor.h
> > +++ b/arch/powerpc/include/asm/processor.h
> > @@ -374,9 +374,18 @@ static inline unsigned long __pack_fe01(unsigned
> > int fpmode)
> >  
> >  #endif
> >  
> > -/* Check that a certain kernel stack pointer is valid in task_struct
> > p */
> > -int validate_sp(unsigned long sp, struct task_struct *p,
> > -                       unsigned long nbytes);
> > +/*
> > + * Check that a certain kernel stack pointer is a valid (minimum
> > sized)
> > + * stack frame in task_struct p.
> > + */
> > +int validate_sp(unsigned long sp, struct task_struct *p);
> > +
> > +/*
> > + * validate the stack frame of a particular minimum size, used for
> > when we are
> > + * looking at a certain object in the stack beyond the minimum.
> > + */
> > +int validate_sp_size(unsigned long sp, struct task_struct *p,
> > +                    unsigned long nbytes);
> >  
> >  /*
> >   * Prefetch macros.
> > diff --git a/arch/powerpc/kernel/process.c
> > b/arch/powerpc/kernel/process.c
> > index 6cb3982a11ef..b5defea32e75 100644
> > --- a/arch/powerpc/kernel/process.c
> > +++ b/arch/powerpc/kernel/process.c
> > @@ -2128,9 +2128,12 @@ static inline int
> > valid_emergency_stack(unsigned long sp, struct task_struct *p,
> >         return 0;
> >  }
> >  
> > -
> > -int validate_sp(unsigned long sp, struct task_struct *p,
> > -                      unsigned long nbytes)
> > +/*
> > + * validate the stack frame of a particular minimum size, used for
> > when we are
> > + * looking at a certain object in the stack beyond the minimum.
> > + */
> > +int validate_sp_size(unsigned long sp, struct task_struct *p,
> > +                    unsigned long nbytes)
> >  {
> >         unsigned long stack_page = (unsigned long)task_stack_page(p);
> >  
> > @@ -2146,7 +2149,10 @@ int validate_sp(unsigned long sp, struct
> > task_struct *p,
> >         return valid_emergency_stack(sp, p, nbytes);
> >  }
> >  
> > -EXPORT_SYMBOL(validate_sp);
> > +int validate_sp(unsigned long sp, struct task_struct *p)
> > +{
> > +       return validate_sp(sp, p, STACK_FRAME_OVERHEAD);
>
> Hi Nick, I assume this supposed to be validate_sp_size()?  Did you get
> this to compile?

Oops, yeah I think I sent a slightly stale version of the series. I
did fix that one.

Thanks,
Nick

[RFC PATCH 15/19] powerpc: allow minimum sized kernel stack frames

[Nicholas Piggin]

This affects only 64-bit ELFv2 kernels, and reduces the minimum
asm-created stack frame size from 112 to 32 byte on those kernels.

Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
---
 arch/powerpc/kernel/head_40x.S       | 2 +-
 arch/powerpc/kernel/head_44x.S       | 6 +++---
 arch/powerpc/kernel/head_64.S        | 6 +++---
 arch/powerpc/kernel/head_85xx.S      | 4 ++--
 arch/powerpc/kernel/head_8xx.S       | 2 +-
 arch/powerpc/kernel/head_book3s_32.S | 4 ++--
 arch/powerpc/kernel/irq.c            | 4 ++--
 arch/powerpc/kernel/misc_32.S        | 2 +-
 arch/powerpc/kernel/misc_64.S        | 4 ++--
 arch/powerpc/kernel/process.c        | 2 +-
 arch/powerpc/kernel/smp.c            | 2 +-
 arch/powerpc/kernel/stacktrace.c     | 2 +-
 12 files changed, 20 insertions(+), 20 deletions(-)

diff --git a/arch/powerpc/kernel/head_40x.S b/arch/powerpc/kernel/head_40x.S
index 088f500896c7..918547b93b5e 100644
--- a/arch/powerpc/kernel/head_40x.S
+++ b/arch/powerpc/kernel/head_40x.S
@@ -602,7 +602,7 @@ start_here:
 	lis	r1,init_thread_union@ha
 	addi	r1,r1,init_thread_union@l
 	li	r0,0
-	stwu	r0,THREAD_SIZE-STACK_FRAME_OVERHEAD(r1)
+	stwu	r0,THREAD_SIZE-STACK_FRAME_MIN_SIZE(r1)
 
 	bl	early_init	/* We have to do this with MMU on */
 
diff --git a/arch/powerpc/kernel/head_44x.S b/arch/powerpc/kernel/head_44x.S
index f15cb9fdb692..63a85c16fef4 100644
--- a/arch/powerpc/kernel/head_44x.S
+++ b/arch/powerpc/kernel/head_44x.S
@@ -109,7 +109,7 @@ _GLOBAL(_start);
 	lis	r1,init_thread_union@h
 	ori	r1,r1,init_thread_union@l
 	li	r0,0
-	stwu	r0,THREAD_SIZE-STACK_FRAME_OVERHEAD(r1)
+	stwu	r0,THREAD_SIZE-STACK_FRAME_MIN_SIZE(r1)
 
 	bl	early_init
 
@@ -1012,7 +1012,7 @@ _GLOBAL(start_secondary_47x)
 	 */
 	lis	r1,temp_boot_stack@h
 	ori	r1,r1,temp_boot_stack@l
-	addi	r1,r1,1024-STACK_FRAME_OVERHEAD
+	addi	r1,r1,1024-STACK_FRAME_MIN_SIZE
 	li	r0,0
 	stw	r0,0(r1)
 	bl	mmu_init_secondary
@@ -1025,7 +1025,7 @@ _GLOBAL(start_secondary_47x)
 	lwz	r1,TASK_STACK(r2)
 
 	/* Current stack pointer */
-	addi	r1,r1,THREAD_SIZE-STACK_FRAME_OVERHEAD
+	addi	r1,r1,THREAD_SIZE-STACK_FRAME_MIN_SIZE
 	li	r0,0
 	stw	r0,0(r1)
 
diff --git a/arch/powerpc/kernel/head_64.S b/arch/powerpc/kernel/head_64.S
index dedcc6fe2263..b513d13bf79e 100644
--- a/arch/powerpc/kernel/head_64.S
+++ b/arch/powerpc/kernel/head_64.S
@@ -424,7 +424,7 @@ generic_secondary_common_init:
 
 	/* Create a temp kernel stack for use before relocation is on.	*/
 	ld	r1,PACAEMERGSP(r13)
-	subi	r1,r1,STACK_FRAME_OVERHEAD
+	subi	r1,r1,STACK_FRAME_MIN_SIZE
 
 	/* See if we need to call a cpu state restore handler */
 	LOAD_REG_ADDR(r23, cur_cpu_spec)
@@ -780,7 +780,7 @@ _GLOBAL(pmac_secondary_start)
 
 	/* Create a temp kernel stack for use before relocation is on.	*/
 	ld	r1,PACAEMERGSP(r13)
-	subi	r1,r1,STACK_FRAME_OVERHEAD
+	subi	r1,r1,STACK_FRAME_MIN_SIZE
 
 	b	__secondary_start
 
@@ -958,7 +958,7 @@ start_here_multiplatform:
 	LOAD_REG_IMMEDIATE(r1,THREAD_SIZE)
 	add	r1,r3,r1
 	li	r0,0
-	stdu	r0,-STACK_FRAME_OVERHEAD(r1)
+	stdu	r0,-STACK_FRAME_MIN_SIZE(r1)
 
 	/*
 	 * Do very early kernel initializations, including initial hash table
diff --git a/arch/powerpc/kernel/head_85xx.S b/arch/powerpc/kernel/head_85xx.S
index 24f39abf81df..d9bd377dec91 100644
--- a/arch/powerpc/kernel/head_85xx.S
+++ b/arch/powerpc/kernel/head_85xx.S
@@ -229,7 +229,7 @@ set_ivor:
 	lis	r1,init_thread_union@h
 	ori	r1,r1,init_thread_union@l
 	li	r0,0
-	stwu	r0,THREAD_SIZE-STACK_FRAME_OVERHEAD(r1)
+	stwu	r0,THREAD_SIZE-STACK_FRAME_MIN_SIZE(r1)
 
 #ifdef CONFIG_SMP
 	stw	r24, TASK_CPU(r2)
@@ -1044,7 +1044,7 @@ __secondary_start:
 	lwz	r1,TASK_STACK(r2)
 
 	/* stack */
-	addi	r1,r1,THREAD_SIZE-STACK_FRAME_OVERHEAD
+	addi	r1,r1,THREAD_SIZE-STACK_FRAME_MIN_SIZE
 	li	r0,0
 	stw	r0,0(r1)
 
diff --git a/arch/powerpc/kernel/head_8xx.S b/arch/powerpc/kernel/head_8xx.S
index 0b05f2be66b9..cf546d0e5c40 100644
--- a/arch/powerpc/kernel/head_8xx.S
+++ b/arch/powerpc/kernel/head_8xx.S
@@ -537,7 +537,7 @@ start_here:
 	ori	r0, r0, STACK_END_MAGIC@l
 	stw	r0, 0(r1)
 	li	r0,0
-	stwu	r0,THREAD_SIZE-STACK_FRAME_OVERHEAD(r1)
+	stwu	r0,THREAD_SIZE-STACK_FRAME_MIN_SIZE(r1)
 
 	lis	r6, swapper_pg_dir@ha
 	tophys(r6,r6)
diff --git a/arch/powerpc/kernel/head_book3s_32.S b/arch/powerpc/kernel/head_book3s_32.S
index 519b60695167..40854d092dd3 100644
--- a/arch/powerpc/kernel/head_book3s_32.S
+++ b/arch/powerpc/kernel/head_book3s_32.S
@@ -840,7 +840,7 @@ __secondary_start:
 	lwz	r1,TASK_STACK(r1)
 
 	/* stack */
-	addi	r1,r1,THREAD_SIZE-STACK_FRAME_OVERHEAD
+	addi	r1,r1,THREAD_SIZE-STACK_FRAME_MIN_SIZE
 	li	r0,0
 	tophys(r3,r1)
 	stw	r0,0(r3)
@@ -966,7 +966,7 @@ END_MMU_FTR_SECTION_IFCLR(MMU_FTR_HPTE_TABLE)
 	lis	r1,init_thread_union@ha
 	addi	r1,r1,init_thread_union@l
 	li	r0,0
-	stwu	r0,THREAD_SIZE-STACK_FRAME_OVERHEAD(r1)
+	stwu	r0,THREAD_SIZE-STACK_FRAME_MIN_SIZE(r1)
 /*
  * Do early platform-specific initialization,
  * and set up the MMU.
diff --git a/arch/powerpc/kernel/irq.c b/arch/powerpc/kernel/irq.c
index 9ede61a5a469..c5b9ce887483 100644
--- a/arch/powerpc/kernel/irq.c
+++ b/arch/powerpc/kernel/irq.c
@@ -210,7 +210,7 @@ static __always_inline void call_do_softirq(const void *sp)
 		 PPC_LL "	%%r1, 0(%%r1)		;"
 		 : // Outputs
 		 : // Inputs
-		   [sp] "b" (sp), [offset] "i" (THREAD_SIZE - STACK_FRAME_OVERHEAD),
+		   [sp] "b" (sp), [offset] "i" (THREAD_SIZE - STACK_FRAME_MIN_SIZE),
 		   [callee] "i" (__do_softirq)
 		 : // Clobbers
 		   "lr", "xer", "ctr", "memory", "cr0", "cr1", "cr5", "cr6",
@@ -264,7 +264,7 @@ static __always_inline void call_do_irq(struct pt_regs *regs, void *sp)
 		 : // Outputs
 		   "+r" (r3)
 		 : // Inputs
-		   [sp] "b" (sp), [offset] "i" (THREAD_SIZE - STACK_FRAME_OVERHEAD),
+		   [sp] "b" (sp), [offset] "i" (THREAD_SIZE - STACK_FRAME_MIN_SIZE),
 		   [callee] "i" (__do_irq)
 		 : // Clobbers
 		   "lr", "xer", "ctr", "memory", "cr0", "cr1", "cr5", "cr6",
diff --git a/arch/powerpc/kernel/misc_32.S b/arch/powerpc/kernel/misc_32.S
index e5127b19fec2..daf8f87d2372 100644
--- a/arch/powerpc/kernel/misc_32.S
+++ b/arch/powerpc/kernel/misc_32.S
@@ -382,7 +382,7 @@ EXPORT_SYMBOL(__bswapdi2)
 _GLOBAL(start_secondary_resume)
 	/* Reset stack */
 	rlwinm	r1, r1, 0, 0, 31 - THREAD_SHIFT
-	addi	r1,r1,THREAD_SIZE-STACK_FRAME_OVERHEAD
+	addi	r1,r1,THREAD_SIZE-STACK_FRAME_MIN_SIZE
 	li	r3,0
 	stw	r3,0(r1)		/* Zero the stack frame pointer	*/
 	bl	start_secondary
diff --git a/arch/powerpc/kernel/misc_64.S b/arch/powerpc/kernel/misc_64.S
index 36184cada00b..4bb6dd30c556 100644
--- a/arch/powerpc/kernel/misc_64.S
+++ b/arch/powerpc/kernel/misc_64.S
@@ -384,7 +384,7 @@ _GLOBAL(kexec_sequence)
 	std	r0,16(r1)
 
 	/* switch stacks to newstack -- &kexec_stack.stack */
-	stdu	r1,THREAD_SIZE-STACK_FRAME_OVERHEAD(r3)
+	stdu	r1,THREAD_SIZE-STACK_FRAME_MIN_SIZE(r3)
 	mr	r1,r3
 
 	li	r0,0
@@ -401,7 +401,7 @@ _GLOBAL(kexec_sequence)
 	std	r26,-48(r1)
 	std	r25,-56(r1)
 
-	stdu	r1,-STACK_FRAME_OVERHEAD-64(r1)
+	stdu	r1,-STACK_FRAME_MIN_SIZE-64(r1)
 
 	/* save args into preserved regs */
 	mr	r31,r3			/* newstack (both) */
diff --git a/arch/powerpc/kernel/process.c b/arch/powerpc/kernel/process.c
index b5defea32e75..20fdb84759f0 100644
--- a/arch/powerpc/kernel/process.c
+++ b/arch/powerpc/kernel/process.c
@@ -2151,7 +2151,7 @@ int validate_sp_size(unsigned long sp, struct task_struct *p,
 
 int validate_sp(unsigned long sp, struct task_struct *p)
 {
-	return validate_sp(sp, p, STACK_FRAME_OVERHEAD);
+	return validate_sp(sp, p, STACK_FRAME_MIN_SIZE);
 }
 
 static unsigned long ___get_wchan(struct task_struct *p)
diff --git a/arch/powerpc/kernel/smp.c b/arch/powerpc/kernel/smp.c
index 0da6e59161cd..6b90f10a6c81 100644
--- a/arch/powerpc/kernel/smp.c
+++ b/arch/powerpc/kernel/smp.c
@@ -1249,7 +1249,7 @@ static void cpu_idle_thread_init(unsigned int cpu, struct task_struct *idle)
 #ifdef CONFIG_PPC64
 	paca_ptrs[cpu]->__current = idle;
 	paca_ptrs[cpu]->kstack = (unsigned long)task_stack_page(idle) +
-				 THREAD_SIZE - STACK_FRAME_OVERHEAD;
+				 THREAD_SIZE - STACK_FRAME_MIN_SIZE;
 #endif
 	task_thread_info(idle)->cpu = cpu;
 	secondary_current = current_set[cpu] = idle;
diff --git a/arch/powerpc/kernel/stacktrace.c b/arch/powerpc/kernel/stacktrace.c
index 1dbbf30f265e..5de8597eaab8 100644
--- a/arch/powerpc/kernel/stacktrace.c
+++ b/arch/powerpc/kernel/stacktrace.c
@@ -94,7 +94,7 @@ int __no_sanitize_address arch_stack_walk_reliable(stack_trace_consume_fn consum
 		 * idle tasks have a custom stack layout,
 		 * c.f. cpu_idle_thread_init().
 		 */
-		stack_end -= STACK_FRAME_OVERHEAD;
+		stack_end -= STACK_FRAME_MIN_SIZE;
 	}
 
 	if (task == current)
-- 
2.37.2

[RFC PATCH 16/19] powerpc/64: ELFv2 use minimal stack frames in int and switch frame sizes

[Nicholas Piggin]

This requires an extra 16 bytes beyond the minimum frame size for the
aligned regs marker for the int frame. The switch frame needs to match
that because they share some offset definitions.

Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
---
 arch/powerpc/include/asm/ptrace.h | 21 +++++++++++++++------
 1 file changed, 15 insertions(+), 6 deletions(-)

diff --git a/arch/powerpc/include/asm/ptrace.h b/arch/powerpc/include/asm/ptrace.h
index 412ef0749775..a9dfce62a5eb 100644
--- a/arch/powerpc/include/asm/ptrace.h
+++ b/arch/powerpc/include/asm/ptrace.h
@@ -96,7 +96,6 @@ struct pt_regs
 };
 #endif
 
-
 // Always displays as "REGS" in memory dumps
 #ifdef CONFIG_CPU_BIG_ENDIAN
 #define STACK_FRAME_REGS_MARKER	ASM_CONST(0x52454753)
@@ -120,16 +119,26 @@ struct pt_regs
 
 #define STACK_FRAME_OVERHEAD	112	/* size of minimum stack frame */
 #define STACK_FRAME_LR_SAVE	2	/* Location of LR in stack frame */
+
+#ifdef CONFIG_PPC64_ELF_ABI_V2
+#define STACK_FRAME_MIN_SIZE	32
+#define STACK_USER_INT_FRAME_SIZE	(sizeof(struct pt_regs) + STACK_FRAME_MIN_SIZE + 16)
+#define STACK_INT_FRAME_REGS	(STACK_FRAME_MIN_SIZE + 16)
+#define STACK_INT_FRAME_MARKER	STACK_FRAME_MIN_SIZE
+#define STACK_SWITCH_FRAME_SIZE (sizeof(struct pt_regs) + STACK_FRAME_MIN_SIZE + 16)
+#define STACK_SWITCH_FRAME_REGS	(STACK_FRAME_MIN_SIZE + 16)
+#else
+/*
+ * The ELFv1 ABI specifies 48 bytes plus a minimum 64 byte parameter save
+ * area. This parameter area is not used by calls to C from interrupt entry,
+ * so the second from last one of those is used for the frame marker.
+ */
+#define STACK_FRAME_MIN_SIZE	112
 #define STACK_USER_INT_FRAME_SIZE	(sizeof(struct pt_regs) + STACK_FRAME_OVERHEAD)
 #define STACK_INT_FRAME_REGS	STACK_FRAME_OVERHEAD
 #define STACK_INT_FRAME_MARKER	(STACK_FRAME_OVERHEAD - 16)
 #define STACK_SWITCH_FRAME_SIZE	(sizeof(struct pt_regs) + STACK_FRAME_OVERHEAD)
 #define STACK_SWITCH_FRAME_REGS	STACK_FRAME_OVERHEAD
-
-#ifdef CONFIG_PPC64_ELF_ABI_V2
-#define STACK_FRAME_MIN_SIZE	32
-#else
-#define STACK_FRAME_MIN_SIZE	STACK_FRAME_OVERHEAD
 #endif
 
 /* Size of dummy stack frame allocated when calling signal handler. */
-- 
2.37.2

[RFC PATCH 17/19] powerpc: remove STACK_FRAME_OVERHEAD

[Nicholas Piggin]

This is equal to STACK_FRAME_MIN_SIZE on 32-bit and 64-bit ELFv1, and no
longer used in 64-bit ELFv2, so replace it with that constant.

Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
---
 arch/powerpc/include/asm/ptrace.h | 24 +++++++++++-------------
 1 file changed, 11 insertions(+), 13 deletions(-)

diff --git a/arch/powerpc/include/asm/ptrace.h b/arch/powerpc/include/asm/ptrace.h
index a9dfce62a5eb..a53c580388e2 100644
--- a/arch/powerpc/include/asm/ptrace.h
+++ b/arch/powerpc/include/asm/ptrace.h
@@ -117,7 +117,6 @@ struct pt_regs
 #define USER_REDZONE_SIZE	512
 #define KERNEL_REDZONE_SIZE	288
 
-#define STACK_FRAME_OVERHEAD	112	/* size of minimum stack frame */
 #define STACK_FRAME_LR_SAVE	2	/* Location of LR in stack frame */
 
 #ifdef CONFIG_PPC64_ELF_ABI_V2
@@ -134,11 +133,11 @@ struct pt_regs
  * so the second from last one of those is used for the frame marker.
  */
 #define STACK_FRAME_MIN_SIZE	112
-#define STACK_USER_INT_FRAME_SIZE	(sizeof(struct pt_regs) + STACK_FRAME_OVERHEAD)
-#define STACK_INT_FRAME_REGS	STACK_FRAME_OVERHEAD
-#define STACK_INT_FRAME_MARKER	(STACK_FRAME_OVERHEAD - 16)
-#define STACK_SWITCH_FRAME_SIZE	(sizeof(struct pt_regs) + STACK_FRAME_OVERHEAD)
-#define STACK_SWITCH_FRAME_REGS	STACK_FRAME_OVERHEAD
+#define STACK_USER_INT_FRAME_SIZE	(sizeof(struct pt_regs) + STACK_FRAME_MIN_SIZE)
+#define STACK_INT_FRAME_REGS	STACK_FRAME_MIN_SIZE
+#define STACK_INT_FRAME_MARKER	(STACK_FRAME_MIN_SIZE - 16)
+#define STACK_SWITCH_FRAME_SIZE	(sizeof(struct pt_regs) + STACK_FRAME_MIN_SIZE)
+#define STACK_SWITCH_FRAME_REGS	STACK_FRAME_MIN_SIZE
 #endif
 
 /* Size of dummy stack frame allocated when calling signal handler. */
@@ -149,14 +148,13 @@ struct pt_regs
 
 #define USER_REDZONE_SIZE	0
 #define KERNEL_REDZONE_SIZE	0
-#define STACK_FRAME_OVERHEAD	16	/* size of minimum stack frame */
+#define STACK_FRAME_MIN_SIZE	16
 #define STACK_FRAME_LR_SAVE	1	/* Location of LR in stack frame */
-#define STACK_USER_INT_FRAME_SIZE	(sizeof(struct pt_regs) + STACK_FRAME_OVERHEAD)
-#define STACK_INT_FRAME_REGS	STACK_FRAME_OVERHEAD
-#define STACK_INT_FRAME_MARKER	(STACK_FRAME_OVERHEAD - 8)
-#define STACK_FRAME_MIN_SIZE	STACK_FRAME_OVERHEAD
-#define STACK_SWITCH_FRAME_SIZE	(sizeof(struct pt_regs) + STACK_FRAME_OVERHEAD)
-#define STACK_SWITCH_FRAME_REGS	STACK_FRAME_OVERHEAD
+#define STACK_USER_INT_FRAME_SIZE	(sizeof(struct pt_regs) + STACK_FRAME_MIN_SIZE)
+#define STACK_INT_FRAME_REGS	STACK_FRAME_MIN_SIZE
+#define STACK_INT_FRAME_MARKER	(STACK_FRAME_MIN_SIZE - 8)
+#define STACK_SWITCH_FRAME_SIZE	(sizeof(struct pt_regs) + STACK_FRAME_MIN_SIZE)
+#define STACK_SWITCH_FRAME_REGS	STACK_FRAME_MIN_SIZE
 
 /* Size of stack frame allocated when calling signal handler. */
 #define __SIGNAL_FRAMESIZE	64
-- 
2.37.2

[RFC PATCH 18/19] powerpc: change stack marker memory operations to 32-bit

[Nicholas Piggin]

The marker is a 32-bit constant across all platforms now, so use 32-bit
memory accesses.

Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
---
 arch/powerpc/include/asm/ptrace.h    | 1 -
 arch/powerpc/kernel/exceptions-64e.S | 2 +-
 arch/powerpc/kernel/exceptions-64s.S | 2 +-
 arch/powerpc/kernel/interrupt_64.S   | 6 +++---
 arch/powerpc/kernel/process.c        | 6 +++---
 arch/powerpc/kernel/stacktrace.c     | 2 +-
 arch/powerpc/perf/callchain.c        | 2 +-
 7 files changed, 10 insertions(+), 11 deletions(-)

diff --git a/arch/powerpc/include/asm/ptrace.h b/arch/powerpc/include/asm/ptrace.h
index a53c580388e2..ab41d47761c9 100644
--- a/arch/powerpc/include/asm/ptrace.h
+++ b/arch/powerpc/include/asm/ptrace.h
@@ -162,7 +162,6 @@ struct pt_regs
 #endif /* __powerpc64__ */
 
 #define STACK_INT_FRAME_SIZE	(KERNEL_REDZONE_SIZE + STACK_USER_INT_FRAME_SIZE)
-#define STACK_INT_FRAME_MARKER_LONGS	(STACK_INT_FRAME_MARKER/sizeof(long))
 
 #ifndef __ASSEMBLY__
 #include <asm/paca.h>
diff --git a/arch/powerpc/kernel/exceptions-64e.S b/arch/powerpc/kernel/exceptions-64e.S
index d74c2a53af13..86cf656c6777 100644
--- a/arch/powerpc/kernel/exceptions-64e.S
+++ b/arch/powerpc/kernel/exceptions-64e.S
@@ -391,7 +391,7 @@ exc_##n##_common:							    \
 	std	r10,_CCR(r1);		/* store orig CR in stackframe */   \
 	std	r9,GPR1(r1);		/* store stack frame back link */   \
 	std	r11,SOFTE(r1);		/* and save it to stackframe */     \
-	std	r12,STACK_INT_FRAME_MARKER(r1); /* mark the frame */	    \
+	stw	r12,STACK_INT_FRAME_MARKER(r1); /* mark the frame */	    \
 	std	r3,_TRAP(r1);		/* set trap number		*/  \
 	std	r0,RESULT(r1);		/* clear regs->result */	    \
 	SAVE_NVGPRS(r1);
diff --git a/arch/powerpc/kernel/exceptions-64s.S b/arch/powerpc/kernel/exceptions-64s.S
index 39f08ec56126..85bbdfa7e4d3 100644
--- a/arch/powerpc/kernel/exceptions-64s.S
+++ b/arch/powerpc/kernel/exceptions-64s.S
@@ -591,7 +591,7 @@ END_FTR_SECTION_IFSET(CPU_FTR_CFAR)
 	li	r10,0
 	LOAD_REG_IMMEDIATE(r11, STACK_FRAME_REGS_MARKER)
 	std	r10,RESULT(r1)		/* clear regs->result		*/
-	std	r11,STACK_INT_FRAME_MARKER(r1) /* mark the frame	*/
+	stw	r11,STACK_INT_FRAME_MARKER(r1) /* mark the frame	*/
 .endm
 
 /*
diff --git a/arch/powerpc/kernel/interrupt_64.S b/arch/powerpc/kernel/interrupt_64.S
index 2306b979e71f..ee8cfe17ca37 100644
--- a/arch/powerpc/kernel/interrupt_64.S
+++ b/arch/powerpc/kernel/interrupt_64.S
@@ -78,7 +78,7 @@ _ASM_NOKPROBE_SYMBOL(system_call_vectored_\name)
 	std	r12,_CCR(r1)
 	std	r3,ORIG_GPR3(r1)
 	LOAD_REG_IMMEDIATE(r11, STACK_FRAME_REGS_MARKER)
-	std	r11,STACK_INT_FRAME_MARKER(r1)		/* "regs" marker */
+	stw	r11,STACK_INT_FRAME_MARKER(r1)		/* "regs" marker */
 	/* Calling convention has r3 = regs, r4 = orig r0 */
 	addi	r3,r1,STACK_INT_FRAME_REGS
 	mr	r4,r0
@@ -251,7 +251,7 @@ END_BTB_FLUSH_SECTION
 	std	r12,_CCR(r1)
 	std	r3,ORIG_GPR3(r1)
 	LOAD_REG_IMMEDIATE(r11, STACK_FRAME_REGS_MARKER)
-	std	r11,STACK_INT_FRAME_MARKER(r1)		/* "regs" marker */
+	stw	r11,STACK_INT_FRAME_MARKER(r1)		/* "regs" marker */
 	/* Calling convention has r3 = regs, r4 = orig r0 */
 	addi	r3,r1,STACK_INT_FRAME_REGS
 	mr	r4,r0
@@ -628,7 +628,7 @@ ALT_FTR_SECTION_END_IFCLR(CPU_FTR_STCX_CHECKS_ADDRESS)
 	 * Leaving a stale STACK_FRAME_REGS_MARKER on the stack can confuse
 	 * the reliable stack unwinder later on. Clear it.
 	 */
-	std	r0,STACK_INT_FRAME_MARKER(r1)
+	stw	r0,STACK_INT_FRAME_MARKER(r1)
 
 	REST_GPRS(2, 5, r1)
 
diff --git a/arch/powerpc/kernel/process.c b/arch/powerpc/kernel/process.c
index 20fdb84759f0..e447bf02f996 100644
--- a/arch/powerpc/kernel/process.c
+++ b/arch/powerpc/kernel/process.c
@@ -1728,7 +1728,7 @@ int copy_thread(struct task_struct *p, const struct kernel_clone_args *args)
 
 	/* Create initial stack frame. */
 	sp -= STACK_USER_INT_FRAME_SIZE;
-	*(unsigned long *)(sp + STACK_INT_FRAME_MARKER) = STACK_FRAME_REGS_MARKER;
+	*(u32 *)(sp + STACK_INT_FRAME_MARKER) = STACK_FRAME_REGS_MARKER;
 
 	/* Copy registers */
 	childregs = (struct pt_regs *)(sp + STACK_INT_FRAME_REGS);
@@ -2246,8 +2246,8 @@ void __no_sanitize_address show_stack(struct task_struct *tsk,
 		 * could hold a pt_regs, if that does not fit then it can't
 		 * have regs.
 		 */
-		if (validate_sp_size(sp, tsk, STACK_SWITCH_FRAME_SIZE)
-		    && stack[STACK_INT_FRAME_MARKER_LONGS] == STACK_FRAME_REGS_MARKER) {
+		if (validate_sp_size(sp, tsk, STACK_SWITCH_FRAME_SIZE) &&
+		    *(u32 *)(sp + STACK_INT_FRAME_MARKER) == STACK_FRAME_REGS_MARKER) {
 			struct pt_regs *regs = (struct pt_regs *)
 				(sp + STACK_INT_FRAME_REGS);
 
diff --git a/arch/powerpc/kernel/stacktrace.c b/arch/powerpc/kernel/stacktrace.c
index 5de8597eaab8..7d975943f5ec 100644
--- a/arch/powerpc/kernel/stacktrace.c
+++ b/arch/powerpc/kernel/stacktrace.c
@@ -136,7 +136,7 @@ int __no_sanitize_address arch_stack_walk_reliable(stack_trace_consume_fn consum
 
 		/* Mark stacktraces with exception frames as unreliable. */
 		if (sp <= stack_end - STACK_INT_FRAME_SIZE &&
-		    stack[STACK_INT_FRAME_MARKER_LONGS] == STACK_FRAME_REGS_MARKER) {
+		    *(u32 *)(sp + STACK_INT_FRAME_MARKER) == STACK_FRAME_REGS_MARKER) {
 			return -EINVAL;
 		}
 
diff --git a/arch/powerpc/perf/callchain.c b/arch/powerpc/perf/callchain.c
index 6b4434dd0ff3..72eef9accdbb 100644
--- a/arch/powerpc/perf/callchain.c
+++ b/arch/powerpc/perf/callchain.c
@@ -62,7 +62,7 @@ perf_callchain_kernel(struct perf_callchain_entry_ctx *entry, struct pt_regs *re
 
 		if (next_sp == sp + STACK_INT_FRAME_SIZE &&
 		    validate_sp_size(sp, current, STACK_INT_FRAME_SIZE) &&
-		    fp[STACK_INT_FRAME_MARKER_LONGS] == STACK_FRAME_REGS_MARKER) {
+		    *(u32 *)(sp + STACK_INT_FRAME_MARKER) == STACK_FRAME_REGS_MARKER) {
 			/*
 			 * This looks like an interrupt frame for an
 			 * interrupt that occurred in the kernel
-- 
2.37.2

[RFC PATCH 19/19] powerpc/64: ELFv2 use reserved word in the stack frame for the regs marker

[Nicholas Piggin]

There are 4 unused bytes in the minimum frame in the ELFv2 ABI. At the
risk of causing a future ABI incompatibility, use this and save 16 bytes
from interrupt and switch frames.

Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
---
 arch/powerpc/include/asm/ptrace.h | 15 ++++++++++-----
 1 file changed, 10 insertions(+), 5 deletions(-)

diff --git a/arch/powerpc/include/asm/ptrace.h b/arch/powerpc/include/asm/ptrace.h
index ab41d47761c9..cc308dce7ace 100644
--- a/arch/powerpc/include/asm/ptrace.h
+++ b/arch/powerpc/include/asm/ptrace.h
@@ -121,11 +121,16 @@ struct pt_regs
 
 #ifdef CONFIG_PPC64_ELF_ABI_V2
 #define STACK_FRAME_MIN_SIZE	32
-#define STACK_USER_INT_FRAME_SIZE	(sizeof(struct pt_regs) + STACK_FRAME_MIN_SIZE + 16)
-#define STACK_INT_FRAME_REGS	(STACK_FRAME_MIN_SIZE + 16)
-#define STACK_INT_FRAME_MARKER	STACK_FRAME_MIN_SIZE
-#define STACK_SWITCH_FRAME_SIZE (sizeof(struct pt_regs) + STACK_FRAME_MIN_SIZE + 16)
-#define STACK_SWITCH_FRAME_REGS	(STACK_FRAME_MIN_SIZE + 16)
+#define STACK_USER_INT_FRAME_SIZE	(sizeof(struct pt_regs) + STACK_FRAME_MIN_SIZE)
+#define STACK_INT_FRAME_REGS	STACK_FRAME_MIN_SIZE
+#define STACK_INT_FRAME_MARKER	12	/* Reserved word next to CR save word */
+
+/*
+ * The switch frame has to match the format of the int frames up to pt_regs,
+ * because fields are accessed with the int frame pt_regs offsets.
+ */
+#define STACK_SWITCH_FRAME_SIZE (sizeof(struct pt_regs) + STACK_FRAME_MIN_SIZE)
+#define STACK_SWITCH_FRAME_REGS	STACK_FRAME_MIN_SIZE
 #else
 /*
  * The ELFv1 ABI specifies 48 bytes plus a minimum 64 byte parameter save
-- 
2.37.2