selftests/powerpc: Move core_busy_loop() into asm

There is at least one bug in core_busy_loop(), we use r0, but it's
not in the clobber list. We were getting away with this it seems but
that was luck.

It's also fishy to be touching the stack, even if we do it below the
stack pointer. It seems we get away with it, but looking at the
generated code that may just be luck.

So move it into assembler, do all the stack handling by hand. We create
a stack frame to save the non-volatiles in, so we can muck around with
them.

Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>

1 files changed, 271 insertions, 0 deletions

diff --git a/tools/testing/selftests/powerpc/pmu/ebb/busy_loop.S b/tools/testing/selftests/powerpc/pmu/ebb/busy_loop.S
new file mode 100644
index 000000000000..c7e4093f1cd3
--- /dev/null
+++ b/tools/testing/selftests/powerpc/pmu/ebb/busy_loop.S
@@ -0,0 +1,271 @@
+/*
+ * Copyright 2014, Michael Ellerman, IBM Corp.
+ * Licensed under GPLv2.
+ */
+
+#include <ppc-asm.h>
+
+	.text
+
+FUNC_START(core_busy_loop)
+	stdu	%r1, -168(%r1)
+	std	r14, 160(%r1)
+	std	r15, 152(%r1)
+	std	r16, 144(%r1)
+	std	r17, 136(%r1)
+	std	r18, 128(%r1)
+	std	r19, 120(%r1)
+	std	r20, 112(%r1)
+	std	r21, 104(%r1)
+	std	r22, 96(%r1)
+	std	r23, 88(%r1)
+	std	r24, 80(%r1)
+	std	r25, 72(%r1)
+	std	r26, 64(%r1)
+	std	r27, 56(%r1)
+	std	r28, 48(%r1)
+	std	r29, 40(%r1)
+	std	r30, 32(%r1)
+	std	r31, 24(%r1)
+
+	li	 r3, 0x3030
+	std	 r3, -96(%r1)
+	li	 r4, 0x4040
+	std	 r4, -104(%r1)
+	li	 r5, 0x5050
+	std	 r5, -112(%r1)
+	li	 r6, 0x6060
+	std	 r6, -120(%r1)
+	li	 r7, 0x7070
+	std	 r7, -128(%r1)
+	li	 r8, 0x0808
+	std	 r8, -136(%r1)
+	li	 r9, 0x0909
+	std	 r9, -144(%r1)
+	li	r10, 0x1010
+	std	r10, -152(%r1)
+	li	r11, 0x1111
+	std	r11, -160(%r1)
+	li	r14, 0x1414
+	std	r14, -168(%r1)
+	li	r15, 0x1515
+	std	r15, -176(%r1)
+	li	r16, 0x1616
+	std	r16, -184(%r1)
+	li	r17, 0x1717
+	std	r17, -192(%r1)
+	li	r18, 0x1818
+	std	r18, -200(%r1)
+	li	r19, 0x1919
+	std	r19, -208(%r1)
+	li	r20, 0x2020
+	std	r20, -216(%r1)
+	li	r21, 0x2121
+	std	r21, -224(%r1)
+	li	r22, 0x2222
+	std	r22, -232(%r1)
+	li	r23, 0x2323
+	std	r23, -240(%r1)
+	li	r24, 0x2424
+	std	r24, -248(%r1)
+	li	r25, 0x2525
+	std	r25, -256(%r1)
+	li	r26, 0x2626
+	std	r26, -264(%r1)
+	li	r27, 0x2727
+	std	r27, -272(%r1)
+	li	r28, 0x2828
+	std	r28, -280(%r1)
+	li	r29, 0x2929
+	std	r29, -288(%r1)
+	li	r30, 0x3030
+	li	r31, 0x3131
+
+	li	r3, 0
+0:	addi	r3, r3, 1
+	cmpwi	r3, 100
+	blt	0b
+
+	/* Return 1 (fail) unless we get through all the checks */
+	li	r3, 1
+
+	/* Check none of our registers have been corrupted */
+	cmpwi	r4,  0x4040
+	bne	1f
+	cmpwi	r5,  0x5050
+	bne	1f
+	cmpwi	r6,  0x6060
+	bne	1f
+	cmpwi	r7,  0x7070
+	bne	1f
+	cmpwi	r8,  0x0808
+	bne	1f
+	cmpwi	r9,  0x0909
+	bne	1f
+	cmpwi	r10, 0x1010
+	bne	1f
+	cmpwi	r11, 0x1111
+	bne	1f
+	cmpwi	r14, 0x1414
+	bne	1f
+	cmpwi	r15, 0x1515
+	bne	1f
+	cmpwi	r16, 0x1616
+	bne	1f
+	cmpwi	r17, 0x1717
+	bne	1f
+	cmpwi	r18, 0x1818
+	bne	1f
+	cmpwi	r19, 0x1919
+	bne	1f
+	cmpwi	r20, 0x2020
+	bne	1f
+	cmpwi	r21, 0x2121
+	bne	1f
+	cmpwi	r22, 0x2222
+	bne	1f
+	cmpwi	r23, 0x2323
+	bne	1f
+	cmpwi	r24, 0x2424
+	bne	1f
+	cmpwi	r25, 0x2525
+	bne	1f
+	cmpwi	r26, 0x2626
+	bne	1f
+	cmpwi	r27, 0x2727
+	bne	1f
+	cmpwi	r28, 0x2828
+	bne	1f
+	cmpwi	r29, 0x2929
+	bne	1f
+	cmpwi	r30, 0x3030
+	bne	1f
+	cmpwi	r31, 0x3131
+	bne	1f
+
+	/* Load junk into all our registers before we reload them from the stack. */
+	li	r3,  0xde
+	li	r4,  0xad
+	li	r5,  0xbe
+	li	r6,  0xef
+	li	r7,  0xde
+	li	r8,  0xad
+	li	r9,  0xbe
+	li	r10, 0xef
+	li	r11, 0xde
+	li	r14, 0xad
+	li	r15, 0xbe
+	li	r16, 0xef
+	li	r17, 0xde
+	li	r18, 0xad
+	li	r19, 0xbe
+	li	r20, 0xef
+	li	r21, 0xde
+	li	r22, 0xad
+	li	r23, 0xbe
+	li	r24, 0xef
+	li	r25, 0xde
+	li	r26, 0xad
+	li	r27, 0xbe
+	li	r28, 0xef
+	li	r29, 0xdd
+
+	ld	r3,	-96(%r1)
+	cmpwi	r3,  0x3030
+	bne	1f
+	ld	r4,	-104(%r1)
+	cmpwi	r4,  0x4040
+	bne	1f
+	ld	r5,	-112(%r1)
+	cmpwi	r5,  0x5050
+	bne	1f
+	ld	r6,	-120(%r1)
+	cmpwi	r6,  0x6060
+	bne	1f
+	ld	r7,	-128(%r1)
+	cmpwi	r7,  0x7070
+	bne	1f
+	ld	r8,	-136(%r1)
+	cmpwi	r8,  0x0808
+	bne	1f
+	ld	r9,	-144(%r1)
+	cmpwi	r9,  0x0909
+	bne	1f
+	ld	r10, -152(%r1)
+	cmpwi	r10, 0x1010
+	bne	1f
+	ld	r11, -160(%r1)
+	cmpwi	r11, 0x1111
+	bne	1f
+	ld	r14, -168(%r1)
+	cmpwi	r14, 0x1414
+	bne	1f
+	ld	r15, -176(%r1)
+	cmpwi	r15, 0x1515
+	bne	1f
+	ld	r16, -184(%r1)
+	cmpwi	r16, 0x1616
+	bne	1f
+	ld	r17, -192(%r1)
+	cmpwi	r17, 0x1717
+	bne	1f
+	ld	r18, -200(%r1)
+	cmpwi	r18, 0x1818
+	bne	1f
+	ld	r19, -208(%r1)
+	cmpwi	r19, 0x1919
+	bne	1f
+	ld	r20, -216(%r1)
+	cmpwi	r20, 0x2020
+	bne	1f
+	ld	r21, -224(%r1)
+	cmpwi	r21, 0x2121
+	bne	1f
+	ld	r22, -232(%r1)
+	cmpwi	r22, 0x2222
+	bne	1f
+	ld	r23, -240(%r1)
+	cmpwi	r23, 0x2323
+	bne	1f
+	ld	r24, -248(%r1)
+	cmpwi	r24, 0x2424
+	bne	1f
+	ld	r25, -256(%r1)
+	cmpwi	r25, 0x2525
+	bne	1f
+	ld	r26, -264(%r1)
+	cmpwi	r26, 0x2626
+	bne	1f
+	ld	r27, -272(%r1)
+	cmpwi	r27, 0x2727
+	bne	1f
+	ld	r28, -280(%r1)
+	cmpwi	r28, 0x2828
+	bne	1f
+	ld	r29, -288(%r1)
+	cmpwi	r29, 0x2929
+	bne	1f
+
+	/* Load 0 (success) to return */
+	li	r3, 0
+
+1:	ld	r14, 160(%r1)
+	ld	r15, 152(%r1)
+	ld	r16, 144(%r1)
+	ld	r17, 136(%r1)
+	ld	r18, 128(%r1)
+	ld	r19, 120(%r1)
+	ld	r20, 112(%r1)
+	ld	r21, 104(%r1)
+	ld	r22, 96(%r1)
+	ld	r23, 88(%r1)
+	ld	r24, 80(%r1)
+	ld	r25, 72(%r1)
+	ld	r26, 64(%r1)
+	ld	r27, 56(%r1)
+	ld	r28, 48(%r1)
+	ld	r29, 40(%r1)
+	ld	r30, 32(%r1)
+	ld	r31, 24(%r1)
+	addi	%r1, %r1, 168
+	blr