[PATCH 01/13] selftests/powerpc: Ensure 16-byte stack pointer alignment

[PATCH 01/13] selftests/powerpc: Ensure 16-byte stack pointer alignment

[Michael Ellerman]

The PUSH/POP_BASIC_STACK helpers in basic_asm.h do not ensure that the
stack pointer is always 16-byte aligned, which is required per the ABI.

Fix the macros to do the alignment if the caller fails to.

Currently only one caller passes a non-aligned size, tm_signal_self(),
which hasn't been caught in testing, presumably because it's a leaf
function.

Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
---
 tools/testing/selftests/powerpc/include/basic_asm.h | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/tools/testing/selftests/powerpc/include/basic_asm.h b/tools/testing/selftests/powerpc/include/basic_asm.h
index 886dc026fe7a..807e83e821ec 100644
--- a/tools/testing/selftests/powerpc/include/basic_asm.h
+++ b/tools/testing/selftests/powerpc/include/basic_asm.h
@@ -58,7 +58,7 @@
 #define PUSH_BASIC_STACK(_extra) \
 	mflr	r0; \
 	std	r0, STACK_FRAME_LR_POS(%r1); \
-	stdu	%r1, -(_extra + STACK_FRAME_MIN_SIZE)(%r1); \
+	stdu	%r1, -(((_extra + 15) & ~15) + STACK_FRAME_MIN_SIZE)(%r1); \
 	mfcr	r0; \
 	stw	r0, STACK_FRAME_CR_POS(%r1); \
 	std	%r2, STACK_FRAME_TOC_POS(%r1);
@@ -67,7 +67,7 @@
 	ld	%r2, STACK_FRAME_TOC_POS(%r1); \
 	lwz	r0, STACK_FRAME_CR_POS(%r1); \
 	mtcr	r0; \
-	addi	%r1, %r1, (_extra + STACK_FRAME_MIN_SIZE); \
+	addi	%r1, %r1, (((_extra + 15) & ~15) + STACK_FRAME_MIN_SIZE); \
 	ld	r0, STACK_FRAME_LR_POS(%r1); \
 	mtlr	r0;
 
-- 
2.35.3

[PATCH 02/13] selftests/powerpc/ptrace: Set LOCAL_HDRS

[Michael Ellerman]

Set LOCAL_HDRS so header changes cause rebuilds. The lib.mk logic adds
all the headers in LOCAL_HDRS as dependencies, so there's no need to
also list them explicitly.

Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
---
 tools/testing/selftests/powerpc/ptrace/Makefile | 5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)

diff --git a/tools/testing/selftests/powerpc/ptrace/Makefile b/tools/testing/selftests/powerpc/ptrace/Makefile
index a500639da97a..0b0652d88b1b 100644
--- a/tools/testing/selftests/powerpc/ptrace/Makefile
+++ b/tools/testing/selftests/powerpc/ptrace/Makefile
@@ -4,12 +4,13 @@ TEST_GEN_PROGS := ptrace-gpr ptrace-tm-gpr ptrace-tm-spd-gpr \
               ptrace-tm-spd-vsx ptrace-tm-spr ptrace-hwbreak ptrace-pkey core-pkey \
               perf-hwbreak ptrace-syscall ptrace-perf-hwbreak
 
+LOCAL_HDRS += $(patsubst %,$(selfdir)/powerpc/ptrace/%,$(wildcard *.h))
+
 top_srcdir = ../../../../..
 include ../../lib.mk
 
 CFLAGS += -m64 -I../../../../../usr/include -I../tm -mhtm -fno-pie
 
-$(OUTPUT)/ptrace-pkey $(OUTPUT)/core-pkey: child.h
 $(OUTPUT)/ptrace-pkey $(OUTPUT)/core-pkey: LDLIBS += -pthread
 
-$(TEST_GEN_PROGS): ../harness.c ../utils.c ../lib/reg.S ptrace.h
+$(TEST_GEN_PROGS): ../harness.c ../utils.c ../lib/reg.S
-- 
2.35.3

[PATCH 03/13] selftests/powerpc/ptrace: Split CFLAGS better

[Michael Ellerman]

Currently all ptrace tests are built 64-bit and with TM enabled.

Only the TM tests need TM enabled, so split those out into a separate
variable so that can be specified precisely.

Split the rest of the tests into a variable, and add -m64 to CFLAGS for
those tests, so that in a subsequent patch some tests can be made to
build 32-bit.

Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
---
 .../testing/selftests/powerpc/ptrace/Makefile | 33 ++++++++++++++++---
 1 file changed, 28 insertions(+), 5 deletions(-)

diff --git a/tools/testing/selftests/powerpc/ptrace/Makefile b/tools/testing/selftests/powerpc/ptrace/Makefile
index 0b0652d88b1b..8611b0670587 100644
--- a/tools/testing/selftests/powerpc/ptrace/Makefile
+++ b/tools/testing/selftests/powerpc/ptrace/Makefile
@@ -1,15 +1,38 @@
 # SPDX-License-Identifier: GPL-2.0
-TEST_GEN_PROGS := ptrace-gpr ptrace-tm-gpr ptrace-tm-spd-gpr \
-              ptrace-tar ptrace-tm-tar ptrace-tm-spd-tar ptrace-vsx ptrace-tm-vsx \
-              ptrace-tm-spd-vsx ptrace-tm-spr ptrace-hwbreak ptrace-pkey core-pkey \
-              perf-hwbreak ptrace-syscall ptrace-perf-hwbreak
+
+TM_TESTS := ptrace-tm-gpr
+TM_TESTS += ptrace-tm-spd-gpr
+TM_TESTS += ptrace-tm-spd-tar
+TM_TESTS += ptrace-tm-spd-vsx
+TM_TESTS += ptrace-tm-spr
+TM_TESTS += ptrace-tm-tar
+TM_TESTS += ptrace-tm-vsx
+
+TESTS_64 := $(TM_TESTS)
+TESTS_64 += core-pkey
+TESTS_64 += perf-hwbreak
+TESTS_64 += ptrace-gpr
+TESTS_64 += ptrace-hwbreak
+TESTS_64 += ptrace-perf-hwbreak
+TESTS_64 += ptrace-pkey
+TESTS_64 += ptrace-syscall
+TESTS_64 += ptrace-tar
+TESTS_64 += ptrace-vsx
+
+TEST_GEN_PROGS := $(TESTS_64)
 
 LOCAL_HDRS += $(patsubst %,$(selfdir)/powerpc/ptrace/%,$(wildcard *.h))
 
 top_srcdir = ../../../../..
 include ../../lib.mk
 
-CFLAGS += -m64 -I../../../../../usr/include -I../tm -mhtm -fno-pie
+TM_TESTS := $(patsubst %,$(OUTPUT)/%,$(TM_TESTS))
+TESTS_64 := $(patsubst %,$(OUTPUT)/%,$(TESTS_64))
+
+$(TESTS_64): CFLAGS += -m64
+$(TM_TESTS): CFLAGS += -I../tm -mhtm
+
+CFLAGS += -I../../../../../usr/include -fno-pie
 
 $(OUTPUT)/ptrace-pkey $(OUTPUT)/core-pkey: LDLIBS += -pthread
 
-- 
2.35.3

[PATCH 04/13] selftests/powerpc: Don't save CR by default in asm helpers

[Michael Ellerman]

Thare are some asm helpers for creating/popping stack frames in
basic_asm.h. They always save/restore CR, but none of the selftests
tests touch non-volatile CR fields, so it's unnecessary to save them by
default.

Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
---
 tools/testing/selftests/powerpc/include/basic_asm.h | 4 ----
 1 file changed, 4 deletions(-)

diff --git a/tools/testing/selftests/powerpc/include/basic_asm.h b/tools/testing/selftests/powerpc/include/basic_asm.h
index 807e83e821ec..faec7b11c041 100644
--- a/tools/testing/selftests/powerpc/include/basic_asm.h
+++ b/tools/testing/selftests/powerpc/include/basic_asm.h
@@ -59,14 +59,10 @@
 	mflr	r0; \
 	std	r0, STACK_FRAME_LR_POS(%r1); \
 	stdu	%r1, -(((_extra + 15) & ~15) + STACK_FRAME_MIN_SIZE)(%r1); \
-	mfcr	r0; \
-	stw	r0, STACK_FRAME_CR_POS(%r1); \
 	std	%r2, STACK_FRAME_TOC_POS(%r1);
 
 #define POP_BASIC_STACK(_extra) \
 	ld	%r2, STACK_FRAME_TOC_POS(%r1); \
-	lwz	r0, STACK_FRAME_CR_POS(%r1); \
-	mtcr	r0; \
 	addi	%r1, %r1, (((_extra + 15) & ~15) + STACK_FRAME_MIN_SIZE); \
 	ld	r0, STACK_FRAME_LR_POS(%r1); \
 	mtlr	r0;
-- 
2.35.3

[PATCH 05/13] selftests/powerpc: Don't save TOC by default in asm helpers

[Michael Ellerman]

Thare are some asm helpers for creating/popping stack frames in
basic_asm.h. They always save/restore r2 (TOC pointer), but none of the
selftests change r2, so it's unnecessary to save it by default.

Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
---
 tools/testing/selftests/powerpc/include/basic_asm.h | 4 +---
 1 file changed, 1 insertion(+), 3 deletions(-)

diff --git a/tools/testing/selftests/powerpc/include/basic_asm.h b/tools/testing/selftests/powerpc/include/basic_asm.h
index faec7b11c041..be380aa779f8 100644
--- a/tools/testing/selftests/powerpc/include/basic_asm.h
+++ b/tools/testing/selftests/powerpc/include/basic_asm.h
@@ -58,11 +58,9 @@
 #define PUSH_BASIC_STACK(_extra) \
 	mflr	r0; \
 	std	r0, STACK_FRAME_LR_POS(%r1); \
-	stdu	%r1, -(((_extra + 15) & ~15) + STACK_FRAME_MIN_SIZE)(%r1); \
-	std	%r2, STACK_FRAME_TOC_POS(%r1);
+	stdu	%r1, -(((_extra + 15) & ~15) + STACK_FRAME_MIN_SIZE)(%r1);
 
 #define POP_BASIC_STACK(_extra) \
-	ld	%r2, STACK_FRAME_TOC_POS(%r1); \
 	addi	%r1, %r1, (((_extra + 15) & ~15) + STACK_FRAME_MIN_SIZE); \
 	ld	r0, STACK_FRAME_LR_POS(%r1); \
 	mtlr	r0;
-- 
2.35.3

[PATCH 06/13] selftests/powerpc: Add 32-bit support to asm helpers

[Michael Ellerman]

Add support for 32-bit builds to the asm helpers.

Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
---
 .../selftests/powerpc/include/basic_asm.h     | 47 +++++++++++++++----
 1 file changed, 38 insertions(+), 9 deletions(-)

diff --git a/tools/testing/selftests/powerpc/include/basic_asm.h b/tools/testing/selftests/powerpc/include/basic_asm.h
index be380aa779f8..2d7f6e592dd9 100644
--- a/tools/testing/selftests/powerpc/include/basic_asm.h
+++ b/tools/testing/selftests/powerpc/include/basic_asm.h
@@ -5,6 +5,16 @@
 #include <ppc-asm.h>
 #include <asm/unistd.h>
 
+#ifdef __powerpc64__
+#define PPC_LL		ld
+#define PPC_STL		std
+#define PPC_STLU	stdu
+#else
+#define PPC_LL		lwz
+#define PPC_STL		stw
+#define PPC_STLU	stwu
+#endif
+
 #define LOAD_REG_IMMEDIATE(reg, expr) \
 	lis	reg, (expr)@highest;	\
 	ori	reg, reg, (expr)@higher;	\
@@ -14,16 +24,20 @@
 
 /*
  * Note: These macros assume that variables being stored on the stack are
- * doublewords, while this is usually the case it may not always be the
+ * sizeof(long), while this is usually the case it may not always be the
  * case for each use case.
  */
+#ifdef  __powerpc64__
+
+// ABIv2
 #if defined(_CALL_ELF) && _CALL_ELF == 2
 #define STACK_FRAME_MIN_SIZE 32
 #define STACK_FRAME_TOC_POS  24
 #define __STACK_FRAME_PARAM(_param)  (32 + ((_param)*8))
 #define __STACK_FRAME_LOCAL(_num_params, _var_num)  \
 	((STACK_FRAME_PARAM(_num_params)) + ((_var_num)*8))
-#else
+
+#else // ABIv1 below
 #define STACK_FRAME_MIN_SIZE 112
 #define STACK_FRAME_TOC_POS  40
 #define __STACK_FRAME_PARAM(i)  (48 + ((i)*8))
@@ -34,7 +48,24 @@
  */
 #define __STACK_FRAME_LOCAL(_num_params, _var_num)  \
 	(112 + ((_var_num)*8))
-#endif
+
+
+#endif // ABIv2
+
+// Common 64-bit
+#define STACK_FRAME_LR_POS   16
+#define STACK_FRAME_CR_POS   8
+
+#else // 32-bit below
+
+#define STACK_FRAME_MIN_SIZE 16
+#define STACK_FRAME_LR_POS   4
+
+#define __STACK_FRAME_PARAM(_param)  (STACK_FRAME_MIN_SIZE + ((_param)*4))
+#define __STACK_FRAME_LOCAL(_num_params, _var_num)  \
+	((STACK_FRAME_PARAM(_num_params)) + ((_var_num)*4))
+
+#endif // __powerpc64__
 
 /* Parameter x saved to the stack */
 #define STACK_FRAME_PARAM(var)    __STACK_FRAME_PARAM(var)
@@ -42,8 +73,6 @@
 /* Local variable x saved to the stack after x parameters */
 #define STACK_FRAME_LOCAL(num_params, var)    \
 	__STACK_FRAME_LOCAL(num_params, var)
-#define STACK_FRAME_LR_POS   16
-#define STACK_FRAME_CR_POS   8
 
 /*
  * It is very important to note here that _extra is the extra amount of
@@ -56,13 +85,13 @@
  * preprocessed incorrectly, hence r0.
  */
 #define PUSH_BASIC_STACK(_extra) \
-	mflr	r0; \
-	std	r0, STACK_FRAME_LR_POS(%r1); \
-	stdu	%r1, -(((_extra + 15) & ~15) + STACK_FRAME_MIN_SIZE)(%r1);
+	mflr	 r0; \
+	PPC_STL	 r0, STACK_FRAME_LR_POS(%r1); \
+	PPC_STLU %r1, -(((_extra + 15) & ~15) + STACK_FRAME_MIN_SIZE)(%r1);
 
 #define POP_BASIC_STACK(_extra) \
 	addi	%r1, %r1, (((_extra + 15) & ~15) + STACK_FRAME_MIN_SIZE); \
-	ld	r0, STACK_FRAME_LR_POS(%r1); \
+	PPC_LL	r0, STACK_FRAME_LR_POS(%r1); \
 	mtlr	r0;
 
 #endif /* _SELFTESTS_POWERPC_BASIC_ASM_H */
-- 
2.35.3

[PATCH 07/13] selftests/powerpc/ptrace: Drop unused load_fpr_single_precision()

[Michael Ellerman]

This function is never called, drop it.

Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
---
 tools/testing/selftests/powerpc/include/reg.h |  1 -
 tools/testing/selftests/powerpc/lib/reg.S     | 37 -------------------
 2 files changed, 38 deletions(-)

diff --git a/tools/testing/selftests/powerpc/include/reg.h b/tools/testing/selftests/powerpc/include/reg.h
index c422be8a42b2..bc7f610c2c6d 100644
--- a/tools/testing/selftests/powerpc/include/reg.h
+++ b/tools/testing/selftests/powerpc/include/reg.h
@@ -160,7 +160,6 @@
 #ifndef __ASSEMBLER__
 void store_gpr(unsigned long *addr);
 void load_gpr(unsigned long *addr);
-void load_fpr_single_precision(float *addr);
 void store_fpr_single_precision(float *addr);
 #endif /* end of __ASSEMBLER__ */
 
diff --git a/tools/testing/selftests/powerpc/lib/reg.S b/tools/testing/selftests/powerpc/lib/reg.S
index 9304ea7d59b9..dd37b8e6f84c 100644
--- a/tools/testing/selftests/powerpc/lib/reg.S
+++ b/tools/testing/selftests/powerpc/lib/reg.S
@@ -53,43 +53,6 @@ FUNC_START(store_gpr)
 	blr
 FUNC_END(store_gpr)
 
-/* Single Precision Float - float buf[32] */
-FUNC_START(load_fpr_single_precision)
-	lfs 0, 0*4(3)
-	lfs 1, 1*4(3)
-	lfs 2, 2*4(3)
-	lfs 3, 3*4(3)
-	lfs 4, 4*4(3)
-	lfs 5, 5*4(3)
-	lfs 6, 6*4(3)
-	lfs 7, 7*4(3)
-	lfs 8, 8*4(3)
-	lfs 9, 9*4(3)
-	lfs 10, 10*4(3)
-	lfs 11, 11*4(3)
-	lfs 12, 12*4(3)
-	lfs 13, 13*4(3)
-	lfs 14, 14*4(3)
-	lfs 15, 15*4(3)
-	lfs 16, 16*4(3)
-	lfs 17, 17*4(3)
-	lfs 18, 18*4(3)
-	lfs 19, 19*4(3)
-	lfs 20, 20*4(3)
-	lfs 21, 21*4(3)
-	lfs 22, 22*4(3)
-	lfs 23, 23*4(3)
-	lfs 24, 24*4(3)
-	lfs 25, 25*4(3)
-	lfs 26, 26*4(3)
-	lfs 27, 27*4(3)
-	lfs 28, 28*4(3)
-	lfs 29, 29*4(3)
-	lfs 30, 30*4(3)
-	lfs 31, 31*4(3)
-	blr
-FUNC_END(load_fpr_single_precision)
-
 /* Single Precision Float - float buf[32] */
 FUNC_START(store_fpr_single_precision)
 	stfs 0, 0*4(3)
-- 
2.35.3

[PATCH 08/13] selftests/powerpc/ptrace: Convert to load/store doubles

[Michael Ellerman]

Some of the ptrace tests check the contents of floating pointer
registers. Currently these use float, which is always 4 bytes, but the
ptrace API supports saving/restoring 8 bytes per register, so switch to
using doubles to exercise the code more fully.

Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
---
 tools/testing/selftests/powerpc/include/reg.h | 68 +++++++++---------
 tools/testing/selftests/powerpc/lib/reg.S     | 70 +++++++++----------
 .../selftests/powerpc/ptrace/ptrace-gpr.c     | 16 ++---
 .../selftests/powerpc/ptrace/ptrace-gpr.h     | 14 ++--
 .../selftests/powerpc/ptrace/ptrace-tm-gpr.c  | 18 ++---
 .../powerpc/ptrace/ptrace-tm-spd-gpr.c        | 20 +++---
 .../testing/selftests/powerpc/ptrace/ptrace.h | 14 ++--
 7 files changed, 111 insertions(+), 109 deletions(-)

diff --git a/tools/testing/selftests/powerpc/include/reg.h b/tools/testing/selftests/powerpc/include/reg.h
index bc7f610c2c6d..7f0dc2f62496 100644
--- a/tools/testing/selftests/powerpc/include/reg.h
+++ b/tools/testing/selftests/powerpc/include/reg.h
@@ -123,44 +123,44 @@
 		"li 30, %[" #_asm_symbol_name_immed "];" \
 		"li 31, %[" #_asm_symbol_name_immed "];"
 
-#define ASM_LOAD_FPR_SINGLE_PRECISION(_asm_symbol_name_addr) \
-		"lfs 0, 0(%[" #_asm_symbol_name_addr "]);" \
-		"lfs 1, 0(%[" #_asm_symbol_name_addr "]);" \
-		"lfs 2, 0(%[" #_asm_symbol_name_addr "]);" \
-		"lfs 3, 0(%[" #_asm_symbol_name_addr "]);" \
-		"lfs 4, 0(%[" #_asm_symbol_name_addr "]);" \
-		"lfs 5, 0(%[" #_asm_symbol_name_addr "]);" \
-		"lfs 6, 0(%[" #_asm_symbol_name_addr "]);" \
-		"lfs 7, 0(%[" #_asm_symbol_name_addr "]);" \
-		"lfs 8, 0(%[" #_asm_symbol_name_addr "]);" \
-		"lfs 9, 0(%[" #_asm_symbol_name_addr "]);" \
-		"lfs 10, 0(%[" #_asm_symbol_name_addr "]);" \
-		"lfs 11, 0(%[" #_asm_symbol_name_addr "]);" \
-		"lfs 12, 0(%[" #_asm_symbol_name_addr "]);" \
-		"lfs 13, 0(%[" #_asm_symbol_name_addr "]);" \
-		"lfs 14, 0(%[" #_asm_symbol_name_addr "]);" \
-		"lfs 15, 0(%[" #_asm_symbol_name_addr "]);" \
-		"lfs 16, 0(%[" #_asm_symbol_name_addr "]);" \
-		"lfs 17, 0(%[" #_asm_symbol_name_addr "]);" \
-		"lfs 18, 0(%[" #_asm_symbol_name_addr "]);" \
-		"lfs 19, 0(%[" #_asm_symbol_name_addr "]);" \
-		"lfs 20, 0(%[" #_asm_symbol_name_addr "]);" \
-		"lfs 21, 0(%[" #_asm_symbol_name_addr "]);" \
-		"lfs 22, 0(%[" #_asm_symbol_name_addr "]);" \
-		"lfs 23, 0(%[" #_asm_symbol_name_addr "]);" \
-		"lfs 24, 0(%[" #_asm_symbol_name_addr "]);" \
-		"lfs 25, 0(%[" #_asm_symbol_name_addr "]);" \
-		"lfs 26, 0(%[" #_asm_symbol_name_addr "]);" \
-		"lfs 27, 0(%[" #_asm_symbol_name_addr "]);" \
-		"lfs 28, 0(%[" #_asm_symbol_name_addr "]);" \
-		"lfs 29, 0(%[" #_asm_symbol_name_addr "]);" \
-		"lfs 30, 0(%[" #_asm_symbol_name_addr "]);" \
-		"lfs 31, 0(%[" #_asm_symbol_name_addr "]);"
+#define ASM_LOAD_FPR(_asm_symbol_name_addr) \
+		"lfd 0, 0(%[" #_asm_symbol_name_addr "]);" \
+		"lfd 1, 0(%[" #_asm_symbol_name_addr "]);" \
+		"lfd 2, 0(%[" #_asm_symbol_name_addr "]);" \
+		"lfd 3, 0(%[" #_asm_symbol_name_addr "]);" \
+		"lfd 4, 0(%[" #_asm_symbol_name_addr "]);" \
+		"lfd 5, 0(%[" #_asm_symbol_name_addr "]);" \
+		"lfd 6, 0(%[" #_asm_symbol_name_addr "]);" \
+		"lfd 7, 0(%[" #_asm_symbol_name_addr "]);" \
+		"lfd 8, 0(%[" #_asm_symbol_name_addr "]);" \
+		"lfd 9, 0(%[" #_asm_symbol_name_addr "]);" \
+		"lfd 10, 0(%[" #_asm_symbol_name_addr "]);" \
+		"lfd 11, 0(%[" #_asm_symbol_name_addr "]);" \
+		"lfd 12, 0(%[" #_asm_symbol_name_addr "]);" \
+		"lfd 13, 0(%[" #_asm_symbol_name_addr "]);" \
+		"lfd 14, 0(%[" #_asm_symbol_name_addr "]);" \
+		"lfd 15, 0(%[" #_asm_symbol_name_addr "]);" \
+		"lfd 16, 0(%[" #_asm_symbol_name_addr "]);" \
+		"lfd 17, 0(%[" #_asm_symbol_name_addr "]);" \
+		"lfd 18, 0(%[" #_asm_symbol_name_addr "]);" \
+		"lfd 19, 0(%[" #_asm_symbol_name_addr "]);" \
+		"lfd 20, 0(%[" #_asm_symbol_name_addr "]);" \
+		"lfd 21, 0(%[" #_asm_symbol_name_addr "]);" \
+		"lfd 22, 0(%[" #_asm_symbol_name_addr "]);" \
+		"lfd 23, 0(%[" #_asm_symbol_name_addr "]);" \
+		"lfd 24, 0(%[" #_asm_symbol_name_addr "]);" \
+		"lfd 25, 0(%[" #_asm_symbol_name_addr "]);" \
+		"lfd 26, 0(%[" #_asm_symbol_name_addr "]);" \
+		"lfd 27, 0(%[" #_asm_symbol_name_addr "]);" \
+		"lfd 28, 0(%[" #_asm_symbol_name_addr "]);" \
+		"lfd 29, 0(%[" #_asm_symbol_name_addr "]);" \
+		"lfd 30, 0(%[" #_asm_symbol_name_addr "]);" \
+		"lfd 31, 0(%[" #_asm_symbol_name_addr "]);"
 
 #ifndef __ASSEMBLER__
 void store_gpr(unsigned long *addr);
 void load_gpr(unsigned long *addr);
-void store_fpr_single_precision(float *addr);
+void store_fpr(double *addr);
 #endif /* end of __ASSEMBLER__ */
 
 #endif /* _SELFTESTS_POWERPC_REG_H */
diff --git a/tools/testing/selftests/powerpc/lib/reg.S b/tools/testing/selftests/powerpc/lib/reg.S
index dd37b8e6f84c..6d1af4a9a6b4 100644
--- a/tools/testing/selftests/powerpc/lib/reg.S
+++ b/tools/testing/selftests/powerpc/lib/reg.S
@@ -53,42 +53,42 @@ FUNC_START(store_gpr)
 	blr
 FUNC_END(store_gpr)
 
-/* Single Precision Float - float buf[32] */
-FUNC_START(store_fpr_single_precision)
-	stfs 0, 0*4(3)
-	stfs 1, 1*4(3)
-	stfs 2, 2*4(3)
-	stfs 3, 3*4(3)
-	stfs 4, 4*4(3)
-	stfs 5, 5*4(3)
-	stfs 6, 6*4(3)
-	stfs 7, 7*4(3)
-	stfs 8, 8*4(3)
-	stfs 9, 9*4(3)
-	stfs 10, 10*4(3)
-	stfs 11, 11*4(3)
-	stfs 12, 12*4(3)
-	stfs 13, 13*4(3)
-	stfs 14, 14*4(3)
-	stfs 15, 15*4(3)
-	stfs 16, 16*4(3)
-	stfs 17, 17*4(3)
-	stfs 18, 18*4(3)
-	stfs 19, 19*4(3)
-	stfs 20, 20*4(3)
-	stfs 21, 21*4(3)
-	stfs 22, 22*4(3)
-	stfs 23, 23*4(3)
-	stfs 24, 24*4(3)
-	stfs 25, 25*4(3)
-	stfs 26, 26*4(3)
-	stfs 27, 27*4(3)
-	stfs 28, 28*4(3)
-	stfs 29, 29*4(3)
-	stfs 30, 30*4(3)
-	stfs 31, 31*4(3)
+/* Double Precision Float - double buf[32] */
+FUNC_START(store_fpr)
+	stfd  0,  0*8(3)
+	stfd  1,  1*8(3)
+	stfd  2,  2*8(3)
+	stfd  3,  3*8(3)
+	stfd  4,  4*8(3)
+	stfd  5,  5*8(3)
+	stfd  6,  6*8(3)
+	stfd  7,  7*8(3)
+	stfd  8,  8*8(3)
+	stfd  9,  9*8(3)
+	stfd 10, 10*8(3)
+	stfd 11, 11*8(3)
+	stfd 12, 12*8(3)
+	stfd 13, 13*8(3)
+	stfd 14, 14*8(3)
+	stfd 15, 15*8(3)
+	stfd 16, 16*8(3)
+	stfd 17, 17*8(3)
+	stfd 18, 18*8(3)
+	stfd 19, 19*8(3)
+	stfd 20, 20*8(3)
+	stfd 21, 21*8(3)
+	stfd 22, 22*8(3)
+	stfd 23, 23*8(3)
+	stfd 24, 24*8(3)
+	stfd 25, 25*8(3)
+	stfd 26, 26*8(3)
+	stfd 27, 27*8(3)
+	stfd 28, 28*8(3)
+	stfd 29, 29*8(3)
+	stfd 30, 30*8(3)
+	stfd 31, 31*8(3)
 	blr
-FUNC_END(store_fpr_single_precision)
+FUNC_END(store_fpr)
 
 /* VMX/VSX registers - unsigned long buf[128] */
 FUNC_START(loadvsx)
diff --git a/tools/testing/selftests/powerpc/ptrace/ptrace-gpr.c b/tools/testing/selftests/powerpc/ptrace/ptrace-gpr.c
index 17cd480c8780..1468e89c044b 100644
--- a/tools/testing/selftests/powerpc/ptrace/ptrace-gpr.c
+++ b/tools/testing/selftests/powerpc/ptrace/ptrace-gpr.c
@@ -12,20 +12,20 @@
 int shm_id;
 int *cptr, *pptr;
 
-float a = FPR_1;
-float b = FPR_2;
-float c = FPR_3;
+double a = FPR_1;
+double b = FPR_2;
+double c = FPR_3;
 
 void gpr(void)
 {
 	unsigned long gpr_buf[18];
-	float fpr_buf[32];
+	double fpr_buf[32];
 
 	cptr = (int *)shmat(shm_id, NULL, 0);
 
 	asm __volatile__(
 		ASM_LOAD_GPR_IMMED(gpr_1)
-		ASM_LOAD_FPR_SINGLE_PRECISION(flt_1)
+		ASM_LOAD_FPR(flt_1)
 		:
 		: [gpr_1]"i"(GPR_1), [flt_1] "b" (&a)
 		: "memory", "r6", "r7", "r8", "r9", "r10",
@@ -41,12 +41,12 @@ void gpr(void)
 
 	shmdt((void *)cptr);
 	store_gpr(gpr_buf);
-	store_fpr_single_precision(fpr_buf);
+	store_fpr(fpr_buf);
 
 	if (validate_gpr(gpr_buf, GPR_3))
 		exit(1);
 
-	if (validate_fpr_float(fpr_buf, c))
+	if (validate_fpr_double(fpr_buf, c))
 		exit(1);
 
 	exit(0);
@@ -55,7 +55,7 @@ void gpr(void)
 int trace_gpr(pid_t child)
 {
 	unsigned long gpr[18];
-	unsigned long fpr[32];
+	__u64 fpr[32];
 
 	FAIL_IF(start_trace(child));
 	FAIL_IF(show_gpr(child, gpr));
diff --git a/tools/testing/selftests/powerpc/ptrace/ptrace-gpr.h b/tools/testing/selftests/powerpc/ptrace/ptrace-gpr.h
index c5cd53181e2e..a5470b88bd08 100644
--- a/tools/testing/selftests/powerpc/ptrace/ptrace-gpr.h
+++ b/tools/testing/selftests/powerpc/ptrace/ptrace-gpr.h
@@ -12,10 +12,10 @@
 #define FPR_3	0.003
 #define FPR_4	0.004
 
-#define FPR_1_REP 0x3f50624de0000000
-#define FPR_2_REP 0x3f60624de0000000
-#define FPR_3_REP 0x3f689374c0000000
-#define FPR_4_REP 0x3f70624de0000000
+#define FPR_1_REP 0x3f50624dd2f1a9fcull
+#define FPR_2_REP 0x3f60624dd2f1a9fcull
+#define FPR_3_REP 0x3f689374bc6a7efaull
+#define FPR_4_REP 0x3f70624dd2f1a9fcull
 
 /* Buffer must have 18 elements */
 int validate_gpr(unsigned long *gpr, unsigned long val)
@@ -36,13 +36,13 @@ int validate_gpr(unsigned long *gpr, unsigned long val)
 }
 
 /* Buffer must have 32 elements */
-int validate_fpr(unsigned long *fpr, unsigned long val)
+int validate_fpr(__u64 *fpr, __u64 val)
 {
 	int i, found = 1;
 
 	for (i = 0; i < 32; i++) {
 		if (fpr[i] != val) {
-			printf("FPR[%d]: %lx Expected: %lx\n", i, fpr[i], val);
+			printf("FPR[%d]: %llx Expected: %llx\n", i, fpr[i], val);
 			found = 0;
 		}
 	}
@@ -53,7 +53,7 @@ int validate_fpr(unsigned long *fpr, unsigned long val)
 }
 
 /* Buffer must have 32 elements */
-int validate_fpr_float(float *fpr, float val)
+int validate_fpr_double(double *fpr, double val)
 {
 	int i, found = 1;
 
diff --git a/tools/testing/selftests/powerpc/ptrace/ptrace-tm-gpr.c b/tools/testing/selftests/powerpc/ptrace/ptrace-tm-gpr.c
index 67ca297c5cca..5dc152b162df 100644
--- a/tools/testing/selftests/powerpc/ptrace/ptrace-tm-gpr.c
+++ b/tools/testing/selftests/powerpc/ptrace/ptrace-tm-gpr.c
@@ -12,15 +12,15 @@
 int shm_id;
 unsigned long *cptr, *pptr;
 
-float a = FPR_1;
-float b = FPR_2;
-float c = FPR_3;
+double a = FPR_1;
+double b = FPR_2;
+double c = FPR_3;
 
 void tm_gpr(void)
 {
 	unsigned long gpr_buf[18];
 	unsigned long result, texasr;
-	float fpr_buf[32];
+	double fpr_buf[32];
 
 	printf("Starting the child\n");
 	cptr = (unsigned long *)shmat(shm_id, NULL, 0);
@@ -29,12 +29,12 @@ void tm_gpr(void)
 	cptr[1] = 0;
 	asm __volatile__(
 		ASM_LOAD_GPR_IMMED(gpr_1)
-		ASM_LOAD_FPR_SINGLE_PRECISION(flt_1)
+		ASM_LOAD_FPR(flt_1)
 		"1: ;"
 		"tbegin.;"
 		"beq 2f;"
 		ASM_LOAD_GPR_IMMED(gpr_2)
-		ASM_LOAD_FPR_SINGLE_PRECISION(flt_2)
+		ASM_LOAD_FPR(flt_2)
 		"tsuspend.;"
 		"li 7, 1;"
 		"stw 7, 0(%[cptr1]);"
@@ -70,12 +70,12 @@ void tm_gpr(void)
 
 		shmdt((void *)cptr);
 		store_gpr(gpr_buf);
-		store_fpr_single_precision(fpr_buf);
+		store_fpr(fpr_buf);
 
 		if (validate_gpr(gpr_buf, GPR_3))
 			exit(1);
 
-		if (validate_fpr_float(fpr_buf, c))
+		if (validate_fpr_double(fpr_buf, c))
 			exit(1);
 
 		exit(0);
@@ -87,7 +87,7 @@ void tm_gpr(void)
 int trace_tm_gpr(pid_t child)
 {
 	unsigned long gpr[18];
-	unsigned long fpr[32];
+	__u64 fpr[32];
 
 	FAIL_IF(start_trace(child));
 	FAIL_IF(show_gpr(child, gpr));
diff --git a/tools/testing/selftests/powerpc/ptrace/ptrace-tm-spd-gpr.c b/tools/testing/selftests/powerpc/ptrace/ptrace-tm-spd-gpr.c
index 6f2bce1b6c5d..458cc1a70ccf 100644
--- a/tools/testing/selftests/powerpc/ptrace/ptrace-tm-spd-gpr.c
+++ b/tools/testing/selftests/powerpc/ptrace/ptrace-tm-spd-gpr.c
@@ -12,10 +12,10 @@
 int shm_id;
 int *cptr, *pptr;
 
-float a = FPR_1;
-float b = FPR_2;
-float c = FPR_3;
-float d = FPR_4;
+double a = FPR_1;
+double b = FPR_2;
+double c = FPR_3;
+double d = FPR_4;
 
 __attribute__((used)) void wait_parent(void)
 {
@@ -28,7 +28,7 @@ void tm_spd_gpr(void)
 {
 	unsigned long gpr_buf[18];
 	unsigned long result, texasr;
-	float fpr_buf[32];
+	double fpr_buf[32];
 
 	cptr = (int *)shmat(shm_id, NULL, 0);
 
@@ -36,7 +36,7 @@ void tm_spd_gpr(void)
 	cptr[2] = 0;
 	asm __volatile__(
 		ASM_LOAD_GPR_IMMED(gpr_1)
-		ASM_LOAD_FPR_SINGLE_PRECISION(flt_1)
+		ASM_LOAD_FPR(flt_1)
 
 		"1: ;"
 		"tbegin.;"
@@ -45,7 +45,7 @@ void tm_spd_gpr(void)
 		ASM_LOAD_GPR_IMMED(gpr_2)
 		"tsuspend.;"
 		ASM_LOAD_GPR_IMMED(gpr_4)
-		ASM_LOAD_FPR_SINGLE_PRECISION(flt_4)
+		ASM_LOAD_FPR(flt_4)
 
 		"bl wait_parent;"
 		"tresume.;"
@@ -77,12 +77,12 @@ void tm_spd_gpr(void)
 
 		shmdt((void *)cptr);
 		store_gpr(gpr_buf);
-		store_fpr_single_precision(fpr_buf);
+		store_fpr(fpr_buf);
 
 		if (validate_gpr(gpr_buf, GPR_3))
 			exit(1);
 
-		if (validate_fpr_float(fpr_buf, c))
+		if (validate_fpr_double(fpr_buf, c))
 			exit(1);
 		exit(0);
 	}
@@ -93,7 +93,7 @@ void tm_spd_gpr(void)
 int trace_tm_spd_gpr(pid_t child)
 {
 	unsigned long gpr[18];
-	unsigned long fpr[32];
+	__u64 fpr[32];
 
 	FAIL_IF(start_trace(child));
 	FAIL_IF(show_gpr(child, gpr));
diff --git a/tools/testing/selftests/powerpc/ptrace/ptrace.h b/tools/testing/selftests/powerpc/ptrace/ptrace.h
index 5181ad9b4b6c..4672e848604f 100644
--- a/tools/testing/selftests/powerpc/ptrace/ptrace.h
+++ b/tools/testing/selftests/powerpc/ptrace/ptrace.h
@@ -4,6 +4,9 @@
  *
  * Copyright (C) 2015 Anshuman Khandual, IBM Corporation.
  */
+
+#define __SANE_USERSPACE_TYPES__
+
 #include <inttypes.h>
 #include <unistd.h>
 #include <stdlib.h>
@@ -30,8 +33,8 @@
 #define TEST_FAIL 1
 
 struct fpr_regs {
-	unsigned long fpr[32];
-	unsigned long fpscr;
+	__u64 fpr[32];
+	__u64 fpscr;
 };
 
 struct tm_spr_regs {
@@ -318,7 +321,7 @@ int write_ckpt_tar_registers(pid_t child, unsigned long tar,
 }
 
 /* FPR */
-int show_fpr(pid_t child, unsigned long *fpr)
+int show_fpr(pid_t child, __u64 *fpr)
 {
 	struct fpr_regs *regs;
 	int ret, i;
@@ -337,7 +340,7 @@ int show_fpr(pid_t child, unsigned long *fpr)
 	return TEST_PASS;
 }
 
-int write_fpr(pid_t child, unsigned long val)
+int write_fpr(pid_t child, __u64 val)
 {
 	struct fpr_regs *regs;
 	int ret, i;
@@ -360,7 +363,7 @@ int write_fpr(pid_t child, unsigned long val)
 	return TEST_PASS;
 }
 
-int show_ckpt_fpr(pid_t child, unsigned long *fpr)
+int show_ckpt_fpr(pid_t child, __u64 *fpr)
 {
 	struct fpr_regs *regs;
 	struct iovec iov;
@@ -742,4 +745,3 @@ void analyse_texasr(unsigned long texasr)
 }
 
 void store_gpr(unsigned long *addr);
-void store_fpr(float *addr);
-- 
2.35.3

[PATCH 09/13] selftests/powerpc/ptrace: Build the ptrace-gpr test as 32-bit when possible

[Michael Ellerman]

The ptrace-gpr test can now be built 32-bit, so do that if that's the
compiler default rather than forcing a 64-bit build.

Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
---
 tools/testing/selftests/powerpc/ptrace/Makefile | 5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)

diff --git a/tools/testing/selftests/powerpc/ptrace/Makefile b/tools/testing/selftests/powerpc/ptrace/Makefile
index 8611b0670587..3434a624ed77 100644
--- a/tools/testing/selftests/powerpc/ptrace/Makefile
+++ b/tools/testing/selftests/powerpc/ptrace/Makefile
@@ -11,7 +11,6 @@ TM_TESTS += ptrace-tm-vsx
 TESTS_64 := $(TM_TESTS)
 TESTS_64 += core-pkey
 TESTS_64 += perf-hwbreak
-TESTS_64 += ptrace-gpr
 TESTS_64 += ptrace-hwbreak
 TESTS_64 += ptrace-perf-hwbreak
 TESTS_64 += ptrace-pkey
@@ -19,7 +18,9 @@ TESTS_64 += ptrace-syscall
 TESTS_64 += ptrace-tar
 TESTS_64 += ptrace-vsx
 
-TEST_GEN_PROGS := $(TESTS_64)
+TESTS += ptrace-gpr
+
+TEST_GEN_PROGS := $(TESTS) $(TESTS_64)
 
 LOCAL_HDRS += $(patsubst %,$(selfdir)/powerpc/ptrace/%,$(wildcard *.h))
 
-- 
2.35.3

[PATCH 10/13] selftests/powerpc/ptrace: Do more of ptrace-gpr in asm

[Michael Ellerman]

The ptrace-gpr test includes some inline asm to load GPR and FPR
registers. It then goes back to C to wait for the parent to trace it and
then checks register contents.

The split between inline asm and C is fragile, it relies on the compiler
not using any non-volatile GPRs after the inline asm block. It also
requires a very large and unwieldy inline asm block.

So convert the logic to set registers, wait, and store registers to a
single asm function, meaning there's no window for the compiler to
intervene.

Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
---
 .../selftests/powerpc/include/basic_asm.h     |  8 +++
 .../testing/selftests/powerpc/ptrace/Makefile |  1 +
 .../selftests/powerpc/ptrace/ptrace-gpr.S     | 52 +++++++++++++++++++
 .../selftests/powerpc/ptrace/ptrace-gpr.c     | 29 +++++------
 4 files changed, 73 insertions(+), 17 deletions(-)
 create mode 100644 tools/testing/selftests/powerpc/ptrace/ptrace-gpr.S

diff --git a/tools/testing/selftests/powerpc/include/basic_asm.h b/tools/testing/selftests/powerpc/include/basic_asm.h
index 2d7f6e592dd9..26cde8ea1f49 100644
--- a/tools/testing/selftests/powerpc/include/basic_asm.h
+++ b/tools/testing/selftests/powerpc/include/basic_asm.h
@@ -94,4 +94,12 @@
 	PPC_LL	r0, STACK_FRAME_LR_POS(%r1); \
 	mtlr	r0;
 
+.macro OP_REGS op, reg_width, start_reg, end_reg, base_reg, base_reg_offset=0, skip=0
+	.set i, \start_reg
+	.rept (\end_reg - \start_reg + 1)
+	\op	i, (\reg_width * (i - \skip) + \base_reg_offset)(\base_reg)
+	.set i, i + 1
+	.endr
+.endm
+
 #endif /* _SELFTESTS_POWERPC_BASIC_ASM_H */
diff --git a/tools/testing/selftests/powerpc/ptrace/Makefile b/tools/testing/selftests/powerpc/ptrace/Makefile
index 3434a624ed77..2f02cb54224d 100644
--- a/tools/testing/selftests/powerpc/ptrace/Makefile
+++ b/tools/testing/selftests/powerpc/ptrace/Makefile
@@ -35,6 +35,7 @@ $(TM_TESTS): CFLAGS += -I../tm -mhtm
 
 CFLAGS += -I../../../../../usr/include -fno-pie
 
+$(OUTPUT)/ptrace-gpr: ptrace-gpr.S
 $(OUTPUT)/ptrace-pkey $(OUTPUT)/core-pkey: LDLIBS += -pthread
 
 $(TEST_GEN_PROGS): ../harness.c ../utils.c ../lib/reg.S
diff --git a/tools/testing/selftests/powerpc/ptrace/ptrace-gpr.S b/tools/testing/selftests/powerpc/ptrace/ptrace-gpr.S
new file mode 100644
index 000000000000..070e8443e3cc
--- /dev/null
+++ b/tools/testing/selftests/powerpc/ptrace/ptrace-gpr.S
@@ -0,0 +1,52 @@
+/* SPDX-License-Identifier: GPL-2.0-or-later */
+/*
+ * test helper assembly functions
+ *
+ * Copyright (C) 2016 Simon Guo, IBM Corporation.
+ * Copyright 2022 Michael Ellerman, IBM Corporation.
+ */
+#include "basic_asm.h"
+
+#define GPR_SIZE	__SIZEOF_LONG__
+#define FIRST_GPR	14
+#define NUM_GPRS	(32 - FIRST_GPR)
+#define STACK_SIZE	(NUM_GPRS * GPR_SIZE)
+
+// gpr_child_loop(int *read_flag, int *write_flag,
+//		  unsigned long *gpr_buf, double *fpr_buf);
+FUNC_START(gpr_child_loop)
+	// r3 = read_flag
+	// r4 = write_flag
+	// r5 = gpr_buf
+	// r6 = fpr_buf
+	PUSH_BASIC_STACK(STACK_SIZE)
+
+	// Save non-volatile GPRs
+	OP_REGS PPC_STL, GPR_SIZE, FIRST_GPR, 31, %r1, STACK_FRAME_LOCAL(0, 0), FIRST_GPR
+
+	// Load GPRs with expected values
+	OP_REGS PPC_LL, GPR_SIZE, FIRST_GPR, 31, r5, 0, FIRST_GPR
+
+	// Load FPRs with expected values
+	OP_REGS lfd, 8, 0, 31, r6
+
+	// Signal to parent that we're ready
+	li	r0, 1
+	stw	r0, 0(r4)
+
+	// Wait for parent to finish
+1:	lwz	r0, 0(r3)
+	cmpwi	r0, 0
+	beq	1b	// Loop while flag is zero
+
+	// Save GPRs back to caller buffer
+	OP_REGS PPC_STL, GPR_SIZE, FIRST_GPR, 31, r5, 0, FIRST_GPR
+
+	// Save FPRs
+	OP_REGS stfd, 8, 0, 31, r6
+
+	// Reload non-volatile GPRs
+	OP_REGS PPC_LL, GPR_SIZE, FIRST_GPR, 31, %r1, STACK_FRAME_LOCAL(0, 0), FIRST_GPR
+
+	POP_BASIC_STACK(STACK_SIZE)
+	blr
diff --git a/tools/testing/selftests/powerpc/ptrace/ptrace-gpr.c b/tools/testing/selftests/powerpc/ptrace/ptrace-gpr.c
index 1468e89c044b..4e7a7eb01e3a 100644
--- a/tools/testing/selftests/powerpc/ptrace/ptrace-gpr.c
+++ b/tools/testing/selftests/powerpc/ptrace/ptrace-gpr.c
@@ -16,32 +16,27 @@ double a = FPR_1;
 double b = FPR_2;
 double c = FPR_3;
 
+extern void gpr_child_loop(int *read_flag, int *write_flag,
+			   unsigned long *gpr_buf, double *fpr_buf);
+
 void gpr(void)
 {
-	unsigned long gpr_buf[18];
+	unsigned long gpr_buf[32];
 	double fpr_buf[32];
+	int i;
 
 	cptr = (int *)shmat(shm_id, NULL, 0);
+	memset(gpr_buf, 0, sizeof(gpr_buf));
+	memset(fpr_buf, 0, sizeof(fpr_buf));
 
-	asm __volatile__(
-		ASM_LOAD_GPR_IMMED(gpr_1)
-		ASM_LOAD_FPR(flt_1)
-		:
-		: [gpr_1]"i"(GPR_1), [flt_1] "b" (&a)
-		: "memory", "r6", "r7", "r8", "r9", "r10",
-		"r11", "r12", "r13", "r14", "r15", "r16", "r17",
-		"r18", "r19", "r20", "r21", "r22", "r23", "r24",
-		"r25", "r26", "r27", "r28", "r29", "r30", "r31"
-		);
-
-	cptr[1] = 1;
+	for (i = 0; i < 32; i++) {
+		gpr_buf[i] = GPR_1;
+		fpr_buf[i] = a;
+	}
 
-	while (!cptr[0])
-		asm volatile("" : : : "memory");
+	gpr_child_loop(&cptr[0], &cptr[1], gpr_buf, fpr_buf);
 
 	shmdt((void *)cptr);
-	store_gpr(gpr_buf);
-	store_fpr(fpr_buf);
 
 	if (validate_gpr(gpr_buf, GPR_3))
 		exit(1);
-- 
2.35.3

[PATCH 11/13] selftests/powerpc/ptrace: Make child errors more obvious

[Michael Ellerman]

Use the FAIL_IF() macro so that errors in the child report a line
number, rather than just silently exiting.

Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
---
 tools/testing/selftests/powerpc/ptrace/ptrace-gpr.c | 13 +++++--------
 1 file changed, 5 insertions(+), 8 deletions(-)

diff --git a/tools/testing/selftests/powerpc/ptrace/ptrace-gpr.c b/tools/testing/selftests/powerpc/ptrace/ptrace-gpr.c
index 4e7a7eb01e3a..b574ea26395c 100644
--- a/tools/testing/selftests/powerpc/ptrace/ptrace-gpr.c
+++ b/tools/testing/selftests/powerpc/ptrace/ptrace-gpr.c
@@ -19,7 +19,7 @@ double c = FPR_3;
 extern void gpr_child_loop(int *read_flag, int *write_flag,
 			   unsigned long *gpr_buf, double *fpr_buf);
 
-void gpr(void)
+static int child(void)
 {
 	unsigned long gpr_buf[32];
 	double fpr_buf[32];
@@ -38,13 +38,10 @@ void gpr(void)
 
 	shmdt((void *)cptr);
 
-	if (validate_gpr(gpr_buf, GPR_3))
-		exit(1);
+	FAIL_IF(validate_gpr(gpr_buf, GPR_3));
+	FAIL_IF(validate_fpr_double(fpr_buf, c));
 
-	if (validate_fpr_double(fpr_buf, c))
-		exit(1);
-
-	exit(0);
+	return 0;
 }
 
 int trace_gpr(pid_t child)
@@ -76,7 +73,7 @@ int ptrace_gpr(void)
 		return TEST_FAIL;
 	}
 	if (pid == 0)
-		gpr();
+		exit(child());
 
 	if (pid) {
 		pptr = (int *)shmat(shm_id, NULL, 0);
-- 
2.35.3

[PATCH 12/13] selftests/powerpc/ptrace: Use more interesting values

[Michael Ellerman]

The ptrace-gpr test uses fixed values to test that registers can be
read/written via ptrace. In particular it sets all GPRs to 1, which
means the test could miss some types of bugs - eg. if the kernel was
only returning the low word.

So generate some random values at startup and use those instead.

Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
---
 .../selftests/powerpc/ptrace/ptrace-gpr.c     | 71 +++++++++++++++----
 1 file changed, 57 insertions(+), 14 deletions(-)

diff --git a/tools/testing/selftests/powerpc/ptrace/ptrace-gpr.c b/tools/testing/selftests/powerpc/ptrace/ptrace-gpr.c
index b574ea26395c..c5dcb8c02616 100644
--- a/tools/testing/selftests/powerpc/ptrace/ptrace-gpr.c
+++ b/tools/testing/selftests/powerpc/ptrace/ptrace-gpr.c
@@ -7,18 +7,18 @@
 #include "ptrace.h"
 #include "ptrace-gpr.h"
 #include "reg.h"
+#include <time.h>
 
 /* Tracer and Tracee Shared Data */
 int shm_id;
 int *cptr, *pptr;
 
-double a = FPR_1;
-double b = FPR_2;
-double c = FPR_3;
-
 extern void gpr_child_loop(int *read_flag, int *write_flag,
 			   unsigned long *gpr_buf, double *fpr_buf);
 
+unsigned long child_gpr_val, parent_gpr_val;
+double child_fpr_val, parent_fpr_val;
+
 static int child(void)
 {
 	unsigned long gpr_buf[32];
@@ -30,16 +30,16 @@ static int child(void)
 	memset(fpr_buf, 0, sizeof(fpr_buf));
 
 	for (i = 0; i < 32; i++) {
-		gpr_buf[i] = GPR_1;
-		fpr_buf[i] = a;
+		gpr_buf[i] = child_gpr_val;
+		fpr_buf[i] = child_fpr_val;
 	}
 
 	gpr_child_loop(&cptr[0], &cptr[1], gpr_buf, fpr_buf);
 
 	shmdt((void *)cptr);
 
-	FAIL_IF(validate_gpr(gpr_buf, GPR_3));
-	FAIL_IF(validate_fpr_double(fpr_buf, c));
+	FAIL_IF(validate_gpr(gpr_buf, parent_gpr_val));
+	FAIL_IF(validate_fpr_double(fpr_buf, parent_fpr_val));
 
 	return 0;
 }
@@ -47,24 +47,67 @@ static int child(void)
 int trace_gpr(pid_t child)
 {
 	unsigned long gpr[18];
-	__u64 fpr[32];
+	__u64 tmp, fpr[32];
 
 	FAIL_IF(start_trace(child));
 	FAIL_IF(show_gpr(child, gpr));
-	FAIL_IF(validate_gpr(gpr, GPR_1));
+	FAIL_IF(validate_gpr(gpr, child_gpr_val));
 	FAIL_IF(show_fpr(child, fpr));
-	FAIL_IF(validate_fpr(fpr, FPR_1_REP));
-	FAIL_IF(write_gpr(child, GPR_3));
-	FAIL_IF(write_fpr(child, FPR_3_REP));
+
+	memcpy(&tmp, &child_fpr_val, sizeof(tmp));
+	FAIL_IF(validate_fpr(fpr, tmp));
+
+	FAIL_IF(write_gpr(child, parent_gpr_val));
+
+	memcpy(&tmp, &parent_fpr_val, sizeof(tmp));
+	FAIL_IF(write_fpr(child, tmp));
+
 	FAIL_IF(stop_trace(child));
 
 	return TEST_PASS;
 }
 
+#ifndef __LONG_WIDTH__
+#define __LONG_WIDTH__ (sizeof(long) * 8)
+#endif
+
+static uint64_t rand_reg(void)
+{
+	uint64_t result;
+	long r;
+
+	r = random();
+
+	// Small values are typical
+	result = r & 0xffff;
+	if (r & 0x10000)
+		return result;
+
+	// Pointers tend to have high bits set
+	result |= random() << (__LONG_WIDTH__ - 31);
+	if (r & 0x100000)
+		return result;
+
+	// And sometimes we want a full 64-bit value
+	result ^= random() << 16;
+
+	return result;
+}
+
 int ptrace_gpr(void)
 {
-	pid_t pid;
+	unsigned long seed;
 	int ret, status;
+	pid_t pid;
+
+	seed = getpid() ^ time(NULL);
+	printf("srand(%lu)\n", seed);
+	srand(seed);
+
+	child_gpr_val = rand_reg();
+	child_fpr_val = rand_reg();
+	parent_gpr_val = rand_reg();
+	parent_fpr_val = rand_reg();
 
 	shm_id = shmget(IPC_PRIVATE, sizeof(int) * 2, 0777|IPC_CREAT);
 	pid = fork();
-- 
2.35.3

[PATCH 13/13] selftests/powerpc/ptrace: Add peek/poke of FPRs

[Michael Ellerman]

Currently the ptrace-gpr test only tests the GET/SET(FP)REGS ptrace
APIs. But there's an alternate (older) API, called PEEK/POKEUSR.

Add some minimal testing of PEEK/POKEUSR of the FPRs. This is sufficient
to detect the bug that was fixed recently in the 32-bit ptrace FPR
handling.

Depends-on: 8e1278444446 ("powerpc/32: Fix overread/overwrite of thread_struct via ptrace")
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
---
 .../selftests/powerpc/ptrace/ptrace-gpr.c     | 24 ++++++-
 .../testing/selftests/powerpc/ptrace/ptrace.h | 65 +++++++++++++++++++
 2 files changed, 87 insertions(+), 2 deletions(-)

diff --git a/tools/testing/selftests/powerpc/ptrace/ptrace-gpr.c b/tools/testing/selftests/powerpc/ptrace/ptrace-gpr.c
index c5dcb8c02616..9ed87d297799 100644
--- a/tools/testing/selftests/powerpc/ptrace/ptrace-gpr.c
+++ b/tools/testing/selftests/powerpc/ptrace/ptrace-gpr.c
@@ -46,22 +46,42 @@ static int child(void)
 
 int trace_gpr(pid_t child)
 {
+	__u64 tmp, fpr[32], *peeked_fprs;
 	unsigned long gpr[18];
-	__u64 tmp, fpr[32];
 
 	FAIL_IF(start_trace(child));
+
+	// Check child GPRs match what we expect using GETREGS
 	FAIL_IF(show_gpr(child, gpr));
 	FAIL_IF(validate_gpr(gpr, child_gpr_val));
-	FAIL_IF(show_fpr(child, fpr));
 
+	// Check child FPRs match what we expect using GETFPREGS
+	FAIL_IF(show_fpr(child, fpr));
 	memcpy(&tmp, &child_fpr_val, sizeof(tmp));
 	FAIL_IF(validate_fpr(fpr, tmp));
 
+	// Check child FPRs match what we expect using PEEKUSR
+	peeked_fprs = peek_fprs(child);
+	FAIL_IF(!peeked_fprs);
+	FAIL_IF(validate_fpr(peeked_fprs, tmp));
+	free(peeked_fprs);
+
+	// Write child GPRs using SETREGS
 	FAIL_IF(write_gpr(child, parent_gpr_val));
 
+	// Write child FPRs using SETFPREGS
 	memcpy(&tmp, &parent_fpr_val, sizeof(tmp));
 	FAIL_IF(write_fpr(child, tmp));
 
+	// Check child FPRs match what we just set, using PEEKUSR
+	peeked_fprs = peek_fprs(child);
+	FAIL_IF(!peeked_fprs);
+	FAIL_IF(validate_fpr(peeked_fprs, tmp));
+
+	// Write child FPRs using POKEUSR
+	FAIL_IF(poke_fprs(child, (unsigned long *)peeked_fprs));
+
+	// Child will check its FPRs match before exiting
 	FAIL_IF(stop_trace(child));
 
 	return TEST_PASS;
diff --git a/tools/testing/selftests/powerpc/ptrace/ptrace.h b/tools/testing/selftests/powerpc/ptrace/ptrace.h
index 4672e848604f..4e0233c0f2b3 100644
--- a/tools/testing/selftests/powerpc/ptrace/ptrace.h
+++ b/tools/testing/selftests/powerpc/ptrace/ptrace.h
@@ -23,6 +23,7 @@
 #include <sys/ipc.h>
 #include <sys/shm.h>
 #include <sys/user.h>
+#include <sys/syscall.h>
 #include <linux/elf.h>
 #include <linux/types.h>
 #include <linux/auxvec.h>
@@ -440,6 +441,70 @@ int show_gpr(pid_t child, unsigned long *gpr)
 	return TEST_PASS;
 }
 
+long sys_ptrace(enum __ptrace_request request, pid_t pid, unsigned long addr, unsigned long data)
+{
+	return syscall(__NR_ptrace, request, pid, (void *)addr, data);
+}
+
+// 33 because of FPSCR
+#define PT_NUM_FPRS	(33 * (sizeof(__u64) / sizeof(unsigned long)))
+
+__u64 *peek_fprs(pid_t child)
+{
+	unsigned long *fprs, *p, addr;
+	long ret;
+	int i;
+
+	fprs = malloc(sizeof(unsigned long) * PT_NUM_FPRS);
+	if (!fprs) {
+		perror("malloc() failed");
+		return NULL;
+	}
+
+	for (i = 0, p = fprs; i < PT_NUM_FPRS; i++, p++) {
+		addr = sizeof(unsigned long) * (PT_FPR0 + i);
+		ret = sys_ptrace(PTRACE_PEEKUSER, child, addr, (unsigned long)p);
+		if (ret) {
+			perror("ptrace(PTRACE_PEEKUSR) failed");
+			return NULL;
+		}
+	}
+
+	addr = sizeof(unsigned long) * (PT_FPR0 + i);
+	ret = sys_ptrace(PTRACE_PEEKUSER, child, addr, (unsigned long)&addr);
+	if (!ret) {
+		printf("ptrace(PTRACE_PEEKUSR) succeeded unexpectedly!\n");
+		return NULL;
+	}
+
+	return (__u64 *)fprs;
+}
+
+int poke_fprs(pid_t child, unsigned long *fprs)
+{
+	unsigned long *p, addr;
+	long ret;
+	int i;
+
+	for (i = 0, p = fprs; i < PT_NUM_FPRS; i++, p++) {
+		addr = sizeof(unsigned long) * (PT_FPR0 + i);
+		ret = sys_ptrace(PTRACE_POKEUSER, child, addr, *p);
+		if (ret) {
+			perror("ptrace(PTRACE_POKEUSR) failed");
+			return -1;
+		}
+	}
+
+	addr = sizeof(unsigned long) * (PT_FPR0 + i);
+	ret = sys_ptrace(PTRACE_POKEUSER, child, addr, addr);
+	if (!ret) {
+		printf("ptrace(PTRACE_POKEUSR) succeeded unexpectedly!\n");
+		return -1;
+	}
+
+	return 0;
+}
+
 int write_gpr(pid_t child, unsigned long val)
 {
 	struct pt_regs *regs;
-- 
2.35.3

Re: [PATCH 01/13] selftests/powerpc: Ensure 16-byte stack pointer alignment

[Michael Ellerman]

On Tue, 28 Jun 2022 00:02:27 +1000, Michael Ellerman wrote:
> The PUSH/POP_BASIC_STACK helpers in basic_asm.h do not ensure that the
> stack pointer is always 16-byte aligned, which is required per the ABI.
> 
> Fix the macros to do the alignment if the caller fails to.
> 
> Currently only one caller passes a non-aligned size, tm_signal_self(),
> which hasn't been caught in testing, presumably because it's a leaf
> function.
> 
> [...]

Applied to powerpc/next.

[01/13] selftests/powerpc: Ensure 16-byte stack pointer alignment
        https://git.kernel.org/powerpc/c/fd19a1f72aa7bf687609e0810e644fe5b3846342
[02/13] selftests/powerpc/ptrace: Set LOCAL_HDRS
        https://git.kernel.org/powerpc/c/cf4baafd7846b3def67057a09b7603a6b566417a
[03/13] selftests/powerpc/ptrace: Split CFLAGS better
        https://git.kernel.org/powerpc/c/3c20a1d07c377d7260ca853e216cc85bbd7857fa
[04/13] selftests/powerpc: Don't save CR by default in asm helpers
        https://git.kernel.org/powerpc/c/8f2e02394dc907f5e0140bfab80a9aa11e3449ed
[05/13] selftests/powerpc: Don't save TOC by default in asm helpers
        https://git.kernel.org/powerpc/c/cfbc0723d18f5aeab4308c66d7d1992317eed7c9
[06/13] selftests/powerpc: Add 32-bit support to asm helpers
        https://git.kernel.org/powerpc/c/bd4d3042e7570fc024b5ff15e895363e4bf5a78f
[07/13] selftests/powerpc/ptrace: Drop unused load_fpr_single_precision()
        https://git.kernel.org/powerpc/c/af9f3f31f6cc8e3f637f19189e83d99f3fdd96ad
[08/13] selftests/powerpc/ptrace: Convert to load/store doubles
        https://git.kernel.org/powerpc/c/53fa86e7ece54cbb1fae1443bd6b348088d8ce7e
[09/13] selftests/powerpc/ptrace: Build the ptrace-gpr test as 32-bit when possible
        https://git.kernel.org/powerpc/c/149a497d5fda3e996a00437260a4c170e43909c8
[10/13] selftests/powerpc/ptrace: Do more of ptrace-gpr in asm
        https://git.kernel.org/powerpc/c/611e385087efc2cc3a7033aedd3f84ad0cf2a703
[11/13] selftests/powerpc/ptrace: Make child errors more obvious
        https://git.kernel.org/powerpc/c/7b1513d02edf4a6334618070641f47abbbef0cef
[12/13] selftests/powerpc/ptrace: Use more interesting values
        https://git.kernel.org/powerpc/c/c5a814cc992002c36fa5b7db5fbd55efb7430386
[13/13] selftests/powerpc/ptrace: Add peek/poke of FPRs
        https://git.kernel.org/powerpc/c/6c9c7d8fbc3a2a0cfed0e7a5b39581847b632f0b

cheers