Resguardo inicial

linux-7.2.3/tools/testing/selftests/powerpc/Makefile

@@ -0,0 +1,76 @@
+# SPDX-License-Identifier: GPL-2.0
+# Makefile for powerpc selftests
+
+# ARCH can be overridden by the user for cross compiling
+ARCH ?= $(shell uname -m)
+ARCH := $(shell echo $(ARCH) | sed -e s/ppc.*/powerpc/)
+
+ifeq ($(ARCH),powerpc)
+
+SUB_DIRS = alignment		\
+	   benchmarks		\
+	   cache_shape		\
+	   copyloops		\
+	   dexcr		\
+	   dscr			\
+	   mm			\
+	   nx-gzip		\
+	   pmu			\
+	   signal		\
+	   primitives		\
+	   stringloops		\
+	   switch_endian	\
+	   syscalls		\
+	   tm			\
+	   eeh			\
+	   vphn         \
+	   math		\
+	   papr_attributes	\
+	   papr_vpd		\
+	   papr_sysparm		\
+	   ptrace	\
+	   security	\
+	   mce
+
+endif
+
+all: $(SUB_DIRS)
+
+$(SUB_DIRS):
+	BUILD_TARGET=$(OUTPUT)/$@; mkdir -p $$BUILD_TARGET; $(MAKE) OUTPUT=$$BUILD_TARGET -k -C $@ all
+
+include ../lib.mk
+include ./flags.mk
+
+override define RUN_TESTS
+	+@for TARGET in $(SUB_DIRS); do \
+		BUILD_TARGET=$(OUTPUT)/$$TARGET;	\
+		$(MAKE) OUTPUT=$$BUILD_TARGET -C $$TARGET run_tests;\
+	done;
+endef
+
+override define INSTALL_RULE
+	+@for TARGET in $(SUB_DIRS); do \
+		BUILD_TARGET=$(OUTPUT)/$$TARGET;	\
+		$(MAKE) OUTPUT=$$BUILD_TARGET INSTALL_PATH=$$INSTALL_PATH/$$TARGET -C $$TARGET install;\
+	done;
+endef
+
+emit_tests:
+	+@for TARGET in $(SUB_DIRS); do \
+		BUILD_TARGET=$(OUTPUT)/$$TARGET;	\
+		$(MAKE) OUTPUT=$$BUILD_TARGET COLLECTION=$(COLLECTION)/$$TARGET -s -C $$TARGET $@;\
+	done;
+
+override define CLEAN
+	+@for TARGET in $(SUB_DIRS); do \
+		BUILD_TARGET=$(OUTPUT)/$$TARGET;	\
+		$(MAKE) OUTPUT=$$BUILD_TARGET -C $$TARGET clean; \
+	done;
+	rm -f tags
+endef
+
+tags:
+	find . -name '*.c' -o -name '*.h' | xargs ctags
+
+.PHONY: tags $(SUB_DIRS) emit_tests

linux-7.2.3/tools/testing/selftests/powerpc/alignment/.gitignore

@@ -0,0 +1,3 @@
+# SPDX-License-Identifier: GPL-2.0-only
+copy_first_unaligned
+alignment_handler

linux-7.2.3/tools/testing/selftests/powerpc/alignment/Makefile

@@ -0,0 +1,8 @@
+# SPDX-License-Identifier: GPL-2.0-only
+TEST_GEN_PROGS := copy_first_unaligned alignment_handler
+
+top_srcdir = ../../../../..
+include ../../lib.mk
+include ../flags.mk
+
+$(TEST_GEN_PROGS): ../harness.c ../utils.c

linux-7.2.3/tools/testing/selftests/powerpc/alignment/alignment_handler.c

@@ -0,0 +1,680 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ * Test the powerpc alignment handler on POWER8/POWER9
+ *
+ * Copyright (C) 2017 IBM Corporation (Michael Neuling, Andrew Donnellan)
+ */
+
+/*
+ * This selftest exercises the powerpc alignment fault handler.
+ *
+ * We create two sets of source and destination buffers, one in regular memory,
+ * the other cache-inhibited (by default we use /dev/fb0 for this, but an
+ * alterative path for cache-inhibited memory may be provided, e.g. memtrace).
+ *
+ * We initialise the source buffers, then use whichever set of load/store
+ * instructions is under test to copy bytes from the source buffers to the
+ * destination buffers. For the regular buffers, these instructions will
+ * execute normally. For the cache-inhibited buffers, these instructions
+ * will trap and cause an alignment fault, and the alignment fault handler
+ * will emulate the particular instruction under test. We then compare the
+ * destination buffers to ensure that the native and emulated cases give the
+ * same result.
+ *
+ * TODO:
+ *   - Any FIXMEs below
+ *   - Test VSX regs < 32 and > 32
+ *   - Test all loads and stores
+ *   - Check update forms do update register
+ *   - Test alignment faults over page boundary
+ *
+ * Some old binutils may not support all the instructions.
+ */
+
+
+#include <sys/mman.h>
+#include <sys/types.h>
+#include <sys/stat.h>
+#include <fcntl.h>
+#include <unistd.h>
+#include <stdbool.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include <assert.h>
+#include <getopt.h>
+#include <setjmp.h>
+#include <signal.h>
+
+#include "utils.h"
+#include "instructions.h"
+
+int bufsize;
+int debug;
+int testing;
+volatile int gotsig;
+bool prefixes_enabled;
+char *cipath = "/dev/fb0";
+long cioffset;
+
+void sighandler(int sig, siginfo_t *info, void *ctx)
+{
+	ucontext_t *ucp = ctx;
+
+	if (!testing) {
+		signal(sig, SIG_DFL);
+		kill(0, sig);
+	}
+	gotsig = sig;
+#ifdef __powerpc64__
+	if (prefixes_enabled) {
+		u32 inst = *(u32 *)ucp->uc_mcontext.gp_regs[PT_NIP];
+		ucp->uc_mcontext.gp_regs[PT_NIP] += ((inst >> 26 == 1) ? 8 : 4);
+	} else {
+		ucp->uc_mcontext.gp_regs[PT_NIP] += 4;
+	}
+#else
+	ucp->uc_mcontext.uc_regs->gregs[PT_NIP] += 4;
+#endif
+}
+
+#define XFORM(reg, n)  " " #reg " ,%"#n",%2 ;"
+#define DFORM(reg, n)  " " #reg " ,0(%"#n") ;"
+
+#define TEST(name, ld_op, st_op, form, ld_reg, st_reg)		\
+	void test_##name(char *s, char *d)			\
+	{							\
+		asm volatile(					\
+			#ld_op form(ld_reg, 0)			\
+			#st_op form(st_reg, 1)			\
+			:: "r"(s), "r"(d), "r"(0)		\
+			: "memory", "vs0", "vs32", "r31");	\
+	}							\
+	rc |= do_test(#name, test_##name)
+
+#define TESTP(name, ld_op, st_op, ld_reg, st_reg)		\
+	void test_##name(char *s, char *d)			\
+	{							\
+		asm volatile(					\
+			ld_op(ld_reg, %0, 0, 0)			\
+			st_op(st_reg, %1, 0, 0)			\
+			:: "r"(s), "r"(d), "r"(0)		\
+			: "memory", "vs0", "vs32", "r31");	\
+	}							\
+	rc |= do_test(#name, test_##name)
+
+#define LOAD_VSX_XFORM_TEST(op) TEST(op, op, stxvd2x, XFORM, 32, 32)
+#define STORE_VSX_XFORM_TEST(op) TEST(op, lxvd2x, op, XFORM, 32, 32)
+#define LOAD_VSX_DFORM_TEST(op) TEST(op, op, stxv, DFORM, 32, 32)
+#define STORE_VSX_DFORM_TEST(op) TEST(op, lxv, op, DFORM, 32, 32)
+#define LOAD_VMX_XFORM_TEST(op) TEST(op, op, stxvd2x, XFORM, 0, 32)
+#define STORE_VMX_XFORM_TEST(op) TEST(op, lxvd2x, op, XFORM, 32, 0)
+#define LOAD_VMX_DFORM_TEST(op) TEST(op, op, stxv, DFORM, 0, 32)
+#define STORE_VMX_DFORM_TEST(op) TEST(op, lxv, op, DFORM, 32, 0)
+
+#define LOAD_XFORM_TEST(op) TEST(op, op, stdx, XFORM, 31, 31)
+#define STORE_XFORM_TEST(op) TEST(op, ldx, op, XFORM, 31, 31)
+#define LOAD_DFORM_TEST(op) TEST(op, op, std, DFORM, 31, 31)
+#define STORE_DFORM_TEST(op) TEST(op, ld, op, DFORM, 31, 31)
+
+#define LOAD_FLOAT_DFORM_TEST(op)  TEST(op, op, stfd, DFORM, 0, 0)
+#define STORE_FLOAT_DFORM_TEST(op) TEST(op, lfd, op, DFORM, 0, 0)
+#define LOAD_FLOAT_XFORM_TEST(op)  TEST(op, op, stfdx, XFORM, 0, 0)
+#define STORE_FLOAT_XFORM_TEST(op) TEST(op, lfdx, op, XFORM, 0, 0)
+
+#define LOAD_MLS_PREFIX_TEST(op) TESTP(op, op, PSTD, 31, 31)
+#define STORE_MLS_PREFIX_TEST(op) TESTP(op, PLD, op, 31, 31)
+
+#define LOAD_8LS_PREFIX_TEST(op) TESTP(op, op, PSTD, 31, 31)
+#define STORE_8LS_PREFIX_TEST(op) TESTP(op, PLD, op, 31, 31)
+
+#define LOAD_FLOAT_MLS_PREFIX_TEST(op) TESTP(op, op, PSTFD, 0, 0)
+#define STORE_FLOAT_MLS_PREFIX_TEST(op) TESTP(op, PLFD, op, 0, 0)
+
+#define LOAD_VSX_8LS_PREFIX_TEST(op, tail) TESTP(op, op, PSTXV ## tail, 0, 32)
+#define STORE_VSX_8LS_PREFIX_TEST(op, tail) TESTP(op, PLXV ## tail, op, 32, 0)
+
+/* FIXME: Unimplemented tests: */
+// STORE_DFORM_TEST(stq)   /* FIXME: need two registers for quad */
+// STORE_DFORM_TEST(stswi) /* FIXME: string instruction */
+
+// STORE_XFORM_TEST(stwat) /* AMO can't emulate or run on CI */
+// STORE_XFORM_TEST(stdat) /* ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ */
+
+
+/* preload byte by byte */
+void preload_data(void *dst, int offset, int width)
+{
+	char *c = dst;
+	int i;
+
+	c += offset;
+
+	for (i = 0 ; i < width ; i++)
+		c[i] = i;
+}
+
+int test_memcpy(void *dst, void *src, int size, int offset,
+		void (*test_func)(char *, char *))
+{
+	char *s, *d;
+
+	s = src;
+	s += offset;
+	d = dst;
+	d += offset;
+
+	assert(size == 16);
+	gotsig = 0;
+	testing = 1;
+
+	test_func(s, d); /* run the actual test */
+
+	testing = 0;
+	if (gotsig) {
+		if (debug)
+			printf("  Got signal %i\n", gotsig);
+		return 1;
+	}
+	return 0;
+}
+
+void dumpdata(char *s1, char *s2, int n, char *test_name)
+{
+	int i;
+
+	printf("  %s: unexpected result:\n", test_name);
+	printf("    mem:");
+	for (i = 0; i < n; i++)
+		printf(" %02x", s1[i]);
+	printf("\n");
+	printf("    ci: ");
+	for (i = 0; i < n; i++)
+		printf(" %02x", s2[i]);
+	printf("\n");
+}
+
+int test_memcmp(void *s1, void *s2, int n, int offset, char *test_name)
+{
+	char *s1c, *s2c;
+
+	s1c = s1;
+	s1c += offset;
+	s2c = s2;
+	s2c += offset;
+
+	if (memcmp(s1c, s2c, n)) {
+		if (debug) {
+			printf("\n  Compare failed. Offset:%i length:%i\n",
+			       offset, n);
+			dumpdata(s1c, s2c, n, test_name);
+		}
+		return 1;
+	}
+	return 0;
+}
+
+/*
+ * Do two memcpy tests using the same instructions. One cachable
+ * memory and the other doesn't.
+ */
+int do_test(char *test_name, void (*test_func)(char *, char *))
+{
+	int offset, width, fd, rc, r;
+	void *mem0, *mem1, *ci0, *ci1;
+
+	printf("\tDoing %s:\t", test_name);
+
+	fd = open(cipath, O_RDWR);
+	if (fd < 0) {
+		printf("\n");
+		perror("Can't open ci file now?");
+		return 1;
+	}
+
+	ci0 = mmap(NULL, bufsize, PROT_WRITE | PROT_READ, MAP_SHARED,
+		   fd, cioffset);
+	ci1 = mmap(NULL, bufsize, PROT_WRITE | PROT_READ, MAP_SHARED,
+		   fd, cioffset + bufsize);
+
+	if ((ci0 == MAP_FAILED) || (ci1 == MAP_FAILED)) {
+		printf("\n");
+		perror("mmap failed");
+		SKIP_IF(1);
+	}
+
+	rc = posix_memalign(&mem0, bufsize, bufsize);
+	if (rc) {
+		printf("\n");
+		return rc;
+	}
+
+	rc = posix_memalign(&mem1, bufsize, bufsize);
+	if (rc) {
+		printf("\n");
+		free(mem0);
+		return rc;
+	}
+
+	rc = 0;
+	/*
+	 * offset = 0 is aligned but tests the workaround for the P9N
+	 * DD2.1 vector CI load issue (see 5080332c2c89 "powerpc/64s:
+	 * Add workaround for P9 vector CI load issue")
+	 */
+	for (offset = 0; offset < 16; offset++) {
+		width = 16; /* vsx == 16 bytes */
+		r = 0;
+
+		/* load pattern into memory byte by byte */
+		preload_data(ci0, offset, width);
+		preload_data(mem0, offset, width); // FIXME: remove??
+		memcpy(ci0, mem0, bufsize);
+		memcpy(ci1, mem1, bufsize); /* initialise output to the same */
+
+		/* sanity check */
+		test_memcmp(mem0, ci0, width, offset, test_name);
+
+		r |= test_memcpy(ci1,  ci0,  width, offset, test_func);
+		r |= test_memcpy(mem1, mem0, width, offset, test_func);
+		if (r && !debug) {
+			printf("FAILED: Got signal");
+			rc = 1;
+			break;
+		}
+
+		r |= test_memcmp(mem1, ci1, width, offset, test_name);
+		if (r && !debug) {
+			printf("FAILED: Wrong Data");
+			rc = 1;
+			break;
+		}
+	}
+
+	if (rc == 0)
+		printf("PASSED");
+
+	printf("\n");
+
+	munmap(ci0, bufsize);
+	munmap(ci1, bufsize);
+	free(mem0);
+	free(mem1);
+	close(fd);
+
+	return rc;
+}
+
+static bool can_open_cifile(void)
+{
+	int fd;
+
+	fd = open(cipath, O_RDWR);
+	if (fd < 0)
+		return false;
+
+	close(fd);
+	return true;
+}
+
+int test_alignment_handler_vsx_206(void)
+{
+	int rc = 0;
+
+	SKIP_IF(!can_open_cifile());
+	SKIP_IF(!have_hwcap(PPC_FEATURE_ARCH_2_06));
+
+	printf("VSX: 2.06B\n");
+	LOAD_VSX_XFORM_TEST(lxvd2x);
+	LOAD_VSX_XFORM_TEST(lxvw4x);
+	LOAD_VSX_XFORM_TEST(lxsdx);
+	LOAD_VSX_XFORM_TEST(lxvdsx);
+	STORE_VSX_XFORM_TEST(stxvd2x);
+	STORE_VSX_XFORM_TEST(stxvw4x);
+	STORE_VSX_XFORM_TEST(stxsdx);
+	return rc;
+}
+
+int test_alignment_handler_vsx_207(void)
+{
+	int rc = 0;
+
+	SKIP_IF(!can_open_cifile());
+	SKIP_IF(!have_hwcap2(PPC_FEATURE2_ARCH_2_07));
+
+	printf("VSX: 2.07B\n");
+	LOAD_VSX_XFORM_TEST(lxsspx);
+	LOAD_VSX_XFORM_TEST(lxsiwax);
+	LOAD_VSX_XFORM_TEST(lxsiwzx);
+	STORE_VSX_XFORM_TEST(stxsspx);
+	STORE_VSX_XFORM_TEST(stxsiwx);
+	return rc;
+}
+
+int test_alignment_handler_vsx_300(void)
+{
+	int rc = 0;
+
+	SKIP_IF(!can_open_cifile());
+
+	SKIP_IF(!have_hwcap2(PPC_FEATURE2_ARCH_3_00));
+	printf("VSX: 3.00B\n");
+	LOAD_VMX_DFORM_TEST(lxsd);
+	LOAD_VSX_XFORM_TEST(lxsibzx);
+	LOAD_VSX_XFORM_TEST(lxsihzx);
+	LOAD_VMX_DFORM_TEST(lxssp);
+	LOAD_VSX_DFORM_TEST(lxv);
+	LOAD_VSX_XFORM_TEST(lxvb16x);
+	LOAD_VSX_XFORM_TEST(lxvh8x);
+	LOAD_VSX_XFORM_TEST(lxvx);
+	LOAD_VSX_XFORM_TEST(lxvwsx);
+	LOAD_VSX_XFORM_TEST(lxvl);
+	LOAD_VSX_XFORM_TEST(lxvll);
+	STORE_VMX_DFORM_TEST(stxsd);
+	STORE_VSX_XFORM_TEST(stxsibx);
+	STORE_VSX_XFORM_TEST(stxsihx);
+	STORE_VMX_DFORM_TEST(stxssp);
+	STORE_VSX_DFORM_TEST(stxv);
+	STORE_VSX_XFORM_TEST(stxvb16x);
+	STORE_VSX_XFORM_TEST(stxvh8x);
+	STORE_VSX_XFORM_TEST(stxvx);
+	STORE_VSX_XFORM_TEST(stxvl);
+	STORE_VSX_XFORM_TEST(stxvll);
+	return rc;
+}
+
+int test_alignment_handler_vsx_prefix(void)
+{
+	int rc = 0;
+
+	SKIP_IF(!can_open_cifile());
+	SKIP_IF(!have_hwcap2(PPC_FEATURE2_ARCH_3_1));
+
+	printf("VSX: PREFIX\n");
+	LOAD_VSX_8LS_PREFIX_TEST(PLXSD, 0);
+	LOAD_VSX_8LS_PREFIX_TEST(PLXSSP, 0);
+	LOAD_VSX_8LS_PREFIX_TEST(PLXV0, 0);
+	LOAD_VSX_8LS_PREFIX_TEST(PLXV1, 1);
+	STORE_VSX_8LS_PREFIX_TEST(PSTXSD, 0);
+	STORE_VSX_8LS_PREFIX_TEST(PSTXSSP, 0);
+	STORE_VSX_8LS_PREFIX_TEST(PSTXV0, 0);
+	STORE_VSX_8LS_PREFIX_TEST(PSTXV1, 1);
+	return rc;
+}
+
+int test_alignment_handler_integer(void)
+{
+	int rc = 0;
+
+	SKIP_IF(!can_open_cifile());
+
+	printf("Integer\n");
+	LOAD_DFORM_TEST(lbz);
+	LOAD_DFORM_TEST(lbzu);
+	LOAD_XFORM_TEST(lbzx);
+	LOAD_XFORM_TEST(lbzux);
+	LOAD_DFORM_TEST(lhz);
+	LOAD_DFORM_TEST(lhzu);
+	LOAD_XFORM_TEST(lhzx);
+	LOAD_XFORM_TEST(lhzux);
+	LOAD_DFORM_TEST(lha);
+	LOAD_DFORM_TEST(lhau);
+	LOAD_XFORM_TEST(lhax);
+	LOAD_XFORM_TEST(lhaux);
+	LOAD_XFORM_TEST(lhbrx);
+	LOAD_DFORM_TEST(lwz);
+	LOAD_DFORM_TEST(lwzu);
+	LOAD_XFORM_TEST(lwzx);
+	LOAD_XFORM_TEST(lwzux);
+	LOAD_DFORM_TEST(lwa);
+	LOAD_XFORM_TEST(lwax);
+	LOAD_XFORM_TEST(lwaux);
+	LOAD_XFORM_TEST(lwbrx);
+	LOAD_DFORM_TEST(ld);
+	LOAD_DFORM_TEST(ldu);
+	LOAD_XFORM_TEST(ldx);
+	LOAD_XFORM_TEST(ldux);
+	STORE_DFORM_TEST(stb);
+	STORE_XFORM_TEST(stbx);
+	STORE_DFORM_TEST(stbu);
+	STORE_XFORM_TEST(stbux);
+	STORE_DFORM_TEST(sth);
+	STORE_XFORM_TEST(sthx);
+	STORE_DFORM_TEST(sthu);
+	STORE_XFORM_TEST(sthux);
+	STORE_XFORM_TEST(sthbrx);
+	STORE_DFORM_TEST(stw);
+	STORE_XFORM_TEST(stwx);
+	STORE_DFORM_TEST(stwu);
+	STORE_XFORM_TEST(stwux);
+	STORE_XFORM_TEST(stwbrx);
+	STORE_DFORM_TEST(std);
+	STORE_XFORM_TEST(stdx);
+	STORE_DFORM_TEST(stdu);
+	STORE_XFORM_TEST(stdux);
+
+#ifdef __BIG_ENDIAN__
+	LOAD_DFORM_TEST(lmw);
+	STORE_DFORM_TEST(stmw);
+#endif
+
+	return rc;
+}
+
+int test_alignment_handler_integer_206(void)
+{
+	int rc = 0;
+
+	SKIP_IF(!can_open_cifile());
+	SKIP_IF(!have_hwcap(PPC_FEATURE_ARCH_2_06));
+
+	printf("Integer: 2.06\n");
+
+	LOAD_XFORM_TEST(ldbrx);
+	STORE_XFORM_TEST(stdbrx);
+
+	return rc;
+}
+
+int test_alignment_handler_integer_prefix(void)
+{
+	int rc = 0;
+
+	SKIP_IF(!can_open_cifile());
+	SKIP_IF(!have_hwcap2(PPC_FEATURE2_ARCH_3_1));
+
+	printf("Integer: PREFIX\n");
+	LOAD_MLS_PREFIX_TEST(PLBZ);
+	LOAD_MLS_PREFIX_TEST(PLHZ);
+	LOAD_MLS_PREFIX_TEST(PLHA);
+	LOAD_MLS_PREFIX_TEST(PLWZ);
+	LOAD_8LS_PREFIX_TEST(PLWA);
+	LOAD_8LS_PREFIX_TEST(PLD);
+	STORE_MLS_PREFIX_TEST(PSTB);
+	STORE_MLS_PREFIX_TEST(PSTH);
+	STORE_MLS_PREFIX_TEST(PSTW);
+	STORE_8LS_PREFIX_TEST(PSTD);
+	return rc;
+}
+
+int test_alignment_handler_vmx(void)
+{
+	int rc = 0;
+
+	SKIP_IF(!can_open_cifile());
+	SKIP_IF(!have_hwcap(PPC_FEATURE_HAS_ALTIVEC));
+
+	printf("VMX\n");
+	LOAD_VMX_XFORM_TEST(lvx);
+
+	/*
+	 * FIXME: These loads only load part of the register, so our
+	 * testing method doesn't work. Also they don't take alignment
+	 * faults, so it's kinda pointless anyway
+	 *
+	 LOAD_VMX_XFORM_TEST(lvebx)
+	 LOAD_VMX_XFORM_TEST(lvehx)
+	 LOAD_VMX_XFORM_TEST(lvewx)
+	 LOAD_VMX_XFORM_TEST(lvxl)
+	*/
+	STORE_VMX_XFORM_TEST(stvx);
+	STORE_VMX_XFORM_TEST(stvebx);
+	STORE_VMX_XFORM_TEST(stvehx);
+	STORE_VMX_XFORM_TEST(stvewx);
+	STORE_VMX_XFORM_TEST(stvxl);
+	return rc;
+}
+
+int test_alignment_handler_fp(void)
+{
+	int rc = 0;
+
+	SKIP_IF(!can_open_cifile());
+
+	printf("Floating point\n");
+	LOAD_FLOAT_DFORM_TEST(lfd);
+	LOAD_FLOAT_XFORM_TEST(lfdx);
+	LOAD_FLOAT_DFORM_TEST(lfdu);
+	LOAD_FLOAT_XFORM_TEST(lfdux);
+	LOAD_FLOAT_DFORM_TEST(lfs);
+	LOAD_FLOAT_XFORM_TEST(lfsx);
+	LOAD_FLOAT_DFORM_TEST(lfsu);
+	LOAD_FLOAT_XFORM_TEST(lfsux);
+	STORE_FLOAT_DFORM_TEST(stfd);
+	STORE_FLOAT_XFORM_TEST(stfdx);
+	STORE_FLOAT_DFORM_TEST(stfdu);
+	STORE_FLOAT_XFORM_TEST(stfdux);
+	STORE_FLOAT_DFORM_TEST(stfs);
+	STORE_FLOAT_XFORM_TEST(stfsx);
+	STORE_FLOAT_DFORM_TEST(stfsu);
+	STORE_FLOAT_XFORM_TEST(stfsux);
+	STORE_FLOAT_XFORM_TEST(stfiwx);
+
+	return rc;
+}
+
+int test_alignment_handler_fp_205(void)
+{
+	int rc = 0;
+
+	SKIP_IF(!can_open_cifile());
+	SKIP_IF(!have_hwcap(PPC_FEATURE_ARCH_2_05));
+
+	printf("Floating point: 2.05\n");
+
+	LOAD_FLOAT_DFORM_TEST(lfdp);
+	LOAD_FLOAT_XFORM_TEST(lfdpx);
+	LOAD_FLOAT_XFORM_TEST(lfiwax);
+	STORE_FLOAT_DFORM_TEST(stfdp);
+	STORE_FLOAT_XFORM_TEST(stfdpx);
+
+	return rc;
+}
+
+int test_alignment_handler_fp_206(void)
+{
+	int rc = 0;
+
+	SKIP_IF(!can_open_cifile());
+	SKIP_IF(!have_hwcap(PPC_FEATURE_ARCH_2_06));
+
+	printf("Floating point: 2.06\n");
+
+	LOAD_FLOAT_XFORM_TEST(lfiwzx);
+
+	return rc;
+}
+
+
+int test_alignment_handler_fp_prefix(void)
+{
+	int rc = 0;
+
+	SKIP_IF(!can_open_cifile());
+	SKIP_IF(!have_hwcap2(PPC_FEATURE2_ARCH_3_1));
+
+	printf("Floating point: PREFIX\n");
+	LOAD_FLOAT_DFORM_TEST(lfs);
+	LOAD_FLOAT_MLS_PREFIX_TEST(PLFS);
+	LOAD_FLOAT_MLS_PREFIX_TEST(PLFD);
+	STORE_FLOAT_MLS_PREFIX_TEST(PSTFS);
+	STORE_FLOAT_MLS_PREFIX_TEST(PSTFD);
+	return rc;
+}
+
+void usage(char *prog)
+{
+	printf("Usage: %s [options] [path [offset]]\n", prog);
+	printf("  -d	Enable debug error output\n");
+	printf("\n");
+	printf("This test requires a POWER8, POWER9 or POWER10 CPU ");
+	printf("and either a usable framebuffer at /dev/fb0 or ");
+	printf("the path to usable cache inhibited memory and optional ");
+	printf("offset to be provided\n");
+}
+
+int main(int argc, char *argv[])
+{
+
+	struct sigaction sa;
+	int rc = 0;
+	int option = 0;
+
+	while ((option = getopt(argc, argv, "d")) != -1) {
+		switch (option) {
+		case 'd':
+			debug++;
+			break;
+		default:
+			usage(argv[0]);
+			exit(1);
+		}
+	}
+	argc -= optind;
+	argv += optind;
+
+	if (argc > 0)
+		cipath = argv[0];
+	if (argc > 1)
+		cioffset = strtol(argv[1], 0, 0x10);
+
+	bufsize = getpagesize();
+
+	sa.sa_sigaction = sighandler;
+	sigemptyset(&sa.sa_mask);
+	sa.sa_flags = SA_SIGINFO;
+	if (sigaction(SIGSEGV, &sa, NULL) == -1
+	    || sigaction(SIGBUS, &sa, NULL) == -1
+	    || sigaction(SIGILL, &sa, NULL) == -1) {
+		perror("sigaction");
+		exit(1);
+	}
+
+	prefixes_enabled = have_hwcap2(PPC_FEATURE2_ARCH_3_1);
+
+	rc |= test_harness(test_alignment_handler_vsx_206,
+			   "test_alignment_handler_vsx_206");
+	rc |= test_harness(test_alignment_handler_vsx_207,
+			   "test_alignment_handler_vsx_207");
+	rc |= test_harness(test_alignment_handler_vsx_300,
+			   "test_alignment_handler_vsx_300");
+	rc |= test_harness(test_alignment_handler_vsx_prefix,
+			   "test_alignment_handler_vsx_prefix");
+	rc |= test_harness(test_alignment_handler_integer,
+			   "test_alignment_handler_integer");
+	rc |= test_harness(test_alignment_handler_integer_206,
+			   "test_alignment_handler_integer_206");
+	rc |= test_harness(test_alignment_handler_integer_prefix,
+			   "test_alignment_handler_integer_prefix");
+	rc |= test_harness(test_alignment_handler_vmx,
+			   "test_alignment_handler_vmx");
+	rc |= test_harness(test_alignment_handler_fp,
+			   "test_alignment_handler_fp");
+	rc |= test_harness(test_alignment_handler_fp_205,
+			   "test_alignment_handler_fp_205");
+	rc |= test_harness(test_alignment_handler_fp_206,
+			   "test_alignment_handler_fp_206");
+	rc |= test_harness(test_alignment_handler_fp_prefix,
+			   "test_alignment_handler_fp_prefix");
+	return rc;
+}

linux-7.2.3/tools/testing/selftests/powerpc/alignment/copy_first_unaligned.c

@@ -0,0 +1,67 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ * Copyright 2016, Chris Smart, IBM Corporation.
+ *
+ * Calls to copy_first which are not 128-byte aligned should be
+ * caught and sent a SIGBUS.
+ */
+
+#include <signal.h>
+#include <string.h>
+#include <unistd.h>
+#include "utils.h"
+#include "instructions.h"
+
+unsigned int expected_instruction = PPC_INST_COPY_FIRST;
+unsigned int instruction_mask = 0xfc2007fe;
+
+void signal_action_handler(int signal_num, siginfo_t *info, void *ptr)
+{
+	ucontext_t *ctx = ptr;
+#ifdef __powerpc64__
+	unsigned int *pc = (unsigned int *)ctx->uc_mcontext.gp_regs[PT_NIP];
+#else
+	unsigned int *pc = (unsigned int *)ctx->uc_mcontext.uc_regs->gregs[PT_NIP];
+#endif
+
+	/*
+	 * Check that the signal was on the correct instruction, using a
+	 * mask because the compiler assigns the register at RB.
+	 */
+	if ((*pc & instruction_mask) == expected_instruction)
+		_exit(0); /* We hit the right instruction */
+
+	_exit(1);
+}
+
+void setup_signal_handler(void)
+{
+	struct sigaction signal_action;
+
+	memset(&signal_action, 0, sizeof(signal_action));
+	signal_action.sa_sigaction = signal_action_handler;
+	signal_action.sa_flags = SA_SIGINFO;
+	sigaction(SIGBUS, &signal_action, NULL);
+}
+
+char cacheline_buf[128] __cacheline_aligned;
+
+int test_copy_first_unaligned(void)
+{
+	/* Only run this test on a P9 or later */
+	SKIP_IF(!have_hwcap2(PPC_FEATURE2_ARCH_3_00));
+
+	/* Register our signal handler with SIGBUS */
+	setup_signal_handler();
+
+	/* +1 makes buf unaligned */
+	copy_first(cacheline_buf+1);
+
+	/* We should not get here */
+	return 1;
+}
+
+int main(int argc, char *argv[])
+{
+	return test_harness(test_copy_first_unaligned, "test_copy_first_unaligned");
+}

linux-7.2.3/tools/testing/selftests/powerpc/alignment/settings

@@ -0,0 +1 @@
+timeout=130

linux-7.2.3/tools/testing/selftests/powerpc/benchmarks/.gitignore

@@ -0,0 +1,8 @@
+# SPDX-License-Identifier: GPL-2.0-only
+gettimeofday
+context_switch
+fork
+exec_target
+mmap_bench
+futex_bench
+null_syscall

linux-7.2.3/tools/testing/selftests/powerpc/benchmarks/Makefile

@@ -0,0 +1,21 @@
+# SPDX-License-Identifier: GPL-2.0
+TEST_GEN_PROGS := gettimeofday context_switch fork mmap_bench futex_bench null_syscall
+TEST_GEN_FILES := exec_target
+
+TEST_FILES := settings
+
+top_srcdir = ../../../../..
+include ../../lib.mk
+include ../flags.mk
+
+CFLAGS += -O2
+
+$(TEST_GEN_PROGS): ../harness.c
+
+$(OUTPUT)/context_switch: ../utils.c
+$(OUTPUT)/context_switch: CFLAGS += -maltivec -mvsx -mabi=altivec
+$(OUTPUT)/context_switch: LDLIBS += -lpthread
+
+$(OUTPUT)/fork: LDLIBS += -lpthread
+
+$(OUTPUT)/exec_target: CFLAGS += -nostartfiles

linux-7.2.3/tools/testing/selftests/powerpc/benchmarks/context_switch.c

@@ -0,0 +1,508 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ * Context switch microbenchmark.
+ *
+ * Copyright (C) 2015 Anton Blanchard <anton@au.ibm.com>, IBM
+ */
+
+#define _GNU_SOURCE
+#include <errno.h>
+#include <sched.h>
+#include <string.h>
+#include <stdio.h>
+#include <unistd.h>
+#include <stdlib.h>
+#include <getopt.h>
+#include <signal.h>
+#include <assert.h>
+#include <pthread.h>
+#include <limits.h>
+#include <sys/time.h>
+#include <sys/syscall.h>
+#include <sys/sysinfo.h>
+#include <sys/types.h>
+#include <sys/shm.h>
+#include <linux/futex.h>
+#ifdef __powerpc__
+#include <altivec.h>
+#endif
+#include "utils.h"
+
+static unsigned int timeout = 30;
+
+static int touch_vdso;
+struct timeval tv;
+
+static int touch_fp = 1;
+double fp;
+
+static int touch_vector = 1;
+vector int a, b, c;
+
+#ifdef __powerpc__
+static int touch_altivec = 1;
+
+/*
+ * Note: LTO (Link Time Optimisation) doesn't play well with this function
+ * attribute. Be very careful enabling LTO for this test.
+ */
+static void __attribute__((__target__("no-vsx"))) altivec_touch_fn(void)
+{
+	c = a + b;
+}
+#endif
+
+static void touch(void)
+{
+	if (touch_vdso)
+		gettimeofday(&tv, NULL);
+
+	if (touch_fp)
+		fp += 0.1;
+
+#ifdef __powerpc__
+	if (touch_altivec)
+		altivec_touch_fn();
+#endif
+
+	if (touch_vector)
+		c = a + b;
+
+	asm volatile("# %0 %1 %2": : "r"(&tv), "r"(&fp), "r"(&c));
+}
+
+static void start_thread_on(void *(*fn)(void *), void *arg, unsigned long cpu)
+{
+	int rc;
+	pthread_t tid;
+	cpu_set_t cpuset;
+	pthread_attr_t attr;
+
+	CPU_ZERO(&cpuset);
+	CPU_SET(cpu, &cpuset);
+
+	rc = pthread_attr_init(&attr);
+	if (rc) {
+		errno = rc;
+		perror("pthread_attr_init");
+		exit(1);
+	}
+
+	rc = pthread_attr_setaffinity_np(&attr, sizeof(cpu_set_t), &cpuset);
+	if (rc)	{
+		errno = rc;
+		perror("pthread_attr_setaffinity_np");
+		exit(1);
+	}
+
+	rc = pthread_create(&tid, &attr, fn, arg);
+	if (rc) {
+		errno = rc;
+		perror("pthread_create");
+		exit(1);
+	}
+}
+
+static void start_process_on(void *(*fn)(void *), void *arg, unsigned long cpu)
+{
+	int pid, ncpus;
+	cpu_set_t *cpuset;
+	size_t size;
+
+	pid = fork();
+	if (pid == -1) {
+		perror("fork");
+		exit(1);
+	}
+
+	if (pid)
+		return;
+
+	ncpus = get_nprocs();
+	size = CPU_ALLOC_SIZE(ncpus);
+	cpuset = CPU_ALLOC(ncpus);
+	if (!cpuset) {
+		perror("malloc");
+		exit(1);
+	}
+	CPU_ZERO_S(size, cpuset);
+	CPU_SET_S(cpu, size, cpuset);
+
+	if (sched_setaffinity(0, size, cpuset)) {
+		perror("sched_setaffinity");
+		CPU_FREE(cpuset);
+		exit(1);
+	}
+
+	CPU_FREE(cpuset);
+	fn(arg);
+
+	exit(0);
+}
+
+static unsigned long iterations;
+static unsigned long iterations_prev;
+
+static void sigalrm_handler(int junk)
+{
+	unsigned long i = iterations;
+
+	printf("%ld\n", i - iterations_prev);
+	iterations_prev = i;
+
+	if (--timeout == 0)
+		kill(0, SIGUSR1);
+
+	alarm(1);
+}
+
+static void sigusr1_handler(int junk)
+{
+	exit(0);
+}
+
+struct actions {
+	void (*setup)(int, int);
+	void *(*thread1)(void *);
+	void *(*thread2)(void *);
+};
+
+#define READ 0
+#define WRITE 1
+
+static int pipe_fd1[2];
+static int pipe_fd2[2];
+
+static void pipe_setup(int cpu1, int cpu2)
+{
+	if (pipe(pipe_fd1) || pipe(pipe_fd2))
+		exit(1);
+}
+
+static void *pipe_thread1(void *arg)
+{
+	signal(SIGALRM, sigalrm_handler);
+	alarm(1);
+
+	while (1) {
+		assert(read(pipe_fd1[READ], &c, 1) == 1);
+		touch();
+
+		assert(write(pipe_fd2[WRITE], &c, 1) == 1);
+		touch();
+
+		iterations += 2;
+	}
+
+	return NULL;
+}
+
+static void *pipe_thread2(void *arg)
+{
+	while (1) {
+		assert(write(pipe_fd1[WRITE], &c, 1) == 1);
+		touch();
+
+		assert(read(pipe_fd2[READ], &c, 1) == 1);
+		touch();
+	}
+
+	return NULL;
+}
+
+static struct actions pipe_actions = {
+	.setup = pipe_setup,
+	.thread1 = pipe_thread1,
+	.thread2 = pipe_thread2,
+};
+
+static void yield_setup(int cpu1, int cpu2)
+{
+	if (cpu1 != cpu2) {
+		fprintf(stderr, "Both threads must be on the same CPU for yield test\n");
+		exit(1);
+	}
+}
+
+static void *yield_thread1(void *arg)
+{
+	signal(SIGALRM, sigalrm_handler);
+	alarm(1);
+
+	while (1) {
+		sched_yield();
+		touch();
+
+		iterations += 2;
+	}
+
+	return NULL;
+}
+
+static void *yield_thread2(void *arg)
+{
+	while (1) {
+		sched_yield();
+		touch();
+	}
+
+	return NULL;
+}
+
+static struct actions yield_actions = {
+	.setup = yield_setup,
+	.thread1 = yield_thread1,
+	.thread2 = yield_thread2,
+};
+
+static long sys_futex(void *addr1, int op, int val1, struct timespec *timeout,
+		      void *addr2, int val3)
+{
+	return syscall(SYS_futex, addr1, op, val1, timeout, addr2, val3);
+}
+
+static unsigned long cmpxchg(unsigned long *p, unsigned long expected,
+			     unsigned long desired)
+{
+	unsigned long exp = expected;
+
+	__atomic_compare_exchange_n(p, &exp, desired, 0,
+				    __ATOMIC_SEQ_CST, __ATOMIC_SEQ_CST);
+	return exp;
+}
+
+static unsigned long xchg(unsigned long *p, unsigned long val)
+{
+	return __atomic_exchange_n(p, val, __ATOMIC_SEQ_CST);
+}
+
+static int processes;
+
+static int mutex_lock(unsigned long *m)
+{
+	int c;
+	int flags = FUTEX_WAIT;
+	if (!processes)
+		flags |= FUTEX_PRIVATE_FLAG;
+
+	c = cmpxchg(m, 0, 1);
+	if (!c)
+		return 0;
+
+	if (c == 1)
+		c = xchg(m, 2);
+
+	while (c) {
+		sys_futex(m, flags, 2, NULL, NULL, 0);
+		c = xchg(m, 2);
+	}
+
+	return 0;
+}
+
+static int mutex_unlock(unsigned long *m)
+{
+	int flags = FUTEX_WAKE;
+	if (!processes)
+		flags |= FUTEX_PRIVATE_FLAG;
+
+	if (*m == 2)
+		*m = 0;
+	else if (xchg(m, 0) == 1)
+		return 0;
+
+	sys_futex(m, flags, 1, NULL, NULL, 0);
+
+	return 0;
+}
+
+static unsigned long *m1, *m2;
+
+static void futex_setup(int cpu1, int cpu2)
+{
+	if (!processes) {
+		static unsigned long _m1, _m2;
+		m1 = &_m1;
+		m2 = &_m2;
+	} else {
+		int shmid;
+		void *shmaddr;
+
+		shmid = shmget(IPC_PRIVATE, getpagesize(), SHM_R | SHM_W);
+		if (shmid < 0) {
+			perror("shmget");
+			exit(1);
+		}
+
+		shmaddr = shmat(shmid, NULL, 0);
+		if (shmaddr == (char *)-1) {
+			perror("shmat");
+			shmctl(shmid, IPC_RMID, NULL);
+			exit(1);
+		}
+
+		shmctl(shmid, IPC_RMID, NULL);
+
+		m1 = shmaddr;
+		m2 = shmaddr + sizeof(*m1);
+	}
+
+	*m1 = 0;
+	*m2 = 0;
+
+	mutex_lock(m1);
+	mutex_lock(m2);
+}
+
+static void *futex_thread1(void *arg)
+{
+	signal(SIGALRM, sigalrm_handler);
+	alarm(1);
+
+	while (1) {
+		mutex_lock(m2);
+		mutex_unlock(m1);
+
+		iterations += 2;
+	}
+
+	return NULL;
+}
+
+static void *futex_thread2(void *arg)
+{
+	while (1) {
+		mutex_unlock(m2);
+		mutex_lock(m1);
+	}
+
+	return NULL;
+}
+
+static struct actions futex_actions = {
+	.setup = futex_setup,
+	.thread1 = futex_thread1,
+	.thread2 = futex_thread2,
+};
+
+static struct option options[] = {
+	{ "test", required_argument, 0, 't' },
+	{ "process", no_argument, &processes, 1 },
+	{ "timeout", required_argument, 0, 's' },
+	{ "vdso", no_argument, &touch_vdso, 1 },
+	{ "no-fp", no_argument, &touch_fp, 0 },
+#ifdef __powerpc__
+	{ "no-altivec", no_argument, &touch_altivec, 0 },
+#endif
+	{ "no-vector", no_argument, &touch_vector, 0 },
+	{ 0, },
+};
+
+static void usage(void)
+{
+	fprintf(stderr, "Usage: context_switch2 <options> CPU1 CPU2\n\n");
+	fprintf(stderr, "\t\t--test=X\tpipe, futex or yield (default)\n");
+	fprintf(stderr, "\t\t--process\tUse processes (default threads)\n");
+	fprintf(stderr, "\t\t--timeout=X\tDuration in seconds to run (default 30)\n");
+	fprintf(stderr, "\t\t--vdso\t\ttouch VDSO\n");
+	fprintf(stderr, "\t\t--no-fp\t\tDon't touch FP\n");
+#ifdef __powerpc__
+	fprintf(stderr, "\t\t--no-altivec\tDon't touch altivec\n");
+#endif
+	fprintf(stderr, "\t\t--no-vector\tDon't touch vector\n");
+}
+
+int main(int argc, char *argv[])
+{
+	signed char c;
+	struct actions *actions = &yield_actions;
+	int cpu1;
+	int cpu2;
+	static void (*start_fn)(void *(*fn)(void *), void *arg, unsigned long cpu);
+
+	while (1) {
+		int option_index = 0;
+
+		c = getopt_long(argc, argv, "", options, &option_index);
+
+		if (c == -1)
+			break;
+
+		switch (c) {
+		case 0:
+			if (options[option_index].flag != 0)
+				break;
+
+			usage();
+			exit(1);
+			break;
+
+		case 't':
+			if (!strcmp(optarg, "pipe")) {
+				actions = &pipe_actions;
+			} else if (!strcmp(optarg, "yield")) {
+				actions = &yield_actions;
+			} else if (!strcmp(optarg, "futex")) {
+				actions = &futex_actions;
+			} else {
+				usage();
+				exit(1);
+			}
+			break;
+
+		case 's':
+			timeout = atoi(optarg);
+			break;
+
+		default:
+			usage();
+			exit(1);
+		}
+	}
+
+	if (processes)
+		start_fn = start_process_on;
+	else
+		start_fn = start_thread_on;
+
+	if (((argc - optind) != 2)) {
+		cpu1 = cpu2 = pick_online_cpu();
+	} else {
+		cpu1 = atoi(argv[optind++]);
+		cpu2 = atoi(argv[optind++]);
+	}
+
+	printf("Using %s with ", processes ? "processes" : "threads");
+
+	if (actions == &pipe_actions)
+		printf("pipe");
+	else if (actions == &yield_actions)
+		printf("yield");
+	else
+		printf("futex");
+
+	if (!have_hwcap(PPC_FEATURE_HAS_ALTIVEC))
+		touch_altivec = 0;
+
+	if (!have_hwcap(PPC_FEATURE_HAS_VSX))
+		touch_vector = 0;
+
+	printf(" on cpus %d/%d touching FP:%s altivec:%s vector:%s vdso:%s\n",
+	       cpu1, cpu2, touch_fp ?  "yes" : "no", touch_altivec ? "yes" : "no",
+	       touch_vector ? "yes" : "no", touch_vdso ? "yes" : "no");
+
+	/* Create a new process group so we can signal everyone for exit */
+	setpgid(getpid(), getpid());
+
+	signal(SIGUSR1, sigusr1_handler);
+
+	actions->setup(cpu1, cpu2);
+
+	start_fn(actions->thread1, NULL, cpu1);
+	start_fn(actions->thread2, NULL, cpu2);
+
+	while (1)
+		sleep(3600);
+
+	return 0;
+}

linux-7.2.3/tools/testing/selftests/powerpc/benchmarks/exec_target.c

@@ -0,0 +1,28 @@
+// SPDX-License-Identifier: GPL-2.0+
+
+/*
+ * Part of fork context switch microbenchmark.
+ *
+ * Copyright 2018, Anton Blanchard, IBM Corp.
+ */
+
+#define _GNU_SOURCE
+#include <sys/syscall.h>
+
+void _start(void)
+{
+	asm volatile (
+		"li %%r0, %[sys_exit];"
+		"li %%r3, 0;"
+		"sc;"
+		:
+		: [sys_exit] "i" (SYS_exit)
+		/*
+		 * "sc" will clobber r0, r3-r13, cr0, ctr, xer and memory.
+		 * Even though sys_exit never returns, handle clobber
+		 * registers.
+		 */
+		: "r0", "r3", "r4", "r5", "r6", "r7", "r8", "r9", "r10",
+		  "r11", "r12", "r13", "cr0", "ctr", "xer", "memory"
+	);
+}

linux-7.2.3/tools/testing/selftests/powerpc/benchmarks/fork.c

@@ -0,0 +1,325 @@
+// SPDX-License-Identifier: GPL-2.0+
+
+/*
+ * Context switch microbenchmark.
+ *
+ * Copyright 2018, Anton Blanchard, IBM Corp.
+ */
+
+#define _GNU_SOURCE
+#include <assert.h>
+#include <errno.h>
+#include <getopt.h>
+#include <limits.h>
+#include <linux/futex.h>
+#include <pthread.h>
+#include <sched.h>
+#include <signal.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include <sys/shm.h>
+#include <sys/syscall.h>
+#include <sys/time.h>
+#include <sys/types.h>
+#include <sys/wait.h>
+#include <unistd.h>
+
+static unsigned int timeout = 30;
+
+static void set_cpu(int cpu)
+{
+	cpu_set_t cpuset;
+
+	if (cpu == -1)
+		return;
+
+	CPU_ZERO(&cpuset);
+	CPU_SET(cpu, &cpuset);
+
+	if (sched_setaffinity(0, sizeof(cpuset), &cpuset)) {
+		perror("sched_setaffinity");
+		exit(1);
+	}
+}
+
+static void start_process_on(void *(*fn)(void *), void *arg, int cpu)
+{
+	int pid;
+
+	pid = fork();
+	if (pid == -1) {
+		perror("fork");
+		exit(1);
+	}
+
+	if (pid)
+		return;
+
+	set_cpu(cpu);
+
+	fn(arg);
+
+	exit(0);
+}
+
+static int cpu;
+static int do_fork = 0;
+static int do_vfork = 0;
+static int do_exec = 0;
+static char *exec_file;
+static int exec_target = 0;
+static unsigned long iterations;
+static unsigned long iterations_prev;
+
+static void run_exec(void)
+{
+	char *const argv[] = { "./exec_target", NULL };
+
+	if (execve("./exec_target", argv, NULL) == -1) {
+		perror("execve");
+		exit(1);
+	}
+}
+
+static void bench_fork(void)
+{
+	while (1) {
+		pid_t pid = fork();
+		if (pid == -1) {
+			perror("fork");
+			exit(1);
+		}
+		if (pid == 0) {
+			if (do_exec)
+				run_exec();
+			_exit(0);
+		}
+		pid = waitpid(pid, NULL, 0);
+		if (pid == -1) {
+			perror("waitpid");
+			exit(1);
+		}
+		iterations++;
+	}
+}
+
+static void bench_vfork(void)
+{
+	while (1) {
+		pid_t pid = vfork();
+		if (pid == -1) {
+			perror("fork");
+			exit(1);
+		}
+		if (pid == 0) {
+			if (do_exec)
+				run_exec();
+			_exit(0);
+		}
+		pid = waitpid(pid, NULL, 0);
+		if (pid == -1) {
+			perror("waitpid");
+			exit(1);
+		}
+		iterations++;
+	}
+}
+
+static void *null_fn(void *arg)
+{
+	pthread_exit(NULL);
+}
+
+static void bench_thread(void)
+{
+	pthread_t tid;
+	cpu_set_t cpuset;
+	pthread_attr_t attr;
+	int rc;
+
+	rc = pthread_attr_init(&attr);
+	if (rc) {
+		errno = rc;
+		perror("pthread_attr_init");
+		exit(1);
+	}
+
+	if (cpu != -1) {
+		CPU_ZERO(&cpuset);
+		CPU_SET(cpu, &cpuset);
+
+		rc = pthread_attr_setaffinity_np(&attr, sizeof(cpu_set_t), &cpuset);
+		if (rc) {
+			errno = rc;
+			perror("pthread_attr_setaffinity_np");
+			exit(1);
+		}
+	}
+
+	while (1) {
+		rc = pthread_create(&tid, &attr, null_fn, NULL);
+		if (rc) {
+			errno = rc;
+			perror("pthread_create");
+			exit(1);
+		}
+		rc = pthread_join(tid, NULL);
+		if (rc) {
+			errno = rc;
+			perror("pthread_join");
+			exit(1);
+		}
+		iterations++;
+	}
+}
+
+static void sigalrm_handler(int junk)
+{
+	unsigned long i = iterations;
+
+	printf("%ld\n", i - iterations_prev);
+	iterations_prev = i;
+
+	if (--timeout == 0)
+		kill(0, SIGUSR1);
+
+	alarm(1);
+}
+
+static void sigusr1_handler(int junk)
+{
+	exit(0);
+}
+
+static void *bench_proc(void *arg)
+{
+	signal(SIGALRM, sigalrm_handler);
+	alarm(1);
+
+	if (do_fork)
+		bench_fork();
+	else if (do_vfork)
+		bench_vfork();
+	else
+		bench_thread();
+
+	return NULL;
+}
+
+static struct option options[] = {
+	{ "fork", no_argument, &do_fork, 1 },
+	{ "vfork", no_argument, &do_vfork, 1 },
+	{ "exec", no_argument, &do_exec, 1 },
+	{ "timeout", required_argument, 0, 's' },
+	{ "exec-target", no_argument, &exec_target, 1 },
+	{ NULL },
+};
+
+static void usage(void)
+{
+	fprintf(stderr, "Usage: fork <options> CPU\n\n");
+	fprintf(stderr, "\t\t--fork\tUse fork() (default threads)\n");
+	fprintf(stderr, "\t\t--vfork\tUse vfork() (default threads)\n");
+	fprintf(stderr, "\t\t--exec\tAlso exec() (default no exec)\n");
+	fprintf(stderr, "\t\t--timeout=X\tDuration in seconds to run (default 30)\n");
+	fprintf(stderr, "\t\t--exec-target\tInternal option for exec workload\n");
+}
+
+int main(int argc, char *argv[])
+{
+	signed char c;
+
+	while (1) {
+		int option_index = 0;
+
+		c = getopt_long(argc, argv, "", options, &option_index);
+
+		if (c == -1)
+			break;
+
+		switch (c) {
+		case 0:
+			if (options[option_index].flag != 0)
+				break;
+
+			usage();
+			exit(1);
+			break;
+
+		case 's':
+			timeout = atoi(optarg);
+			break;
+
+		default:
+			usage();
+			exit(1);
+		}
+	}
+
+	if (do_fork && do_vfork) {
+		usage();
+		exit(1);
+	}
+	if (do_exec && !do_fork && !do_vfork) {
+		usage();
+		exit(1);
+	}
+
+	if (do_exec) {
+		char *dirname = strdup(argv[0]);
+		int i;
+		i = strlen(dirname) - 1;
+		while (i) {
+			if (dirname[i] == '/') {
+				dirname[i] = '\0';
+				if (chdir(dirname) == -1) {
+					perror("chdir");
+					exit(1);
+				}
+				break;
+			}
+			i--;
+		}
+	}
+
+	if (exec_target) {
+		exit(0);
+	}
+
+	if (((argc - optind) != 1)) {
+		cpu = -1;
+	} else {
+		cpu = atoi(argv[optind++]);
+	}
+
+	if (do_exec)
+		exec_file = argv[0];
+
+	set_cpu(cpu);
+
+	printf("Using ");
+	if (do_fork)
+		printf("fork");
+	else if (do_vfork)
+		printf("vfork");
+	else
+		printf("clone");
+
+	if (do_exec)
+		printf(" + exec");
+
+	printf(" on cpu %d\n", cpu);
+
+	/* Create a new process group so we can signal everyone for exit */
+	setpgid(getpid(), getpid());
+
+	signal(SIGUSR1, sigusr1_handler);
+
+	start_process_on(bench_proc, NULL, cpu);
+
+	while (1)
+		sleep(3600);
+
+	return 0;
+}

linux-7.2.3/tools/testing/selftests/powerpc/benchmarks/futex_bench.c

@@ -0,0 +1,43 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * Copyright 2016, Anton Blanchard, Michael Ellerman, IBM Corp.
+ */
+
+#define _GNU_SOURCE
+
+#include <stdio.h>
+#include <sys/syscall.h>
+#include <time.h>
+#include <unistd.h>
+#include <linux/futex.h>
+
+#include "utils.h"
+
+#define ITERATIONS 100000000
+
+#define futex(A, B, C, D, E, F)	 syscall(__NR_futex, A, B, C, D, E, F)
+
+int test_futex(void)
+{
+	struct timespec ts_start, ts_end;
+	unsigned long i = ITERATIONS;
+
+	clock_gettime(CLOCK_MONOTONIC, &ts_start);
+
+	while (i--) {
+		unsigned int addr = 0;
+		futex(&addr, FUTEX_WAKE, 1, NULL, NULL, 0);
+	}
+
+	clock_gettime(CLOCK_MONOTONIC, &ts_end);
+
+	printf("time = %.6f\n", ts_end.tv_sec - ts_start.tv_sec + (ts_end.tv_nsec - ts_start.tv_nsec) / 1e9);
+
+	return 0;
+}
+
+int main(void)
+{
+	test_harness_set_timeout(300);
+	return test_harness(test_futex, "futex_bench");
+}

linux-7.2.3/tools/testing/selftests/powerpc/benchmarks/gettimeofday.c

@@ -0,0 +1,33 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * Copyright 2015, Anton Blanchard, IBM Corp.
+ */
+
+#include <sys/time.h>
+#include <stdio.h>
+
+#include "utils.h"
+
+static int test_gettimeofday(void)
+{
+	int i;
+
+	struct timeval tv_start, tv_end, tv_diff;
+
+	gettimeofday(&tv_start, NULL);
+
+	for(i = 0; i < 100000000; i++) {
+		gettimeofday(&tv_end, NULL);
+	}
+
+	timersub(&tv_end, &tv_start, &tv_diff);
+
+	printf("time = %.6f\n", tv_diff.tv_sec + (tv_diff.tv_usec) * 1e-6);
+
+	return 0;
+}
+
+int main(void)
+{
+	return test_harness(test_gettimeofday, "gettimeofday");
+}

linux-7.2.3/tools/testing/selftests/powerpc/benchmarks/mmap_bench.c

@@ -0,0 +1,90 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * Copyright 2016, Anton Blanchard, Michael Ellerman, IBM Corp.
+ */
+
+#include <stdio.h>
+#include <stdlib.h>
+#include <sys/mman.h>
+#include <time.h>
+#include <getopt.h>
+
+#include "utils.h"
+
+#define ITERATIONS 5000000
+
+#define MEMSIZE (1UL << 27)
+#define PAGE_SIZE (1UL << 16)
+#define CHUNK_COUNT (MEMSIZE/PAGE_SIZE)
+
+static int pg_fault;
+static int iterations = ITERATIONS;
+
+static struct option options[] = {
+	{ "pgfault", no_argument, &pg_fault, 1 },
+	{ "iterations", required_argument, 0, 'i' },
+	{ 0, },
+};
+
+static void usage(void)
+{
+	printf("mmap_bench <--pgfault> <--iterations count>\n");
+}
+
+int test_mmap(void)
+{
+	struct timespec ts_start, ts_end;
+	unsigned long i = iterations;
+
+	clock_gettime(CLOCK_MONOTONIC, &ts_start);
+
+	while (i--) {
+		char *c = mmap(NULL, MEMSIZE, PROT_READ|PROT_WRITE,
+			       MAP_PRIVATE|MAP_ANONYMOUS, -1, 0);
+		FAIL_IF(c == MAP_FAILED);
+		if (pg_fault) {
+			int count;
+			for (count = 0; count < CHUNK_COUNT; count++)
+				c[count << 16] = 'c';
+		}
+		munmap(c, MEMSIZE);
+	}
+
+	clock_gettime(CLOCK_MONOTONIC, &ts_end);
+
+	printf("time = %.6f\n", ts_end.tv_sec - ts_start.tv_sec + (ts_end.tv_nsec - ts_start.tv_nsec) / 1e9);
+
+	return 0;
+}
+
+int main(int argc, char *argv[])
+{
+	signed char c;
+	while (1) {
+		int option_index = 0;
+
+		c = getopt_long(argc, argv, "", options, &option_index);
+
+		if (c == -1)
+			break;
+
+		switch (c) {
+		case 0:
+			if (options[option_index].flag != 0)
+				break;
+
+			usage();
+			exit(1);
+			break;
+		case 'i':
+			iterations = atoi(optarg);
+			break;
+		default:
+			usage();
+			exit(1);
+		}
+	}
+
+	test_harness_set_timeout(300);
+	return test_harness(test_mmap, "mmap_bench");
+}

linux-7.2.3/tools/testing/selftests/powerpc/benchmarks/null_syscall.c

@@ -0,0 +1,154 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ * Test null syscall performance
+ *
+ * Copyright (C) 2009-2015 Anton Blanchard, IBM
+ */
+
+#define NR_LOOPS 10000000
+
+#include <string.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <unistd.h>
+#include <time.h>
+#include <sys/types.h>
+#include <sys/time.h>
+#include <sys/syscall.h>
+#include <signal.h>
+
+static volatile int soak_done;
+unsigned long long clock_frequency;
+unsigned long long timebase_frequency;
+double timebase_multiplier;
+
+static inline unsigned long mftb(void)
+{
+	unsigned long low;
+
+	asm volatile("mftb %0" : "=r" (low));
+
+	return low;
+}
+
+static void sigalrm_handler(int unused)
+{
+	soak_done = 1;
+}
+
+/*
+ * Use a timer instead of busy looping on clock_gettime() so we don't
+ * pollute profiles with glibc and VDSO hits.
+ */
+static void cpu_soak_usecs(unsigned long usecs)
+{
+	struct itimerval val;
+
+	memset(&val, 0, sizeof(val));
+	val.it_value.tv_usec = usecs;
+
+	signal(SIGALRM, sigalrm_handler);
+	setitimer(ITIMER_REAL, &val, NULL);
+
+	while (1) {
+		if (soak_done)
+			break;
+	}
+
+	signal(SIGALRM, SIG_DFL);
+}
+
+/*
+ * This only works with recent kernels where cpufreq modifies
+ * /proc/cpuinfo dynamically.
+ */
+static void get_proc_frequency(void)
+{
+	FILE *f;
+	char line[128];
+	char *p, *end;
+	unsigned long v;
+	double d;
+	char *override;
+
+	/* Try to get out of low power/low frequency mode */
+	cpu_soak_usecs(0.25 * 1000000);
+
+	f = fopen("/proc/cpuinfo", "r");
+	if (f == NULL)
+		return;
+
+	timebase_frequency = 0;
+
+	while (fgets(line, sizeof(line), f) != NULL) {
+		if (strncmp(line, "timebase", 8) == 0) {
+			p = strchr(line, ':');
+			if (p != NULL) {
+				v = strtoull(p + 1, &end, 0);
+				if (end != p + 1)
+					timebase_frequency = v;
+			}
+		}
+
+		if (((strncmp(line, "clock", 5) == 0) ||
+		     (strncmp(line, "cpu MHz", 7) == 0))) {
+			p = strchr(line, ':');
+			if (p != NULL) {
+				d = strtod(p + 1, &end);
+				if (end != p + 1) {
+					/* Find fastest clock frequency */
+					if ((d * 1000000ULL) > clock_frequency)
+						clock_frequency = d * 1000000ULL;
+				}
+			}
+		}
+	}
+
+	fclose(f);
+
+	override = getenv("FREQUENCY");
+	if (override)
+		clock_frequency = strtoull(override, NULL, 10);
+
+	if (timebase_frequency)
+		timebase_multiplier = (double)clock_frequency
+					/ timebase_frequency;
+	else
+		timebase_multiplier = 1;
+}
+
+static void do_null_syscall(unsigned long nr)
+{
+	unsigned long i;
+
+	for (i = 0; i < nr; i++)
+		syscall(__NR_gettid);
+}
+
+#define TIME(A, STR) \
+
+int main(void)
+{
+	unsigned long tb_start, tb_now;
+	struct timespec tv_start, tv_now;
+	unsigned long long elapsed_ns, elapsed_tb;
+
+	get_proc_frequency();
+
+	clock_gettime(CLOCK_MONOTONIC, &tv_start);
+	tb_start = mftb();
+
+	do_null_syscall(NR_LOOPS);
+
+	clock_gettime(CLOCK_MONOTONIC, &tv_now);
+	tb_now = mftb();
+
+	elapsed_ns = (tv_now.tv_sec - tv_start.tv_sec) * 1000000000ULL +
+			(tv_now.tv_nsec - tv_start.tv_nsec);
+	elapsed_tb = tb_now - tb_start;
+
+	printf("%10.2f ns %10.2f cycles\n", (float)elapsed_ns / NR_LOOPS,
+			(float)elapsed_tb * timebase_multiplier / NR_LOOPS);
+
+	return 0;
+}

linux-7.2.3/tools/testing/selftests/powerpc/benchmarks/settings

@@ -0,0 +1 @@
+timeout=0

linux-7.2.3/tools/testing/selftests/powerpc/cache_shape/.gitignore

@@ -0,0 +1,2 @@
+# SPDX-License-Identifier: GPL-2.0-only
+cache_shape

linux-7.2.3/tools/testing/selftests/powerpc/cache_shape/Makefile

@@ -0,0 +1,8 @@
+# SPDX-License-Identifier: GPL-2.0
+TEST_GEN_PROGS := cache_shape
+
+top_srcdir = ../../../../..
+include ../../lib.mk
+include ../flags.mk
+
+$(TEST_GEN_PROGS): ../harness.c ../utils.c

linux-7.2.3/tools/testing/selftests/powerpc/cache_shape/cache_shape.c

@@ -0,0 +1,121 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ * Copyright 2017, Michael Ellerman, IBM Corp.
+ */
+
+#include <elf.h>
+#include <errno.h>
+#include <fcntl.h>
+#include <link.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include <sys/stat.h>
+#include <sys/types.h>
+#include <sys/wait.h>
+#include <unistd.h>
+
+#include "utils.h"
+
+#ifndef AT_L1I_CACHESIZE
+#define AT_L1I_CACHESIZE	40
+#define AT_L1I_CACHEGEOMETRY	41
+#define AT_L1D_CACHESIZE	42
+#define AT_L1D_CACHEGEOMETRY	43
+#define AT_L2_CACHESIZE		44
+#define AT_L2_CACHEGEOMETRY	45
+#define AT_L3_CACHESIZE		46
+#define AT_L3_CACHEGEOMETRY	47
+#endif
+
+static void print_size(const char *label, uint32_t val)
+{
+	printf("%s cache size: %#10x %10dB %10dK\n", label, val, val, val / 1024);
+}
+
+static void print_geo(const char *label, uint32_t val)
+{
+	uint16_t assoc;
+
+	printf("%s line size:  %#10x       ", label, val & 0xFFFF);
+
+	assoc = val >> 16;
+	if (assoc)
+		printf("%u-way", assoc);
+	else
+		printf("fully");
+
+	printf(" associative\n");
+}
+
+static int test_cache_shape()
+{
+	static char buffer[4096];
+	ElfW(auxv_t) *p;
+	int found;
+
+	FAIL_IF(read_auxv(buffer, sizeof(buffer)));
+
+	found = 0;
+
+	p = find_auxv_entry(AT_L1I_CACHESIZE, buffer);
+	if (p) {
+		found++;
+		print_size("L1I ", (uint32_t)p->a_un.a_val);
+	}
+
+	p = find_auxv_entry(AT_L1I_CACHEGEOMETRY, buffer);
+	if (p) {
+		found++;
+		print_geo("L1I ", (uint32_t)p->a_un.a_val);
+	}
+
+	p = find_auxv_entry(AT_L1D_CACHESIZE, buffer);
+	if (p) {
+		found++;
+		print_size("L1D ", (uint32_t)p->a_un.a_val);
+	}
+
+	p = find_auxv_entry(AT_L1D_CACHEGEOMETRY, buffer);
+	if (p) {
+		found++;
+		print_geo("L1D ", (uint32_t)p->a_un.a_val);
+	}
+
+	p = find_auxv_entry(AT_L2_CACHESIZE, buffer);
+	if (p) {
+		found++;
+		print_size("L2  ", (uint32_t)p->a_un.a_val);
+	}
+
+	p = find_auxv_entry(AT_L2_CACHEGEOMETRY, buffer);
+	if (p) {
+		found++;
+		print_geo("L2  ", (uint32_t)p->a_un.a_val);
+	}
+
+	p = find_auxv_entry(AT_L3_CACHESIZE, buffer);
+	if (p) {
+		found++;
+		print_size("L3  ", (uint32_t)p->a_un.a_val);
+	}
+
+	p = find_auxv_entry(AT_L3_CACHEGEOMETRY, buffer);
+	if (p) {
+		found++;
+		print_geo("L3  ", (uint32_t)p->a_un.a_val);
+	}
+
+	/* If we found none we're probably on a system where they don't exist */
+	SKIP_IF(found == 0);
+
+	/* But if we found any, we expect to find them all */
+	FAIL_IF(found != 8);
+
+	return 0;
+}
+
+int main(void)
+{
+	return test_harness(test_cache_shape, "cache_shape");
+}

linux-7.2.3/tools/testing/selftests/powerpc/cache_shape/settings

@@ -0,0 +1 @@
+timeout=130

linux-7.2.3/tools/testing/selftests/powerpc/copyloops/.gitignore

@@ -0,0 +1,16 @@
+# SPDX-License-Identifier: GPL-2.0-only
+copyuser_64_t0
+copyuser_64_t1
+copyuser_64_t2
+copyuser_p7
+copyuser_p7_vmx
+memcpy_64_t0
+memcpy_64_t1
+memcpy_64_t2
+memcpy_p7_t0
+memcpy_p7_t1
+copyuser_64_exc_t0
+copyuser_64_exc_t1
+copyuser_64_exc_t2
+copy_mc_64
+memmove_64

linux-7.2.3/tools/testing/selftests/powerpc/copyloops/Makefile

@@ -0,0 +1,71 @@
+# SPDX-License-Identifier: GPL-2.0
+TEST_GEN_PROGS := copyuser_64_t0 copyuser_64_t1 copyuser_64_t2 \
+		copyuser_p7 copyuser_p7_vmx \
+		memcpy_64_t0 memcpy_64_t1 memcpy_64_t2 \
+		memcpy_p7_t0 memcpy_p7_t1 copy_mc_64 \
+		copyuser_64_exc_t0 copyuser_64_exc_t1 copyuser_64_exc_t2 \
+		memmove_64
+
+EXTRA_SOURCES := validate.c ../harness.c stubs.S
+
+top_srcdir = ../../../../..
+include ../../lib.mk
+include ../flags.mk
+
+# The loops are all 64-bit code
+CFLAGS += -m64
+CFLAGS += -I$(CURDIR)
+CFLAGS += -D SELFTEST
+CFLAGS += -maltivec
+CFLAGS += -mcpu=power4
+
+# Use our CFLAGS for the implicit .S rule & set the asm machine type
+ASFLAGS = $(CFLAGS) -Wa,-mpower4
+
+$(OUTPUT)/copyuser_64_t%:	copyuser_64.S $(EXTRA_SOURCES)
+	$(CC) $(CPPFLAGS) $(CFLAGS) \
+		-D COPY_LOOP=test___copy_tofrom_user_base \
+		-D SELFTEST_CASE=$(subst copyuser_64_t,,$(notdir $@)) \
+		-o $@ $^
+
+$(OUTPUT)/copyuser_p7:	copyuser_power7.S $(EXTRA_SOURCES)
+	$(CC) $(CPPFLAGS) $(CFLAGS) \
+		-D COPY_LOOP=test___copy_tofrom_user_power7 \
+		-o $@ $^
+
+$(OUTPUT)/copyuser_p7_vmx:       copyuser_power7.S $(EXTRA_SOURCES) ../utils.c
+	$(CC) $(CPPFLAGS) $(CFLAGS) \
+		-D COPY_LOOP=test___copy_tofrom_user_power7_vmx \
+		-D VMX_TEST \
+		-o $@ $^
+
+# Strictly speaking, we only need the memcpy_64 test cases for big-endian
+$(OUTPUT)/memcpy_64_t%:	memcpy_64.S $(EXTRA_SOURCES)
+	$(CC) $(CPPFLAGS) $(CFLAGS) \
+		-D COPY_LOOP=test_memcpy \
+		-D SELFTEST_CASE=$(subst memcpy_64_t,,$(notdir $@)) \
+		-o $@ $^
+
+$(OUTPUT)/memcpy_p7_t%:	memcpy_power7.S $(EXTRA_SOURCES)
+	$(CC) $(CPPFLAGS) $(CFLAGS) \
+		-D COPY_LOOP=test_memcpy_power7 \
+		-D SELFTEST_CASE=$(subst memcpy_p7_t,,$(notdir $@)) \
+		-o $@ $^
+
+$(OUTPUT)/copy_mc_64: copy_mc_64.S $(EXTRA_SOURCES)
+	$(CC) $(CPPFLAGS) $(CFLAGS) \
+		-D COPY_LOOP=test_copy_mc_generic \
+		-o $@ $^
+
+$(OUTPUT)/copyuser_64_exc_t%: copyuser_64.S exc_validate.c ../harness.c \
+		copy_tofrom_user_reference.S stubs.S
+	$(CC) $(CPPFLAGS) $(CFLAGS) \
+		-D COPY_LOOP=test___copy_tofrom_user_base \
+		-D SELFTEST_CASE=$(subst copyuser_64_exc_t,,$(notdir $@)) \
+		-o $@ $^
+
+$(OUTPUT)/memmove_64: mem_64.S memcpy_64.S memmove_validate.c ../harness.c \
+		memcpy_stubs.S
+	$(CC) $(CPPFLAGS) $(CFLAGS) \
+		-D TEST_MEMMOVE=test_memmove \
+		-o $@ $^

linux-7.2.3/tools/testing/selftests/powerpc/copyloops/asm/ppc_asm.h

@@ -0,0 +1,62 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+#ifndef __SELFTESTS_POWERPC_PPC_ASM_H
+#define __SELFTESTS_POWERPC_PPC_ASM_H
+#include <ppc-asm.h>
+
+#define CONFIG_ALTIVEC
+
+#define r1	1
+
+#define R14 r14
+#define R15 r15
+#define R16 r16
+#define R17 r17
+#define R18 r18
+#define R19 r19
+#define R20 r20
+#define R21 r21
+#define R22 r22
+#define R29 r29
+#define R30 r30
+#define R31 r31
+
+#define STACKFRAMESIZE	256
+#define STK_REG(i)	(112 + ((i)-14)*8)
+
+#define _GLOBAL(A) FUNC_START(test_ ## A)
+#define _GLOBAL_TOC(A) _GLOBAL(A)
+#define _GLOBAL_TOC_KASAN(A) _GLOBAL(A)
+#define _GLOBAL_KASAN(A) _GLOBAL(A)
+#define CFUNC(name) name
+
+#define PPC_MTOCRF(A, B)	mtocrf A, B
+
+#define EX_TABLE(x, y)			\
+	.section __ex_table,"a";	\
+	.8byte	x, y;			\
+	.previous
+
+#define BEGIN_FTR_SECTION		.if test_feature
+#define FTR_SECTION_ELSE		.else
+#define ALT_FTR_SECTION_END_IFCLR(x)	.endif
+#define ALT_FTR_SECTION_END_IFSET(x)	.endif
+#define ALT_FTR_SECTION_END(x, y)	.endif
+#define END_FTR_SECTION_IFCLR(x)	.endif
+#define END_FTR_SECTION_IFSET(x)	.endif
+
+/* Default to taking the first of any alternative feature sections */
+test_feature = 1
+
+#define DCBT_SETUP_STREAMS(from, from_parms, to, to_parms, scratch)	\
+	lis	scratch,0x8000;	/* GO=1 */				\
+	clrldi	scratch,scratch,32;					\
+	/* setup read stream 0 */					\
+	dcbt	0,from,0b01000;		/* addr from */			\
+	dcbt	0,from_parms,0b01010;	/* length and depth from */	\
+	/* setup write stream 1 */					\
+	dcbtst	0,to,0b01000;		/* addr to */			\
+	dcbtst	0,to_parms,0b01010;	/* length and depth to */	\
+	eieio;								\
+	dcbt	0,scratch,0b01010;	/* all streams GO */
+
+#endif /* __SELFTESTS_POWERPC_PPC_ASM_H */

linux-7.2.3/tools/testing/selftests/powerpc/copyloops/copy_mc_64.S

@@ -0,0 +1 @@
+../../../../../arch/powerpc/lib/copy_mc_64.S

linux-7.2.3/tools/testing/selftests/powerpc/copyloops/copy_tofrom_user_reference.S

@@ -0,0 +1,24 @@
+#include <asm/ppc_asm.h>
+
+_GLOBAL(copy_tofrom_user_reference)
+	cmpdi	r5,0
+	beq	4f
+
+	mtctr	r5
+
+1:	lbz	r6,0(r4)
+2:	stb	r6,0(r3)
+	addi	r3,r3,1
+	addi	r4,r4,1
+	bdnz	1b
+
+3:	mfctr	r3
+	blr
+
+4:	mr	r3,r5
+	blr
+
+.section __ex_table,"a"
+	.llong	1b,3b
+	.llong	2b,3b
+.text

linux-7.2.3/tools/testing/selftests/powerpc/copyloops/copyuser_64.S

@@ -0,0 +1 @@
+../../../../../arch/powerpc/lib/copyuser_64.S

linux-7.2.3/tools/testing/selftests/powerpc/copyloops/copyuser_power7.S

@@ -0,0 +1 @@
+../../../../../arch/powerpc/lib/copyuser_power7.S

linux-7.2.3/tools/testing/selftests/powerpc/copyloops/exc_validate.c

@@ -0,0 +1,124 @@
+#include <stdlib.h>
+#include <string.h>
+#include <stdio.h>
+#include <signal.h>
+#include <unistd.h>
+#include <sys/mman.h>
+
+#include "utils.h"
+
+extern char __start___ex_table[];
+extern char __stop___ex_table[];
+
+#if defined(__powerpc64__)
+#define UCONTEXT_NIA(UC)	(UC)->uc_mcontext.gp_regs[PT_NIP]
+#elif defined(__powerpc__)
+#define UCONTEXT_NIA(UC)	(UC)->uc_mcontext.uc_regs->gregs[PT_NIP]
+#else
+#error implement UCONTEXT_NIA
+#endif
+
+static void segv_handler(int signr, siginfo_t *info, void *ptr)
+{
+	ucontext_t *uc = (ucontext_t *)ptr;
+	unsigned long addr = (unsigned long)info->si_addr;
+	unsigned long *ip = &UCONTEXT_NIA(uc);
+	unsigned long *ex_p = (unsigned long *)__start___ex_table;
+
+	while (ex_p < (unsigned long *)__stop___ex_table) {
+		unsigned long insn, fixup;
+
+		insn = *ex_p++;
+		fixup = *ex_p++;
+
+		if (insn == *ip) {
+			*ip = fixup;
+			return;
+		}
+	}
+
+	printf("No exception table match for NIA %lx ADDR %lx\n", *ip, addr);
+	abort();
+}
+
+static void setup_segv_handler(void)
+{
+	struct sigaction action;
+
+	memset(&action, 0, sizeof(action));
+	action.sa_sigaction = segv_handler;
+	action.sa_flags = SA_SIGINFO;
+	sigaction(SIGSEGV, &action, NULL);
+}
+
+unsigned long COPY_LOOP(void *to, const void *from, unsigned long size);
+unsigned long test_copy_tofrom_user_reference(void *to, const void *from, unsigned long size);
+
+static int total_passed;
+static int total_failed;
+
+static void do_one_test(char *dstp, char *srcp, unsigned long len)
+{
+	unsigned long got, expected;
+
+	got = COPY_LOOP(dstp, srcp, len);
+	expected = test_copy_tofrom_user_reference(dstp, srcp, len);
+
+	if (got != expected) {
+		total_failed++;
+		printf("FAIL from=%p to=%p len=%ld returned %ld, expected %ld\n",
+		       srcp, dstp, len, got, expected);
+		//abort();
+	} else
+		total_passed++;
+}
+
+//#define MAX_LEN 512
+#define MAX_LEN 16
+
+int test_copy_exception(void)
+{
+	int page_size;
+	static char *p, *q;
+	unsigned long src, dst, len;
+
+	page_size = getpagesize();
+	p = mmap(NULL, page_size * 2, PROT_READ|PROT_WRITE,
+		MAP_PRIVATE|MAP_ANONYMOUS, -1, 0);
+
+	if (p == MAP_FAILED) {
+		perror("mmap");
+		exit(1);
+	}
+
+	memset(p, 0, page_size);
+
+	setup_segv_handler();
+
+	if (mprotect(p + page_size, page_size, PROT_NONE)) {
+		perror("mprotect");
+		exit(1);
+	}
+
+	q = p + page_size - MAX_LEN;
+
+	for (src = 0; src < MAX_LEN; src++) {
+		for (dst = 0; dst < MAX_LEN; dst++) {
+			for (len = 0; len < MAX_LEN+1; len++) {
+				// printf("from=%p to=%p len=%ld\n", q+dst, q+src, len);
+				do_one_test(q+dst, q+src, len);
+			}
+		}
+	}
+
+	printf("Totals:\n");
+	printf("  Pass: %d\n", total_passed);
+	printf("  Fail: %d\n", total_failed);
+
+	return 0;
+}
+
+int main(void)
+{
+	return test_harness(test_copy_exception, str(COPY_LOOP));
+}

linux-7.2.3/tools/testing/selftests/powerpc/copyloops/linux/export.h

@@ -0,0 +1,4 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+#define EXPORT_SYMBOL(x)
+#define EXPORT_SYMBOL_GPL(x)
+#define EXPORT_SYMBOL_KASAN(x)

linux-7.2.3/tools/testing/selftests/powerpc/copyloops/mem_64.S

@@ -0,0 +1 @@
+../../../../../arch/powerpc/lib/mem_64.S

linux-7.2.3/tools/testing/selftests/powerpc/copyloops/memcpy_64.S

@@ -0,0 +1 @@
+../../../../../arch/powerpc/lib/memcpy_64.S

linux-7.2.3/tools/testing/selftests/powerpc/copyloops/memcpy_power7.S

@@ -0,0 +1 @@
+../../../../../arch/powerpc/lib/memcpy_power7.S

linux-7.2.3/tools/testing/selftests/powerpc/copyloops/memcpy_stubs.S

@@ -0,0 +1,8 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+#include <asm/ppc_asm.h>
+
+FUNC_START(memcpy)
+	b test_memcpy
+
+FUNC_START(backwards_memcpy)
+	b test_backwards_memcpy

linux-7.2.3/tools/testing/selftests/powerpc/copyloops/memmove_validate.c

@@ -0,0 +1,58 @@
+// SPDX-License-Identifier: GPL-2.0
+#include <malloc.h>
+#include <stdlib.h>
+#include <string.h>
+#include <assert.h>
+#include "utils.h"
+
+void *TEST_MEMMOVE(const void *s1, const void *s2, size_t n);
+
+#define BUF_LEN 65536
+#define MAX_OFFSET 512
+
+size_t max(size_t a, size_t b)
+{
+	if (a >= b)
+		return a;
+	return b;
+}
+
+static int testcase_run(void)
+{
+	size_t i, src_off, dst_off, len;
+
+	char *usermap = memalign(BUF_LEN, BUF_LEN);
+	char *kernelmap = memalign(BUF_LEN, BUF_LEN);
+
+	assert(usermap != NULL);
+	assert(kernelmap != NULL);
+
+	memset(usermap, 0, BUF_LEN);
+	memset(kernelmap, 0, BUF_LEN);
+
+	for (i = 0; i < BUF_LEN; i++) {
+		usermap[i] = i & 0xff;
+		kernelmap[i] = i & 0xff;
+	}
+
+	for (src_off = 0; src_off < MAX_OFFSET; src_off++) {
+		for (dst_off = 0; dst_off < MAX_OFFSET; dst_off++) {
+			for (len = 1; len < MAX_OFFSET - max(src_off, dst_off); len++) {
+
+				memmove(usermap + dst_off, usermap + src_off, len);
+				TEST_MEMMOVE(kernelmap + dst_off, kernelmap + src_off, len);
+				if (memcmp(usermap, kernelmap, MAX_OFFSET) != 0) {
+					printf("memmove failed at %ld %ld %ld\n",
+							src_off, dst_off, len);
+					abort();
+				}
+			}
+		}
+	}
+	return 0;
+}
+
+int main(void)
+{
+	return test_harness(testcase_run, "memmove");
+}

linux-7.2.3/tools/testing/selftests/powerpc/copyloops/settings

@@ -0,0 +1 @@
+timeout=130

linux-7.2.3/tools/testing/selftests/powerpc/copyloops/stubs.S

@@ -0,0 +1,11 @@
+#include <asm/ppc_asm.h>
+
+FUNC_START(enter_vmx_ops)
+	li	r3,1
+	blr
+
+FUNC_START(exit_vmx_ops)
+	blr
+
+FUNC_START(__copy_tofrom_user_base)
+	blr

linux-7.2.3/tools/testing/selftests/powerpc/copyloops/validate.c

@@ -0,0 +1,113 @@
+// SPDX-License-Identifier: GPL-2.0
+#include <malloc.h>
+#include <string.h>
+#include <stdlib.h>
+#include <stdbool.h>
+
+#include "utils.h"
+
+#define MAX_LEN 8192
+#define MAX_OFFSET 16
+#define MIN_REDZONE 128
+#define BUFLEN (MAX_LEN+MAX_OFFSET+2*MIN_REDZONE)
+#define POISON 0xa5
+
+#ifdef VMX_TEST
+#define VMX_COPY_THRESHOLD 3328
+#endif
+
+unsigned long COPY_LOOP(void *to, const void *from, unsigned long size);
+
+static void do_one(char *src, char *dst, unsigned long src_off,
+		   unsigned long dst_off, unsigned long len, void *redzone,
+		   void *fill)
+{
+	char *srcp, *dstp;
+	unsigned long ret;
+	unsigned long i;
+
+	srcp = src + MIN_REDZONE + src_off;
+	dstp = dst + MIN_REDZONE + dst_off;
+
+	memset(src, POISON, BUFLEN);
+	memset(dst, POISON, BUFLEN);
+	memcpy(srcp, fill, len);
+
+	ret = COPY_LOOP(dstp, srcp, len);
+	if (ret && ret != (unsigned long)dstp) {
+		printf("(%p,%p,%ld) returned %ld\n", dstp, srcp, len, ret);
+		abort();
+	}
+
+	if (memcmp(dstp, srcp, len)) {
+		printf("(%p,%p,%ld) miscompare\n", dstp, srcp, len);
+		printf("src: ");
+		for (i = 0; i < len; i++)
+			printf("%02x ", srcp[i]);
+		printf("\ndst: ");
+		for (i = 0; i < len; i++)
+			printf("%02x ", dstp[i]);
+		printf("\n");
+		abort();
+	}
+
+	if (memcmp(dst, redzone, dstp - dst)) {
+		printf("(%p,%p,%ld) redzone before corrupted\n",
+		       dstp, srcp, len);
+		abort();
+	}
+
+	if (memcmp(dstp+len, redzone, dst+BUFLEN-(dstp+len))) {
+		printf("(%p,%p,%ld) redzone after corrupted\n",
+		       dstp, srcp, len);
+		abort();
+	}
+}
+
+int test_copy_loop(void)
+{
+	char *src, *dst, *redzone, *fill;
+	unsigned long len, src_off, dst_off;
+	unsigned long i;
+
+	src = memalign(BUFLEN, BUFLEN);
+	dst = memalign(BUFLEN, BUFLEN);
+	redzone = malloc(BUFLEN);
+	fill = malloc(BUFLEN);
+
+	if (!src || !dst || !redzone || !fill) {
+		fprintf(stderr, "malloc failed\n");
+		exit(1);
+	}
+
+	memset(redzone, POISON, BUFLEN);
+
+	/* Fill with sequential bytes */
+	for (i = 0; i < BUFLEN; i++)
+		fill[i] = i & 0xff;
+#ifdef VMX_TEST
+	/* Force sizes above kernel VMX threshold (3328) */
+	for (len = VMX_COPY_THRESHOLD + 1; len < MAX_LEN; len++) {
+#else
+	for (len = 1; len < MAX_LEN; len++) {
+#endif
+		for (src_off = 0; src_off < MAX_OFFSET; src_off++) {
+			for (dst_off = 0; dst_off < MAX_OFFSET; dst_off++) {
+				do_one(src, dst, src_off, dst_off, len,
+				       redzone, fill);
+			}
+		}
+	}
+
+	return 0;
+}
+
+int main(void)
+{
+#ifdef VMX_TEST
+	/* Skip if Altivec not present */
+	SKIP_IF_MSG(!have_hwcap(PPC_FEATURE_HAS_ALTIVEC), "ALTIVEC not supported");
+#endif
+
+	return test_harness(test_copy_loop, str(COPY_LOOP));
+}

linux-7.2.3/tools/testing/selftests/powerpc/dexcr/.gitignore

@@ -0,0 +1,4 @@
+dexcr_test
+hashchk_test
+chdexcr
+lsdexcr

linux-7.2.3/tools/testing/selftests/powerpc/dexcr/Makefile

@@ -0,0 +1,12 @@
+TEST_GEN_PROGS := dexcr_test hashchk_test
+TEST_GEN_FILES := lsdexcr chdexcr
+
+include ../../lib.mk
+include ../flags.mk
+
+CFLAGS += $(KHDR_INCLUDES)
+
+$(OUTPUT)/hashchk_test: CFLAGS += -fno-pie -no-pie $(call cc-option,-mno-rop-protect)
+
+$(TEST_GEN_PROGS): ../harness.c ../utils.c ./dexcr.c
+$(TEST_GEN_FILES): ../utils.c ./dexcr.c

linux-7.2.3/tools/testing/selftests/powerpc/dexcr/chdexcr.c

@@ -0,0 +1,112 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+
+#include <errno.h>
+#include <stddef.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include <sys/prctl.h>
+
+#include "dexcr.h"
+#include "utils.h"
+
+static void die(const char *msg)
+{
+	printf("%s\n", msg);
+	exit(1);
+}
+
+static void help(void)
+{
+	printf("Invoke a provided program with a custom DEXCR on-exec reset value\n"
+	       "\n"
+	       "usage: chdexcr [CHDEXCR OPTIONS] -- PROGRAM [ARGS...]\n"
+	       "\n"
+	       "Each configurable DEXCR aspect is exposed as an option.\n"
+	       "\n"
+	       "The normal option sets the aspect in the DEXCR. The --no- variant\n"
+	       "clears that aspect. For example, --ibrtpd sets the IBRTPD aspect bit,\n"
+	       "so indirect branch prediction will be disabled in the provided program.\n"
+	       "Conversely, --no-ibrtpd clears the aspect bit, so indirect branch\n"
+	       "prediction may occur.\n"
+	       "\n"
+	       "CHDEXCR OPTIONS:\n");
+
+	for (int i = 0; i < ARRAY_SIZE(aspects); i++) {
+		const struct dexcr_aspect *aspect = &aspects[i];
+
+		if (aspect->prctl == -1)
+			continue;
+
+		printf("  --%-6s / --no-%-6s : %s\n", aspect->opt, aspect->opt, aspect->desc);
+	}
+}
+
+static const struct dexcr_aspect *opt_to_aspect(const char *opt)
+{
+	for (int i = 0; i < ARRAY_SIZE(aspects); i++)
+		if (aspects[i].prctl != -1 && !strcmp(aspects[i].opt, opt))
+			return &aspects[i];
+
+	return NULL;
+}
+
+static int apply_option(const char *option)
+{
+	const struct dexcr_aspect *aspect;
+	const char *opt = NULL;
+	const char *set_prefix = "--";
+	const char *clear_prefix = "--no-";
+	unsigned long ctrl = 0;
+	int err;
+
+	if (!strcmp(option, "-h") || !strcmp(option, "--help")) {
+		help();
+		exit(0);
+	}
+
+	/* Strip out --(no-) prefix and determine ctrl value */
+	if (!strncmp(option, clear_prefix, strlen(clear_prefix))) {
+		opt = &option[strlen(clear_prefix)];
+		ctrl |= PR_PPC_DEXCR_CTRL_CLEAR_ONEXEC;
+	} else if (!strncmp(option, set_prefix, strlen(set_prefix))) {
+		opt = &option[strlen(set_prefix)];
+		ctrl |= PR_PPC_DEXCR_CTRL_SET_ONEXEC;
+	}
+
+	if (!opt || !*opt)
+		return 1;
+
+	aspect = opt_to_aspect(opt);
+	if (!aspect)
+		die("unknown aspect");
+
+	err = pr_set_dexcr(aspect->prctl, ctrl);
+	if (err)
+		die("failed to apply option");
+
+	return 0;
+}
+
+int main(int argc, char *const argv[])
+{
+	int i;
+
+	if (!dexcr_exists())
+		die("DEXCR not detected on this hardware");
+
+	for (i = 1; i < argc; i++)
+		if (apply_option(argv[i]))
+			break;
+
+	if (i < argc && !strcmp(argv[i], "--"))
+		i++;
+
+	if (i >= argc)
+		die("missing command");
+
+	execvp(argv[i], &argv[i]);
+	perror("execve");
+
+	return errno;
+}

linux-7.2.3/tools/testing/selftests/powerpc/dexcr/dexcr.c

@@ -0,0 +1,172 @@
+// SPDX-License-Identifier: GPL-2.0+
+
+#include <errno.h>
+#include <setjmp.h>
+#include <signal.h>
+#include <sys/prctl.h>
+#include <sys/types.h>
+#include <sys/wait.h>
+
+#include "dexcr.h"
+#include "reg.h"
+#include "utils.h"
+
+static jmp_buf generic_signal_jump_buf;
+
+static void generic_signal_handler(int signum, siginfo_t *info, void *context)
+{
+	longjmp(generic_signal_jump_buf, 0);
+}
+
+bool dexcr_exists(void)
+{
+	struct sigaction old;
+	volatile bool exists;
+
+	old = push_signal_handler(SIGILL, generic_signal_handler);
+	if (setjmp(generic_signal_jump_buf))
+		goto out;
+
+	/*
+	 * If the SPR is not recognised by the hardware it triggers
+	 * a hypervisor emulation interrupt. If the kernel does not
+	 * recognise/try to emulate it, we receive a SIGILL signal.
+	 *
+	 * If we do not receive a signal, assume we have the SPR or the
+	 * kernel is trying to emulate it correctly.
+	 */
+	exists = false;
+	mfspr(SPRN_DEXCR_RO);
+	exists = true;
+
+out:
+	pop_signal_handler(SIGILL, old);
+	return exists;
+}
+
+unsigned int pr_which_to_aspect(unsigned long which)
+{
+	switch (which) {
+	case PR_PPC_DEXCR_SBHE:
+		return DEXCR_PR_SBHE;
+	case PR_PPC_DEXCR_IBRTPD:
+		return DEXCR_PR_IBRTPD;
+	case PR_PPC_DEXCR_SRAPD:
+		return DEXCR_PR_SRAPD;
+	case PR_PPC_DEXCR_NPHIE:
+		return DEXCR_PR_NPHIE;
+	default:
+		FAIL_IF_EXIT_MSG(true, "unknown PR aspect");
+	}
+}
+
+int pr_get_dexcr(unsigned long which)
+{
+	return prctl(PR_PPC_GET_DEXCR, which, 0UL, 0UL, 0UL);
+}
+
+int pr_set_dexcr(unsigned long which, unsigned long ctrl)
+{
+	return prctl(PR_PPC_SET_DEXCR, which, ctrl, 0UL, 0UL);
+}
+
+bool pr_dexcr_aspect_supported(unsigned long which)
+{
+	if (pr_get_dexcr(which) == -1)
+		return errno == ENODEV;
+
+	return true;
+}
+
+bool pr_dexcr_aspect_editable(unsigned long which)
+{
+	return pr_get_dexcr(which) & PR_PPC_DEXCR_CTRL_EDITABLE;
+}
+
+/*
+ * Just test if a bad hashchk triggers a signal, without checking
+ * for support or if the NPHIE aspect is enabled.
+ */
+bool hashchk_triggers(void)
+{
+	struct sigaction old;
+	volatile bool triggers;
+
+	old = push_signal_handler(SIGILL, generic_signal_handler);
+	if (setjmp(generic_signal_jump_buf))
+		goto out;
+
+	triggers = true;
+	do_bad_hashchk();
+	triggers = false;
+
+out:
+	pop_signal_handler(SIGILL, old);
+	return triggers;
+}
+
+unsigned int get_dexcr(enum dexcr_source source)
+{
+	switch (source) {
+	case DEXCR:
+		return mfspr(SPRN_DEXCR_RO);
+	case HDEXCR:
+		return mfspr(SPRN_HDEXCR_RO);
+	case EFFECTIVE:
+		return mfspr(SPRN_DEXCR_RO) | mfspr(SPRN_HDEXCR_RO);
+	default:
+		FAIL_IF_EXIT_MSG(true, "bad enum dexcr_source");
+	}
+}
+
+void await_child_success(pid_t pid)
+{
+	int wstatus;
+
+	FAIL_IF_EXIT_MSG(pid == -1, "fork failed");
+	FAIL_IF_EXIT_MSG(waitpid(pid, &wstatus, 0) == -1, "wait failed");
+	FAIL_IF_EXIT_MSG(!WIFEXITED(wstatus), "child did not exit cleanly");
+	FAIL_IF_EXIT_MSG(WEXITSTATUS(wstatus) != 0, "child exit error");
+}
+
+/*
+ * Perform a hashst instruction. The following components determine the result
+ *
+ * 1. The LR value (any register technically)
+ * 2. The SP value (also any register, but it must be a valid address)
+ * 3. A secret key managed by the kernel
+ *
+ * The result is stored to the address held in SP.
+ */
+void hashst(unsigned long lr, void *sp)
+{
+	asm volatile ("addi 31, %0, 0;"		/* set r31 (pretend LR) to lr */
+		      "addi 30, %1, 8;"		/* set r30 (pretend SP) to sp + 8 */
+		      PPC_RAW_HASHST(31, -8, 30)	/* compute hash into stack location */
+		      : : "r" (lr), "r" (sp) : "r31", "r30", "memory");
+}
+
+/*
+ * Perform a hashchk instruction. A hash is computed as per hashst(),
+ * however the result is not stored to memory. Instead the existing
+ * value is read and compared against the computed hash.
+ *
+ * If they match, execution continues.
+ * If they differ, an interrupt triggers.
+ */
+void hashchk(unsigned long lr, void *sp)
+{
+	asm volatile ("addi 31, %0, 0;"		/* set r31 (pretend LR) to lr */
+		      "addi 30, %1, 8;"		/* set r30 (pretend SP) to sp + 8 */
+		      PPC_RAW_HASHCHK(31, -8, 30)	/* check hash at stack location */
+		      : : "r" (lr), "r" (sp) : "r31", "r30", "memory");
+}
+
+void do_bad_hashchk(void)
+{
+	unsigned long hash = 0;
+
+	hashst(0, &hash);
+	hash += 1;
+	hashchk(0, &hash);
+}

linux-7.2.3/tools/testing/selftests/powerpc/dexcr/dexcr.h

@@ -0,0 +1,106 @@
+/* SPDX-License-Identifier: GPL-2.0-or-later */
+/*
+ * POWER Dynamic Execution Control Facility (DEXCR)
+ *
+ * This header file contains helper functions and macros
+ * required for all the DEXCR related test cases.
+ */
+#ifndef _SELFTESTS_POWERPC_DEXCR_DEXCR_H
+#define _SELFTESTS_POWERPC_DEXCR_DEXCR_H
+
+#include <stdbool.h>
+#include <sys/prctl.h>
+#include <sys/types.h>
+
+#include "reg.h"
+
+#define DEXCR_PR_BIT(aspect)	__MASK(63 - (32 + (aspect)))
+#define DEXCR_PR_SBHE		DEXCR_PR_BIT(0)
+#define DEXCR_PR_IBRTPD		DEXCR_PR_BIT(3)
+#define DEXCR_PR_SRAPD		DEXCR_PR_BIT(4)
+#define DEXCR_PR_NPHIE		DEXCR_PR_BIT(5)
+
+#define PPC_RAW_HASH_ARGS(b, i, a) \
+	((((i) >> 3) & 0x1F) << 21 | (a) << 16 | (b) << 11 | (((i) >> 8) & 0x1))
+#define PPC_RAW_HASHST(b, i, a) \
+	str(.long (0x7C0005A4 | PPC_RAW_HASH_ARGS(b, i, a));)
+#define PPC_RAW_HASHCHK(b, i, a) \
+	str(.long (0x7C0005E4 | PPC_RAW_HASH_ARGS(b, i, a));)
+
+struct dexcr_aspect {
+	const char *name;	/* Short display name */
+	const char *opt;	/* Option name for chdexcr */
+	const char *desc;	/* Expanded aspect meaning */
+	unsigned int index;	/* Aspect bit index in DEXCR */
+	unsigned long prctl;	/* 'which' value for get/set prctl */
+};
+
+static const struct dexcr_aspect aspects[] = {
+	{
+		.name = "SBHE",
+		.opt = "sbhe",
+		.desc = "Speculative branch hint enable",
+		.index = 0,
+		.prctl = PR_PPC_DEXCR_SBHE,
+	},
+	{
+		.name = "IBRTPD",
+		.opt = "ibrtpd",
+		.desc = "Indirect branch recurrent target prediction disable",
+		.index = 3,
+		.prctl = PR_PPC_DEXCR_IBRTPD,
+	},
+	{
+		.name = "SRAPD",
+		.opt = "srapd",
+		.desc = "Subroutine return address prediction disable",
+		.index = 4,
+		.prctl = PR_PPC_DEXCR_SRAPD,
+	},
+	{
+		.name = "NPHIE",
+		.opt = "nphie",
+		.desc = "Non-privileged hash instruction enable",
+		.index = 5,
+		.prctl = PR_PPC_DEXCR_NPHIE,
+	},
+	{
+		.name = "PHIE",
+		.opt = "phie",
+		.desc = "Privileged hash instruction enable",
+		.index = 6,
+		.prctl = -1,
+	},
+};
+
+bool dexcr_exists(void);
+
+bool pr_dexcr_aspect_supported(unsigned long which);
+
+bool pr_dexcr_aspect_editable(unsigned long which);
+
+int pr_get_dexcr(unsigned long pr_aspect);
+
+int pr_set_dexcr(unsigned long pr_aspect, unsigned long ctrl);
+
+unsigned int pr_which_to_aspect(unsigned long which);
+
+bool hashchk_triggers(void);
+
+enum dexcr_source {
+	DEXCR,		/* Userspace DEXCR value */
+	HDEXCR,		/* Hypervisor enforced DEXCR value */
+	EFFECTIVE,	/* Bitwise OR of UDEXCR and ENFORCED DEXCR bits */
+};
+
+unsigned int get_dexcr(enum dexcr_source source);
+
+void await_child_success(pid_t pid);
+
+void hashst(unsigned long lr, void *sp);
+
+void hashchk(unsigned long lr, void *sp);
+
+void do_bad_hashchk(void);
+
+#endif  /* _SELFTESTS_POWERPC_DEXCR_DEXCR_H */

linux-7.2.3/tools/testing/selftests/powerpc/dexcr/dexcr_test.c

@@ -0,0 +1,215 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+
+#include <errno.h>
+#include <fcntl.h>
+#include <stdlib.h>
+#include <string.h>
+#include <sys/prctl.h>
+#include <unistd.h>
+
+#include "dexcr.h"
+#include "utils.h"
+
+/*
+ * Helper function for testing the behaviour of a newly exec-ed process
+ */
+static int dexcr_prctl_onexec_test_child(unsigned long which, const char *status)
+{
+	unsigned long dexcr = mfspr(SPRN_DEXCR_RO);
+	unsigned long aspect = pr_which_to_aspect(which);
+	int ctrl = pr_get_dexcr(which);
+
+	if (!strcmp(status, "set")) {
+		FAIL_IF_EXIT_MSG(!(ctrl & PR_PPC_DEXCR_CTRL_SET),
+				 "setting aspect across exec not applied");
+
+		FAIL_IF_EXIT_MSG(!(ctrl & PR_PPC_DEXCR_CTRL_SET_ONEXEC),
+				 "setting aspect across exec not inherited");
+
+		FAIL_IF_EXIT_MSG(!(aspect & dexcr), "setting aspect across exec did not take effect");
+	} else if (!strcmp(status, "clear")) {
+		FAIL_IF_EXIT_MSG(!(ctrl & PR_PPC_DEXCR_CTRL_CLEAR),
+				 "clearing aspect across exec not applied");
+
+		FAIL_IF_EXIT_MSG(!(ctrl & PR_PPC_DEXCR_CTRL_CLEAR_ONEXEC),
+				 "clearing aspect across exec not inherited");
+
+		FAIL_IF_EXIT_MSG(aspect & dexcr, "clearing aspect across exec did not take effect");
+	} else {
+		FAIL_IF_EXIT_MSG(true, "unknown expected status");
+	}
+
+	return 0;
+}
+
+/*
+ * Test that the given prctl value can be manipulated freely
+ */
+static int dexcr_prctl_aspect_test(unsigned long which)
+{
+	unsigned long aspect = pr_which_to_aspect(which);
+	pid_t pid;
+	int ctrl;
+	int err;
+	int errno_save;
+
+	SKIP_IF_MSG(!dexcr_exists(), "DEXCR not supported");
+	SKIP_IF_MSG(!pr_dexcr_aspect_supported(which), "DEXCR aspect not supported");
+	SKIP_IF_MSG(!pr_dexcr_aspect_editable(which), "DEXCR aspect not editable with prctl");
+
+	/* We reject invalid combinations of arguments */
+	err = pr_set_dexcr(which, PR_PPC_DEXCR_CTRL_SET | PR_PPC_DEXCR_CTRL_CLEAR);
+	errno_save = errno;
+	FAIL_IF_MSG(err != -1, "simultaneous set and clear should be rejected");
+	FAIL_IF_MSG(errno_save != EINVAL, "simultaneous set and clear should be rejected with EINVAL");
+
+	err = pr_set_dexcr(which, PR_PPC_DEXCR_CTRL_SET_ONEXEC | PR_PPC_DEXCR_CTRL_CLEAR_ONEXEC);
+	errno_save = errno;
+	FAIL_IF_MSG(err != -1, "simultaneous set and clear on exec should be rejected");
+	FAIL_IF_MSG(errno_save != EINVAL, "simultaneous set and clear on exec should be rejected with EINVAL");
+
+	/* We set the aspect */
+	err = pr_set_dexcr(which, PR_PPC_DEXCR_CTRL_SET);
+	FAIL_IF_MSG(err, "PR_PPC_DEXCR_CTRL_SET failed");
+
+	ctrl = pr_get_dexcr(which);
+	FAIL_IF_MSG(!(ctrl & PR_PPC_DEXCR_CTRL_SET), "config value not PR_PPC_DEXCR_CTRL_SET");
+	FAIL_IF_MSG(ctrl & PR_PPC_DEXCR_CTRL_CLEAR, "config value unexpected clear flag");
+	FAIL_IF_MSG(!(aspect & mfspr(SPRN_DEXCR_RO)), "setting aspect did not take effect");
+
+	/* We clear the aspect */
+	err = pr_set_dexcr(which, PR_PPC_DEXCR_CTRL_CLEAR);
+	FAIL_IF_MSG(err, "PR_PPC_DEXCR_CTRL_CLEAR failed");
+
+	ctrl = pr_get_dexcr(which);
+	FAIL_IF_MSG(!(ctrl & PR_PPC_DEXCR_CTRL_CLEAR), "config value not PR_PPC_DEXCR_CTRL_CLEAR");
+	FAIL_IF_MSG(ctrl & PR_PPC_DEXCR_CTRL_SET, "config value unexpected set flag");
+	FAIL_IF_MSG(aspect & mfspr(SPRN_DEXCR_RO), "clearing aspect did not take effect");
+
+	/* We make it set on exec (doesn't change our current value) */
+	err = pr_set_dexcr(which, PR_PPC_DEXCR_CTRL_SET_ONEXEC);
+	FAIL_IF_MSG(err, "PR_PPC_DEXCR_CTRL_SET_ONEXEC failed");
+
+	ctrl = pr_get_dexcr(which);
+	FAIL_IF_MSG(!(ctrl & PR_PPC_DEXCR_CTRL_CLEAR), "process aspect should still be cleared");
+	FAIL_IF_MSG(!(ctrl & PR_PPC_DEXCR_CTRL_SET_ONEXEC), "config value not PR_PPC_DEXCR_CTRL_SET_ONEXEC");
+	FAIL_IF_MSG(ctrl & PR_PPC_DEXCR_CTRL_CLEAR_ONEXEC, "config value unexpected clear on exec flag");
+	FAIL_IF_MSG(aspect & mfspr(SPRN_DEXCR_RO), "scheduling aspect to set on exec should not change it now");
+
+	/* We make it clear on exec (doesn't change our current value) */
+	err = pr_set_dexcr(which, PR_PPC_DEXCR_CTRL_CLEAR_ONEXEC);
+	FAIL_IF_MSG(err, "PR_PPC_DEXCR_CTRL_CLEAR_ONEXEC failed");
+
+	ctrl = pr_get_dexcr(which);
+	FAIL_IF_MSG(!(ctrl & PR_PPC_DEXCR_CTRL_CLEAR), "process aspect config should still be cleared");
+	FAIL_IF_MSG(!(ctrl & PR_PPC_DEXCR_CTRL_CLEAR_ONEXEC), "config value not PR_PPC_DEXCR_CTRL_CLEAR_ONEXEC");
+	FAIL_IF_MSG(ctrl & PR_PPC_DEXCR_CTRL_SET_ONEXEC, "config value unexpected set on exec flag");
+	FAIL_IF_MSG(aspect & mfspr(SPRN_DEXCR_RO), "process aspect should still be cleared");
+
+	/* We allow setting the current and on-exec value in a single call */
+	err = pr_set_dexcr(which, PR_PPC_DEXCR_CTRL_SET | PR_PPC_DEXCR_CTRL_CLEAR_ONEXEC);
+	FAIL_IF_MSG(err, "PR_PPC_DEXCR_CTRL_SET | PR_PPC_DEXCR_CTRL_CLEAR_ONEXEC failed");
+
+	ctrl = pr_get_dexcr(which);
+	FAIL_IF_MSG(!(ctrl & PR_PPC_DEXCR_CTRL_SET), "config value not PR_PPC_DEXCR_CTRL_SET");
+	FAIL_IF_MSG(!(ctrl & PR_PPC_DEXCR_CTRL_CLEAR_ONEXEC), "config value not PR_PPC_DEXCR_CTRL_CLEAR_ONEXEC");
+	FAIL_IF_MSG(!(aspect & mfspr(SPRN_DEXCR_RO)), "process aspect should be set");
+
+	err = pr_set_dexcr(which, PR_PPC_DEXCR_CTRL_CLEAR | PR_PPC_DEXCR_CTRL_SET_ONEXEC);
+	FAIL_IF_MSG(err, "PR_PPC_DEXCR_CTRL_CLEAR | PR_PPC_DEXCR_CTRL_SET_ONEXEC failed");
+
+	ctrl = pr_get_dexcr(which);
+	FAIL_IF_MSG(!(ctrl & PR_PPC_DEXCR_CTRL_CLEAR), "config value not PR_PPC_DEXCR_CTRL_CLEAR");
+	FAIL_IF_MSG(!(ctrl & PR_PPC_DEXCR_CTRL_SET_ONEXEC), "config value not PR_PPC_DEXCR_CTRL_SET_ONEXEC");
+	FAIL_IF_MSG(aspect & mfspr(SPRN_DEXCR_RO), "process aspect should be clear");
+
+	/* Verify the onexec value is applied across exec */
+	pid = fork();
+	if (!pid) {
+		char which_str[32] = {};
+		char *args[] = { "dexcr_prctl_onexec_test_child", which_str, "set", NULL };
+		unsigned int ctrl = pr_get_dexcr(which);
+
+		sprintf(which_str, "%lu", which);
+
+		FAIL_IF_EXIT_MSG(!(ctrl & PR_PPC_DEXCR_CTRL_SET_ONEXEC),
+				 "setting aspect on exec not copied across fork");
+
+		FAIL_IF_EXIT_MSG(mfspr(SPRN_DEXCR_RO) & aspect,
+				 "setting aspect on exec wrongly applied to fork");
+
+		execve("/proc/self/exe", args, NULL);
+		_exit(errno);
+	}
+	await_child_success(pid);
+
+	err = pr_set_dexcr(which, PR_PPC_DEXCR_CTRL_SET | PR_PPC_DEXCR_CTRL_CLEAR_ONEXEC);
+	FAIL_IF_MSG(err, "PR_PPC_DEXCR_CTRL_SET | PR_PPC_DEXCR_CTRL_CLEAR_ONEXEC failed");
+
+	pid = fork();
+	if (!pid) {
+		char which_str[32] = {};
+		char *args[] = { "dexcr_prctl_onexec_test_child", which_str, "clear", NULL };
+		unsigned int ctrl = pr_get_dexcr(which);
+
+		sprintf(which_str, "%lu", which);
+
+		FAIL_IF_EXIT_MSG(!(ctrl & PR_PPC_DEXCR_CTRL_CLEAR_ONEXEC),
+				 "clearing aspect on exec not copied across fork");
+
+		FAIL_IF_EXIT_MSG(!(mfspr(SPRN_DEXCR_RO) & aspect),
+				 "clearing aspect on exec wrongly applied to fork");
+
+		execve("/proc/self/exe", args, NULL);
+		_exit(errno);
+	}
+	await_child_success(pid);
+
+	return 0;
+}
+
+static int dexcr_prctl_ibrtpd_test(void)
+{
+	return dexcr_prctl_aspect_test(PR_PPC_DEXCR_IBRTPD);
+}
+
+static int dexcr_prctl_srapd_test(void)
+{
+	return dexcr_prctl_aspect_test(PR_PPC_DEXCR_SRAPD);
+}
+
+static int dexcr_prctl_nphie_test(void)
+{
+	return dexcr_prctl_aspect_test(PR_PPC_DEXCR_NPHIE);
+}
+
+int main(int argc, char *argv[])
+{
+	int err = 0;
+
+	/*
+	 * Some tests require checking what happens across exec, so we may be
+	 * invoked as the child of a particular test
+	 */
+	if (argc > 1) {
+		if (argc == 3 && !strcmp(argv[0], "dexcr_prctl_onexec_test_child")) {
+			unsigned long which;
+
+			err = parse_ulong(argv[1], strlen(argv[1]), &which, 10);
+			FAIL_IF_MSG(err, "failed to parse which value for child");
+
+			return dexcr_prctl_onexec_test_child(which, argv[2]);
+		}
+
+		FAIL_IF_MSG(true, "unknown test case");
+	}
+
+	/*
+	 * Otherwise we are the main test invocation and run the full suite
+	 */
+	err |= test_harness(dexcr_prctl_ibrtpd_test, "dexcr_prctl_ibrtpd");
+	err |= test_harness(dexcr_prctl_srapd_test, "dexcr_prctl_srapd");
+	err |= test_harness(dexcr_prctl_nphie_test, "dexcr_prctl_nphie");
+
+	return err;
+}

linux-7.2.3/tools/testing/selftests/powerpc/dexcr/hashchk_test.c

@@ -0,0 +1,233 @@
+// SPDX-License-Identifier: GPL-2.0+
+
+#define _GNU_SOURCE
+
+#include <errno.h>
+#include <fcntl.h>
+#include <limits.h>
+#include <sched.h>
+#include <setjmp.h>
+#include <signal.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include <sys/mman.h>
+#include <sys/prctl.h>
+#include <unistd.h>
+
+#include "dexcr.h"
+#include "utils.h"
+
+static int require_nphie(void)
+{
+	SKIP_IF_MSG(!dexcr_exists(), "DEXCR not supported");
+
+	pr_set_dexcr(PR_PPC_DEXCR_NPHIE, PR_PPC_DEXCR_CTRL_SET | PR_PPC_DEXCR_CTRL_SET_ONEXEC);
+
+	if (get_dexcr(EFFECTIVE) & DEXCR_PR_NPHIE)
+		return 0;
+
+	SKIP_IF_MSG(!(get_dexcr(EFFECTIVE) & DEXCR_PR_NPHIE),
+		    "Failed to enable DEXCR[NPHIE]");
+
+	return 0;
+}
+
+static jmp_buf hashchk_detected_buf;
+static const char *hashchk_failure_msg;
+
+static void hashchk_handler(int signum, siginfo_t *info, void *context)
+{
+	if (signum != SIGILL)
+		hashchk_failure_msg = "wrong signal received";
+	else if (info->si_code != ILL_ILLOPN)
+		hashchk_failure_msg = "wrong signal code received";
+
+	longjmp(hashchk_detected_buf, 0);
+}
+
+/*
+ * Check that hashchk triggers when DEXCR[NPHIE] is enabled
+ * and is detected as such by the kernel exception handler
+ */
+static int hashchk_detected_test(void)
+{
+	struct sigaction old;
+	int err;
+
+	err = require_nphie();
+	if (err)
+		return err;
+
+	old = push_signal_handler(SIGILL, hashchk_handler);
+	if (setjmp(hashchk_detected_buf))
+		goto out;
+
+	hashchk_failure_msg = NULL;
+	do_bad_hashchk();
+	hashchk_failure_msg = "hashchk failed to trigger";
+
+out:
+	pop_signal_handler(SIGILL, old);
+	FAIL_IF_MSG(hashchk_failure_msg, hashchk_failure_msg);
+	return 0;
+}
+
+#define HASH_COUNT 8
+
+static unsigned long hash_values[HASH_COUNT + 1];
+
+static void fill_hash_values(void)
+{
+	for (unsigned long i = 0; i < HASH_COUNT; i++)
+		hashst(i, &hash_values[i]);
+
+	/* Used to ensure the checks uses the same addresses as the hashes */
+	hash_values[HASH_COUNT] = (unsigned long)&hash_values;
+}
+
+static unsigned int count_hash_values_matches(void)
+{
+	unsigned long matches = 0;
+
+	for (unsigned long i = 0; i < HASH_COUNT; i++) {
+		unsigned long orig_hash = hash_values[i];
+		hash_values[i] = 0;
+
+		hashst(i, &hash_values[i]);
+
+		if (hash_values[i] == orig_hash)
+			matches++;
+	}
+
+	return matches;
+}
+
+static int hashchk_exec_child(void)
+{
+	ssize_t count;
+
+	fill_hash_values();
+
+	count = write(STDOUT_FILENO, hash_values, sizeof(hash_values));
+	return count == sizeof(hash_values) ? 0 : EOVERFLOW;
+}
+
+static char *hashchk_exec_child_args[] = { "hashchk_exec_child", NULL };
+
+/*
+ * Check that new programs get different keys so a malicious process
+ * can't recreate a victim's hash values.
+ */
+static int hashchk_exec_random_key_test(void)
+{
+	pid_t pid;
+	int err;
+	int pipefd[2];
+
+	err = require_nphie();
+	if (err)
+		return err;
+
+	FAIL_IF_MSG(pipe(pipefd), "failed to create pipe");
+
+	pid = fork();
+	if (pid == 0) {
+		if (dup2(pipefd[1], STDOUT_FILENO) == -1)
+			_exit(errno);
+
+		execve("/proc/self/exe", hashchk_exec_child_args, NULL);
+		_exit(errno);
+	}
+
+	await_child_success(pid);
+	FAIL_IF_MSG(read(pipefd[0], hash_values, sizeof(hash_values)) != sizeof(hash_values),
+		    "missing expected child output");
+
+	/* Verify the child used the same hash_values address */
+	FAIL_IF_EXIT_MSG(hash_values[HASH_COUNT] != (unsigned long)&hash_values,
+			 "bad address check");
+
+	/* If all hashes are the same it means (most likely) same key */
+	FAIL_IF_MSG(count_hash_values_matches() == HASH_COUNT, "shared key detected");
+
+	return 0;
+}
+
+/*
+ * Check that forks share the same key so that existing hash values
+ * remain valid.
+ */
+static int hashchk_fork_share_key_test(void)
+{
+	pid_t pid;
+	int err;
+
+	err = require_nphie();
+	if (err)
+		return err;
+
+	fill_hash_values();
+
+	pid = fork();
+	if (pid == 0) {
+		if (count_hash_values_matches() != HASH_COUNT)
+			_exit(1);
+		_exit(0);
+	}
+
+	await_child_success(pid);
+	return 0;
+}
+
+#define STACK_SIZE (1024 * 1024)
+
+static int hashchk_clone_child_fn(void *args)
+{
+	fill_hash_values();
+	return 0;
+}
+
+/*
+ * Check that threads share the same key so that existing hash values
+ * remain valid.
+ */
+static int hashchk_clone_share_key_test(void)
+{
+	void *child_stack;
+	pid_t pid;
+	int err;
+
+	err = require_nphie();
+	if (err)
+		return err;
+
+	child_stack = mmap(NULL, STACK_SIZE, PROT_READ | PROT_WRITE,
+			   MAP_PRIVATE | MAP_ANONYMOUS | MAP_STACK, -1, 0);
+
+	FAIL_IF_MSG(child_stack == MAP_FAILED, "failed to map child stack");
+
+	pid = clone(hashchk_clone_child_fn, child_stack + STACK_SIZE,
+		    CLONE_VM | SIGCHLD, NULL);
+
+	await_child_success(pid);
+	FAIL_IF_MSG(count_hash_values_matches() != HASH_COUNT,
+		    "different key detected");
+
+	return 0;
+}
+
+int main(int argc, char *argv[])
+{
+	int err = 0;
+
+	if (argc >= 1 && !strcmp(argv[0], hashchk_exec_child_args[0]))
+		return hashchk_exec_child();
+
+	err |= test_harness(hashchk_detected_test, "hashchk_detected");
+	err |= test_harness(hashchk_exec_random_key_test, "hashchk_exec_random_key");
+	err |= test_harness(hashchk_fork_share_key_test, "hashchk_fork_share_key");
+	err |= test_harness(hashchk_clone_share_key_test, "hashchk_clone_share_key");
+
+	return err;
+}

linux-7.2.3/tools/testing/selftests/powerpc/dexcr/lsdexcr.c

@@ -0,0 +1,172 @@
+// SPDX-License-Identifier: GPL-2.0+
+
+#include <stddef.h>
+#include <stdio.h>
+#include <string.h>
+#include <sys/prctl.h>
+
+#include "dexcr.h"
+#include "utils.h"
+
+static unsigned int dexcr;
+static unsigned int hdexcr;
+static unsigned int effective;
+
+static void print_list(const char *list[], size_t len)
+{
+	for (size_t i = 0; i < len; i++) {
+		printf("%s", list[i]);
+		if (i + 1 < len)
+			printf(", ");
+	}
+}
+
+static void print_dexcr(char *name, unsigned int bits)
+{
+	const char *enabled_aspects[ARRAY_SIZE(aspects) + 1] = {NULL};
+	size_t j = 0;
+
+	printf("%s: 0x%08x", name, bits);
+
+	if (bits == 0) {
+		printf("\n");
+		return;
+	}
+
+	for (size_t i = 0; i < ARRAY_SIZE(aspects); i++) {
+		unsigned int mask = DEXCR_PR_BIT(aspects[i].index);
+
+		if (bits & mask) {
+			enabled_aspects[j++] = aspects[i].name;
+			bits &= ~mask;
+		}
+	}
+
+	if (bits)
+		enabled_aspects[j++] = "unknown";
+
+	printf(" (");
+	print_list(enabled_aspects, j);
+	printf(")\n");
+}
+
+static void print_aspect(const struct dexcr_aspect *aspect)
+{
+	const char *attributes[8] = {NULL};
+	size_t j = 0;
+	unsigned long mask;
+
+	mask = DEXCR_PR_BIT(aspect->index);
+	if (dexcr & mask)
+		attributes[j++] = "set";
+	if (hdexcr & mask)
+		attributes[j++] = "set (hypervisor)";
+	if (!(effective & mask))
+		attributes[j++] = "clear";
+
+	printf("%12s %c (%d): ", aspect->name, effective & mask ? '*' : ' ', aspect->index);
+	print_list(attributes, j);
+	printf("  \t(%s)\n", aspect->desc);
+}
+
+static void print_aspect_config(const struct dexcr_aspect *aspect)
+{
+	const char *reason = NULL;
+	const char *reason_hyp = NULL;
+	const char *reason_prctl = "no prctl";
+	bool actual = effective & DEXCR_PR_BIT(aspect->index);
+	bool expected = actual;  /* Assume it's fine if we don't expect a specific set/clear value */
+
+	if (actual)
+		reason = "set by unknown";
+	else
+		reason = "cleared by unknown";
+
+	if (aspect->prctl != -1) {
+		int ctrl = pr_get_dexcr(aspect->prctl);
+
+		if (ctrl < 0) {
+			reason_prctl = "failed to read prctl";
+		} else {
+			if (ctrl & PR_PPC_DEXCR_CTRL_SET) {
+				reason_prctl = "set by prctl";
+				expected = true;
+			} else if (ctrl & PR_PPC_DEXCR_CTRL_CLEAR) {
+				reason_prctl = "cleared by prctl";
+				expected = false;
+			} else {
+				reason_prctl = "unknown prctl";
+			}
+
+			reason = reason_prctl;
+		}
+	}
+
+	if (hdexcr & DEXCR_PR_BIT(aspect->index)) {
+		reason_hyp = "set by hypervisor";
+		reason = reason_hyp;
+		expected = true;
+	} else {
+		reason_hyp = "not modified by hypervisor";
+	}
+
+	printf("%12s (%d): %-28s (%s, %s)\n",
+	       aspect->name,
+	       aspect->index,
+	       reason,
+	       reason_hyp,
+	       reason_prctl);
+
+	/*
+	 * The checks are not atomic, so this can technically trigger if the
+	 * hypervisor makes a change while we are checking each source. It's
+	 * far more likely to be a bug if we see this though.
+	 */
+	if (actual != expected)
+		printf("                : ! actual %s does not match config\n", aspect->name);
+}
+
+int main(int argc, char *argv[])
+{
+	if (!dexcr_exists()) {
+		printf("DEXCR not detected on this hardware\n");
+		return 1;
+	}
+
+	dexcr = get_dexcr(DEXCR);
+	hdexcr = get_dexcr(HDEXCR);
+	effective = dexcr | hdexcr;
+
+	printf("current status:\n");
+
+	print_dexcr("    DEXCR", dexcr);
+	print_dexcr("   HDEXCR", hdexcr);
+	print_dexcr("Effective", effective);
+	printf("\n");
+
+	for (size_t i = 0; i < ARRAY_SIZE(aspects); i++)
+		print_aspect(&aspects[i]);
+	printf("\n");
+
+	if (effective & DEXCR_PR_NPHIE) {
+		printf("DEXCR[NPHIE] enabled: hashst/hashchk ");
+		if (hashchk_triggers())
+			printf("working\n");
+		else
+			printf("failed to trigger\n");
+	} else {
+		printf("DEXCR[NPHIE] disabled: hashst/hashchk ");
+		if (hashchk_triggers())
+			printf("unexpectedly triggered\n");
+		else
+			printf("ignored\n");
+	}
+	printf("\n");
+
+	printf("configuration:\n");
+	for (size_t i = 0; i < ARRAY_SIZE(aspects); i++)
+		print_aspect_config(&aspects[i]);
+	printf("\n");
+
+	return 0;
+}

linux-7.2.3/tools/testing/selftests/powerpc/dexcr/settings

@@ -0,0 +1 @@
+timeout=130

linux-7.2.3/tools/testing/selftests/powerpc/dscr/.gitignore

@@ -0,0 +1,8 @@
+# SPDX-License-Identifier: GPL-2.0-only
+dscr_default_test
+dscr_explicit_test
+dscr_inherit_exec_test
+dscr_inherit_test
+dscr_sysfs_test
+dscr_sysfs_thread_test
+dscr_user_test

linux-7.2.3/tools/testing/selftests/powerpc/dscr/Makefile

@@ -0,0 +1,13 @@
+# SPDX-License-Identifier: GPL-2.0
+TEST_GEN_PROGS := dscr_default_test dscr_explicit_test dscr_user_test	\
+	      dscr_inherit_test dscr_inherit_exec_test dscr_sysfs_test	\
+	      dscr_sysfs_thread_test
+
+top_srcdir = ../../../../..
+include ../../lib.mk
+include ../flags.mk
+
+$(OUTPUT)/dscr_default_test: LDLIBS += -lpthread
+$(OUTPUT)/dscr_explicit_test: LDLIBS += -lpthread
+
+$(TEST_GEN_PROGS): ../harness.c ../utils.c

linux-7.2.3/tools/testing/selftests/powerpc/dscr/dscr.h

@@ -0,0 +1,89 @@
+/* SPDX-License-Identifier: GPL-2.0-only */
+/*
+ * POWER Data Stream Control Register (DSCR)
+ *
+ * This header file contains helper functions and macros
+ * required for all the DSCR related test cases.
+ *
+ * Copyright 2012, Anton Blanchard, IBM Corporation.
+ * Copyright 2015, Anshuman Khandual, IBM Corporation.
+ */
+#ifndef _SELFTESTS_POWERPC_DSCR_DSCR_H
+#define _SELFTESTS_POWERPC_DSCR_DSCR_H
+
+#include <unistd.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include <fcntl.h>
+#include <dirent.h>
+#include <pthread.h>
+#include <sched.h>
+#include <sys/types.h>
+#include <sys/stat.h>
+#include <sys/wait.h>
+
+#include "reg.h"
+#include "utils.h"
+
+#define THREADS		100	/* Max threads */
+#define COUNT		100	/* Max iterations */
+#define DSCR_MAX	16	/* Max DSCR value */
+#define LEN_MAX		100	/* Max name length */
+
+#define DSCR_DEFAULT	"/sys/devices/system/cpu/dscr_default"
+#define CPU_PATH	"/sys/devices/system/cpu/"
+
+#define rmb()  asm volatile("lwsync":::"memory")
+#define wmb()  asm volatile("lwsync":::"memory")
+
+#define READ_ONCE(x) (*(volatile typeof(x) *)&(x))
+
+/* Prilvilege state DSCR access */
+inline unsigned long get_dscr(void)
+{
+	return mfspr(SPRN_DSCR_PRIV);
+}
+
+inline void set_dscr(unsigned long val)
+{
+	mtspr(SPRN_DSCR_PRIV, val);
+}
+
+/* Problem state DSCR access */
+inline unsigned long get_dscr_usr(void)
+{
+	return mfspr(SPRN_DSCR);
+}
+
+inline void set_dscr_usr(unsigned long val)
+{
+	mtspr(SPRN_DSCR, val);
+}
+
+/* Default DSCR access */
+unsigned long get_default_dscr(void)
+{
+	int err;
+	unsigned long val;
+
+	err = read_ulong(DSCR_DEFAULT, &val, 16);
+	if (err) {
+		perror("read() failed");
+		exit(1);
+	}
+	return val;
+}
+
+void set_default_dscr(unsigned long val)
+{
+	int err;
+
+	err = write_ulong(DSCR_DEFAULT, val, 16);
+	if (err) {
+		perror("write() failed");
+		exit(1);
+	}
+}
+
+#endif	/* _SELFTESTS_POWERPC_DSCR_DSCR_H */

linux-7.2.3/tools/testing/selftests/powerpc/dscr/dscr_default_test.c

@@ -0,0 +1,169 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * POWER Data Stream Control Register (DSCR) default test
+ *
+ * This test modifies the system wide default DSCR through
+ * it's sysfs interface and then verifies that all threads
+ * see the correct changed DSCR value immediately.
+ *
+ * Copyright 2012, Anton Blanchard, IBM Corporation.
+ * Copyright 2015, Anshuman Khandual, IBM Corporation.
+ */
+
+#define _GNU_SOURCE
+
+#include "dscr.h"
+
+#include <pthread.h>
+#include <semaphore.h>
+#include <unistd.h>
+
+static void *dscr_default_lockstep_writer(void *arg)
+{
+	sem_t *reader_sem = (sem_t *)arg;
+	sem_t *writer_sem = (sem_t *)arg + 1;
+	unsigned long expected_dscr = 0;
+
+	for (int i = 0; i < COUNT; i++) {
+		FAIL_IF_EXIT(sem_wait(writer_sem));
+
+		set_default_dscr(expected_dscr);
+		expected_dscr = (expected_dscr + 1) % DSCR_MAX;
+
+		FAIL_IF_EXIT(sem_post(reader_sem));
+	}
+
+	return NULL;
+}
+
+int dscr_default_lockstep_test(void)
+{
+	pthread_t writer;
+	sem_t rw_semaphores[2];
+	sem_t *reader_sem = &rw_semaphores[0];
+	sem_t *writer_sem = &rw_semaphores[1];
+	unsigned long expected_dscr = 0;
+
+	SKIP_IF(!have_hwcap2(PPC_FEATURE2_DSCR));
+
+	FAIL_IF(sem_init(reader_sem, 0, 0));
+	FAIL_IF(sem_init(writer_sem, 0, 1));  /* writer starts first */
+	FAIL_IF(bind_to_cpu(BIND_CPU_ANY) < 0);
+	FAIL_IF(pthread_create(&writer, NULL, dscr_default_lockstep_writer, (void *)rw_semaphores));
+
+	for (int i = 0; i < COUNT ; i++) {
+		FAIL_IF(sem_wait(reader_sem));
+
+		FAIL_IF(get_dscr() != expected_dscr);
+		FAIL_IF(get_dscr_usr() != expected_dscr);
+
+		expected_dscr = (expected_dscr + 1) % DSCR_MAX;
+
+		FAIL_IF(sem_post(writer_sem));
+	}
+
+	FAIL_IF(pthread_join(writer, NULL));
+	FAIL_IF(sem_destroy(reader_sem));
+	FAIL_IF(sem_destroy(writer_sem));
+
+	return 0;
+}
+
+struct random_thread_args {
+	pthread_t thread_id;
+	unsigned long *expected_system_dscr;
+	pthread_rwlock_t *rw_lock;
+	pthread_barrier_t *barrier;
+};
+
+static void *dscr_default_random_thread(void *in)
+{
+	struct random_thread_args *args = (struct random_thread_args *)in;
+	unsigned long *expected_dscr_p = args->expected_system_dscr;
+	pthread_rwlock_t *rw_lock = args->rw_lock;
+	int err;
+
+	srand(gettid());
+
+	err = pthread_barrier_wait(args->barrier);
+	FAIL_IF_EXIT(err != 0 && err != PTHREAD_BARRIER_SERIAL_THREAD);
+
+	for (int i = 0; i < COUNT; i++) {
+		unsigned long expected_dscr;
+		unsigned long current_dscr;
+		unsigned long current_dscr_usr;
+
+		FAIL_IF_EXIT(pthread_rwlock_rdlock(rw_lock));
+		expected_dscr = *expected_dscr_p;
+		current_dscr = get_dscr();
+		current_dscr_usr = get_dscr_usr();
+		FAIL_IF_EXIT(pthread_rwlock_unlock(rw_lock));
+
+		FAIL_IF_EXIT(current_dscr != expected_dscr);
+		FAIL_IF_EXIT(current_dscr_usr != expected_dscr);
+
+		if (rand() % 10 == 0) {
+			unsigned long next_dscr;
+
+			FAIL_IF_EXIT(pthread_rwlock_wrlock(rw_lock));
+			next_dscr = (*expected_dscr_p + 1) % DSCR_MAX;
+			set_default_dscr(next_dscr);
+			*expected_dscr_p = next_dscr;
+			FAIL_IF_EXIT(pthread_rwlock_unlock(rw_lock));
+		}
+	}
+
+	pthread_exit((void *)0);
+}
+
+int dscr_default_random_test(void)
+{
+	struct random_thread_args threads[THREADS];
+	unsigned long expected_system_dscr = 0;
+	pthread_rwlockattr_t rwlock_attr;
+	pthread_rwlock_t rw_lock;
+	pthread_barrier_t barrier;
+
+	SKIP_IF(!have_hwcap2(PPC_FEATURE2_DSCR));
+
+	FAIL_IF(pthread_rwlockattr_setkind_np(&rwlock_attr,
+					      PTHREAD_RWLOCK_PREFER_WRITER_NONRECURSIVE_NP));
+	FAIL_IF(pthread_rwlock_init(&rw_lock, &rwlock_attr));
+	FAIL_IF(pthread_barrier_init(&barrier, NULL, THREADS));
+
+	set_default_dscr(expected_system_dscr);
+
+	for (int i = 0; i < THREADS; i++) {
+		threads[i].expected_system_dscr = &expected_system_dscr;
+		threads[i].rw_lock = &rw_lock;
+		threads[i].barrier = &barrier;
+
+		FAIL_IF(pthread_create(&threads[i].thread_id, NULL,
+				       dscr_default_random_thread, (void *)&threads[i]));
+	}
+
+	for (int i = 0; i < THREADS; i++)
+		FAIL_IF(pthread_join(threads[i].thread_id, NULL));
+
+	FAIL_IF(pthread_barrier_destroy(&barrier));
+	FAIL_IF(pthread_rwlock_destroy(&rw_lock));
+
+	return 0;
+}
+
+int main(int argc, char *argv[])
+{
+	unsigned long orig_dscr_default = 0;
+	int err = 0;
+
+	if (have_hwcap2(PPC_FEATURE2_DSCR))
+		orig_dscr_default = get_default_dscr();
+
+	err |= test_harness(dscr_default_lockstep_test, "dscr_default_lockstep_test");
+	err |= test_harness(dscr_default_random_test, "dscr_default_random_test");
+
+	if (have_hwcap2(PPC_FEATURE2_DSCR))
+		set_default_dscr(orig_dscr_default);
+
+	return err;
+}

linux-7.2.3/tools/testing/selftests/powerpc/dscr/dscr_explicit_test.c

@@ -0,0 +1,173 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * POWER Data Stream Control Register (DSCR) explicit test
+ *
+ * This test modifies the DSCR value using mtspr instruction and
+ * verifies the change with mfspr instruction. It uses both the
+ * privilege state SPR and the problem state SPR for this purpose.
+ *
+ * When using the privilege state SPR, the instructions such as
+ * mfspr or mtspr are privileged and the kernel emulates them
+ * for us. Instructions using problem state SPR can be executed
+ * directly without any emulation if the HW supports them. Else
+ * they also get emulated by the kernel.
+ *
+ * Copyright 2012, Anton Blanchard, IBM Corporation.
+ * Copyright 2015, Anshuman Khandual, IBM Corporation.
+ */
+
+#define _GNU_SOURCE
+
+#include "dscr.h"
+#include "utils.h"
+
+#include <pthread.h>
+#include <sched.h>
+#include <semaphore.h>
+
+void *dscr_explicit_lockstep_thread(void *args)
+{
+	sem_t *prev = (sem_t *)args;
+	sem_t *next = (sem_t *)args + 1;
+	unsigned long expected_dscr = 0;
+
+	set_dscr(expected_dscr);
+	srand(gettid());
+
+	for (int i = 0; i < COUNT; i++) {
+		FAIL_IF_EXIT(sem_wait(prev));
+
+		FAIL_IF_EXIT(expected_dscr != get_dscr());
+		FAIL_IF_EXIT(expected_dscr != get_dscr_usr());
+
+		expected_dscr = (expected_dscr + 1) % DSCR_MAX;
+		set_dscr(expected_dscr);
+
+		FAIL_IF_EXIT(sem_post(next));
+	}
+
+	return NULL;
+}
+
+int dscr_explicit_lockstep_test(void)
+{
+	pthread_t thread;
+	sem_t semaphores[2];
+	sem_t *prev = &semaphores[1];  /* reversed prev/next than for the other thread */
+	sem_t *next = &semaphores[0];
+	unsigned long expected_dscr = 0;
+
+	SKIP_IF(!have_hwcap2(PPC_FEATURE2_DSCR));
+
+	srand(gettid());
+	set_dscr(expected_dscr);
+
+	FAIL_IF(sem_init(prev, 0, 0));
+	FAIL_IF(sem_init(next, 0, 1));  /* other thread starts first */
+	FAIL_IF(bind_to_cpu(BIND_CPU_ANY) < 0);
+	FAIL_IF(pthread_create(&thread, NULL, dscr_explicit_lockstep_thread, (void *)semaphores));
+
+	for (int i = 0; i < COUNT; i++) {
+		FAIL_IF(sem_wait(prev));
+
+		FAIL_IF(expected_dscr != get_dscr());
+		FAIL_IF(expected_dscr != get_dscr_usr());
+
+		expected_dscr = (expected_dscr - 1) % DSCR_MAX;
+		set_dscr(expected_dscr);
+
+		FAIL_IF(sem_post(next));
+	}
+
+	FAIL_IF(pthread_join(thread, NULL));
+	FAIL_IF(sem_destroy(prev));
+	FAIL_IF(sem_destroy(next));
+
+	return 0;
+}
+
+struct random_thread_args {
+	pthread_t thread_id;
+	bool do_yields;
+	pthread_barrier_t *barrier;
+};
+
+void *dscr_explicit_random_thread(void *in)
+{
+	struct random_thread_args *args = (struct random_thread_args *)in;
+	unsigned long expected_dscr = 0;
+	int err;
+
+	srand(gettid());
+
+	err = pthread_barrier_wait(args->barrier);
+	FAIL_IF_EXIT(err != 0 && err != PTHREAD_BARRIER_SERIAL_THREAD);
+
+	for (int i = 0; i < COUNT; i++) {
+		expected_dscr = rand() % DSCR_MAX;
+		set_dscr(expected_dscr);
+
+		for (int j = rand() % 5; j > 0; --j) {
+			FAIL_IF_EXIT(get_dscr() != expected_dscr);
+			FAIL_IF_EXIT(get_dscr_usr() != expected_dscr);
+
+			if (args->do_yields && rand() % 2)
+				sched_yield();
+		}
+
+		expected_dscr = rand() % DSCR_MAX;
+		set_dscr_usr(expected_dscr);
+
+		for (int j = rand() % 5; j > 0; --j) {
+			FAIL_IF_EXIT(get_dscr() != expected_dscr);
+			FAIL_IF_EXIT(get_dscr_usr() != expected_dscr);
+
+			if (args->do_yields && rand() % 2)
+				sched_yield();
+		}
+	}
+
+	return NULL;
+}
+
+int dscr_explicit_random_test(void)
+{
+	struct random_thread_args threads[THREADS];
+	pthread_barrier_t barrier;
+
+	SKIP_IF(!have_hwcap2(PPC_FEATURE2_DSCR));
+
+	FAIL_IF(pthread_barrier_init(&barrier, NULL, THREADS));
+
+	for (int i = 0; i < THREADS; i++) {
+		threads[i].do_yields = i % 2 == 0;
+		threads[i].barrier = &barrier;
+
+		FAIL_IF(pthread_create(&threads[i].thread_id, NULL,
+				       dscr_explicit_random_thread, (void *)&threads[i]));
+	}
+
+	for (int i = 0; i < THREADS; i++)
+		FAIL_IF(pthread_join(threads[i].thread_id, NULL));
+
+	FAIL_IF(pthread_barrier_destroy(&barrier));
+
+	return 0;
+}
+
+int main(int argc, char *argv[])
+{
+	unsigned long orig_dscr_default = 0;
+	int err = 0;
+
+	if (have_hwcap2(PPC_FEATURE2_DSCR))
+		orig_dscr_default = get_default_dscr();
+
+	err |= test_harness(dscr_explicit_lockstep_test, "dscr_explicit_lockstep_test");
+	err |= test_harness(dscr_explicit_random_test, "dscr_explicit_random_test");
+
+	if (have_hwcap2(PPC_FEATURE2_DSCR))
+		set_default_dscr(orig_dscr_default);
+
+	return err;
+}

linux-7.2.3/tools/testing/selftests/powerpc/dscr/dscr_inherit_exec_test.c

@@ -0,0 +1,108 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * POWER Data Stream Control Register (DSCR) fork exec test
+ *
+ * This testcase modifies the DSCR using mtspr, forks & execs and
+ * verifies that the child is using the changed DSCR using mfspr.
+ *
+ * When using the privilege state SPR, the instructions such as
+ * mfspr or mtspr are privileged and the kernel emulates them
+ * for us. Instructions using problem state SPR can be executed
+ * directly without any emulation if the HW supports them. Else
+ * they also get emulated by the kernel.
+ *
+ * Copyright 2012, Anton Blanchard, IBM Corporation.
+ * Copyright 2015, Anshuman Khandual, IBM Corporation.
+ */
+#include "dscr.h"
+
+static char *prog;
+
+static void do_exec(unsigned long parent_dscr)
+{
+	unsigned long cur_dscr, cur_dscr_usr;
+
+	cur_dscr = get_dscr();
+	cur_dscr_usr = get_dscr_usr();
+
+	if (cur_dscr != parent_dscr) {
+		fprintf(stderr, "Parent DSCR %ld was not inherited "
+				"over exec (kernel value)\n", parent_dscr);
+		exit(1);
+	}
+
+	if (cur_dscr_usr != parent_dscr) {
+		fprintf(stderr, "Parent DSCR %ld was not inherited "
+				"over exec (user value)\n", parent_dscr);
+		exit(1);
+	}
+	exit(0);
+}
+
+int dscr_inherit_exec(void)
+{
+	unsigned long i, dscr = 0;
+	pid_t pid;
+
+	SKIP_IF(!have_hwcap2(PPC_FEATURE2_DSCR));
+
+	for (i = 0; i < COUNT; i++) {
+		dscr++;
+		if (dscr > DSCR_MAX)
+			dscr = 0;
+
+		if (dscr == get_default_dscr())
+			continue;
+
+		if (i % 2 == 0)
+			set_dscr_usr(dscr);
+		else
+			set_dscr(dscr);
+
+		pid = fork();
+		if (pid == -1) {
+			perror("fork() failed");
+			exit(1);
+		} else if (pid) {
+			int status;
+
+			if (waitpid(pid, &status, 0) == -1) {
+				perror("waitpid() failed");
+				exit(1);
+			}
+
+			if (!WIFEXITED(status)) {
+				fprintf(stderr, "Child didn't exit cleanly\n");
+				exit(1);
+			}
+
+			if (WEXITSTATUS(status) != 0) {
+				fprintf(stderr, "Child didn't exit cleanly\n");
+				return 1;
+			}
+		} else {
+			char dscr_str[16];
+
+			sprintf(dscr_str, "%ld", dscr);
+			execlp(prog, prog, "exec", dscr_str, NULL);
+			exit(1);
+		}
+	}
+	return 0;
+}
+
+int main(int argc, char *argv[])
+{
+	if (argc == 3 && !strcmp(argv[1], "exec")) {
+		unsigned long parent_dscr;
+
+		parent_dscr = atoi(argv[2]);
+		do_exec(parent_dscr);
+	} else if (argc != 1) {
+		fprintf(stderr, "Usage: %s\n", argv[0]);
+		exit(1);
+	}
+
+	prog = argv[0];
+	return test_harness(dscr_inherit_exec, "dscr_inherit_exec_test");
+}

linux-7.2.3/tools/testing/selftests/powerpc/dscr/dscr_inherit_test.c

@@ -0,0 +1,86 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * POWER Data Stream Control Register (DSCR) fork test
+ *
+ * This testcase modifies the DSCR using mtspr, forks and then
+ * verifies that the child process has the correct changed DSCR
+ * value using mfspr.
+ *
+ * When using the privilege state SPR, the instructions such as
+ * mfspr or mtspr are privileged and the kernel emulates them
+ * for us. Instructions using problem state SPR can be executed
+ * directly without any emulation if the HW supports them. Else
+ * they also get emulated by the kernel.
+ *
+ * Copyright 2012, Anton Blanchard, IBM Corporation.
+ * Copyright 2015, Anshuman Khandual, IBM Corporation.
+ */
+#include "dscr.h"
+
+int dscr_inherit(void)
+{
+	unsigned long i, dscr = 0;
+	pid_t pid;
+
+	SKIP_IF(!have_hwcap2(PPC_FEATURE2_DSCR));
+
+	srand(getpid());
+	set_dscr(dscr);
+
+	for (i = 0; i < COUNT; i++) {
+		unsigned long cur_dscr, cur_dscr_usr;
+
+		dscr++;
+		if (dscr > DSCR_MAX)
+			dscr = 0;
+
+		if (i % 2 == 0)
+			set_dscr_usr(dscr);
+		else
+			set_dscr(dscr);
+
+		pid = fork();
+		if (pid == -1) {
+			perror("fork() failed");
+			exit(1);
+		} else if (pid) {
+			int status;
+
+			if (waitpid(pid, &status, 0) == -1) {
+				perror("waitpid() failed");
+				exit(1);
+			}
+
+			if (!WIFEXITED(status)) {
+				fprintf(stderr, "Child didn't exit cleanly\n");
+				exit(1);
+			}
+
+			if (WEXITSTATUS(status) != 0) {
+				fprintf(stderr, "Child didn't exit cleanly\n");
+				return 1;
+			}
+		} else {
+			cur_dscr = get_dscr();
+			if (cur_dscr != dscr) {
+				fprintf(stderr, "Kernel DSCR should be %ld "
+					"but is %ld\n", dscr, cur_dscr);
+				exit(1);
+			}
+
+			cur_dscr_usr = get_dscr_usr();
+			if (cur_dscr_usr != dscr) {
+				fprintf(stderr, "User DSCR should be %ld "
+					"but is %ld\n", dscr, cur_dscr_usr);
+				exit(1);
+			}
+			exit(0);
+		}
+	}
+	return 0;
+}
+
+int main(int argc, char *argv[])
+{
+	return test_harness(dscr_inherit, "dscr_inherit_test");
+}

linux-7.2.3/tools/testing/selftests/powerpc/dscr/dscr_sysfs_test.c

@@ -0,0 +1,89 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * POWER Data Stream Control Register (DSCR) sysfs interface test
+ *
+ * This test updates to system wide DSCR default through the sysfs interface
+ * and then verifies that all the CPU specific DSCR defaults are updated as
+ * well verified from their sysfs interfaces.
+ *
+ * Copyright 2015, Anshuman Khandual, IBM Corporation.
+ */
+#include "dscr.h"
+
+static int check_cpu_dscr_default(char *file, unsigned long val)
+{
+	unsigned long cpu_dscr;
+	int err;
+
+	err = read_ulong(file, &cpu_dscr, 16);
+	if (err)
+		return err;
+
+	if (cpu_dscr != val) {
+		printf("DSCR match failed: %ld (system) %ld (cpu)\n",
+					val, cpu_dscr);
+		return 1;
+	}
+	return 0;
+}
+
+static int check_all_cpu_dscr_defaults(unsigned long val)
+{
+	DIR *sysfs;
+	struct dirent *dp;
+	char file[LEN_MAX];
+
+	sysfs = opendir(CPU_PATH);
+	if (!sysfs) {
+		perror("opendir() failed");
+		return 1;
+	}
+
+	while ((dp = readdir(sysfs))) {
+		int len;
+
+		if (!(dp->d_type & DT_DIR))
+			continue;
+		if (!strcmp(dp->d_name, "cpuidle"))
+			continue;
+		if (!strstr(dp->d_name, "cpu"))
+			continue;
+
+		len = snprintf(file, LEN_MAX, "%s%s/dscr", CPU_PATH, dp->d_name);
+		if (len >= LEN_MAX)
+			continue;
+		if (access(file, F_OK))
+			continue;
+
+		if (check_cpu_dscr_default(file, val)) {
+			closedir(sysfs);
+			return 1;
+		}
+	}
+	closedir(sysfs);
+	return 0;
+}
+
+int dscr_sysfs(void)
+{
+	unsigned long orig_dscr_default;
+
+	SKIP_IF(!have_hwcap2(PPC_FEATURE2_DSCR));
+
+	orig_dscr_default = get_default_dscr();
+	for (int i = 0; i < DSCR_MAX; i++) {
+		set_default_dscr(i);
+		if (check_all_cpu_dscr_defaults(i))
+			goto fail;
+	}
+	set_default_dscr(orig_dscr_default);
+	return 0;
+fail:
+	set_default_dscr(orig_dscr_default);
+	return 1;
+}
+
+int main(int argc, char *argv[])
+{
+	return test_harness(dscr_sysfs, "dscr_sysfs_test");
+}

linux-7.2.3/tools/testing/selftests/powerpc/dscr/dscr_sysfs_thread_test.c

@@ -0,0 +1,79 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * POWER Data Stream Control Register (DSCR) sysfs thread test
+ *
+ * This test updates the system wide DSCR default value through
+ * sysfs interface which should then update all the CPU specific
+ * DSCR default values which must also be then visible to threads
+ * executing on individual CPUs on the system.
+ *
+ * Copyright 2015, Anshuman Khandual, IBM Corporation.
+ */
+#define _GNU_SOURCE
+#include "dscr.h"
+
+static int test_thread_dscr(unsigned long val)
+{
+	unsigned long cur_dscr, cur_dscr_usr;
+
+	cur_dscr = get_dscr();
+	cur_dscr_usr = get_dscr_usr();
+
+	if (val != cur_dscr) {
+		printf("[cpu %d] Kernel DSCR should be %ld but is %ld\n",
+					sched_getcpu(), val, cur_dscr);
+		return 1;
+	}
+
+	if (val != cur_dscr_usr) {
+		printf("[cpu %d] User DSCR should be %ld but is %ld\n",
+					sched_getcpu(), val, cur_dscr_usr);
+		return 1;
+	}
+	return 0;
+}
+
+static int check_cpu_dscr_thread(unsigned long val)
+{
+	cpu_set_t mask;
+	int cpu;
+
+	for (cpu = 0; cpu < CPU_SETSIZE; cpu++) {
+		CPU_ZERO(&mask);
+		CPU_SET(cpu, &mask);
+		if (sched_setaffinity(0, sizeof(mask), &mask))
+			continue;
+
+		if (test_thread_dscr(val))
+			return 1;
+	}
+	return 0;
+
+}
+
+int dscr_sysfs_thread(void)
+{
+	unsigned long orig_dscr_default;
+	int i, j;
+
+	SKIP_IF(!have_hwcap2(PPC_FEATURE2_DSCR));
+
+	orig_dscr_default = get_default_dscr();
+	for (i = 0; i < COUNT; i++) {
+		for (j = 0; j < DSCR_MAX; j++) {
+			set_default_dscr(j);
+			if (check_cpu_dscr_thread(j))
+				goto fail;
+		}
+	}
+	set_default_dscr(orig_dscr_default);
+	return 0;
+fail:
+	set_default_dscr(orig_dscr_default);
+	return 1;
+}
+
+int main(int argc, char *argv[])
+{
+	return test_harness(dscr_sysfs_thread, "dscr_sysfs_thread_test");
+}

linux-7.2.3/tools/testing/selftests/powerpc/dscr/dscr_user_test.c

@@ -0,0 +1,60 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * POWER Data Stream Control Register (DSCR) SPR test
+ *
+ * This test modifies the DSCR value through both the SPR number
+ * based mtspr instruction and then makes sure that the same is
+ * reflected through mfspr instruction using either of the SPR
+ * numbers.
+ *
+ * When using the privilege state SPR, the instructions such as
+ * mfspr or mtspr are privileged and the kernel emulates them
+ * for us. Instructions using problem state SPR can be executed
+ * directly without any emulation if the HW supports them. Else
+ * they also get emulated by the kernel.
+ *
+ * Copyright 2013, Anton Blanchard, IBM Corporation.
+ * Copyright 2015, Anshuman Khandual, IBM Corporation.
+ */
+#include "dscr.h"
+
+static int check_dscr(char *str)
+{
+	unsigned long cur_dscr, cur_dscr_usr;
+
+	cur_dscr = get_dscr();
+	cur_dscr_usr = get_dscr_usr();
+	if (cur_dscr != cur_dscr_usr) {
+		printf("%s set, kernel get %lx != user get %lx\n",
+					str, cur_dscr, cur_dscr_usr);
+		return 1;
+	}
+	return 0;
+}
+
+int dscr_user(void)
+{
+	int i;
+
+	SKIP_IF(!have_hwcap2(PPC_FEATURE2_DSCR));
+
+	check_dscr("");
+
+	for (i = 0; i < COUNT; i++) {
+		set_dscr(i);
+		if (check_dscr("kernel"))
+			return 1;
+	}
+
+	for (i = 0; i < COUNT; i++) {
+		set_dscr_usr(i);
+		if (check_dscr("user"))
+			return 1;
+	}
+	return 0;
+}
+
+int main(int argc, char *argv[])
+{
+	return test_harness(dscr_user, "dscr_user_test");
+}

linux-7.2.3/tools/testing/selftests/powerpc/dscr/settings

@@ -0,0 +1 @@
+timeout=130

linux-7.2.3/tools/testing/selftests/powerpc/eeh/Makefile

@@ -0,0 +1,10 @@
+# SPDX-License-Identifier: GPL-2.0
+noarg:
+	$(MAKE) -C ../
+
+TEST_PROGS := eeh-basic.sh
+TEST_FILES := eeh-functions.sh settings
+
+top_srcdir = ../../../../..
+include ../../lib.mk
+include ../flags.mk

linux-7.2.3/tools/testing/selftests/powerpc/eeh/eeh-basic.sh

@@ -0,0 +1,57 @@
+#!/bin/sh
+# SPDX-License-Identifier: GPL-2.0-only
+
+. ./eeh-functions.sh
+
+eeh_test_prep # NB: may exit
+
+pre_lspci=`mktemp`
+lspci > $pre_lspci
+
+# record the devices that we break in here. Assuming everything
+# goes to plan we should get them back once the recover process
+# is finished.
+devices=""
+
+# Build up a list of candidate devices.
+for dev in `ls -1 /sys/bus/pci/devices/ | grep '\.0$'` ; do
+	if ! eeh_can_break $dev ; then
+		continue;
+	fi
+
+	# Skip VFs for now since we don't have a reliable way to break them.
+	if [ -e "/sys/bus/pci/devices/$dev/physfn" ] ; then
+		echo "$dev, Skipped: virtfn"
+		continue;
+	fi
+
+	echo "$dev, Added"
+
+	# Add to this list of device to check
+	devices="$devices $dev"
+done
+
+dev_count="$(echo $devices | wc -w)"
+echo "Found ${dev_count} breakable devices..."
+
+failed=0
+for dev in $devices ; do
+	echo "Breaking $dev..."
+
+	if ! pe_ok $dev ; then
+		echo "Skipping $dev, Initial PE state is not ok"
+		failed="$((failed + 1))"
+		continue;
+	fi
+
+	if ! eeh_one_dev $dev ; then
+		failed="$((failed + 1))"
+	fi
+done
+
+echo "$failed devices failed to recover ($dev_count tested)"
+lspci | diff -u $pre_lspci -
+rm -f $pre_lspci
+
+test "$failed" -eq 0
+exit $?

linux-7.2.3/tools/testing/selftests/powerpc/eeh/eeh-functions.sh

@@ -0,0 +1,245 @@
+#!/bin/sh
+# SPDX-License-Identifier: GPL-2.0-only
+
+export KSELFTESTS_SKIP=4
+
+log() {
+	echo >/dev/stderr $*
+}
+
+pe_ok() {
+	local dev="$1"
+	local path="/sys/bus/pci/devices/$dev/eeh_pe_state"
+
+	# if a driver doesn't support the error handling callbacks then the
+	# device is recovered by removing and re-probing it. This causes the
+	# sysfs directory to disappear so read the PE state once and squash
+	# any potential error messages
+	local eeh_state="$(cat $path 2>/dev/null)"
+	if [ -z "$eeh_state" ]; then
+		return 1;
+	fi
+
+	local fw_state="$(echo $eeh_state | cut -d' ' -f1)"
+	local sw_state="$(echo $eeh_state | cut -d' ' -f2)"
+
+	# If EEH_PE_ISOLATED or EEH_PE_RECOVERING are set then the PE is in an
+	# error state or being recovered. Either way, not ok.
+	if [ "$((sw_state & 0x3))" -ne 0 ] ; then
+		return 1
+	fi
+
+	# A functioning PE should have the EEH_STATE_MMIO_ACTIVE and
+	# EEH_STATE_DMA_ACTIVE flags set. For some goddamn stupid reason
+	# the platform backends set these when the PE is in reset. The
+	# RECOVERING check above should stop any false positives though.
+	if [ "$((fw_state & 0x18))" -ne "$((0x18))" ] ; then
+		return 1
+	fi
+
+	return 0;
+}
+
+eeh_supported() {
+	test -e /proc/powerpc/eeh && \
+	grep -q 'EEH Subsystem is enabled' /proc/powerpc/eeh
+}
+
+eeh_test_prep() {
+	if ! eeh_supported ; then
+		echo "EEH not supported on this system, skipping"
+		exit $KSELFTESTS_SKIP;
+	fi
+
+	if [ ! -e "/sys/kernel/debug/powerpc/eeh_dev_check" ] && \
+	   [ ! -e "/sys/kernel/debug/powerpc/eeh_dev_break" ] ; then
+		log "debugfs EEH testing files are missing. Is debugfs mounted?"
+		exit $KSELFTESTS_SKIP;
+	fi
+
+	# Bump the max freeze count to something absurd so we don't
+	# trip over it while breaking things.
+	echo 5000 > /sys/kernel/debug/powerpc/eeh_max_freezes
+}
+
+eeh_can_break() {
+	# skip bridges since we can't recover them (yet...)
+	if [ -e "/sys/bus/pci/devices/$dev/pci_bus" ] ; then
+		log "$dev, Skipped: bridge"
+		return 1;
+	fi
+
+	# The ahci driver doesn't support error recovery. If the ahci device
+	# happens to be hosting the root filesystem, and then we go and break
+	# it the system will generally go down. We should probably fix that
+	# at some point
+	if [ "ahci" = "$(basename $(realpath /sys/bus/pci/devices/$dev/driver))" ] ; then
+		log "$dev, Skipped: ahci doesn't support recovery"
+		return 1;
+	fi
+
+	# Don't inject errosr into an already-frozen PE. This happens with
+	# PEs that contain multiple PCI devices (e.g. multi-function cards)
+	# and injecting new errors during the recovery process will probably
+	# result in the recovery failing and the device being marked as
+	# failed.
+	if ! pe_ok $dev ; then
+		log "$dev, Skipped: Bad initial PE state"
+		return 1;
+	fi
+
+	return 0
+}
+
+eeh_one_dev() {
+	local dev="$1"
+
+	# Using this function from the command line is sometimes useful for
+	# testing so check that the argument is a well-formed sysfs device
+	# name.
+	if ! test -e /sys/bus/pci/devices/$dev/ ; then
+		log "Error: '$dev' must be a sysfs device name (DDDD:BB:DD.F)"
+		return 1;
+	fi
+
+	# Break it
+	echo $dev >/sys/kernel/debug/powerpc/eeh_dev_break
+
+	# Force an EEH device check. If the kernel has already
+	# noticed the EEH (due to a driver poll or whatever), this
+	# is a no-op.
+	echo $dev >/sys/kernel/debug/powerpc/eeh_dev_check
+
+	# Default to a 60s timeout when waiting for a device to recover. This
+	# is an arbitrary default which can be overridden by setting the
+	# EEH_MAX_WAIT environmental variable when required.
+
+	# The current record holder for longest recovery time is:
+	#  "Adaptec Series 8 12G SAS/PCIe 3" at 39 seconds
+	max_wait=${EEH_MAX_WAIT:=60}
+
+	for i in `seq 0 ${max_wait}` ; do
+		if pe_ok $dev ; then
+			break;
+		fi
+		log "$dev, waited $i/${max_wait}"
+		sleep 1
+	done
+
+	if ! pe_ok $dev ; then
+		log "$dev, Failed to recover!"
+		return 1;
+	fi
+
+	log "$dev, Recovered after $i seconds"
+	return 0;
+}
+
+eeh_has_driver() {
+	test -e /sys/bus/pci/devices/$1/driver;
+	return $?
+}
+
+eeh_can_recover() {
+	# we'll get an IO error if the device's current driver doesn't support
+	# error recovery
+	echo $1 > '/sys/kernel/debug/powerpc/eeh_dev_can_recover' 2>/dev/null
+
+	return $?
+}
+
+eeh_find_all_pfs() {
+	devices=""
+
+	# SR-IOV on pseries requires hypervisor support, so check for that
+	is_pseries=""
+	if grep -q pSeries /proc/cpuinfo ; then
+		if [ ! -f /proc/device-tree/rtas/ibm,open-sriov-allow-unfreeze ] ||
+		   [ ! -f /proc/device-tree/rtas/ibm,open-sriov-map-pe-number ] ; then
+			return 1;
+		fi
+
+		is_pseries="true"
+	fi
+
+	for dev in `ls -1 /sys/bus/pci/devices/` ; do
+		sysfs="/sys/bus/pci/devices/$dev"
+		if [ ! -e "$sysfs/sriov_numvfs" ] ; then
+			continue
+		fi
+
+		# skip unsupported PFs on pseries
+		if [ -z "$is_pseries" ] &&
+		   [ ! -f "$sysfs/of_node/ibm,is-open-sriov-pf" ] &&
+		   [ ! -f "$sysfs/of_node/ibm,open-sriov-vf-bar-info" ] ; then
+			continue;
+		fi
+
+		# no driver, no vfs
+		if ! eeh_has_driver $dev ; then
+			continue
+		fi
+
+		devices="$devices $dev"
+	done
+
+	if [ -z "$devices" ] ; then
+		return 1;
+	fi
+
+	echo $devices
+	return 0;
+}
+
+# attempts to enable one VF on each PF so we can do VF specific tests.
+# stdout: list of enabled VFs, one per line
+# return code: 0 if vfs are found, 1 otherwise
+eeh_enable_vfs() {
+	pf_list="$(eeh_find_all_pfs)"
+
+	vfs=0
+	for dev in $pf_list ; do
+		pf_sysfs="/sys/bus/pci/devices/$dev"
+
+		# make sure we have a single VF
+		echo 0 > "$pf_sysfs/sriov_numvfs"
+		echo 1 > "$pf_sysfs/sriov_numvfs"
+		if [ "$?" != 0 ] ; then
+			log "Unable to enable VFs on $pf, skipping"
+			continue;
+		fi
+
+		vf="$(basename $(realpath "$pf_sysfs/virtfn0"))"
+		if [ $? != 0 ] ; then
+			log "unable to find enabled vf on $pf"
+			echo 0 > "$pf_sysfs/sriov_numvfs"
+			continue;
+		fi
+
+		if ! eeh_can_break $vf ; then
+			log "skipping "
+
+			echo 0 > "$pf_sysfs/sriov_numvfs"
+			continue;
+		fi
+
+		vfs="$((vfs + 1))"
+		echo $vf
+	done
+
+	test "$vfs" != 0
+	return $?
+}
+
+eeh_disable_vfs() {
+	pf_list="$(eeh_find_all_pfs)"
+	if [ -z "$pf_list" ] ; then
+		return 1;
+	fi
+
+	for dev in $pf_list ; do
+		echo 0 > "/sys/bus/pci/devices/$dev/sriov_numvfs"
+	done
+
+	return 0;
+}

linux-7.2.3/tools/testing/selftests/powerpc/eeh/eeh-vf-aware.sh

@@ -0,0 +1,45 @@
+#!/bin/sh
+# SPDX-License-Identifier: GPL-2.0-only
+
+. ./eeh-functions.sh
+
+eeh_test_prep # NB: may exit
+
+vf_list="$(eeh_enable_vfs)";
+if $? != 0 ; then
+	log "No usable VFs found. Skipping EEH unaware VF test"
+	exit $KSELFTESTS_SKIP;
+fi
+
+log "Enabled VFs: $vf_list"
+
+tested=0
+passed=0
+for vf in $vf_list ; do
+	log "Testing $vf"
+
+	if ! eeh_can_recover $vf ; then
+		log "Driver for $vf doesn't support error recovery, skipping"
+		continue;
+	fi
+
+	tested="$((tested + 1))"
+
+	log "Breaking $vf..."
+	if ! eeh_one_dev $vf ; then
+		log "$vf failed to recover"
+		continue;
+	fi
+
+	passed="$((passed + 1))"
+done
+
+eeh_disable_vfs
+
+if [ "$tested" == 0 ] ; then
+	echo "No VFs with EEH aware drivers found, skipping"
+	exit $KSELFTESTS_SKIP
+fi
+
+test "$failed" != 0
+exit $?;

linux-7.2.3/tools/testing/selftests/powerpc/eeh/eeh-vf-unaware.sh

@@ -0,0 +1,35 @@
+#!/bin/sh
+# SPDX-License-Identifier: GPL-2.0-only
+
+. ./eeh-functions.sh
+
+eeh_test_prep # NB: may exit
+
+vf_list="$(eeh_enable_vfs)";
+if $? != 0 ; then
+	log "No usable VFs found. Skipping EEH unaware VF test"
+	exit $KSELFTESTS_SKIP;
+fi
+
+log "Enabled VFs: $vf_list"
+
+failed=0
+for vf in $vf_list ; do
+	log "Testing $vf"
+
+	if eeh_can_recover $vf ; then
+		log "Driver for $vf supports error recovery. Unbinding..."
+		echo "$vf" > /sys/bus/pci/devices/$vf/driver/unbind
+	fi
+
+	log "Breaking $vf..."
+	if ! eeh_one_dev $vf ; then
+		log "$vf failed to recover"
+		failed="$((failed + 1))"
+	fi
+done
+
+eeh_disable_vfs
+
+test "$failed" != 0
+exit $?;

linux-7.2.3/tools/testing/selftests/powerpc/eeh/settings

@@ -0,0 +1 @@
+timeout=300

linux-7.2.3/tools/testing/selftests/powerpc/flags.mk

@@ -0,0 +1,9 @@
+#This checks for any ENV variables and add those.
+
+ifeq ($(GIT_VERSION),)
+GIT_VERSION := $(shell git describe --always --long --dirty || echo "unknown")
+export GIT_VERSION
+endif
+
+CFLAGS := -std=gnu99 -O2 -Wall -Werror -DGIT_VERSION='"$(GIT_VERSION)"' -I$(selfdir)/powerpc/include $(USERCFLAGS)
+export CFLAGS

linux-7.2.3/tools/testing/selftests/powerpc/harness.c

@@ -0,0 +1,133 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * Copyright 2013, Michael Ellerman, IBM Corp.
+ */
+
+#include <errno.h>
+#include <signal.h>
+#include <stdbool.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <sys/types.h>
+#include <sys/wait.h>
+#include <unistd.h>
+#include <elf.h>
+#include <fcntl.h>
+#include <link.h>
+#include <sys/stat.h>
+
+#include "subunit.h"
+#include "utils.h"
+
+#define KILL_TIMEOUT	5
+
+/* Setting timeout to -1 disables the alarm */
+static uint64_t timeout = 120;
+
+int run_test(int (test_function)(void), const char *name)
+{
+	bool terminated;
+	int rc, status;
+	pid_t pid;
+
+	/* Make sure output is flushed before forking */
+	fflush(stdout);
+
+	pid = fork();
+	if (pid == 0) {
+		setpgid(0, 0);
+		exit(test_function());
+	} else if (pid == -1) {
+		perror("fork");
+		return 1;
+	}
+
+	setpgid(pid, pid);
+
+	if (timeout != -1)
+		/* Wake us up in timeout seconds */
+		alarm(timeout);
+	terminated = false;
+
+wait:
+	rc = waitpid(pid, &status, 0);
+	if (rc == -1) {
+		if (errno != EINTR) {
+			printf("unknown error from waitpid\n");
+			return 1;
+		}
+
+		if (terminated) {
+			printf("!! force killing %s\n", name);
+			kill(-pid, SIGKILL);
+			return 1;
+		} else {
+			printf("!! killing %s\n", name);
+			kill(-pid, SIGTERM);
+			terminated = true;
+			alarm(KILL_TIMEOUT);
+			goto wait;
+		}
+	}
+
+	/* Kill anything else in the process group that is still running */
+	kill(-pid, SIGTERM);
+
+	if (WIFEXITED(status))
+		status = WEXITSTATUS(status);
+	else {
+		if (WIFSIGNALED(status))
+			printf("!! child died by signal %d\n", WTERMSIG(status));
+		else
+			printf("!! child died by unknown cause\n");
+
+		status = 1; /* Signal or other */
+	}
+
+	return status;
+}
+
+static void sig_handler(int signum)
+{
+	/* Just wake us up from waitpid */
+}
+
+static struct sigaction sig_action = {
+	.sa_handler = sig_handler,
+};
+
+void test_harness_set_timeout(uint64_t time)
+{
+	timeout = time;
+}
+
+int test_harness(int (test_function)(void), const char *name)
+{
+	int rc;
+
+	test_start(name);
+	test_set_git_version(GIT_VERSION);
+
+	if (sigaction(SIGINT, &sig_action, NULL)) {
+		perror("sigaction (sigint)");
+		test_error(name);
+		return 1;
+	}
+
+	if (sigaction(SIGALRM, &sig_action, NULL)) {
+		perror("sigaction (sigalrm)");
+		test_error(name);
+		return 1;
+	}
+
+	rc = run_test(test_function, name);
+
+	if (rc == MAGIC_SKIP_RETURN_VALUE) {
+		test_skip(name);
+		/* so that skipped test is not marked as failed */
+		rc = 0;
+	} else
+		test_finish(name, rc);
+
+	return rc;
+}

linux-7.2.3/tools/testing/selftests/powerpc/include/basic_asm.h

@@ -0,0 +1,105 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+#ifndef _SELFTESTS_POWERPC_BASIC_ASM_H
+#define _SELFTESTS_POWERPC_BASIC_ASM_H
+
+#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;	\
+	rldicr	reg, reg, 32, 31;	\
+	oris	reg, reg, (expr)@high;	\
+	ori	reg, reg, (expr)@l;
+
+/*
+ * Note: These macros assume that variables being stored on the stack are
+ * 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 // ABIv1 below
+#define STACK_FRAME_MIN_SIZE 112
+#define STACK_FRAME_TOC_POS  40
+#define __STACK_FRAME_PARAM(i)  (48 + ((i)*8))
+
+/*
+ * Caveat: if a function passed more than 8 doublewords, the caller will have
+ * made more space... which would render the 112 incorrect.
+ */
+#define __STACK_FRAME_LOCAL(_num_params, _var_num)  \
+	(112 + ((_var_num)*8))
+
+
+#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)
+
+/* Local variable x saved to the stack after x parameters */
+#define STACK_FRAME_LOCAL(num_params, var)    \
+	__STACK_FRAME_LOCAL(num_params, var)
+
+/*
+ * It is very important to note here that _extra is the extra amount of
+ * stack space needed. This space can be accessed using STACK_FRAME_PARAM()
+ * or STACK_FRAME_LOCAL() macros.
+ *
+ * r1 and r2 are not defined in ppc-asm.h (instead they are defined as sp
+ * and toc). Kernel programmers tend to prefer rX even for r1 and r2, hence
+ * %1 and %r2. r0 is defined in ppc-asm.h and therefore %r0 gets
+ * preprocessed incorrectly, hence r0.
+ */
+#define PUSH_BASIC_STACK(_extra) \
+	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); \
+	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 */

linux-7.2.3/tools/testing/selftests/powerpc/include/fpu_asm.h

@@ -0,0 +1,76 @@
+/* SPDX-License-Identifier: GPL-2.0-or-later */
+/*
+ * Copyright 2016, Cyril Bur, IBM Corp.
+ */
+
+#ifndef _SELFTESTS_POWERPC_FPU_ASM_H
+#define _SELFTESTS_POWERPC_FPU_ASM_H
+#include "basic_asm.h"
+
+#define PUSH_FPU(stack_size) \
+	stfd	f31,(stack_size + STACK_FRAME_MIN_SIZE)(%r1); \
+	stfd	f30,(stack_size + STACK_FRAME_MIN_SIZE - 8)(%r1); \
+	stfd	f29,(stack_size + STACK_FRAME_MIN_SIZE - 16)(%r1); \
+	stfd	f28,(stack_size + STACK_FRAME_MIN_SIZE - 24)(%r1); \
+	stfd	f27,(stack_size + STACK_FRAME_MIN_SIZE - 32)(%r1); \
+	stfd	f26,(stack_size + STACK_FRAME_MIN_SIZE - 40)(%r1); \
+	stfd	f25,(stack_size + STACK_FRAME_MIN_SIZE - 48)(%r1); \
+	stfd	f24,(stack_size + STACK_FRAME_MIN_SIZE - 56)(%r1); \
+	stfd	f23,(stack_size + STACK_FRAME_MIN_SIZE - 64)(%r1); \
+	stfd	f22,(stack_size + STACK_FRAME_MIN_SIZE - 72)(%r1); \
+	stfd	f21,(stack_size + STACK_FRAME_MIN_SIZE - 80)(%r1); \
+	stfd	f20,(stack_size + STACK_FRAME_MIN_SIZE - 88)(%r1); \
+	stfd	f19,(stack_size + STACK_FRAME_MIN_SIZE - 96)(%r1); \
+	stfd	f18,(stack_size + STACK_FRAME_MIN_SIZE - 104)(%r1); \
+	stfd	f17,(stack_size + STACK_FRAME_MIN_SIZE - 112)(%r1); \
+	stfd	f16,(stack_size + STACK_FRAME_MIN_SIZE - 120)(%r1); \
+	stfd	f15,(stack_size + STACK_FRAME_MIN_SIZE - 128)(%r1); \
+	stfd	f14,(stack_size + STACK_FRAME_MIN_SIZE - 136)(%r1);
+
+#define POP_FPU(stack_size) \
+	lfd	f31,(stack_size + STACK_FRAME_MIN_SIZE)(%r1); \
+	lfd	f30,(stack_size + STACK_FRAME_MIN_SIZE - 8)(%r1); \
+	lfd	f29,(stack_size + STACK_FRAME_MIN_SIZE - 16)(%r1); \
+	lfd	f28,(stack_size + STACK_FRAME_MIN_SIZE - 24)(%r1); \
+	lfd	f27,(stack_size + STACK_FRAME_MIN_SIZE - 32)(%r1); \
+	lfd	f26,(stack_size + STACK_FRAME_MIN_SIZE - 40)(%r1); \
+	lfd	f25,(stack_size + STACK_FRAME_MIN_SIZE - 48)(%r1); \
+	lfd	f24,(stack_size + STACK_FRAME_MIN_SIZE - 56)(%r1); \
+	lfd	f23,(stack_size + STACK_FRAME_MIN_SIZE - 64)(%r1); \
+	lfd	f22,(stack_size + STACK_FRAME_MIN_SIZE - 72)(%r1); \
+	lfd	f21,(stack_size + STACK_FRAME_MIN_SIZE - 80)(%r1); \
+	lfd	f20,(stack_size + STACK_FRAME_MIN_SIZE - 88)(%r1); \
+	lfd	f19,(stack_size + STACK_FRAME_MIN_SIZE - 96)(%r1); \
+	lfd	f18,(stack_size + STACK_FRAME_MIN_SIZE - 104)(%r1); \
+	lfd	f17,(stack_size + STACK_FRAME_MIN_SIZE - 112)(%r1); \
+	lfd	f16,(stack_size + STACK_FRAME_MIN_SIZE - 120)(%r1); \
+	lfd	f15,(stack_size + STACK_FRAME_MIN_SIZE - 128)(%r1); \
+	lfd	f14,(stack_size + STACK_FRAME_MIN_SIZE - 136)(%r1);
+
+/*
+ * Careful calling this, it will 'clobber' fpu (by design)
+ * Don't call this from C
+ */
+FUNC_START(load_fpu)
+	lfd	f14,0(r3)
+	lfd	f15,8(r3)
+	lfd	f16,16(r3)
+	lfd	f17,24(r3)
+	lfd	f18,32(r3)
+	lfd	f19,40(r3)
+	lfd	f20,48(r3)
+	lfd	f21,56(r3)
+	lfd	f22,64(r3)
+	lfd	f23,72(r3)
+	lfd	f24,80(r3)
+	lfd	f25,88(r3)
+	lfd	f26,96(r3)
+	lfd	f27,104(r3)
+	lfd	f28,112(r3)
+	lfd	f29,120(r3)
+	lfd	f30,128(r3)
+	lfd	f31,136(r3)
+	blr
+FUNC_END(load_fpu)
+
+#endif /* _SELFTESTS_POWERPC_FPU_ASM_H */

linux-7.2.3/tools/testing/selftests/powerpc/include/gpr_asm.h

@@ -0,0 +1,92 @@
+/* SPDX-License-Identifier: GPL-2.0-or-later */
+/*
+ * Copyright 2016, Cyril Bur, IBM Corp.
+ */
+
+#ifndef _SELFTESTS_POWERPC_GPR_ASM_H
+#define _SELFTESTS_POWERPC_GPR_ASM_H
+
+#include "basic_asm.h"
+
+#define __PUSH_NVREGS(top_pos); \
+	std r31,(top_pos)(%r1); \
+	std r30,(top_pos - 8)(%r1); \
+	std r29,(top_pos - 16)(%r1); \
+	std r28,(top_pos - 24)(%r1); \
+	std r27,(top_pos - 32)(%r1); \
+	std r26,(top_pos - 40)(%r1); \
+	std r25,(top_pos - 48)(%r1); \
+	std r24,(top_pos - 56)(%r1); \
+	std r23,(top_pos - 64)(%r1); \
+	std r22,(top_pos - 72)(%r1); \
+	std r21,(top_pos - 80)(%r1); \
+	std r20,(top_pos - 88)(%r1); \
+	std r19,(top_pos - 96)(%r1); \
+	std r18,(top_pos - 104)(%r1); \
+	std r17,(top_pos - 112)(%r1); \
+	std r16,(top_pos - 120)(%r1); \
+	std r15,(top_pos - 128)(%r1); \
+	std r14,(top_pos - 136)(%r1)
+
+#define __POP_NVREGS(top_pos); \
+	ld r31,(top_pos)(%r1); \
+	ld r30,(top_pos - 8)(%r1); \
+	ld r29,(top_pos - 16)(%r1); \
+	ld r28,(top_pos - 24)(%r1); \
+	ld r27,(top_pos - 32)(%r1); \
+	ld r26,(top_pos - 40)(%r1); \
+	ld r25,(top_pos - 48)(%r1); \
+	ld r24,(top_pos - 56)(%r1); \
+	ld r23,(top_pos - 64)(%r1); \
+	ld r22,(top_pos - 72)(%r1); \
+	ld r21,(top_pos - 80)(%r1); \
+	ld r20,(top_pos - 88)(%r1); \
+	ld r19,(top_pos - 96)(%r1); \
+	ld r18,(top_pos - 104)(%r1); \
+	ld r17,(top_pos - 112)(%r1); \
+	ld r16,(top_pos - 120)(%r1); \
+	ld r15,(top_pos - 128)(%r1); \
+	ld r14,(top_pos - 136)(%r1)
+
+#define PUSH_NVREGS(stack_size) \
+	__PUSH_NVREGS(stack_size + STACK_FRAME_MIN_SIZE)
+
+/* 18 NV FPU REGS */
+#define PUSH_NVREGS_BELOW_FPU(stack_size) \
+	__PUSH_NVREGS(stack_size + STACK_FRAME_MIN_SIZE - (18 * 8))
+
+#define POP_NVREGS(stack_size) \
+	__POP_NVREGS(stack_size + STACK_FRAME_MIN_SIZE)
+
+/* 18 NV FPU REGS */
+#define POP_NVREGS_BELOW_FPU(stack_size) \
+	__POP_NVREGS(stack_size + STACK_FRAME_MIN_SIZE - (18 * 8))
+
+/*
+ * Careful calling this, it will 'clobber' NVGPRs (by design)
+ * Don't call this from C
+ */
+FUNC_START(load_gpr)
+	ld	r14,0(r3)
+	ld	r15,8(r3)
+	ld	r16,16(r3)
+	ld	r17,24(r3)
+	ld	r18,32(r3)
+	ld	r19,40(r3)
+	ld	r20,48(r3)
+	ld	r21,56(r3)
+	ld	r22,64(r3)
+	ld	r23,72(r3)
+	ld	r24,80(r3)
+	ld	r25,88(r3)
+	ld	r26,96(r3)
+	ld	r27,104(r3)
+	ld	r28,112(r3)
+	ld	r29,120(r3)
+	ld	r30,128(r3)
+	ld	r31,136(r3)
+	blr
+FUNC_END(load_gpr)
+
+
+#endif /* _SELFTESTS_POWERPC_GPR_ASM_H */

linux-7.2.3/tools/testing/selftests/powerpc/include/instructions.h

@@ -0,0 +1,146 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+#ifndef _SELFTESTS_POWERPC_INSTRUCTIONS_H
+#define _SELFTESTS_POWERPC_INSTRUCTIONS_H
+
+#include <stdio.h>
+#include <stdlib.h>
+
+/* This defines the "copy" instruction from Power ISA 3.0 Book II, section 4.4. */
+#define __COPY(RA, RB, L) \
+	(0x7c00060c | (RA) << (31-15) | (RB) << (31-20) | (L) << (31-10))
+#define COPY(RA, RB, L) \
+	.long __COPY((RA), (RB), (L))
+
+static inline void copy(void *i)
+{
+	asm volatile(str(COPY(0, %0, 0))";"
+			:
+			: "b" (i)
+			: "memory"
+		    );
+}
+
+static inline void copy_first(void *i)
+{
+	asm volatile(str(COPY(0, %0, 1))";"
+			:
+			: "b" (i)
+			: "memory"
+		    );
+}
+
+/* This defines the "paste" instruction from Power ISA 3.0 Book II, section 4.4. */
+#define __PASTE(RA, RB, L, RC) \
+	(0x7c00070c | (RA) << (31-15) | (RB) << (31-20) | (L) << (31-10) | (RC) << (31-31))
+#define PASTE(RA, RB, L, RC) \
+	.long __PASTE((RA), (RB), (L), (RC))
+
+static inline int paste(void *i)
+{
+	int cr;
+
+	asm volatile(str(PASTE(0, %1, 0, 0))";"
+			"mfcr %0;"
+			: "=r" (cr)
+			: "b" (i)
+			: "memory"
+		    );
+	return cr;
+}
+
+static inline int paste_last(void *i)
+{
+	int cr;
+
+	asm volatile(str(PASTE(0, %1, 1, 1))";"
+			"mfcr %0;"
+			: "=r" (cr)
+			: "b" (i)
+			: "memory"
+		    );
+	return cr;
+}
+
+#define PPC_INST_COPY                  __COPY(0, 0, 0)
+#define PPC_INST_COPY_FIRST            __COPY(0, 0, 1)
+#define PPC_INST_PASTE                 __PASTE(0, 0, 0, 0)
+#define PPC_INST_PASTE_LAST            __PASTE(0, 0, 1, 1)
+
+/* This defines the prefixed load/store instructions */
+#ifdef __ASSEMBLER__
+#  define stringify_in_c(...)	__VA_ARGS__
+#else
+#  define __stringify_in_c(...)	#__VA_ARGS__
+#  define stringify_in_c(...)	__stringify_in_c(__VA_ARGS__) " "
+#endif
+
+#define __PPC_RA(a)	(((a) & 0x1f) << 16)
+#define __PPC_RS(s)	(((s) & 0x1f) << 21)
+#define __PPC_RT(t)	__PPC_RS(t)
+#define __PPC_PREFIX_R(r)	(((r) & 0x1) << 20)
+
+#define PPC_PREFIX_MLS			0x06000000
+#define PPC_PREFIX_8LS			0x04000000
+
+#define PPC_INST_LBZ			0x88000000
+#define PPC_INST_LHZ			0xa0000000
+#define PPC_INST_LHA			0xa8000000
+#define PPC_INST_LWZ			0x80000000
+#define PPC_INST_STB			0x98000000
+#define PPC_INST_STH			0xb0000000
+#define PPC_INST_STW			0x90000000
+#define PPC_INST_STD			0xf8000000
+#define PPC_INST_LFS			0xc0000000
+#define PPC_INST_LFD			0xc8000000
+#define PPC_INST_STFS			0xd0000000
+#define PPC_INST_STFD			0xd8000000
+
+#define PREFIX_MLS(instr, t, a, r, d)	stringify_in_c(.balign 64, , 4;)		\
+					stringify_in_c(.long PPC_PREFIX_MLS |		\
+						       __PPC_PREFIX_R(r) |		\
+						       (((d) >> 16) & 0x3ffff);)	\
+					stringify_in_c(.long (instr)  |			\
+						       __PPC_RT(t) |			\
+						       __PPC_RA(a) |			\
+						       ((d) & 0xffff);\n)
+
+#define PREFIX_8LS(instr, t, a, r, d)	stringify_in_c(.balign 64, , 4;)		\
+					stringify_in_c(.long PPC_PREFIX_8LS |		\
+						       __PPC_PREFIX_R(r) |		\
+						       (((d) >> 16) & 0x3ffff);)	\
+					stringify_in_c(.long (instr)  |			\
+						       __PPC_RT(t) |			\
+						       __PPC_RA(a) |			\
+						       ((d) & 0xffff);\n)
+
+/* Prefixed Integer Load/Store instructions */
+#define PLBZ(t, a, r, d)		PREFIX_MLS(PPC_INST_LBZ, t, a, r, d)
+#define PLHZ(t, a, r, d)		PREFIX_MLS(PPC_INST_LHZ, t, a, r, d)
+#define PLHA(t, a, r, d)		PREFIX_MLS(PPC_INST_LHA, t, a, r, d)
+#define PLWZ(t, a, r, d)		PREFIX_MLS(PPC_INST_LWZ, t, a, r, d)
+#define PLWA(t, a, r, d)		PREFIX_8LS(0xa4000000, t, a, r, d)
+#define PLD(t, a, r, d)			PREFIX_8LS(0xe4000000, t, a, r, d)
+#define PLQ(t, a, r, d)			PREFIX_8LS(0xe0000000, t, a, r, d)
+#define PSTB(s, a, r, d)		PREFIX_MLS(PPC_INST_STB, s, a, r, d)
+#define PSTH(s, a, r, d)		PREFIX_MLS(PPC_INST_STH, s, a, r, d)
+#define PSTW(s, a, r, d)		PREFIX_MLS(PPC_INST_STW, s, a, r, d)
+#define PSTD(s, a, r, d)		PREFIX_8LS(0xf4000000, s, a, r, d)
+#define PSTQ(s, a, r, d)		PREFIX_8LS(0xf0000000, s, a, r, d)
+
+/* Prefixed Floating-Point Load/Store Instructions */
+#define PLFS(frt, a, r, d)		PREFIX_MLS(PPC_INST_LFS, frt, a, r, d)
+#define PLFD(frt, a, r, d)		PREFIX_MLS(PPC_INST_LFD, frt, a, r, d)
+#define PSTFS(frs, a, r, d)		PREFIX_MLS(PPC_INST_STFS, frs, a, r, d)
+#define PSTFD(frs, a, r, d)		PREFIX_MLS(PPC_INST_STFD, frs, a, r, d)
+
+/* Prefixed VSX Load/Store Instructions */
+#define PLXSD(vrt, a, r, d)		PREFIX_8LS(0xa8000000, vrt, a, r, d)
+#define PLXSSP(vrt, a, r, d)		PREFIX_8LS(0xac000000, vrt, a, r, d)
+#define PLXV0(s, a, r, d)		PREFIX_8LS(0xc8000000, s, a, r, d)
+#define PLXV1(s, a, r, d)		PREFIX_8LS(0xcc000000, s, a, r, d)
+#define PSTXSD(vrs, a, r, d)		PREFIX_8LS(0xb8000000, vrs, a, r, d)
+#define PSTXSSP(vrs, a, r, d)		PREFIX_8LS(0xbc000000, vrs, a, r, d)
+#define PSTXV0(s, a, r, d)		PREFIX_8LS(0xd8000000, s, a, r, d)
+#define PSTXV1(s, a, r, d)		PREFIX_8LS(0xdc000000, s, a, r, d)
+
+#endif /* _SELFTESTS_POWERPC_INSTRUCTIONS_H */

linux-7.2.3/tools/testing/selftests/powerpc/include/pkeys.h

@@ -0,0 +1,147 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+/*
+ * Copyright 2020, Sandipan Das, IBM Corp.
+ */
+
+#ifndef _SELFTESTS_POWERPC_PKEYS_H
+#define _SELFTESTS_POWERPC_PKEYS_H
+
+#include <sys/mman.h>
+
+#include "reg.h"
+#include "utils.h"
+
+/*
+ * Older versions of libc use the Intel-specific access rights.
+ * Hence, override the definitions as they might be incorrect.
+ */
+#undef PKEY_DISABLE_ACCESS
+#define PKEY_DISABLE_ACCESS	0x3
+
+#undef PKEY_DISABLE_WRITE
+#define PKEY_DISABLE_WRITE	0x2
+
+#undef PKEY_DISABLE_EXECUTE
+#define PKEY_DISABLE_EXECUTE	0x4
+
+#undef PKEY_UNRESTRICTED
+#define PKEY_UNRESTRICTED	0x0
+
+/* Older versions of libc do not define this */
+#ifndef SEGV_PKUERR
+#define SEGV_PKUERR	4
+#endif
+
+#define SI_PKEY_OFFSET	0x20
+
+#define __NR_pkey_mprotect	386
+#define __NR_pkey_alloc		384
+#define __NR_pkey_free		385
+
+#ifndef NT_PPC_PKEY
+#define NT_PPC_PKEY		0x110
+#endif
+
+#define PKEY_BITS_PER_PKEY	2
+#define NR_PKEYS		32
+#define PKEY_BITS_MASK		((1UL << PKEY_BITS_PER_PKEY) - 1)
+
+#define AMR_BITS_PER_PKEY 2
+#define PKEY_REG_BITS (sizeof(u64) * 8)
+#define pkeyshift(pkey) (PKEY_REG_BITS - ((pkey + 1) * AMR_BITS_PER_PKEY))
+
+inline unsigned long pkeyreg_get(void)
+{
+	return mfspr(SPRN_AMR);
+}
+
+inline void pkeyreg_set(unsigned long amr)
+{
+	set_amr(amr);
+}
+
+void pkey_set_rights(int pkey, unsigned long rights)
+{
+	unsigned long amr, shift;
+
+	shift = (NR_PKEYS - pkey - 1) * PKEY_BITS_PER_PKEY;
+	amr = pkeyreg_get();
+	amr &= ~(PKEY_BITS_MASK << shift);
+	amr |= (rights & PKEY_BITS_MASK) << shift;
+	pkeyreg_set(amr);
+}
+
+int sys_pkey_mprotect(void *addr, size_t len, int prot, int pkey)
+{
+	return syscall(__NR_pkey_mprotect, addr, len, prot, pkey);
+}
+
+int sys_pkey_alloc(unsigned long flags, unsigned long rights)
+{
+	return syscall(__NR_pkey_alloc, flags, rights);
+}
+
+int sys_pkey_free(int pkey)
+{
+	return syscall(__NR_pkey_free, pkey);
+}
+
+int pkeys_unsupported(void)
+{
+	bool hash_mmu = false;
+	int pkey;
+
+	/* Protection keys are currently supported on Hash MMU only */
+	FAIL_IF(using_hash_mmu(&hash_mmu));
+	SKIP_IF(!hash_mmu);
+
+	/* Check if the system call is supported */
+	pkey = sys_pkey_alloc(0, PKEY_UNRESTRICTED);
+	SKIP_IF(pkey < 0);
+	sys_pkey_free(pkey);
+
+	return 0;
+}
+
+int siginfo_pkey(siginfo_t *si)
+{
+	/*
+	 * In older versions of libc, siginfo_t does not have si_pkey as
+	 * a member.
+	 */
+#ifdef si_pkey
+	return si->si_pkey;
+#else
+	return *((int *)(((char *) si) + SI_PKEY_OFFSET));
+#endif
+}
+
+#define pkey_rights(r) ({						\
+	static char buf[4] = "rwx";					\
+	unsigned int amr_bits;						\
+	if ((r) & PKEY_DISABLE_EXECUTE)					\
+		buf[2] = '-';						\
+	amr_bits = (r) & PKEY_BITS_MASK;				\
+	if (amr_bits & PKEY_DISABLE_WRITE)				\
+		buf[1] = '-';						\
+	if (amr_bits & PKEY_DISABLE_ACCESS & ~PKEY_DISABLE_WRITE)	\
+		buf[0] = '-';						\
+	buf;								\
+})
+
+unsigned long next_pkey_rights(unsigned long rights)
+{
+	if (rights == PKEY_DISABLE_ACCESS)
+		return PKEY_DISABLE_EXECUTE;
+	else if (rights == (PKEY_DISABLE_ACCESS | PKEY_DISABLE_EXECUTE))
+		return 0;
+
+	if ((rights & PKEY_BITS_MASK) == 0)
+		rights |= PKEY_DISABLE_WRITE;
+	else if ((rights & PKEY_BITS_MASK) == PKEY_DISABLE_WRITE)
+		rights |= PKEY_DISABLE_ACCESS;
+
+	return rights;
+}
+
+#endif /* _SELFTESTS_POWERPC_PKEYS_H */

linux-7.2.3/tools/testing/selftests/powerpc/include/reg.h

@@ -0,0 +1,174 @@
+/* SPDX-License-Identifier: GPL-2.0-only */
+/*
+ * Copyright 2014, Michael Ellerman, IBM Corp.
+ */
+
+#ifndef _SELFTESTS_POWERPC_REG_H
+#define _SELFTESTS_POWERPC_REG_H
+
+#define __stringify_1(x)        #x
+#define __stringify(x)          __stringify_1(x)
+
+#define mfspr(rn)	({unsigned long rval; \
+			 asm volatile("mfspr %0," _str(rn) \
+				    : "=r" (rval)); rval; })
+#define mtspr(rn, v)	asm volatile("mtspr " _str(rn) ",%0" : \
+				    : "r" ((unsigned long)(v)) \
+				    : "memory")
+
+#define mb()		asm volatile("sync" : : : "memory");
+#define barrier()	asm volatile("" : : : "memory");
+
+#define SPRN_HDEXCR_RO 455	/* Userspace readonly view of SPRN_HDEXCR (471) */
+
+#define SPRN_MMCR2     769
+#define SPRN_MMCRA     770
+#define SPRN_MMCR0     779
+#define   MMCR0_PMAO   0x00000080
+#define   MMCR0_PMAE   0x04000000
+#define   MMCR0_FC     0x80000000
+#define SPRN_EBBHR     804
+#define SPRN_EBBRR     805
+#define SPRN_BESCR     806     /* Branch event status & control register */
+#define SPRN_BESCRS    800     /* Branch event status & control set (1 bits set to 1) */
+#define SPRN_BESCRSU   801     /* Branch event status & control set upper */
+#define SPRN_BESCRR    802     /* Branch event status & control REset (1 bits set to 0) */
+#define SPRN_BESCRRU   803     /* Branch event status & control REset upper */
+
+#define BESCR_PMEO     0x1     /* PMU Event-based exception Occurred */
+#define BESCR_PME      (0x1ul << 32) /* PMU Event-based exception Enable */
+
+#define SPRN_PMC1      771
+#define SPRN_PMC2      772
+#define SPRN_PMC3      773
+#define SPRN_PMC4      774
+#define SPRN_PMC5      775
+#define SPRN_PMC6      776
+
+#define SPRN_SIAR      780
+#define SPRN_SDAR      781
+#define SPRN_SIER      768
+
+#define SPRN_DEXCR_RO  812	/* Userspace readonly view of SPRN_DEXCR (828) */
+
+#define SPRN_TEXASR     0x82    /* Transaction Exception and Status Register */
+#define SPRN_TFIAR      0x81    /* Transaction Failure Inst Addr    */
+#define SPRN_TFHAR      0x80    /* Transaction Failure Handler Addr */
+#define SPRN_TAR        0x32f	/* Target Address Register */
+
+#define PVR_VER(pvr)	(((pvr) >>  16) & 0xFFFF)
+#define SPRN_PVR	0x11F
+
+#define PVR_CFG(pvr)    (((pvr) >>  8) & 0xF)   /* Configuration field */
+#define PVR_MAJ(pvr)    (((pvr) >>  4) & 0xF)   /* Major revision field */
+#define PVR_MIN(pvr)    (((pvr) >>  0) & 0xF)   /* Minor revision field */
+
+#define SPRN_DSCR_PRIV 0x11	/* Privilege State DSCR */
+#define SPRN_DSCR      0x03	/* Data Stream Control Register */
+#define SPRN_PPR       896	/* Program Priority Register */
+#define SPRN_AMR       13	/* Authority Mask Register - problem state */
+
+#define set_amr(v)	asm volatile("isync;" \
+				     "mtspr " __stringify(SPRN_AMR) ",%0;" \
+				     "isync" : \
+				    : "r" ((unsigned long)(v)) \
+				    : "memory")
+
+/* TEXASR register bits */
+#define TEXASR_FC	0xFE00000000000000
+#define TEXASR_FP	0x0100000000000000
+#define TEXASR_DA	0x0080000000000000
+#define TEXASR_NO	0x0040000000000000
+#define TEXASR_FO	0x0020000000000000
+#define TEXASR_SIC	0x0010000000000000
+#define TEXASR_NTC	0x0008000000000000
+#define TEXASR_TC	0x0004000000000000
+#define TEXASR_TIC	0x0002000000000000
+#define TEXASR_IC	0x0001000000000000
+#define TEXASR_IFC	0x0000800000000000
+#define TEXASR_ABT	0x0000000100000000
+#define TEXASR_SPD	0x0000000080000000
+#define TEXASR_HV	0x0000000020000000
+#define TEXASR_PR	0x0000000010000000
+#define TEXASR_FS	0x0000000008000000
+#define TEXASR_TE	0x0000000004000000
+#define TEXASR_ROT	0x0000000002000000
+
+/* MSR register bits */
+#define MSR_HV 		(1ul << 60)	/* Hypervisor state */
+#define MSR_TS_S_LG     33              /* Trans Mem state: Suspended */
+#define MSR_TS_T_LG	34              /* Trans Mem state: Active */
+
+#define __MASK(X)       (1UL<<(X))
+
+/* macro to check TM MSR bits */
+#define MSR_TS_S        __MASK(MSR_TS_S_LG)   /* Transaction Suspended */
+#define MSR_TS_T	__MASK(MSR_TS_T_LG)   /* Transaction Transactional */
+
+/* Vector Instructions */
+#define VSX_XX1(xs, ra, rb)	(((xs) & 0x1f) << 21 | ((ra) << 16) |  \
+				 ((rb) << 11) | (((xs) >> 5)))
+#define STXVD2X(xs, ra, rb)	.long (0x7c000798 | VSX_XX1((xs), (ra), (rb)))
+#define LXVD2X(xs, ra, rb)	.long (0x7c000698 | VSX_XX1((xs), (ra), (rb)))
+
+#define ASM_LOAD_GPR_IMMED(_asm_symbol_name_immed) \
+		"li 14, %[" #_asm_symbol_name_immed "];" \
+		"li 15, %[" #_asm_symbol_name_immed "];" \
+		"li 16, %[" #_asm_symbol_name_immed "];" \
+		"li 17, %[" #_asm_symbol_name_immed "];" \
+		"li 18, %[" #_asm_symbol_name_immed "];" \
+		"li 19, %[" #_asm_symbol_name_immed "];" \
+		"li 20, %[" #_asm_symbol_name_immed "];" \
+		"li 21, %[" #_asm_symbol_name_immed "];" \
+		"li 22, %[" #_asm_symbol_name_immed "];" \
+		"li 23, %[" #_asm_symbol_name_immed "];" \
+		"li 24, %[" #_asm_symbol_name_immed "];" \
+		"li 25, %[" #_asm_symbol_name_immed "];" \
+		"li 26, %[" #_asm_symbol_name_immed "];" \
+		"li 27, %[" #_asm_symbol_name_immed "];" \
+		"li 28, %[" #_asm_symbol_name_immed "];" \
+		"li 29, %[" #_asm_symbol_name_immed "];" \
+		"li 30, %[" #_asm_symbol_name_immed "];" \
+		"li 31, %[" #_asm_symbol_name_immed "];"
+
+#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(double *addr);
+#endif /* end of __ASSEMBLER__ */
+
+#endif /* _SELFTESTS_POWERPC_REG_H */

linux-7.2.3/tools/testing/selftests/powerpc/include/subunit.h

@@ -0,0 +1,52 @@
+/* SPDX-License-Identifier: GPL-2.0-only */
+/*
+ * Copyright 2013, Michael Ellerman, IBM Corp.
+ */
+
+#ifndef _SELFTESTS_POWERPC_SUBUNIT_H
+#define _SELFTESTS_POWERPC_SUBUNIT_H
+
+static inline void test_start(const char *name)
+{
+	printf("test: %s\n", name);
+}
+
+static inline void test_failure_detail(const char *name, const char *detail)
+{
+	printf("failure: %s [%s]\n", name, detail);
+}
+
+static inline void test_failure(const char *name)
+{
+	printf("failure: %s\n", name);
+}
+
+static inline void test_error(const char *name)
+{
+	printf("error: %s\n", name);
+}
+
+static inline void test_skip(const char *name)
+{
+	printf("skip: %s\n", name);
+}
+
+static inline void test_success(const char *name)
+{
+	printf("success: %s\n", name);
+}
+
+static inline void test_finish(const char *name, int status)
+{
+	if (status)
+		test_failure(name);
+	else
+		test_success(name);
+}
+
+static inline void test_set_git_version(const char *value)
+{
+	printf("tags: git_version:%s\n", value);
+}
+
+#endif /* _SELFTESTS_POWERPC_SUBUNIT_H */

linux-7.2.3/tools/testing/selftests/powerpc/include/utils.h

@@ -0,0 +1,211 @@
+/* SPDX-License-Identifier: GPL-2.0-only */
+/*
+ * Copyright 2013, Michael Ellerman, IBM Corp.
+ */
+
+#ifndef _SELFTESTS_POWERPC_UTILS_H
+#define _SELFTESTS_POWERPC_UTILS_H
+
+#define __cacheline_aligned __attribute__((aligned(128)))
+
+#include <stdint.h>
+#include <stdio.h>
+#include <stdbool.h>
+#include <sys/signal.h>
+#include <linux/auxvec.h>
+#include <linux/perf_event.h>
+#include <asm/cputable.h>
+#include "reg.h"
+#include <unistd.h>
+
+#ifndef ARRAY_SIZE
+# define ARRAY_SIZE(x) (sizeof(x) / sizeof((x)[0]))
+#endif
+
+/* Avoid headaches with PRI?64 - just use %ll? always */
+typedef unsigned long long u64;
+typedef   signed long long s64;
+
+/* Just for familiarity */
+typedef uint32_t u32;
+typedef uint16_t u16;
+typedef uint8_t u8;
+
+void test_harness_set_timeout(uint64_t time);
+int test_harness(int (test_function)(void), const char *name);
+
+int read_auxv(char *buf, ssize_t buf_size);
+void *find_auxv_entry(int type, char *auxv);
+void *get_auxv_entry(int type);
+
+#define BIND_CPU_ANY	(-1)
+
+int pick_online_cpu(void);
+int bind_to_cpu(int cpu);
+
+int parse_intmax(const char *buffer, size_t count, intmax_t *result, int base);
+int parse_uintmax(const char *buffer, size_t count, uintmax_t *result, int base);
+int parse_int(const char *buffer, size_t count, int *result, int base);
+int parse_uint(const char *buffer, size_t count, unsigned int *result, int base);
+int parse_long(const char *buffer, size_t count, long *result, int base);
+int parse_ulong(const char *buffer, size_t count, unsigned long *result, int base);
+
+int read_file(const char *path, char *buf, size_t count, size_t *len);
+int write_file(const char *path, const char *buf, size_t count);
+int read_file_alloc(const char *path, char **buf, size_t *len);
+int read_long(const char *path, long *result, int base);
+int write_long(const char *path, long result, int base);
+int read_ulong(const char *path, unsigned long *result, int base);
+int write_ulong(const char *path, unsigned long result, int base);
+int read_debugfs_file(const char *debugfs_file, char *buf, size_t count);
+int write_debugfs_file(const char *debugfs_file, const char *buf, size_t count);
+int read_debugfs_int(const char *debugfs_file, int *result);
+int write_debugfs_int(const char *debugfs_file, int result);
+int read_sysfs_file(char *debugfs_file, char *result, size_t result_size);
+int perf_event_open_counter(unsigned int type,
+			    unsigned long config, int group_fd);
+int perf_event_enable(int fd);
+int perf_event_disable(int fd);
+int perf_event_reset(int fd);
+
+struct perf_event_read {
+	__u64 nr;
+	__u64 l1d_misses;
+};
+
+#if !defined(__GLIBC_PREREQ) || !__GLIBC_PREREQ(2, 30)
+#include <sys/syscall.h>
+
+static inline pid_t gettid(void)
+{
+	return syscall(SYS_gettid);
+}
+#endif
+
+static inline bool have_hwcap(unsigned long ftr)
+{
+	return ((unsigned long)get_auxv_entry(AT_HWCAP) & ftr) == ftr;
+}
+
+#ifdef AT_HWCAP2
+static inline bool have_hwcap2(unsigned long ftr2)
+{
+	return ((unsigned long)get_auxv_entry(AT_HWCAP2) & ftr2) == ftr2;
+}
+#else
+static inline bool have_hwcap2(unsigned long ftr2)
+{
+	return false;
+}
+#endif
+
+static inline char *auxv_base_platform(void)
+{
+	return ((char *)get_auxv_entry(AT_BASE_PLATFORM));
+}
+
+static inline char *auxv_platform(void)
+{
+	return ((char *)get_auxv_entry(AT_PLATFORM));
+}
+
+bool is_ppc64le(void);
+int using_hash_mmu(bool *using_hash);
+
+struct sigaction push_signal_handler(int sig, void (*fn)(int, siginfo_t *, void *));
+struct sigaction pop_signal_handler(int sig, struct sigaction old_handler);
+
+/* Yes, this is evil */
+#define FAIL_IF(x)						\
+do {								\
+	if ((x)) {						\
+		fprintf(stderr,					\
+		"[FAIL] Test FAILED on line %d\n", __LINE__);	\
+		return 1;					\
+	}							\
+} while (0)
+
+#define FAIL_IF_MSG(x, msg)					\
+do {								\
+	if ((x)) {						\
+		fprintf(stderr,					\
+		"[FAIL] Test FAILED on line %d: %s\n", 		\
+		__LINE__, msg);					\
+		return 1;					\
+	}							\
+} while (0)
+
+#define FAIL_IF_EXIT(x)						\
+do {								\
+	if ((x)) {						\
+		fprintf(stderr,					\
+		"[FAIL] Test FAILED on line %d\n", __LINE__);	\
+		_exit(1);					\
+	}							\
+} while (0)
+
+#define FAIL_IF_EXIT_MSG(x, msg)				\
+do {								\
+	if ((x)) {						\
+		fprintf(stderr,					\
+		"[FAIL] Test FAILED on line %d: %s\n", 		\
+		__LINE__, msg);					\
+		_exit(1);					\
+	}							\
+} while (0)
+
+/* The test harness uses this, yes it's gross */
+#define MAGIC_SKIP_RETURN_VALUE	99
+
+#define SKIP_IF(x)						\
+do {								\
+	if ((x)) {						\
+		fprintf(stderr,					\
+		"[SKIP] Test skipped on line %d\n", __LINE__);	\
+		return MAGIC_SKIP_RETURN_VALUE;			\
+	}							\
+} while (0)
+
+#define SKIP_IF_MSG(x, msg)					\
+do {								\
+	if ((x)) {						\
+		fprintf(stderr,					\
+		"[SKIP] Test skipped on line %d: %s\n",		\
+		 __LINE__, msg);				\
+		return MAGIC_SKIP_RETURN_VALUE;			\
+	}							\
+} while (0)
+
+#define _str(s) #s
+#define str(s) _str(s)
+
+#define sigsafe_err(msg)	({ \
+		ssize_t nbytes __attribute__((unused)); \
+		nbytes = write(STDERR_FILENO, msg, strlen(msg)); })
+
+/* POWER9 feature */
+#ifndef PPC_FEATURE2_ARCH_3_00
+#define PPC_FEATURE2_ARCH_3_00 0x00800000
+#endif
+
+/* POWER10 feature */
+#ifndef PPC_FEATURE2_ARCH_3_1
+#define PPC_FEATURE2_ARCH_3_1 0x00040000
+#endif
+
+/* POWER10 features */
+#ifndef PPC_FEATURE2_MMA
+#define PPC_FEATURE2_MMA 0x00020000
+#endif
+
+#if defined(__powerpc64__)
+#define UCONTEXT_NIA(UC)	(UC)->uc_mcontext.gp_regs[PT_NIP]
+#define UCONTEXT_MSR(UC)	(UC)->uc_mcontext.gp_regs[PT_MSR]
+#elif defined(__powerpc__)
+#define UCONTEXT_NIA(UC)	(UC)->uc_mcontext.uc_regs->gregs[PT_NIP]
+#define UCONTEXT_MSR(UC)	(UC)->uc_mcontext.uc_regs->gregs[PT_MSR]
+#else
+#error implement UCONTEXT_NIA
+#endif
+
+#endif /* _SELFTESTS_POWERPC_UTILS_H */

linux-7.2.3/tools/testing/selftests/powerpc/include/vmx_asm.h

@@ -0,0 +1,92 @@
+/* SPDX-License-Identifier: GPL-2.0-or-later */
+/*
+ * Copyright 2015, Cyril Bur, IBM Corp.
+ */
+
+#include "basic_asm.h"
+
+/* POS MUST BE 16 ALIGNED! */
+#define PUSH_VMX(pos,reg) \
+	li	reg,pos; \
+	stvx	v20,reg,%r1; \
+	addi	reg,reg,16; \
+	stvx	v21,reg,%r1; \
+	addi	reg,reg,16; \
+	stvx	v22,reg,%r1; \
+	addi	reg,reg,16; \
+	stvx	v23,reg,%r1; \
+	addi	reg,reg,16; \
+	stvx	v24,reg,%r1; \
+	addi	reg,reg,16; \
+	stvx	v25,reg,%r1; \
+	addi	reg,reg,16; \
+	stvx	v26,reg,%r1; \
+	addi	reg,reg,16; \
+	stvx	v27,reg,%r1; \
+	addi	reg,reg,16; \
+	stvx	v28,reg,%r1; \
+	addi	reg,reg,16; \
+	stvx	v29,reg,%r1; \
+	addi	reg,reg,16; \
+	stvx	v30,reg,%r1; \
+	addi	reg,reg,16; \
+	stvx	v31,reg,%r1;
+
+/* POS MUST BE 16 ALIGNED! */
+#define POP_VMX(pos,reg) \
+	li	reg,pos; \
+	lvx	v20,reg,%r1; \
+	addi	reg,reg,16; \
+	lvx	v21,reg,%r1; \
+	addi	reg,reg,16; \
+	lvx	v22,reg,%r1; \
+	addi	reg,reg,16; \
+	lvx	v23,reg,%r1; \
+	addi	reg,reg,16; \
+	lvx	v24,reg,%r1; \
+	addi	reg,reg,16; \
+	lvx	v25,reg,%r1; \
+	addi	reg,reg,16; \
+	lvx	v26,reg,%r1; \
+	addi	reg,reg,16; \
+	lvx	v27,reg,%r1; \
+	addi	reg,reg,16; \
+	lvx	v28,reg,%r1; \
+	addi	reg,reg,16; \
+	lvx	v29,reg,%r1; \
+	addi	reg,reg,16; \
+	lvx	v30,reg,%r1; \
+	addi	reg,reg,16; \
+	lvx	v31,reg,%r1;
+
+/*
+ * Careful this will 'clobber' vmx (by design)
+ * Don't call this from C
+ */
+FUNC_START(load_vmx)
+	li	r5,0
+	lvx	v20,r5,r3
+	addi	r5,r5,16
+	lvx	v21,r5,r3
+	addi	r5,r5,16
+	lvx	v22,r5,r3
+	addi	r5,r5,16
+	lvx	v23,r5,r3
+	addi	r5,r5,16
+	lvx	v24,r5,r3
+	addi	r5,r5,16
+	lvx	v25,r5,r3
+	addi	r5,r5,16
+	lvx	v26,r5,r3
+	addi	r5,r5,16
+	lvx	v27,r5,r3
+	addi	r5,r5,16
+	lvx	v28,r5,r3
+	addi	r5,r5,16
+	lvx	v29,r5,r3
+	addi	r5,r5,16
+	lvx	v30,r5,r3
+	addi	r5,r5,16
+	lvx	v31,r5,r3
+	blr
+FUNC_END(load_vmx)

linux-7.2.3/tools/testing/selftests/powerpc/include/vsx_asm.h

@@ -0,0 +1,67 @@
+/* SPDX-License-Identifier: GPL-2.0-or-later */
+/*
+ * Copyright 2015, Cyril Bur, IBM Corp.
+ */
+
+#include "basic_asm.h"
+
+/*
+ * Careful this will 'clobber' vsx (by design), VSX are always
+ * volatile though so unlike vmx this isn't so much of an issue
+ * Still should avoid calling from C
+ */
+FUNC_START(load_vsx)
+	li	r5,0
+	lxvd2x	vs20,r5,r3
+	addi	r5,r5,16
+	lxvd2x	vs21,r5,r3
+	addi	r5,r5,16
+	lxvd2x	vs22,r5,r3
+	addi	r5,r5,16
+	lxvd2x	vs23,r5,r3
+	addi	r5,r5,16
+	lxvd2x	vs24,r5,r3
+	addi	r5,r5,16
+	lxvd2x	vs25,r5,r3
+	addi	r5,r5,16
+	lxvd2x	vs26,r5,r3
+	addi	r5,r5,16
+	lxvd2x	vs27,r5,r3
+	addi	r5,r5,16
+	lxvd2x	vs28,r5,r3
+	addi	r5,r5,16
+	lxvd2x	vs29,r5,r3
+	addi	r5,r5,16
+	lxvd2x	vs30,r5,r3
+	addi	r5,r5,16
+	lxvd2x	vs31,r5,r3
+	blr
+FUNC_END(load_vsx)
+
+FUNC_START(store_vsx)
+	li	r5,0
+	stxvd2x	vs20,r5,r3
+	addi	r5,r5,16
+	stxvd2x	vs21,r5,r3
+	addi	r5,r5,16
+	stxvd2x	vs22,r5,r3
+	addi	r5,r5,16
+	stxvd2x	vs23,r5,r3
+	addi	r5,r5,16
+	stxvd2x	vs24,r5,r3
+	addi	r5,r5,16
+	stxvd2x	vs25,r5,r3
+	addi	r5,r5,16
+	stxvd2x	vs26,r5,r3
+	addi	r5,r5,16
+	stxvd2x	vs27,r5,r3
+	addi	r5,r5,16
+	stxvd2x	vs28,r5,r3
+	addi	r5,r5,16
+	stxvd2x	vs29,r5,r3
+	addi	r5,r5,16
+	stxvd2x	vs30,r5,r3
+	addi	r5,r5,16
+	stxvd2x	vs31,r5,r3
+	blr
+FUNC_END(store_vsx)

linux-7.2.3/tools/testing/selftests/powerpc/lib/reg.S

@@ -0,0 +1,356 @@
+/* SPDX-License-Identifier: GPL-2.0-or-later */
+/*
+ * test helper assembly functions
+ *
+ * Copyright (C) 2016 Simon Guo, IBM Corporation.
+ */
+#include <ppc-asm.h>
+#include "reg.h"
+
+
+/* Non volatile GPR - unsigned long buf[18] */
+FUNC_START(load_gpr)
+	ld	14, 0*8(3)
+	ld	15, 1*8(3)
+	ld	16, 2*8(3)
+	ld	17, 3*8(3)
+	ld	18, 4*8(3)
+	ld	19, 5*8(3)
+	ld	20, 6*8(3)
+	ld	21, 7*8(3)
+	ld	22, 8*8(3)
+	ld	23, 9*8(3)
+	ld	24, 10*8(3)
+	ld	25, 11*8(3)
+	ld	26, 12*8(3)
+	ld	27, 13*8(3)
+	ld	28, 14*8(3)
+	ld	29, 15*8(3)
+	ld	30, 16*8(3)
+	ld	31, 17*8(3)
+	blr
+FUNC_END(load_gpr)
+
+FUNC_START(store_gpr)
+	std	14, 0*8(3)
+	std	15, 1*8(3)
+	std	16, 2*8(3)
+	std	17, 3*8(3)
+	std	18, 4*8(3)
+	std	19, 5*8(3)
+	std	20, 6*8(3)
+	std	21, 7*8(3)
+	std	22, 8*8(3)
+	std	23, 9*8(3)
+	std	24, 10*8(3)
+	std	25, 11*8(3)
+	std	26, 12*8(3)
+	std	27, 13*8(3)
+	std	28, 14*8(3)
+	std	29, 15*8(3)
+	std	30, 16*8(3)
+	std	31, 17*8(3)
+	blr
+FUNC_END(store_gpr)
+
+/* 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)
+
+/* VMX/VSX registers - unsigned long buf[128] */
+FUNC_START(loadvsx)
+	lis	4, 0
+	LXVD2X	(0,(4),(3))
+	addi	4, 4, 16
+	LXVD2X	(1,(4),(3))
+	addi	4, 4, 16
+	LXVD2X	(2,(4),(3))
+	addi	4, 4, 16
+	LXVD2X	(3,(4),(3))
+	addi	4, 4, 16
+	LXVD2X	(4,(4),(3))
+	addi	4, 4, 16
+	LXVD2X	(5,(4),(3))
+	addi	4, 4, 16
+	LXVD2X	(6,(4),(3))
+	addi	4, 4, 16
+	LXVD2X	(7,(4),(3))
+	addi	4, 4, 16
+	LXVD2X	(8,(4),(3))
+	addi	4, 4, 16
+	LXVD2X	(9,(4),(3))
+	addi	4, 4, 16
+	LXVD2X	(10,(4),(3))
+	addi	4, 4, 16
+	LXVD2X	(11,(4),(3))
+	addi	4, 4, 16
+	LXVD2X	(12,(4),(3))
+	addi	4, 4, 16
+	LXVD2X	(13,(4),(3))
+	addi	4, 4, 16
+	LXVD2X	(14,(4),(3))
+	addi	4, 4, 16
+	LXVD2X	(15,(4),(3))
+	addi	4, 4, 16
+	LXVD2X	(16,(4),(3))
+	addi	4, 4, 16
+	LXVD2X	(17,(4),(3))
+	addi	4, 4, 16
+	LXVD2X	(18,(4),(3))
+	addi	4, 4, 16
+	LXVD2X	(19,(4),(3))
+	addi	4, 4, 16
+	LXVD2X	(20,(4),(3))
+	addi	4, 4, 16
+	LXVD2X	(21,(4),(3))
+	addi	4, 4, 16
+	LXVD2X	(22,(4),(3))
+	addi	4, 4, 16
+	LXVD2X	(23,(4),(3))
+	addi	4, 4, 16
+	LXVD2X	(24,(4),(3))
+	addi	4, 4, 16
+	LXVD2X	(25,(4),(3))
+	addi	4, 4, 16
+	LXVD2X	(26,(4),(3))
+	addi	4, 4, 16
+	LXVD2X	(27,(4),(3))
+	addi	4, 4, 16
+	LXVD2X	(28,(4),(3))
+	addi	4, 4, 16
+	LXVD2X	(29,(4),(3))
+	addi	4, 4, 16
+	LXVD2X	(30,(4),(3))
+	addi	4, 4, 16
+	LXVD2X	(31,(4),(3))
+	addi	4, 4, 16
+	LXVD2X	(32,(4),(3))
+	addi	4, 4, 16
+	LXVD2X	(33,(4),(3))
+	addi	4, 4, 16
+	LXVD2X	(34,(4),(3))
+	addi	4, 4, 16
+	LXVD2X	(35,(4),(3))
+	addi	4, 4, 16
+	LXVD2X	(36,(4),(3))
+	addi	4, 4, 16
+	LXVD2X	(37,(4),(3))
+	addi	4, 4, 16
+	LXVD2X	(38,(4),(3))
+	addi	4, 4, 16
+	LXVD2X	(39,(4),(3))
+	addi	4, 4, 16
+	LXVD2X	(40,(4),(3))
+	addi	4, 4, 16
+	LXVD2X	(41,(4),(3))
+	addi	4, 4, 16
+	LXVD2X	(42,(4),(3))
+	addi	4, 4, 16
+	LXVD2X	(43,(4),(3))
+	addi	4, 4, 16
+	LXVD2X	(44,(4),(3))
+	addi	4, 4, 16
+	LXVD2X	(45,(4),(3))
+	addi	4, 4, 16
+	LXVD2X	(46,(4),(3))
+	addi	4, 4, 16
+	LXVD2X	(47,(4),(3))
+	addi	4, 4, 16
+	LXVD2X	(48,(4),(3))
+	addi	4, 4, 16
+	LXVD2X	(49,(4),(3))
+	addi	4, 4, 16
+	LXVD2X	(50,(4),(3))
+	addi	4, 4, 16
+	LXVD2X	(51,(4),(3))
+	addi	4, 4, 16
+	LXVD2X	(52,(4),(3))
+	addi	4, 4, 16
+	LXVD2X	(53,(4),(3))
+	addi	4, 4, 16
+	LXVD2X	(54,(4),(3))
+	addi	4, 4, 16
+	LXVD2X	(55,(4),(3))
+	addi	4, 4, 16
+	LXVD2X	(56,(4),(3))
+	addi	4, 4, 16
+	LXVD2X	(57,(4),(3))
+	addi	4, 4, 16
+	LXVD2X	(58,(4),(3))
+	addi	4, 4, 16
+	LXVD2X	(59,(4),(3))
+	addi	4, 4, 16
+	LXVD2X	(60,(4),(3))
+	addi	4, 4, 16
+	LXVD2X	(61,(4),(3))
+	addi	4, 4, 16
+	LXVD2X	(62,(4),(3))
+	addi	4, 4, 16
+	LXVD2X	(63,(4),(3))
+	blr
+FUNC_END(loadvsx)
+
+FUNC_START(storevsx)
+	lis	4, 0
+	STXVD2X	(0,(4),(3))
+	addi	4, 4, 16
+	STXVD2X	(1,(4),(3))
+	addi	4, 4, 16
+	STXVD2X	(2,(4),(3))
+	addi	4, 4, 16
+	STXVD2X	(3,(4),(3))
+	addi	4, 4, 16
+	STXVD2X	(4,(4),(3))
+	addi	4, 4, 16
+	STXVD2X	(5,(4),(3))
+	addi	4, 4, 16
+	STXVD2X	(6,(4),(3))
+	addi	4, 4, 16
+	STXVD2X	(7,(4),(3))
+	addi	4, 4, 16
+	STXVD2X	(8,(4),(3))
+	addi	4, 4, 16
+	STXVD2X	(9,(4),(3))
+	addi	4, 4, 16
+	STXVD2X	(10,(4),(3))
+	addi	4, 4, 16
+	STXVD2X	(11,(4),(3))
+	addi	4, 4, 16
+	STXVD2X	(12,(4),(3))
+	addi	4, 4, 16
+	STXVD2X	(13,(4),(3))
+	addi	4, 4, 16
+	STXVD2X	(14,(4),(3))
+	addi	4, 4, 16
+	STXVD2X	(15,(4),(3))
+	addi	4, 4, 16
+	STXVD2X	(16,(4),(3))
+	addi	4, 4, 16
+	STXVD2X	(17,(4),(3))
+	addi	4, 4, 16
+	STXVD2X	(18,(4),(3))
+	addi	4, 4, 16
+	STXVD2X	(19,(4),(3))
+	addi	4, 4, 16
+	STXVD2X	(20,(4),(3))
+	addi	4, 4, 16
+	STXVD2X	(21,(4),(3))
+	addi	4, 4, 16
+	STXVD2X	(22,(4),(3))
+	addi	4, 4, 16
+	STXVD2X	(23,(4),(3))
+	addi	4, 4, 16
+	STXVD2X	(24,(4),(3))
+	addi	4, 4, 16
+	STXVD2X	(25,(4),(3))
+	addi	4, 4, 16
+	STXVD2X	(26,(4),(3))
+	addi	4, 4, 16
+	STXVD2X	(27,(4),(3))
+	addi	4, 4, 16
+	STXVD2X	(28,(4),(3))
+	addi	4, 4, 16
+	STXVD2X	(29,(4),(3))
+	addi	4, 4, 16
+	STXVD2X	(30,(4),(3))
+	addi	4, 4, 16
+	STXVD2X	(31,(4),(3))
+	addi	4, 4, 16
+	STXVD2X	(32,(4),(3))
+	addi	4, 4, 16
+	STXVD2X	(33,(4),(3))
+	addi	4, 4, 16
+	STXVD2X	(34,(4),(3))
+	addi	4, 4, 16
+	STXVD2X	(35,(4),(3))
+	addi	4, 4, 16
+	STXVD2X	(36,(4),(3))
+	addi	4, 4, 16
+	STXVD2X	(37,(4),(3))
+	addi	4, 4, 16
+	STXVD2X	(38,(4),(3))
+	addi	4, 4, 16
+	STXVD2X	(39,(4),(3))
+	addi	4, 4, 16
+	STXVD2X	(40,(4),(3))
+	addi	4, 4, 16
+	STXVD2X	(41,(4),(3))
+	addi	4, 4, 16
+	STXVD2X	(42,(4),(3))
+	addi	4, 4, 16
+	STXVD2X	(43,(4),(3))
+	addi	4, 4, 16
+	STXVD2X	(44,(4),(3))
+	addi	4, 4, 16
+	STXVD2X	(45,(4),(3))
+	addi	4, 4, 16
+	STXVD2X	(46,(4),(3))
+	addi	4, 4, 16
+	STXVD2X	(47,(4),(3))
+	addi	4, 4, 16
+	STXVD2X	(48,(4),(3))
+	addi	4, 4, 16
+	STXVD2X	(49,(4),(3))
+	addi	4, 4, 16
+	STXVD2X	(50,(4),(3))
+	addi	4, 4, 16
+	STXVD2X	(51,(4),(3))
+	addi	4, 4, 16
+	STXVD2X	(52,(4),(3))
+	addi	4, 4, 16
+	STXVD2X	(53,(4),(3))
+	addi	4, 4, 16
+	STXVD2X	(54,(4),(3))
+	addi	4, 4, 16
+	STXVD2X	(55,(4),(3))
+	addi	4, 4, 16
+	STXVD2X	(56,(4),(3))
+	addi	4, 4, 16
+	STXVD2X	(57,(4),(3))
+	addi	4, 4, 16
+	STXVD2X	(58,(4),(3))
+	addi	4, 4, 16
+	STXVD2X	(59,(4),(3))
+	addi	4, 4, 16
+	STXVD2X	(60,(4),(3))
+	addi	4, 4, 16
+	STXVD2X	(61,(4),(3))
+	addi	4, 4, 16
+	STXVD2X	(62,(4),(3))
+	addi	4, 4, 16
+	STXVD2X	(63,(4),(3))
+	blr
+FUNC_END(storevsx)

linux-7.2.3/tools/testing/selftests/powerpc/lib/settings

@@ -0,0 +1 @@
+timeout=130

linux-7.2.3/tools/testing/selftests/powerpc/math/.gitignore

@@ -0,0 +1,10 @@
+# SPDX-License-Identifier: GPL-2.0-only
+fpu_syscall
+vmx_syscall
+fpu_preempt
+vmx_preempt
+fpu_signal
+vmx_signal
+vsx_preempt
+fpu_denormal
+mma

linux-7.2.3/tools/testing/selftests/powerpc/math/Makefile

@@ -0,0 +1,22 @@
+# SPDX-License-Identifier: GPL-2.0
+TEST_GEN_PROGS := fpu_syscall fpu_preempt fpu_signal fpu_denormal vmx_syscall vmx_preempt vmx_signal vsx_preempt mma
+
+top_srcdir = ../../../../..
+include ../../lib.mk
+include ../flags.mk
+
+$(TEST_GEN_PROGS): ../harness.c
+$(TEST_GEN_PROGS): CFLAGS += -O2 -g -pthread -m64 -maltivec
+
+$(OUTPUT)/fpu_syscall: fpu_asm.S
+$(OUTPUT)/fpu_preempt: fpu_asm.S
+$(OUTPUT)/fpu_signal:  fpu_asm.S
+
+$(OUTPUT)/vmx_syscall: vmx_asm.S ../utils.c
+$(OUTPUT)/vmx_preempt: vmx_asm.S ../utils.c
+$(OUTPUT)/vmx_signal: vmx_asm.S ../utils.c
+
+$(OUTPUT)/vsx_preempt: CFLAGS += -mvsx
+$(OUTPUT)/vsx_preempt: vsx_asm.S ../utils.c
+
+$(OUTPUT)/mma: mma.c mma.S ../utils.c

linux-7.2.3/tools/testing/selftests/powerpc/math/fpu.h

@@ -0,0 +1,25 @@
+/* SPDX-License-Identifier: GPL-2.0-or-later */
+/*
+ * Copyright 2023, Michael Ellerman, IBM Corporation.
+ */
+
+#ifndef _SELFTESTS_POWERPC_FPU_H
+#define _SELFTESTS_POWERPC_FPU_H
+
+static inline void randomise_darray(double *darray, int num)
+{
+	long val;
+
+	for (int i = 0; i < num; i++) {
+		val = random();
+		if (val & 1)
+			val *= -1;
+
+		if (val & 2)
+			darray[i] = 1.0 / val;
+		else
+			darray[i] = val * val;
+	}
+}
+
+#endif /* _SELFTESTS_POWERPC_FPU_H */

linux-7.2.3/tools/testing/selftests/powerpc/math/fpu_asm.S

@@ -0,0 +1,165 @@
+/* SPDX-License-Identifier: GPL-2.0-or-later */
+/*
+ * Copyright 2015, Cyril Bur, IBM Corp.
+ */
+
+#include "basic_asm.h"
+#include "fpu_asm.h"
+
+FUNC_START(check_fpu)
+	mr r4,r3
+	li	r3,1 # assume a bad result
+	lfd	f0,0(r4)
+	fcmpu	cr1,f0,f14
+	bne	cr1,1f
+	lfd	f0,8(r4)
+	fcmpu	cr1,f0,f15
+	bne	cr1,1f
+	lfd	f0,16(r4)
+	fcmpu	cr1,f0,f16
+	bne	cr1,1f
+	lfd	f0,24(r4)
+	fcmpu	cr1,f0,f17
+	bne	cr1,1f
+	lfd	f0,32(r4)
+	fcmpu	cr1,f0,f18
+	bne	cr1,1f
+	lfd	f0,40(r4)
+	fcmpu	cr1,f0,f19
+	bne	cr1,1f
+	lfd	f0,48(r4)
+	fcmpu	cr1,f0,f20
+	bne	cr1,1f
+	lfd	f0,56(r4)
+	fcmpu	cr1,f0,f21
+	bne	cr1,1f
+	lfd	f0,64(r4)
+	fcmpu	cr1,f0,f22
+	bne	cr1,1f
+	lfd	f0,72(r4)
+	fcmpu	cr1,f0,f23
+	bne	cr1,1f
+	lfd	f0,80(r4)
+	fcmpu	cr1,f0,f24
+	bne	cr1,1f
+	lfd	f0,88(r4)
+	fcmpu	cr1,f0,f25
+	bne	cr1,1f
+	lfd	f0,96(r4)
+	fcmpu	cr1,f0,f26
+	bne	cr1,1f
+	lfd	f0,104(r4)
+	fcmpu	cr1,f0,f27
+	bne	cr1,1f
+	lfd	f0,112(r4)
+	fcmpu	cr1,f0,f28
+	bne	cr1,1f
+	lfd	f0,120(r4)
+	fcmpu	cr1,f0,f29
+	bne	cr1,1f
+	lfd	f0,128(r4)
+	fcmpu	cr1,f0,f30
+	bne	cr1,1f
+	lfd	f0,136(r4)
+	fcmpu	cr1,f0,f31
+	bne	cr1,1f
+	li	r3,0 # Success!!!
+1:	blr
+
+
+// int check_all_fprs(double darray[32])
+FUNC_START(check_all_fprs)
+	PUSH_BASIC_STACK(8)
+	mr	r4, r3	// r4 = darray
+	li	r3, 1	// prepare for failure
+
+	stfd	f31, STACK_FRAME_LOCAL(0, 0)(sp) // backup f31
+
+	// Check regs f0-f30, using f31 as scratch
+	.set i, 0
+	.rept 31
+	lfd	f31, (8 * i)(r4)	// load expected value
+	fcmpu	cr0, i, f31		// compare
+	bne	cr0, 1f			// bail if mismatch
+	.set i, i + 1
+	.endr
+
+	lfd	f31, STACK_FRAME_LOCAL(0, 0)(sp) // reload f31
+	stfd	f30, STACK_FRAME_LOCAL(0, 0)(sp) // backup f30
+
+	lfd	f30, (8 * 31)(r4)	// load expected value of f31
+	fcmpu	cr0, f30, f31		// compare
+	bne	cr0, 1f			// bail if mismatch
+
+	lfd	f30, STACK_FRAME_LOCAL(0, 0)(sp) // reload f30
+
+	// Success
+	li	r3, 0
+
+1:	POP_BASIC_STACK(8)
+	blr
+FUNC_END(check_all_fprs)
+
+FUNC_START(test_fpu)
+	# r3 holds pointer to where to put the result of fork
+	# r4 holds pointer to the pid
+	# f14-f31 are non volatiles
+	PUSH_BASIC_STACK(256)
+	PUSH_FPU(256)
+	std	r3,STACK_FRAME_PARAM(0)(sp) # Address of darray
+	std r4,STACK_FRAME_PARAM(1)(sp) # Address of pid
+
+	// Load FPRs with expected values
+	OP_REGS lfd, 8, 0, 31, r3
+
+	li	r0,__NR_fork
+	sc
+
+	# pass the result of the fork to the caller
+	ld	r9,STACK_FRAME_PARAM(1)(sp)
+	std	r3,0(r9)
+
+	ld r3,STACK_FRAME_PARAM(0)(sp)
+	bl check_all_fprs
+	nop
+
+	POP_FPU(256)
+	POP_BASIC_STACK(256)
+	blr
+FUNC_END(test_fpu)
+
+# int preempt_fpu(double *darray, int *threads_running, int *running)
+# On starting will (atomically) decrement not_ready as a signal that the FPU
+# has been loaded with darray. Will proceed to check the validity of the FPU
+# registers while running is not zero.
+FUNC_START(preempt_fpu)
+	PUSH_BASIC_STACK(256)
+	PUSH_FPU(256)
+	std r3,STACK_FRAME_PARAM(0)(sp) # double *darray
+	std r4,STACK_FRAME_PARAM(1)(sp) # int *threads_starting
+	std r5,STACK_FRAME_PARAM(2)(sp) # int *running
+
+	// Load FPRs with expected values
+	OP_REGS lfd, 8, 0, 31, r3
+
+	sync
+	# Atomic DEC
+	ld r3,STACK_FRAME_PARAM(1)(sp)
+1:	lwarx r4,0,r3
+	addi r4,r4,-1
+	stwcx. r4,0,r3
+	bne- 1b
+
+2:	ld r3,STACK_FRAME_PARAM(0)(sp)
+	bl check_all_fprs
+	cmpdi r3,0
+	bne 3f
+	ld r4,STACK_FRAME_PARAM(2)(sp)
+	ld r5,0(r4)
+	cmpwi r5,0
+	bne 2b
+
+3:	POP_FPU(256)
+	POP_BASIC_STACK(256)
+	blr
+FUNC_END(preempt_fpu)

linux-7.2.3/tools/testing/selftests/powerpc/math/fpu_denormal.c

@@ -0,0 +1,38 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ * Copyright IBM Corp. 2020
+ *
+ * This test attempts to cause a FP denormal exception on POWER8 CPUs. Unfortunately
+ * if the denormal handler is not configured or working properly, this can cause a bad
+ * crash in kernel mode when the kernel tries to save FP registers when the process
+ * exits.
+ */
+
+#include <stdio.h>
+#include <string.h>
+
+#include "utils.h"
+
+static int test_denormal_fpu(void)
+{
+	unsigned int m32;
+	unsigned long m64;
+	volatile float f;
+	volatile double d;
+
+	/* try to induce lfs <denormal> ; stfd */
+
+	m32 = 0x00715fcf; /* random denormal */
+	memcpy((float *)&f, &m32, sizeof(f));
+	d = f;
+	memcpy(&m64, (double *)&d, sizeof(d));
+
+	FAIL_IF((long)(m64 != 0x380c57f3c0000000)); /* renormalised value */
+
+	return 0;
+}
+
+int main(int argc, char *argv[])
+{
+	return test_harness(test_denormal_fpu, "fpu_denormal");
+}

linux-7.2.3/tools/testing/selftests/powerpc/math/fpu_preempt.c

@@ -0,0 +1,105 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ * Copyright 2015, Cyril Bur, IBM Corp.
+ * Copyright 2023, Michael Ellerman, IBM Corp.
+ *
+ * This test attempts to see if the FPU registers change across preemption.
+ * There is no way to be sure preemption happened so this test just uses many
+ * threads and a long wait. As such, a successful test doesn't mean much but
+ * a failure is bad.
+ */
+
+#include <stdio.h>
+#include <unistd.h>
+#include <sys/syscall.h>
+#include <sys/time.h>
+#include <sys/types.h>
+#include <sys/wait.h>
+#include <stdlib.h>
+#include <pthread.h>
+
+#include "utils.h"
+#include "fpu.h"
+
+/* Time to wait for workers to get preempted (seconds) */
+#define PREEMPT_TIME 60
+/*
+ * Factor by which to multiply number of online CPUs for total number of
+ * worker threads
+ */
+#define THREAD_FACTOR 8
+
+
+__thread double darray[32];
+
+int threads_starting;
+int running;
+
+extern int preempt_fpu(double *darray, int *threads_starting, int *running);
+
+void *preempt_fpu_c(void *p)
+{
+	long rc;
+
+	srand(pthread_self());
+	randomise_darray(darray, ARRAY_SIZE(darray));
+	rc = preempt_fpu(darray, &threads_starting, &running);
+
+	return (void *)rc;
+}
+
+int test_preempt_fpu(void)
+{
+	int i, rc, threads;
+	pthread_t *tids;
+
+	threads = sysconf(_SC_NPROCESSORS_ONLN) * THREAD_FACTOR;
+	tids = malloc((threads) * sizeof(pthread_t));
+	FAIL_IF(!tids);
+
+	running = true;
+	threads_starting = threads;
+	for (i = 0; i < threads; i++) {
+		rc = pthread_create(&tids[i], NULL, preempt_fpu_c, NULL);
+		FAIL_IF(rc);
+	}
+
+	setbuf(stdout, NULL);
+	/* Not really necessary but nice to wait for every thread to start */
+	printf("\tWaiting for all workers to start...");
+	while(threads_starting)
+		asm volatile("": : :"memory");
+	printf("done\n");
+
+	printf("\tWaiting for %d seconds to let some workers get preempted...", PREEMPT_TIME);
+	sleep(PREEMPT_TIME);
+	printf("done\n");
+
+	printf("\tStopping workers...");
+	/*
+	 * Working are checking this value every loop. In preempt_fpu 'cmpwi r5,0; bne 2b'.
+	 * r5 will have loaded the value of running.
+	 */
+	running = 0;
+	for (i = 0; i < threads; i++) {
+		void *rc_p;
+		pthread_join(tids[i], &rc_p);
+
+		/*
+		 * Harness will say the fail was here, look at why preempt_fpu
+		 * returned
+		 */
+		if ((long) rc_p)
+			printf("oops\n");
+		FAIL_IF((long) rc_p);
+	}
+	printf("done\n");
+
+	free(tids);
+	return 0;
+}
+
+int main(int argc, char *argv[])
+{
+	return test_harness(test_preempt_fpu, "fpu_preempt");
+}

linux-7.2.3/tools/testing/selftests/powerpc/math/fpu_signal.c

@@ -0,0 +1,127 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ * Copyright 2015, Cyril Bur, IBM Corp.
+ *
+ * This test attempts to see if the FPU registers are correctly reported in a
+ * signal context. Each worker just spins checking its FPU registers, at some
+ * point a signal will interrupt it and C code will check the signal context
+ * ensuring it is also the same.
+ */
+
+#include <stdio.h>
+#include <unistd.h>
+#include <sys/syscall.h>
+#include <sys/time.h>
+#include <sys/types.h>
+#include <sys/wait.h>
+#include <stdlib.h>
+#include <pthread.h>
+
+#include "utils.h"
+#include "fpu.h"
+
+/* Number of times each thread should receive the signal */
+#define ITERATIONS 10
+/*
+ * Factor by which to multiply number of online CPUs for total number of
+ * worker threads
+ */
+#define THREAD_FACTOR 8
+
+__thread double darray[32];
+
+bool bad_context;
+int threads_starting;
+int running;
+
+extern long preempt_fpu(double *darray, int *threads_starting, int *running);
+
+void signal_fpu_sig(int sig, siginfo_t *info, void *context)
+{
+	int i;
+	ucontext_t *uc = context;
+	mcontext_t *mc = &uc->uc_mcontext;
+
+	// Don't check f30/f31, they're used as scratches in check_all_fprs()
+	for (i = 0; i < 30; i++) {
+		if (mc->fp_regs[i] != darray[i]) {
+			bad_context = true;
+			break;
+		}
+	}
+}
+
+void *signal_fpu_c(void *p)
+{
+	long rc;
+	struct sigaction act;
+	act.sa_sigaction = signal_fpu_sig;
+	act.sa_flags = SA_SIGINFO;
+	rc = sigaction(SIGUSR1, &act, NULL);
+	if (rc)
+		return p;
+
+	srand(pthread_self());
+	randomise_darray(darray, ARRAY_SIZE(darray));
+	rc = preempt_fpu(darray, &threads_starting, &running);
+
+	return (void *) rc;
+}
+
+int test_signal_fpu(void)
+{
+	int i, j, rc, threads;
+	void *rc_p;
+	pthread_t *tids;
+
+	threads = sysconf(_SC_NPROCESSORS_ONLN) * THREAD_FACTOR;
+	tids = malloc(threads * sizeof(pthread_t));
+	FAIL_IF(!tids);
+
+	running = true;
+	threads_starting = threads;
+	for (i = 0; i < threads; i++) {
+		rc = pthread_create(&tids[i], NULL, signal_fpu_c, NULL);
+		FAIL_IF(rc);
+	}
+
+	setbuf(stdout, NULL);
+	printf("\tWaiting for all workers to start...");
+	while (threads_starting)
+		asm volatile("": : :"memory");
+	printf("done\n");
+
+	printf("\tSending signals to all threads %d times...", ITERATIONS);
+	for (i = 0; i < ITERATIONS; i++) {
+		for (j = 0; j < threads; j++) {
+			pthread_kill(tids[j], SIGUSR1);
+		}
+		sleep(1);
+	}
+	printf("done\n");
+
+	printf("\tStopping workers...");
+	running = 0;
+	for (i = 0; i < threads; i++) {
+		pthread_join(tids[i], &rc_p);
+
+		/*
+		 * Harness will say the fail was here, look at why signal_fpu
+		 * returned
+		 */
+		if ((long) rc_p || bad_context)
+			printf("oops\n");
+		if (bad_context)
+			fprintf(stderr, "\t!! bad_context is true\n");
+		FAIL_IF((long) rc_p || bad_context);
+	}
+	printf("done\n");
+
+	free(tids);
+	return 0;
+}
+
+int main(int argc, char *argv[])
+{
+	return test_harness(test_signal_fpu, "fpu_signal");
+}

linux-7.2.3/tools/testing/selftests/powerpc/math/fpu_syscall.c

@@ -0,0 +1,88 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ * Copyright 2015, Cyril Bur, IBM Corp.
+ *
+ * This test attempts to see if the FPU registers change across a syscall (fork).
+ */
+
+#include <stdio.h>
+#include <unistd.h>
+#include <sys/syscall.h>
+#include <sys/time.h>
+#include <sys/types.h>
+#include <sys/wait.h>
+#include <stdlib.h>
+
+#include "utils.h"
+#include "fpu.h"
+
+extern int test_fpu(double *darray, pid_t *pid);
+
+double darray[32];
+
+int syscall_fpu(void)
+{
+	pid_t fork_pid;
+	int i;
+	int ret;
+	int child_ret;
+
+	randomise_darray(darray, ARRAY_SIZE(darray));
+
+	for (i = 0; i < 1000; i++) {
+		/* test_fpu will fork() */
+		ret = test_fpu(darray, &fork_pid);
+		if (fork_pid == -1)
+			return -1;
+		if (fork_pid == 0)
+			exit(ret);
+		waitpid(fork_pid, &child_ret, 0);
+		if (ret || child_ret)
+			return 1;
+	}
+
+	return 0;
+}
+
+int test_syscall_fpu(void)
+{
+	/*
+	 * Setup an environment with much context switching
+	 */
+	pid_t pid2;
+	pid_t pid = fork();
+	int ret;
+	int child_ret;
+	FAIL_IF(pid == -1);
+
+	pid2 = fork();
+	/* Can't FAIL_IF(pid2 == -1); because already forked once */
+	if (pid2 == -1) {
+		/*
+		 * Couldn't fork, ensure test is a fail
+		 */
+		child_ret = ret = 1;
+	} else {
+		ret = syscall_fpu();
+		if (pid2)
+			waitpid(pid2, &child_ret, 0);
+		else
+			exit(ret);
+	}
+
+	ret |= child_ret;
+
+	if (pid)
+		waitpid(pid, &child_ret, 0);
+	else
+		exit(ret);
+
+	FAIL_IF(ret || child_ret);
+	return 0;
+}
+
+int main(int argc, char *argv[])
+{
+	return test_harness(test_syscall_fpu, "syscall_fpu");
+
+}

linux-7.2.3/tools/testing/selftests/powerpc/math/mma.S

@@ -0,0 +1,36 @@
+/* SPDX-License-Identifier: GPL-2.0-or-later
+ *
+ * Test basic matrix multiply assist (MMA) functionality if available.
+ *
+ * Copyright 2020, Alistair Popple, IBM Corp.
+ */
+	.global test_mma
+test_mma:
+	/* Load accumulator via VSX registers from image passed in r3 */
+	lxvh8x	4,0,3
+	lxvh8x	5,0,4
+
+	/* Clear and prime the accumulator (xxsetaccz) */
+	.long	0x7c030162
+
+	/* Prime the accumulator with MMA VSX move to accumulator
+	* X-form (xxmtacc) (not needed due to above zeroing) */
+	//.long 0x7c010162
+
+	/* xvi16ger2s */
+	.long	0xec042958
+
+	/* Deprime the accumulator - xxmfacc 0 */
+	.long 0x7c000162
+
+	/* Store result in image passed in r5 */
+	stxvw4x	0,0,5
+	addi	5,5,16
+	stxvw4x	1,0,5
+	addi	5,5,16
+	stxvw4x	2,0,5
+	addi	5,5,16
+	stxvw4x	3,0,5
+	addi	5,5,16
+
+	blr

linux-7.2.3/tools/testing/selftests/powerpc/math/mma.c

@@ -0,0 +1,48 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ * Test basic matrix multiply assist (MMA) functionality if available.
+ *
+ * Copyright 2020, Alistair Popple, IBM Corp.
+ */
+#include <stdio.h>
+#include <stdint.h>
+
+#include "utils.h"
+
+extern void test_mma(uint16_t (*)[8], uint16_t (*)[8], uint32_t (*)[4*4]);
+
+static int mma(void)
+{
+	int i;
+	int rc = 0;
+	uint16_t x[] = {1, 0, 2, 0, 3, 0, 4, 0};
+	uint16_t y[] = {1, 0, 2, 0, 3, 0, 4, 0};
+	uint32_t z[4*4];
+	uint32_t exp[4*4] = {1, 2, 3, 4,
+			     2, 4, 6, 8,
+			     3, 6, 9, 12,
+			     4, 8, 12, 16};
+
+	SKIP_IF_MSG(!have_hwcap2(PPC_FEATURE2_ARCH_3_1), "Need ISAv3.1");
+	SKIP_IF_MSG(!have_hwcap2(PPC_FEATURE2_MMA), "Need MMA");
+
+	test_mma(&x, &y, &z);
+
+	for (i = 0; i < 16; i++) {
+		printf("MMA[%d] = %d ", i, z[i]);
+
+		if (z[i] == exp[i]) {
+			printf(" (Correct)\n");
+		} else {
+			printf(" (Incorrect)\n");
+			rc = 1;
+		}
+	}
+
+	return rc;
+}
+
+int main(int argc, char *argv[])
+{
+	return test_harness(mma, "mma");
+}

linux-7.2.3/tools/testing/selftests/powerpc/math/settings

@@ -0,0 +1 @@
+timeout=130

linux-7.2.3/tools/testing/selftests/powerpc/math/vmx_asm.S

@@ -0,0 +1,148 @@
+/* SPDX-License-Identifier: GPL-2.0-or-later */
+/*
+ * Copyright 2015, Cyril Bur, IBM Corp.
+ */
+
+#include "basic_asm.h"
+#include "vmx_asm.h"
+
+# Should be safe from C, only touches r4, r5 and v0,v1,v2
+FUNC_START(check_vmx)
+	PUSH_BASIC_STACK(32)
+	mr r4,r3
+	li	r3,1 # assume a bad result
+	li	r5,0
+	lvx	v0,r5,r4
+	vcmpequd.	v1,v0,v20
+	vmr	v2,v1
+
+	addi	r5,r5,16
+	lvx	v0,r5,r4
+	vcmpequd.	v1,v0,v21
+	vand	v2,v2,v1
+
+	addi	r5,r5,16
+	lvx	v0,r5,r4
+	vcmpequd.	v1,v0,v22
+	vand	v2,v2,v1
+
+	addi	r5,r5,16
+	lvx	v0,r5,r4
+	vcmpequd.	v1,v0,v23
+	vand	v2,v2,v1
+
+	addi	r5,r5,16
+	lvx	v0,r5,r4
+	vcmpequd.	v1,v0,v24
+	vand	v2,v2,v1
+
+	addi	r5,r5,16
+	lvx	v0,r5,r4
+	vcmpequd.	v1,v0,v25
+	vand	v2,v2,v1
+
+	addi	r5,r5,16
+	lvx	v0,r5,r4
+	vcmpequd.	v1,v0,v26
+	vand	v2,v2,v1
+
+	addi	r5,r5,16
+	lvx	v0,r5,r4
+	vcmpequd.	v1,v0,v27
+	vand	v2,v2,v1
+
+	addi	r5,r5,16
+	lvx	v0,r5,r4
+	vcmpequd.	v1,v0,v28
+	vand	v2,v2,v1
+
+	addi	r5,r5,16
+	lvx	v0,r5,r4
+	vcmpequd.	v1,v0,v29
+	vand	v2,v2,v1
+
+	addi	r5,r5,16
+	lvx	v0,r5,r4
+	vcmpequd.	v1,v0,v30
+	vand	v2,v2,v1
+
+	addi	r5,r5,16
+	lvx	v0,r5,r4
+	vcmpequd.	v1,v0,v31
+	vand	v2,v2,v1
+
+	li	r5,STACK_FRAME_LOCAL(0,0)
+	stvx	v2,r5,sp
+	ldx	r0,r5,sp
+	cmpdi	r0,0xffffffffffffffff
+	bne	1f
+	li	r3,0
+1:	POP_BASIC_STACK(32)
+	blr
+FUNC_END(check_vmx)
+
+# Safe from C
+FUNC_START(test_vmx)
+	# r3 holds pointer to where to put the result of fork
+	# r4 holds pointer to the pid
+	# v20-v31 are non-volatile
+	PUSH_BASIC_STACK(512)
+	std	r3,STACK_FRAME_PARAM(0)(sp) # Address of varray
+	std r4,STACK_FRAME_PARAM(1)(sp) # address of pid
+	PUSH_VMX(STACK_FRAME_LOCAL(2,0),r4)
+
+	bl load_vmx
+	nop
+
+	li	r0,__NR_fork
+	sc
+	# Pass the result of fork back to the caller
+	ld	r9,STACK_FRAME_PARAM(1)(sp)
+	std	r3,0(r9)
+
+	ld r3,STACK_FRAME_PARAM(0)(sp)
+	bl check_vmx
+	nop
+
+	POP_VMX(STACK_FRAME_LOCAL(2,0),r4)
+	POP_BASIC_STACK(512)
+	blr
+FUNC_END(test_vmx)
+
+# int preempt_vmx(vector int *varray, int *threads_starting, int *running)
+# On starting will (atomically) decrement threads_starting as a signal that
+# the VMX have been loaded with varray. Will proceed to check the validity of
+# the VMX registers while running is not zero.
+FUNC_START(preempt_vmx)
+	PUSH_BASIC_STACK(512)
+	std r3,STACK_FRAME_PARAM(0)(sp) # vector int *varray
+	std r4,STACK_FRAME_PARAM(1)(sp) # int *threads_starting
+	std r5,STACK_FRAME_PARAM(2)(sp) # int *running
+	# VMX need to write to 16 byte aligned addresses, skip STACK_FRAME_LOCAL(3,0)
+	PUSH_VMX(STACK_FRAME_LOCAL(4,0),r4)
+
+	bl load_vmx
+	nop
+
+	sync
+	# Atomic DEC
+	ld r3,STACK_FRAME_PARAM(1)(sp)
+1:	lwarx r4,0,r3
+	addi r4,r4,-1
+	stwcx. r4,0,r3
+	bne- 1b
+
+2:	ld r3,STACK_FRAME_PARAM(0)(sp)
+	bl check_vmx
+	nop
+	cmpdi r3,0
+	bne 3f
+	ld r4,STACK_FRAME_PARAM(2)(sp)
+	ld r5,0(r4)
+	cmpwi r5,0
+	bne 2b
+
+3:	POP_VMX(STACK_FRAME_LOCAL(4,0),r4)
+	POP_BASIC_STACK(512)
+	blr
+FUNC_END(preempt_vmx)

linux-7.2.3/tools/testing/selftests/powerpc/math/vmx_preempt.c

@@ -0,0 +1,113 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ * Copyright 2015, Cyril Bur, IBM Corp.
+ *
+ * This test attempts to see if the VMX registers change across preemption.
+ * Two things should be noted here a) The check_vmx function in asm only checks
+ * the non volatile registers as it is reused from the syscall test b) There is
+ * no way to be sure preemption happened so this test just uses many threads
+ * and a long wait. As such, a successful test doesn't mean much but a failure
+ * is bad.
+ */
+
+#include <stdio.h>
+#include <unistd.h>
+#include <sys/syscall.h>
+#include <sys/time.h>
+#include <sys/types.h>
+#include <sys/wait.h>
+#include <stdlib.h>
+#include <pthread.h>
+
+#include "utils.h"
+
+/* Time to wait for workers to get preempted (seconds) */
+#define PREEMPT_TIME 20
+/*
+ * Factor by which to multiply number of online CPUs for total number of
+ * worker threads
+ */
+#define THREAD_FACTOR 8
+
+__thread vector int varray[] = {{1, 2, 3, 4}, {5, 6, 7, 8}, {9, 10,11,12},
+	{13,14,15,16},{17,18,19,20},{21,22,23,24},
+	{25,26,27,28},{29,30,31,32},{33,34,35,36},
+	{37,38,39,40},{41,42,43,44},{45,46,47,48}};
+
+int threads_starting;
+int running;
+
+extern int preempt_vmx(vector int *varray, int *threads_starting, int *running);
+
+void *preempt_vmx_c(void *p)
+{
+	int i, j;
+	long rc;
+
+	srand(pthread_self());
+	for (i = 0; i < 12; i++)
+		for (j = 0; j < 4; j++)
+			varray[i][j] = rand();
+
+	rc = preempt_vmx(varray, &threads_starting, &running);
+
+	return (void *)rc;
+}
+
+int test_preempt_vmx(void)
+{
+	int i, rc, threads;
+	pthread_t *tids;
+
+	// vcmpequd used in vmx_asm.S is v2.07
+	SKIP_IF(!have_hwcap2(PPC_FEATURE2_ARCH_2_07));
+
+	threads = sysconf(_SC_NPROCESSORS_ONLN) * THREAD_FACTOR;
+	tids = malloc(threads * sizeof(pthread_t));
+	FAIL_IF(!tids);
+
+	running = true;
+	threads_starting = threads;
+	for (i = 0; i < threads; i++) {
+		rc = pthread_create(&tids[i], NULL, preempt_vmx_c, NULL);
+		FAIL_IF(rc);
+	}
+
+	setbuf(stdout, NULL);
+	/* Not really nessesary but nice to wait for every thread to start */
+	printf("\tWaiting for all workers to start...");
+	while(threads_starting)
+		asm volatile("": : :"memory");
+	printf("done\n");
+
+	printf("\tWaiting for %d seconds to let some workers get preempted...", PREEMPT_TIME);
+	sleep(PREEMPT_TIME);
+	printf("done\n");
+
+	printf("\tStopping workers...");
+	/*
+	 * Working are checking this value every loop. In preempt_vmx 'cmpwi r5,0; bne 2b'.
+	 * r5 will have loaded the value of running.
+	 */
+	running = 0;
+	for (i = 0; i < threads; i++) {
+		void *rc_p;
+		pthread_join(tids[i], &rc_p);
+
+		/*
+		 * Harness will say the fail was here, look at why preempt_vmx
+		 * returned
+		 */
+		if ((long) rc_p)
+			printf("oops\n");
+		FAIL_IF((long) rc_p);
+	}
+	printf("done\n");
+
+	return 0;
+}
+
+int main(int argc, char *argv[])
+{
+	return test_harness(test_preempt_vmx, "vmx_preempt");
+}

linux-7.2.3/tools/testing/selftests/powerpc/math/vmx_signal.c

@@ -0,0 +1,156 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ * Copyright 2015, Cyril Bur, IBM Corp.
+ *
+ * This test attempts to see if the VMX registers are correctly reported in a
+ * signal context. Each worker just spins checking its VMX registers, at some
+ * point a signal will interrupt it and C code will check the signal context
+ * ensuring it is also the same.
+ */
+
+#include <stdio.h>
+#include <unistd.h>
+#include <sys/syscall.h>
+#include <sys/time.h>
+#include <sys/types.h>
+#include <sys/wait.h>
+#include <stdlib.h>
+#include <string.h>
+#include <pthread.h>
+#include <altivec.h>
+
+#include "utils.h"
+
+/* Number of times each thread should receive the signal */
+#define ITERATIONS 10
+/*
+ * Factor by which to multiply number of online CPUs for total number of
+ * worker threads
+ */
+#define THREAD_FACTOR 8
+
+__thread vector int varray[] = {{1, 2, 3, 4}, {5, 6, 7, 8}, {9, 10,11,12},
+	{13,14,15,16},{17,18,19,20},{21,22,23,24},
+	{25,26,27,28},{29,30,31,32},{33,34,35,36},
+	{37,38,39,40},{41,42,43,44},{45,46,47,48}};
+
+bool bad_context;
+int running;
+int threads_starting;
+
+extern int preempt_vmx(vector int *varray, int *threads_starting, int *sentinal);
+
+void signal_vmx_sig(int sig, siginfo_t *info, void *context)
+{
+	int i;
+	ucontext_t *uc = context;
+	mcontext_t *mc = &uc->uc_mcontext;
+
+	/* Only the non volatiles were loaded up */
+	for (i = 20; i < 32; i++) {
+		if (memcmp(mc->v_regs->vrregs[i], &varray[i - 20], 16)) {
+			int j;
+			/*
+			 * Shouldn't printf() in a signal handler, however, this is a
+			 * test and we've detected failure. Understanding what failed
+			 * is paramount. All that happens after this is tests exit with
+			 * failure.
+			 */
+			printf("VMX mismatch at reg %d!\n", i);
+			printf("Reg | Actual                  | Expected\n");
+			for (j = 20; j < 32; j++) {
+				printf("%d  | 0x%04x%04x%04x%04x      | 0x%04x%04x%04x%04x\n", j, mc->v_regs->vrregs[j][0],
+					   mc->v_regs->vrregs[j][1], mc->v_regs->vrregs[j][2], mc->v_regs->vrregs[j][3],
+					   varray[j - 20][0], varray[j - 20][1], varray[j - 20][2], varray[j - 20][3]);
+			}
+			bad_context = true;
+			break;
+		}
+	}
+}
+
+void *signal_vmx_c(void *p)
+{
+	int i, j;
+	long rc;
+	struct sigaction act;
+	act.sa_sigaction = signal_vmx_sig;
+	act.sa_flags = SA_SIGINFO;
+	rc = sigaction(SIGUSR1, &act, NULL);
+	if (rc)
+		return p;
+
+	srand(pthread_self());
+	for (i = 0; i < 12; i++)
+		for (j = 0; j < 4; j++)
+			varray[i][j] = rand();
+
+	rc = preempt_vmx(varray, &threads_starting, &running);
+
+	return (void *) rc;
+}
+
+int test_signal_vmx(void)
+{
+	int i, j, rc, threads;
+	void *rc_p;
+	pthread_t *tids;
+
+	// vcmpequd used in vmx_asm.S is v2.07
+	SKIP_IF(!have_hwcap2(PPC_FEATURE2_ARCH_2_07));
+
+	threads = sysconf(_SC_NPROCESSORS_ONLN) * THREAD_FACTOR;
+	tids = malloc(threads * sizeof(pthread_t));
+	FAIL_IF(!tids);
+
+	running = true;
+	threads_starting = threads;
+	for (i = 0; i < threads; i++) {
+		rc = pthread_create(&tids[i], NULL, signal_vmx_c, NULL);
+		FAIL_IF(rc);
+	}
+
+	setbuf(stdout, NULL);
+	printf("\tWaiting for %d workers to start... %d", threads, threads_starting);
+	while (threads_starting) {
+		asm volatile("": : :"memory");
+		usleep(1000);
+		printf(", %d", threads_starting);
+	}
+	printf(" ...done\n");
+
+	printf("\tSending signals to all threads %d times...", ITERATIONS);
+	for (i = 0; i < ITERATIONS; i++) {
+		for (j = 0; j < threads; j++) {
+			pthread_kill(tids[j], SIGUSR1);
+		}
+		sleep(1);
+	}
+	printf("done\n");
+
+	printf("\tKilling workers...");
+	running = 0;
+	for (i = 0; i < threads; i++) {
+		pthread_join(tids[i], &rc_p);
+
+		/*
+		 * Harness will say the fail was here, look at why signal_vmx
+		 * returned
+		 */
+		if ((long) rc_p || bad_context)
+			printf("oops\n");
+		if (bad_context)
+			fprintf(stderr, "\t!! bad_context is true\n");
+		FAIL_IF((long) rc_p || bad_context);
+	}
+	printf("done\n");
+
+	free(tids);
+	return 0;
+}
+
+int main(int argc, char *argv[])
+{
+	test_harness_set_timeout(360);
+	return test_harness(test_signal_vmx, "vmx_signal");
+}

linux-7.2.3/tools/testing/selftests/powerpc/math/vmx_syscall.c

@@ -0,0 +1,92 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ * Copyright 2015, Cyril Bur, IBM Corp.
+ *
+ * This test attempts to see if the VMX registers change across a syscall (fork).
+ */
+
+#include <altivec.h>
+#include <stdio.h>
+#include <unistd.h>
+#include <sys/syscall.h>
+#include <sys/time.h>
+#include <stdlib.h>
+#include <sys/types.h>
+#include <sys/wait.h>
+#include "utils.h"
+
+vector int varray[] = {{1, 2, 3, 4}, {5, 6, 7, 8}, {9, 10,11,12},
+	{13,14,15,16},{17,18,19,20},{21,22,23,24},
+	{25,26,27,28},{29,30,31,32},{33,34,35,36},
+	{37,38,39,40},{41,42,43,44},{45,46,47,48}};
+
+extern int test_vmx(vector int *varray, pid_t *pid);
+
+int vmx_syscall(void)
+{
+	pid_t fork_pid;
+	int i;
+	int ret;
+	int child_ret;
+	for (i = 0; i < 1000; i++) {
+		/* test_vmx will fork() */
+		ret = test_vmx(varray, &fork_pid);
+		if (fork_pid == -1)
+			return -1;
+		if (fork_pid == 0)
+			exit(ret);
+		waitpid(fork_pid, &child_ret, 0);
+		if (ret || child_ret)
+			return 1;
+	}
+
+	return 0;
+}
+
+int test_vmx_syscall(void)
+{
+	/*
+	 * Setup an environment with much context switching
+	 */
+	pid_t pid2;
+	pid_t pid;
+	int ret;
+	int child_ret;
+
+	// vcmpequd used in vmx_asm.S is v2.07
+	SKIP_IF(!have_hwcap2(PPC_FEATURE2_ARCH_2_07));
+
+	pid = fork();
+	FAIL_IF(pid == -1);
+
+	pid2 = fork();
+	ret = vmx_syscall();
+	/* Can't FAIL_IF(pid2 == -1); because we've already forked */
+	if (pid2 == -1) {
+		/*
+		 * Couldn't fork, ensure child_ret is set and is a fail
+		 */
+		ret = child_ret = 1;
+	} else {
+		if (pid2)
+			waitpid(pid2, &child_ret, 0);
+		else
+			exit(ret);
+	}
+
+	ret |= child_ret;
+
+	if (pid)
+		waitpid(pid, &child_ret, 0);
+	else
+		exit(ret);
+
+	FAIL_IF(ret || child_ret);
+	return 0;
+}
+
+int main(int argc, char *argv[])
+{
+	return test_harness(test_vmx_syscall, "vmx_syscall");
+
+}

linux-7.2.3/tools/testing/selftests/powerpc/math/vsx_asm.S

@@ -0,0 +1,57 @@
+/* SPDX-License-Identifier: GPL-2.0-or-later */
+/*
+ * Copyright 2015, Cyril Bur, IBM Corp.
+ */
+
+#include "basic_asm.h"
+#include "vsx_asm.h"
+
+#long check_vsx(vector int *r3);
+#This function wraps storeing VSX regs to the end of an array and a
+#call to a comparison function in C which boils down to a memcmp()
+FUNC_START(check_vsx)
+	PUSH_BASIC_STACK(32)
+	std	r3,STACK_FRAME_PARAM(0)(sp)
+	addi r3, r3, 16 * 12 #Second half of array
+	bl store_vsx
+	ld r3,STACK_FRAME_PARAM(0)(sp)
+	bl vsx_memcmp
+	POP_BASIC_STACK(32)
+	blr
+FUNC_END(check_vsx)
+
+# int preempt_vmx(vector int *varray, int *threads_starting,
+#                 int *running);
+# On starting will (atomically) decrement threads_starting as a signal
+# that the VMX have been loaded with varray. Will proceed to check the
+# validity of the VMX registers while running is not zero.
+FUNC_START(preempt_vsx)
+	PUSH_BASIC_STACK(512)
+	std r3,STACK_FRAME_PARAM(0)(sp) # vector int *varray
+	std r4,STACK_FRAME_PARAM(1)(sp) # int *threads_starting
+	std r5,STACK_FRAME_PARAM(2)(sp) # int *running
+
+	bl load_vsx
+	nop
+
+	sync
+	# Atomic DEC
+	ld r3,STACK_FRAME_PARAM(1)(sp)
+1:	lwarx r4,0,r3
+	addi r4,r4,-1
+	stwcx. r4,0,r3
+	bne- 1b
+
+2:	ld r3,STACK_FRAME_PARAM(0)(sp)
+	bl check_vsx
+	nop
+	cmpdi r3,0
+	bne 3f
+	ld r4,STACK_FRAME_PARAM(2)(sp)
+	ld r5,0(r4)
+	cmpwi r5,0
+	bne 2b
+
+3:	POP_BASIC_STACK(512)
+	blr
+FUNC_END(preempt_vsx)

linux-7.2.3/tools/testing/selftests/powerpc/math/vsx_preempt.c

@@ -0,0 +1,145 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ * Copyright 2015, Cyril Bur, IBM Corp.
+ *
+ * This test attempts to see if the VSX registers change across preemption.
+ * There is no way to be sure preemption happened so this test just
+ * uses many threads and a long wait. As such, a successful test
+ * doesn't mean much but a failure is bad.
+ */
+
+#include <stdio.h>
+#include <string.h>
+#include <unistd.h>
+#include <sys/syscall.h>
+#include <sys/time.h>
+#include <sys/types.h>
+#include <sys/wait.h>
+#include <stdlib.h>
+#include <pthread.h>
+
+#include "utils.h"
+
+/* Time to wait for workers to get preempted (seconds) */
+#define PREEMPT_TIME 20
+/*
+ * Factor by which to multiply number of online CPUs for total number of
+ * worker threads
+ */
+#define THREAD_FACTOR 8
+
+/*
+ * Ensure there is twice the number of non-volatile VMX regs!
+ * check_vmx() is going to use the other half as space to put the live
+ * registers before calling vsx_memcmp()
+ */
+__thread vector int varray[24] = {
+	{1, 2, 3, 4 }, {5, 6, 7, 8 }, {9, 10,11,12},
+	{13,14,15,16}, {17,18,19,20}, {21,22,23,24},
+	{25,26,27,28}, {29,30,31,32}, {33,34,35,36},
+	{37,38,39,40}, {41,42,43,44}, {45,46,47,48}
+};
+
+int threads_starting;
+int running;
+
+extern long preempt_vsx(vector int *varray, int *threads_starting, int *running);
+
+long vsx_memcmp(vector int *a) {
+	vector int zero = {0, 0, 0, 0};
+	int i;
+
+	FAIL_IF(a != varray);
+
+	for(i = 0; i < 12; i++) {
+		if (memcmp(&a[i + 12], &zero, sizeof(vector int)) == 0) {
+			fprintf(stderr, "Detected zero from the VSX reg %d\n", i + 12);
+			return 2;
+		}
+	}
+
+	if (memcmp(a, &a[12], 12 * sizeof(vector int))) {
+		long *p = (long *)a;
+		fprintf(stderr, "VSX mismatch\n");
+		for (i = 0; i < 24; i=i+2)
+			fprintf(stderr, "%d: 0x%08lx%08lx | 0x%08lx%08lx\n",
+					i/2 + i%2 + 20, p[i], p[i + 1], p[i + 24], p[i + 25]);
+		return 1;
+	}
+	return 0;
+}
+
+void *preempt_vsx_c(void *p)
+{
+	int i, j;
+	long rc;
+	srand(pthread_self());
+	for (i = 0; i < 12; i++)
+		for (j = 0; j < 4; j++) {
+			varray[i][j] = rand();
+			/* Don't want zero because it hides kernel problems */
+			if (varray[i][j] == 0)
+				j--;
+		}
+	rc = preempt_vsx(varray, &threads_starting, &running);
+	if (rc == 2)
+		fprintf(stderr, "Caught zeros in VSX compares\n");
+	return (void *)rc;
+}
+
+int test_preempt_vsx(void)
+{
+	int i, rc, threads;
+	pthread_t *tids;
+
+	SKIP_IF(!have_hwcap(PPC_FEATURE_HAS_VSX));
+
+	threads = sysconf(_SC_NPROCESSORS_ONLN) * THREAD_FACTOR;
+	tids = malloc(threads * sizeof(pthread_t));
+	FAIL_IF(!tids);
+
+	running = true;
+	threads_starting = threads;
+	for (i = 0; i < threads; i++) {
+		rc = pthread_create(&tids[i], NULL, preempt_vsx_c, NULL);
+		FAIL_IF(rc);
+	}
+
+	setbuf(stdout, NULL);
+	/* Not really nessesary but nice to wait for every thread to start */
+	printf("\tWaiting for %d workers to start...", threads_starting);
+	while(threads_starting)
+		asm volatile("": : :"memory");
+	printf("done\n");
+
+	printf("\tWaiting for %d seconds to let some workers get preempted...", PREEMPT_TIME);
+	sleep(PREEMPT_TIME);
+	printf("done\n");
+
+	printf("\tStopping workers...");
+	/*
+	 * Working are checking this value every loop. In preempt_vsx 'cmpwi r5,0; bne 2b'.
+	 * r5 will have loaded the value of running.
+	 */
+	running = 0;
+	for (i = 0; i < threads; i++) {
+		void *rc_p;
+		pthread_join(tids[i], &rc_p);
+
+		/*
+		 * Harness will say the fail was here, look at why preempt_vsx
+		 * returned
+		 */
+		if ((long) rc_p)
+			printf("oops\n");
+		FAIL_IF((long) rc_p);
+	}
+	printf("done\n");
+
+	return 0;
+}
+
+int main(int argc, char *argv[])
+{
+	return test_harness(test_preempt_vsx, "vsx_preempt");
+}

linux-7.2.3/tools/testing/selftests/powerpc/mce/.gitignore

@@ -0,0 +1 @@
+inject-ra-err

linux-7.2.3/tools/testing/selftests/powerpc/mce/Makefile

@@ -0,0 +1,8 @@
+#SPDX-License-Identifier: GPL-2.0-or-later
+
+TEST_GEN_PROGS := inject-ra-err
+
+include ../../lib.mk
+include ../flags.mk
+
+$(TEST_GEN_PROGS): ../harness.c