| lists.openwall.net | lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC | |
|
Open Source and information security mailing list archives
| ||
|
Message-Id: <20171012230326.19984-15-mathieu.desnoyers@efficios.com>
Date: Thu, 12 Oct 2017 19:03:26 -0400
From: Mathieu Desnoyers <mathieu.desnoyers@...icios.com>
To: "Paul E. McKenney" <paulmck@...ux.vnet.ibm.com>,
Boqun Feng <boqun.feng@...il.com>,
Peter Zijlstra <peterz@...radead.org>,
Paul Turner <pjt@...gle.com>, Andrew Hunter <ahh@...gle.com>,
Andy Lutomirski <luto@...capital.net>,
Dave Watson <davejwatson@...com>,
Josh Triplett <josh@...htriplett.org>,
Will Deacon <will.deacon@....com>
Cc: linux-kernel@...r.kernel.org,
Mathieu Desnoyers <mathieu.desnoyers@...icios.com>,
Russell King <linux@....linux.org.uk>,
Catalin Marinas <catalin.marinas@....com>,
Thomas Gleixner <tglx@...utronix.de>,
Andi Kleen <andi@...stfloor.org>, Chris Lameter <cl@...ux.com>,
Ingo Molnar <mingo@...hat.com>,
"H. Peter Anvin" <hpa@...or.com>, Ben Maurer <bmaurer@...com>,
Steven Rostedt <rostedt@...dmis.org>,
Linus Torvalds <torvalds@...ux-foundation.org>,
Andrew Morton <akpm@...ux-foundation.org>,
Shuah Khan <shuah@...nel.org>, linux-kselftest@...r.kernel.org,
linux-api@...r.kernel.org
Subject: [RFC PATCH for 4.15 14/14] Restartable sequences: Provide self-tests
Implements two basic tests of RSEQ functionality, and one more
exhaustive parameterizable test.
The first, "basic_test" only asserts that RSEQ works moderately
correctly.
E.g. that:
- The CPUID pointer works
- Code infinitely looping within a critical section will eventually be
interrupted.
- Critical sections are interrupted by signals.
"basic_percpu_ops_test" is a slightly more "realistic" variant,
implementing a few simple per-cpu operations and testing their
correctness.
"param_test" is a parametrizable restartable sequences test. See
the "--help" output for usage.
As part of those tests, a helper library "rseq" implements a user-space
API around restartable sequences. It uses the cpu_opv system call as
fallback when single-stepped by a debugger. It exposes the instruction
pointer addresses where the rseq assembly blocks begin and end, as well
as the associated abort instruction pointer, in the __rseq_table
section. This section allows debuggers may know where to place
breakpoints when single-stepping through assembly blocks which may be
aborted at any point by the kernel.
The following rseq APIs are implemented in this helper library:
- rseq_register_current_thread()/rseq_unregister_current_thread():
register/unregister current thread's use of rseq,
- rseq_current_cpu_raw():
current CPU number,
- rseq_start():
beginning of a restartable sequence,
- rseq_cpu_at_start():
CPU number at start of restartable sequence,
- rseq_finish():
End of restartable sequence made of zero or more loads, completed by
a word-sized store,
- rseq_finish2():
End of restartable sequence made of zero or more loads, one
speculative word-sized store, completed by a word-sized store,
- rseq_finish2_release():
End of restartable sequence made of zero or more loads, one
speculative word-sized store, completed by a word-sized store with
release semantic,
- rseq_finish_memcpy():
End of restartable sequence made of zero or more loads, a
speculative copy of a variable length memory region, completed by a
word-sized store.
- rseq_finish_memcpy_release():
End of restartable sequence made of zero or more loads, a
speculative copy of a variable length memory region, completed by a
word-sized store with release semantic.
PowerPC tests have been implemented by Boqun Feng.
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@...icios.com>
CC: Russell King <linux@....linux.org.uk>
CC: Catalin Marinas <catalin.marinas@....com>
CC: Will Deacon <will.deacon@....com>
CC: Thomas Gleixner <tglx@...utronix.de>
CC: Paul Turner <pjt@...gle.com>
CC: Andrew Hunter <ahh@...gle.com>
CC: Peter Zijlstra <peterz@...radead.org>
CC: Andy Lutomirski <luto@...capital.net>
CC: Andi Kleen <andi@...stfloor.org>
CC: Dave Watson <davejwatson@...com>
CC: Chris Lameter <cl@...ux.com>
CC: Ingo Molnar <mingo@...hat.com>
CC: "H. Peter Anvin" <hpa@...or.com>
CC: Ben Maurer <bmaurer@...com>
CC: Steven Rostedt <rostedt@...dmis.org>
CC: "Paul E. McKenney" <paulmck@...ux.vnet.ibm.com>
CC: Josh Triplett <josh@...htriplett.org>
CC: Linus Torvalds <torvalds@...ux-foundation.org>
CC: Andrew Morton <akpm@...ux-foundation.org>
CC: Boqun Feng <boqun.feng@...il.com>
CC: Shuah Khan <shuah@...nel.org>
CC: linux-kselftest@...r.kernel.org
CC: linux-api@...r.kernel.org
---
MAINTAINERS | 1 +
tools/testing/selftests/rseq/.gitignore | 4 +
tools/testing/selftests/rseq/Makefile | 13 +
.../testing/selftests/rseq/basic_percpu_ops_test.c | 319 +++++
tools/testing/selftests/rseq/basic_test.c | 97 ++
tools/testing/selftests/rseq/param_test.c | 1246 ++++++++++++++++++++
tools/testing/selftests/rseq/rseq-arm.h | 159 +++
tools/testing/selftests/rseq/rseq-ppc.h | 266 +++++
tools/testing/selftests/rseq/rseq-x86.h | 304 +++++
tools/testing/selftests/rseq/rseq.c | 78 ++
tools/testing/selftests/rseq/rseq.h | 298 +++++
11 files changed, 2785 insertions(+)
create mode 100644 tools/testing/selftests/rseq/.gitignore
create mode 100644 tools/testing/selftests/rseq/Makefile
create mode 100644 tools/testing/selftests/rseq/basic_percpu_ops_test.c
create mode 100644 tools/testing/selftests/rseq/basic_test.c
create mode 100644 tools/testing/selftests/rseq/param_test.c
create mode 100644 tools/testing/selftests/rseq/rseq-arm.h
create mode 100644 tools/testing/selftests/rseq/rseq-ppc.h
create mode 100644 tools/testing/selftests/rseq/rseq-x86.h
create mode 100644 tools/testing/selftests/rseq/rseq.c
create mode 100644 tools/testing/selftests/rseq/rseq.h
diff --git a/MAINTAINERS b/MAINTAINERS
index 9134a3234737..a79b0b473e7f 100644
--- a/MAINTAINERS
+++ b/MAINTAINERS
@@ -11242,6 +11242,7 @@ S: Supported
F: kernel/rseq.c
F: include/uapi/linux/rseq.h
F: include/trace/events/rseq.h
+F: tools/testing/selftests/rseq/
RFKILL
M: Johannes Berg <johannes@...solutions.net>
diff --git a/tools/testing/selftests/rseq/.gitignore b/tools/testing/selftests/rseq/.gitignore
new file mode 100644
index 000000000000..9409c3db99b2
--- /dev/null
+++ b/tools/testing/selftests/rseq/.gitignore
@@ -0,0 +1,4 @@
+basic_percpu_ops_test
+basic_test
+basic_rseq_op_test
+param_test
diff --git a/tools/testing/selftests/rseq/Makefile b/tools/testing/selftests/rseq/Makefile
new file mode 100644
index 000000000000..7f0153556b80
--- /dev/null
+++ b/tools/testing/selftests/rseq/Makefile
@@ -0,0 +1,13 @@
+CFLAGS += -O2 -Wall -g -I./ -I../cpu-opv/ -I../../../../usr/include/
+LDFLAGS += -lpthread
+
+TESTS = basic_test basic_percpu_ops_test param_test
+
+all: $(TESTS)
+%: %.c rseq.h rseq-*.h rseq.c ../cpu-opv/cpu-op.c ../cpu-opv/cpu-op.h
+ $(CC) $(CFLAGS) -o $@ $^ $(LDFLAGS)
+
+include ../lib.mk
+
+clean:
+ $(RM) $(TESTS)
diff --git a/tools/testing/selftests/rseq/basic_percpu_ops_test.c b/tools/testing/selftests/rseq/basic_percpu_ops_test.c
new file mode 100644
index 000000000000..5771470862bf
--- /dev/null
+++ b/tools/testing/selftests/rseq/basic_percpu_ops_test.c
@@ -0,0 +1,319 @@
+#define _GNU_SOURCE
+#include <assert.h>
+#include <pthread.h>
+#include <sched.h>
+#include <stdint.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+
+#include "rseq.h"
+#include "cpu-op.h"
+
+#define ARRAY_SIZE(arr) (sizeof(arr) / sizeof((arr)[0]))
+
+struct percpu_lock_entry {
+ intptr_t v;
+} __attribute__((aligned(128)));
+
+struct percpu_lock {
+ struct percpu_lock_entry c[CPU_SETSIZE];
+};
+
+struct test_data_entry {
+ intptr_t count;
+} __attribute__((aligned(128)));
+
+struct spinlock_test_data {
+ struct percpu_lock lock;
+ struct test_data_entry c[CPU_SETSIZE];
+ int reps;
+};
+
+struct percpu_list_node {
+ intptr_t data;
+ struct percpu_list_node *next;
+};
+
+struct percpu_list_entry {
+ struct percpu_list_node *head;
+} __attribute__((aligned(128)));
+
+struct percpu_list {
+ struct percpu_list_entry c[CPU_SETSIZE];
+};
+
+/* A simple percpu spinlock. Returns the cpu lock was acquired on. */
+int rseq_percpu_lock(struct percpu_lock *lock)
+{
+ int cpu;
+
+ for (;;) {
+ struct rseq_state rseq_state;
+ intptr_t expect = 0, n = 1;
+ int ret;
+
+ /* Try fast path. */
+ rseq_state = rseq_start();
+ cpu = rseq_cpu_at_start(rseq_state);
+ if (unlikely(lock->c[cpu].v != 0))
+ continue; /* Retry.*/
+ if (likely(rseq_finish(&lock->c[cpu].v, 1, rseq_state)))
+ break;
+ /* Fallback on cpu_opv system call. */
+ cpu = rseq_current_cpu_raw();
+ ret = cpu_op_cmpstore(&lock->c[cpu].v, &expect, &n,
+ sizeof(intptr_t), cpu);
+ if (likely(!ret))
+ break;
+ assert(ret >= 0 || errno == EAGAIN);
+ }
+ /*
+ * Acquire semantic when taking lock after control dependency.
+ * Matches smp_store_release().
+ */
+ smp_acquire__after_ctrl_dep();
+ return cpu;
+}
+
+void rseq_percpu_unlock(struct percpu_lock *lock, int cpu)
+{
+ assert(lock->c[cpu].v == 1);
+ /*
+ * Release lock, with release semantic. Matches
+ * smp_acquire__after_ctrl_dep().
+ */
+ smp_store_release(&lock->c[cpu].v, 0);
+}
+
+void *test_percpu_spinlock_thread(void *arg)
+{
+ struct spinlock_test_data *data = arg;
+ int i, cpu;
+
+ if (rseq_register_current_thread())
+ abort();
+ for (i = 0; i < data->reps; i++) {
+ cpu = rseq_percpu_lock(&data->lock);
+ data->c[cpu].count++;
+ rseq_percpu_unlock(&data->lock, cpu);
+ }
+ if (rseq_unregister_current_thread())
+ abort();
+
+ return NULL;
+}
+
+/*
+ * A simple test which implements a sharded counter using a per-cpu
+ * lock. Obviously real applications might prefer to simply use a
+ * per-cpu increment; however, this is reasonable for a test and the
+ * lock can be extended to synchronize more complicated operations.
+ */
+void test_percpu_spinlock(void)
+{
+ const int num_threads = 200;
+ int i;
+ uint64_t sum;
+ pthread_t test_threads[num_threads];
+ struct spinlock_test_data data;
+
+ memset(&data, 0, sizeof(data));
+ data.reps = 5000;
+
+ for (i = 0; i < num_threads; i++)
+ pthread_create(&test_threads[i], NULL,
+ test_percpu_spinlock_thread, &data);
+
+ for (i = 0; i < num_threads; i++)
+ pthread_join(test_threads[i], NULL);
+
+ sum = 0;
+ for (i = 0; i < CPU_SETSIZE; i++)
+ sum += data.c[i].count;
+
+ assert(sum == (uint64_t)data.reps * num_threads);
+}
+
+int percpu_list_push(struct percpu_list *list, struct percpu_list_node *node)
+{
+ struct rseq_state rseq_state;
+ intptr_t *targetptr, newval, expect;
+ int cpu;
+
+ /* Try fast path. */
+ rseq_state = rseq_start();
+ cpu = rseq_cpu_at_start(rseq_state);
+ newval = (intptr_t)node;
+ targetptr = (intptr_t *)&list->c[cpu].head;
+ node->next = list->c[cpu].head;
+ if (unlikely(!rseq_finish(targetptr, newval, rseq_state))) {
+ /* Fallback on cpu_opv system call. */
+ for (;;) {
+ int ret;
+
+ cpu = rseq_current_cpu_raw();
+ /* Load list->c[cpu].head with single-copy atomicity. */
+ expect = (intptr_t)READ_ONCE(list->c[cpu].head);
+ newval = (intptr_t)node;
+ targetptr = (intptr_t *)&list->c[cpu].head;
+ node->next = (struct percpu_list_node *)expect;
+ ret = cpu_op_cmpstore(targetptr, &expect, &newval,
+ sizeof(intptr_t), cpu);
+ if (likely(!ret))
+ break;
+ assert(ret >= 0 || errno == EAGAIN);
+ }
+ }
+ return cpu;
+}
+
+/*
+ * Unlike a traditional lock-less linked list; the availability of a
+ * rseq primitive allows us to implement pop without concerns over
+ * ABA-type races.
+ */
+struct percpu_list_node *percpu_list_pop(struct percpu_list *list)
+{
+ struct percpu_list_node *head, *next;
+ struct rseq_state rseq_state;
+ intptr_t *targetptr, newval, expect;
+ int cpu;
+
+ /* Try fast path. */
+ rseq_state = rseq_start();
+ cpu = rseq_cpu_at_start(rseq_state);
+ /* Load head with single-copy atomicity. */
+ head = READ_ONCE(list->c[cpu].head);
+ if (!head)
+ return NULL;
+ /* Load head->next with single-copy atomicity. */
+ next = READ_ONCE(head->next);
+ newval = (intptr_t)next;
+ targetptr = (intptr_t *)&list->c[cpu].head;
+ if (unlikely(!rseq_finish(targetptr, newval, rseq_state))) {
+ /* Fallback on cpu_opv system call. */
+ for (;;) {
+ int ret;
+
+ cpu = rseq_current_cpu_raw();
+ /* Load head with single-copy atomicity. */
+ head = READ_ONCE(list->c[cpu].head);
+ if (!head)
+ return NULL;
+ expect = (intptr_t)head;
+ /* Load head->next with single-copy atomicity. */
+ next = READ_ONCE(head->next);
+ newval = (intptr_t)next;
+ targetptr = (intptr_t *)&list->c[cpu].head;
+ ret = cpu_op_2cmp1store(targetptr, &expect, &newval,
+ &head->next, &next,
+ sizeof(intptr_t), cpu);
+ if (likely(!ret))
+ break;
+ assert(ret >= 0 || errno == EAGAIN);
+ }
+ }
+
+ return head;
+}
+
+void *test_percpu_list_thread(void *arg)
+{
+ int i;
+ struct percpu_list *list = (struct percpu_list *)arg;
+
+ if (rseq_register_current_thread())
+ abort();
+
+ for (i = 0; i < 100000; i++) {
+ struct percpu_list_node *node = percpu_list_pop(list);
+
+ sched_yield(); /* encourage shuffling */
+ if (node)
+ percpu_list_push(list, node);
+ }
+
+ if (rseq_unregister_current_thread())
+ abort();
+
+ return NULL;
+}
+
+/* Simultaneous modification to a per-cpu linked list from many threads. */
+void test_percpu_list(void)
+{
+ int i, j;
+ uint64_t sum = 0, expected_sum = 0;
+ struct percpu_list list;
+ pthread_t test_threads[200];
+ cpu_set_t allowed_cpus;
+
+ memset(&list, 0, sizeof(list));
+
+ /* Generate list entries for every usable cpu. */
+ sched_getaffinity(0, sizeof(allowed_cpus), &allowed_cpus);
+ for (i = 0; i < CPU_SETSIZE; i++) {
+ if (!CPU_ISSET(i, &allowed_cpus))
+ continue;
+ for (j = 1; j <= 100; j++) {
+ struct percpu_list_node *node;
+
+ expected_sum += j;
+
+ node = malloc(sizeof(*node));
+ assert(node);
+ node->data = j;
+ node->next = list.c[i].head;
+ list.c[i].head = node;
+ }
+ }
+
+ for (i = 0; i < 200; i++)
+ assert(pthread_create(&test_threads[i], NULL,
+ test_percpu_list_thread, &list) == 0);
+
+ for (i = 0; i < 200; i++)
+ pthread_join(test_threads[i], NULL);
+
+ for (i = 0; i < CPU_SETSIZE; i++) {
+ cpu_set_t pin_mask;
+ struct percpu_list_node *node;
+
+ if (!CPU_ISSET(i, &allowed_cpus))
+ continue;
+
+ CPU_ZERO(&pin_mask);
+ CPU_SET(i, &pin_mask);
+ sched_setaffinity(0, sizeof(pin_mask), &pin_mask);
+
+ while ((node = percpu_list_pop(&list))) {
+ sum += node->data;
+ free(node);
+ }
+ }
+
+ /*
+ * All entries should now be accounted for (unless some external
+ * actor is interfering with our allowed affinity while this
+ * test is running).
+ */
+ assert(sum == expected_sum);
+}
+
+int main(int argc, char **argv)
+{
+ if (rseq_register_current_thread())
+ goto error;
+ printf("spinlock\n");
+ test_percpu_spinlock();
+ printf("percpu_list\n");
+ test_percpu_list();
+ if (rseq_unregister_current_thread())
+ goto error;
+ return 0;
+
+error:
+ return -1;
+}
+
diff --git a/tools/testing/selftests/rseq/basic_test.c b/tools/testing/selftests/rseq/basic_test.c
new file mode 100644
index 000000000000..236bbe2610af
--- /dev/null
+++ b/tools/testing/selftests/rseq/basic_test.c
@@ -0,0 +1,97 @@
+/*
+ * Basic test coverage for critical regions and rseq_current_cpu().
+ */
+
+#define _GNU_SOURCE
+#include <assert.h>
+#include <sched.h>
+#include <signal.h>
+#include <stdio.h>
+#include <string.h>
+#include <sys/time.h>
+
+#include "rseq.h"
+
+volatile int signals_delivered;
+volatile __thread struct rseq_state sigtest_start;
+
+void test_cpu_pointer(void)
+{
+ cpu_set_t affinity, test_affinity;
+ int i;
+
+ sched_getaffinity(0, sizeof(affinity), &affinity);
+ CPU_ZERO(&test_affinity);
+ for (i = 0; i < CPU_SETSIZE; i++) {
+ if (CPU_ISSET(i, &affinity)) {
+ CPU_SET(i, &test_affinity);
+ sched_setaffinity(0, sizeof(test_affinity),
+ &test_affinity);
+ assert(rseq_current_cpu() == sched_getcpu());
+ assert(rseq_current_cpu() == i);
+ CPU_CLR(i, &test_affinity);
+ }
+ }
+ sched_setaffinity(0, sizeof(affinity), &affinity);
+}
+
+/*
+ * This depends solely on some environmental event triggering a counter
+ * increase.
+ */
+void test_critical_section(void)
+{
+ struct rseq_state start;
+ uint32_t event_counter;
+
+ start = rseq_start();
+ event_counter = start.event_counter;
+ do {
+ start = rseq_start();
+ } while (start.event_counter == event_counter);
+}
+
+void test_signal_interrupt_handler(int signo)
+{
+ struct rseq_state current;
+
+ current = rseq_start();
+ /*
+ * The potential critical section bordered by 'start' must be
+ * invalid.
+ */
+ assert(current.event_counter != sigtest_start.event_counter);
+ signals_delivered++;
+}
+
+void test_signal_interrupts(void)
+{
+ struct itimerval it = { { 0, 1 }, { 0, 1 } };
+ struct itimerval stop_it = { { 0, 0 }, { 0, 0 } };
+
+ setitimer(ITIMER_PROF, &it, NULL);
+ signal(SIGPROF, test_signal_interrupt_handler);
+
+ do {
+ sigtest_start = rseq_start();
+ } while (signals_delivered < 10);
+ setitimer(ITIMER_PROF, &stop_it, NULL);
+}
+
+int main(int argc, char **argv)
+{
+ if (rseq_register_current_thread())
+ goto init_thread_error;
+ printf("testing current cpu\n");
+ test_cpu_pointer();
+ printf("testing critical section\n");
+ test_critical_section();
+ printf("testing critical section is interrupted by signal\n");
+ test_signal_interrupts();
+ if (rseq_unregister_current_thread())
+ goto init_thread_error;
+ return 0;
+
+init_thread_error:
+ return -1;
+}
diff --git a/tools/testing/selftests/rseq/param_test.c b/tools/testing/selftests/rseq/param_test.c
new file mode 100644
index 000000000000..a68fa0886d50
--- /dev/null
+++ b/tools/testing/selftests/rseq/param_test.c
@@ -0,0 +1,1246 @@
+#define _GNU_SOURCE
+#include <assert.h>
+#include <pthread.h>
+#include <sched.h>
+#include <stdint.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include <syscall.h>
+#include <unistd.h>
+#include <poll.h>
+#include <sys/types.h>
+#include <signal.h>
+#include <errno.h>
+
+#include "cpu-op.h"
+
+static inline pid_t gettid(void)
+{
+ return syscall(__NR_gettid);
+}
+
+#define NR_INJECT 9
+static int loop_cnt[NR_INJECT + 1];
+
+static int opt_modulo;
+
+static int opt_yield, opt_signal, opt_sleep,
+ opt_disable_rseq, opt_threads = 200,
+ opt_reps = 5000, opt_disable_mod = 0, opt_test = 's';
+
+static __thread unsigned int signals_delivered;
+
+#ifndef BENCHMARK
+
+static __thread unsigned int yield_mod_cnt, nr_retry;
+
+#define printf_nobench(fmt, ...) printf(fmt, ## __VA_ARGS__)
+
+#define RSEQ_INJECT_INPUT \
+ , [loop_cnt_1]"m"(loop_cnt[1]) \
+ , [loop_cnt_2]"m"(loop_cnt[2]) \
+ , [loop_cnt_3]"m"(loop_cnt[3]) \
+ , [loop_cnt_4]"m"(loop_cnt[4]) \
+ , [loop_cnt_5]"m"(loop_cnt[5])
+
+#if defined(__x86_64__) || defined(__i386__)
+
+#define INJECT_ASM_REG "eax"
+
+#define RSEQ_INJECT_CLOBBER \
+ , INJECT_ASM_REG
+
+#define RSEQ_INJECT_ASM(n) \
+ "mov %[loop_cnt_" #n "], %%" INJECT_ASM_REG "\n\t" \
+ "test %%" INJECT_ASM_REG ",%%" INJECT_ASM_REG "\n\t" \
+ "jz 333f\n\t" \
+ "222:\n\t" \
+ "dec %%" INJECT_ASM_REG "\n\t" \
+ "jnz 222b\n\t" \
+ "333:\n\t"
+
+#elif defined(__ARMEL__)
+
+#define INJECT_ASM_REG "r4"
+
+#define RSEQ_INJECT_CLOBBER \
+ , INJECT_ASM_REG
+
+#define RSEQ_INJECT_ASM(n) \
+ "ldr " INJECT_ASM_REG ", %[loop_cnt_" #n "]\n\t" \
+ "cmp " INJECT_ASM_REG ", #0\n\t" \
+ "beq 333f\n\t" \
+ "222:\n\t" \
+ "subs " INJECT_ASM_REG ", #1\n\t" \
+ "bne 222b\n\t" \
+ "333:\n\t"
+
+#elif __PPC__
+#define INJECT_ASM_REG "r18"
+
+#define RSEQ_INJECT_CLOBBER \
+ , INJECT_ASM_REG
+
+#define RSEQ_INJECT_ASM(n) \
+ "lwz %%" INJECT_ASM_REG ", %[loop_cnt_" #n "]\n\t" \
+ "cmpwi %%" INJECT_ASM_REG ", 0\n\t" \
+ "beq 333f\n\t" \
+ "222:\n\t" \
+ "subic. %%" INJECT_ASM_REG ", %%" INJECT_ASM_REG ", 1\n\t" \
+ "bne 222b\n\t" \
+ "333:\n\t"
+#else
+#error unsupported target
+#endif
+
+#define RSEQ_INJECT_FAILED \
+ nr_retry++;
+
+#define RSEQ_INJECT_C(n) \
+{ \
+ int loc_i, loc_nr_loops = loop_cnt[n]; \
+ \
+ for (loc_i = 0; loc_i < loc_nr_loops; loc_i++) { \
+ barrier(); \
+ } \
+ if (loc_nr_loops == -1 && opt_modulo) { \
+ if (yield_mod_cnt == opt_modulo - 1) { \
+ if (opt_sleep > 0) \
+ poll(NULL, 0, opt_sleep); \
+ if (opt_yield) \
+ sched_yield(); \
+ if (opt_signal) \
+ raise(SIGUSR1); \
+ yield_mod_cnt = 0; \
+ } else { \
+ yield_mod_cnt++; \
+ } \
+ } \
+}
+
+#else
+
+#define printf_nobench(fmt, ...)
+
+#endif /* BENCHMARK */
+
+#include "rseq.h"
+
+struct percpu_lock_entry {
+ intptr_t v;
+} __attribute__((aligned(128)));
+
+struct percpu_lock {
+ struct percpu_lock_entry c[CPU_SETSIZE];
+};
+
+struct test_data_entry {
+ intptr_t count;
+} __attribute__((aligned(128)));
+
+struct spinlock_test_data {
+ struct percpu_lock lock;
+ struct test_data_entry c[CPU_SETSIZE];
+};
+
+struct spinlock_thread_test_data {
+ struct spinlock_test_data *data;
+ int reps;
+ int reg;
+};
+
+struct inc_test_data {
+ struct test_data_entry c[CPU_SETSIZE];
+};
+
+struct inc_thread_test_data {
+ struct inc_test_data *data;
+ int reps;
+ int reg;
+};
+
+struct percpu_list_node {
+ intptr_t data;
+ struct percpu_list_node *next;
+};
+
+struct percpu_list_entry {
+ struct percpu_list_node *head;
+} __attribute__((aligned(128)));
+
+struct percpu_list {
+ struct percpu_list_entry c[CPU_SETSIZE];
+};
+
+#define BUFFER_ITEM_PER_CPU 100
+
+struct percpu_buffer_node {
+ intptr_t data;
+};
+
+struct percpu_buffer_entry {
+ intptr_t offset;
+ intptr_t buflen;
+ struct percpu_buffer_node **array;
+} __attribute__((aligned(128)));
+
+struct percpu_buffer {
+ struct percpu_buffer_entry c[CPU_SETSIZE];
+};
+
+#define MEMCPY_BUFFER_ITEM_PER_CPU 100
+
+struct percpu_memcpy_buffer_node {
+ intptr_t data1;
+ uint64_t data2;
+};
+
+struct percpu_memcpy_buffer_entry {
+ intptr_t offset;
+ intptr_t buflen;
+ struct percpu_memcpy_buffer_node *array;
+} __attribute__((aligned(128)));
+
+struct percpu_memcpy_buffer {
+ struct percpu_memcpy_buffer_entry c[CPU_SETSIZE];
+};
+
+/* A simple percpu spinlock. Returns the cpu lock was acquired on. */
+static int rseq_percpu_lock(struct percpu_lock *lock)
+{
+ int cpu;
+
+ for (;;) {
+#ifndef SKIP_FASTPATH
+ struct rseq_state rseq_state;
+
+ /* Try fast path. */
+ rseq_state = rseq_start();
+ cpu = rseq_cpu_at_start(rseq_state);
+ if (unlikely(lock->c[cpu].v != 0))
+ continue; /* Retry.*/
+ if (likely(rseq_finish(&lock->c[cpu].v, 1, rseq_state)))
+ break;
+ else
+#endif
+ {
+ /* Fallback on cpu_opv system call. */
+ intptr_t expect = 0, n = 1;
+ int ret;
+
+ cpu = rseq_current_cpu_raw();
+ ret = cpu_op_cmpstore(&lock->c[cpu].v, &expect, &n,
+ sizeof(intptr_t), cpu);
+ if (likely(!ret))
+ break;
+ assert(ret >= 0 || errno == EAGAIN);
+ }
+ }
+ /*
+ * Acquire semantic when taking lock after control dependency.
+ * Matches smp_store_release().
+ */
+ smp_acquire__after_ctrl_dep();
+ return cpu;
+}
+
+static void rseq_percpu_unlock(struct percpu_lock *lock, int cpu)
+{
+ assert(lock->c[cpu].v == 1);
+ /*
+ * Release lock, with release semantic. Matches
+ * smp_acquire__after_ctrl_dep().
+ */
+ smp_store_release(&lock->c[cpu].v, 0);
+}
+
+void *test_percpu_spinlock_thread(void *arg)
+{
+ struct spinlock_thread_test_data *thread_data = arg;
+ struct spinlock_test_data *data = thread_data->data;
+ int i, cpu;
+
+ if (!opt_disable_rseq && thread_data->reg
+ && rseq_register_current_thread())
+ abort();
+ for (i = 0; i < thread_data->reps; i++) {
+ cpu = rseq_percpu_lock(&data->lock);
+ data->c[cpu].count++;
+ rseq_percpu_unlock(&data->lock, cpu);
+#ifndef BENCHMARK
+ if (i != 0 && !(i % (thread_data->reps / 10)))
+ printf("tid %d: count %d\n", (int) gettid(), i);
+#endif
+ }
+ printf_nobench("tid %d: number of retry: %d, signals delivered: %u\n",
+ (int) gettid(), nr_retry, signals_delivered);
+ if (rseq_unregister_current_thread())
+ abort();
+ return NULL;
+}
+
+/*
+ * A simple test which implements a sharded counter using a per-cpu
+ * lock. Obviously real applications might prefer to simply use a
+ * per-cpu increment; however, this is reasonable for a test and the
+ * lock can be extended to synchronize more complicated operations.
+ */
+void test_percpu_spinlock(void)
+{
+ const int num_threads = opt_threads;
+ int i, ret;
+ uint64_t sum;
+ pthread_t test_threads[num_threads];
+ struct spinlock_test_data data;
+ struct spinlock_thread_test_data thread_data[num_threads];
+
+ memset(&data, 0, sizeof(data));
+ for (i = 0; i < num_threads; i++) {
+ thread_data[i].reps = opt_reps;
+ if (opt_disable_mod <= 0 || (i % opt_disable_mod))
+ thread_data[i].reg = 1;
+ else
+ thread_data[i].reg = 0;
+ thread_data[i].data = &data;
+ ret = pthread_create(&test_threads[i], NULL,
+ test_percpu_spinlock_thread, &thread_data[i]);
+ if (ret) {
+ errno = ret;
+ perror("pthread_create");
+ abort();
+ }
+ }
+
+ for (i = 0; i < num_threads; i++) {
+ pthread_join(test_threads[i], NULL);
+ if (ret) {
+ errno = ret;
+ perror("pthread_join");
+ abort();
+ }
+ }
+
+ sum = 0;
+ for (i = 0; i < CPU_SETSIZE; i++)
+ sum += data.c[i].count;
+
+ assert(sum == (uint64_t)opt_reps * num_threads);
+}
+
+void *test_percpu_inc_thread(void *arg)
+{
+ struct inc_thread_test_data *thread_data = arg;
+ struct inc_test_data *data = thread_data->data;
+ int i;
+
+ if (!opt_disable_rseq && thread_data->reg
+ && rseq_register_current_thread())
+ abort();
+ for (i = 0; i < thread_data->reps; i++) {
+ int cpu;
+
+#ifndef SKIP_FASTPATH
+ struct rseq_state rseq_state;
+ intptr_t *targetptr, newval;
+
+ /* Try fast path. */
+ rseq_state = rseq_start();
+ cpu = rseq_cpu_at_start(rseq_state);
+ newval = (intptr_t)data->c[cpu].count + 1;
+ targetptr = (intptr_t *)&data->c[cpu].count;
+ if (unlikely(!rseq_finish(targetptr, newval, rseq_state)))
+#endif
+ {
+ for (;;) {
+ /* Fallback on cpu_opv system call. */
+ int ret;
+
+ cpu = rseq_current_cpu_raw();
+ ret = cpu_op_add(&data->c[cpu].count, 1,
+ sizeof(intptr_t), cpu);
+ if (likely(!ret))
+ break;
+ assert(ret >= 0 || errno == EAGAIN);
+ }
+ }
+
+#ifndef BENCHMARK
+ if (i != 0 && !(i % (thread_data->reps / 10)))
+ printf("tid %d: count %d\n", (int) gettid(), i);
+#endif
+ }
+ printf_nobench("tid %d: number of retry: %d, signals delivered: %u\n",
+ (int) gettid(), nr_retry, signals_delivered);
+ if (rseq_unregister_current_thread())
+ abort();
+ return NULL;
+}
+
+void test_percpu_inc(void)
+{
+ const int num_threads = opt_threads;
+ int i, ret;
+ uint64_t sum;
+ pthread_t test_threads[num_threads];
+ struct inc_test_data data;
+ struct inc_thread_test_data thread_data[num_threads];
+
+ memset(&data, 0, sizeof(data));
+ for (i = 0; i < num_threads; i++) {
+ thread_data[i].reps = opt_reps;
+ if (opt_disable_mod <= 0 || (i % opt_disable_mod))
+ thread_data[i].reg = 1;
+ else
+ thread_data[i].reg = 0;
+ thread_data[i].data = &data;
+ ret = pthread_create(&test_threads[i], NULL,
+ test_percpu_inc_thread, &thread_data[i]);
+ if (ret) {
+ errno = ret;
+ perror("pthread_create");
+ abort();
+ }
+ }
+
+ for (i = 0; i < num_threads; i++) {
+ pthread_join(test_threads[i], NULL);
+ if (ret) {
+ errno = ret;
+ perror("pthread_join");
+ abort();
+ }
+ }
+
+ sum = 0;
+ for (i = 0; i < CPU_SETSIZE; i++)
+ sum += data.c[i].count;
+
+ assert(sum == (uint64_t)opt_reps * num_threads);
+}
+
+int percpu_list_push(struct percpu_list *list, struct percpu_list_node *node)
+{
+ intptr_t *targetptr, newval, expect;
+ int cpu;
+#ifndef SKIP_FASTPATH
+ struct rseq_state rseq_state;
+
+ /* Try fast path. */
+ rseq_state = rseq_start();
+ cpu = rseq_cpu_at_start(rseq_state);
+ newval = (intptr_t)node;
+ targetptr = (intptr_t *)&list->c[cpu].head;
+ node->next = list->c[cpu].head;
+ if (unlikely(!rseq_finish(targetptr, newval, rseq_state)))
+#endif
+ {
+ /* Fallback on cpu_opv system call. */
+ for (;;) {
+ int ret;
+
+ cpu = rseq_current_cpu_raw();
+ /* Load list->c[cpu].head with single-copy atomicity. */
+ expect = (intptr_t)READ_ONCE(list->c[cpu].head);
+ newval = (intptr_t)node;
+ targetptr = (intptr_t *)&list->c[cpu].head;
+ node->next = (struct percpu_list_node *)expect;
+ ret = cpu_op_cmpstore(targetptr, &expect, &newval,
+ sizeof(intptr_t), cpu);
+ if (likely(!ret))
+ break;
+ assert(ret >= 0 || errno == EAGAIN);
+ }
+ }
+ return cpu;
+}
+
+/*
+ * Unlike a traditional lock-less linked list; the availability of a
+ * rseq primitive allows us to implement pop without concerns over
+ * ABA-type races.
+ */
+struct percpu_list_node *percpu_list_pop(struct percpu_list *list)
+{
+ struct percpu_list_node *head, *next;
+ intptr_t *targetptr, newval, expect;
+ int cpu;
+#ifndef SKIP_FASTPATH
+ struct rseq_state rseq_state;
+
+ /* Try fast path. */
+ rseq_state = rseq_start();
+ cpu = rseq_cpu_at_start(rseq_state);
+ /* Load list->c[cpu].head with single-copy atomicity. */
+ head = READ_ONCE(list->c[cpu].head);
+ if (!head)
+ return NULL;
+ /* Load head->next with single-copy atomicity. */
+ next = READ_ONCE(head->next);
+ newval = (intptr_t)next;
+ targetptr = (intptr_t *)&list->c[cpu].head;
+ if (unlikely(!rseq_finish(targetptr, newval, rseq_state)))
+#endif
+ {
+ /* Fallback on cpu_opv system call. */
+ for (;;) {
+ int ret;
+
+ cpu = rseq_current_cpu_raw();
+ /* Load list->c[cpu].head with single-copy atomicity. */
+ head = READ_ONCE(list->c[cpu].head);
+ if (!head)
+ return NULL;
+ expect = (intptr_t)head;
+ /* Load head->next with single-copy atomicity. */
+ next = READ_ONCE(head->next);
+ newval = (intptr_t)next;
+ targetptr = (intptr_t *)&list->c[cpu].head;
+ ret = cpu_op_2cmp1store(targetptr, &expect, &newval,
+ &head->next, &next,
+ sizeof(intptr_t), cpu);
+ if (likely(!ret))
+ break;
+ assert(ret >= 0 || errno == EAGAIN);
+ }
+ }
+
+ return head;
+}
+
+void *test_percpu_list_thread(void *arg)
+{
+ int i;
+ struct percpu_list *list = (struct percpu_list *)arg;
+
+ if (rseq_register_current_thread())
+ abort();
+
+ for (i = 0; i < opt_reps; i++) {
+ struct percpu_list_node *node = percpu_list_pop(list);
+
+ if (opt_yield)
+ sched_yield(); /* encourage shuffling */
+ if (node)
+ percpu_list_push(list, node);
+ }
+
+ if (rseq_unregister_current_thread())
+ abort();
+
+ return NULL;
+}
+
+/* Simultaneous modification to a per-cpu linked list from many threads. */
+void test_percpu_list(void)
+{
+ const int num_threads = opt_threads;
+ int i, j, ret;
+ uint64_t sum = 0, expected_sum = 0;
+ struct percpu_list list;
+ pthread_t test_threads[num_threads];
+ cpu_set_t allowed_cpus;
+
+ memset(&list, 0, sizeof(list));
+
+ /* Generate list entries for every usable cpu. */
+ sched_getaffinity(0, sizeof(allowed_cpus), &allowed_cpus);
+ for (i = 0; i < CPU_SETSIZE; i++) {
+ if (!CPU_ISSET(i, &allowed_cpus))
+ continue;
+ for (j = 1; j <= 100; j++) {
+ struct percpu_list_node *node;
+
+ expected_sum += j;
+
+ node = malloc(sizeof(*node));
+ assert(node);
+ node->data = j;
+ node->next = list.c[i].head;
+ list.c[i].head = node;
+ }
+ }
+
+ for (i = 0; i < num_threads; i++) {
+ ret = pthread_create(&test_threads[i], NULL,
+ test_percpu_list_thread, &list);
+ if (ret) {
+ errno = ret;
+ perror("pthread_create");
+ abort();
+ }
+ }
+
+ for (i = 0; i < num_threads; i++) {
+ pthread_join(test_threads[i], NULL);
+ if (ret) {
+ errno = ret;
+ perror("pthread_join");
+ abort();
+ }
+ }
+
+ for (i = 0; i < CPU_SETSIZE; i++) {
+ cpu_set_t pin_mask;
+ struct percpu_list_node *node;
+
+ if (!CPU_ISSET(i, &allowed_cpus))
+ continue;
+
+ CPU_ZERO(&pin_mask);
+ CPU_SET(i, &pin_mask);
+ sched_setaffinity(0, sizeof(pin_mask), &pin_mask);
+
+ while ((node = percpu_list_pop(&list))) {
+ sum += node->data;
+ free(node);
+ }
+ }
+
+ /*
+ * All entries should now be accounted for (unless some external
+ * actor is interfering with our allowed affinity while this
+ * test is running).
+ */
+ assert(sum == expected_sum);
+}
+
+bool percpu_buffer_push(struct percpu_buffer *buffer,
+ struct percpu_buffer_node *node)
+{
+ intptr_t *targetptr_spec, newval_spec;
+ intptr_t *targetptr_final, newval_final;
+ int cpu;
+ intptr_t offset;
+#ifndef SKIP_FASTPATH
+ struct rseq_state rseq_state;
+
+ /* Try fast path. */
+ rseq_state = rseq_start();
+ cpu = rseq_cpu_at_start(rseq_state);
+ /* Load offset with single-copy atomicity. */
+ offset = READ_ONCE(buffer->c[cpu].offset);
+ if (offset == buffer->c[cpu].buflen)
+ return false;
+ newval_spec = (intptr_t)node;
+ targetptr_spec = (intptr_t *)&buffer->c[cpu].array[offset];
+ newval_final = offset + 1;
+ targetptr_final = &buffer->c[cpu].offset;
+ if (unlikely(!rseq_finish2(targetptr_spec, newval_spec,
+ targetptr_final, newval_final, rseq_state)))
+#endif
+ {
+ /* Fallback on cpu_opv system call. */
+ for (;;) {
+ int ret;
+
+ cpu = rseq_current_cpu_raw();
+ /* Load offset with single-copy atomicity. */
+ offset = READ_ONCE(buffer->c[cpu].offset);
+ if (offset == buffer->c[cpu].buflen)
+ return false;
+ newval_spec = (intptr_t)node;
+ targetptr_spec = (intptr_t *)&buffer->c[cpu].array[offset];
+ newval_final = offset + 1;
+ targetptr_final = &buffer->c[cpu].offset;
+ ret = cpu_op_1cmp2store(targetptr_final, &offset, &newval_final,
+ targetptr_spec, &newval_spec,
+ sizeof(intptr_t), cpu);
+ if (likely(!ret))
+ break;
+ assert(ret >= 0 || errno == EAGAIN);
+ }
+ }
+ return true;
+}
+
+struct percpu_buffer_node *percpu_buffer_pop(struct percpu_buffer *buffer)
+{
+ struct percpu_buffer_node *head;
+ intptr_t *targetptr, newval;
+ int cpu;
+ intptr_t offset;
+#ifndef SKIP_FASTPATH
+ struct rseq_state rseq_state;
+
+ /* Try fast path. */
+ rseq_state = rseq_start();
+ cpu = rseq_cpu_at_start(rseq_state);
+ /* Load offset with single-copy atomicity. */
+ offset = READ_ONCE(buffer->c[cpu].offset);
+ if (offset == 0)
+ return NULL;
+ head = buffer->c[cpu].array[offset - 1];
+ newval = offset - 1;
+ targetptr = (intptr_t *)&buffer->c[cpu].offset;
+ if (unlikely(!rseq_finish(targetptr, newval, rseq_state)))
+#endif
+ {
+ /* Fallback on cpu_opv system call. */
+ for (;;) {
+ int ret;
+
+ cpu = rseq_current_cpu_raw();
+ /* Load offset with single-copy atomicity. */
+ offset = READ_ONCE(buffer->c[cpu].offset);
+ if (offset == 0)
+ return NULL;
+ head = buffer->c[cpu].array[offset - 1];
+ newval = offset - 1;
+ targetptr = (intptr_t *)&buffer->c[cpu].offset;
+ ret = cpu_op_2cmp1store(targetptr, &offset, &newval,
+ &buffer->c[cpu].array[offset - 1], &head,
+ sizeof(intptr_t), cpu);
+ if (likely(!ret))
+ break;
+ assert(ret >= 0 || errno == EAGAIN);
+ }
+ }
+ return head;
+}
+
+void *test_percpu_buffer_thread(void *arg)
+{
+ int i;
+ struct percpu_buffer *buffer = (struct percpu_buffer *)arg;
+
+ if (rseq_register_current_thread())
+ abort();
+
+ for (i = 0; i < opt_reps; i++) {
+ struct percpu_buffer_node *node = percpu_buffer_pop(buffer);
+
+ if (opt_yield)
+ sched_yield(); /* encourage shuffling */
+ if (node) {
+ if (!percpu_buffer_push(buffer, node)) {
+ /* Should increase buffer size. */
+ abort();
+ }
+ }
+ }
+
+ if (rseq_unregister_current_thread())
+ abort();
+
+ return NULL;
+}
+
+/* Simultaneous modification to a per-cpu buffer from many threads. */
+void test_percpu_buffer(void)
+{
+ const int num_threads = opt_threads;
+ int i, j, ret;
+ uint64_t sum = 0, expected_sum = 0;
+ struct percpu_buffer buffer;
+ pthread_t test_threads[num_threads];
+ cpu_set_t allowed_cpus;
+
+ memset(&buffer, 0, sizeof(buffer));
+
+ /* Generate list entries for every usable cpu. */
+ sched_getaffinity(0, sizeof(allowed_cpus), &allowed_cpus);
+ for (i = 0; i < CPU_SETSIZE; i++) {
+ if (!CPU_ISSET(i, &allowed_cpus))
+ continue;
+ /* Worse-case is every item in same CPU. */
+ buffer.c[i].array =
+ malloc(sizeof(*buffer.c[i].array) * CPU_SETSIZE
+ * BUFFER_ITEM_PER_CPU);
+ assert(buffer.c[i].array);
+ buffer.c[i].buflen = CPU_SETSIZE * BUFFER_ITEM_PER_CPU;
+ for (j = 1; j <= BUFFER_ITEM_PER_CPU; j++) {
+ struct percpu_buffer_node *node;
+
+ expected_sum += j;
+
+ /*
+ * We could theoretically put the word-sized
+ * "data" directly in the buffer. However, we
+ * want to model objects that would not fit
+ * within a single word, so allocate an object
+ * for each node.
+ */
+ node = malloc(sizeof(*node));
+ assert(node);
+ node->data = j;
+ buffer.c[i].array[j - 1] = node;
+ buffer.c[i].offset++;
+ }
+ }
+
+ for (i = 0; i < num_threads; i++) {
+ ret = pthread_create(&test_threads[i], NULL,
+ test_percpu_buffer_thread, &buffer);
+ if (ret) {
+ errno = ret;
+ perror("pthread_create");
+ abort();
+ }
+ }
+
+ for (i = 0; i < num_threads; i++) {
+ pthread_join(test_threads[i], NULL);
+ if (ret) {
+ errno = ret;
+ perror("pthread_join");
+ abort();
+ }
+ }
+
+ for (i = 0; i < CPU_SETSIZE; i++) {
+ cpu_set_t pin_mask;
+ struct percpu_buffer_node *node;
+
+ if (!CPU_ISSET(i, &allowed_cpus))
+ continue;
+
+ CPU_ZERO(&pin_mask);
+ CPU_SET(i, &pin_mask);
+ sched_setaffinity(0, sizeof(pin_mask), &pin_mask);
+
+ while ((node = percpu_buffer_pop(&buffer))) {
+ sum += node->data;
+ free(node);
+ }
+ free(buffer.c[i].array);
+ }
+
+ /*
+ * All entries should now be accounted for (unless some external
+ * actor is interfering with our allowed affinity while this
+ * test is running).
+ */
+ assert(sum == expected_sum);
+}
+
+bool percpu_memcpy_buffer_push(struct percpu_memcpy_buffer *buffer,
+ struct percpu_memcpy_buffer_node item)
+{
+ char *destptr, *srcptr;
+ size_t copylen;
+ intptr_t *targetptr_final, newval_final;
+ int cpu;
+ intptr_t offset;
+#ifndef SKIP_FASTPATH
+ struct rseq_state rseq_state;
+
+ /* Try fast path. */
+ rseq_state = rseq_start();
+ cpu = rseq_cpu_at_start(rseq_state);
+ /* Load offset with single-copy atomicity. */
+ offset = READ_ONCE(buffer->c[cpu].offset);
+ if (offset == buffer->c[cpu].buflen)
+ return false;
+ destptr = (char *)&buffer->c[cpu].array[offset];
+ srcptr = (char *)&item;
+ copylen = sizeof(item);
+ newval_final = offset + 1;
+ targetptr_final = &buffer->c[cpu].offset;
+ if (unlikely(!rseq_finish_memcpy(destptr, srcptr, copylen,
+ targetptr_final, newval_final, rseq_state)))
+#endif
+ {
+ /* Fallback on cpu_opv system call. */
+ for (;;) {
+ int ret;
+
+ cpu = rseq_current_cpu_raw();
+ /* Load offset with single-copy atomicity. */
+ offset = READ_ONCE(buffer->c[cpu].offset);
+ if (offset == buffer->c[cpu].buflen)
+ return false;
+ destptr = (char *)&buffer->c[cpu].array[offset];
+ srcptr = (char *)&item;
+ copylen = sizeof(item);
+ newval_final = offset + 1;
+ targetptr_final = &buffer->c[cpu].offset;
+ /* copylen must be <= PAGE_SIZE. */
+ ret = cpu_op_cmpstorememcpy(targetptr_final, &offset, &newval_final,
+ sizeof(intptr_t), destptr, srcptr, copylen, cpu);
+ if (likely(!ret))
+ break;
+ assert(ret >= 0 || errno == EAGAIN);
+ }
+ }
+ return true;
+}
+
+bool percpu_memcpy_buffer_pop(struct percpu_memcpy_buffer *buffer,
+ struct percpu_memcpy_buffer_node *item)
+{
+ char *destptr, *srcptr;
+ size_t copylen;
+ intptr_t *targetptr_final, newval_final;
+ int cpu;
+ intptr_t offset;
+#ifndef SKIP_FASTPATH
+ struct rseq_state rseq_state;
+
+ /* Try fast path. */
+ rseq_state = rseq_start();
+ cpu = rseq_cpu_at_start(rseq_state);
+ /* Load offset with single-copy atomicity. */
+ offset = READ_ONCE(buffer->c[cpu].offset);
+ if (offset == 0)
+ return false;
+ destptr = (char *)item;
+ srcptr = (char *)&buffer->c[cpu].array[offset - 1];
+ copylen = sizeof(*item);
+ newval_final = offset - 1;
+ targetptr_final = &buffer->c[cpu].offset;
+ if (unlikely(!rseq_finish_memcpy(destptr, srcptr, copylen,
+ targetptr_final, newval_final, rseq_state)))
+#endif
+ {
+ /* Fallback on cpu_opv system call. */
+ for (;;) {
+ int ret;
+
+ cpu = rseq_current_cpu_raw();
+ /* Load offset with single-copy atomicity. */
+ offset = READ_ONCE(buffer->c[cpu].offset);
+ if (offset == 0)
+ return false;
+ destptr = (char *)item;
+ srcptr = (char *)&buffer->c[cpu].array[offset - 1];
+ copylen = sizeof(*item);
+ newval_final = offset - 1;
+ targetptr_final = &buffer->c[cpu].offset;
+ /* copylen must be <= PAGE_SIZE. */
+ ret = cpu_op_cmpstorememcpy(targetptr_final, &offset, &newval_final,
+ sizeof(intptr_t), destptr, srcptr, copylen, cpu);
+ if (likely(!ret))
+ break;
+ assert(ret >= 0 || errno == EAGAIN);
+ }
+ }
+ return true;
+}
+
+void *test_percpu_memcpy_buffer_thread(void *arg)
+{
+ int i;
+ struct percpu_memcpy_buffer *buffer = (struct percpu_memcpy_buffer *)arg;
+
+ if (rseq_register_current_thread())
+ abort();
+
+ for (i = 0; i < opt_reps; i++) {
+ struct percpu_memcpy_buffer_node item;
+ bool result;
+
+ result = percpu_memcpy_buffer_pop(buffer, &item);
+ if (opt_yield)
+ sched_yield(); /* encourage shuffling */
+ if (result) {
+ if (!percpu_memcpy_buffer_push(buffer, item)) {
+ /* Should increase buffer size. */
+ abort();
+ }
+ }
+ }
+
+ if (rseq_unregister_current_thread())
+ abort();
+
+ return NULL;
+}
+
+/* Simultaneous modification to a per-cpu buffer from many threads. */
+void test_percpu_memcpy_buffer(void)
+{
+ const int num_threads = opt_threads;
+ int i, j, ret;
+ uint64_t sum = 0, expected_sum = 0;
+ struct percpu_memcpy_buffer buffer;
+ pthread_t test_threads[num_threads];
+ cpu_set_t allowed_cpus;
+
+ memset(&buffer, 0, sizeof(buffer));
+
+ /* Generate list entries for every usable cpu. */
+ sched_getaffinity(0, sizeof(allowed_cpus), &allowed_cpus);
+ for (i = 0; i < CPU_SETSIZE; i++) {
+ if (!CPU_ISSET(i, &allowed_cpus))
+ continue;
+ /* Worse-case is every item in same CPU. */
+ buffer.c[i].array =
+ malloc(sizeof(*buffer.c[i].array) * CPU_SETSIZE
+ * MEMCPY_BUFFER_ITEM_PER_CPU);
+ assert(buffer.c[i].array);
+ buffer.c[i].buflen = CPU_SETSIZE * MEMCPY_BUFFER_ITEM_PER_CPU;
+ for (j = 1; j <= MEMCPY_BUFFER_ITEM_PER_CPU; j++) {
+ expected_sum += 2 * j + 1;
+
+ /*
+ * We could theoretically put the word-sized
+ * "data" directly in the buffer. However, we
+ * want to model objects that would not fit
+ * within a single word, so allocate an object
+ * for each node.
+ */
+ buffer.c[i].array[j - 1].data1 = j;
+ buffer.c[i].array[j - 1].data2 = j + 1;
+ buffer.c[i].offset++;
+ }
+ }
+
+ for (i = 0; i < num_threads; i++) {
+ ret = pthread_create(&test_threads[i], NULL,
+ test_percpu_memcpy_buffer_thread, &buffer);
+ if (ret) {
+ errno = ret;
+ perror("pthread_create");
+ abort();
+ }
+ }
+
+ for (i = 0; i < num_threads; i++) {
+ pthread_join(test_threads[i], NULL);
+ if (ret) {
+ errno = ret;
+ perror("pthread_join");
+ abort();
+ }
+ }
+
+ for (i = 0; i < CPU_SETSIZE; i++) {
+ cpu_set_t pin_mask;
+ struct percpu_memcpy_buffer_node item;
+
+ if (!CPU_ISSET(i, &allowed_cpus))
+ continue;
+
+ CPU_ZERO(&pin_mask);
+ CPU_SET(i, &pin_mask);
+ sched_setaffinity(0, sizeof(pin_mask), &pin_mask);
+
+ while (percpu_memcpy_buffer_pop(&buffer, &item)) {
+ sum += item.data1;
+ sum += item.data2;
+ }
+ free(buffer.c[i].array);
+ }
+
+ /*
+ * All entries should now be accounted for (unless some external
+ * actor is interfering with our allowed affinity while this
+ * test is running).
+ */
+ assert(sum == expected_sum);
+}
+
+static void test_signal_interrupt_handler(int signo)
+{
+ signals_delivered++;
+}
+
+static int set_signal_handler(void)
+{
+ int ret = 0;
+ struct sigaction sa;
+ sigset_t sigset;
+
+ ret = sigemptyset(&sigset);
+ if (ret < 0) {
+ perror("sigemptyset");
+ return ret;
+ }
+
+ sa.sa_handler = test_signal_interrupt_handler;
+ sa.sa_mask = sigset;
+ sa.sa_flags = 0;
+ ret = sigaction(SIGUSR1, &sa, NULL);
+ if (ret < 0) {
+ perror("sigaction");
+ return ret;
+ }
+
+ printf_nobench("Signal handler set for SIGUSR1\n");
+
+ return ret;
+}
+
+static void show_usage(int argc, char **argv)
+{
+ printf("Usage : %s <OPTIONS>\n",
+ argv[0]);
+ printf("OPTIONS:\n");
+ printf(" [-1 loops] Number of loops for delay injection 1\n");
+ printf(" [-2 loops] Number of loops for delay injection 2\n");
+ printf(" [-3 loops] Number of loops for delay injection 3\n");
+ printf(" [-4 loops] Number of loops for delay injection 4\n");
+ printf(" [-5 loops] Number of loops for delay injection 5\n");
+ printf(" [-6 loops] Number of loops for delay injection 6 (-1 to enable -m)\n");
+ printf(" [-7 loops] Number of loops for delay injection 7 (-1 to enable -m)\n");
+ printf(" [-8 loops] Number of loops for delay injection 8 (-1 to enable -m)\n");
+ printf(" [-9 loops] Number of loops for delay injection 9 (-1 to enable -m)\n");
+ printf(" [-m N] Yield/sleep/kill every modulo N (default 0: disabled) (>= 0)\n");
+ printf(" [-y] Yield\n");
+ printf(" [-k] Kill thread with signal\n");
+ printf(" [-s S] S: =0: disabled (default), >0: sleep time (ms)\n");
+ printf(" [-t N] Number of threads (default 200)\n");
+ printf(" [-r N] Number of repetitions per thread (default 5000)\n");
+ printf(" [-d] Disable rseq system call (no initialization)\n");
+ printf(" [-D M] Disable rseq for each M threads\n");
+ printf(" [-T test] Choose test: (s)pinlock, (l)ist, (b)uffer, (m)emcpy, (i)ncrement\n");
+ printf(" [-h] Show this help.\n");
+ printf("\n");
+}
+
+int main(int argc, char **argv)
+{
+ int i;
+
+ if (set_signal_handler())
+ goto error;
+ for (i = 1; i < argc; i++) {
+ if (argv[i][0] != '-')
+ continue;
+ switch (argv[i][1]) {
+ case '1':
+ case '2':
+ case '3':
+ case '4':
+ case '5':
+ case '6':
+ case '7':
+ case '8':
+ case '9':
+ if (argc < i + 2) {
+ show_usage(argc, argv);
+ goto error;
+ }
+ loop_cnt[argv[i][1] - '0'] = atol(argv[i + 1]);
+ i++;
+ break;
+ case 'm':
+ if (argc < i + 2) {
+ show_usage(argc, argv);
+ goto error;
+ }
+ opt_modulo = atol(argv[i + 1]);
+ if (opt_modulo < 0) {
+ show_usage(argc, argv);
+ goto error;
+ }
+ i++;
+ break;
+ case 's':
+ if (argc < i + 2) {
+ show_usage(argc, argv);
+ goto error;
+ }
+ opt_sleep = atol(argv[i + 1]);
+ if (opt_sleep < 0) {
+ show_usage(argc, argv);
+ goto error;
+ }
+ i++;
+ break;
+ case 'y':
+ opt_yield = 1;
+ break;
+ case 'k':
+ opt_signal = 1;
+ break;
+ case 'd':
+ opt_disable_rseq = 1;
+ break;
+ case 'D':
+ if (argc < i + 2) {
+ show_usage(argc, argv);
+ goto error;
+ }
+ opt_disable_mod = atol(argv[i + 1]);
+ if (opt_disable_mod < 0) {
+ show_usage(argc, argv);
+ goto error;
+ }
+ i++;
+ break;
+ case 't':
+ if (argc < i + 2) {
+ show_usage(argc, argv);
+ goto error;
+ }
+ opt_threads = atol(argv[i + 1]);
+ if (opt_threads < 0) {
+ show_usage(argc, argv);
+ goto error;
+ }
+ i++;
+ break;
+ case 'r':
+ if (argc < i + 2) {
+ show_usage(argc, argv);
+ goto error;
+ }
+ opt_reps = atol(argv[i + 1]);
+ if (opt_reps < 0) {
+ show_usage(argc, argv);
+ goto error;
+ }
+ i++;
+ break;
+ case 'h':
+ show_usage(argc, argv);
+ goto end;
+ case 'T':
+ if (argc < i + 2) {
+ show_usage(argc, argv);
+ goto error;
+ }
+ opt_test = *argv[i + 1];
+ switch (opt_test) {
+ case 's':
+ case 'l':
+ case 'i':
+ case 'b':
+ case 'm':
+ break;
+ default:
+ show_usage(argc, argv);
+ goto error;
+ }
+ i++;
+ break;
+ default:
+ show_usage(argc, argv);
+ goto error;
+ }
+ }
+
+ if (!opt_disable_rseq && rseq_register_current_thread())
+ goto error;
+ switch (opt_test) {
+ case 's':
+ printf_nobench("spinlock\n");
+ test_percpu_spinlock();
+ break;
+ case 'l':
+ printf_nobench("linked list\n");
+ test_percpu_list();
+ break;
+ case 'b':
+ printf_nobench("buffer\n");
+ test_percpu_buffer();
+ break;
+ case 'm':
+ printf_nobench("memcpy buffer\n");
+ test_percpu_memcpy_buffer();
+ break;
+ case 'i':
+ printf_nobench("counter increment\n");
+ test_percpu_inc();
+ break;
+ }
+ if (rseq_unregister_current_thread())
+ abort();
+end:
+ return 0;
+
+error:
+ return -1;
+}
diff --git a/tools/testing/selftests/rseq/rseq-arm.h b/tools/testing/selftests/rseq/rseq-arm.h
new file mode 100644
index 000000000000..b5f57d250071
--- /dev/null
+++ b/tools/testing/selftests/rseq/rseq-arm.h
@@ -0,0 +1,159 @@
+/*
+ * rseq-arm.h
+ *
+ * (C) Copyright 2016 - Mathieu Desnoyers <mathieu.desnoyers@...icios.com>
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ */
+
+#define smp_mb() __asm__ __volatile__ ("dmb" : : : "memory")
+#define smp_rmb() __asm__ __volatile__ ("dmb" : : : "memory")
+#define smp_wmb() __asm__ __volatile__ ("dmb" : : : "memory")
+
+#define smp_load_acquire(p) \
+__extension__ ({ \
+ __typeof(*p) ____p1 = READ_ONCE(*p); \
+ smp_mb(); \
+ ____p1; \
+})
+
+#define smp_acquire__after_ctrl_dep() smp_rmb()
+
+#define smp_store_release(p, v) \
+do { \
+ smp_mb(); \
+ WRITE_ONCE(*p, v); \
+} while (0)
+
+#define has_fast_acquire_release() 0
+#define has_single_copy_load_64() 1
+
+/*
+ * The __rseq_table section can be used by debuggers to better handle
+ * single-stepping through the restartable critical sections.
+ */
+
+#define RSEQ_FINISH_ASM(_target_final, _to_write_final, _start_value, \
+ _failure, _spec_store, _spec_input, \
+ _final_store, _final_input, _extra_clobber, \
+ _setup, _teardown, _scratch) \
+do { \
+ _scratch \
+ __asm__ __volatile__ goto ( \
+ ".pushsection __rseq_table, \"aw\"\n\t" \
+ ".balign 32\n\t" \
+ ".word 1f, 0x0, 2f, 0x0, 5f, 0x0, 0x0, 0x0\n\t" \
+ ".popsection\n\t" \
+ "1:\n\t" \
+ _setup \
+ RSEQ_INJECT_ASM(1) \
+ "adr r0, 3f\n\t" \
+ "str r0, [%[rseq_cs]]\n\t" \
+ RSEQ_INJECT_ASM(2) \
+ "ldr r0, %[current_event_counter]\n\t" \
+ "cmp %[start_event_counter], r0\n\t" \
+ "bne 5f\n\t" \
+ RSEQ_INJECT_ASM(3) \
+ _spec_store \
+ _final_store \
+ "2:\n\t" \
+ RSEQ_INJECT_ASM(5) \
+ _teardown \
+ "b 4f\n\t" \
+ ".balign 32\n\t" \
+ "3:\n\t" \
+ ".word 1b, 0x0, 2b, 0x0, 5f, 0x0, 0x0, 0x0\n\t" \
+ "5:\n\t" \
+ _teardown \
+ "b %l[failure]\n\t" \
+ "4:\n\t" \
+ : /* gcc asm goto does not allow outputs */ \
+ : [start_event_counter]"r"((_start_value).event_counter), \
+ [current_event_counter]"m"((_start_value).rseqp->u.e.event_counter), \
+ [rseq_cs]"r"(&(_start_value).rseqp->rseq_cs) \
+ _spec_input \
+ _final_input \
+ RSEQ_INJECT_INPUT \
+ : "r0", "memory", "cc" \
+ _extra_clobber \
+ RSEQ_INJECT_CLOBBER \
+ : _failure \
+ ); \
+} while (0)
+
+#define RSEQ_FINISH_FINAL_STORE_ASM() \
+ "str %[to_write_final], [%[target_final]]\n\t"
+
+#define RSEQ_FINISH_FINAL_STORE_RELEASE_ASM() \
+ "dmb\n\t" \
+ RSEQ_FINISH_FINAL_STORE_ASM()
+
+#define RSEQ_FINISH_FINAL_STORE_INPUT(_target_final, _to_write_final) \
+ , [to_write_final]"r"(_to_write_final), \
+ [target_final]"r"(_target_final)
+
+#define RSEQ_FINISH_SPECULATIVE_STORE_ASM() \
+ "str %[to_write_spec], [%[target_spec]]\n\t" \
+ RSEQ_INJECT_ASM(4)
+
+#define RSEQ_FINISH_SPECULATIVE_STORE_INPUT(_target_spec, _to_write_spec) \
+ , [to_write_spec]"r"(_to_write_spec), \
+ [target_spec]"r"(_target_spec)
+
+/* TODO: implement a faster memcpy. */
+#define RSEQ_FINISH_MEMCPY_STORE_ASM() \
+ "cmp %[len_memcpy], #0\n\t" \
+ "beq 333f\n\t" \
+ "222:\n\t" \
+ "ldrb %%r0, [%[to_write_memcpy]]\n\t" \
+ "strb %%r0, [%[target_memcpy]]\n\t" \
+ "adds %[to_write_memcpy], #1\n\t" \
+ "adds %[target_memcpy], #1\n\t" \
+ "subs %[len_memcpy], #1\n\t" \
+ "bne 222b\n\t" \
+ "333:\n\t" \
+ RSEQ_INJECT_ASM(4)
+
+#define RSEQ_FINISH_MEMCPY_STORE_INPUT(_target_memcpy, _to_write_memcpy, _len_memcpy) \
+ , [to_write_memcpy]"r"(_to_write_memcpy), \
+ [target_memcpy]"r"(_target_memcpy), \
+ [len_memcpy]"r"(_len_memcpy), \
+ [rseq_scratch0]"m"(rseq_scratch[0]), \
+ [rseq_scratch1]"m"(rseq_scratch[1]), \
+ [rseq_scratch2]"m"(rseq_scratch[2])
+
+/* We can use r0. */
+#define RSEQ_FINISH_MEMCPY_CLOBBER()
+
+#define RSEQ_FINISH_MEMCPY_SCRATCH() \
+ uint32_t rseq_scratch[3];
+
+/*
+ * We need to save and restore those input registers so they can be
+ * modified within the assembly.
+ */
+#define RSEQ_FINISH_MEMCPY_SETUP() \
+ "str %[to_write_memcpy], %[rseq_scratch0]\n\t" \
+ "str %[target_memcpy], %[rseq_scratch1]\n\t" \
+ "str %[len_memcpy], %[rseq_scratch2]\n\t"
+
+#define RSEQ_FINISH_MEMCPY_TEARDOWN() \
+ "ldr %[len_memcpy], %[rseq_scratch2]\n\t" \
+ "ldr %[target_memcpy], %[rseq_scratch1]\n\t" \
+ "ldr %[to_write_memcpy], %[rseq_scratch0]\n\t"
diff --git a/tools/testing/selftests/rseq/rseq-ppc.h b/tools/testing/selftests/rseq/rseq-ppc.h
new file mode 100644
index 000000000000..94c8ba0b4311
--- /dev/null
+++ b/tools/testing/selftests/rseq/rseq-ppc.h
@@ -0,0 +1,266 @@
+/*
+ * rseq-ppc.h
+ *
+ * (C) Copyright 2016 - Mathieu Desnoyers <mathieu.desnoyers@...icios.com>
+ * (C) Copyright 2016 - Boqun Feng <boqun.feng@...il.com>
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ */
+
+#define smp_mb() __asm__ __volatile__ ("sync" : : : "memory")
+#define smp_lwsync() __asm__ __volatile__ ("lwsync" : : : "memory")
+#define smp_rmb() smp_lwsync()
+#define smp_wmb() smp_lwsync()
+
+#define smp_load_acquire(p) \
+__extension__ ({ \
+ __typeof(*p) ____p1 = READ_ONCE(*p); \
+ smp_lwsync(); \
+ ____p1; \
+})
+
+#define smp_acquire__after_ctrl_dep() smp_lwsync()
+
+#define smp_store_release(p, v) \
+do { \
+ smp_lwsync(); \
+ WRITE_ONCE(*p, v); \
+} while (0)
+
+#define has_fast_acquire_release() 0
+
+#ifdef __PPC64__
+#define has_single_copy_load_64() 1
+#else
+#define has_single_copy_load_64() 0
+#endif
+
+/*
+ * The __rseq_table section can be used by debuggers to better handle
+ * single-stepping through the restartable critical sections.
+ */
+
+#ifdef __PPC64__
+
+#define RSEQ_FINISH_ASM(_target_final, _to_write_final, _start_value, \
+ _failure, _spec_store, _spec_input, \
+ _final_store, _final_input, _extra_clobber, \
+ _setup, _teardown, _scratch) \
+ __asm__ __volatile__ goto ( \
+ ".pushsection __rseq_table, \"aw\"\n\t" \
+ ".balign 32\n\t" \
+ "3:\n\t" \
+ ".quad 1f, 2f, 4f\n\t" \
+ ".long 0x0, 0x0\n\t" \
+ ".popsection\n\t" \
+ "1:\n\t" \
+ _setup \
+ RSEQ_INJECT_ASM(1) \
+ "lis %%r17, (3b)@highest\n\t" \
+ "ori %%r17, %%r17, (3b)@higher\n\t" \
+ "rldicr %%r17, %%r17, 32, 31\n\t" \
+ "oris %%r17, %%r17, (3b)@h\n\t" \
+ "ori %%r17, %%r17, (3b)@l\n\t" \
+ "std %%r17, 0(%[rseq_cs])\n\t" \
+ RSEQ_INJECT_ASM(2) \
+ "lwz %%r17, %[current_event_counter]\n\t" \
+ "cmpw cr7, %[start_event_counter], %%r17\n\t" \
+ "bne- cr7, 4f\n\t" \
+ RSEQ_INJECT_ASM(3) \
+ _spec_store \
+ _final_store \
+ "2:\n\t" \
+ RSEQ_INJECT_ASM(5) \
+ _teardown \
+ "b 5f\n\t" \
+ "4:\n\t" \
+ _teardown \
+ "b %l[failure]\n\t" \
+ "5:\n\t" \
+ : /* gcc asm goto does not allow outputs */ \
+ : [start_event_counter]"r"((_start_value).event_counter), \
+ [current_event_counter]"m"((_start_value).rseqp->u.e.event_counter), \
+ [rseq_cs]"b"(&(_start_value).rseqp->rseq_cs) \
+ _spec_input \
+ _final_input \
+ RSEQ_INJECT_INPUT \
+ : "r17", "memory", "cc" \
+ _extra_clobber \
+ RSEQ_INJECT_CLOBBER \
+ : _failure \
+ )
+
+#define RSEQ_FINISH_FINAL_STORE_ASM() \
+ "std %[to_write_final], 0(%[target_final])\n\t"
+
+#define RSEQ_FINISH_FINAL_STORE_RELEASE_ASM() \
+ "lwsync\n\t" \
+ RSEQ_FINISH_FINAL_STORE_ASM()
+
+#define RSEQ_FINISH_FINAL_STORE_INPUT(_target_final, _to_write_final) \
+ , [to_write_final]"r"(_to_write_final), \
+ [target_final]"b"(_target_final)
+
+#define RSEQ_FINISH_SPECULATIVE_STORE_ASM() \
+ "std %[to_write_spec], 0(%[target_spec])\n\t" \
+ RSEQ_INJECT_ASM(4)
+
+#define RSEQ_FINISH_SPECULATIVE_STORE_INPUT(_target_spec, _to_write_spec) \
+ , [to_write_spec]"r"(_to_write_spec), \
+ [target_spec]"b"(_target_spec)
+
+/* TODO: implement a faster memcpy. */
+#define RSEQ_FINISH_MEMCPY_STORE_ASM() \
+ "cmpdi %%r19, 0\n\t" \
+ "beq 333f\n\t" \
+ "addi %%r20, %%r20, -1\n\t" \
+ "addi %%r21, %%r21, -1\n\t" \
+ "222:\n\t" \
+ "lbzu %%r18, 1(%%r20)\n\t" \
+ "stbu %%r18, 1(%%r21)\n\t" \
+ "addi %%r19, %%r19, -1\n\t" \
+ "cmpdi %%r19, 0\n\t" \
+ "bne 222b\n\t" \
+ "333:\n\t" \
+ RSEQ_INJECT_ASM(4)
+
+#define RSEQ_FINISH_MEMCPY_STORE_INPUT(_target_memcpy, _to_write_memcpy, _len_memcpy) \
+ , [to_write_memcpy]"r"(_to_write_memcpy), \
+ [target_memcpy]"r"(_target_memcpy), \
+ [len_memcpy]"r"(_len_memcpy)
+
+#define RSEQ_FINISH_MEMCPY_CLOBBER() \
+ , "r18", "r19", "r20", "r21"
+
+#define RSEQ_FINISH_MEMCPY_SCRATCH()
+
+/*
+ * We use extra registers to hold the input registers, and we don't need to
+ * save and restore the input registers.
+ */
+#define RSEQ_FINISH_MEMCPY_SETUP() \
+ "mr %%r19, %[len_memcpy]\n\t" \
+ "mr %%r20, %[to_write_memcpy]\n\t" \
+ "mr %%r21, %[target_memcpy]\n\t" \
+
+#define RSEQ_FINISH_MEMCPY_TEARDOWN()
+
+#else /* #ifdef __PPC64__ */
+
+#define RSEQ_FINISH_ASM(_target_final, _to_write_final, _start_value, \
+ _failure, _spec_store, _spec_input, \
+ _final_store, _final_input, _extra_clobber, \
+ _setup, _teardown, _scratch) \
+ __asm__ __volatile__ goto ( \
+ ".pushsection __rseq_table, \"aw\"\n\t" \
+ ".balign 32\n\t" \
+ "3:\n\t" \
+ /* 32-bit only supported on BE */ \
+ ".long 0x0, 1f, 0x0, 2f, 0x0, 4f, 0x0, 0x0\n\t" \
+ ".popsection\n\t" \
+ "1:\n\t" \
+ _setup \
+ RSEQ_INJECT_ASM(1) \
+ "lis %%r17, (3b)@ha\n\t" \
+ "addi %%r17, %%r17, (3b)@l\n\t" \
+ "stw %%r17, 0(%[rseq_cs])\n\t" \
+ RSEQ_INJECT_ASM(2) \
+ "lwz %%r17, %[current_event_counter]\n\t" \
+ "cmpw cr7, %[start_event_counter], %%r17\n\t" \
+ "bne- cr7, 4f\n\t" \
+ RSEQ_INJECT_ASM(3) \
+ _spec_store \
+ _final_store \
+ "2:\n\t" \
+ RSEQ_INJECT_ASM(5) \
+ _teardown \
+ "b 5f\n\t" \
+ "4:\n\t" \
+ _teardown \
+ "b %l[failure]\n\t" \
+ "5:\n\t" \
+ : /* gcc asm goto does not allow outputs */ \
+ : [start_event_counter]"r"((_start_value).event_counter), \
+ [current_event_counter]"m"((_start_value).rseqp->u.e.event_counter), \
+ [rseq_cs]"b"(&(_start_value).rseqp->rseq_cs) \
+ _spec_input \
+ _final_input \
+ RSEQ_INJECT_INPUT \
+ : "r17", "memory", "cc" \
+ _extra_clobber \
+ RSEQ_INJECT_CLOBBER \
+ : _failure \
+ )
+
+#define RSEQ_FINISH_FINAL_STORE_ASM() \
+ "stw %[to_write_final], 0(%[target_final])\n\t"
+
+#define RSEQ_FINISH_FINAL_STORE_RELEASE_ASM() \
+ "lwsync\n\t" \
+ RSEQ_FINISH_FINAL_STORE_ASM()
+
+#define RSEQ_FINISH_FINAL_STORE_INPUT(_target_final, _to_write_final) \
+ , [to_write_final]"r"(_to_write_final), \
+ [target_final]"b"(_target_final)
+
+#define RSEQ_FINISH_SPECULATIVE_STORE_ASM() \
+ "stw %[to_write_spec], 0(%[target_spec])\n\t" \
+ RSEQ_INJECT_ASM(4)
+
+#define RSEQ_FINISH_SPECULATIVE_STORE_INPUT(_target_spec, _to_write_spec) \
+ , [to_write_spec]"r"(_to_write_spec), \
+ [target_spec]"b"(_target_spec)
+
+/* TODO: implement a faster memcpy. */
+#define RSEQ_FINISH_MEMCPY_STORE_ASM() \
+ "cmpwi %%r19, 0\n\t" \
+ "beq 333f\n\t" \
+ "addi %%r20, %%r20, -1\n\t" \
+ "addi %%r21, %%r21, -1\n\t" \
+ "222:\n\t" \
+ "lbzu %%r18, 1(%%r20)\n\t" \
+ "stbu %%r18, 1(%%r21)\n\t" \
+ "addi %%r19, %%r19, -1\n\t" \
+ "cmpwi %%r19, 0\n\t" \
+ "bne 222b\n\t" \
+ "333:\n\t" \
+ RSEQ_INJECT_ASM(4)
+
+#define RSEQ_FINISH_MEMCPY_STORE_INPUT(_target_memcpy, _to_write_memcpy, _len_memcpy) \
+ , [to_write_memcpy]"r"(_to_write_memcpy), \
+ [target_memcpy]"r"(_target_memcpy), \
+ [len_memcpy]"r"(_len_memcpy)
+
+#define RSEQ_FINISH_MEMCPY_CLOBBER() \
+ , "r18", "r19", "r20", "r21"
+
+#define RSEQ_FINISH_MEMCPY_SCRATCH()
+
+/*
+ * We use extra registers to hold the input registers, and we don't need to
+ * save and restore the input registers.
+ */
+#define RSEQ_FINISH_MEMCPY_SETUP() \
+ "mr %%r19, %[len_memcpy]\n\t" \
+ "mr %%r20, %[to_write_memcpy]\n\t" \
+ "mr %%r21, %[target_memcpy]\n\t" \
+
+#define RSEQ_FINISH_MEMCPY_TEARDOWN()
+
+#endif /* #else #ifdef __PPC64__ */
diff --git a/tools/testing/selftests/rseq/rseq-x86.h b/tools/testing/selftests/rseq/rseq-x86.h
new file mode 100644
index 000000000000..2896186eef9b
--- /dev/null
+++ b/tools/testing/selftests/rseq/rseq-x86.h
@@ -0,0 +1,304 @@
+/*
+ * rseq-x86.h
+ *
+ * (C) Copyright 2016 - Mathieu Desnoyers <mathieu.desnoyers@...icios.com>
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ */
+
+#ifdef __x86_64__
+
+#define smp_mb() __asm__ __volatile__ ("mfence" : : : "memory")
+#define smp_rmb() barrier()
+#define smp_wmb() barrier()
+
+#define smp_load_acquire(p) \
+__extension__ ({ \
+ __typeof(*p) ____p1 = READ_ONCE(*p); \
+ barrier(); \
+ ____p1; \
+})
+
+#define smp_acquire__after_ctrl_dep() smp_rmb()
+
+#define smp_store_release(p, v) \
+do { \
+ barrier(); \
+ WRITE_ONCE(*p, v); \
+} while (0)
+
+#define has_fast_acquire_release() 1
+#define has_single_copy_load_64() 1
+
+/*
+ * The __rseq_table section can be used by debuggers to better handle
+ * single-stepping through the restartable critical sections.
+ */
+#define RSEQ_FINISH_ASM(_target_final, _to_write_final, _start_value, \
+ _failure, _spec_store, _spec_input, \
+ _final_store, _final_input, _extra_clobber, \
+ _setup, _teardown, _scratch) \
+do { \
+ _scratch \
+ __asm__ __volatile__ goto ( \
+ ".pushsection __rseq_table, \"aw\"\n\t" \
+ ".balign 32\n\t" \
+ "3:\n\t" \
+ ".quad 1f, 2f, 4f\n\t" \
+ ".long 0x0, 0x0\n\t" \
+ ".popsection\n\t" \
+ "1:\n\t" \
+ _setup \
+ RSEQ_INJECT_ASM(1) \
+ "leaq 3b(%%rip), %%rax\n\t" \
+ "movq %%rax, %[rseq_cs]\n\t" \
+ RSEQ_INJECT_ASM(2) \
+ "cmpl %[start_event_counter], %[current_event_counter]\n\t" \
+ "jnz 4f\n\t" \
+ RSEQ_INJECT_ASM(3) \
+ _spec_store \
+ _final_store \
+ "2:\n\t" \
+ RSEQ_INJECT_ASM(5) \
+ _teardown \
+ ".pushsection __rseq_failure, \"a\"\n\t" \
+ "4:\n\t" \
+ _teardown \
+ "jmp %l[failure]\n\t" \
+ ".popsection\n\t" \
+ : /* gcc asm goto does not allow outputs */ \
+ : [start_event_counter]"r"((_start_value).event_counter), \
+ [current_event_counter]"m"((_start_value).rseqp->u.e.event_counter), \
+ [rseq_cs]"m"((_start_value).rseqp->rseq_cs) \
+ _spec_input \
+ _final_input \
+ RSEQ_INJECT_INPUT \
+ : "memory", "cc", "rax" \
+ _extra_clobber \
+ RSEQ_INJECT_CLOBBER \
+ : _failure \
+ ); \
+} while (0)
+
+#define RSEQ_FINISH_FINAL_STORE_ASM() \
+ "movq %[to_write_final], %[target_final]\n\t"
+
+/* x86-64 is TSO */
+#define RSEQ_FINISH_FINAL_STORE_RELEASE_ASM() \
+ RSEQ_FINISH_FINAL_STORE_ASM()
+
+#define RSEQ_FINISH_FINAL_STORE_INPUT(_target_final, _to_write_final) \
+ , [to_write_final]"r"(_to_write_final), \
+ [target_final]"m"(*(_target_final))
+
+#define RSEQ_FINISH_SPECULATIVE_STORE_ASM() \
+ "movq %[to_write_spec], %[target_spec]\n\t" \
+ RSEQ_INJECT_ASM(4)
+
+#define RSEQ_FINISH_SPECULATIVE_STORE_INPUT(_target_spec, _to_write_spec) \
+ , [to_write_spec]"r"(_to_write_spec), \
+ [target_spec]"m"(*(_target_spec))
+
+/* TODO: implement a faster memcpy. */
+#define RSEQ_FINISH_MEMCPY_STORE_ASM() \
+ "test %[len_memcpy], %[len_memcpy]\n\t" \
+ "jz 333f\n\t" \
+ "222:\n\t" \
+ "movb (%[to_write_memcpy]), %%al\n\t" \
+ "movb %%al, (%[target_memcpy])\n\t" \
+ "inc %[to_write_memcpy]\n\t" \
+ "inc %[target_memcpy]\n\t" \
+ "dec %[len_memcpy]\n\t" \
+ "jnz 222b\n\t" \
+ "333:\n\t" \
+ RSEQ_INJECT_ASM(4)
+
+#define RSEQ_FINISH_MEMCPY_STORE_INPUT(_target_memcpy, _to_write_memcpy, _len_memcpy) \
+ , [to_write_memcpy]"r"(_to_write_memcpy), \
+ [target_memcpy]"r"(_target_memcpy), \
+ [len_memcpy]"r"(_len_memcpy), \
+ [rseq_scratch0]"m"(rseq_scratch[0]), \
+ [rseq_scratch1]"m"(rseq_scratch[1]), \
+ [rseq_scratch2]"m"(rseq_scratch[2])
+
+#define RSEQ_FINISH_MEMCPY_CLOBBER() \
+ , "rax"
+
+#define RSEQ_FINISH_MEMCPY_SCRATCH() \
+ uint64_t rseq_scratch[3];
+
+/*
+ * We need to save and restore those input registers so they can be
+ * modified within the assembly.
+ */
+#define RSEQ_FINISH_MEMCPY_SETUP() \
+ "movq %[to_write_memcpy], %[rseq_scratch0]\n\t" \
+ "movq %[target_memcpy], %[rseq_scratch1]\n\t" \
+ "movq %[len_memcpy], %[rseq_scratch2]\n\t"
+
+#define RSEQ_FINISH_MEMCPY_TEARDOWN() \
+ "movq %[rseq_scratch2], %[len_memcpy]\n\t" \
+ "movq %[rseq_scratch1], %[target_memcpy]\n\t" \
+ "movq %[rseq_scratch0], %[to_write_memcpy]\n\t"
+
+#elif __i386__
+
+/*
+ * Support older 32-bit architectures that do not implement fence
+ * instructions.
+ */
+#define smp_mb() \
+ __asm__ __volatile__ ("lock; addl $0,0(%%esp)" : : : "memory")
+#define smp_rmb() \
+ __asm__ __volatile__ ("lock; addl $0,0(%%esp)" : : : "memory")
+#define smp_wmb() \
+ __asm__ __volatile__ ("lock; addl $0,0(%%esp)" : : : "memory")
+
+#define smp_load_acquire(p) \
+__extension__ ({ \
+ __typeof(*p) ____p1 = READ_ONCE(*p); \
+ smp_mb(); \
+ ____p1; \
+})
+
+#define smp_acquire__after_ctrl_dep() smp_rmb()
+
+#define smp_store_release(p, v) \
+do { \
+ smp_mb(); \
+ WRITE_ONCE(*p, v); \
+} while (0)
+
+#define has_fast_acquire_release() 0
+#define has_single_copy_load_64() 0
+
+/*
+ * Use eax as scratch register and take memory operands as input to
+ * lessen register pressure. Especially needed when compiling
+ * do_rseq_memcpy() in O0.
+ */
+#define RSEQ_FINISH_ASM(_target_final, _to_write_final, _start_value, \
+ _failure, _spec_store, _spec_input, \
+ _final_store, _final_input, _extra_clobber, \
+ _setup, _teardown, _scratch) \
+do { \
+ _scratch \
+ __asm__ __volatile__ goto ( \
+ ".pushsection __rseq_table, \"aw\"\n\t" \
+ ".balign 32\n\t" \
+ "3:\n\t" \
+ ".long 1f, 0x0, 2f, 0x0, 4f, 0x0, 0x0, 0x0\n\t" \
+ ".popsection\n\t" \
+ "1:\n\t" \
+ _setup \
+ RSEQ_INJECT_ASM(1) \
+ "movl $3b, %[rseq_cs]\n\t" \
+ RSEQ_INJECT_ASM(2) \
+ "movl %[start_event_counter], %%eax\n\t" \
+ "cmpl %%eax, %[current_event_counter]\n\t" \
+ "jnz 4f\n\t" \
+ RSEQ_INJECT_ASM(3) \
+ _spec_store \
+ _final_store \
+ "2:\n\t" \
+ RSEQ_INJECT_ASM(5) \
+ _teardown \
+ ".pushsection __rseq_failure, \"a\"\n\t" \
+ "4:\n\t" \
+ _teardown \
+ "jmp %l[failure]\n\t" \
+ ".popsection\n\t" \
+ : /* gcc asm goto does not allow outputs */ \
+ : [start_event_counter]"m"((_start_value).event_counter), \
+ [current_event_counter]"m"((_start_value).rseqp->u.e.event_counter), \
+ [rseq_cs]"m"((_start_value).rseqp->rseq_cs) \
+ _spec_input \
+ _final_input \
+ RSEQ_INJECT_INPUT \
+ : "memory", "cc", "eax" \
+ _extra_clobber \
+ RSEQ_INJECT_CLOBBER \
+ : _failure \
+ ); \
+} while (0)
+
+#define RSEQ_FINISH_FINAL_STORE_ASM() \
+ "movl %[to_write_final], %%eax\n\t" \
+ "movl %%eax, %[target_final]\n\t"
+
+#define RSEQ_FINISH_FINAL_STORE_RELEASE_ASM() \
+ "lock; addl $0,0(%%esp)\n\t" \
+ RSEQ_FINISH_FINAL_STORE_ASM()
+
+#define RSEQ_FINISH_FINAL_STORE_INPUT(_target_final, _to_write_final) \
+ , [to_write_final]"m"(_to_write_final), \
+ [target_final]"m"(*(_target_final))
+
+#define RSEQ_FINISH_SPECULATIVE_STORE_ASM() \
+ "movl %[to_write_spec], %%eax\n\t" \
+ "movl %%eax, %[target_spec]\n\t" \
+ RSEQ_INJECT_ASM(4)
+
+#define RSEQ_FINISH_SPECULATIVE_STORE_INPUT(_target_spec, _to_write_spec) \
+ , [to_write_spec]"m"(_to_write_spec), \
+ [target_spec]"m"(*(_target_spec))
+
+/* TODO: implement a faster memcpy. */
+#define RSEQ_FINISH_MEMCPY_STORE_ASM() \
+ "movl %[len_memcpy], %%eax\n\t" \
+ "test %%eax, %%eax\n\t" \
+ "jz 333f\n\t" \
+ "222:\n\t" \
+ "movb (%[to_write_memcpy]), %%al\n\t" \
+ "movb %%al, (%[target_memcpy])\n\t" \
+ "inc %[to_write_memcpy]\n\t" \
+ "inc %[target_memcpy]\n\t" \
+ "decl %[rseq_scratch2]\n\t" \
+ "jnz 222b\n\t" \
+ "333:\n\t" \
+ RSEQ_INJECT_ASM(4)
+
+#define RSEQ_FINISH_MEMCPY_STORE_INPUT(_target_memcpy, _to_write_memcpy, _len_memcpy) \
+ , [to_write_memcpy]"r"(_to_write_memcpy), \
+ [target_memcpy]"r"(_target_memcpy), \
+ [len_memcpy]"m"(_len_memcpy), \
+ [rseq_scratch0]"m"(rseq_scratch[0]), \
+ [rseq_scratch1]"m"(rseq_scratch[1]), \
+ [rseq_scratch2]"m"(rseq_scratch[2])
+
+#define RSEQ_FINISH_MEMCPY_CLOBBER()
+
+#define RSEQ_FINISH_MEMCPY_SCRATCH() \
+ uint32_t rseq_scratch[3];
+
+/*
+ * We need to save and restore those input registers so they can be
+ * modified within the assembly.
+ */
+#define RSEQ_FINISH_MEMCPY_SETUP() \
+ "movl %[to_write_memcpy], %[rseq_scratch0]\n\t" \
+ "movl %[target_memcpy], %[rseq_scratch1]\n\t" \
+ "movl %[len_memcpy], %%eax\n\t" \
+ "movl %%eax, %[rseq_scratch2]\n\t"
+
+#define RSEQ_FINISH_MEMCPY_TEARDOWN() \
+ "movl %[rseq_scratch1], %[target_memcpy]\n\t" \
+ "movl %[rseq_scratch0], %[to_write_memcpy]\n\t"
+
+#endif
diff --git a/tools/testing/selftests/rseq/rseq.c b/tools/testing/selftests/rseq/rseq.c
new file mode 100644
index 000000000000..79eba7f20064
--- /dev/null
+++ b/tools/testing/selftests/rseq/rseq.c
@@ -0,0 +1,78 @@
+/*
+ * rseq.c
+ *
+ * Copyright (C) 2016 Mathieu Desnoyers <mathieu.desnoyers@...icios.com>
+ *
+ * This library is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU Lesser General Public
+ * License as published by the Free Software Foundation; only
+ * version 2.1 of the License.
+ *
+ * This library is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ * Lesser General Public License for more details.
+ */
+
+#define _GNU_SOURCE
+#include <errno.h>
+#include <sched.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include <unistd.h>
+#include <syscall.h>
+#include <assert.h>
+#include <signal.h>
+
+#include "rseq.h"
+
+#define ARRAY_SIZE(arr) (sizeof(arr) / sizeof((arr)[0]))
+
+__attribute__((weak)) __thread volatile struct rseq __rseq_abi = {
+ .u.e.cpu_id = -1,
+};
+
+static int sys_rseq(volatile struct rseq *rseq_abi, int flags)
+{
+ return syscall(__NR_rseq, rseq_abi, flags);
+}
+
+int rseq_register_current_thread(void)
+{
+ int rc;
+
+ rc = sys_rseq(&__rseq_abi, 0);
+ if (rc) {
+ fprintf(stderr, "Error: sys_rseq(...) failed(%d): %s\n",
+ errno, strerror(errno));
+ return -1;
+ }
+ assert(rseq_current_cpu() >= 0);
+ return 0;
+}
+
+int rseq_unregister_current_thread(void)
+{
+ int rc;
+
+ rc = sys_rseq(NULL, 0);
+ if (rc) {
+ fprintf(stderr, "Error: sys_rseq(...) failed(%d): %s\n",
+ errno, strerror(errno));
+ return -1;
+ }
+ return 0;
+}
+
+int rseq_fallback_current_cpu(void)
+{
+ int cpu;
+
+ cpu = sched_getcpu();
+ if (cpu < 0) {
+ perror("sched_getcpu()");
+ abort();
+ }
+ return cpu;
+}
diff --git a/tools/testing/selftests/rseq/rseq.h b/tools/testing/selftests/rseq/rseq.h
new file mode 100644
index 000000000000..b0015f255ffc
--- /dev/null
+++ b/tools/testing/selftests/rseq/rseq.h
@@ -0,0 +1,298 @@
+/*
+ * rseq.h
+ *
+ * (C) Copyright 2016 - Mathieu Desnoyers <mathieu.desnoyers@...icios.com>
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to deal
+ * in the Software without restriction, including without limitation the rights
+ * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+ * copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ */
+
+#ifndef RSEQ_H
+#define RSEQ_H
+
+#include <stdint.h>
+#include <stdbool.h>
+#include <pthread.h>
+#include <signal.h>
+#include <sched.h>
+#include <errno.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <sched.h>
+#include <linux/rseq.h>
+
+/*
+ * Empty code injection macros, override when testing.
+ * It is important to consider that the ASM injection macros need to be
+ * fully reentrant (e.g. do not modify the stack).
+ */
+#ifndef RSEQ_INJECT_ASM
+#define RSEQ_INJECT_ASM(n)
+#endif
+
+#ifndef RSEQ_INJECT_C
+#define RSEQ_INJECT_C(n)
+#endif
+
+#ifndef RSEQ_INJECT_INPUT
+#define RSEQ_INJECT_INPUT
+#endif
+
+#ifndef RSEQ_INJECT_CLOBBER
+#define RSEQ_INJECT_CLOBBER
+#endif
+
+#ifndef RSEQ_INJECT_FAILED
+#define RSEQ_INJECT_FAILED
+#endif
+
+extern __thread volatile struct rseq __rseq_abi;
+
+#define likely(x) __builtin_expect(!!(x), 1)
+#define unlikely(x) __builtin_expect(!!(x), 0)
+#define barrier() __asm__ __volatile__("" : : : "memory")
+
+#define ACCESS_ONCE(x) (*(__volatile__ __typeof__(x) *)&(x))
+#define WRITE_ONCE(x, v) __extension__ ({ ACCESS_ONCE(x) = (v); })
+#define READ_ONCE(x) ACCESS_ONCE(x)
+
+#if defined(__x86_64__) || defined(__i386__)
+#include <rseq-x86.h>
+#elif defined(__ARMEL__)
+#include <rseq-arm.h>
+#elif defined(__PPC__)
+#include <rseq-ppc.h>
+#else
+#error unsupported target
+#endif
+
+/* State returned by rseq_start, passed as argument to rseq_finish. */
+struct rseq_state {
+ volatile struct rseq *rseqp;
+ int32_t cpu_id; /* cpu_id at start. */
+ uint32_t event_counter; /* event_counter at start. */
+};
+
+/*
+ * Register rseq for the current thread. This needs to be called once
+ * by any thread which uses restartable sequences, before they start
+ * using restartable sequences. If initialization is not invoked, or if
+ * it fails, the restartable critical sections will fall-back on locking
+ * (rseq_lock).
+ */
+int rseq_register_current_thread(void);
+
+/*
+ * Unregister rseq for current thread.
+ */
+int rseq_unregister_current_thread(void);
+
+/*
+ * Restartable sequence fallback for reading the current CPU number.
+ */
+int rseq_fallback_current_cpu(void);
+
+static inline int32_t rseq_cpu_at_start(struct rseq_state start_value)
+{
+ return start_value.cpu_id;
+}
+
+static inline int32_t rseq_current_cpu_raw(void)
+{
+ return ACCESS_ONCE(__rseq_abi.u.e.cpu_id);
+}
+
+static inline int32_t rseq_current_cpu(void)
+{
+ int32_t cpu;
+
+ cpu = rseq_current_cpu_raw();
+ if (unlikely(cpu < 0))
+ cpu = rseq_fallback_current_cpu();
+ return cpu;
+}
+
+static inline __attribute__((always_inline))
+struct rseq_state rseq_start(void)
+{
+ struct rseq_state result;
+
+ result.rseqp = &__rseq_abi;
+ if (has_single_copy_load_64()) {
+ union rseq_cpu_event u;
+
+ u.v = ACCESS_ONCE(result.rseqp->u.v);
+ result.event_counter = u.e.event_counter;
+ result.cpu_id = u.e.cpu_id;
+ } else {
+ result.event_counter =
+ ACCESS_ONCE(result.rseqp->u.e.event_counter);
+ /* load event_counter before cpu_id. */
+ RSEQ_INJECT_C(6)
+ result.cpu_id = ACCESS_ONCE(result.rseqp->u.e.cpu_id);
+ }
+ RSEQ_INJECT_C(7)
+ /*
+ * Ensure the compiler does not re-order loads of protected
+ * values before we load the event counter.
+ */
+ barrier();
+ return result;
+}
+
+enum rseq_finish_type {
+ RSEQ_FINISH_SINGLE,
+ RSEQ_FINISH_TWO,
+ RSEQ_FINISH_MEMCPY,
+};
+
+/*
+ * p_spec and to_write_spec are used for a speculative write attempted
+ * near the end of the restartable sequence. A rseq_finish2 may fail
+ * even after this write takes place.
+ *
+ * p_final and to_write_final are used for the final write. If this
+ * write takes place, the rseq_finish2 is guaranteed to succeed.
+ */
+static inline __attribute__((always_inline))
+bool __rseq_finish(intptr_t *p_spec, intptr_t to_write_spec,
+ void *p_memcpy, void *to_write_memcpy, size_t len_memcpy,
+ intptr_t *p_final, intptr_t to_write_final,
+ struct rseq_state start_value,
+ enum rseq_finish_type type, bool release)
+{
+ RSEQ_INJECT_C(9)
+
+ switch (type) {
+ case RSEQ_FINISH_SINGLE:
+ RSEQ_FINISH_ASM(p_final, to_write_final, start_value, failure,
+ /* no speculative write */, /* no speculative write */,
+ RSEQ_FINISH_FINAL_STORE_ASM(),
+ RSEQ_FINISH_FINAL_STORE_INPUT(p_final, to_write_final),
+ /* no extra clobber */, /* no arg */, /* no arg */,
+ /* no arg */
+ );
+ break;
+ case RSEQ_FINISH_TWO:
+ if (release) {
+ RSEQ_FINISH_ASM(p_final, to_write_final, start_value, failure,
+ RSEQ_FINISH_SPECULATIVE_STORE_ASM(),
+ RSEQ_FINISH_SPECULATIVE_STORE_INPUT(p_spec, to_write_spec),
+ RSEQ_FINISH_FINAL_STORE_RELEASE_ASM(),
+ RSEQ_FINISH_FINAL_STORE_INPUT(p_final, to_write_final),
+ /* no extra clobber */, /* no arg */, /* no arg */,
+ /* no arg */
+ );
+ } else {
+ RSEQ_FINISH_ASM(p_final, to_write_final, start_value, failure,
+ RSEQ_FINISH_SPECULATIVE_STORE_ASM(),
+ RSEQ_FINISH_SPECULATIVE_STORE_INPUT(p_spec, to_write_spec),
+ RSEQ_FINISH_FINAL_STORE_ASM(),
+ RSEQ_FINISH_FINAL_STORE_INPUT(p_final, to_write_final),
+ /* no extra clobber */, /* no arg */, /* no arg */,
+ /* no arg */
+ );
+ }
+ break;
+ case RSEQ_FINISH_MEMCPY:
+ if (release) {
+ RSEQ_FINISH_ASM(p_final, to_write_final, start_value, failure,
+ RSEQ_FINISH_MEMCPY_STORE_ASM(),
+ RSEQ_FINISH_MEMCPY_STORE_INPUT(p_memcpy, to_write_memcpy, len_memcpy),
+ RSEQ_FINISH_FINAL_STORE_RELEASE_ASM(),
+ RSEQ_FINISH_FINAL_STORE_INPUT(p_final, to_write_final),
+ RSEQ_FINISH_MEMCPY_CLOBBER(),
+ RSEQ_FINISH_MEMCPY_SETUP(),
+ RSEQ_FINISH_MEMCPY_TEARDOWN(),
+ RSEQ_FINISH_MEMCPY_SCRATCH()
+ );
+ } else {
+ RSEQ_FINISH_ASM(p_final, to_write_final, start_value, failure,
+ RSEQ_FINISH_MEMCPY_STORE_ASM(),
+ RSEQ_FINISH_MEMCPY_STORE_INPUT(p_memcpy, to_write_memcpy, len_memcpy),
+ RSEQ_FINISH_FINAL_STORE_ASM(),
+ RSEQ_FINISH_FINAL_STORE_INPUT(p_final, to_write_final),
+ RSEQ_FINISH_MEMCPY_CLOBBER(),
+ RSEQ_FINISH_MEMCPY_SETUP(),
+ RSEQ_FINISH_MEMCPY_TEARDOWN(),
+ RSEQ_FINISH_MEMCPY_SCRATCH()
+ );
+ }
+ break;
+ }
+ return true;
+failure:
+ RSEQ_INJECT_FAILED
+ return false;
+}
+
+static inline __attribute__((always_inline))
+bool rseq_finish(intptr_t *p, intptr_t to_write,
+ struct rseq_state start_value)
+{
+ return __rseq_finish(NULL, 0,
+ NULL, NULL, 0,
+ p, to_write, start_value,
+ RSEQ_FINISH_SINGLE, false);
+}
+
+static inline __attribute__((always_inline))
+bool rseq_finish2(intptr_t *p_spec, intptr_t to_write_spec,
+ intptr_t *p_final, intptr_t to_write_final,
+ struct rseq_state start_value)
+{
+ return __rseq_finish(p_spec, to_write_spec,
+ NULL, NULL, 0,
+ p_final, to_write_final, start_value,
+ RSEQ_FINISH_TWO, false);
+}
+
+static inline __attribute__((always_inline))
+bool rseq_finish2_release(intptr_t *p_spec, intptr_t to_write_spec,
+ intptr_t *p_final, intptr_t to_write_final,
+ struct rseq_state start_value)
+{
+ return __rseq_finish(p_spec, to_write_spec,
+ NULL, NULL, 0,
+ p_final, to_write_final, start_value,
+ RSEQ_FINISH_TWO, true);
+}
+
+static inline __attribute__((always_inline))
+bool rseq_finish_memcpy(void *p_memcpy, void *to_write_memcpy,
+ size_t len_memcpy, intptr_t *p_final, intptr_t to_write_final,
+ struct rseq_state start_value)
+{
+ return __rseq_finish(NULL, 0,
+ p_memcpy, to_write_memcpy, len_memcpy,
+ p_final, to_write_final, start_value,
+ RSEQ_FINISH_MEMCPY, false);
+}
+
+static inline __attribute__((always_inline))
+bool rseq_finish_memcpy_release(void *p_memcpy, void *to_write_memcpy,
+ size_t len_memcpy, intptr_t *p_final, intptr_t to_write_final,
+ struct rseq_state start_value)
+{
+ return __rseq_finish(NULL, 0,
+ p_memcpy, to_write_memcpy, len_memcpy,
+ p_final, to_write_final, start_value,
+ RSEQ_FINISH_MEMCPY, true);
+}
+
+#endif /* RSEQ_H_ */
--
2.11.0
Powered by blists - more mailing lists