1 files changed, 268 insertions, 0 deletions
diff --git a/storage/tokudb/PerconaFT/src/tests/blocking-next-prev-deadlock.cc b/storage/tokudb/PerconaFT/src/tests/blocking-next-prev-deadlock.cc
new file mode 100644
index 00000000..781708cc
--- /dev/null
+++ b/storage/tokudb/PerconaFT/src/tests/blocking-next-prev-deadlock.cc
@@ -0,0 +1,268 @@
+/* -*- mode: C++; c-basic-offset: 4; indent-tabs-mode: nil -*- */
+// vim: ft=cpp:expandtab:ts=8:sw=4:softtabstop=4:
+#ident "$Id$"
+/*======
+This file is part of PerconaFT.
+
+
+Copyright (c) 2006, 2015, Percona and/or its affiliates. All rights reserved.
+
+    PerconaFT is free software: you can redistribute it and/or modify
+    it under the terms of the GNU General Public License, version 2,
+    as published by the Free Software Foundation.
+
+    PerconaFT 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 General Public License for more details.
+
+    You should have received a copy of the GNU General Public License
+    along with PerconaFT.  If not, see <http://www.gnu.org/licenses/>.
+
+----------------------------------------
+
+    PerconaFT is free software: you can redistribute it and/or modify
+    it under the terms of the GNU Affero General Public License, version 3,
+    as published by the Free Software Foundation.
+
+    PerconaFT 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 Affero General Public License for more details.
+
+    You should have received a copy of the GNU Affero General Public License
+    along with PerconaFT.  If not, see <http://www.gnu.org/licenses/>.
+======= */
+
+#ident "Copyright (c) 2006, 2015, Percona and/or its affiliates. All rights reserved."
+
+// verify that two transactions doing cursor next and prev operations detect a deadlock when
+// using write locking cursor operations.
+
+#include "test.h"
+#include "toku_pthread.h"
+
+static uint64_t get_key(DBT *key) {
+    uint64_t k = 0;
+    assert(key->size == sizeof k);
+    memcpy(&k, key->data, key->size);
+    return htonl(k);
+}
+
+static void populate(DB_ENV *db_env, DB *db, uint64_t nrows) {
+    int r;
+
+    DB_TXN *txn = NULL;
+    r = db_env->txn_begin(db_env, NULL, &txn, 0); assert(r == 0);
+
+    for (uint64_t i = 0; i < nrows; i++) {
+
+        uint64_t k = htonl(i);
+        uint64_t v = i;
+        DBT key = { .data = &k, .size = sizeof k };
+        DBT val = { .data = &v, .size = sizeof v };
+        r = db->put(db, txn, &key, &val, 0); assert(r == 0);
+    }
+
+    r = txn->commit(txn, 0); assert(r == 0);
+}
+
+struct my_callback_context {
+    DBT key;
+    DBT val;
+};
+
+static int blocking_next_callback(DBT const *a UU(), DBT const *b UU(), void *e UU()) {
+    DBT const *found_key = a;
+    DBT const *found_val = b;
+    struct my_callback_context *context = (struct my_callback_context *) e;
+    copy_dbt(&context->key, found_key);
+    copy_dbt(&context->val, found_val);
+    return 0;
+}
+
+static void blocking_next(DB_ENV *db_env, DB *db, uint64_t nrows UU(), long sleeptime) {
+    int r;
+
+    struct my_callback_context context;
+    dbt_init_realloc(&context.key);
+    dbt_init_realloc(&context.val);
+
+    DB_TXN *txn = NULL;
+    r = db_env->txn_begin(db_env, NULL, &txn, 0); assert(r == 0);
+
+    DBC *cursor = NULL;
+    r = db->cursor(db, txn, &cursor, 0); assert(r == 0);
+
+    uint64_t i;
+    for (i = 0; ; i++) {
+        r = cursor->c_getf_next(cursor, DB_RMW, blocking_next_callback, &context);
+        if (r != 0)
+            break;
+        if (verbose)
+            printf("%lu next %" PRIu64 "\n", (unsigned long) toku_pthread_self(), get_key(&context.key));
+        usleep(sleeptime);
+    }
+
+    if (verbose)
+        printf("%lu next=%d\n", (unsigned long) toku_pthread_self(), r);
+    assert(r == DB_NOTFOUND || r == DB_LOCK_DEADLOCK);
+
+    int rr = cursor->c_close(cursor); assert(rr == 0);
+
+    if (r == DB_NOTFOUND) {
+        if (verbose) printf("%lu commit\n", (unsigned long) toku_pthread_self());
+        r = txn->commit(txn, 0); 
+    } else {
+        if (verbose) printf("%lu abort\n", (unsigned long) toku_pthread_self());
+        r = txn->abort(txn);
+    }
+    assert(r == 0);
+
+    toku_free(context.key.data);
+    toku_free(context.val.data);
+}
+
+static void blocking_prev(DB_ENV *db_env, DB *db, uint64_t nrows UU(), long sleeptime) {
+    int r;
+
+    struct my_callback_context context;
+    dbt_init_realloc(&context.key);
+    dbt_init_realloc(&context.val);
+
+    DB_TXN *txn = NULL;
+    r = db_env->txn_begin(db_env, NULL, &txn, 0); assert(r == 0);
+
+    DBC *cursor = NULL;
+    r = db->cursor(db, txn, &cursor, 0); assert(r == 0);
+
+    uint64_t i;
+    for (i = 0; ; i++) {
+        r = cursor->c_getf_prev(cursor, DB_RMW, blocking_next_callback, &context);
+        if (r != 0)
+            break;
+        if (verbose)
+            printf("%lu prev %" PRIu64 "\n", (unsigned long) toku_pthread_self(), get_key(&context.key));
+        usleep(sleeptime);
+    }
+
+    if (verbose)
+        printf("%lu prev=%d\n", (unsigned long) toku_pthread_self(), r);
+    assert(r == DB_NOTFOUND || r == DB_LOCK_DEADLOCK);
+
+    int rr = cursor->c_close(cursor); assert(rr == 0);
+
+    if (r == DB_NOTFOUND) {
+        if (verbose) printf("%lu commit\n", (unsigned long) toku_pthread_self());
+        r = txn->commit(txn, 0); 
+    } else {
+        if (verbose) printf("%lu abort\n", (unsigned long) toku_pthread_self());
+        r = txn->abort(txn);
+    }
+    assert(r == 0);
+
+    toku_free(context.key.data);
+    toku_free(context.val.data);
+}
+
+struct blocking_next_args {
+    DB_ENV *db_env;
+    DB *db;
+    uint64_t nrows;
+    long sleeptime;
+};
+
+static void *blocking_next_thread(void *arg) {
+    struct blocking_next_args *a = (struct blocking_next_args *) arg;
+    blocking_next(a->db_env, a->db, a->nrows, a->sleeptime);
+    return arg;
+}
+
+static void run_test(DB_ENV *db_env, DB *db, int nthreads, uint64_t nrows, long sleeptime) {
+    int r;
+    toku_pthread_t tids[nthreads];
+    struct blocking_next_args a = {db_env, db, nrows, sleeptime};
+    for (int i = 0; i < nthreads - 1; i++) {
+        r = toku_pthread_create(
+            toku_uninstrumented, &tids[i], nullptr, blocking_next_thread, &a);
+        assert(r == 0);
+    }
+    blocking_prev(db_env, db, nrows, sleeptime);
+    for (int i = 0; i < nthreads - 1; i++) {
+        void *ret;
+        r = toku_pthread_join(tids[i], &ret); assert(r == 0);
+    }
+}
+
+int test_main(int argc, char * const argv[]) {
+    uint64_t cachesize = 0;
+    uint32_t pagesize = 0;
+    uint64_t nrows = 10;
+    int nthreads = 2;
+    long sleeptime = 100000;
+    const char *db_env_dir = TOKU_TEST_FILENAME;
+    const char *db_filename = "test.db";
+    int db_env_open_flags = DB_CREATE | DB_PRIVATE | DB_INIT_MPOOL | DB_INIT_TXN | DB_INIT_LOCK | DB_INIT_LOG | DB_THREAD;
+
+    // parse_args(argc, argv);
+    for (int i = 1; i < argc; i++) {
+        if (strcmp(argv[i], "-v") == 0 || strcmp(argv[i], "--verbose") == 0) {
+            verbose++;
+            continue;
+        }
+        if (strcmp(argv[i], "-q") == 0 || strcmp(argv[i], "--quiet") == 0) {
+            if (verbose > 0)
+                verbose--;
+            continue;
+        }
+        if (strcmp(argv[i], "--nrows") == 0 && i+1 < argc) {
+            nrows = atoll(argv[++i]);
+            continue;
+        }
+        if (strcmp(argv[i], "--nthreads") == 0 && i+1 < argc) {
+            nthreads = atoi(argv[++i]);
+            continue;
+        }
+        if (strcmp(argv[i], "--sleeptime") == 0 && i+1 < argc) {
+            sleeptime = atol(argv[++i]);
+            continue;
+        }
+        assert(0);
+    }
+
+    // setup env
+    int r;
+    char rm_cmd[strlen(db_env_dir) + strlen("rm -rf ") + 1];
+    snprintf(rm_cmd, sizeof(rm_cmd), "rm -rf %s", db_env_dir);
+    r = system(rm_cmd); assert(r == 0);
+
+    r = toku_os_mkdir(db_env_dir, S_IRWXU | S_IRGRP | S_IXGRP | S_IROTH | S_IXOTH); assert(r == 0);
+
+    DB_ENV *db_env = NULL;
+    r = db_env_create(&db_env, 0); assert(r == 0);
+    if (cachesize) {
+        const uint64_t gig = 1 << 30;
+        r = db_env->set_cachesize(db_env, cachesize / gig, cachesize % gig, 1); assert(r == 0);
+    }
+    r = db_env->open(db_env, db_env_dir, db_env_open_flags, S_IRUSR | S_IWUSR | S_IRGRP | S_IROTH); assert(r == 0);
+    r = db_env->set_lock_timeout(db_env, 30 * 1000, nullptr); assert(r == 0);
+
+    // create the db
+    DB *db = NULL;
+    r = db_create(&db, db_env, 0); assert(r == 0);
+    if (pagesize) {
+        r = db->set_pagesize(db, pagesize); assert(r == 0);
+    }
+    r = db->open(db, NULL, db_filename, NULL, DB_BTREE, DB_CREATE|DB_AUTO_COMMIT|DB_THREAD, S_IRUSR | S_IWUSR | S_IRGRP | S_IROTH); assert(r == 0);
+
+    // populate the db
+    populate(db_env, db, nrows);
+
+    run_test(db_env, db, nthreads, nrows, sleeptime);
+
+    // close env
+    r = db->close(db, 0); assert(r == 0); db = NULL;
+    r = db_env->close(db_env, 0); assert(r == 0); db_env = NULL;
+
+    return 0;
+}