Fix logic bug in optimizer, and fix up DEBUG_DEPS compile.
[ccan] / ccan / tdb / tools / replay_trace.c
index 4af09dc337d7475a213486a0fc5288311f375ef9..db5a07be1cf7c7dd4b3cd7ae18a5770ca1d5ca63 100644 (file)
@@ -1,27 +1,58 @@
 #include <ccan/tdb/tdb.h>
 #include <ccan/grab_file/grab_file.h>
+#include <ccan/hash/hash.h>
 #include <ccan/talloc/talloc.h>
 #include <ccan/str_talloc/str_talloc.h>
 #include <ccan/str/str.h>
+#include <ccan/list/list.h>
 #include <err.h>
 #include <ctype.h>
+#include <string.h>
+#include <unistd.h>
+#include <sys/types.h>
+#include <sys/wait.h>
 #include <sys/time.h>
+#include <errno.h>
+#include <signal.h>
+#include <assert.h>
 
 #define STRINGIFY2(x) #x
 #define STRINGIFY(x) STRINGIFY2(x)
 
+/* Avoid mod by zero */
+static unsigned int total_keys = 1;
+
+/* #define DEBUG_DEPS 1 */
+
+/* Traversals block transactions in the current implementation. */
+#define TRAVERSALS_TAKE_TRANSACTION_LOCK 1
+
+struct pipe {
+       int fd[2];
+};
+static struct pipe *pipes;
+
+static void __attribute__((noreturn)) fail(const char *filename,
+                                          unsigned int line,
+                                          const char *fmt, ...)
+{
+       va_list ap;
+
+       va_start(ap, fmt);
+       fprintf(stderr, "%s:%u: FAIL: ", filename, line);
+       vfprintf(stderr, fmt, ap);
+       fprintf(stderr, "\n");
+       va_end(ap);
+       exit(1);
+}
+       
 /* Try or die. */
-#define try(expr, op)                                                  \
+#define try(expr, expect)                                              \
        do {                                                            \
                int ret = (expr);                                       \
-               if (ret < 0) {                                          \
-                       if (tdb_error(tdb) != -op.ret)                  \
-                               errx(1, "Line %u: " STRINGIFY(expr)     \
-                                    "= %i: %s",                        \
-                                    i+1, ret, tdb_errorstr(tdb));      \
-               } else if (ret != op.ret)                               \
-                       errx(1, "Line %u: " STRINGIFY(expr) "= %i: %s", \
-                            i+1, ret, tdb_errorstr(tdb));              \
+               if (ret != (expect))                                    \
+                       fail(filename[file], i+1,                       \
+                            STRINGIFY(expr) "= %i", ret);              \
        } while (0)
 
 /* Try or imitate results. */
@@ -29,8 +60,9 @@
        do {                                                            \
                int ret = expr;                                         \
                if (ret != expect) {                                    \
-                       warnx("Line %u: %s gave %i not %i",             \
-                             i+1, STRINGIFY(expr), ret, expect);       \
+                       fprintf(stderr, "%s:%u: %s gave %i not %i",     \
+                               filename[file], i+1, STRINGIFY(expr),   \
+                               ret, expect);                           \
                        if (expect == 0)                                \
                                force;                                  \
                        else                                            \
                }                                                       \
        } while (0)
 
+static bool key_eq(TDB_DATA a, TDB_DATA b)
+{
+       if (a.dsize != b.dsize)
+               return false;
+       return memcmp(a.dptr, b.dptr, a.dsize) == 0;
+}
+
+/* This is based on the hash algorithm from gdbm */
+static unsigned int hash_key(TDB_DATA *key)
+{
+       uint32_t value; /* Used to compute the hash value.  */
+       uint32_t   i;   /* Used to cycle through random values. */
+
+       /* Set the initial value from the key size. */
+       for (value = 0x238F13AF ^ key->dsize, i=0; i < key->dsize; i++)
+               value = (value + (key->dptr[i] << (i*5 % 24)));
+
+       return (1103515243 * value + 12345);  
+}
+
 enum op_type {
        OP_TDB_LOCKALL,
        OP_TDB_LOCKALL_MARK,
@@ -54,7 +106,6 @@ enum op_type {
        OP_TDB_CHAINUNLOCK,
        OP_TDB_CHAINLOCK_READ,
        OP_TDB_CHAINUNLOCK_READ,
-       OP_TDB_INCREMENT_SEQNUM_NONBLOCK,
        OP_TDB_PARSE_RECORD,
        OP_TDB_EXISTS,
        OP_TDB_STORE,
@@ -72,65 +123,409 @@ enum op_type {
        OP_TDB_NEXTKEY,
        OP_TDB_FETCH,
        OP_TDB_DELETE,
-       OP_TDB_CLOSE,
 };
 
 struct op {
+       unsigned int serial;
        enum op_type op;
        TDB_DATA key;
        TDB_DATA data;
-       int flag;
        int ret;
+
+       /* Who is waiting for us? */
+       struct list_head post;
+       /* What are we waiting for? */
+       struct list_head pre;
+
+       /* If I'm part of a group (traverse/transaction) where is
+        * start?  (Otherwise, 0) */
+       unsigned int group_start;
+
+       union {
+               int flag; /* open and store */
+               struct traverse *trav; /* traverse start */
+               struct {  /* append */
+                       TDB_DATA pre;
+                       TDB_DATA post;
+               } append;
+               unsigned int transaction_end; /* transaction start */
+       };
 };
 
-static unsigned char hex_char(unsigned int line, char c)
+static unsigned char hex_char(const char *filename, unsigned int line, char c)
 {
        c = toupper(c);
        if (c >= 'A' && c <= 'F')
                return c - 'A' + 10;
        if (c >= '0' && c <= '9')
                return c - '0';
-       errx(1, "Line %u: invalid hex character '%c'", line, c);
+       fail(filename, line, "invalid hex character '%c'", c);
 }
 
 /* TDB data is <size>:<%02x>* */
 static TDB_DATA make_tdb_data(const void *ctx,
-                             unsigned int line, const char *word)
+                             const char *filename, unsigned int line,
+                             const char *word)
 {
        TDB_DATA data;
        unsigned int i;
        const char *p;
 
+       if (streq(word, "NULL"))
+               return tdb_null;
+
        data.dsize = atoi(word);
        data.dptr = talloc_array(ctx, unsigned char, data.dsize);
        p = strchr(word, ':');
        if (!p)
-               errx(1, "Line %u: Invalid tdb data '%s'", line, word);
+               fail(filename, line, "invalid tdb data '%s'", word);
        p++;
        for (i = 0; i < data.dsize; i++)
-               data.dptr[i] = hex_char(line, p[i*2])*16
-                       + hex_char(line, p[i*2+1]);
+               data.dptr[i] = hex_char(filename, line, p[i*2])*16
+                       + hex_char(filename, line, p[i*2+1]);
+
        return data;
 }
 
-static void add_op(struct op **op, unsigned int i,
-                  enum op_type type, const char *key, const char *data,
-                  int flag, int ret)
+static void add_op(const char *filename, struct op **op, unsigned int i,
+                  unsigned int serial, enum op_type type)
 {
        struct op *new;
        *op = talloc_realloc(NULL, *op, struct op, i+1);
        new = (*op) + i;
        new->op = type;
-       new->flag = flag;
-       new->ret = ret;
-       if (key)
-               new->key = make_tdb_data(*op, i+1, key);
-       else
-               new->key = tdb_null;
-       if (data)
-               new->data = make_tdb_data(*op, i+1, data);
-       else
-               new->data = tdb_null;
+       new->serial = serial;
+       new->ret = 0;
+       new->group_start = 0;
+}
+
+static void op_add_nothing(const char *filename,
+                          struct op op[], unsigned int op_num, char *words[])
+{
+       if (words[2])
+               fail(filename, op_num+1, "Expected no arguments");
+       op[op_num].key = tdb_null;
+}
+
+static void op_add_key(const char *filename,
+                      struct op op[], unsigned int op_num, char *words[])
+{
+       if (words[2] == NULL || words[3])
+               fail(filename, op_num+1, "Expected just a key");
+
+       op[op_num].key = make_tdb_data(op, filename, op_num+1, words[2]);
+       if (op[op_num].op != OP_TDB_TRAVERSE)
+               total_keys++;
+}
+
+static void op_add_key_ret(const char *filename,
+                          struct op op[], unsigned int op_num, char *words[])
+{
+       if (!words[2] || !words[3] || !words[4] || words[5]
+           || !streq(words[3], "="))
+               fail(filename, op_num+1, "Expected <key> = <ret>");
+       op[op_num].ret = atoi(words[4]);
+       op[op_num].key = make_tdb_data(op, filename, op_num+1, words[2]);
+       /* May only be a unique key if it fails */
+       if (op[op_num].ret != 0)
+               total_keys++;
+}
+
+static void op_add_key_data(const char *filename,
+                           struct op op[], unsigned int op_num, char *words[])
+{
+       if (!words[2] || !words[3] || !words[4] || words[5]
+           || !streq(words[3], "="))
+               fail(filename, op_num+1, "Expected <key> = <data>");
+       op[op_num].key = make_tdb_data(op, filename, op_num+1, words[2]);
+       op[op_num].data = make_tdb_data(op, filename, op_num+1, words[4]);
+       /* May only be a unique key if it fails */
+       if (!op[op_num].data.dptr)
+               total_keys++;
+}
+
+/* <serial> tdb_store <rec> <rec> <flag> = <ret> */
+static void op_add_store(const char *filename,
+                        struct op op[], unsigned int op_num, char *words[])
+{
+       if (!words[2] || !words[3] || !words[4] || !words[5] || !words[6]
+           || words[7] || !streq(words[5], "="))
+               fail(filename, op_num+1, "Expect <key> <data> <flag> = <ret>");
+
+       op[op_num].flag = strtoul(words[4], NULL, 0);
+       op[op_num].ret = atoi(words[6]);
+       op[op_num].key = make_tdb_data(op, filename, op_num+1, words[2]);
+       op[op_num].data = make_tdb_data(op, filename, op_num+1, words[3]);
+       total_keys++;
+}
+
+/* <serial> tdb_append <rec> <rec> = <rec> */
+static void op_add_append(const char *filename,
+                         struct op op[], unsigned int op_num, char *words[])
+{
+       if (!words[2] || !words[3] || !words[4] || !words[5] || words[6]
+           || !streq(words[4], "="))
+               fail(filename, op_num+1, "Expect <key> <data> = <rec>");
+
+       op[op_num].key = make_tdb_data(op, filename, op_num+1, words[2]);
+       op[op_num].data = make_tdb_data(op, filename, op_num+1, words[3]);
+
+       op[op_num].append.post
+               = make_tdb_data(op, filename, op_num+1, words[5]);
+
+       /* By subtraction, figure out what previous data was. */
+       op[op_num].append.pre.dptr = op[op_num].append.post.dptr;
+       op[op_num].append.pre.dsize
+               = op[op_num].append.post.dsize - op[op_num].data.dsize;
+       total_keys++;
+}
+
+/* <serial> tdb_get_seqnum = <ret> */
+static void op_add_seqnum(const char *filename,
+                         struct op op[], unsigned int op_num, char *words[])
+{
+       if (!words[2] || !words[3] || words[4] || !streq(words[2], "="))
+               fail(filename, op_num+1, "Expect = <ret>");
+
+       op[op_num].key = tdb_null;
+       op[op_num].ret = atoi(words[3]);
+}
+
+static void op_add_traverse(const char *filename,
+                           struct op op[], unsigned int op_num, char *words[])
+{
+       if (words[2])
+               fail(filename, op_num+1, "Expect no arguments");
+
+       op[op_num].key = tdb_null;
+       op[op_num].trav = NULL;
+}
+
+static void op_add_transaction(const char *filename, struct op op[],
+                              unsigned int op_num, char *words[])
+{
+       if (words[2])
+               fail(filename, op_num+1, "Expect no arguments");
+
+       op[op_num].key = tdb_null;
+       op[op_num].transaction_end = 0;
+}
+
+static void op_analyze_transaction(const char *filename,
+                                  struct op op[], unsigned int op_num,
+                                  char *words[])
+{
+       int i, start;
+
+       op[op_num].key = tdb_null;
+
+       if (words[2])
+               fail(filename, op_num+1, "Expect no arguments");
+
+       for (i = op_num-1; i >= 0; i--) {
+               if (op[i].op == OP_TDB_TRANSACTION_START &&
+                   !op[i].transaction_end)
+                       break;
+       }
+
+       if (i < 0)
+               fail(filename, op_num+1, "no transaction start found");
+
+       start = i;
+       op[start].transaction_end = op_num;
+
+       /* This rolls in nested transactions.  I think that's right. */
+       for (i++; i <= op_num; i++)
+               op[i].group_start = start;
+}
+
+struct traverse_hash {
+       TDB_DATA key;
+       unsigned int index;
+};
+
+/* A traverse is a hash of keys, each one associated with ops. */
+struct traverse {
+       /* How many traversal callouts should I do? */
+       unsigned int num;
+
+       /* Where is traversal end op? */
+       unsigned int end;
+
+       /* For trivial traversals. */
+       struct traverse_hash *hash;
+};
+
+/* A trivial traversal is one which doesn't terminate early and only
+ * plays with its own record.  We can reliably replay these even if
+ * traverse order changes. */
+static bool is_trivial_traverse(struct op op[], unsigned int end)
+{
+#if 0
+       unsigned int i;
+       TDB_DATA cur = tdb_null;
+
+       if (op[end].ret != 0)
+               return false;
+
+       for (i = 0; i < end; i++) {
+               if (!op[i].key.dptr)
+                       continue;
+               if (op[i].op == OP_TDB_TRAVERSE)
+                       cur = op[i].key;
+               if (!key_eq(cur, op[i].key))
+                       return false;
+       }
+       return true;
+#endif
+       /* With multiple things happening at once, no traverse is trivial. */
+       return false;
+}
+
+static void op_analyze_traverse(const char *filename,
+                               struct op op[], unsigned int op_num,
+                               char *words[])
+{
+       int i, start;
+       struct traverse *trav = talloc(op, struct traverse);
+
+       op[op_num].key = tdb_null;
+
+       /* = %u means traverse function terminated. */
+       if (words[2]) {
+               if (!streq(words[2], "=") || !words[3] || words[4])
+                       fail(filename, op_num+1, "expect = <num>");
+               op[op_num].ret = atoi(words[3]);
+       } else
+               op[op_num].ret = 0;
+
+       trav->num = 0;
+       trav->end = op_num;
+       for (i = op_num-1; i >= 0; i--) {
+               if (op[i].op == OP_TDB_TRAVERSE)
+                       trav->num++;
+               if (op[i].op != OP_TDB_TRAVERSE_READ_START
+                   && op[i].op != OP_TDB_TRAVERSE_START)
+                       continue;
+               if (op[i].trav)
+                       continue;
+               break;
+       }
+
+       if (i < 0)
+               fail(filename, op_num+1, "no traversal start found");
+
+       start = i;
+       op[start].trav = trav;
+
+       for (i = start; i <= op_num; i++)
+               op[i].group_start = start;
+
+       if (is_trivial_traverse(op+i, op_num-i)) {
+               /* Fill in a plentiful hash table. */
+               op[start].trav->hash = talloc_zero_array(op[i].trav,
+                                                        struct traverse_hash,
+                                                        trav->num * 2);
+               for (i = start; i < op_num; i++) {
+                       unsigned int h;
+                       if (op[i].op != OP_TDB_TRAVERSE)
+                               continue;
+                       h = hash_key(&op[i].key) % (trav->num * 2);
+                       while (trav->hash[h].index)
+                               h = (h + 1) % (trav->num * 2);
+                       trav->hash[h].index = i+1;
+                       trav->hash[h].key = op[i].key;
+               }
+       } else
+               trav->hash = NULL;
+}
+
+/* Keep -Wmissing-declarations happy: */
+const struct op_table *
+find_keyword (register const char *str, register unsigned int len);
+
+#include "keywords.c"
+
+struct depend {
+       /* We can have more than one */
+       struct list_node pre_list;
+       struct list_node post_list;
+       unsigned int needs_file;
+       unsigned int needs_opnum;
+       unsigned int satisfies_file;
+       unsigned int satisfies_opnum;
+};
+
+static void check_deps(const char *filename, struct op op[], unsigned int num)
+{
+#ifdef DEBUG_DEPS
+       unsigned int i;
+
+       for (i = 1; i < num; i++)
+               if (!list_empty(&op[i].pre))
+                       fail(filename, i+1, "Still has dependencies");
+#endif
+}
+
+static void dump_pre(char *filename[], unsigned int file,
+                    struct op op[], unsigned int i)
+{
+       struct depend *dep;
+
+       printf("%s:%u still waiting for:\n", filename[file], i+1);
+       list_for_each(&op[i].pre, dep, pre_list)
+               printf("    %s:%u\n",
+                      filename[dep->satisfies_file], dep->satisfies_opnum+1);
+       check_deps(filename[file], op, i);
+}
+
+/* We simply read/write pointers, since we all are children. */
+static void do_pre(char *filename[], unsigned int file, int pre_fd,
+                  struct op op[], unsigned int i)
+{
+       while (!list_empty(&op[i].pre)) {
+               struct depend *dep;
+
+#if DEBUG_DEPS
+               printf("%s:%u:waiting for pre\n", filename[file], i+1);
+               fflush(stdout);
+#endif
+               alarm(10);
+               while (read(pre_fd, &dep, sizeof(dep)) != sizeof(dep)) {
+                       if (errno == EINTR) {
+                               dump_pre(filename, file, op, i);
+                               exit(1);
+                       } else
+                               errx(1, "Reading from pipe");
+               }
+               alarm(0);
+
+#if DEBUG_DEPS
+               printf("%s:%u:got pre %u from %s:%u\n", filename[file], i+1,
+                      dep->needs_opnum+1, filename[dep->satisfies_file],
+                      dep->satisfies_opnum+1);
+               fflush(stdout);
+#endif
+               /* This could be any op, not just this one. */
+               talloc_free(dep);
+       }
+}
+
+static void do_post(char *filename[], unsigned int file,
+                   const struct op op[], unsigned int i)
+{
+       struct depend *dep;
+
+       list_for_each(&op[i].post, dep, post_list) {
+#if DEBUG_DEPS
+               printf("%s:%u:sending to file %s:%u\n", filename[file], i+1,
+                      filename[dep->needs_file], dep->needs_opnum+1);
+#endif
+               if (write(pipes[dep->needs_file].fd[1], &dep, sizeof(dep))
+                   != sizeof(dep))
+                       err(1, "%s:%u failed to tell file %s",
+                           filename[file], i+1, filename[dep->needs_file]);
+       }
 }
 
 static int get_len(TDB_DATA key, TDB_DATA data, void *private_data)
@@ -138,99 +533,153 @@ static int get_len(TDB_DATA key, TDB_DATA data, void *private_data)
        return data.dsize;
 }
 
+static unsigned run_ops(struct tdb_context *tdb,
+                       int pre_fd,
+                       char *filename[],
+                       unsigned int file,
+                       struct op op[],
+                       unsigned int start, unsigned int stop);
+
 struct traverse_info {
-       const struct op *op;
-       unsigned int start_line;
+       struct op *op;
+       char **filename;
+       unsigned file;
+       int pre_fd;
+       unsigned int start;
        unsigned int i;
-       unsigned int num;
 };
 
-static unsigned run_ops(struct tdb_context *tdb, const struct op op[],
-                       unsigned int start, unsigned int stop);
+/* Trivial case: do whatever they did for this key. */
+static int trivial_traverse(struct tdb_context *tdb,
+                           TDB_DATA key, TDB_DATA data,
+                           void *_tinfo)
+{
+       struct traverse_info *tinfo = _tinfo;
+       struct traverse *trav = tinfo->op[tinfo->start].trav;
+       unsigned int h = hash_key(&key) % (trav->num * 2);
 
-static int traverse(struct tdb_context *tdb, TDB_DATA key, TDB_DATA data,
-                   void *_tinfo)
+       while (trav->hash[h].index) {
+               if (key_eq(trav->hash[h].key, key)) {
+                       run_ops(tdb, tinfo->pre_fd, tinfo->filename,
+                               tinfo->file, tinfo->op, trav->hash[h].index,
+                               trav->end);
+                       tinfo->i++;
+                       return 0;
+               }
+               h = (h + 1) % (trav->num * 2);
+       }
+       fail(tinfo->filename[tinfo->file], tinfo->start + 1,
+            "unexpected traverse key");
+}
+
+/* More complex.  Just do whatever's they did at the n'th entry. */
+static int nontrivial_traverse(struct tdb_context *tdb,
+                              TDB_DATA key, TDB_DATA data,
+                              void *_tinfo)
 {
        struct traverse_info *tinfo = _tinfo;
+       struct traverse *trav = tinfo->op[tinfo->start].trav;
 
-       if (tinfo->i == tinfo->num)
-               errx(1, "Transaction starting line %u did not terminate",
-                    tinfo->start_line);
+       if (tinfo->i == trav->end) {
+               /* This can happen if traverse expects to be empty. */
+               if (tinfo->start + 1 == trav->end)
+                       return 1;
+               fail(tinfo->filename[tinfo->file], tinfo->start + 1,
+                    "traverse did not terminate");
+       }
 
        if (tinfo->op[tinfo->i].op != OP_TDB_TRAVERSE)
-               errx(1, "Transaction starting line %u terminatd early",
-                    tinfo->start_line);
-
-       /* Check we have right key. */
-       if (key.dsize != tinfo->op[tinfo->i].key.dsize
-           || memcmp(key.dptr, tinfo->op[tinfo->i].key.dptr, key.dsize))
-               errx(1, "Line %u: bad traverse key", tinfo->i+1);
-       if (data.dsize != tinfo->op[tinfo->i].data.dsize
-           || memcmp(data.dptr, tinfo->op[tinfo->i].data.dptr, data.dsize))
-               errx(1, "Line %u: bad traverse data", tinfo->i+1);
-
-       tinfo->i++;
+               fail(tinfo->filename[tinfo->file], tinfo->start + 1,
+                    "%s:%u:traverse terminated early");
+
        /* Run any normal ops. */
-       tinfo->i = run_ops(tdb, tinfo->op, tinfo->i, tinfo->num);
+       tinfo->i = run_ops(tdb, tinfo->pre_fd, tinfo->filename, tinfo->file,
+                          tinfo->op, tinfo->i+1, trav->end);
 
-       if (tinfo->op[tinfo->i].op == OP_TDB_TRAVERSE_END)
+       if (tinfo->i == trav->end)
                return 1;
+
        return 0;
 }
 
 static unsigned op_traverse(struct tdb_context *tdb,
-                           const struct op op[],
-                           unsigned int line,
-                           unsigned int num)
+                           int pre_fd,
+                           char *filename[],
+                           unsigned int file,
+                           int (*traversefn)(struct tdb_context *,
+                                             tdb_traverse_func, void *),
+                           struct op op[],
+                           unsigned int start)
 {
-       struct traverse_info tinfo = { op, line, line, num };
+       struct traverse *trav = op[start].trav;
+       struct traverse_info tinfo = { op, filename, file, pre_fd,
+                                      start, start+1 };
+
+       /* Trivial case. */
+       if (trav->hash) {
+               int ret = traversefn(tdb, trivial_traverse, &tinfo);
+               if (ret != trav->num)
+                       fail(filename[file], start+1,
+                            "short traversal %i", ret);
+               return trav->end;
+       }
+
+       traversefn(tdb, nontrivial_traverse, &tinfo);
 
-       tdb_traverse(tdb, traverse, &tinfo);
-       if (tinfo.i < num && op[tinfo.i].op != OP_TDB_TRAVERSE_END)
-               errx(1, "%u: Short traversal", line);
-       return tinfo.i;
+       /* Traversing in wrong order can have strange effects: eg. if
+        * original traverse went A (delete A), B, we might do B
+        * (delete A).  So if we have ops left over, we do it now. */
+       while (tinfo.i != trav->end) {
+               if (op[tinfo.i].op == OP_TDB_TRAVERSE)
+                       tinfo.i++;
+               else
+                       tinfo.i = run_ops(tdb, pre_fd, filename, file, op,
+                                         tinfo.i, trav->end);
+       }
+
+       return trav->end;
 }
 
-static unsigned op_read_traverse(struct tdb_context *tdb,
-                                const struct op op[],
-                                unsigned int line,
-                                unsigned int num)
+static void break_out(int sig)
 {
-       struct traverse_info tinfo = { op, line, line, num };
-
-       tdb_traverse_read(tdb, traverse, &tinfo);
-       if (tinfo.i < num && op[tinfo.i].op != OP_TDB_TRAVERSE_END)
-               errx(1, "%u: Short traversal", line);
-       return tinfo.i;
 }
 
 static __attribute__((noinline))
-unsigned run_ops(struct tdb_context *tdb, const struct op op[],
-                       unsigned int start, unsigned int stop)
+unsigned run_ops(struct tdb_context *tdb,
+                int pre_fd,
+                char *filename[],
+                unsigned int file,
+                struct op op[], unsigned int start, unsigned int stop)
 {
        unsigned int i;
-       TDB_DATA data;
+       struct sigaction sa;
 
+       sa.sa_handler = break_out;
+       sa.sa_flags = 0;
+
+       sigaction(SIGALRM, &sa, NULL);
        for (i = start; i < stop; i++) {
+               do_pre(filename, file, pre_fd, op, i);
+
                switch (op[i].op) {
                case OP_TDB_LOCKALL:
-                       try(tdb_lockall(tdb), op[i]);
+                       try(tdb_lockall(tdb), op[i].ret);
                        break;
                case OP_TDB_LOCKALL_MARK:
-                       try(tdb_lockall_mark(tdb), op[i]);
+                       try(tdb_lockall_mark(tdb), op[i].ret);
                        break;
                case OP_TDB_LOCKALL_UNMARK:
-                       try(tdb_lockall_unmark(tdb), op[i]);
+                       try(tdb_lockall_unmark(tdb), op[i].ret);
                        break;
                case OP_TDB_LOCKALL_NONBLOCK:
                        unreliable(tdb_lockall_nonblock(tdb), op[i].ret,
                                   tdb_lockall(tdb), tdb_unlockall(tdb));
                        break;
                case OP_TDB_UNLOCKALL:
-                       try(tdb_unlockall(tdb), op[i]);
+                       try(tdb_unlockall(tdb), op[i].ret);
                        break;
                case OP_TDB_LOCKALL_READ:
-                       try(tdb_lockall_read(tdb), op[i]);
+                       try(tdb_lockall_read(tdb), op[i].ret);
                        break;
                case OP_TDB_LOCKALL_READ_NONBLOCK:
                        unreliable(tdb_lockall_read_nonblock(tdb), op[i].ret,
@@ -238,10 +687,10 @@ unsigned run_ops(struct tdb_context *tdb, const struct op op[],
                                   tdb_unlockall_read(tdb));
                        break;
                case OP_TDB_UNLOCKALL_READ:
-                       try(tdb_unlockall_read(tdb), op[i]);
+                       try(tdb_unlockall_read(tdb), op[i].ret);
                        break;
                case OP_TDB_CHAINLOCK:
-                       try(tdb_chainlock(tdb, op[i].key), op[i]);
+                       try(tdb_chainlock(tdb, op[i].key), op[i].ret);
                        break;
                case OP_TDB_CHAINLOCK_NONBLOCK:
                        unreliable(tdb_chainlock_nonblock(tdb, op[i].key),
@@ -250,260 +699,824 @@ unsigned run_ops(struct tdb_context *tdb, const struct op op[],
                                   tdb_chainunlock(tdb, op[i].key));
                        break;
                case OP_TDB_CHAINLOCK_MARK:
-                       try(tdb_chainlock_mark(tdb, op[i].key), op[i]);
+                       try(tdb_chainlock_mark(tdb, op[i].key), op[i].ret);
                        break;
                case OP_TDB_CHAINLOCK_UNMARK:
-                       try(tdb_chainlock_unmark(tdb, op[i].key), op[i]);
+                       try(tdb_chainlock_unmark(tdb, op[i].key), op[i].ret);
                        break;
                case OP_TDB_CHAINUNLOCK:
-                       try(tdb_chainunlock(tdb, op[i].key), op[i]);
+                       try(tdb_chainunlock(tdb, op[i].key), op[i].ret);
                        break;
                case OP_TDB_CHAINLOCK_READ:
-                       try(tdb_chainlock_read(tdb, op[i].key), op[i]);
+                       try(tdb_chainlock_read(tdb, op[i].key), op[i].ret);
                        break;
                case OP_TDB_CHAINUNLOCK_READ:
-                       try(tdb_chainunlock_read(tdb, op[i].key), op[i]);
-                       break;
-               case OP_TDB_INCREMENT_SEQNUM_NONBLOCK:
-                       tdb_increment_seqnum_nonblock(tdb);
+                       try(tdb_chainunlock_read(tdb, op[i].key), op[i].ret);
                        break;
                case OP_TDB_PARSE_RECORD:
-                       try(tdb_parse_record(tdb, op[i].key, get_len, NULL), op[i]);
+                       try(tdb_parse_record(tdb, op[i].key, get_len, NULL),
+                           op[i].ret);
                        break;
                case OP_TDB_EXISTS:
-                       try(tdb_exists(tdb, op[i].key), op[i]);
+                       try(tdb_exists(tdb, op[i].key), op[i].ret);
                        break;
                case OP_TDB_STORE:
-                       try(tdb_store(tdb, op[i].key, op[i].data, op[i].flag), op[i]);
+                       try(tdb_store(tdb, op[i].key, op[i].data, op[i].flag),
+                           op[i].ret);
                        break;
                case OP_TDB_APPEND:
-                       try(tdb_append(tdb, op[i].key, op[i].data), op[i]);
+                       try(tdb_append(tdb, op[i].key, op[i].data), op[i].ret);
                        break;
                case OP_TDB_GET_SEQNUM:
-                       try(tdb_get_seqnum(tdb), op[i]);
+                       try(tdb_get_seqnum(tdb), op[i].ret);
                        break;
                case OP_TDB_WIPE_ALL:
-                       try(tdb_wipe_all(tdb), op[i]);
+                       try(tdb_wipe_all(tdb), op[i].ret);
                        break;
                case OP_TDB_TRANSACTION_START:
-                       try(tdb_transaction_start(tdb), op[i]);
+                       try(tdb_transaction_start(tdb), op[i].ret);
                        break;
                case OP_TDB_TRANSACTION_CANCEL:
-                       try(tdb_transaction_cancel(tdb), op[i]);
+                       try(tdb_transaction_cancel(tdb), op[i].ret);
                        break;
                case OP_TDB_TRANSACTION_COMMIT:
-                       try(tdb_transaction_commit(tdb), op[i]);
+                       try(tdb_transaction_commit(tdb), op[i].ret);
                        break;
                case OP_TDB_TRAVERSE_READ_START:
-                       i = op_read_traverse(tdb, op, i+1, stop);
+                       i = op_traverse(tdb, pre_fd, filename, file,
+                                       tdb_traverse_read, op, i);
                        break;
                case OP_TDB_TRAVERSE_START:
-                       i = op_traverse(tdb, op, i+1, stop);
+                       i = op_traverse(tdb, pre_fd, filename, file,
+                                       tdb_traverse, op, i);
                        break;
                case OP_TDB_TRAVERSE:
-               case OP_TDB_TRAVERSE_END:
-                       /* Either of these mean we're in a traversal,
-                        * finished this iteration. */
+                       /* Terminate: we're in a traverse, and we've
+                        * done our ops. */
                        return i;
+               case OP_TDB_TRAVERSE_END:
+                       fail(filename[file], i+1, "unepxected end traverse");
+               /* FIXME: These must be treated like traverse. */
                case OP_TDB_FIRSTKEY:
-                       data = tdb_firstkey(tdb);
-                       if (data.dsize != op[i].data.dsize
-                           || memcmp(data.dptr, op[i].data.dptr, data.dsize))
-                               errx(1, "Line %u: bad firstkey", i+1);
+                       if (!key_eq(tdb_firstkey(tdb), op[i].data))
+                               fail(filename[file], i+1, "bad firstkey");
                        break;
                case OP_TDB_NEXTKEY:
-                       data = tdb_nextkey(tdb, op[i].key);
-                       if (data.dsize != op[i].data.dsize
-                           || memcmp(data.dptr, op[i].data.dptr, data.dsize))
-                               errx(1, "Line %u: bad nextkey", i+1);
+                       if (!key_eq(tdb_nextkey(tdb, op[i].key), op[i].data))
+                               fail(filename[file], i+1, "bad nextkey");
                        break;
-               case OP_TDB_FETCH:
-                       data = tdb_fetch(tdb, op[i].key);
-                       if (data.dsize != op[i].data.dsize
-                           || memcmp(data.dptr, op[i].data.dptr, data.dsize))
-                               errx(1, "Line %u: bad fetch", i+1);
+               case OP_TDB_FETCH: {
+                       TDB_DATA f = tdb_fetch(tdb, op[i].key);
+                       if (!key_eq(f, op[i].data))
+                               fail(filename[file], i+1, "bad fetch %u",
+                                    f.dsize);
                        break;
+               }
                case OP_TDB_DELETE:
-                       try(tdb_delete(tdb, op[i].key), op[i]);
-                       break;
-               case OP_TDB_CLOSE:
-                       errx(1, "Line %u: unexpected close", i+1);
+                       try(tdb_delete(tdb, op[i].key), op[i].ret);
                        break;
                }
+               do_post(filename, file, op, i);
        }
        return i;
 }
 
-int main(int argc, char *argv[])
+static struct op *load_tracefile(const char *filename, unsigned int *num,
+                                unsigned int *hashsize,
+                                unsigned int *tdb_flags,
+                                unsigned int *open_flags)
 {
-       const char *file;
-       char **lines;
        unsigned int i;
-       struct tdb_context *tdb = NULL;
        struct op *op = talloc_array(NULL, struct op, 1);
-       struct timeval start, end;
-
-       if (argc != 3)
-               errx(1, "Usage: %s <tracefile> <tdbfile>", argv[0]);
+       char **words;
+       char **lines;
+       char *file;
 
-       file = grab_file(NULL, argv[1], NULL);
+       file = grab_file(NULL, filename, NULL);
        if (!file)
-               err(1, "Reading %s", argv[1]);
+               err(1, "Reading %s", filename);
 
        lines = strsplit(file, file, "\n", NULL);
+       if (!lines[0])
+               errx(1, "%s is empty", filename);
 
-       for (i = 0; lines[i]; i++) {
-               char **words = strsplit(lines, lines[i], " ", NULL);
-               if (!tdb && !streq(words[0], "tdb_open"))
-                       errx(1, "Line %u is not tdb_open", i+1);
-
-               if (streq(words[0], "tdb_open")) {
-                       if (tdb)
-                               errx(1, "Line %u: tdb_open again?", i+1);
-                       tdb = tdb_open(argv[2], atoi(words[2]),
-                                      strtoul(words[3], NULL, 0),
-                                      strtoul(words[4], NULL, 0), 0600);
-                       if (!tdb)
-                               err(1, "Opening tdb %s", argv[2]);
-               } else if (streq(words[0], "tdb_lockall")) {
-                       add_op(&op, i, OP_TDB_LOCKALL, NULL, NULL, 0, 0);
-               } else if (streq(words[0], "tdb_lockall_mark")) {
-                       add_op(&op, i, OP_TDB_LOCKALL_MARK, NULL, NULL, 0, 0);
-               } else if (streq(words[0], "tdb_lockall_unmark")) {
-                       add_op(&op, i, OP_TDB_LOCKALL_UNMARK, NULL, NULL, 0, 0);
-               } else if (streq(words[0], "tdb_lockall_nonblock")) {
-                       add_op(&op, i, OP_TDB_LOCKALL_NONBLOCK, NULL, NULL, 0,
-                              atoi(words[1]));
-               } else if (streq(words[0], "tdb_unlockall")) {
-                       add_op(&op, i, OP_TDB_UNLOCKALL, NULL, NULL, 0, 0);
-               } else if (streq(words[0], "tdb_lockall_read")) {
-                       add_op(&op, i, OP_TDB_LOCKALL_READ, NULL, NULL, 0, 0);
-               } else if (streq(words[0], "tdb_lockall_read_nonblock")) {
-                       add_op(&op, i, OP_TDB_LOCKALL_READ_NONBLOCK, NULL, NULL,
-                              0, atoi(words[1]));
-               } else if (streq(words[0], "tdb_unlockall_read\n")) {
-                       add_op(&op, i, OP_TDB_UNLOCKALL_READ, NULL, NULL, 0, 0);
-               } else if (streq(words[0], "tdb_chainlock")) {
-                       add_op(&op, i, OP_TDB_CHAINLOCK, words[1], NULL, 0, 0);
-               } else if (streq(words[0], "tdb_chainlock_nonblock")) {
-                       add_op(&op, i, OP_TDB_CHAINLOCK_NONBLOCK,
-                              words[1], NULL, 0, atoi(words[3]));
-               } else if (streq(words[0], "tdb_chainlock_mark")) {
-                       add_op(&op, i, OP_TDB_CHAINLOCK_MARK, words[1], NULL,
-                              0, 0);
-               } else if (streq(words[0], "tdb_chainlock_unmark")) {
-                       add_op(&op, i, OP_TDB_CHAINLOCK_UNMARK, words[1], NULL,
-                              0, 0);
-               } else if (streq(words[0], "tdb_chainunlock")) {
-                       add_op(&op, i, OP_TDB_CHAINUNLOCK, words[1], NULL,
-                              0, 0);
-               } else if (streq(words[0], "tdb_chainlock_read")) {
-                       add_op(&op, i, OP_TDB_CHAINLOCK_READ, words[1],
-                              NULL, 0, 0);
-               } else if (streq(words[0], "tdb_chainunlock_read")) {
-                       add_op(&op, i, OP_TDB_CHAINUNLOCK_READ, words[1],
-                              NULL, 0, 0);
-               } else if (streq(words[0], "tdb_close")) {
-                       add_op(&op, i, OP_TDB_CLOSE, NULL, NULL, 0, 0);
-               } else if (streq(words[0], "tdb_increment_seqnum_nonblock")) {
-                       add_op(&op, i, OP_TDB_INCREMENT_SEQNUM_NONBLOCK,
-                              NULL, NULL, 0, 0);
-               } else if (streq(words[0], "tdb_fetch")) {
-                       if (streq(words[3], "ENOENT"))
-                               add_op(&op, i, OP_TDB_FETCH, words[1], NULL, 0,
-                                      -TDB_ERR_NOEXIST);
-                       else
-                               add_op(&op, i, OP_TDB_FETCH, words[1], words[3],
-                                      0, 0);
-               } else if (streq(words[0], "tdb_parse_record")) {
-                       if (streq(words[3], "ENOENT"))
-                               add_op(&op, i, OP_TDB_PARSE_RECORD,
-                                      words[1], NULL, 0, -TDB_ERR_NOEXIST);
-                       else
-                               add_op(&op, i, OP_TDB_PARSE_RECORD,
-                                      words[1], NULL, 0, atoi(words[3]));
-               } else if (streq(words[0], "tdb_exists")) {
-                       add_op(&op, i, OP_TDB_EXISTS, words[1], NULL, 0,
-                              atoi(words[3]));
-               } else if (streq(words[0], "tdb_delete")) {
-                       add_op(&op, i, OP_TDB_DELETE, words[1], NULL, 0,
-                              streq(words[3], "ENOENT")
-                              ? -TDB_ERR_NOEXIST : 0);
-               } else if (streq(words[0], "tdb_store")) {
-                       int flag;
-
-                       if (streq(words[1], "insert"))
-                               flag = TDB_INSERT;
-                       else if (streq(words[1], "modify"))
-                               flag = TDB_MODIFY;
-                       else if (streq(words[1], "normal"))
-                               flag = 0;
-                       else
-                               errx(1, "Line %u: invalid tdb_store", i+1);
-
-                       if (streq(words[5], "EEXIST"))
-                               add_op(&op, i, OP_TDB_STORE, words[2], words[3],
-                                      flag, -TDB_ERR_EXISTS);
-                       else if (streq(words[5], "ENOENT"))
-                               add_op(&op, i, OP_TDB_STORE, words[2], words[3],
-                                      flag, -TDB_ERR_NOEXIST);
-                       else
-                               add_op(&op, i, OP_TDB_STORE, words[2], words[3],
-                                      flag, 0);
-               } else if (streq(words[0], "tdb_append")) {
-                       add_op(&op, i, OP_TDB_STORE, words[1], words[2], 0, 0);
-               } else if (streq(words[0], "tdb_get_seqnum")) {
-                       add_op(&op, i, OP_TDB_GET_SEQNUM, NULL, NULL, 0,
-                              atoi(words[2]));
-               } else if (streq(words[0], "tdb_wipe_all")) {
-                       add_op(&op, i, OP_TDB_WIPE_ALL, NULL, NULL, 0, 0);
-               } else if (streq(words[0], "tdb_transaction_start")) {
-                       add_op(&op, i, OP_TDB_TRANSACTION_START, NULL, NULL,
-                              0, 0);
-               } else if (streq(words[0], "tdb_transaction_cancel")) {
-                       add_op(&op, i, OP_TDB_TRANSACTION_CANCEL, NULL, NULL,
-                              0, 0);
-               } else if (streq(words[0], "tdb_transaction_commit")) {
-                       add_op(&op, i, OP_TDB_TRANSACTION_COMMIT, NULL, NULL,
-                              0, 0);
-               } else if (streq(words[0], "tdb_traverse_read_start")) {
-                       add_op(&op, i, OP_TDB_TRAVERSE_READ_START, NULL, NULL,
-                              0, 0);
-               } else if (streq(words[0], "tdb_traverse_start")) {
-                       add_op(&op, i, OP_TDB_TRAVERSE_START, NULL, NULL,
-                              0, 0);
-               } else if (streq(words[0], "tdb_traverse_end")) {
-                       add_op(&op, i, OP_TDB_TRAVERSE_END, NULL, NULL,
-                              0, atoi(words[2]));
-               } else if (streq(words[0], "traverse")) {
-                       add_op(&op, i, OP_TDB_TRAVERSE, words[1], words[2],
-                              0, 0);
-               } else if (streq(words[0], "tdb_firstkey")) {
-                       if (streq(words[2], "ENOENT"))
-                               add_op(&op, i, OP_TDB_FIRSTKEY, NULL, NULL,
-                                      0, -TDB_ERR_NOEXIST);
+       words = strsplit(lines, lines[0], " ", NULL);
+       if (!streq(words[1], "tdb_open"))
+               fail(filename, 1, "does not start with tdb_open");
+
+       *hashsize = atoi(words[2]);
+       *tdb_flags = strtoul(words[3], NULL, 0);
+       *open_flags = strtoul(words[4], NULL, 0);
+
+       for (i = 1; lines[i]; i++) {
+               const struct op_table *opt;
+
+               words = strsplit(lines, lines[i], " ", NULL);
+               if (!words[0] || !words[1])
+                       fail(filename, i+1, "Expected serial number and op");
+              
+               opt = find_keyword(words[1], strlen(words[1]));
+               if (!opt) {
+                       if (streq(words[1], "tdb_close")) {
+                               if (lines[i+1])
+                                       fail(filename, i+2,
+                                            "lines after tdb_close");
+                               *num = i;
+                               talloc_free(lines);
+                               return op;
+                       }
+                       fail(filename, i+1, "Unknown operation '%s'", words[1]);
+               }
+
+               add_op(filename, &op, i, atoi(words[0]), opt->type);
+               opt->enhance_op(filename, op, i, words);
+       }
+
+       fprintf(stderr, "%s:%u:last operation is not tdb_close: incomplete?",
+             filename, i);
+       talloc_free(lines);
+       *num = i - 1;
+       return op;
+}
+
+/* We remember all the keys we've ever seen, and who has them. */
+struct key_user {
+       unsigned int file;
+       unsigned int op_num;
+};
+
+struct keyinfo {
+       TDB_DATA key;
+       unsigned int num_users;
+       struct key_user *user;
+};
+
+static const TDB_DATA must_not_exist;
+static const TDB_DATA must_exist;
+static const TDB_DATA not_exists_or_empty;
+
+/* NULL means doesn't care if it exists or not, &must_exist means
+ * it must exist but we don't care what, &must_not_exist means it must
+ * not exist, otherwise the data it needs. */
+static const TDB_DATA *needs(const struct op *op)
+{
+       switch (op->op) {
+       /* FIXME: Pull forward deps, since we can deadlock */
+       case OP_TDB_CHAINLOCK:
+       case OP_TDB_CHAINLOCK_NONBLOCK:
+       case OP_TDB_CHAINLOCK_MARK:
+       case OP_TDB_CHAINLOCK_UNMARK:
+       case OP_TDB_CHAINUNLOCK:
+       case OP_TDB_CHAINLOCK_READ:
+       case OP_TDB_CHAINUNLOCK_READ:
+               return NULL;
+
+       case OP_TDB_APPEND:
+               if (op->append.pre.dsize == 0)
+                       return &not_exists_or_empty;
+               return &op->append.pre;
+
+       case OP_TDB_STORE:
+               if (op->flag == TDB_INSERT) {
+                       if (op->ret < 0)
+                               return &must_exist;
                        else
-                               add_op(&op, i, OP_TDB_FIRSTKEY, NULL, words[2],
-                                      0, 0);
-               } else if (streq(words[0], "tdb_nextkey")) {
-                       if (streq(words[3], "ENOENT"))
-                               add_op(&op, i, OP_TDB_NEXTKEY, words[1], NULL,
-                                      0, -TDB_ERR_NOEXIST);
+                               return &must_not_exist;
+               } else if (op->flag == TDB_MODIFY) {
+                       if (op->ret < 0)
+                               return &must_not_exist;
                        else
-                               add_op(&op, i, OP_TDB_NEXTKEY,
-                                      words[1], words[3], 0, 0);
-               } else
-                       errx(1, "Line %u: unknown op '%s'", i+1, words[0]);
+                               return &must_exist;
+               }
+               /* No flags?  Don't care */
+               return NULL;
+
+       case OP_TDB_EXISTS:
+               if (op->ret == 1)
+                       return &must_exist;
+               else
+                       return &must_not_exist;
+
+       case OP_TDB_PARSE_RECORD:
+               if (op->ret < 0)
+                       return &must_not_exist;
+               return &must_exist;
+
+       /* FIXME: handle these. */
+       case OP_TDB_WIPE_ALL:
+       case OP_TDB_FIRSTKEY:
+       case OP_TDB_NEXTKEY:
+       case OP_TDB_GET_SEQNUM:
+       case OP_TDB_TRAVERSE:
+       case OP_TDB_TRANSACTION_COMMIT:
+       case OP_TDB_TRANSACTION_CANCEL:
+       case OP_TDB_TRANSACTION_START:
+               return NULL;
+
+       case OP_TDB_FETCH:
+               if (!op->data.dptr)
+                       return &must_not_exist;
+               return &op->data;
+
+       case OP_TDB_DELETE:
+               if (op->ret < 0)
+                       return &must_not_exist;
+               return &must_exist;
+
+       default:
+               errx(1, "Unexpected op %i", op->op);
+       }
+       
+}
+
+/* What's the data after this op?  pre if nothing changed. */
+static const TDB_DATA *gives(const struct op *op, const TDB_DATA *pre)
+{
+       /* Failed ops don't change state of db. */
+       if (op->ret < 0)
+               return pre;
+
+       if (op->op == OP_TDB_DELETE || op->op == OP_TDB_WIPE_ALL)
+               return &tdb_null;
+
+       if (op->op == OP_TDB_APPEND)
+               return &op->append.post;
+
+       if (op->op == OP_TDB_STORE)
+               return &op->data;
+
+       return pre;
+}
+
+static struct keyinfo *hash_ops(struct op *op[], unsigned int num_ops[],
+                               unsigned int num)
+{
+       unsigned int i, j, h;
+       struct keyinfo *hash;
+
+       hash = talloc_zero_array(op[0], struct keyinfo, total_keys*2);
+       for (i = 0; i < num; i++) {
+               for (j = 1; j < num_ops[i]; j++) {
+                       /* We can't do this on allocation, due to realloc. */
+                       list_head_init(&op[i][j].post);
+                       list_head_init(&op[i][j].pre);
+
+                       if (!op[i][j].key.dptr)
+                               continue;
+
+                       /* We don't wait for traverse keys */
+                       /* FIXME: We should, for trivial traversals. */
+                       if (op[i][j].op == OP_TDB_TRAVERSE)
+                               continue;
+
+                       h = hash_key(&op[i][j].key) % (total_keys * 2);
+                       while (!key_eq(hash[h].key, op[i][j].key)) {
+                               if (!hash[h].key.dptr) {
+                                       hash[h].key = op[i][j].key;
+                                       break;
+                               }
+                               h = (h + 1) % (total_keys * 2);
+                       }
+                       /* Might as well save some memory if we can. */
+                       if (op[i][j].key.dptr != hash[h].key.dptr) {
+                               talloc_free(op[i][j].key.dptr);
+                               op[i][j].key.dptr = hash[h].key.dptr;
+                       }
+                       hash[h].user = talloc_realloc(hash, hash[h].user,
+                                                    struct key_user,
+                                                    hash[h].num_users+1);
+                       hash[h].user[hash[h].num_users].op_num = j;
+                       hash[h].user[hash[h].num_users].file = i;
+                       hash[h].num_users++;
+               }
+       }
+
+       return hash;
+}
+
+static bool satisfies(const TDB_DATA *data, const TDB_DATA *need)
+{
+       /* Don't need anything?  Cool. */
+       if (!need)
+               return true;
+
+       /* This should be tdb_null or a real value. */
+       assert(data != &must_exist);
+       assert(data != &must_not_exist);
+       assert(data != &not_exists_or_empty);
+
+       /* must_not_exist == must_not_exist, must_exist == must_exist, or
+          not_exists_or_empty == not_exists_or_empty. */
+       if (data->dsize == need->dsize && data->dptr == need->dptr)
+               return true;
+
+       /* Must not exist?  data must not exist. */
+       if (need == &must_not_exist)
+               return data->dptr == NULL;
+
+       /* Must exist? */
+       if (need == &must_exist)
+               return data->dptr != NULL;
+
+       /* Either noexist or empty. */
+       if (need == &not_exists_or_empty)
+               return data->dsize == 0;
+
+       /* Needs something specific. */
+       return key_eq(*data, *need);
+}
+
+static void move_to_front(struct key_user res[], unsigned int elem)
+{
+       if (elem != 0) {
+               struct key_user tmp = res[elem];
+               memmove(res + 1, res, elem*sizeof(res[0]));
+               res[0] = tmp;
+       }
+}
+
+static void restore_to_pos(struct key_user res[], unsigned int elem)
+{
+       if (elem != 0) {
+               struct key_user tmp = res[0];
+               memmove(res, res + 1, elem*sizeof(res[0]));
+               res[elem] = tmp;
+       }
+}
+
+static bool sort_deps(char *filename[], struct op *op[],
+                     struct key_user res[], unsigned num,
+                     const TDB_DATA *data, unsigned num_files)
+{
+       unsigned int i, files_done;
+       struct op *this_op;
+       bool done[num_files];
+
+       /* Nothing left?  We're sorted. */
+       if (num == 0)
+               return true;
+
+       memset(done, 0, sizeof(done));
+
+       /* Since ops within a trace file are ordered, we just need to figure
+        * out which file to try next.  Since we don't take into account
+        * inter-key relationships (which exist by virtue of trace file order),
+        * we minimize the chance of harm by trying to keep in serial order. */
+       for (files_done = 0, i = 0; i < num && files_done < num_files; i++) {
+               if (done[res[i].file])
+                       continue;
+
+               this_op = &op[res[i].file][res[i].op_num];
+               /* Is what we have good enough for this op? */
+               if (satisfies(data, needs(this_op))) {
+                       move_to_front(res, i);
+                       if (sort_deps(filename, op, res+1, num-1,
+                                     gives(this_op, data), num_files))
+                               return true;
+                       restore_to_pos(res, i);
+               }
+               done[res[i].file] = true;
+               files_done++;
+       }
+
+       /* No combination worked. */
+       return false;
+}
+
+static void check_dep_sorting(struct key_user user[], unsigned num_users,
+                             unsigned num_files)
+{
+#if DEBUG_DEPS
+       unsigned int i;
+       unsigned minima[num_files];
+
+       memset(minima, 0, sizeof(minima));
+       for (i = 0; i < num_users; i++) {
+               assert(minima[user[i].file] < user[i].op_num);
+               minima[user[i].file] = user[i].op_num;
+       }
+#endif
+}
+
+/* All these ops have the same serial number.  Which comes first?
+ *
+ * This can happen both because read ops or failed write ops don't
+ * change serial number, and also due to race since we access the
+ * number unlocked (the race can cause less detectable ordering problems,
+ * in which case we'll deadlock and report: fix manually in that case).
+ */
+static void figure_deps(char *filename[], struct op *op[],
+                       struct key_user user[], unsigned num_users,
+                       unsigned num_files)
+{
+       /* We assume database starts empty. */
+       const struct TDB_DATA *data = &tdb_null;
+
+       if (!sort_deps(filename, op, user, num_users, data, num_files))
+               fail(filename[user[0].file], user[0].op_num+1,
+                    "Could not resolve inter-dependencies");
+
+       check_dep_sorting(user, num_users, num_files);
+}
+
+static void sort_ops(struct keyinfo hash[], char *filename[], struct op *op[],
+                    unsigned int num)
+{
+       unsigned int h;
+
+       /* Gcc nexted function extension.  How cool is this? */
+       int compare_serial(const void *_a, const void *_b)
+       {
+               const struct key_user *a = _a, *b = _b;
+
+               /* First, maintain order within any trace file. */
+               if (a->file == b->file)
+                       return a->op_num - b->op_num;
+
+               /* Otherwise, arrange by serial order. */
+               return op[a->file][a->op_num].serial
+                       - op[b->file][b->op_num].serial;
+       }
+
+       /* Now sort into serial order. */
+       for (h = 0; h < total_keys * 2; h++) {
+               struct key_user *user = hash[h].user;
+
+               qsort(user, hash[h].num_users, sizeof(user[0]), compare_serial);
+               figure_deps(filename, op, user, hash[h].num_users, num);
+       }
+}
+
+static int destroy_depend(struct depend *dep)
+{
+       list_del(&dep->pre_list);
+       list_del(&dep->post_list);
+       return 0;
+}
+
+static void add_dependency(void *ctx,
+                          struct op *op[],
+                          char *filename[],
+                          unsigned int needs_file,
+                          unsigned int needs_opnum,
+                          unsigned int satisfies_file,
+                          unsigned int satisfies_opnum)
+{
+       struct depend *dep;
+       unsigned int needs_start, sat_start;
+
+       /* We don't depend on ourselves. */
+       if (needs_file == satisfies_file) {
+               assert(satisfies_opnum < needs_opnum);
+               return;
+       }
+
+#if DEBUG_DEPS
+       printf("%s:%u: depends on %s:%u\n",
+              filename[needs_file], needs_opnum+1,
+              filename[satisfies_file], satisfies_opnum+1);
+#endif
+
+       needs_start = op[needs_file][needs_opnum].group_start;
+       sat_start = op[satisfies_file][satisfies_opnum].group_start;
+
+       /* If needs is in a transaction, we need it before start. */
+       if (needs_start) {
+               switch (op[needs_file][needs_start].op) {
+               case OP_TDB_TRANSACTION_START:
+                       needs_opnum = needs_start;
+#ifdef DEBUG_DEPS
+                       printf("  -> Back to %u\n", needs_start+1);
+                       fflush(stdout);
+#endif
+                       break;
+               default:
+                       break;
+               }
+       }
+
+       /* If satisfies is in a transaction, we wait until after commit. */
+       /* FIXME: If transaction is cancelled, don't need dependency. */
+       if (sat_start) {
+               if (op[satisfies_file][sat_start].op
+                   == OP_TDB_TRANSACTION_START) {
+                       satisfies_opnum
+                               = op[satisfies_file][sat_start].transaction_end;
+#ifdef DEBUG_DEPS
+                       printf("  -> Depends on %u\n", satisfies_opnum+1);
+                       fflush(stdout);
+#endif
+               }
+       }
+
+       dep = talloc(ctx, struct depend);
+       dep->needs_file = needs_file;
+       dep->needs_opnum = needs_opnum;
+       dep->satisfies_file = satisfies_file;
+       dep->satisfies_opnum = satisfies_opnum;
+       list_add(&op[satisfies_file][satisfies_opnum].post, &dep->post_list);
+       list_add(&op[needs_file][needs_opnum].pre, &dep->pre_list);
+       talloc_set_destructor(dep, destroy_depend);
+}
+
+#if TRAVERSALS_TAKE_TRANSACTION_LOCK
+struct traverse_dep {
+       unsigned int file;
+       unsigned int op_num;
+       const struct op *op;
+};
+
+/* Sort by which one runs first. */
+static int compare_traverse_dep(const void *_a, const void *_b)
+{
+       const struct traverse_dep *a = _a, *b = _b;
+       const struct traverse *trava = a->op->trav, *travb = b->op->trav;
+
+       if (a->op->serial != b->op->serial)
+               return a->op->serial - b->op->serial;
+
+       /* If they have same serial, it means one didn't make any changes.
+        * Thus sort by end in that case. */
+       return a->op[trava->end - a->op_num].serial
+               - b->op[travb->end - b->op_num].serial;
+}
+
+/* Traversals can deadlock against each other.  Force order. */
+static void make_traverse_depends(char *filename[],
+                                 struct op *op[], unsigned int num_ops[],
+                                 unsigned int num)
+{
+       unsigned int i, j, num_traversals = 0;
+       struct traverse_dep *dep;
+
+       dep = talloc_array(NULL, struct traverse_dep, 1);
+
+       /* Count them. */
+       for (i = 0; i < num; i++) {
+               for (j = 0; j < num_ops[i]; j++) {
+                       if (op[i][j].op == OP_TDB_TRAVERSE_START
+                           || op[i][j].op == OP_TDB_TRAVERSE_READ_START) {
+                               dep = talloc_realloc(NULL, dep,
+                                                    struct traverse_dep,
+                                                    num_traversals+1);
+                               dep[num_traversals].file = i;
+                               dep[num_traversals].op_num = j;
+                               dep[num_traversals].op = &op[i][j];
+                               num_traversals++;
+                       }
+               }
        }
+       qsort(dep, num_traversals, sizeof(dep[0]), compare_traverse_dep);
+       for (i = 1; i < num_traversals; i++) {
+               /* i depends on end of traverse i-1. */
+               add_dependency(NULL, op, filename, dep[i].file, dep[i].op_num,
+                              dep[i-1].file, dep[i-1].op->trav->end);
+       }
+       talloc_free(dep);
+}
+#endif /* TRAVERSALS_TAKE_TRANSACTION_LOCK */
+
+static bool changes_db(const struct op *op)
+{
+       return gives(op, NULL) != NULL;
+}
+
+static void depend_on_previous(struct op *op[],
+                              char *filename[],
+                              unsigned int num,
+                              struct key_user user[],
+                              unsigned int i,
+                              int prev)
+{
+       bool deps[num];
+       int j;
+
+       if (i == 0)
+               return;
+
+       if (prev == i - 1) {
+               /* Just depend on previous. */
+               add_dependency(NULL, op, filename,
+                              user[i].file, user[i].op_num,
+                              user[prev].file, user[prev].op_num);
+               return;
+       }
+
+       /* We have to wait for the readers.  Find last one in *each* file. */
+       memset(deps, 0, sizeof(deps));
+       deps[user[i].file] = true;
+       for (j = i - 1; j > prev; j--) {
+               if (!deps[user[j].file]) {
+                       add_dependency(NULL, op, filename,
+                                      user[i].file, user[i].op_num,
+                                      user[j].file, user[j].op_num);
+                       deps[user[j].file] = true;
+               }
+       }
+}
+
+/* This is simple, but not complete.  We don't take into account
+ * indirect dependencies. */
+static void optimize_dependencies(struct op *op[], unsigned int num_ops[],
+                                 unsigned int num)
+{
+       unsigned int i, j;
+
+       /* There can only be one real dependency on each file */
+       for (i = 0; i < num; i++) {
+               for (j = 1; j < num_ops[i]; j++) {
+                       struct depend *dep, *next;
+                       struct depend *prev[num];
+
+                       memset(prev, 0, sizeof(prev));
+
+                       list_for_each_safe(&op[i][j].pre, dep, next, pre_list) {
+                               if (!prev[dep->satisfies_file]) {
+                                       prev[dep->satisfies_file] = dep;
+                                       continue;
+                               }
+                               if (prev[dep->satisfies_file]->satisfies_opnum
+                                   < dep->satisfies_opnum) {
+                                       talloc_free(prev[dep->satisfies_file]);
+                                       prev[dep->satisfies_file] = dep;
+                               } else
+                                       talloc_free(dep);
+                       }
+               }
+       }
+
+       for (i = 0; i < num; i++) {
+               int deps[num];
 
+               for (j = 0; j < num; j++)
+                       deps[j] = -1;
+
+               for (j = 1; j < num_ops[i]; j++) {
+                       struct depend *dep, *next;
+
+                       list_for_each_safe(&op[i][j].pre, dep, next, pre_list) {
+                               if (deps[dep->satisfies_file]
+                                   >= (int)dep->satisfies_opnum)
+                                       talloc_free(dep);
+                               else
+                                       deps[dep->satisfies_file]
+                                               = dep->satisfies_opnum;
+                       }
+               }
+       }
+}
+
+static void derive_dependencies(char *filename[],
+                               struct op *op[], unsigned int num_ops[],
+                               unsigned int num)
+{
+       struct keyinfo *hash;
+       unsigned int h, i;
+
+       /* Create hash table for faster key lookup. */
+       hash = hash_ops(op, num_ops, num);
+
+       /* Sort them by serial number. */
+       sort_ops(hash, filename, op, num);
+
+       /* Create dependencies back to the last change, rather than
+        * creating false dependencies by naively making each one
+        * depend on the previous.  This has two purposes: it makes
+        * later optimization simpler, and it also avoids deadlock with
+        * same sequence number ops inside traversals (if one
+        * traversal doesn't write anything, two ops can have the same
+        * sequence number yet we can create a traversal dependency
+        * the other way). */
+       for (h = 0; h < total_keys * 2; h++) {
+               int prev = -1;
+
+               if (hash[h].num_users < 2)
+                       continue;
+
+               for (i = 0; i < hash[h].num_users; i++) {
+                       if (changes_db(&op[hash[h].user[i].file]
+                                      [hash[h].user[i].op_num])) {
+                               depend_on_previous(op, filename, num,
+                                                  hash[h].user, i, prev);
+                               prev = i;
+                       } else if (prev >= 0)
+                               add_dependency(hash, op, filename,
+                                              hash[h].user[i].file,
+                                              hash[h].user[i].op_num,
+                                              hash[h].user[prev].file,
+                                              hash[h].user[prev].op_num);
+               }
+       }
+
+#if TRAVERSALS_TAKE_TRANSACTION_LOCK
+       make_traverse_depends(filename, op, num_ops, num);
+#endif
+
+       optimize_dependencies(op, num_ops, num);
+}
+
+int main(int argc, char *argv[])
+{
+       struct timeval start, end;
+       unsigned int i, num_ops[argc], hashsize[argc], tdb_flags[argc], open_flags[argc];
+       struct op *op[argc];
+       int fds[2];
+       char c;
+       bool ok = true;
+
+       if (argc < 3)
+               errx(1, "Usage: %s <tdbfile> <tracefile>...", argv[0]);
+
+       pipes = talloc_array(NULL, struct pipe, argc - 2);
+       for (i = 0; i < argc - 2; i++) {
+               printf("Loading tracefile %s...", argv[2+i]);
+               fflush(stdout);
+               op[i] = load_tracefile(argv[2+i], &num_ops[i], &hashsize[i],
+                                      &tdb_flags[i], &open_flags[i]);
+               if (pipe(pipes[i].fd) != 0)
+                       err(1, "creating pipe");
+               printf("done\n");
+       }
+
+       printf("Calculating inter-dependencies...");
+       fflush(stdout);
+       derive_dependencies(argv+2, op, num_ops, i);
+       printf("done\n");
+
+       /* Don't fork for single arg case: simple debugging. */
+       if (argc == 3) {
+               struct tdb_context *tdb;
+               tdb = tdb_open_ex(argv[1], hashsize[0], tdb_flags[0],
+                                 open_flags[0], 0600,
+                                 NULL, hash_key);
+               printf("Single threaded run...");
+               fflush(stdout);
+
+               run_ops(tdb, pipes[0].fd[0], argv+2, 0, op[0], 1, num_ops[0]);
+               check_deps(argv[2], op[0], num_ops[0]);
+
+               printf("done\n");
+               exit(0);
+       }
+
+       if (pipe(fds) != 0)
+               err(1, "creating pipe");
+
+       for (i = 0; i < argc - 2; i++) {
+               struct tdb_context *tdb;
+
+               switch (fork()) {
+               case -1:
+                       err(1, "fork failed");
+               case 0:
+                       close(fds[1]);
+                       tdb = tdb_open_ex(argv[1], hashsize[i], tdb_flags[i],
+                                         open_flags[i], 0600,
+                                         NULL, hash_key);
+                       if (!tdb)
+                               err(1, "Opening tdb %s", argv[1]);
+
+                       /* This catches parent exiting. */
+                       if (read(fds[0], &c, 1) != 1)
+                               exit(1);
+                       run_ops(tdb, pipes[i].fd[0], argv+2, i, op[i], 1,
+                               num_ops[i]);
+                       check_deps(argv[2+i], op[i], num_ops[i]);
+                       exit(0);
+               default:
+                       break;
+               }
+       }
+
+       /* Let everything settle. */
+       sleep(1);
+
+       printf("Starting run...");
+       fflush(stdout);
        gettimeofday(&start, NULL);
-       run_ops(tdb, op, 1, i-1);
+       /* Tell them all to go!  Any write of sufficient length will do. */
+       if (write(fds[1], hashsize, i) != i)
+               err(1, "Writing to wakeup pipe");
+
+       for (i = 0; i < argc - 2; i++) {
+               int status;
+               wait(&status);
+               if (!WIFEXITED(status)) {
+                       warnx("Child died with signal %i", WTERMSIG(status));
+                       ok = false;
+               } else if (WEXITSTATUS(status) != 0)
+                       /* Assume child spat out error. */
+                       ok = false;
+       }
+       if (!ok)
+               exit(1);
+
        gettimeofday(&end, NULL);
-       if (op[i-1].op != OP_TDB_CLOSE)
-               warnx("Last operation is not tdb_close: incomplete?");
-       tdb_close(tdb);
+       printf("done\n");
+
        end.tv_sec -= start.tv_sec;
        printf("Time replaying: %lu usec\n",
               end.tv_sec * 1000000UL + (end.tv_usec - start.tv_usec));
+       
        exit(0);
 }