From: Andre Noll <maan@systemlinux.org>
Date: Mon, 26 May 2008 09:24:34 +0000 (+0200)
Subject: Merge commit 'meins/master'
X-Git-Tag: v0.0.2~42
X-Git-Url: http://git.tuebingen.mpg.de/?p=adu.git;a=commitdiff_plain;h=2cfa8c8aaa4662f44bc7e7cd4d3591ed6e1326bd;hp=3ebfb15c8a53bd1e1fbd0d5f8d4f00c7eeb76fd0

Merge commit 'meins/master'

Conflicts:

	Makefile
	adu.c
---

diff --git a/Makefile b/Makefile
index 3cd9cc4..b590b29 100644
--- a/Makefile
+++ b/Makefile
@@ -1,4 +1,4 @@
-objects := adu.o string.o
+objects := adu.o string.o cmdline.o
 all: adu
 
 DEBUG_CPPFLAGS += -Wno-sign-compare -g -Wunused -Wundef -W
@@ -21,6 +21,13 @@ Makefile.deps: $(wildcard *.c *.h)
 adu: $(objects)
 	$(CC) -o $@ $(objects) -lcrypto -losl
 
+cmdline.o: cmdline.c cmdline.h
+	$(CC) -c $(CPPFLAGS) $<
+
+cmdline.c cmdline.h: adu.ggo
+	gengetopt --conf-parser < $<
+
+
 %.o: %.c Makefile
 	$(CC) -c $(CPPFLAGS) $(DEBUG_CPPFLAGS) $<
 
diff --git a/adu.c b/adu.c
index 598919a..9edcc58 100644
--- a/adu.c
+++ b/adu.c
@@ -2,13 +2,30 @@
 #include <dirent.h> /* readdir() */
 
 #include "gcc-compat.h"
+#include "cmdline.h"
 #include "fd.h"
 #include "string.h"
 #include "error.h"
+#include "portable_io.h"
 
 DEFINE_ERRLIST;
 
+#define DATABASE_DIR "/tmp/adu"
+#define UID_LIST DATABASE_DIR "/" "uid_list"
 
+/** Command line and config file options. */
+static struct gengetopt_args_info conf;
+
+struct user_info {
+	uint32_t uid;
+	struct osl_table *table;
+	uint64_t files;
+	uint64_t bytes;
+	uint64_t dirs;
+	struct osl_table_description *desc;
+};
+
+static struct user_info *uid_hash_table;
 
 /** evaluates to 1 if x < y, to -1 if x > y and to 0 if x == y */
 #define NUM_COMPARE(x, y) ((int)((x) < (y)) - (int)((x) > (y)))
@@ -30,7 +47,7 @@ __printf_2_3 void __log(int ll, const char* fmt,...)
 	time_t t1;
 	char str[255] = "";
 
-	if (ll < 4)
+	if (ll < conf.loglevel_arg)
 		return;
 	outfd = stderr;
 	time(&t1);
@@ -74,18 +91,42 @@ static int size_compare(const struct osl_object *obj1, const struct osl_object *
  * are taken into account.
  *
  * \return It returns an integer less than, equal to, or greater than zero if
- * \a obj1 is found, respectively, to be less than, to match, or be greater than
- * obj2.
+ * \a obj1 is found, respectively, to be less than, to match, or be greater
+ * than obj2.
  *
  * \sa strcmp(3), strncmp(3), osl_compare_func.
  */
-int string_compare(const struct osl_object *obj1, const struct osl_object *obj2)
+static int string_compare(const struct osl_object *obj1,
+		const struct osl_object *obj2)
 {
 	const char *str1 = (const char *)obj1->data;
 	const char *str2 = (const char *)obj2->data;
 	return strncmp(str1, str2, MIN(obj1->size, obj2->size));
 }
 
+/**
+ * Compare two osl objects pointing to unsigned integers of 64 bit size.
+ *
+ * \param obj1 Pointer to the first integer.
+ * \param obj2 Pointer to the second integer.
+ *
+ * \return The values required for an osl compare function.
+ *
+ * \sa osl_compare_func, osl_hash_compare().
+ */
+static int uint64_compare(const struct osl_object *obj1,
+		const struct osl_object *obj2)
+{
+	uint64_t d1 = read_u64((const char *)obj1->data);
+	uint64_t d2 = read_u64((const char *)obj2->data);
+
+	if (d1 < d2)
+		return 1;
+	if (d1 > d2)
+		return -1;
+	return 0;
+}
+
 /** The columns of the directory table. */
 enum dir_table_columns {
 	/** The name of the directory. */
@@ -111,8 +152,8 @@ static struct osl_column_description dir_table_cols[] = {
 		.storage_type = OSL_MAPPED_STORAGE,
 		.storage_flags = OSL_RBTREE | OSL_FIXED_SIZE | OSL_UNIQUE,
 		.name = "num",
-		.compare_function = uint32_compare,
-		.data_size = sizeof(uint32_t)
+		.compare_function = uint64_compare,
+		.data_size = sizeof(uint64_t)
 	},
 	[DT_BYTES] = {
 		.storage_type = OSL_MAPPED_STORAGE,
@@ -135,59 +176,7 @@ static struct osl_table_description dir_table_desc = {
 	.num_columns = NUM_DT_COLUMNS,
 	.flags = 0,
 	.column_descriptions = dir_table_cols,
-	.dir = "/tmp/adu"
-};
-
-/** The columns of the id table. */
-enum id_table_columns {
-	/** The user id. */
-	IDT_UID,
-	/** The number of bytes of all regular files owned by this id. */
-	IDT_BYTES,
-	/** The number of regular files owned by this id. */
-	IDT_FILES,
-	/** The user table for this uid. */
-	IDT_TABLE,
-	/** Number of columns in this table. */
-	NUM_IDT_COLUMNS
-};
-
-static struct osl_column_description id_table_cols[] = {
-	[IDT_UID] = {
-		.storage_type = OSL_MAPPED_STORAGE,
-		.storage_flags = OSL_RBTREE | OSL_FIXED_SIZE | OSL_UNIQUE,
-		.name = "uid",
-		.compare_function = uint32_compare,
-		.data_size = sizeof(uint32_t)
-	},
-	[IDT_BYTES] = {
-		.storage_type = OSL_MAPPED_STORAGE,
-		.storage_flags = OSL_RBTREE | OSL_FIXED_SIZE,
-		.compare_function = size_compare,
-		.name = "num_bytes",
-		.data_size = sizeof(uint64_t)
-	},
-	[IDT_FILES] = {
-		.storage_type = OSL_MAPPED_STORAGE,
-		.storage_flags = OSL_RBTREE | OSL_FIXED_SIZE,
-		.compare_function = size_compare,
-		.name = "num_filess",
-		.data_size = sizeof(uint64_t)
-	},
-	[IDT_TABLE] = {
-		.storage_type = OSL_NO_STORAGE,
-		.storage_flags = OSL_FIXED_SIZE | OSL_UNIQUE,
-		.name = "user_table",
-		.data_size = sizeof(void *)
-	}
-};
-
-static struct osl_table_description id_table_desc = {
-	.name = "id_table",
-	.num_columns = NUM_IDT_COLUMNS,
-	.flags = 0,
-	.column_descriptions = id_table_cols,
-	.dir = "/tmp/adu"
+	.dir = DATABASE_DIR
 };
 
 /** The columns of the id table. */
@@ -207,17 +196,17 @@ static struct osl_column_description user_table_cols[] = {
 		.storage_type = OSL_MAPPED_STORAGE,
 		.storage_flags = OSL_RBTREE | OSL_FIXED_SIZE | OSL_UNIQUE,
 		.name = "dir_num",
-		.compare_function = uint32_compare,
-		.data_size = sizeof(uint32_t)
+		.compare_function = uint64_compare,
+		.data_size = sizeof(uint64_t)
 	},
-	[IDT_BYTES] = {
+	[UT_BYTES] = {
 		.storage_type = OSL_MAPPED_STORAGE,
 		.storage_flags = OSL_RBTREE | OSL_FIXED_SIZE,
 		.compare_function = size_compare,
 		.name = "num_bytes",
 		.data_size = sizeof(uint64_t)
 	},
-	[IDT_FILES] = {
+	[UT_FILES] = {
 		.storage_type = OSL_MAPPED_STORAGE,
 		.storage_flags = OSL_RBTREE | OSL_FIXED_SIZE,
 		.compare_function = size_compare,
@@ -226,32 +215,14 @@ static struct osl_column_description user_table_cols[] = {
 	},
 };
 
-static struct osl_table_description user_table_desc = {
-	.num_columns = NUM_UT_COLUMNS,
-	.flags = 0,
-	.column_descriptions = user_table_cols,
-	.dir = "/tmp/adu"
-};
 static struct osl_table *dir_table;
-static struct osl_table *id_table;
-
-static int create_tables(void)
-{
-	int ret = osl_create_table(&dir_table_desc);
-	if (ret < 0)
-		return ret;
-	ret = osl_create_table(&id_table_desc);
-	if (ret < 0)
-		return ret;
-	return 1;
-}
 
-int add_directory(char *dirname, uint32_t dir_num, uint64_t *dir_size,
+static int add_directory(char *dirname, uint64_t dir_num, uint64_t *dir_size,
 		uint64_t *dir_files)
 {
 	struct osl_object dir_objects[NUM_DT_COLUMNS];
 
-	INFO_LOG("adding #%u: %s\n", dir_num, dirname);
+	INFO_LOG("adding #%llu: %s\n", (long long unsigned)dir_num, dirname);
 	dir_objects[DT_NAME].data = dirname;
 	dir_objects[DT_NAME].size = strlen(dirname) + 1;
 	dir_objects[DT_NUM].data = &dir_num;
@@ -264,78 +235,130 @@ int add_directory(char *dirname, uint32_t dir_num, uint64_t *dir_size,
 	return osl_add_row(dir_table, dir_objects);
 }
 
-int create_and_open_user_table(uint32_t uid, struct osl_table **t)
+static uint32_t num_uids;
+
+static int open_user_table(struct user_info *ui, int create)
 {
 	int ret;
-	struct osl_table_description *desc = para_malloc(sizeof(*desc));
-
-	desc->num_columns = NUM_UT_COLUMNS;
-	desc->flags = 0;
-	desc->column_descriptions = user_table_cols;
-	desc->dir = para_strdup("/tmp/adu");
-	desc->name = make_message("%u", uid);
-	INFO_LOG("................................. %u\n", uid);
-//	user_table_desc.name = make_message("%u", uid);
-	ret = osl_create_table(desc);
+
+	ui->desc = para_malloc(sizeof(*ui->desc));
+	ui->desc->num_columns = NUM_UT_COLUMNS;
+	ui->desc->flags = 0;
+	ui->desc->column_descriptions = user_table_cols;
+	ui->desc->dir = para_strdup(DATABASE_DIR);
+	ui->desc->name = make_message("%u", (unsigned)ui->uid);
+	num_uids++;
+	INFO_LOG(".............................uid #%u: %u\n",
+		(unsigned)num_uids, (unsigned)ui->uid);
+	if (create) {
+		ret = osl_create_table(ui->desc);
+		if (ret < 0)
+			goto err;
+	}
+	ret = osl_open_table(ui->desc, &ui->table);
 	if (ret < 0)
-		return ret;
-	return osl_open_table(desc, t);
+		goto err;
+	return 1;
+err:
+	free((char *)ui->desc->name);
+	free((char *)ui->desc->dir);
+	free(ui->desc);
+	ui->desc->name = NULL;
+	ui->desc->dir = NULL;
+	ui->desc = NULL;
+	ui->table = NULL;
+	return ret;
 }
 
-static int insert_id_row(uint32_t uid, struct osl_table *t, struct osl_row **row)
+#define uid_hash_bits 8
+static uint32_t uid_hash_table_size = 1 << uid_hash_bits;
+#define PRIME1 0x811c9dc5
+#define PRIME2 0x01000193
+
+static void create_hash_table(void)
 {
-	struct osl_object objects[NUM_IDT_COLUMNS];
-	uint64_t num = 0;
-
-	struct osl_table **table_ptr = para_malloc(sizeof(*table_ptr));
-	*table_ptr = t;
-
-	INFO_LOG("§§§§§§§§§§§§§§§§§§§§§ uid: %d, t: %p\n", uid, t);
-	objects[IDT_UID].data = &uid;
-	objects[IDT_UID].size = sizeof(uid);
-	objects[IDT_BYTES].data = &num;
-	objects[IDT_BYTES].size = sizeof(num);
-	objects[IDT_FILES].data = &num;
-	objects[IDT_FILES].size = sizeof(num);
-	objects[IDT_TABLE].data = table_ptr;
-	objects[IDT_TABLE].size = sizeof(*table_ptr);
-	return osl_add_and_get_row(id_table, objects, row);
+	uid_hash_table = para_calloc(uid_hash_table_size
+		* sizeof(struct user_info));
 }
 
-static int get_user_table(struct osl_row *row, struct osl_table **t)
+static void free_hash_table(void)
 {
-	struct osl_object obj;
+	free(uid_hash_table);
+	uid_hash_table = NULL;
+}
+
+static int create_tables(void)
+{
+	int ret;
 
-	int ret = osl_get_object(id_table, row, IDT_TABLE, &obj);
+	ret = osl_create_table(&dir_table_desc);
 	if (ret < 0)
 		return ret;
-	*t = *(struct osl_table **)obj.data;
-	INFO_LOG("^^^^^^^^^^^^^^^^^^ t: %p\n", *t);
+	create_hash_table();
 	return 1;
 }
 
-static int add_id_bytes(struct osl_row *row, uint64_t *add)
+/*
+ * We use a hash table of size s=2^uid_hash_bits to map the uids into the
+ * interval [0..s]. Hash collisions are treated by open addressing, i.e.
+ * unused slots in the table are used to store different uids that hash to the
+ * same slot.
+ *
+ * If a hash collision occurs, different slots are successively probed in order
+ * to find an unused slot for the new uid. Probing is implemented via a second
+ * hash function that maps the uid to h=(uid * PRIME2) | 1, which is always an
+ * odd number.
+ *
+ * An odd number is sufficient to make sure each entry of the hash table gets
+ * probed for probe_num between 0 and s-1 because s is a power of two, hence
+ * the second hash value never hash a common divisor with the hash table size.
+ * IOW: h is invertible in the ring [0..s].
+ */
+static uint32_t double_hash(uint32_t uid, uint32_t probe_num)
 {
-	uint64_t num;
-	struct osl_object obj1, obj2 = {.data = &num, .size = sizeof(num)};
+	return (uid * PRIME1 + ((uid * PRIME2) | 1) * probe_num)
+		% uid_hash_table_size;
+}
 
-	/* update number of bytes */
-	int ret = osl_get_object(id_table, row, IDT_BYTES, &obj1);
-	if (ret < 0)
-		return ret;
-	num = *(uint64_t *)obj1.data + *add;
-	ret = osl_update_object(id_table, row, IDT_BYTES, &obj2);
-	if (ret < 0)
-		return ret;
-	/* increment number of files */
-	ret = osl_get_object(id_table, row, IDT_FILES, &obj1);
-	if (ret < 0)
-		return ret;
-	num = *(uint64_t *)obj1.data + 1;
-	return osl_update_object(id_table, row, IDT_FILES, &obj2);
+#define FOR_EACH_USER(ui) for (ui = uid_hash_table; ui && ui < uid_hash_table \
+		+ uid_hash_table_size; ui++)
+
+enum search_uid_flags {
+	OPEN_USER_TABLE = 1,
+	CREATE_USER_TABLE = 2,
+};
+
+static int search_uid(uint32_t uid, enum search_uid_flags flags,
+		struct user_info **ui_ptr)
+{
+	uint32_t p;
+
+	for (p = 0; p < uid_hash_table_size; p++) {
+		struct user_info *ui = uid_hash_table + double_hash(uid, p);
+
+		if (!ui->table) {
+			int ret;
+
+			if (!flags)
+				return -E_BAD_UID;
+			ui->uid = uid;
+			ret = open_user_table(ui, flags & CREATE_USER_TABLE);
+			if (ret < 0)
+				return ret;
+			if (ui_ptr)
+				*ui_ptr = ui;
+			return 1;
+		}
+		if (ui->uid != uid)
+			continue;
+		if (ui_ptr)
+			*ui_ptr = ui;
+		return 0;
+	}
+	return flags? -E_HASH_TABLE_OVERFLOW : -E_BAD_UID;
 }
 
-static int update_user_row(struct osl_table *t, uint32_t dir_num,
+static int update_user_row(struct osl_table *t, uint64_t dir_num,
 		uint64_t *add)
 {
 	struct osl_row *row;
@@ -378,7 +401,9 @@ static int update_user_row(struct osl_table *t, uint32_t dir_num,
 	}
 }
 
-static uint32_t dir_num;
+static uint64_t num_dirs;
+static uint64_t num_files;
+static uint64_t num_bytes;
 
 int scan_dir(char *dirname)
 {
@@ -386,9 +411,9 @@ int scan_dir(char *dirname)
 	struct dirent *entry;
 	int ret, cwd_fd, ret2;
 	uint64_t dir_size = 0, dir_files = 0;
-	struct osl_object obj;
+	uint64_t this_dir_num = num_dirs++;
 
-	INFO_LOG("----------------- %s\n", dirname);
+	DEBUG_LOG("----------------- %llu: %s\n", (long long unsigned)num_dirs, dirname);
 	ret = para_opendir(dirname, &dir, &cwd_fd);
 	if (ret < 0) {
 		if (ret != -ERRNO_TO_ERROR(EACCES))
@@ -402,15 +427,17 @@ int scan_dir(char *dirname)
 		struct stat s;
 		uint32_t uid;
 		uint64_t size;
-		struct osl_row *id_row;
-		struct osl_table *user_table;
+		struct user_info *ui;
 
 		if (!strcmp(entry->d_name, "."))
 			continue;
 		if (!strcmp(entry->d_name, ".."))
 			continue;
-		if (lstat(entry->d_name, &s) == -1)
+		if (lstat(entry->d_name, &s) == -1) {
+			WARNING_LOG("lstat error for %s/%s\n", dirname,
+				entry->d_name);
 			continue;
+		}
 		m = s.st_mode;
 		if (!S_ISREG(m) && !S_ISDIR(m))
 			continue;
@@ -425,36 +452,20 @@ int scan_dir(char *dirname)
 		/* regular file */
 		size = s.st_size;
 		dir_size += size;
+		num_bytes += size;
 		dir_files++;
+		num_files++;
 		uid = s.st_uid;
-		INFO_LOG("++++++++++++++++++++++++++ %s, uid: %u\n", entry->d_name, uid);
-		obj.data = &uid;
-		obj.size = sizeof(uid);
-		ret = osl_get_row(id_table, IDT_UID, &obj, &id_row);
-		if (ret < 0 && ret != -E_RB_KEY_NOT_FOUND)
-			goto out;
-		if (ret < 0) {
-			ret = create_and_open_user_table(uid, &user_table);
-			if (ret < 0)
-				goto out;
-			ret = insert_id_row(uid, user_table, &id_row);
-			if (ret < 0)
-				goto out;
-		} else {
-			ret = get_user_table(id_row, &user_table);
-			if (ret < 0)
-				goto out;
-		}
-		ret = add_id_bytes(id_row, &size);
+		ret = search_uid(uid, CREATE_USER_TABLE | OPEN_USER_TABLE, &ui);
 		if (ret < 0)
 			goto out;
-		INFO_LOG("user_table: %p\n", user_table);
-		ret = update_user_row(user_table, dir_num, &size);
-		INFO_LOG("update_user  ret: %d\n", ret);
+		ui->bytes += size;
+		ui->files++;
+		ret = update_user_row(ui->table, this_dir_num, &size);
 		if (ret < 0)
 			goto out;
 	}
-	ret = add_directory(dirname, dir_num++, &dir_size, &dir_files);
+	ret = add_directory(dirname, this_dir_num, &dir_size, &dir_files);
 out:
 	closedir(dir);
 	ret2 = para_fchdir(cwd_fd);
@@ -475,166 +486,361 @@ static int get_dir_name(struct osl_row *row, char **name)
 	return 1;
 }
 
-static int print_dirname_and_size(struct osl_row *row, void *data)
+enum global_stats_flags {
+	GSF_PRINT_DIRNAME = 1,
+	GSF_PRINT_BYTES = 2,
+	GSF_PRINT_FILES = 4,
+	GSF_COMPUTE_SUMMARY = 8,
+};
+
+struct global_stats_info {
+	uint32_t count;
+	enum global_stats_flags flags;
+};
+
+static int global_stats_loop_function(struct osl_row *row, void *data)
 {
-	unsigned *count = data;
+	struct global_stats_info *gsi = data;
 	struct osl_object obj;
-	char *name;
-	int ret;
+	char *dirname;
+	int ret, summary = gsi->flags & GSF_COMPUTE_SUMMARY;
 
-	if ((*count)++ > 100)
+	if (!gsi->count && !summary)
 		return -E_LOOP_COMPLETE;
-	ret = get_dir_name(row, &name);
-	if (ret < 0)
-		return ret;
-	ret = osl_get_object(dir_table, row, DT_BYTES, &obj);
-	if (ret < 0)
-		return ret;
-	printf("%s\t%llu\n", name, *(long long unsigned *)obj.data);
+	if (gsi->count && (gsi->flags & GSF_PRINT_DIRNAME)) {
+		ret = get_dir_name(row, &dirname);
+		if (ret < 0)
+			return ret;
+		printf("%s%s", dirname,
+			(gsi->flags & (GSF_PRINT_FILES | GSF_PRINT_BYTES))?
+				"\t" : "\n"
+		);
+	}
+	if (summary || (gsi->count && (gsi->flags & GSF_PRINT_FILES))) {
+		uint64_t files;
+		ret = osl_get_object(dir_table, row, DT_FILES, &obj);
+		if (ret < 0)
+			return ret;
+		files = *(uint64_t *)obj.data;
+		if (gsi->count && (gsi->flags & GSF_PRINT_FILES))
+			printf("%llu%s", (long long unsigned)files,
+				(gsi->flags & GSF_PRINT_BYTES)? "\t" : "\n");
+		if (summary)
+			num_files += files;
+	}
+	if (summary || (gsi->count && (gsi->flags & GSF_PRINT_BYTES))) {
+		uint64_t bytes;
+		ret = osl_get_object(dir_table, row, DT_BYTES, &obj);
+		if (ret < 0)
+			return ret;
+		bytes = *(uint64_t *)obj.data;
+		if (gsi->count && (gsi->flags & GSF_PRINT_BYTES))
+			printf("%llu\n", (long long unsigned)bytes);
+		if (summary) {
+			num_bytes += bytes;
+			num_dirs++;
+		}
+	}
+	if (gsi->count)
+		gsi->count--;
 	return 1;
 }
 
-static int print_dirname_and_file_count(struct osl_row *row, void *data)
+static void print_id_stats(void)
+{
+	struct user_info *ui;
+
+	printf("--------------------- user summary (uid/dirs/files/bytes):\n");
+	FOR_EACH_USER(ui) {
+		if (!ui->table)
+			continue;
+		printf("%u\t%llu\t%llu\t%llu\n", (unsigned)ui->uid,
+			(long long unsigned)ui->dirs,
+			(long long unsigned)ui->files,
+			(long long unsigned)ui->bytes);
+	}
+}
+
+enum user_stats_flags {
+	USF_PRINT_DIRNAME = 1,
+	USF_PRINT_BYTES = 2,
+	USF_PRINT_FILES = 4,
+	USF_COMPUTE_SUMMARY = 8,
+};
+
+struct user_stats_info {
+	uint32_t count;
+	enum user_stats_flags flags;
+	struct user_info *ui;
+};
+
+static int user_stats_loop_function(struct osl_row *row, void *data)
 {
-	unsigned *count = data;
+	struct user_stats_info *usi = data;
+	struct osl_row *dir_row;
 	struct osl_object obj;
-	char *name;
-	int ret;
+	int ret, summary = usi->flags & GSF_COMPUTE_SUMMARY;
 
-	if ((*count)++ > 100)
+	if (!usi->count && !summary)
 		return -E_LOOP_COMPLETE;
-	ret = get_dir_name(row, &name);
-	if (ret < 0)
-		return ret;
-	ret = osl_get_object(dir_table, row, DT_FILES, &obj);
-	if (ret < 0)
-		return ret;
-	printf("%s\t%llu\n", name, *(long long unsigned *)obj.data);
+	if (usi->count && (usi->flags & USF_PRINT_DIRNAME)) {
+		char *dirname;
+		ret = osl_get_object(usi->ui->table, row, UT_DIR_NUM, &obj);
+		if (ret < 0)
+			return ret;
+		ret = osl_get_row(dir_table, DT_NUM, &obj, &dir_row);
+		if (ret < 0)
+			return ret;
+		ret = osl_get_object(dir_table, dir_row, DT_NAME, &obj);
+		if (ret < 0)
+			return ret;
+		dirname = obj.data;
+		printf("%s%s",
+			dirname,
+			(usi->flags & (USF_PRINT_FILES | USF_PRINT_BYTES))?
+				"\t" : "\n"
+		);
+	}
+	if (summary || (usi->count && (usi->flags & USF_PRINT_FILES))) {
+		uint64_t files;
+		ret = osl_get_object(usi->ui->table, row, UT_FILES, &obj);
+		if (ret < 0)
+			return ret;
+		files = *(uint64_t *)obj.data;
+		if (usi->count && (usi->flags & USF_PRINT_FILES))
+			printf("%llu%s",
+				(long long unsigned)files,
+				(usi->flags & USF_PRINT_BYTES)? "\t" : "\n"
+			);
+		if (summary)
+			usi->ui->files += files;
+	}
+	if (summary || (usi->count && (usi->flags & USF_PRINT_BYTES))) {
+		uint64_t bytes;
+		ret = osl_get_object(usi->ui->table, row, UT_BYTES, &obj);
+		if (ret < 0)
+			return ret;
+		bytes = *(uint64_t *)obj.data;
+		if (usi->count && (usi->flags & USF_PRINT_BYTES))
+			printf("%llu\n", (long long unsigned)bytes);
+		if (summary) {
+			usi->ui->bytes += bytes;
+			usi->ui->dirs++;
+		}
+
+	}
+	if (usi->count)
+		usi->count--;
 	return 1;
 }
 
-static int print_id_stats(struct osl_row *row, __a_unused void *data)
+static void print_user_stats(void)
 {
-	struct osl_object obj;
-	uint32_t uid;
-	uint64_t bytes, files;
-	int ret = osl_get_object(id_table, row, IDT_UID, &obj);
+	struct user_info *ui;
+
+	FOR_EACH_USER(ui) {
+		struct user_stats_info usi = {
+			.count = 10,
+			.ui = ui
+		};
+		if (!ui->table)
+			continue;
+		usi.flags = USF_PRINT_DIRNAME | USF_PRINT_BYTES | USF_COMPUTE_SUMMARY;
+		printf("************************************************ uid %u\n",
+			(unsigned) ui->uid);
+		if (!ui->table)
+			continue;
+		printf("----------------- Largest dirs -------------------\n");
+		osl_rbtree_loop_reverse(ui->table, UT_BYTES, &usi,
+			user_stats_loop_function);
+		printf("---------- dirs containing most files ------------\n");
+		usi.count = 10;
+		usi.flags = USF_PRINT_DIRNAME | USF_PRINT_FILES;
+		osl_rbtree_loop_reverse(ui->table, UT_FILES, &usi,
+			user_stats_loop_function);
+	}
+}
 
-	if (ret < 0)
-		return ret;
-	uid = *(uint32_t *)obj.data;
-	ret = osl_get_object(id_table, row, IDT_BYTES, &obj);
-	if (ret < 0)
+static int print_statistics(void)
+{
+	int ret;
+	struct global_stats_info gsi = {
+		.count = 10,
+		.flags = GSF_PRINT_DIRNAME | GSF_PRINT_BYTES | GSF_COMPUTE_SUMMARY
+	};
+
+	printf("----------------- Largest dirs -------------------\n");
+	ret = osl_rbtree_loop_reverse(dir_table, DT_BYTES, &gsi,
+		global_stats_loop_function);
+	if (ret < 0 && ret != -E_LOOP_COMPLETE)
 		return ret;
-	bytes = *(uint64_t *)obj.data;
-	ret = osl_get_object(id_table, row, IDT_FILES, &obj);
-	if (ret < 0)
+	gsi.count = 10;
+
+	gsi.flags = GSF_PRINT_DIRNAME | GSF_PRINT_FILES;
+	printf("---------- dirs containing most files ------------\n");
+	ret = osl_rbtree_loop_reverse(dir_table, DT_FILES, &gsi,
+		global_stats_loop_function);
+	if (ret < 0 && ret != -E_LOOP_COMPLETE)
 		return ret;
-	files = *(uint64_t *)obj.data;
 
-	printf("%u\t%llu\t%llu\n", (unsigned)uid, (long long unsigned)files,
-		(long long unsigned)bytes);
+	printf("------------------ Global summary (dirs/files/bytes)\n"
+		"%llu\t%llu\t%llu\n",
+		(long long unsigned)num_dirs, (long long unsigned)num_files,
+		(long long unsigned)num_bytes);
+	print_user_stats();
+	print_id_stats();
 	return 1;
 }
 
-struct id_dir_stat_info {
-	unsigned count;
-	struct osl_table *user_table;
-};
+static int write_uid_list(void)
+{
+	char *buf;
+	uint32_t count = 0;
+	struct user_info *ui;
+	size_t size = num_uids * sizeof(uint32_t);
+	int ret;
+
+	if (!num_uids)
+		return 0;
+	buf = para_malloc(size);
+	FOR_EACH_USER(ui) {
+		if (!ui->table)
+			continue;
+		write_u32(buf + count++ * sizeof(uint32_t), ui->uid);
+	}
+	ret = para_write_file(UID_LIST, buf, size);
+	free(buf);
+	return ret;
+}
 
-static int print_big_dir(struct osl_row *row, void *data)
+static int open_dir_table(void)
+{
+	return osl_open_table(&dir_table_desc, &dir_table);
+}
+
+static void close_dir_table(void)
 {
-	struct id_dir_stat_info *info = data;
-	info->count++;
 	int ret;
-	struct osl_row *dir_row;
-	char *dirname;
-	uint64_t bytes;
-	struct osl_object obj;
 
-	if (info->count > 10)
-		return -E_LOOP_COMPLETE;
-	ret = osl_get_object(info->user_table, row, UT_BYTES, &obj);
+	if (!dir_table)
+		return;
+	ret = osl_close_table(dir_table, OSL_MARK_CLEAN);
 	if (ret < 0)
-		return ret;
-	bytes = *(uint64_t *)obj.data;
-	ret = osl_get_object(info->user_table, row, UT_DIR_NUM, &obj);
+		ERROR_LOG("failed to close dir table: %s\n", error_txt(-ret));
+	dir_table = NULL;
+}
+
+static void close_user_table(struct user_info *ui)
+{
+	int ret;
+
+	if (!ui || !ui->table)
+		return;
+	ret = osl_close_table(ui->table, OSL_MARK_CLEAN);
 	if (ret < 0)
-		return ret;
-	ret = osl_get_row(dir_table, DT_NUM, &obj, &dir_row);
+		ERROR_LOG("failed to close user table %u: %s\n",
+			(unsigned) ui->uid, error_txt(-ret));
+	free((char *)ui->desc->name);
+	ui->desc->name = NULL;
+	free((char *)ui->desc->dir);
+	ui->desc->dir = NULL;
+	free(ui->desc);
+	ui->desc = NULL;
+	ui->table = NULL;
+}
+
+static void close_user_tables(void)
+{
+	struct user_info *ui;
+
+	FOR_EACH_USER(ui)
+		close_user_table(ui);
+}
+
+static void close_all_tables(void)
+{
+	close_dir_table();
+	close_user_tables();
+	free_hash_table();
+}
+
+static int com_create()
+{
+	int ret = create_tables();
 	if (ret < 0)
 		return ret;
-	ret = osl_get_object(dir_table, dir_row, DT_NAME, &obj);
+	ret = open_dir_table();
 	if (ret < 0)
 		return ret;
-	dirname = obj.data;
-	printf("%s: %llu\n", dirname, (long long unsigned)bytes);
-	return 1;
+	ret = scan_dir(conf.base_dir_arg);
+	if (ret < 0)
+		goto out;
+	ret = write_uid_list();
+out:
+	close_all_tables();
+	return ret;
 }
 
-static int print_id_dir_stats(struct osl_row *row, __a_unused void *data)
+static int read_uid_file(void)
 {
-	struct osl_object obj;
-	uint32_t uid;
-	int ret = osl_get_object(id_table, row, IDT_UID, &obj);
-	struct id_dir_stat_info info = {.count = 0};
+	char *map;
+	size_t size;
+	int ret = mmap_full_file(UID_LIST, O_RDONLY, (void **)&map, &size, NULL);
+	uint32_t n;
 
 	if (ret < 0)
 		return ret;
-	uid = *(uint32_t *)obj.data;
-
-	ret = osl_get_object(id_table, row, IDT_TABLE, &obj);
-	if (ret < 0)
-		return ret;
-	info.user_table = *(struct osl_table **)obj.data;
-
-	printf("************************* Big dirs owned by uid %u\n", (unsigned) uid);
-	osl_rbtree_loop_reverse(info.user_table, IDT_BYTES, &info, print_big_dir);
-	return 1;
+	num_uids = size / 4;
+	/* hash table size should be a power of two and larger than the number of uids */
+	uid_hash_table_size = 4;
+	while (uid_hash_table_size < num_uids)
+		uid_hash_table_size *= 2;
+	create_hash_table();
+	for (n = 0; n < num_uids; n++) {
+		uint32_t uid = read_u32(map + n * sizeof(uid));
+		ret = search_uid(uid, OPEN_USER_TABLE, NULL);
+		if (ret < 0)
+			goto out;
+	}
+out:
+	para_munmap(map, size);
+	return ret;
 }
 
-static int print_statistics(void)
+static int com_select(void)
 {
-	unsigned count = 0;
 	int ret;
 
-	printf("************************* Biggest dirs\n");
-	ret = osl_rbtree_loop_reverse(dir_table, DT_BYTES, &count, print_dirname_and_size);
-	if (ret < 0 && ret != -E_LOOP_COMPLETE)
-		return ret;
-	count = 0;
-	printf("************************* dirs containing many files\n");
-	ret = osl_rbtree_loop_reverse(dir_table, DT_FILES, &count, print_dirname_and_file_count);
-	if (ret < 0 && ret != -E_LOOP_COMPLETE)
+	ret = open_dir_table();
+	if (ret < 0)
 		return ret;
-
-	printf("************************* dirs stats by owner\n");
-	ret = osl_rbtree_loop(id_table, IDT_BYTES, NULL, print_id_stats);
+	ret = read_uid_file();
 	if (ret < 0)
 		return ret;
-
-	return osl_rbtree_loop(id_table, IDT_BYTES, NULL, print_id_dir_stats);
+	print_statistics();
+	close_all_tables();
+	return 1;
 }
 
-
 int main(int argc, char **argv)
 {
-	int ret = create_tables();
-	if (ret < 0)
-		goto out;
-	ret = osl_open_table(&dir_table_desc, &dir_table);
-	if (ret < 0)
-		goto out;
-	ret = osl_open_table(&id_table_desc, &id_table);
-	if (ret < 0)
-		goto out;
+	int ret;
+	struct cmdline_parser_params params = {
+		.override = 0,
+		.initialize = 1,
+		.check_required = 0,
+		.check_ambiguity = 0,
+		.print_errors = 1
+	};
+
+	cmdline_parser_ext(argc, argv, &conf, &params); /* aborts on errors */
 	ret = -E_SYNTAX;
-	if (argc != 2)
-		goto out;
-	ret = scan_dir(argv[1]);
+	if (conf.select_given)
+		ret = com_select();
+	else
+		ret = com_create();
 	if (ret < 0)
 		goto out;
-	print_statistics();
 out:
 	if (ret < 0) {
 		ERROR_LOG("%s\n", error_txt(-ret));
@@ -642,4 +848,3 @@ out:
 	}
 	return EXIT_SUCCESS;
 }
-
diff --git a/adu.ggo b/adu.ggo
new file mode 100644
index 0000000..ab6e025
--- /dev/null
+++ b/adu.ggo
@@ -0,0 +1,163 @@
+# Copyright (C) 2008 Andre Noll <maan@systemlinux.org>
+#
+# Licensed under the GPL v2. For licencing details see COPYING.
+
+package "adu"
+version "0.0.1"
+purpose "advanced disk usage
+
+adu creates a database containing disk usage statistics of a given
+directory. It allows to query that database to quickly retrieve
+usage patterns of subdirectories and/or files owned by a given user id.
+"
+
+#########################
+section "General options"
+#########################
+
+option "config-file" c
+#~~~~~~~~~~~~~~~~~~~~~
+"(default='~/.adurc')"
+string typestr="filename"
+optional
+details="
+	Options may be given at the command line or in the
+	configuration file. As usual, if an option is given both at
+	the command line and in the configuration file, the command
+	line option takes precedence.
+
+"
+
+option "database-dir" d
+#~~~~~~~~~~~~~~~~~~~~~~
+"directory containing the osl tables"
+string typestr="path"
+required
+details="
+	Full path to the directory containing the osl tables. This
+	directory must exist. It must be writable for the user running
+	adu in --create mode and readable in --select mode.
+
+"
+option "loglevel" l
+#~~~~~~~~~~~~~~~~~~
+"Set loglevel (0-6)"
+int typestr="level"
+default="3"
+optional
+details="
+	Log messages are always written to stderr while normal output
+	goes to stdout. Lower values mean more verbose logging.
+"
+
+option "uid" u
+#~~~~~~~~~~~~~
+"user id(s) to take into account"
+string typestr="uid_spec"
+optional
+multiple
+details="
+	An uid specifier may be a single number, or a range of uids.
+	Example:
+
+	--uid 42    # only consider uid 42
+	--uid 42-   # only consider uids greater or equal than 42
+	--uid 23-42 # only consider uids between 23 and 42, inclusively.
+
+	This option may be given multiple times. An uid is taken into
+	account if it satisfies at least one --uid option.
+"
+
+
+option "paths" p
+#~~~~~~~~~~~~~~~
+"files to take into account"
+string typestr="pattern"
+optional
+details="
+	Shell wildcard pattern that must match a file in order to be
+	included in the database in --create mode or in the output
+	for --select mode. Only the part of the filename below the
+	base directory is matched against the pattern. The default
+	is to take all files into account. See fnmatch(3) for details.
+"
+
+###############
+section "Modes"
+###############
+
+defgroup "mode"
+#==============
+groupdesc="
+	adu may started in one of two possible modes, each of which
+	corresponds to a different command line option. Exactly one
+	of these options must be given.
+
+"
+required
+
+groupoption "create" C
+#~~~~~~~~~~~~~~~~~~~~~
+"Create a new database"
+group="mode"
+details="
+	Traverse the given directory and track disk user on a per-user
+	basis. Results are stored in N + 1 osl tables where N is
+	the number of uids that own at least one regular file in
+	that directory.
+"
+
+groupoption "select" S
+#~~~~~~~~~~~~~~~~~~~~~
+"query a database previously created with --create"
+group="mode"
+details="
+	This option prints statistics about matching subdirectories to
+	stdout. The output depends on the other options, see below.
+"
+
+##############################
+section "Options for --create"
+##############################
+
+option "base-dir" b
+#~~~~~~~~~~~~~~~~~~
+"directory to traverse"
+string typestr="path"
+dependon="create"
+required
+details="
+	The base directory to be traversed recursively. A warning
+	message is printed for each subdirectory that could not be
+	read because of insufficient permission. These directories
+	will be ignored when computing statistics.
+"
+
+
+##############################
+section "Options for --select"
+##############################
+
+option "limit" L
+#~~~~~~~~~~~~~~~
+"Limit output"
+int  typestr="num"
+required
+dependon="select"
+details="
+	Only print num lines of output.
+"
+
+option "units" U
+#~~~~~~~~~~~~~~~
+"select numerical output format"
+string typestr="format"
+optional
+details="
+	Print the number of files/directories and the sizes in
+	the given format.  All	sizes  are output in these units:
+	(h)uman-readable, (b)ytes, (k)ilobytes, (m)egabytes,
+	(g)igabytes, (t)erabytes.  Capitalise to use multiples
+	of 1000 (S.I.) instead of 1024. The default is \"h\",
+	i.e. human-readable.
+"
diff --git a/error.h b/error.h
index 8603291..1067e7f 100644
--- a/error.h
+++ b/error.h
@@ -91,7 +91,9 @@ static inline char *error_txt(int num)
 	_ERROR(EMPTY, "file empty") \
 	_ERROR(MMAP, "mmap error") \
 	_ERROR(SYNTAX, "syntax error") \
-	_ERROR(LOOP_COMPLETE, "loop complete")
+	_ERROR(LOOP_COMPLETE, "loop complete") \
+	_ERROR(HASH_TABLE_OVERFLOW, "hash table too small") \
+	_ERROR(BAD_UID, "uid not found in hash table")
 
 
 /**