Threaded grep

index da854fa..252cb0b 100644
@@ -16,11 +16,270 @@
 #include "quote.h"
 #include "dir.h"
 
+#ifndef NO_PTHREADS
+#include "thread-utils.h"
+#include <pthread.h>
+#endif
+
 static char const * const grep_usage[] = {
 	"git grep [options] [-e] <pattern> [<rev>...] [[--] path...]",
 	NULL
 };
 
+static int use_threads = 1;
+
+#ifndef NO_PTHREADS
+#define THREADS 8
+static pthread_t threads[THREADS];
+
+static void *load_sha1(const unsigned char *sha1, unsigned long *size,
+		       const char *name);
+static void *load_file(const char *filename, size_t *sz);
+
+enum work_type {WORK_SHA1, WORK_FILE};
+
+/* We use one producer thread and THREADS consumer
+ * threads. The producer adds struct work_items to 'todo' and the
+ * consumers pick work items from the same array.
+ */
+struct work_item
+{
+	enum work_type type;
+	char *name;
+
+	/* if type == WORK_SHA1, then 'identifier' is a SHA1,
+	 * otherwise type == WORK_FILE, and 'identifier' is a NUL
+	 * terminated filename.
+	 */
+	void *identifier;
+	char done;
+	struct strbuf out;
+};
+
+/* In the range [todo_done, todo_start) in 'todo' we have work_items
+ * that have been or are processed by a consumer thread. We haven't
+ * written the result for these to stdout yet.
+ *
+ * The work_items in [todo_start, todo_end) are waiting to be picked
+ * up by a consumer thread.
+ *
+ * The ranges are modulo TODO_SIZE.
+ */
+#define TODO_SIZE 128
+static struct work_item todo[TODO_SIZE];
+static int todo_start;
+static int todo_end;
+static int todo_done;
+
+/* Has all work items been added? */
+static int all_work_added;
+
+/* This lock protects all the variables above. */
+static pthread_mutex_t grep_lock;
+
+/* Used to serialize calls to read_sha1_file. */
+static pthread_mutex_t read_sha1_lock;
+
+/* Signalled when a new work_item is added to todo. */
+static pthread_cond_t cond_add;
+
+/* Signalled when the result from one work_item is written to
+ * stdout.
+ */
+static pthread_cond_t cond_write;
+
+/* Signalled when we are finished with everything. */
+static pthread_cond_t cond_result;
+
+static void add_work(enum work_type type, char *name, void *id)
+{
+	pthread_mutex_lock(&grep_lock);
+
+	while ((todo_end+1) % ARRAY_SIZE(todo) == todo_done) {
+		pthread_cond_wait(&cond_write, &grep_lock);
+	}
+
+	todo[todo_end].type = type;
+	todo[todo_end].name = name;
+	todo[todo_end].identifier = id;
+	todo[todo_end].done = 0;
+	strbuf_reset(&todo[todo_end].out);
+	todo_end = (todo_end + 1) % ARRAY_SIZE(todo);
+
+	pthread_cond_signal(&cond_add);
+	pthread_mutex_unlock(&grep_lock);
+}
+
+static struct work_item *get_work()
+{
+	struct work_item *ret;
+
+	pthread_mutex_lock(&grep_lock);
+	while (todo_start == todo_end && !all_work_added) {
+		pthread_cond_wait(&cond_add, &grep_lock);
+	}
+
+	if (todo_start == todo_end && all_work_added) {
+		ret = NULL;
+	} else {
+		ret = &todo[todo_start];
+		todo_start = (todo_start + 1) % ARRAY_SIZE(todo);
+	}
+	pthread_mutex_unlock(&grep_lock);
+	return ret;
+}
+
+static void grep_sha1_async(struct grep_opt *opt, char *name,
+			    const unsigned char *sha1)
+{
+	unsigned char *s;
+	s = xmalloc(20);
+	memcpy(s, sha1, 20);
+	add_work(WORK_SHA1, name, s);
+}
+
+static void grep_file_async(struct grep_opt *opt, char *name,
+			    const char *filename)
+{
+	add_work(WORK_FILE, name, xstrdup(filename));
+}
+
+static void work_done(struct work_item *w)
+{
+	int old_done;
+
+	pthread_mutex_lock(&grep_lock);
+	w->done = 1;
+	old_done = todo_done;
+	for(; todo[todo_done].done && todo_done != todo_start;
+	    todo_done = (todo_done+1) % ARRAY_SIZE(todo)) {
+		w = &todo[todo_done];
+		write_or_die(1, w->out.buf, w->out.len);
+		free(w->name);
+		free(w->identifier);
+	}
+
+	if (old_done != todo_done)
+		pthread_cond_signal(&cond_write);
+
+	if (all_work_added && todo_done == todo_end)
+		pthread_cond_signal(&cond_result);
+
+	pthread_mutex_unlock(&grep_lock);
+}
+
+static void *run(void *arg)
+{
+	int hit = 0;
+	struct grep_opt *opt = arg;
+
+	while (1) {
+		struct work_item *w = get_work();
+		if (!w)
+			break;
+
+		opt->output_priv = w;
+		if (w->type == WORK_SHA1) {
+			unsigned long sz;
+			void* data;
+
+			pthread_mutex_lock(&read_sha1_lock);
+			data = load_sha1(w->identifier, &sz, w->name);
+			pthread_mutex_unlock(&read_sha1_lock);
+
+			if (data) {
+				hit |= grep_buffer(opt, w->name, data, sz);
+				free(data);
+			}
+		} else if (w->type == WORK_FILE) {
+			size_t sz;
+			void* data = load_file(w->identifier, &sz);
+			if (data) {
+				hit |= grep_buffer(opt, w->name, data, sz);
+				free(data);
+			}
+		} else {
+			assert(0);
+		}
+
+		work_done(w);
+	}
+
+	return (void*) (intptr_t) hit;
+}
+
+static void strbuf_out(struct grep_opt *opt, const void *buf, size_t size)
+{
+	struct work_item *w = opt->output_priv;
+	strbuf_add(&w->out, buf, size);
+}
+
+static void start_threads(struct grep_opt *opt)
+{
+	int i;
+
+	pthread_mutex_init(&grep_lock, NULL);
+	pthread_mutex_init(&read_sha1_lock, NULL);
+	pthread_cond_init(&cond_add, NULL);
+	pthread_cond_init(&cond_write, NULL);
+	pthread_cond_init(&cond_result, NULL);
+
+	for (i = 0; i < ARRAY_SIZE(todo); i++) {
+		strbuf_init(&todo[i].out, 0);
+	}
+
+	for (i = 0; i < ARRAY_SIZE(threads); i++) {
+		int err;
+		struct grep_opt *o = grep_opt_dup(opt);
+		o->output = strbuf_out;
+		compile_grep_patterns(o);
+		err = pthread_create(&threads[i], NULL, run, o);
+
+		if (err)
+			die("grep: failed to create thread: %s",
+			    strerror(err));
+	}
+}
+
+static int wait_all()
+{
+	int hit = 0;
+	int i;
+
+	pthread_mutex_lock(&grep_lock);
+	all_work_added = 1;
+
+	/* Wait until all work is done. */
+	while (todo_done != todo_end)
+		pthread_cond_wait(&cond_result, &grep_lock);
+
+	/* Wake up all the consumer threads so they can see that there
+	 * is no more work to do.
+	 */
+	pthread_cond_broadcast(&cond_add);
+	pthread_mutex_unlock(&grep_lock);
+
+	for (i = 0; i < ARRAY_SIZE(threads); i++) {
+		void *h;
+		pthread_join(threads[i], &h);
+		hit |= (int) (intptr_t) h;
+	}
+
+	pthread_mutex_destroy(&grep_lock);
+	pthread_mutex_destroy(&read_sha1_lock);
+	pthread_cond_destroy(&cond_add);
+	pthread_cond_destroy(&cond_write);
+	pthread_cond_destroy(&cond_result);
+
+	return hit;
+}
+#else /* !NO_PTHREADS */
+static int wait_all()
+{
+	return 0;
+}
+#endif
+
 static int grep_config(const char *var, const char *value, void *cb)
 {
 	struct grep_opt *opt = cb;

@@ -144,37 +403,60 @@ static int pathspec_matches(const char **paths, const char *name, int max_depth)
 	return 0;
 }
 
-static int grep_sha1(struct grep_opt *opt, const unsigned char *sha1, const char *name, int tree_name_len)
+static void *load_sha1(const unsigned char *sha1, unsigned long *size,
+		       const char *name)
 {
-	unsigned long size;
-	char *data;
 	enum object_type type;
-	int hit;
-	struct strbuf pathbuf = STRBUF_INIT;
+	char *data = read_sha1_file(sha1, &type, size);
 
-	data = read_sha1_file(sha1, &type, &size);
-	if (!data) {
+	if (!data)
 		error("'%s': unable to read %s", name, sha1_to_hex(sha1));
-		return 0;
-	}
+
+	return data;
+}
+
+static int grep_sha1(struct grep_opt *opt, const unsigned char *sha1,
+		     const char *filename, int tree_name_len)
+{
+	struct strbuf pathbuf = STRBUF_INIT;
+	char *name;
+
 	if (opt->relative && opt->prefix_length) {
-		quote_path_relative(name + tree_name_len, -1, &pathbuf, opt->prefix);
-		strbuf_insert(&pathbuf, 0, name, tree_name_len);
-		name = pathbuf.buf;
+		quote_path_relative(filename + tree_name_len, -1, &pathbuf,
+				    opt->prefix);
+		strbuf_insert(&pathbuf, 0, filename, tree_name_len);
+	} else {
+		strbuf_addstr(&pathbuf, filename);
+	}
+
+	name = strbuf_detach(&pathbuf, NULL);
+
+#ifndef NO_PTHREADS
+	if (use_threads) {
+		grep_sha1_async(opt, name, sha1);
+		return 0;
+	} else
+#endif
+	{
+		int hit;
+		unsigned long sz;
+		void *data = load_sha1(sha1, &sz, name);
+		if (!data)
+			hit = 0;
+		else
+			hit = grep_buffer(opt, name, data, sz);
+
+		free(data);
+		free(name);
+		return hit;
 	}
-	hit = grep_buffer(opt, name, data, size);
-	strbuf_release(&pathbuf);
-	free(data);
-	return hit;
 }
 
-static int grep_file(struct grep_opt *opt, const char *filename)
+static void *load_file(const char *filename, size_t *sz)
 {
 	struct stat st;
-	int i;
 	char *data;
-	size_t sz;
-	struct strbuf buf = STRBUF_INIT;
+	int i;
 
 	if (lstat(filename, &st) < 0) {
 	err_ret:

@@ -184,25 +466,52 @@ static int grep_file(struct grep_opt *opt, const char *filename)
 	}
 	if (!S_ISREG(st.st_mode))
 		return 0;
-	sz = xsize_t(st.st_size);
+	*sz = xsize_t(st.st_size);
 	i = open(filename, O_RDONLY);
 	if (i < 0)
 		goto err_ret;
-	data = xmalloc(sz + 1);
-	if (st.st_size != read_in_full(i, data, sz)) {
+	data = xmalloc(*sz + 1);
+	if (st.st_size != read_in_full(i, data, *sz)) {
 		error("'%s': short read %s", filename, strerror(errno));
 		close(i);
 		free(data);
 		return 0;
 	}
 	close(i);
-	data[sz] = 0;
+	data[*sz] = 0;
+	return data;
+}
+
+static int grep_file(struct grep_opt *opt, const char *filename)
+{
+	struct strbuf buf = STRBUF_INIT;
+	char *name;
+
 	if (opt->relative && opt->prefix_length)
-		filename = quote_path_relative(filename, -1, &buf, opt->prefix);
-	i = grep_buffer(opt, filename, data, sz);
-	strbuf_release(&buf);
-	free(data);
-	return i;
+		quote_path_relative(filename, -1, &buf, opt->prefix);
+	else
+		strbuf_addstr(&buf, filename);
+	name = strbuf_detach(&buf, NULL);
+
+#ifndef NO_PTHREADS
+	if (use_threads) {
+		grep_file_async(opt, name, filename);
+		return 0;
+	} else
+#endif
+	{
+		int hit;
+		size_t sz;
+		void *data = load_file(filename, &sz);
+		if (!data)
+			hit = 0;
+		else
+			hit = grep_buffer(opt, name, data, sz);
+
+		free(data);
+		free(name);
+		return hit;
+	}
 }
 
 static int grep_cache(struct grep_opt *opt, const char **paths, int cached)

@@ -572,6 +881,17 @@ int cmd_grep(int argc, const char **argv, const char *prefix)
 		opt.regflags |= REG_ICASE;
 	if ((opt.regflags != REG_NEWLINE) && opt.fixed)
 		die("cannot mix --fixed-strings and regexp");
+
+#ifndef NO_PTHREADS
+	if (online_cpus() == 1 || !grep_threads_ok(&opt))
+		use_threads = 0;
+
+	if (use_threads)
+		start_threads(&opt);
+#else
+	use_threads = 0;
+#endif
+
 	compile_grep_patterns(&opt);
 
 	/* Check revs and then paths */

@@ -609,17 +929,26 @@ int cmd_grep(int argc, const char **argv, const char *prefix)
 	}
 
 	if (!use_index) {
+		int hit;
 		if (cached)
 			die("--cached cannot be used with --no-index.");
 		if (list.nr)
 			die("--no-index cannot be used with revs.");
-		return !grep_directory(&opt, paths);
+		hit = grep_directory(&opt, paths);
+		if (use_threads)
+			hit |= wait_all();
+		return !hit;
 	}
 
 	if (!list.nr) {
+		int hit;
 		if (!cached)
 			setup_work_tree();
-		return !grep_cache(&opt, paths, cached);
+
+		hit = grep_cache(&opt, paths, cached);
+		if (use_threads)
+			hit |= wait_all();
+		return !hit;
 	}
 
 	if (cached)

@@ -631,6 +960,9 @@ int cmd_grep(int argc, const char **argv, const char *prefix)
 		if (grep_object(&opt, paths, real_obj, list.objects[i].name))
 			hit = 1;
 	}
+
+	if (use_threads)
+		hit |= wait_all();
 	free_grep_patterns(&opt);
 	return !hit;
 }

171 lines not on this page · raw message

index 0c61b00..9703087 100644
@@ -91,6 +91,9 @@ struct grep_opt {
 	unsigned last_shown;
 	int show_hunk_mark;
 	void *priv;
+
+	void (*output)(struct grep_opt *opt, const void *data, size_t size);
+	void *output_priv;
 };
 
 extern void append_grep_pattern(struct grep_opt *opt, const char *pat, const char *origin, int no, enum grep_pat_token t);

@@ -99,4 +102,7 @@ extern void compile_grep_patterns(struct grep_opt *opt);
 extern void free_grep_patterns(struct grep_opt *opt);
 extern int grep_buffer(struct grep_opt *opt, const char *name, char *buf, unsigned long size);
 
+extern struct grep_opt *grep_opt_dup(const struct grep_opt *opt);
+extern int grep_threads_ok(const struct grep_opt *opt);
+
 #endif

index 7ecf222..6cc743d 100644
@@ -75,10 +75,15 @@ static int todo_done;
 static int all_work_added;
 
 /* This lock protects all the variables above. */
-static pthread_mutex_t grep_lock;
+static pthread_mutex_t grep_mutex;
 
 /* Used to serialize calls to read_sha1_file. */
-static pthread_mutex_t read_sha1_lock;
+static pthread_mutex_t read_sha1_mutex;
+
+#define grep_lock() pthread_mutex_lock(&grep_mutex)
+#define grep_unlock() pthread_mutex_unlock(&grep_mutex)
+#define read_sha1_lock() pthread_mutex_lock(&read_sha1_mutex)
+#define read_sha1_unlock() pthread_mutex_unlock(&read_sha1_mutex)
 
 /* Signalled when a new work_item is added to todo. */
 static pthread_cond_t cond_add;

@@ -93,10 +98,10 @@ static pthread_cond_t cond_result;
 
 static void add_work(enum work_type type, char *name, void *id)
 {
-	pthread_mutex_lock(&grep_lock);
+	grep_lock();
 
 	while ((todo_end+1) % ARRAY_SIZE(todo) == todo_done) {
-		pthread_cond_wait(&cond_write, &grep_lock);
+		pthread_cond_wait(&cond_write, &grep_mutex);
 	}
 
 	todo[todo_end].type = type;

@@ -107,16 +112,16 @@ static void add_work(enum work_type type, char *name, void *id)
 	todo_end = (todo_end + 1) % ARRAY_SIZE(todo);
 
 	pthread_cond_signal(&cond_add);
-	pthread_mutex_unlock(&grep_lock);
+	grep_unlock();
 }
 
 static struct work_item *get_work(void)
 {
 	struct work_item *ret;
 
-	pthread_mutex_lock(&grep_lock);
+	grep_lock();
 	while (todo_start == todo_end && !all_work_added) {
-		pthread_cond_wait(&cond_add, &grep_lock);
+		pthread_cond_wait(&cond_add, &grep_mutex);
 	}
 
 	if (todo_start == todo_end && all_work_added) {

@@ -125,7 +130,7 @@ static struct work_item *get_work(void)
 		ret = &todo[todo_start];
 		todo_start = (todo_start + 1) % ARRAY_SIZE(todo);
 	}
-	pthread_mutex_unlock(&grep_lock);
+	grep_unlock();
 	return ret;
 }
 

@@ -148,7 +153,7 @@ static void work_done(struct work_item *w)
 {
 	int old_done;
 
-	pthread_mutex_lock(&grep_lock);
+	grep_lock();
 	w->done = 1;
 	old_done = todo_done;
 	for(; todo[todo_done].done && todo_done != todo_start;

@@ -165,7 +170,7 @@ static void work_done(struct work_item *w)
 	if (all_work_added && todo_done == todo_end)
 		pthread_cond_signal(&cond_result);
 
-	pthread_mutex_unlock(&grep_lock);
+	grep_unlock();
 }
 
 static void *run(void *arg)

@@ -181,11 +186,7 @@ static void *run(void *arg)
 		opt->output_priv = w;
 		if (w->type == WORK_SHA1) {
 			unsigned long sz;
-			void* data;
-
-			pthread_mutex_lock(&read_sha1_lock);
-			data = load_sha1(w->identifier, &sz, w->name);
-			pthread_mutex_unlock(&read_sha1_lock);
+			void* data = load_sha1(w->identifier, &sz, w->name);
 
 			if (data) {
 				hit |= grep_buffer(opt, w->name, data, sz);

@@ -218,8 +219,8 @@ static void start_threads(struct grep_opt *opt)
 {
 	int i;
 
-	pthread_mutex_init(&grep_lock, NULL);
-	pthread_mutex_init(&read_sha1_lock, NULL);
+	pthread_mutex_init(&grep_mutex, NULL);
+	pthread_mutex_init(&read_sha1_mutex, NULL);
 	pthread_cond_init(&cond_add, NULL);
 	pthread_cond_init(&cond_write, NULL);
 	pthread_cond_init(&cond_result, NULL);

@@ -246,18 +247,18 @@ static int wait_all(void)
 	int hit = 0;
 	int i;
 
-	pthread_mutex_lock(&grep_lock);
+	grep_lock();
 	all_work_added = 1;
 
 	/* Wait until all work is done. */
 	while (todo_done != todo_end)
-		pthread_cond_wait(&cond_result, &grep_lock);
+		pthread_cond_wait(&cond_result, &grep_mutex);
 
 	/* Wake up all the consumer threads so they can see that there
 	 * is no more work to do.
 	 */
 	pthread_cond_broadcast(&cond_add);
-	pthread_mutex_unlock(&grep_lock);
+	grep_unlock();
 
 	for (i = 0; i < ARRAY_SIZE(threads); i++) {
 		void *h;

@@ -265,8 +266,8 @@ static int wait_all(void)
 		hit |= (int) (intptr_t) h;
 	}
 
-	pthread_mutex_destroy(&grep_lock);
-	pthread_mutex_destroy(&read_sha1_lock);
+	pthread_mutex_destroy(&grep_mutex);
+	pthread_mutex_destroy(&read_sha1_mutex);
 	pthread_cond_destroy(&cond_add);
 	pthread_cond_destroy(&cond_write);
 	pthread_cond_destroy(&cond_result);

@@ -274,6 +275,9 @@ static int wait_all(void)
 	return hit;
 }
 #else /* !NO_PTHREADS */
+#define read_sha1_lock()
+#define read_sha1_unlock()
+
 static int wait_all(void)
 {
 	return 0;

@@ -407,7 +411,11 @@ static void *load_sha1(const unsigned char *sha1, unsigned long *size,
 		       const char *name)
 {
 	enum object_type type;
-	char *data = read_sha1_file(sha1, &type, size);
+	char *data;
+
+	read_sha1_lock();
+	data = read_sha1_file(sha1, &type, size);
+	read_sha1_unlock();
 
 	if (!data)
 		error("'%s': unable to read %s", name, sha1_to_hex(sha1));

@@ -596,7 +604,10 @@ static int grep_tree(struct grep_opt *opt, const char **paths,
 			void *data;
 			unsigned long size;
 
+			read_sha1_lock();
 			data = read_sha1_file(entry.sha1, &type, &size);
+			read_sha1_unlock();
+
 			if (!data)
 				die("unable to read tree (%s)",
 				    sha1_to_hex(entry.sha1));

index d281a02..60cce46 100644
@@ -675,8 +675,15 @@ static int look_ahead(struct grep_opt *opt,
 
 		if (p->fixed)
 			hit = !fixmatch(p->pattern, bol, p->ignore_case, &m);
-		else
+		else {
+#ifdef REG_STARTEND
+			m.rm_so = 0;
+			m.rm_eo = *left_p;
+			hit = !regexec(&p->regexp, bol, 1, &m, REG_STARTEND);
+#else
 			hit = !regexec(&p->regexp, bol, 1, &m, 0);
+#endif
+		}
 		if (!hit || m.rm_so < 0 || m.rm_eo < 0)
 			continue;
 		if (earliest < 0 || m.rm_so < earliest)

index 8e1f7de..452c2cb 100644
@@ -640,8 +640,15 @@ static int look_ahead(struct grep_opt *opt,
 
 		if (p->fixed)
 			hit = !fixmatch(p->pattern, bol, p->ignore_case, &m);
-		else
+		else {
+#ifdef REG_STARTEND
+			m.rm_so = 0;
+			m.rm_eo = *left_p;
+			hit = !regexec(&p->regexp, bol, 1, &m, REG_STARTEND);
+#else
 			hit = !regexec(&p->regexp, bol, 1, &m, 0);
+#endif
+		}
 		if (!hit || m.rm_so < 0 || m.rm_eo < 0)
 			continue;
 		if (earliest < 0 || m.rm_so < earliest)