Add memory pool library

Add a memory pool library implemented using C macros. The
obj_pool_gen() macro creates a type-specific memory pool.

The memory pool library is distinguished from the existing specialized
allocators in alloc.c by using a contiguous block for all allocations.
This means that on one hand, long-lived pointers have to be written as
offsets, since the base address changes as the pool grows, but on the
other hand, the entire pool can be easily written to the file system.
This could allow the memory pool to persist between runs of an
application.

For the svn importer, such a facility is useful because each svn
revision can copy trees and files from any previous revision.  The
relevant information for all revisions has to persist somehow to
support incremental runs.

[rr: minor cleanups]
[jn: added tests; removed file system backing for now]

Signed-off-by: David Barr <david.barr@cordelta.com>
Signed-off-by: Ramkumar Ramachandra <artagnon@gmail.com>
Signed-off-by: Jonathan Nieder <jrnieder@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>

5 files changed, 260 insertions, 1 deletions

diff --git a/.gitignore b/.gitignore
index 14e2b6bde9..1e64a6a138 100644
--- a/.gitignore
+++ b/.gitignore
@@ -167,6 +167,7 @@
 /test-genrandom
 /test-index-version
 /test-match-trees
+/test-obj-pool
 /test-parse-options
 /test-path-utils
 /test-run-command
diff --git a/Makefile b/Makefile
index 0b6454d3e8..47cbf26421 100644
--- a/Makefile
+++ b/Makefile
@@ -409,6 +409,7 @@ TEST_PROGRAMS_NEED_X += test-delta
 TEST_PROGRAMS_NEED_X += test-dump-cache-tree
 TEST_PROGRAMS_NEED_X += test-genrandom
 TEST_PROGRAMS_NEED_X += test-match-trees
+TEST_PROGRAMS_NEED_X += test-obj-pool
 TEST_PROGRAMS_NEED_X += test-parse-options
 TEST_PROGRAMS_NEED_X += test-path-utils
 TEST_PROGRAMS_NEED_X += test-run-command
@@ -1864,7 +1865,8 @@ xdiff-interface.o $(XDIFF_OBJS): \
 	xdiff/xinclude.h xdiff/xmacros.h xdiff/xdiff.h xdiff/xtypes.h \
 	xdiff/xutils.h xdiff/xprepare.h xdiff/xdiffi.h xdiff/xemit.h
 
-$(VCSSVN_OBJS):
+$(VCSSVN_OBJS): \
+	vcs-svn/obj_pool.h
 endif
 
 exec_cmd.s exec_cmd.o: EXTRA_CPPFLAGS = \
diff --git a/t/t0080-vcs-svn.sh b/t/t0080-vcs-svn.sh
new file mode 100755
index 0000000000..3f29496bf8
--- /dev/null
+++ b/t/t0080-vcs-svn.sh
@@ -0,0 +1,79 @@
+#!/bin/sh
+
+test_description='check infrastructure for svn importer'
+
+. ./test-lib.sh
+uint32_max=4294967295
+
+test_expect_success 'obj pool: store data' '
+	cat <<-\EOF >expected &&
+	0
+	1
+	EOF
+
+	test-obj-pool <<-\EOF >actual &&
+	alloc one 16
+	set one 13
+	test one 13
+	reset one
+	EOF
+	test_cmp expected actual
+'
+
+test_expect_success 'obj pool: NULL is offset ~0' '
+	echo "$uint32_max" >expected &&
+	echo null one | test-obj-pool >actual &&
+	test_cmp expected actual
+'
+
+test_expect_success 'obj pool: out-of-bounds access' '
+	cat <<-EOF >expected &&
+	0
+	0
+	$uint32_max
+	$uint32_max
+	16
+	20
+	$uint32_max
+	EOF
+
+	test-obj-pool <<-\EOF >actual &&
+	alloc one 16
+	alloc two 16
+	offset one 20
+	offset two 20
+	alloc one 5
+	offset one 20
+	free one 1
+	offset one 20
+	reset one
+	reset two
+	EOF
+	test_cmp expected actual
+'
+
+test_expect_success 'obj pool: high-water mark' '
+	cat <<-\EOF >expected &&
+	0
+	0
+	10
+	20
+	20
+	20
+	EOF
+
+	test-obj-pool <<-\EOF >actual &&
+	alloc one 10
+	committed one
+	alloc one 10
+	commit one
+	committed one
+	alloc one 10
+	free one 20
+	committed one
+	reset one
+	EOF
+	test_cmp expected actual
+'
+
+test_done
diff --git a/test-obj-pool.c b/test-obj-pool.c
new file mode 100644
index 0000000000..5018863ef5
--- /dev/null
+++ b/test-obj-pool.c
@@ -0,0 +1,116 @@
+/*
+ * test-obj-pool.c: code to exercise the svn importer's object pool
+ */
+
+#include "cache.h"
+#include "vcs-svn/obj_pool.h"
+
+enum pool { POOL_ONE, POOL_TWO };
+obj_pool_gen(one, int, 1)
+obj_pool_gen(two, int, 4096)
+
+static uint32_t strtouint32(const char *s)
+{
+	char *end;
+	uintmax_t n = strtoumax(s, &end, 10);
+	if (*s == '\0' || (*end != '\n' && *end != '\0'))
+		die("invalid offset: %s", s);
+	return (uint32_t) n;
+}
+
+static void handle_command(const char *command, enum pool pool, const char *arg)
+{
+	switch (*command) {
+	case 'a':
+		if (!prefixcmp(command, "alloc ")) {
+			uint32_t n = strtouint32(arg);
+			printf("%"PRIu32"\n",
+				pool == POOL_ONE ?
+				one_alloc(n) : two_alloc(n));
+			return;
+		}
+	case 'c':
+		if (!prefixcmp(command, "commit ")) {
+			pool == POOL_ONE ? one_commit() : two_commit();
+			return;
+		}
+		if (!prefixcmp(command, "committed ")) {
+			printf("%"PRIu32"\n",
+				pool == POOL_ONE ?
+				one_pool.committed : two_pool.committed);
+			return;
+		}
+	case 'f':
+		if (!prefixcmp(command, "free ")) {
+			uint32_t n = strtouint32(arg);
+			pool == POOL_ONE ? one_free(n) : two_free(n);
+			return;
+		}
+	case 'n':
+		if (!prefixcmp(command, "null ")) {
+			printf("%"PRIu32"\n",
+				pool == POOL_ONE ?
+				one_offset(NULL) : two_offset(NULL));
+			return;
+		}
+	case 'o':
+		if (!prefixcmp(command, "offset ")) {
+			uint32_t n = strtouint32(arg);
+			printf("%"PRIu32"\n",
+				pool == POOL_ONE ?
+				one_offset(one_pointer(n)) :
+				two_offset(two_pointer(n)));
+			return;
+		}
+	case 'r':
+		if (!prefixcmp(command, "reset ")) {
+			pool == POOL_ONE ? one_reset() : two_reset();
+			return;
+		}
+	case 's':
+		if (!prefixcmp(command, "set ")) {
+			uint32_t n = strtouint32(arg);
+			if (pool == POOL_ONE)
+				*one_pointer(n) = 1;
+			else
+				*two_pointer(n) = 1;
+			return;
+		}
+	case 't':
+		if (!prefixcmp(command, "test ")) {
+			uint32_t n = strtouint32(arg);
+			printf("%d\n", pool == POOL_ONE ?
+				*one_pointer(n) : *two_pointer(n));
+			return;
+		}
+	default:
+		die("unrecognized command: %s", command);
+	}
+}
+
+static void handle_line(const char *line)
+{
+	const char *arg = strchr(line, ' ');
+	enum pool pool;
+
+	if (arg && !prefixcmp(arg + 1, "one"))
+		pool = POOL_ONE;
+	else if (arg && !prefixcmp(arg + 1, "two"))
+		pool = POOL_TWO;
+	else
+		die("no pool specified: %s", line);
+
+	handle_command(line, pool, arg + strlen("one "));
+}
+
+int main(int argc, char *argv[])
+{
+	struct strbuf sb = STRBUF_INIT;
+	if (argc != 1)
+		usage("test-obj-str < script");
+
+	while (strbuf_getline(&sb, stdin, '\n') != EOF)
+		handle_line(sb.buf);
+	strbuf_release(&sb);
+	return 0;
+}
diff --git a/vcs-svn/obj_pool.h b/vcs-svn/obj_pool.h
new file mode 100644
index 0000000000..deb6eb8135
--- /dev/null
+++ b/vcs-svn/obj_pool.h
@@ -0,0 +1,61 @@
+/*
+ * Licensed under a two-clause BSD-style license.
+ * See LICENSE for details.
+ */
+
+#ifndef OBJ_POOL_H_
+#define OBJ_POOL_H_
+
+#include "git-compat-util.h"
+
+#define MAYBE_UNUSED __attribute__((__unused__))
+
+#define obj_pool_gen(pre, obj_t, initial_capacity) \
+static struct { \
+	uint32_t committed; \
+	uint32_t size; \
+	uint32_t capacity; \
+	obj_t *base; \
+} pre##_pool = {0, 0, 0, NULL}; \
+static MAYBE_UNUSED uint32_t pre##_alloc(uint32_t count) \
+{ \
+	uint32_t offset; \
+	if (pre##_pool.size + count > pre##_pool.capacity) { \
+		while (pre##_pool.size + count > pre##_pool.capacity) \
+			if (pre##_pool.capacity) \
+				pre##_pool.capacity *= 2; \
+			else \
+				pre##_pool.capacity = initial_capacity; \
+		pre##_pool.base = realloc(pre##_pool.base, \
+					pre##_pool.capacity * sizeof(obj_t)); \
+	} \
+	offset = pre##_pool.size; \
+	pre##_pool.size += count; \
+	return offset; \
+} \
+static MAYBE_UNUSED void pre##_free(uint32_t count) \
+{ \
+	pre##_pool.size -= count; \
+} \
+static MAYBE_UNUSED uint32_t pre##_offset(obj_t *obj) \
+{ \
+	return obj == NULL ? ~0 : obj - pre##_pool.base; \
+} \
+static MAYBE_UNUSED obj_t *pre##_pointer(uint32_t offset) \
+{ \
+	return offset >= pre##_pool.size ? NULL : &pre##_pool.base[offset]; \
+} \
+static MAYBE_UNUSED void pre##_commit(void) \
+{ \
+	pre##_pool.committed = pre##_pool.size; \
+} \
+static MAYBE_UNUSED void pre##_reset(void) \
+{ \
+	free(pre##_pool.base); \
+	pre##_pool.base = NULL; \
+	pre##_pool.size = 0; \
+	pre##_pool.capacity = 0; \
+	pre##_pool.committed = 0; \
+}
+
+#endif