close(fd);
return;
}
- map = mmap(NULL, st.st_size, PROT_READ, MAP_PRIVATE, fd, 0);
+ map = xmmap(NULL, st.st_size, PROT_READ, MAP_PRIVATE, fd, 0);
close(fd);
- if (map == MAP_FAILED)
- return;
link_alt_odb_entries(map, map + st.st_size, '\n', relative_base, depth);
return NULL;
}
-static int pack_used_ctr;
-static unsigned long pack_mapped;
+static unsigned int pack_used_ctr;
+static unsigned int pack_mmap_calls;
+static unsigned int peak_pack_open_windows;
+static unsigned int pack_open_windows;
+static size_t peak_pack_mapped;
+static size_t pack_mapped;
+static size_t page_size;
struct packed_git *packed_git;
+void pack_report()
+{
+ fprintf(stderr,
+ "pack_report: getpagesize() = %10lu\n"
+ "pack_report: core.packedGitWindowSize = %10lu\n"
+ "pack_report: core.packedGitLimit = %10lu\n",
+ page_size,
+ packed_git_window_size,
+ packed_git_limit);
+ fprintf(stderr,
+ "pack_report: pack_used_ctr = %10u\n"
+ "pack_report: pack_mmap_calls = %10u\n"
+ "pack_report: pack_open_windows = %10u / %10u\n"
+ "pack_report: pack_mapped = %10lu / %10lu\n",
+ pack_used_ctr,
+ pack_mmap_calls,
+ pack_open_windows, peak_pack_open_windows,
+ pack_mapped, peak_pack_mapped);
+}
+
static int check_packed_git_idx(const char *path, unsigned long *idx_size_,
void **idx_map_)
{
return -1;
}
idx_size = st.st_size;
- idx_map = mmap(NULL, idx_size, PROT_READ, MAP_PRIVATE, fd, 0);
+ idx_map = xmmap(NULL, idx_size, PROT_READ, MAP_PRIVATE, fd, 0);
close(fd);
- if (idx_map == MAP_FAILED)
- return -1;
index = idx_map;
*idx_map_ = idx_map;
return 0;
}
-static int unuse_one_packed_git(void)
+static void scan_windows(struct packed_git *p,
+ struct packed_git **lru_p,
+ struct pack_window **lru_w,
+ struct pack_window **lru_l)
{
- struct packed_git *p, *lru = NULL;
+ struct pack_window *w, *w_l;
- for (p = packed_git; p; p = p->next) {
- if (!p->windows || p->windows->inuse_cnt)
- continue;
- if (!lru || p->windows->last_used < lru->windows->last_used)
- lru = p;
+ for (w_l = NULL, w = p->windows; w; w = w->next) {
+ if (!w->inuse_cnt) {
+ if (!*lru_w || w->last_used < (*lru_w)->last_used) {
+ *lru_p = p;
+ *lru_w = w;
+ *lru_l = w_l;
+ }
+ }
+ w_l = w;
}
- if (!lru)
- return 0;
- munmap(lru->windows->base, lru->windows->len);
- free(lru->windows);
- lru->windows = NULL;
- close(p->pack_fd);
- p->pack_fd = -1;
- return 1;
+}
+
+static int unuse_one_window(struct packed_git *current)
+{
+ struct packed_git *p, *lru_p = NULL;
+ struct pack_window *lru_w = NULL, *lru_l = NULL;
+
+ if (current)
+ scan_windows(current, &lru_p, &lru_w, &lru_l);
+ for (p = packed_git; p; p = p->next)
+ scan_windows(p, &lru_p, &lru_w, &lru_l);
+ if (lru_p) {
+ munmap(lru_w->base, lru_w->len);
+ pack_mapped -= lru_w->len;
+ if (lru_l)
+ lru_l->next = lru_w->next;
+ else {
+ lru_p->windows = lru_w->next;
+ if (!lru_p->windows && lru_p != current) {
+ close(lru_p->pack_fd);
+ lru_p->pack_fd = -1;
+ }
+ }
+ free(lru_w);
+ pack_open_windows--;
+ return 1;
+ }
+ return 0;
+}
+
+void release_pack_memory(size_t need)
+{
+ size_t cur = pack_mapped;
+ while (need >= (cur - pack_mapped) && unuse_one_window(NULL))
+ ; /* nothing */
}
void unuse_pack(struct pack_window **w_cursor)
struct pack_header hdr;
unsigned char sha1[20];
unsigned char *idx_sha1;
+ long fd_flag;
p->pack_fd = open(p->pack_name, O_RDONLY);
if (p->pack_fd < 0 || fstat(p->pack_fd, &st))
} else if (p->pack_size != st.st_size)
die("packfile %s size changed", p->pack_name);
+ /* We leave these file descriptors open with sliding mmap;
+ * there is no point keeping them open across exec(), though.
+ */
+ fd_flag = fcntl(p->pack_fd, F_GETFD, 0);
+ if (fd_flag < 0)
+ die("cannot determine file descriptor flags");
+ fd_flag |= FD_CLOEXEC;
+ if (fcntl(p->pack_fd, F_SETFD, fd_flag) == -1)
+ die("cannot set FD_CLOEXEC");
+
/* Verify we recognize this pack file format. */
read_or_die(p->pack_fd, &hdr, sizeof(hdr));
if (hdr.hdr_signature != htonl(PACK_SIGNATURE))
die("packfile %s does not match index", p->pack_name);
}
+static int in_window(struct pack_window *win, unsigned long offset)
+{
+ /* We must promise at least 20 bytes (one hash) after the
+ * offset is available from this window, otherwise the offset
+ * is not actually in this window and a different window (which
+ * has that one hash excess) must be used. This is to support
+ * the object header and delta base parsing routines below.
+ */
+ off_t win_off = win->offset;
+ return win_off <= offset
+ && (offset + 20) <= (win_off + win->len);
+}
+
unsigned char* use_pack(struct packed_git *p,
struct pack_window **w_cursor,
unsigned long offset,
unsigned int *left)
{
- struct pack_window *win = p->windows;
+ struct pack_window *win = *w_cursor;
if (p->pack_fd == -1)
open_packed_git(p);
- if (!win) {
- pack_mapped += p->pack_size;
- while (packed_git_limit < pack_mapped && unuse_one_packed_git())
- ; /* nothing */
- win = xcalloc(1, sizeof(*win));
- win->len = p->pack_size;
- win->base = mmap(NULL, p->pack_size, PROT_READ, MAP_PRIVATE, p->pack_fd, 0);
- if (win->base == MAP_FAILED)
- die("packfile %s cannot be mapped.", p->pack_name);
- p->windows = win;
+
+ /* Since packfiles end in a hash of their content and its
+ * pointless to ask for an offset into the middle of that
+ * hash, and the in_window function above wouldn't match
+ * don't allow an offset too close to the end of the file.
+ */
+ if (offset > (p->pack_size - 20))
+ die("offset beyond end of packfile (truncated pack?)");
+
+ if (!win || !in_window(win, offset)) {
+ if (win)
+ win->inuse_cnt--;
+ for (win = p->windows; win; win = win->next) {
+ if (in_window(win, offset))
+ break;
+ }
+ if (!win) {
+ if (!page_size)
+ page_size = getpagesize();
+ win = xcalloc(1, sizeof(*win));
+ win->offset = (offset / page_size) * page_size;
+ win->len = p->pack_size - win->offset;
+ if (win->len > packed_git_window_size)
+ win->len = packed_git_window_size;
+ pack_mapped += win->len;
+ while (packed_git_limit < pack_mapped
+ && unuse_one_window(p))
+ ; /* nothing */
+ win->base = xmmap(NULL, win->len,
+ PROT_READ, MAP_PRIVATE,
+ p->pack_fd, win->offset);
+ if (win->base == MAP_FAILED)
+ die("packfile %s cannot be mapped: %s",
+ p->pack_name,
+ strerror(errno));
+ pack_mmap_calls++;
+ pack_open_windows++;
+ if (pack_mapped > peak_pack_mapped)
+ peak_pack_mapped = pack_mapped;
+ if (pack_open_windows > peak_pack_open_windows)
+ peak_pack_open_windows = pack_open_windows;
+ win->next = p->windows;
+ p->windows = win;
+ }
}
if (win != *w_cursor) {
win->last_used = pack_used_ctr++;
win->inuse_cnt++;
*w_cursor = win;
}
+ offset -= win->offset;
if (left)
*left = win->len - offset;
return win->base + offset;
*/
sha1_file_open_flag = 0;
}
- map = mmap(NULL, st.st_size, PROT_READ, MAP_PRIVATE, fd, 0);
+ map = xmmap(NULL, st.st_size, PROT_READ, MAP_PRIVATE, fd, 0);
close(fd);
- if (map == MAP_FAILED)
- return NULL;
*size = st.st_size;
return map;
}
unsigned char *base_info = use_pack(p, w_curs, offset, NULL);
unsigned long base_offset;
- /* there must be at least 20 bytes left regardless of delta type */
- if (p->pack_size <= offset + 20)
- die("truncated pack file");
-
+ /* use_pack() assured us we have [base_info, base_info + 20)
+ * as a range that we can look at without walking off the
+ * end of the mapped window. Its actually the hash size
+ * that is assured. An OFS_DELTA longer than the hash size
+ * is stupid, as then a REF_DELTA would be smaller to store.
+ */
if (kind == OBJ_OFS_DELTA) {
unsigned used = 0;
unsigned char c = base_info[used++];
if (sizep) {
const unsigned char *data;
- unsigned char delta_head[20];
+ unsigned char delta_head[20], *in;
unsigned long result_size;
z_stream stream;
int st;
memset(&stream, 0, sizeof(stream));
-
- stream.next_in = use_pack(p, w_curs, offset, &stream.avail_in);
stream.next_out = delta_head;
stream.avail_out = sizeof(delta_head);
inflateInit(&stream);
- st = inflate(&stream, Z_FINISH);
+ do {
+ in = use_pack(p, w_curs, offset, &stream.avail_in);
+ stream.next_in = in;
+ st = inflate(&stream, Z_FINISH);
+ offset += stream.next_in - in;
+ } while ((st == Z_OK || st == Z_BUF_ERROR)
+ && stream.total_out < sizeof(delta_head));
inflateEnd(&stream);
if ((st != Z_STREAM_END) &&
stream.total_out != sizeof(delta_head))
unsigned int left;
unsigned long used;
+ /* use_pack() assures us we have [base, base + 20) available
+ * as a range that we can look at at. (Its actually the hash
+ * size that is assurred.) With our object header encoding
+ * the maximum deflated object size is 2^137, which is just
+ * insane, so we know won't exceed what we have been given.
+ */
base = use_pack(p, w_curs, offset, &left);
used = unpack_object_header_gently(base, left, type, sizep);
if (!used)
{
int st;
z_stream stream;
- unsigned char *buffer;
+ unsigned char *buffer, *in;
buffer = xmalloc(size + 1);
buffer[size] = 0;
memset(&stream, 0, sizeof(stream));
- stream.next_in = use_pack(p, w_curs, offset, &stream.avail_in);
stream.next_out = buffer;
stream.avail_out = size;
inflateInit(&stream);
- st = inflate(&stream, Z_FINISH);
+ do {
+ in = use_pack(p, w_curs, offset, &stream.avail_in);
+ stream.next_in = in;
+ st = inflate(&stream, Z_FINISH);
+ offset += stream.next_in - in;
+ } while (st == Z_OK || st == Z_BUF_ERROR);
inflateEnd(&stream);
if ((st != Z_STREAM_END) || stream.total_out != size) {
free(buffer);
buf = "";
if (size)
- buf = mmap(NULL, size, PROT_READ, MAP_PRIVATE, fd, 0);
+ buf = xmmap(NULL, size, PROT_READ, MAP_PRIVATE, fd, 0);
close(fd);
- if (buf == MAP_FAILED)
- return -1;
if (!type)
type = blob_type;