@matheusbsbs / matheus-bagliano / commits / 4f36b8597c

reftable/stack: fix race in up-to-date check

In 6fdfaf15a0 (reftable/stack: use stat info to avoid re-reading stack list, 2024-01-11) we have introduced a new mechanism to avoid re-reading the table list in case stat(3P) figures out that the stack didn't change since the last time we read it. While this change significantly improved performance when writing many refs, it can unfortunately lead to false negatives in very specific scenarios. Given two processes A and B, there is a feasible sequence of events that cause us to accidentally treat the table list as up-to-date even though it changed: 1. A reads the reftable stack and caches its stat info. 2. B updates the stack, appending a new table to "tables.list". This will both use a new inode and result in a different file size, thus invalidating A's cache in theory. 3. B decides to auto-compact the stack and merges two tables. The file size now matches what A has cached again. Furthermore, the filesystem may decide to recycle the inode number of the file we have replaced in (2) because it is not in use anymore. 4. A reloads the reftable stack. Neither the inode number nor the file size changed. If the timestamps did not change either then we think the cached copy of our stack is up-to-date. In fact, the commit introduced three related issues: - Non-POSIX compliant systems may not report proper `st_dev` and `st_ino` values in stat(3P), which made us rely solely on the file's potentially coarse-grained mtime and ctime. - `stat_validity_check()` and friends may end up not comparing `st_dev` and `st_ino` depending on the "core.checkstat" config, again reducing the signal to the mtime and ctime. - `st_ino` can be recycled, rendering the check moot even on POSIX-compliant systems. Given that POSIX defines that "The st_ino and st_dev fields taken together uniquely identify the file within the system", these issues led to the most important signal to establish file identity to be ignored or become useless in some cases. Refactor the code to stop using `stat_validity_check()`. Instead, we manually stat(3P) the file descriptors to make relevant information available. On Windows and MSYS2 the result will have both `st_dev` and `st_ino` set to 0, which allows us to address the first issue by not using the stat-based cache in that case. It also allows us to make sure that we always compare `st_dev` and `st_ino`, addressing the second issue. The third issue of inode recycling can be addressed by keeping the file descriptor of "files.list" open during the lifetime of the reftable stack. As the file will still exist on disk even though it has been unlinked it is impossible for its inode to be recycled as long as the file descriptor is still open. This should address the race in a POSIX-compliant way. The only real downside is that this mechanism cannot be used on non-POSIX-compliant systems like Windows. But we at least have the second-level caching mechanism in place that compares contents of "files.list" with the currently loaded list of tables. This new mechanism performs roughly the same as the previous one that relied on `stat_validity_check()`: Benchmark 1: update-ref: create many refs (HEAD~) Time (mean ± σ): 4.754 s ± 0.026 s [User: 2.204 s, System: 2.549 s] Range (min … max): 4.694 s … 4.802 s 20 runs Benchmark 2: update-ref: create many refs (HEAD) Time (mean ± σ): 4.721 s ± 0.020 s [User: 2.194 s, System: 2.527 s] Range (min … max): 4.691 s … 4.753 s 20 runs Summary update-ref: create many refs (HEAD~) ran 1.01 ± 0.01 times faster than update-ref: create many refs (HEAD) Signed-off-by: Patrick Steinhardt <ps@pks.im> Signed-off-by: Junio C Hamano <gitster@pobox.com>

Patrick Steinhardt committed Jan 18, 2024 at 14:41 UTC 4f36b8597c8f1c20e465ade4159896fb592e34c0

3 files changed +95 -9

reftable/stack.c

+92 -7

@@ -66,6 +66,7 @@ int reftable_new_stack(struct reftable_stack **dest, const char *dir,

66 strbuf_addstr(&list_file_name, "/tables.list");

68 p->list_file = strbuf_detach(&list_file_name, NULL);

69 + p->list_fd = -1;

70 p->reftable_dir = xstrdup(dir);

71 p->config = config;

@@ -175,7 +176,12 @@ void reftable_stack_destroy(struct reftable_stack *st)

176 st->readers_len = 0;

177 FREE_AND_NULL(st->readers);

178 }

178 - stat_validity_clear(&st->list_validity);

179 +

180 + if (st->list_fd >= 0) {

181 + close(st->list_fd);

182 + st->list_fd = -1;

183 + }

184 +

185 FREE_AND_NULL(st->list_file);

186 FREE_AND_NULL(st->reftable_dir);

187 reftable_free(st);

@@ -375,11 +381,59 @@ static int reftable_stack_reload_maybe_reuse(struct reftable_stack *st,

381 sleep_millisec(delay);

382 }

383

378 - stat_validity_update(&st->list_validity, fd);

379 -

384 out:

381 - if (err)

382 - stat_validity_clear(&st->list_validity);

385 + /*

386 + * Invalidate the stat cache. It is sufficient to only close the file

387 + * descriptor and keep the cached stat info because we never use the

388 + * latter when the former is negative.

389 + */

390 + if (st->list_fd >= 0) {

391 + close(st->list_fd);

392 + st->list_fd = -1;

393 + }

394 +

395 + /*

396 + * Cache stat information in case it provides a useful signal to us.

397 + * According to POSIX, "The st_ino and st_dev fields taken together

398 + * uniquely identify the file within the system." That being said,

399 + * Windows is not POSIX compliant and we do not have these fields

400 + * available. So the information we have there is insufficient to

401 + * determine whether two file descriptors point to the same file.

402 + *

403 + * While we could fall back to using other signals like the file's

404 + * mtime, those are not sufficient to avoid races. We thus refrain from

405 + * using the stat cache on such systems and fall back to the secondary

406 + * caching mechanism, which is to check whether contents of the file

407 + * have changed.

408 + *

409 + * On other systems which are POSIX compliant we must keep the file

410 + * descriptor open. This is to avoid a race condition where two

411 + * processes access the reftable stack at the same point in time:

412 + *

413 + * 1. A reads the reftable stack and caches its stat info.

414 + *

415 + * 2. B updates the stack, appending a new table to "tables.list".

416 + * This will both use a new inode and result in a different file

417 + * size, thus invalidating A's cache in theory.

418 + *

419 + * 3. B decides to auto-compact the stack and merges two tables. The

420 + * file size now matches what A has cached again. Furthermore, the

421 + * filesystem may decide to recycle the inode number of the file

422 + * we have replaced in (2) because it is not in use anymore.

423 + *

424 + * 4. A reloads the reftable stack. Neither the inode number nor the

425 + * file size changed. If the timestamps did not change either then

426 + * we think the cached copy of our stack is up-to-date.

427 + *

428 + * By keeping the file descriptor open the inode number cannot be

429 + * recycled, mitigating the race.

430 + */

431 + if (!err && fd >= 0 && !fstat(fd, &st->list_st) &&

432 + st->list_st.st_dev && st->list_st.st_ino) {

433 + st->list_fd = fd;

434 + fd = -1;

435 + }

436 +

437 if (fd >= 0)

438 close(fd);

439 free_names(names);

@@ -396,8 +450,39 @@ static int stack_uptodate(struct reftable_stack *st)

450 int err;

451 int i = 0;

452

399 - if (stat_validity_check(&st->list_validity, st->list_file))

400 - return 0;

453 + /*

454 + * When we have cached stat information available then we use it to

455 + * verify whether the file has been rewritten.

456 + *

457 + * Note that we explicitly do not want to use `stat_validity_check()`

458 + * and friends here because they may end up not comparing the `st_dev`

459 + * and `st_ino` fields. These functions thus cannot guarantee that we

460 + * indeed still have the same file.

461 + */

462 + if (st->list_fd >= 0) {

463 + struct stat list_st;

464 +

465 + if (stat(st->list_file, &list_st) < 0) {

466 + /*

467 + * It's fine for "tables.list" to not exist. In that

468 + * case, we have to refresh when the loaded stack has

469 + * any readers.

470 + */

471 + if (errno == ENOENT)

472 + return !!st->readers_len;

473 + return REFTABLE_IO_ERROR;

474 + }

475 +

476 + /*

477 + * When "tables.list" refers to the same file we can assume

478 + * that it didn't change. This is because we always use

479 + * rename(3P) to update the file and never write to it

480 + * directly.

481 + */

482 + if (st->list_st.st_dev == list_st.st_dev &&

483 + st->list_st.st_ino == list_st.st_ino)

484 + return 0;

485 + }

486

487 err = read_lines(st->list_file, &names);

488 if (err < 0)

reftable/stack.h

+3 -1

@@ -14,8 +14,10 @@ https://developers.google.com/open-source/licenses/bsd

14 #include "reftable-stack.h"

16 struct reftable_stack {

17 - struct stat_validity list_validity;

17 + struct stat list_st;

18 char *list_file;

19 + int list_fd;

20 +

21 char *reftable_dir;

22 int disable_auto_compact;

reftable/system.h

-1

@@ -12,7 +12,6 @@ https://developers.google.com/open-source/licenses/bsd

12 /* This header glues the reftable library to the rest of Git */

14 #include "git-compat-util.h"

15 -#include "statinfo.h"

15 #include "strbuf.h"

16 #include "hash-ll.h" /* hash ID, sizes.*/

17 #include "dir.h" /* remove_dir_recursively, for tests.*/