"git maintenance start" crashed due to an uninitialized variable
reference, which has been corrected.
* ps/maintenance-start-crash-fix:
builtin/gc: fix crash when running `git maintenance start`
On macOS, fsmonitor can fall into a race condition that results in
a client waiting forever to be notified for an event that have
already happened. This problem has been corrected.
* jk/fsmonitor-event-listener-race-fix:
fsmonitor: initialize fs event listener before accepting clients
simple-ipc: split async server initialization and running
Use after free and double freeing at the end in "git log -L... -p"
had been identified and fixed.
* ds/line-log-asan-fix:
line-log: protect inner strbuf from free
It was reported on the mailing list that running `git maintenance start`
immediately segfaults starting with b6c3f8e12c (builtin/maintenance: fix
leak in `get_schedule_cmd()`, 2024-09-26). And indeed, this segfault is
trivial to reproduce up to a point where one is scratching their head
why we didn't catch this regression in our test suite.
The root cause of this error is `get_schedule_cmd()`, which does not
populate the `out` parameter in all cases anymore starting with the
mentioned commit. Callers do assume it to always be populated though and
will e.g. call `strvec_split()` on the returned value, which will of
course segfault when the variable is uninitialized.
So why didn't we catch this trivial regression? The reason is that our
tests always set up the "GIT_TEST_MAINT_SCHEDULER" environment variable
via "t/test-lib.sh", which allows us to override the scheduler command
with a custom one so that we don't accidentally modify the developer's
system. But the faulty code where we don't set the `out` parameter will
only get hit in case that environment variable is _not_ set, which is
never the case when executing our tests.
Fix the regression by again unconditionally allocating the value in the
`out` parameter, if provided. Add a test that unsets the environment
variable to catch future regressions in this area.
Reported-by: Shubham Kanodia <shubham.kanodia10@gmail.com>
Signed-off-by: Patrick Steinhardt <ps@pks.im>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
There's a racy hang in fsmonitor on macOS that we sometimes see in CI.
When we serve a client, what's supposed to happen is:
1. The client thread calls with_lock__wait_for_cookie() in which we
create a cookie file and then wait for a pthread_cond event
2. The filesystem event listener sees the cookie file creation, does
some internal book-keeping, and then triggers the pthread_cond.
But there's a problem: we start the listener that accepts client threads
before we start the fs event thread. So it's possible for us to accept a
client which creates the cookie file and starts waiting before the fs
event thread is initialized, and we miss those filesystem events
entirely. That leaves the client thread hanging forever.
In CI, the symptom is that t9210 (which is testing scalar, which always
enables fsmonitor under the hood) may hang forever in "scalar clone". It
is waiting on "git fetch" which is waiting on the fsmonitor daemon.
The race happens more frequently under load, but you can trigger it
predictably with a sleep like this, which delays the start of the fs
event thread:
--- a/compat/fsmonitor/fsm-listen-darwin.c
+++ b/compat/fsmonitor/fsm-listen-darwin.c
@@ -510,6 +510,7 @@ void fsm_listen__loop(struct fsmonitor_daemon_state *state)
FSEventStreamSetDispatchQueue(data->stream, data->dq);
data->stream_scheduled = 1;
+ sleep(1);
if (!FSEventStreamStart(data->stream)) {
error(_("Failed to start the FSEventStream"));
goto force_error_stop_without_loop;
One solution might be to reverse the order of initialization: start the
fs event thread before we start the thread listening for clients. But
the fsmonitor code explicitly does it in the opposite direction. The fs
event thread wants to refer to the ipc_server_data struct, so we need it
to be initialized first.
A further complication is that we need a signal from the fs event thread
that it is actually ready and listening. And those details happen within
backend-specific fsmonitor code, whereas the initialization is in the
shared code.
So instead, let's use the ipc_server init/start split added in the
previous commit. The generic fsmonitor code will init the ipc_server but
_not_ start it, leaving that to the backend specific code, which now
needs to call ipc_server_start_async() at the right time.
For macOS, that is right after we start the FSEventStream that you can
see in the diff above.
It's not clear to me if Windows suffers from the same problem (and we
simply don't trigger it in CI), or if it is immune. Regardless, the
obvious place to start accepting clients there is right after we've
established the ReadDirectoryChanges watch.
This makes the hangs go away in our macOS CI environment, even when
compiled with the sleep() above.
Helped-by: Koji Nakamaru <koji.nakamaru@gree.net>
Signed-off-by: Jeff King <peff@peff.net>
Acked-by: Koji Nakamaru <koji.nakamaru@gree.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
To start an async ipc server, you call ipc_server_run_async(). That
initializes the ipc_server_data object, and starts all of the threads
running, which may immediately start serving clients.
This can create some awkward timing problems, though. In the fsmonitor
daemon (the sole user of the simple-ipc system), we want to create the
ipc server early in the process, which means we may start serving
clients before the rest of the daemon is fully initialized.
To solve this, let's break run_async() into two parts: an initialization
which allocates all data and spawns the threads (without letting them
run), and a start function which actually lets them begin work. Since we
have two simple-ipc implementations, we have to handle this twice:
- in ipc-unix-socket.c, we have a central listener thread which hands
connections off to worker threads using a work_available mutex. We
can hold that mutex after init, and release it when we're ready to
start.
We do need an extra "started" flag so that we know whether the main
thread is holding the mutex or not (e.g., if we prematurely stop the
server, we want to make sure all of the worker threads are released
to hear about the shutdown).
- in ipc-win32.c, we don't have a central mutex. So we'll introduce a
new startup_barrier mutex, which we'll similarly hold until we're
ready to let the threads proceed.
We again need a "started" flag here to make sure that we release the
barrier mutex when shutting down, so that the sub-threads can
proceed to the finish.
I've renamed the run_async() function to init_async() to make sure we
catch all callers, since they'll now need to call the matching
start_async().
We could leave run_async() as a wrapper that does both, but there's not
much point. There are only two callers, one of which is fsmonitor, which
will want to actually do work between the two calls. And the other is
just a test-tool wrapper.
For now I've added the start_async() calls in fsmonitor where they would
otherwise have happened, so there should be no behavior change with this
patch.
Signed-off-by: Jeff King <peff@peff.net>
Acked-by: Koji Nakamaru <koji.nakamaru@gree.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
macOS with fsmonitor daemon can hang forever when a submodule is
involved, which has been corrected.
* kn/osx-fsmonitor-with-submodules-fix:
fsmonitor OSX: fix hangs for submodules
Usability improvements for running tests in leak-checking mode.
* jk/test-lsan-improvements:
test-lib: check for leak logs after every test
test-lib: show leak-sanitizer logs on --immediate failure
test-lib: stop showing old leak logs
In 6241ce2170 (refs/reftable: reload locked stack when preparing
transaction, 2024-09-24) we have introduced a new test that exercises
how the reftable backend behaves with many concurrent writers all racing
with each other. This test was introduced after a couple of fixes in
this context that should make concurrent writes behave gracefully. As it
turns out though, Windows systems do not yet handle concurrent writes
properly, as we've got two reports for Cygwin and MinGW failing in this
newly added test.
The root cause of this is how we update the "tables.list" file: when
writing a new stack of tables we first write the data into a lockfile
and then rename that file into place. But Windows forbids us from doing
that rename when the target path is open for reading by another process.
And as the test races both readers and writers with each other we are
quite likely to hit this edge case.
This is not a regression: the logic didn't work before the mentioned
commit, and after the commit it performs well on Linux and macOS, and
the situation on Windows should have at least improved a bit. But the
test shows that we need to put more thought into how to make this work
properly there.
Work around the issue by disabling the test on Windows for now. While at
it, increase the locking timeout to address reported timeouts when using
either the address or memory sanitizer, which also tend to significantly
extend the runtime of this test.
This should be revisited after Git v2.47 is out.
Signed-off-by: Patrick Steinhardt <ps@pks.im>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
fsmonitor_classify_path_absolute() expects state->path_gitdir_watch.buf
has no trailing '/' or '.' For a submodule, fsmonitor_run_daemon() sets
the value with trailing "/." (as repo_get_git_dir(the_repository) on
Darwin returns ".") so that fsmonitor_classify_path_absolute() returns
IS_OUTSIDE_CONE.
In this case, fsevent_callback() doesn't update cookie_list so that
fsmonitor_publish() does nothing and with_lock__mark_cookies_seen() is
not invoked.
As with_lock__wait_for_cookie() infinitely waits for state->cookies_cond
that with_lock__mark_cookies_seen() should unlock, the whole daemon
hangs.
Remove trailing "/." from state->path_gitdir_watch.buf for submodules
and add a corresponding test in t7527-builtin-fsmonitor.sh. The test is
disabled for MINGW because hangs treated with this patch occur only for
Darwin and there is no simple way to terminate the win32 fsmonitor
daemon that hangs.
Suggested-by: Johannes Schindelin <johannes.schindelin@gmx.de>
Suggested-by: Junio C Hamano <gitster@pobox.com>
Signed-off-by: Koji Nakamaru <koji.nakamaru@gree.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Commit 253ed9ecff (hash.h: scaffolding for _unsafe hashing variants,
2024-09-26) introduced the concept of having two hash algorithms: a safe
and an unsafe one. When the Makefile knobs do not explicitly request an
unsafe one, we fall back to using the safe algorithm.
However, the fallback to do so forgot one case: we should inherit the
NEEDS_CLONE_HELPER flag from the safe variant. Failing to do so means
that we'll end up defining two clone functions (the algorithm specific
one, and the generic one that just calls memcpy). You'll see an error
like this:
$ make OPENSSL_SHA1=1
[...]
sha1/openssl.h:46:29: error: redefinition of ‘openssl_SHA1_Clone’
46 | #define platform_SHA1_Clone openssl_SHA1_Clone
| ^~~~~~~~~~~~~~~~~~
hash.h:83:40: note: in expansion of macro ‘platform_SHA1_Clone’
83 | # define platform_SHA1_Clone_unsafe platform_SHA1_Clone
| ^~~~~~~~~~~~~~~~~~~
hash.h:101:33: note: in expansion of macro ‘platform_SHA1_Clone_unsafe’
101 | # define git_SHA1_Clone_unsafe platform_SHA1_Clone_unsafe
| ^~~~~~~~~~~~~~~~~~~~~~~~~~
hash.h:133:20: note: in expansion of macro ‘git_SHA1_Clone_unsafe’
133 | static inline void git_SHA1_Clone_unsafe(git_SHA_CTX_unsafe *dst,
| ^~~~~~~~~~~~~~~~~~~~~
sha1/openssl.h:37:20: note: previous definition of ‘openssl_SHA1_Clone’ with type ‘void(struct openssl_SHA1_CTX *, const struct openssl_SHA1_CTX *)’
37 | static inline void openssl_SHA1_Clone(struct openssl_SHA1_CTX *dst,
| ^~~~~~~~~~~~~~~~~~
This only matters when compiling with openssl as the "safe" variant,
since it's the only algorithm that requires a clone helper (and even
then, only if you are using openssl 3.0+). And you should never do that,
because it's not safe. But still, the invocation above used to work and
should continue to do so until we decide to require a
collision-detecting variant for the safe algorithm entirely.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The error message added by 296743a7ca (archive: load index before
pathspec checks, 2024-09-21) is misleading: unpack_trees() is not
touching the working tree at all here, but just loading a tree into
the index. Correct it.
Signed-off-by: René Scharfe <l.s.r@web.de>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The output_prefix() method in line-log.c may call a function pointer via
the diff_options struct. This function pointer returns a strbuf struct
and then its buffer is passed back. However, that implies that the
consumer is responsible to free the string. This is especially true
because the default behavior is to duplicate the empty string.
The existing functions used in the output_prefix pointer include:
1. idiff_prefix_cb() in diff-lib.c. This returns the data pointer, so
the value exists across multiple calls.
2. diff_output_prefix_callback() in graph.c. This uses a static strbuf
struct, so it reuses buffers across calls. These should not be
freed.
3. output_prefix_cb() in range-diff.c. This is similar to the
diff-lib.c case.
In each case, we should not be freeing this buffer. We can convert the
output_prefix() function to return a const char pointer and stop freeing
the result.
This choice is essentially the opposite of what was done in 394affd46d
(line-log: always allocate the output prefix, 2024-06-07).
This was discovered via 'valgrind' while investigating a public report
of a bug in 'git log --graph -L' [1].
[1] https://github.com/git-for-windows/git/issues/5185
This issue would have been caught by the new test, when Git is compiled
with ASan to catch these double frees.
Co-authored-by: Jeff King <peff@peff.net>
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Derrick Stolee <stolee@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Since DEVELOPER=YesPlease build enables -Wunused-parameter warnings
these days, the fallback definition for reencode_string_len() that
did not touch any of its parameters but one needs to be annotated
properly.
Signed-off-by: Mike Hommey <mh@glandium.org>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
