Volume XXII, number 279Tuesday, October 6, 2026Latest message 43 minutes ago

The Git List

News and archive of git@vger.kernel.org, since April 2005

patchmerge: fix leak with merge.defaultToUpstream

3 messages between Jul 28, 2026 and Jul 28, 2026, from Toon Claes, Jeff King, Junio C Hamano.

Plain Markdown or JSON for tools and agents. Diffs are folded; open one to read it.

Toon ClaesJul 28, 2026, 13:00 UTC on lore

By default the setting 'merge.defaultToUpstream' for git-merge(1) is set to 'true', which means when `git merge` is invoked with no arguments it merges the upstream branch configured for the current branch.

With this configuration set to 'true', setup_with_upstream() is called. That function allocates an array of arguments and hands it back to cmd_merge() via its `argv` parameter. This array is never freed, so cmd_merge() leaks it on every invocation.

Track the allocated array in a separate variable and free it at the end.

The leak has been present since 93e535a5b7 (merge: merge with the default upstream branch without argument, 2011-03-24). Although the leak sanitizer was enabled for tests in fc1ddf42af (t: remove TEST_PASSES_SANITIZE_LEAK annotations, 2024-11-21), it went unnoticed because no test calls `git merge` without arguments, exercising the default-to-upstream path. Add such a test in t7600, which fails under the leak sanitizer without this fix.

Signed-off-by: Toon Claes <toon@iotcl.com>
---
I ran into this leak while running `yay`[1] (I use Arch btw). `yay` uses
Git to fetch packages from source, and it happens to call `git merge`
without any revision to merge (it only passes the options `--no-edit`
and `--ff`). Because I have Git with the leak sanitizer enabled in my
$PATH, my `yay` tripped on a leak.
This series fixes the leak and adds a test to reproduce it.
[1]: https://github.com/Jguer/yay
---
 builtin/merge.c  |  7 +++++--
 t/t7600-merge.sh | 17 +++++++++++++++++
 2 files changed, 22 insertions(+), 2 deletions(-)
Show changes to 2 files +22 −2

builtin/merge.c, t/t7600-merge.sh

diff --git a/builtin/merge.c b/builtin/merge.c
index 58d1b7bb07..5b4eb23a83 100644
--- a/builtin/merge.c
+++ b/builtin/merge.c
@@ -1373,7 +1373,7 @@ int cmd_merge(int argc,
 	struct commit_list *common = NULL;
 	const char *best_strategy = NULL, *wt_strategy = NULL;
 	struct commit_list *remoteheads = NULL, *p;
-	void *branch_to_free;
+	void *branch_to_free, *argv_to_free = NULL;
 	int orig_argc = argc;
 	int merge_log_config = -1;
 
@@ -1517,8 +1517,10 @@ int cmd_merge(int argc,
 		option_commit = 1;
 
 	if (!argc) {
-		if (default_to_upstream)
+		if (default_to_upstream) {
 			argc = setup_with_upstream(&argv);
+			argv_to_free = argv;
+		}
 		else
 			die(_("No commit specified and merge.defaultToUpstream not set."));
 	} else if (argc == 1 && !strcmp(argv[0], "-")) {
@@ -1880,6 +1882,7 @@ int cmd_merge(int argc,
 	}
 	strbuf_release(&buf);
 	free(branch_to_free);
+	free(argv_to_free);
 	free(pull_twohead);
 	free(pull_octopus);
 	discard_index(the_repository->index);
diff --git a/t/t7600-merge.sh b/t/t7600-merge.sh
index 7f2a1db16d..e31d261f9d 100755
--- a/t/t7600-merge.sh
+++ b/t/t7600-merge.sh
@@ -1166,4 +1166,21 @@ test_expect_success 'suggested names are not ambiguous' '
 	test_grep remotes/origin/not-local stderr
 '
 
+test_expect_success 'merge with no argument defaults to upstream' '
+	test_when_finished "rm -rf upstream downstream" &&
+	git init upstream &&
+	(
+		cd upstream &&
+		test_commit one &&
+		test_commit two
+	) &&
+	git clone upstream downstream &&
+	(
+		cd downstream &&
+		git reset --hard HEAD^ &&
+		git merge &&
+		test_cmp_rev origin/main HEAD
+	)
+'
+
 test_done

---
base-commit: 13c7afec212fc97ce257d15601659314c6673d6c
change-id: 20260728-toon-fix-merge-leak-6d3bc5af2082
Jeff KingJul 28, 2026, 15:19 UTC in reply to Toon Claes on lore

Re: [PATCH] merge: fix leak with merge.defaultToUpstream

On Tue, Jul 28, 2026 at 03:00:04PM +0200, Toon Claes wrote:
Show 7 quoted lines
> The leak has been present since 93e535a5b7 (merge: merge with the
> default upstream branch without argument, 2011-03-24). Although the leak
> sanitizer was enabled for tests in fc1ddf42af (t: remove
> TEST_PASSES_SANITIZE_LEAK annotations, 2024-11-21), it went unnoticed
> because no test calls `git merge` without arguments, exercising the
> default-to-upstream path. Add such a test in t7600, which fails under
> the leak sanitizer without this fix.

Wow, I'm surprised we didn't cover this case in the test suite. Increasing coverage is good.

Show 9 quoted lines
> @@ -1517,8 +1517,10 @@ int cmd_merge(int argc,
>  		option_commit = 1;
>  
>  	if (!argc) {
> -		if (default_to_upstream)
> +		if (default_to_upstream) {
>  			argc = setup_with_upstream(&argv);
> +			argv_to_free = argv;
> +		}

The fix looks correct to me. This whole argv-juggling is pretty gross (especially the part below which overwrites argv[0]!). I suspect using a separate strvec to hold the heads would be cleaner, but it is probably not worth anybody's time to micro-polish this.

-Peff
Junio C HamanoJul 28, 2026, 18:12 UTC in reply to Jeff King on lore

Re: [PATCH] merge: fix leak with merge.defaultToUpstream

Jeff King <peff@peff.net> writes:
Show 27 quoted lines
> On Tue, Jul 28, 2026 at 03:00:04PM +0200, Toon Claes wrote:
>
>> The leak has been present since 93e535a5b7 (merge: merge with the
>> default upstream branch without argument, 2011-03-24). Although the leak
>> sanitizer was enabled for tests in fc1ddf42af (t: remove
>> TEST_PASSES_SANITIZE_LEAK annotations, 2024-11-21), it went unnoticed
>> because no test calls `git merge` without arguments, exercising the
>> default-to-upstream path. Add such a test in t7600, which fails under
>> the leak sanitizer without this fix.
>
> Wow, I'm surprised we didn't cover this case in the test suite.
> Increasing coverage is good.
>
>> @@ -1517,8 +1517,10 @@ int cmd_merge(int argc,
>>  		option_commit = 1;
>>  
>>  	if (!argc) {
>> -		if (default_to_upstream)
>> +		if (default_to_upstream) {
>>  			argc = setup_with_upstream(&argv);
>> +			argv_to_free = argv;
>> +		}
>
> The fix looks correct to me. This whole argv-juggling is pretty gross
> (especially the part below which overwrites argv[0]!). I suspect using a
> separate strvec to hold the heads would be cleaner, but it is probably
> not worth anybody's time to micro-polish this.
Thanks for heading off my OCD before it showed ;-).
And of course, thanks Toon for noticing and fixing.
Will queue.

Back to recent threads