hooks: fix "invoked hook" regression in a8cc5943338
A message that is not in the archive (sent off-list or not yet received): [email protected]
PATCH 1/1 hooks: fix "invoked hook" regression in a8cc5943338
Fix a regression in a8cc5943338 (hooks: fix an obscure TOCTOU "did we just run a hook?" race, 2022-03-07): The "invoked_hook" variable passed to run_commit_hook() wasn't passed forward to run_hooks_opt(), as push_to_checkout() in that commit correctly did.
Whether we ran the code contingent on having run the hook or not was thus undefined, but in practice on most (all?) modern platforms we'd have run it (almost?) all the time, since stack variables will get initialized to some random value, which most of the time isn't "0".
This bug was revealed by running e.g. "t5537-fetch-shallow.sh" with the --valgrind option. Unfortunately running the whole test suite with --valgrind is really slow, so we didn't have a CI job that spotted this. The --valgrind output was:
==31275== Conditional jump or move depends on uninitialised value(s)
==31275== at 0x43C63F: prepare_to_commit (commit.c:1058)
==31275== by 0x4396A5: cmd_commit (commit.c:1722)
==31275== by 0x407C8A: run_builtin (git.c:465)
==31275== by 0x406741: handle_builtin (git.c:718)
==31275== by 0x407665: run_argv (git.c:785)
==31275== by 0x406500: cmd_main (git.c:916)
==31275== by 0x510839: main (common-main.c:56)
==31275== Uninitialised value was created by a stack allocation
==31275== at 0x43B344: prepare_to_commit (commit.c:719)---
Quoted 3 lines · Jonathan Tan
This commit causes Git to fail Valgrind (tested using "cd t && sh t5537*.sh -i -v --valgrind-only=10"). You can see here that a caller of
run_commit_hook()expectsinvoked_hookto be set, but...
That's a really stupid bug, sorry :( This fixes it.
commit.c | 1 + 1 file changed, 1 insertion(+)
commit.c +1 -0
@@ -1732,5 +1732,6 @@ int run_commit_hook(int editor_is_used, const char *index_file,strvec_push (&opt.args, arg);va_end (args); + opt.invoked_hook = invoked_hook;return run_hooks_opt (name, &opt); }
Signature · 1 line
2.35.1.1441.g1e9a595f48f
Quoted 46 lines · Ævar Arnfjörð Bjarmason
Fix a regression in
a8cc5943338(hooks: fix an obscure TOCTOU "did we just run a hook?" race, 2022-03-07): The "invoked_hook" variable passed torun_commit_hook()wasn't passed forward torun_hooks_opt(), aspush_to_checkout()in that commit correctly did.Whether we ran the code contingent on having run the hook or not was thus undefined, but in practice on most (all?) modern platforms we'd have run it (almost?) all the time, since stack variables will get initialized to some random value, which most of the time isn't "0".
This bug was revealed by running e.g. "
t5537-fetch-shallow.sh" with the --valgrind option. Unfortunately running the whole test suite with --valgrind is really slow, so we didn't have a CI job that spotted this. The --valgrind output was:==31275== Conditional jump or move depends on uninitialised value(s) ==31275== at 0x43C63F: prepare_to_commit (commit.c:1058) ==31275== by 0x4396A5: cmd_commit (commit.c:1722) ==31275== by 0x407C8A: run_builtin (git.c:465) ==31275== by 0x406741: handle_builtin (git.c:718) ==31275== by 0x407665: run_argv (git.c:785) ==31275== by 0x406500: cmd_main (git.c:916) ==31275== by 0x510839: main (common-main.c:56) ==31275== Uninitialised value was created by a stack allocation ==31275== at 0x43B344: prepare_to_commit (commit.c:719)Reported-by Jonathan Tan <[email protected]>Signed-off-by Ævar Arnfjörð Bjarmason <[email protected]>---
Quoted 3 lines · Jonathan Tan
This commit causes Git to fail Valgrind (tested using "cd t && sh t5537*.sh -i -v --valgrind-only=10"). You can see here that a caller of
run_commit_hook()expectsinvoked_hookto be set, but...That's a really stupid bug, sorry :( This fixes it.
commit.c | 1 + 1 file changed, 1 insertion(+)commit.c +0
-0
commit.c int run_commit_hook(int editor_is_used, const char *index_file, · line 1737
Quoted 6 lines
commit.c +1
-0@@ -1732,5 +1732,6 @@ int run_commit_hook(int editor_is_used, const char *index_file,strvec_push (&opt.args, arg);va_end (args); + opt.invoked_hook = invoked_hook;return run_hooks_opt (name, &opt); }
I tried adding a valgrind CI job, but even running --valgrind-only=1 (i.e. only the first test in each file would run with valgrind) took around 5h40m to complete with the patch below.
I wonder if it's still worth it in some form, it would suck to have to wait that long by default for "green" CI, but e.g. for Junio's "master" and "next" (and maybe "seen") maybe it's fine if some of the extended checks take a long time to complete. It would have caught the problem in this case.
We could probably run the entire test suite through valgrind by sharding the test jobs similarly to what the Windows jobs do now, and could also run e.g. SANITIZE=address and any other opt-in long-running CI.
.github/workflows/main.yml +3 -0
@@ -251,6 +251,9 @@ jobs: - jobname: linux-leaks cc: gcc pool: ubuntu-latest + - jobname: linux-valgrind + cc: gcc + pool: ubuntu-latest env: CC: ${{matrix.vector.cc}} CC_PACKAGE: ${{matrix.vector.cc_package}}
ci/install-dependencies.sh +3 -0
@@ -76,6 +76,9 @@ linux-gcc-default) sudo apt-get -q update sudo apt-get -q -y install $UBUNTU_COMMON_PKGS ;; +linux-valgrind) + sudo apt-get -q -y install valgrind + ;; esac if type p4d >/dev/null && type p4 >/dev/null
ci/lib.sh +4 -0
@@ -208,6 +208,10 @@ linux-leaks) export SANITIZE=leak export GIT_TEST_PASSING_SANITIZE_LEAK=true ;; +linux-valgrind) + GIT_TEST_OPTS="$GIT_TEST_OPTS --valgrind-only=1" + export GIT_TEST_OPTS + ;; esac MAKEFLAGS="$MAKEFLAGS CC=${CC:-cc}"
t/test-lib.sh +1 -1
@@ -274,7 +274,7 @@ fi if test -n "$valgrind_only" then test -z "$valgrind" && valgrind=memcheck - test -z "$verbose" && verbose_only="$valgrind_only" + test -z "$verbose$verbose_log" && verbose_only="$valgrind_only" elif test -n "$valgrind" then test -z "$verbose_log" && verbose=t