git/list[1] front-page[2] threads[3] people[4] search[5] about
 

[PATCH] leak tests: add an interface to the LSAN_OPTIONS "suppressions"

From
Ævar Arnfjörð Bjarmason <avarab@gmail.com>
Date
Oct 22, 2021, 10:32 UTC
Message-ID
<patch-1.1-9190f3c128f-20211022T102725Z-avarab@gmail.com>
In-Reply-To
<211022.86sfwtl6uj.gmgdl@evledraar.gmail.com>

Extend the SANITIZE=leak testing mode added in 956d2e4639b (tests: add a test mode for SANITIZE=leak, run it in CI, 2021-09-23) to optionally be able to add a "suppressions" file to the $LSAN_OPTIONS.

This allows for marking tests as passing with "TEST_PASSES_SANITIZE_LEAK=true" when they still have failure due more general widespread memory leaks, such as from the "git log" family of commands. We can now mark the "git -C" tests as passing.

For getting specific tests to pass this is preferable to using UNLEAK() in these codepaths, as I'll have fixes for those leaks soon, and being able to atomically mark relevant tests as passing with "TEST_PASSES_SANITIZE_LEAK=true" helps to explain those changes. See [1] for more details.

1. https://lore.kernel.org/git/211022.86sfwtl6uj.gmgdl@evledraar.gmail.com/
Signed-off-by: Ævar Arnfjörð Bjarmason <avarab@gmail.com>
---
On Fri, Oct 22 2021, Ævar Arnfjörð Bjarmason wrote:
Show 5 quoted lines
> On Fri, Oct 22 2021, Taylor Blau wrote:
>
>> On Thu, Oct 21, 2021 at 01:50:55PM +0200, Ævar Arnfjörð Bjarmason wrote:
>>>
>>> On Wed, Oct 20 2021, Taylor Blau wrote:
[...]
Show 19 quoted lines
> If you want to pick that approach up and run with it I think it would
> probably make sense to factor that suggested test_expect_success out
> into a function in test-lib-functions.sh or whatever, and call it as
> e.g.:
>     
>     TEST_PASSES_SANITIZE_LEAK=true
>      . ./test-lib.sh
>     declare_known_leaks <<-\EOF
>     add_rev_cmdline
>     [...]
>     EOF
>
> Then pipe it through sed 's/^/leak:/' and have it set LSAN_OPTIONS for
> you.
>
> Doing it that way would be less boilerplate for each test that wants it,
> and is also more likely to work with other non-LSAN leak appoaches,
> i.e. as long as something can take a list of lines matching stack traces
> we can feed that to that leak checker's idea of a whitelist.

I just went ahead and hacked that up. If you're OK with that approach it would really help reduce the work for leak changes I've got planned, and as noted gives you the end-state of a passing 5319.

I don't know if it makes more sense for you to base on top of this if/when Junio picks it up, or to integrate it into your series etc. Maybe Junio will chime in ...

 t/t0056-git-C.sh        |  4 ++++
 t/test-lib-functions.sh | 42 +++++++++++++++++++++++++++++++++++++++++
 2 files changed, 46 insertions(+)
diff --git a/t/t0056-git-C.sh b/t/t0056-git-C.sh
index 2630e756dab..490aefa81a1 100755
--- a/t/t0056-git-C.sh
+++ b/t/t0056-git-C.sh
@@ -2,7 +2,11 @@
 
 test_description='"-C <path>" option and its effects on other path-related options'
 
+TEST_PASSES_SANITIZE_LEAK=true
 . ./test-lib.sh
+todo_leaks <<-\EOF
+^cmd_log_init_finish$
+EOF
 
 test_expect_success '"git -C <path>" runs git from the directory <path>' '
 	test_create_repo dir1 &&
diff --git a/t/test-lib-functions.sh b/t/test-lib-functions.sh
index eef2262a360..d89bf5da7dc 100644
--- a/t/test-lib-functions.sh
+++ b/t/test-lib-functions.sh
@@ -243,6 +243,48 @@ debug () {
 	done
 }
 
+# Declare known "general" memory leaks, for use with TEST_PASSES_SANITIZE_LEAK=true.
+#
+# Matches lines in a stack trace that leaks. Intended for
+# LSAN_OPTIONS, but the format is intended to be easy to use with
+# other leak checkers, so the "leak:" prefix is omitted (and added for
+# you).
+#
+# Use it immediately after sourcing test-lib.sh (or equivalent), and
+# after a "TEST_PASSES_SANITIZE_LEAK=true" has been set. E.g:
+#
+#    TEST_PASSES_SANITIZE_LEAK=true
+#    . ./test-lib.sh
+#    todo_leaks <<-\EOF
+#    ^cmd_log_init_finish$
+#    EOF
+#
+# The "^" and "$" anchors don't suggest full regex syntax support,
+# that's the only anchoring (or other metacharacter) understood by
+# LSAN_OPTIONS,.
+#
+# See
+# https://github.com/google/sanitizers/wiki/AddressSanitizerLeakSanitizer#suppressions
+# for the relevant LSAN_OPTIONS documentation.
+todo_leaks () {
+	if ! test_have_prereq SANITIZE_LEAK
+	then
+		return 0
+	fi
+
+	# Try not to interfere with any test logic
+	suppressions=.lsan-suppressions.txt
+	if test -d .git
+	then
+		suppressions=".git/$suppressions"
+	fi
+	suppressions="$PWD/$suppressions"
+
+	sed 's/^/leak:/' >"$suppressions" &&
+	LSAN_OPTIONS="$LSAN_OPTIONS:suppressions=\"$suppressions\"" &&
+	export LSAN_OPTIONS
+}
+
 # Usage: test_commit [options] <message> [<file> [<contents> [<tag>]]]
 #   -C <dir>:
 #	Run all git commands in directory <dir>
-- 
2.33.1.1494.g88b39a443e1
Previous: Ævar Arnfjörð BjarmasonNext: Taylor Blau
Message 51 of 63 in “midx: clean up t5319 under 'SANITIZE=leak'”
  1. 00/11 midx: clean up t5319 under 'SANITIZE=leak'Taylor Blau, Oct 21, 2021
  2. 01/11 midx.c: clean up chunkfile after reading the MIDXTaylor Blau, Oct 21, 2021
  3. Junio C HamanoOct 21, 2021
  4. Ævar Arnfjörð BjarmasonOct 21, 2021
  5. Junio C HamanoOct 21, 2021
  6. Taylor BlauOct 22, 2021
  7. 02/11 midx.c: don't leak MIDX from verify_midx_fileTaylor Blau, Oct 21, 2021
  8. Eric SunshineOct 21, 2021
  9. Junio C HamanoOct 21, 2021
  10. Junio C HamanoOct 21, 2021
  11. 03/11 t/helper/test-read-midx.c: free MIDX within read_midx_file()Taylor Blau, Oct 21, 2021
  12. 04/11 builtin/pack-objects.c: don't leak memory via argumentsTaylor Blau, Oct 21, 2021
  13. 05/11 builtin/repack.c: avoid leaking child argumentsTaylor Blau, Oct 21, 2021
  14. Derrick StoleeOct 21, 2021
  15. Junio C HamanoOct 21, 2021
  16. Junio C HamanoOct 21, 2021
  17. Taylor BlauOct 22, 2021
  18. 06/11 builtin/multi-pack-index.c: don't leak concatenated optionsTaylor Blau, Oct 21, 2021
  19. 07/11 pack-bitmap.c: avoid leaking via midx_bitmap_filename()Taylor Blau, Oct 21, 2021
  20. Junio C HamanoOct 21, 2021
  21. Taylor BlauOct 22, 2021
  22. 08/11 pack-bitmap.c: don't leak type-level bitmapsTaylor Blau, Oct 21, 2021
  23. Junio C HamanoOct 21, 2021
  24. 09/11 pack-bitmap.c: more aggressively free in free_bitmap_index()Taylor Blau, Oct 21, 2021
  25. Eric SunshineOct 21, 2021
  26. Junio C HamanoOct 21, 2021
  27. Taylor BlauOct 22, 2021
  28. Junio C HamanoOct 21, 2021
  29. 11/11 t5319: UNLEAK() the remaining leaksTaylor Blau, Oct 21, 2021
  30. 10/11 pack-bitmap-write.c: don't return without stop_progress()Taylor Blau, Oct 21, 2021
  31. Eric SunshineOct 21, 2021
  32. Ævar Arnfjörð BjarmasonOct 21, 2021
  33. Junio C HamanoOct 21, 2021
  34. Taylor BlauOct 22, 2021
  35. Junio C HamanoOct 23, 2021
  36. SubmittingPatchs: clarify choice of base and testingJunio C Hamano, Oct 23, 2021
  37. Ævar Arnfjörð BjarmasonOct 23, 2021
  38. Junio C HamanoOct 23, 2021
  39. Junio C HamanoOct 23, 2021
  40. Fabian StelzerOct 25, 2021
  41. Junio C HamanoOct 25, 2021
  42. Junio C HamanoOct 25, 2021
  43. Junio C HamanoOct 25, 2021
  44. SubmittingPatchs: clarify choice of base and testingJunio C Hamano, Dec 23, 2021
  45. Elijah NewrenDec 28, 2021
  46. Fabian StelzerDec 30, 2021
  47. Re* [PATCH v2] SubmittingPatchs: clarify choice of base and testingJunio C Hamano, Dec 30, 2021
  48. Ævar Arnfjörð BjarmasonOct 21, 2021
  49. Taylor BlauOct 22, 2021
  50. Ævar Arnfjörð BjarmasonOct 22, 2021
  51. leak tests: add an interface to the LSAN_OPTIONS "suppressions"Ævar Arnfjörð Bjarmason, Oct 22, 2021
  52. Taylor BlauOct 26, 2021
  53. Jeff KingOct 26, 2021
  54. Taylor BlauOct 26, 2021
  55. Jeff KingOct 26, 2021
  56. Ævar Arnfjörð BjarmasonOct 27, 2021
  57. Jeff KingOct 27, 2021
  58. Junio C HamanoOct 27, 2021
  59. Ævar Arnfjörð BjarmasonOct 27, 2021
  60. Jeff KingOct 29, 2021
  61. Jeff KingOct 29, 2021
  62. Ævar Arnfjörð BjarmasonOct 27, 2021
  63. Derrick StoleeOct 21, 2021

Read the whole thread, see it on lore, or plain text.

$ cat FOOTERMessages come from the public archive at lore.kernel.org/git, fetched every hour. The front page is chosen and written each morning by an AI editor and can be wrong; the threads themselves are the record. About and API. For agents: an MCP server at https://gitlist.dev/mcp, and any thread, story or person page as Markdown by adding .md to its URL (or sending Accept: text/markdown). Details in /llms.txt.