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

Re: [PATCH v2 11/27] git: refactor builtin handling to use a `struct strvec`

From
Toon Claes <toon@iotcl.com>
Date
Nov 20, 2024, 10:38 UTC
Message-ID
<8734jmtcjs.fsf@iotcl.com>
In-Reply-To
<20241111-b4-pks-leak-fixes-pt10-v2-11-6154bf91f0b0@pks.im>
Patrick Steinhardt <ps@pks.im> writes:
Show 105 quoted lines
> Similar as with the preceding commit, `handle_builtin()` does not
> properly track lifetimes of the `argv` array and its strings. As it may
> end up modifying the array this can lead to memory leaks in case it
> contains allocated strings.
>
> Refactor the function to use a `struct strvec` instead.
>
> Signed-off-by: Patrick Steinhardt <ps@pks.im>
> ---
>  git.c                  | 66 ++++++++++++++++++++++++--------------------------
>  t/t0211-trace2-perf.sh |  2 +-
>  2 files changed, 32 insertions(+), 36 deletions(-)
>
> diff --git a/git.c b/git.c
> index 88356afe5fb568ccc147f055e3ab253c53a1befa..159dd45b08204c4a89d1dc4ab6990978e2454eb6 100644
> --- a/git.c
> +++ b/git.c
> @@ -696,63 +696,57 @@ void load_builtin_commands(const char *prefix, struct cmdnames *cmds)
>  }
>  
>  #ifdef STRIP_EXTENSION
> -static void strip_extension(const char **argv)
> +static void strip_extension(struct strvec *args)
>  {
>  	size_t len;
>  
> -	if (strip_suffix(argv[0], STRIP_EXTENSION, &len))
> -		argv[0] = xmemdupz(argv[0], len);
> +	if (strip_suffix(args->v[0], STRIP_EXTENSION, &len)) {
> +		char *stripped = xmemdupz(args->v[0], len);
> +		strvec_replace(args, 0, stripped);
> +		free(stripped);
> +	}
>  }
>  #else
>  #define strip_extension(cmd)
>  #endif
>  
> -static void handle_builtin(int argc, const char **argv)
> +static void handle_builtin(struct strvec *args)
>  {
> -	struct strvec args = STRVEC_INIT;
> -	const char **argv_copy = NULL;
>  	const char *cmd;
>  	struct cmd_struct *builtin;
>  
> -	strip_extension(argv);
> -	cmd = argv[0];
> +	strip_extension(args);
> +	cmd = args->v[0];
>  
>  	/* Turn "git cmd --help" into "git help --exclude-guides cmd" */
> -	if (argc > 1 && !strcmp(argv[1], "--help")) {
> -		int i;
> -
> -		argv[1] = argv[0];
> -		argv[0] = cmd = "help";
> -
> -		for (i = 0; i < argc; i++) {
> -			strvec_push(&args, argv[i]);
> -			if (!i)
> -				strvec_push(&args, "--exclude-guides");
> -		}
> +	if (args->nr > 1 && !strcmp(args->v[1], "--help")) {
> +		const char *exclude_guides_arg[] = { "--exclude-guides" };
> +
> +		strvec_replace(args, 1, args->v[0]);
> +		strvec_replace(args, 0, "help");
> +		cmd = "help";
> +		strvec_splice(args, 2, 0, exclude_guides_arg,
> +			      ARRAY_SIZE(exclude_guides_arg));
> +	}
>  
> -		argc++;
> +	builtin = get_builtin(cmd);
> +	if (builtin) {
> +		const char **argv_copy = NULL;
> +		int ret;
>  
>  		/*
>  		 * `run_builtin()` will modify the argv array, so we need to
>  		 * create a shallow copy such that we can free all of its
>  		 * strings.
>  		 */
> -		CALLOC_ARRAY(argv_copy, argc + 1);
> -		COPY_ARRAY(argv_copy, args.v, argc);
> +		if (args->nr)
> +			DUP_ARRAY(argv_copy, args->v, args->nr + 1);
>  
> -		argv = argv_copy;
> -	}
> -
> -	builtin = get_builtin(cmd);
> -	if (builtin) {
> -		int ret = run_builtin(builtin, argc, argv, the_repository);
> -		strvec_clear(&args);
> +		ret = run_builtin(builtin, args->nr, argv_copy, the_repository);
> +		strvec_clear(args);
>  		free(argv_copy);
>  		exit(ret);
>  	}
> -
> -	strvec_clear(&args);
> -	free(argv_copy);
>  }

Just want to give you a little shout out how you've reorganized the code and it now makes a lot more sense how `argv_copy` is used by putting it inside the `if (builtin)` code block.

-- 
Toon
Previous: Patrick SteinhardtNext: Patrick Steinhardt
Message 22 of 45 in “Memory leak fixes (pt.10, final)”
  1. 00/27 Memory leak fixes (pt.10, final)Patrick Steinhardt, Nov 11, 2024
  2. 01/27 builtin/blame: fix leaking blame entries with `--incremental`Patrick Steinhardt, Nov 11, 2024
  3. 02/27 bisect: fix leaking good/bad terms when reading multipe timesPatrick Steinhardt, Nov 11, 2024
  4. 03/27 bisect: fix leaking string in `handle_bad_merge_base()`Patrick Steinhardt, Nov 11, 2024
  5. 04/27 bisect: fix leaking `current_bad_oid`Patrick Steinhardt, Nov 11, 2024
  6. 05/27 bisect: fix multiple leaks in `bisect_next_all()`Patrick Steinhardt, Nov 11, 2024
  7. 06/27 bisect: fix leaking commit list items in `check_merge_base()`Patrick Steinhardt, Nov 11, 2024
  8. 07/27 bisect: fix various cases where we leak commit list itemsPatrick Steinhardt, Nov 11, 2024
  9. Toon ClaesNov 20, 2024
  10. Patrick SteinhardtNov 20, 2024
  11. 08/27 line-log: fix leak when rewriting commit parentsPatrick Steinhardt, Nov 11, 2024
  12. 09/27 strvec: introduce new `strvec_splice()` functionPatrick Steinhardt, Nov 11, 2024
  13. Toon ClaesNov 20, 2024
  14. Patrick SteinhardtNov 20, 2024
  15. Junio C HamanoNov 20, 2024
  16. Jeff KingNov 21, 2024
  17. Jeff KingNov 21, 2024
  18. Doxygen-styled comments [was: Re: [PATCH v2 09/27] strvec: introduce new `strvec_splice()` function]Toon Claes, Nov 21, 2024
  19. Jeff KingNov 21, 2024
  20. 10/27 git: refactor alias handling to use a `struct strvec`Patrick Steinhardt, Nov 11, 2024
  21. 11/27 git: refactor builtin handling to use a `struct strvec`Patrick Steinhardt, Nov 11, 2024
  22. Toon ClaesNov 20, 2024
  23. 12/27 split-index: fix memory leak in `move_cache_to_base_index()`Patrick Steinhardt, Nov 11, 2024
  24. 13/27 builtin/sparse-checkout: fix leaking sanitized patternsPatrick Steinhardt, Nov 11, 2024
  25. 14/27 help: refactor to not use globals for reading configPatrick Steinhardt, Nov 11, 2024
  26. 15/27 help: fix leaking `struct cmdnames`Patrick Steinhardt, Nov 11, 2024
  27. 16/27 help: fix leaking return value from `help_unknown_cmd()`Patrick Steinhardt, Nov 11, 2024
  28. 17/27 builtin/help: fix leaks in `check_git_cmd()`Patrick Steinhardt, Nov 11, 2024
  29. 18/27 builtin/init-db: fix leaking directory pathsPatrick Steinhardt, Nov 11, 2024
  30. 19/27 builtin/branch: fix leaking sorting optionsPatrick Steinhardt, Nov 11, 2024
  31. 20/27 t/helper: fix leaking commit graph in "read-graph" subcommandPatrick Steinhardt, Nov 11, 2024
  32. 21/27 global: drop `UNLEAK()` annotationPatrick Steinhardt, Nov 11, 2024
  33. Jeff KingNov 12, 2024
  34. Patrick SteinhardtNov 12, 2024
  35. Jeff KingNov 12, 2024
  36. 22/27 git-compat-util: drop now-unused `UNLEAK()` macroPatrick Steinhardt, Nov 11, 2024
  37. 23/27 t5601: work around leak sanitizer issuePatrick Steinhardt, Nov 11, 2024
  38. 24/27 t: mark some tests as leak freePatrick Steinhardt, Nov 11, 2024
  39. 25/27 t: remove unneeded !SANITIZE_LEAK prerequisitesPatrick Steinhardt, Nov 11, 2024
  40. 26/27 test-lib: unconditionally enable leak checkingPatrick Steinhardt, Nov 11, 2024
  41. 27/27 t: remove TEST_PASSES_SANITIZE_LEAK annotationsPatrick Steinhardt, Nov 11, 2024
  42. Toon ClaesNov 20, 2024
  43. Patrick SteinhardtNov 20, 2024
  44. Rubén JustoNov 11, 2024
  45. Rubén JustoNov 12, 2024

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.