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

Re: [RFC PATCH v3 2/3] grep: make PCRE2 aware of custom allocator

From
René Scharfe <l.s.r@web.de>
Date
Aug 7, 2019, 05:38 UTC
Message-ID
<ab8a378c-0a60-9554-b2dd-ecb3d05229cb@web.de>
In-Reply-To
<20190806163658.66932-3-carenas@gmail.com>
Am 06.08.19 um 18:36 schrieb Carlo Marcelo Arenas Belón:
Show 20 quoted lines
> 94da9193a6 (grep: add support for PCRE v2, 2017-06-01) didn't include
> a way to override the system allocator, and so it is incompatible with
> USE_NED_ALLOCATOR.  The problem was made visible when an attempt to
> avoid a leak in a data structure that is created by the library was
> passed to NED's free for disposal triggering a segfault in Windows.
>
> PCRE2 requires the use of a general context to override the allocator
> and therefore, there is a lot more code needed than in PCRE1, including
> a couple of wrapper functions.
>
> Extend the grep API with a "destructor" that could be called to cleanup
> any objects that were created and used globally.
>
> Update builtin/grep to use that new API, but any other future
> users should make sure to have matching grep_init/grep_destroy calls
> if they are using the pattern matching functionality (currently only
> relevant when using both NED and PCRE2)
>
> Move some of the logic that was before done per thread (in the workers)
> into an earlier phase to avoid degrading performance
Which logic is moved?  In the patch I basically only see additions.
Show 6 quoted lines
>, but as the use
> of PCRE2 with NED is better understood it is expected more of its
> functions will be instructed to use the custom allocator as well as
> was done in the original code[1] this work was based on.
>
> [1] https://public-inbox.org/git/3397e6797f872aedd18c6d795f4976e1c579514b.1565005867.git.gitgitgadget@gmail.com/

I'm not sure I understand that part. Do you mean there are gaps of knowledge about nedmalloc and/or PCRE2 and/or their interaction? And someone is working on closing those gaps, and is going to submit more patches in the process?

Your patch uses a custom global context only if USE_NED_ALLOCATOR is defined, while [1] does it unconditionally. The latter is easier to debug and requires less preprocessor directives. What's the upside of your approach?

Show 35 quoted lines
>
> Reported-by: Johannes Schindelin <johannes.schindelin@gmx.de>
> Signed-off-by: Carlo Marcelo Arenas Belón <carenas@gmail.com>
> ---
>  builtin/grep.c |  1 +
>  grep.c         | 36 +++++++++++++++++++++++++++++++++++-
>  grep.h         |  1 +
>  3 files changed, 37 insertions(+), 1 deletion(-)
>
> diff --git a/builtin/grep.c b/builtin/grep.c
> index 560051784e..e49c20df60 100644
> --- a/builtin/grep.c
> +++ b/builtin/grep.c
> @@ -1145,5 +1145,6 @@ int cmd_grep(int argc, const char **argv, const char *prefix)
>  		run_pager(&opt, prefix);
>  	clear_pathspec(&pathspec);
>  	free_grep_patterns(&opt);
> +	grep_destroy();
>  	return !hit;
>  }
> diff --git a/grep.c b/grep.c
> index 0154998695..3e5bdf94a6 100644
> --- a/grep.c
> +++ b/grep.c
> @@ -16,6 +16,22 @@ static int grep_source_is_binary(struct grep_source *gs,
>
>  static struct grep_opt grep_defaults;
>
> +#ifdef USE_LIBPCRE2
> +static pcre2_general_context *pcre2_global_context;
> +
> +#ifdef USE_NED_ALLOCATOR
> +static void *pcre2_malloc(PCRE2_SIZE size, MAYBE_UNUSED void *memory_data)
> +{
> +	return malloc(size);
Should this be xmalloc() to get consistent out-of-memory handling?
Show 57 quoted lines
> +}
> +
> +static void pcre2_free(void *pointer, MAYBE_UNUSED void *memory_data)
> +{
> +	return free(pointer);
> +}
> +#endif
> +#endif
> +
>  static const char *color_grep_slots[] = {
>  	[GREP_COLOR_CONTEXT]	    = "context",
>  	[GREP_COLOR_FILENAME]	    = "filename",
> @@ -153,6 +169,7 @@ int grep_config(const char *var, const char *value, void *cb)
>   *
>   * If using PCRE make sure that the library is configured
>   * to use the right allocator (ex: NED)
> + * if any object is created it should be cleaned up in grep_destroy()
>   */
>  void grep_init(struct grep_opt *opt, struct repository *repo, const char *prefix)
>  {
> @@ -188,6 +205,13 @@ void grep_init(struct grep_opt *opt, struct repository *repo, const char *prefix
>  		color_set(opt->colors[i], def->colors[i]);
>  }
>
> +void grep_destroy(void)
> +{
> +#ifdef USE_LIBPCRE2
> +	pcre2_general_context_free(pcre2_global_context);
> +#endif
> +}
> +
>  static void grep_set_pattern_type_option(enum grep_pattern_type pattern_type, struct grep_opt *opt)
>  {
>  	/*
> @@ -319,6 +343,11 @@ void append_header_grep_pattern(struct grep_opt *opt,
>  void append_grep_pattern(struct grep_opt *opt, const char *pat,
>  			 const char *origin, int no, enum grep_pat_token t)
>  {
> +#if defined(USE_LIBPCRE2) && defined(USE_NED_ALLOCATOR)
> +	if (!pcre2_global_context && opt->ignore_case && has_non_ascii(pat))
> +		pcre2_global_context = pcre2_general_context_create(
> +					pcre2_malloc, pcre2_free, NULL);
> +#endif
>  	append_grep_pat(opt, pat, strlen(pat), origin, no, t);
>  }
>
> @@ -507,9 +536,14 @@ static void compile_pcre2_pattern(struct grep_pat *p, const struct grep_opt *opt
>
>  	p->pcre2_compile_context = NULL;
>
> +	/* pcre2_global_context is initialized in append_grep_pattern */
>  	if (opt->ignore_case) {
>  		if (has_non_ascii(p->pattern)) {
> -			character_tables = pcre2_maketables(NULL);
> +#ifdef USE_NED_ALLOCATOR
> +			if (!pcre2_global_context)
> +				BUG("pcre2_global_context uninitialized");

[1] initializes on demand. Why not do that? To avoid race conditions that would lead to occasional double-allocation of the global context?

> +#endif
> +			character_tables = pcre2_maketables(pcre2_global_context);
>  			p->pcre2_compile_context = pcre2_compile_context_create(NULL);

Don't you want to pass pcre2_global_context here as well? And [1] even uses it in some more places.

Oh, that's the "expected more" when "better understood" part from the commit message, right?

Basically I'd expect the custom global context to be used for all PCRE2 allocations; I can't think of a reason for mixing allocators (e.g. system malloc for PCRE2 regexes and nedmalloc for everything else).

Show 15 quoted lines
>  			pcre2_set_character_tables(p->pcre2_compile_context, character_tables);
>  		}
> diff --git a/grep.h b/grep.h
> index 1875880f37..526c2db9ef 100644
> --- a/grep.h
> +++ b/grep.h
> @@ -189,6 +189,7 @@ struct grep_opt {
>  void init_grep_defaults(struct repository *);
>  int grep_config(const char *var, const char *value, void *);
>  void grep_init(struct grep_opt *, struct repository *repo, const char *prefix);
> +void grep_destroy(void);
>  void grep_commit_pattern_type(enum grep_pattern_type, struct grep_opt *opt);
>
>  void append_grep_pat(struct grep_opt *opt, const char *pat, size_t patlen, const char *origin, int no, enum grep_pat_token t);
>
Previous: Carlo Marcelo Arenas BelónNext: Carlo Arenas
Message 19 of 67 in “Fix a problem with PCRE2 and nedmalloc, found via Azure Pipelines”
  1. 0/1 Fix a problem with PCRE2 and nedmalloc, found via Azure PipelinesJohannes Schindelin via GitGitGadget, Aug 5, 2019
  2. 1/1 pcre2: allow overriding the system allocatorJohannes Schindelin via GitGitGadget, Aug 5, 2019
  3. Carlo ArenasAug 5, 2019
  4. Carlo ArenasAug 5, 2019
  5. Johannes SchindelinAug 5, 2019
  6. Johannes SchindelinAug 5, 2019
  7. Junio C HamanoAug 5, 2019
  8. Carlo ArenasAug 6, 2019
  9. 0/3 grep: no leaks (WIP)Carlo Marcelo Arenas Belón, Aug 6, 2019
  10. 1/3 grep: make PCRE1 aware of custom allocatorCarlo Marcelo Arenas Belón, Aug 6, 2019
  11. Johannes SchindelinAug 8, 2019
  12. Carlo ArenasAug 8, 2019
  13. 3/3 grep: avoid leak of chartables in PCRE2Carlo Marcelo Arenas Belón, Aug 6, 2019
  14. 2/3 grep: make PCRE2 aware of custom allocatorCarlo Marcelo Arenas Belón, Aug 6, 2019
  15. Johannes SchindelinAug 8, 2019
  16. Carlo ArenasAug 8, 2019
  17. 0/3 grep: no leaks or crashes (windows testing needed)Carlo Marcelo Arenas Belón, Aug 6, 2019
  18. 2/3 grep: make PCRE2 aware of custom allocatorCarlo Marcelo Arenas Belón, Aug 6, 2019
  19. René ScharfeAug 7, 2019
  20. Carlo ArenasAug 7, 2019
  21. René ScharfeAug 7, 2019
  22. 1/2 nedmalloc: do assignments only after the declaration sectionRené Scharfe, Aug 7, 2019
  23. 2/2 nedmalloc: avoid compiler warning about unused valueRené Scharfe, Aug 7, 2019
  24. Carlo ArenasAug 8, 2019
  25. René ScharfeAug 8, 2019
  26. Carlo ArenasAug 8, 2019
  27. René ScharfeAug 8, 2019
  28. Johannes SchindelinAug 8, 2019
  29. Junio C HamanoAug 7, 2019
  30. 1/3 grep: make PCRE1 aware of custom allocatorCarlo Marcelo Arenas Belón, Aug 6, 2019
  31. 3/3 grep: avoid leak of chartables in PCRE2Carlo Marcelo Arenas Belón, Aug 6, 2019
  32. Junio C HamanoAug 6, 2019
  33. 0/3 grep: no leaks or crashes (windows testing needed)Carlo Marcelo Arenas Belón, Aug 7, 2019
  34. 1/3 grep: make PCRE1 aware of custom allocatorCarlo Marcelo Arenas Belón, Aug 7, 2019
  35. 2/3 grep: make PCRE2 aware of custom allocatorCarlo Marcelo Arenas Belón, Aug 7, 2019
  36. Junio C HamanoAug 7, 2019
  37. 3/3 grep: avoid leak of chartables in PCRE2Carlo Marcelo Arenas Belón, Aug 7, 2019
  38. 0/3 grep: almost no more leaks, hopefully no crashesCarlo Marcelo Arenas Belón, Aug 9, 2019
  39. 1/3 grep: make PCRE1 aware of custom allocatorCarlo Marcelo Arenas Belón, Aug 9, 2019
  40. 3/3 grep: avoid leak of chartables in PCRE2Carlo Marcelo Arenas Belón, Aug 9, 2019
  41. 2/3 grep: make PCRE2 aware of custom allocatorCarlo Marcelo Arenas Belón, Aug 9, 2019
  42. Johannes SchindelinAug 27, 2019
  43. Carlo ArenasAug 27, 2019
  44. Junio C HamanoOct 3, 2019
  45. Johannes SchindelinOct 3, 2019
  46. Carlo ArenasOct 3, 2019
  47. Johannes SchindelinOct 3, 2019
  48. Junio C HamanoOct 3, 2019
  49. Carlo ArenasAug 9, 2019
  50. René ScharfeAug 9, 2019
  51. Junio C HamanoAug 9, 2019
  52. Johannes SchindelinAug 9, 2019
  53. SQUASHCarlo Marcelo Arenas Belón, Aug 10, 2019
  54. René ScharfeAug 10, 2019
  55. Carlo ArenasAug 10, 2019
  56. René ScharfeAug 10, 2019
  57. Carlo ArenasAug 12, 2019
  58. René ScharfeAug 12, 2019
  59. Carlo ArenasAug 12, 2019
  60. Johannes SchindelinAug 10, 2019
  61. Carlo ArenasAug 10, 2019
  62. René ScharfeAug 10, 2019
  63. Carlo ArenasAug 10, 2019
  64. René ScharfeAug 10, 2019
  65. Johannes SchindelinAug 8, 2019
  66. Carlo ArenasAug 9, 2019
  67. Johannes SchindelinAug 9, 2019

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.