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

[PATCH] coccicheck: optionally batch spatch invocations

From
Jeff King <peff@peff.net>
Date
May 6, 2019, 23:43 UTC
Message-ID
<20190506234334.GA13296@sigill.intra.peff.net>
In-Reply-To
<CACsJy8CLjUdHCro8QJfTozMB0xVWppHuFRSLCvFSaeKO_PxAog@mail.gmail.com>
On Mon, May 06, 2019 at 04:34:09PM +0700, Duy Nguyen wrote:
Show 8 quoted lines
> > However, it comes at a cost. The RSS of each spatch process goes from
> > ~50MB to ~1500MB (and peak memory usage may be even higher if make runs
> 
> 1.5G should be fine. Trying...
> 
> Even with no -j, my htop's RES column goes up 6GB and put my laptop in
> "swap every bit of memory out, including the bits handling the screen"
> mode :( I don't think it was even the peak.

Interesting if you have a different version of spatch. I'm using 1.0.4 from Debian unstable.

I had just been eyeballing the values in "top" before, but I actually measured more carefully. My peak was actually ~1900MB.

Show 5 quoted lines
> It's probably a bit too much to ask, but is it possible to handle N
> files at a time (instead of all files), which consumes less memory and
> runs a bit slower, but still better than the default mode? I can see
> it already gets tricky doing complicated stuff in Makefile so "no" is
> perfectly ok.

I almost did this initially but I feared that nobody would actually use it. :) So given at least one person who wants it, I took a look. If we rely on xargs, then it is really not too bad (and is in fact shorter than the current code). I also wrote up a pure-shell version, but it's rather verbose even after taking some shortcuts with whitespace splitting.

So here's what I think we should apply:
-- >8 --
Subject: [PATCH] coccicheck: optionally batch spatch invocations

In our "make coccicheck" rule, we currently feed each source file to its own individual invocation of spatch. This has a few downsides:

  - it repeats any overhead spatch has for starting up and reading the
    patch file
  - any included header files may get processed from multiple
    invocations. This is slow (we see the same header files multiple
    times) and may produce a resulting patch with repeated hunks (which
    cannot be applied without further cleanup)

Ideally we'd just invoke a single instance of spatch per rule-file and feed it all source files. But spatch can be rather memory hungry when run in this way. I measured the peak RSS going from ~90MB for a single file to ~1900MB for all files. Multiplied by multiple rule files being processed at the same time (for "make -j"), this can make things slower or even cause them to fail (e.g., this is reported to happen on our Travis builds).

Instead, let's provide a tunable knob. We'll leave the default at "1", but it can be cranked up to "999" for maximum CPU/memory tradeoff, or people can find points in between that serve their particular machines.

Here are a few numbers running a single rule via:
  SIZES='1 4 16 999'
  RULE=contrib/coccinelle/object_id.cocci
  for i in $SIZES; do
    make clean
    /usr/bin/time -o $i.out --format='%e | %U | %S | %M' \
      make $RULE.patch SPATCH_BATCH_SIZE=$i
  done
  for i in $SIZES; do
    printf '%4d | %s\n' $i "$(cat $i.out)"
  done
which yields:
     1 | 97.73 | 93.38 | 4.33 | 100128
     4 | 52.80 | 51.14 | 1.69 | 135204
    16 | 35.82 | 35.09 | 0.76 | 284124
   999 | 23.30 | 23.13 | 0.20 | 1903852

The implementation is done with xargs, which should be widely available; it's in POSIX, we rely on it already in the test suite. And "coccicheck" is really a developer-only tool anyway, so it's not a big deal if obscure systems can't run it.

Signed-off-by: Jeff King <peff@peff.net>
---
I left the default at 1 for safety. Probably 4 or 16 would be an OK
default, but I don't have any interest in figuring out exactly what
Travis or some hypothetical average machine can handle. I'll be setting
mine to 999. ;)

Making "0" work as "unlimited" might be nice, but xargs doesn't support that and I didn't want to make the recipe any more unreadable than it already is.

 Makefile | 13 ++++++-------
 1 file changed, 6 insertions(+), 7 deletions(-)
diff --git a/Makefile b/Makefile
index 9f1b6e8926..daba958b8f 100644
--- a/Makefile
+++ b/Makefile
@@ -1174,8 +1174,10 @@ PTHREAD_CFLAGS =
 SPARSE_FLAGS ?=
 SP_EXTRA_FLAGS =
 
-# For the 'coccicheck' target
+# For the 'coccicheck' target; setting SPATCH_BATCH_SIZE higher will
+# usually result in less CPU usage at the cost of higher peak memory.
 SPATCH_FLAGS = --all-includes --patch .
+SPATCH_BATCH_SIZE = 1
 
 include config.mak.uname
 -include config.mak.autogen
@@ -2790,12 +2792,9 @@ endif
 
 %.cocci.patch: %.cocci $(COCCI_SOURCES)
 	@echo '    ' SPATCH $<; \
-	ret=0; \
-	for f in $(COCCI_SOURCES); do \
-		$(SPATCH) --sp-file $< $$f $(SPATCH_FLAGS) || \
-			{ ret=$$?; break; }; \
-	done >$@+ 2>$@.log; \
-	if test $$ret != 0; \
+	if ! echo $(COCCI_SOURCES) | xargs -n $(SPATCH_BATCH_SIZE) \
+		$(SPATCH) --sp-file $< $(SPATCH_FLAGS) \
+		>$@+ 2>$@.log; \
 	then \
 		cat $@.log; \
 		exit 1; \
-- 
2.21.0.1314.g224b191707
Previous: Duy NguyenNext: Jacob Keller
Message 34 of 103 in “*.h: remove extern from function declarations”
  1. *.h: remove extern from function declarationsDenton Liu, Apr 12, 2019
  2. Jeff KingApr 13, 2019
  3. Junio C HamanoApr 13, 2019
  4. 0/3 remove extern from function declarationsDenton Liu, Apr 15, 2019
  5. 2/3 *.[ch]: remove extern from function declarations using sedDenton Liu, Apr 15, 2019
  6. 3/3 cocci: prevent extern function declarationsDenton Liu, Apr 15, 2019
  7. 1/3 *.[ch]: remove extern from function declarations using spatchDenton Liu, Apr 15, 2019
  8. Thomas GummererApr 15, 2019
  9. 0/4 remove extern from function declarationsDenton Liu, Apr 17, 2019
  10. 2/4 *.[ch]: remove extern from function declarations using sedDenton Liu, Apr 17, 2019
  11. 1/4 *.[ch]: remove extern from function declarations using spatchDenton Liu, Apr 17, 2019
  12. 4/4 cocci: prevent extern function declarationsDenton Liu, Apr 17, 2019
  13. 3/4 *.[ch]: manually align parameter listsDenton Liu, Apr 17, 2019
  14. cache.h: fix mismerge of 'dl/no-extern-in-func-decl'Denton Liu, Apr 22, 2019
  15. Junio C HamanoApr 22, 2019
  16. Junio C HamanoApr 22, 2019
  17. Jeff KingApr 22, 2019
  18. SZEDER GáborApr 25, 2019
  19. Denton LiuApr 25, 2019
  20. Johannes SchindelinApr 30, 2019
  21. Denton LiuMay 1, 2019
  22. Jeff KingMay 1, 2019
  23. SZEDER GáborMay 2, 2019
  24. Johannes SchindelinMay 3, 2019
  25. SZEDER GáborMay 3, 2019
  26. SZEDER GáborMay 3, 2019
  27. Jeff KingMay 3, 2019
  28. SZEDER GáborMay 3, 2019
  29. Junio C HamanoMay 5, 2019
  30. Jacob KellerMay 5, 2019
  31. Jacob KellerMay 5, 2019
  32. coccicheck: optionally process every source file at onceJeff King, May 6, 2019
  33. Duy NguyenMay 6, 2019
  34. coccicheck: optionally batch spatch invocationsJeff King, May 6, 2019
  35. Jacob KellerMay 7, 2019
  36. Jeff KingMay 7, 2019
  37. Junio C HamanoMay 7, 2019
  38. Jeff KingMay 7, 2019
  39. Jacob KellerMay 7, 2019
  40. Junio C HamanoMay 7, 2019
  41. Jeff KingMay 8, 2019
  42. Denton LiuMay 8, 2019
  43. Jeff KingMay 8, 2019
  44. Duy NguyenMay 7, 2019
  45. SZEDER GáborMay 7, 2019
  46. Makefile: fix bugs in coccicheck and speed it upÆvar Arnfjörð Bjarmason, Mar 2, 2021
  47. Ævar Arnfjörð BjarmasonMar 3, 2021
  48. Denton LiuMar 3, 2021
  49. Junio C HamanoMar 4, 2021
  50. Ævar Arnfjörð BjarmasonMar 5, 2021
  51. Jeff KingMar 5, 2021
  52. Ævar Arnfjörð BjarmasonMar 5, 2021
  53. Jeff KingMar 6, 2021
  54. 1/4 Makefile/coccicheck: add comment heading for all SPATCH flagsÆvar Arnfjörð Bjarmason, Mar 5, 2021
  55. 0/4 Makefile/coccicheck: fix bugs and speed it upÆvar Arnfjörð Bjarmason, Mar 5, 2021
  56. René Scharfe.Mar 5, 2021
  57. Ævar Arnfjörð BjarmasonMar 6, 2021
  58. René Scharfe.Mar 6, 2021
  59. Ævar Arnfjörð BjarmasonMar 6, 2021
  60. René Scharfe.Mar 6, 2021
  61. 3/4 Makefile/coccicheck: allow for setting xargs concurrencyÆvar Arnfjörð Bjarmason, Mar 5, 2021
  62. Jeff KingMar 6, 2021
  63. 4/4 Makefile/coccicheck: set SPATCH_BATCH_SIZE to 8Ævar Arnfjörð Bjarmason, Mar 5, 2021
  64. 5/4 Makefile/coccicheck: use --include-headers-for-typesÆvar Arnfjörð Bjarmason, Mar 6, 2021
  65. SZEDER GáborMar 18, 2021
  66. Ævar Arnfjörð BjarmasonMar 19, 2021
  67. 0/4 Makefile/coccicheck: fix bugs and speed it upÆvar Arnfjörð Bjarmason, Mar 22, 2021
  68. 1/4 Makefile/coccicheck: add comment heading for all SPATCH flagsÆvar Arnfjörð Bjarmason, Mar 22, 2021
  69. René Scharfe.Mar 22, 2021
  70. 4/4 Makefile/coccicheck: set SPATCH_BATCH_SIZE to 8Ævar Arnfjörð Bjarmason, Mar 22, 2021
  71. René Scharfe.Mar 22, 2021
  72. Jeff KingMar 24, 2021
  73. 2/4 Makefile/coccicheck: speed up and fix bug with duplicate hunksÆvar Arnfjörð Bjarmason, Mar 22, 2021
  74. René Scharfe.Mar 22, 2021
  75. Jeff KingMar 24, 2021
  76. Junio C HamanoMar 22, 2021
  77. 3/4 Makefile/coccicheck: allow for setting xargs concurrencyÆvar Arnfjörð Bjarmason, Mar 22, 2021
  78. Jeff KingMar 24, 2021
  79. Ævar Arnfjörð BjarmasonMar 25, 2021
  80. Jeff KingMar 26, 2021
  81. Junio C HamanoMar 27, 2021
  82. Ævar Arnfjörð BjarmasonMar 27, 2021
  83. 2/4 Makefile/coccicheck: speed up and fix bug with duplicate hunksÆvar Arnfjörð Bjarmason, Mar 5, 2021
  84. Jeff KingMar 6, 2021
  85. Ævar Arnfjörð BjarmasonMar 6, 2021
  86. Denton LiuMay 3, 2019
  87. 0/4 remove extern from function declarationsDenton Liu, Apr 23, 2019
  88. 2/4 *.[ch]: remove extern from function declarations using sedDenton Liu, Apr 23, 2019
  89. Junio C HamanoApr 24, 2019
  90. Denton LiuApr 25, 2019
  91. 3/4 *.[ch]: manually align parameter listsDenton Liu, Apr 23, 2019
  92. 4/4 cocci: prevent extern function declarationsDenton Liu, Apr 23, 2019
  93. 1/4 *.[ch]: remove extern from function declarations using spatchDenton Liu, Apr 23, 2019
  94. 0/3 *** SUBJECT HERE ***Denton Liu, Apr 29, 2019
  95. 2/3 *.[ch]: remove extern from function declarations using sedDenton Liu, Apr 29, 2019
  96. 1/3 *.[ch]: remove extern from function declarations using spatchDenton Liu, Apr 29, 2019
  97. 3/3 *.[ch]: manually align parameter listsDenton Liu, Apr 29, 2019
  98. Denton LiuApr 29, 2019
  99. Ævar Arnfjörð BjarmasonMay 6, 2019
  100. Denton LiuMay 6, 2019
  101. René Scharfe.Mar 6, 2021
  102. René Scharfe.Mar 6, 2021
  103. René Scharfe.Mar 13, 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.