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

[PATCH 0/4] hash.h: support choosing a separate SHA-1 for non-cryptographic uses

From
Taylor Blau <me@ttaylorr.com>
Date
Sep 1, 2024, 16:03 UTC
Message-ID
<cover.1725206584.git.me@ttaylorr.com>

This series adds a build-time knob to allow selecting an alternative SHA-1 implementation for non-cryptographic hashing within Git, starting with the `hashwrite()` family of functions.

This series is the result of starting to roll out verbatim multi-pack reuse within GitHub's infrastructure. I noticed that on larger repositories, it is harder thus far to measure a CPU speed-up on clones where multi-pack reuse is enabled.

After some profiling, I noticed that we spend a significant amount of time in hashwrite(), which is not all that surprising. But much of that time is wasted in GitHub's infrastructure, since we are using the same collision-detecting SHA-1 implementation to produce a trailing checksum for the pack which does not need to be cryptographically secure.

This series teaches a new set of build-time knobs: OPENSSL_SHA1_FAST, BLK_SHA1_FAST, and APPLE_COMMON_CRYPTO_SHA1_FAST, which can be used to select an alterantive SHA-1 implementation for non-cryptographic uses within Git.

The series is laid out as follows:
  - The first two patches are preparatory, allowing us to include
    multiple SHA-1 wrapper headers and adding scaffolding functions for
    the _fast() variants, respectively.
  - The third patch introduces the build-time knobs for selecting which
    SHA-1 implementation is used for non-cryptographic purposes.
  - The fourth and final patch updates the hashwrite() implementation to
    use the _fast() variants.

From the final commit, a clone of the kernel which took 19.219 seconds on my machine with no build-time modifications now only takes 11.597 seconds when compiling with OPENSSL_SHA1_FAST=1, for a ~39.7% speed-up. On a freshly-repacked copy of git.git, the time goes from 758ms to 333ms for a ~56% speed-up.

Thanks in advance for your review!
Taylor Blau (4):
  sha1: do not redefine `platform_SHA_CTX` and friends
  hash.h: scaffolding for _fast hashing variants
  Makefile: allow specifying a SHA-1 for non-cryptographic uses
  csum-file.c: use fast SHA-1 implementation when available
 Makefile          | 25 ++++++++++++++++++
 block-sha1/sha1.h |  2 ++
 csum-file.c       | 18 ++++++-------
 hash.h            | 67 +++++++++++++++++++++++++++++++++++++++++++++++
 object-file.c     | 42 +++++++++++++++++++++++++++++
 sha1/openssl.h    |  2 ++
 sha1dc_git.h      |  3 +++
 7 files changed, 150 insertions(+), 9 deletions(-)
base-commit: 159f2d50e75c17382c9f4eb7cbda671a6fa612d1
-- 
2.46.0.425.ge8f5cbd280c
Next: Taylor Blau
Message 1 of 99 in “hash.h: support choosing a separate SHA-1 for non-cryptographic uses”
  1. 0/4 hash.h: support choosing a separate SHA-1 for non-cryptographic usesTaylor Blau, Sep 1, 2024
  2. 1/4 sha1: do not redefine `platform_SHA_CTX` and friendsTaylor Blau, Sep 1, 2024
  3. Patrick SteinhardtSep 2, 2024
  4. Taylor BlauSep 3, 2024
  5. 2/4 hash.h: scaffolding for _fast hashing variantsTaylor Blau, Sep 1, 2024
  6. Patrick SteinhardtSep 2, 2024
  7. Junio C HamanoSep 3, 2024
  8. Taylor BlauSep 3, 2024
  9. Junio C HamanoSep 3, 2024
  10. Taylor BlauSep 3, 2024
  11. Patrick SteinhardtSep 4, 2024
  12. Junio C HamanoSep 4, 2024
  13. Taylor BlauSep 3, 2024
  14. 3/4 Makefile: allow specifying a SHA-1 for non-cryptographic usesTaylor Blau, Sep 1, 2024
  15. Patrick SteinhardtSep 2, 2024
  16. Taylor BlauSep 3, 2024
  17. 4/4 csum-file.c: use fast SHA-1 implementation when availableTaylor Blau, Sep 1, 2024
  18. Patrick SteinhardtSep 2, 2024
  19. brian m. carlsonSep 3, 2024
  20. Taylor BlauSep 3, 2024
  21. Junio C HamanoSep 2, 2024
  22. Taylor BlauSep 3, 2024
  23. Junio C HamanoSep 3, 2024
  24. brian m. carlsonSep 2, 2024
  25. Taylor BlauSep 3, 2024
  26. Junio C HamanoSep 3, 2024
  27. brian m. carlsonSep 4, 2024
  28. Jeff KingSep 5, 2024
  29. Junio C HamanoSep 5, 2024
  30. Taylor BlauSep 5, 2024
  31. Junio C HamanoSep 5, 2024
  32. Taylor BlauSep 5, 2024
  33. Taylor BlauSep 5, 2024
  34. Taylor BlauSep 5, 2024
  35. Jeff KingSep 5, 2024
  36. Junio C HamanoSep 5, 2024
  37. 0/4 hash.h: support choosing a separate SHA-1 for non-cryptographic usesTaylor Blau, Sep 5, 2024
  38. 1/4 sha1: do not redefine `platform_SHA_CTX` and friendsTaylor Blau, Sep 5, 2024
  39. 2/4 hash.h: scaffolding for _fast hashing variantsTaylor Blau, Sep 5, 2024
  40. 3/4 Makefile: allow specifying a SHA-1 for non-cryptographic usesTaylor Blau, Sep 5, 2024
  41. 4/4 csum-file.c: use fast SHA-1 implementation when availableTaylor Blau, Sep 5, 2024
  42. 0/9 hash.h: support choosing a separate SHA-1 for non-cryptographic usesTaylor Blau, Sep 6, 2024
  43. 1/9 finalize_object_file(): check for name collision before renamingTaylor Blau, Sep 6, 2024
  44. 2/9 finalize_object_file(): refactor unlink_or_warn() placementTaylor Blau, Sep 6, 2024
  45. 3/9 finalize_object_file(): implement collision checkTaylor Blau, Sep 6, 2024
  46. Junio C HamanoSep 6, 2024
  47. Chris TorekSep 6, 2024
  48. Jeff KingSep 10, 2024
  49. Junio C HamanoSep 10, 2024
  50. Patrick SteinhardtSep 16, 2024
  51. Taylor BlauSep 16, 2024
  52. Taylor BlauSep 16, 2024
  53. Junio C HamanoSep 17, 2024
  54. 4/9 pack-objects: use finalize_object_file() to rename pack/idx/etcTaylor Blau, Sep 6, 2024
  55. 5/9 i5500-git-daemon.sh: use compile-able version of Git without OpenSSLTaylor Blau, Sep 6, 2024
  56. Jeff KingSep 11, 2024
  57. Jeff KingSep 11, 2024
  58. Junio C HamanoSep 12, 2024
  59. Junio C HamanoSep 11, 2024
  60. Jeff KingSep 11, 2024
  61. 6/9 sha1: do not redefine `platform_SHA_CTX` and friendsTaylor Blau, Sep 6, 2024
  62. 7/9 hash.h: scaffolding for _fast hashing variantsTaylor Blau, Sep 6, 2024
  63. 8/9 Makefile: allow specifying a SHA-1 for non-cryptographic usesTaylor Blau, Sep 6, 2024
  64. 9/9 csum-file.c: use fast SHA-1 implementation when availableTaylor Blau, Sep 6, 2024
  65. Junio C HamanoSep 6, 2024
  66. 0/8 hash.h: support choosing a separate SHA-1 for non-cryptographic usesTaylor Blau, Sep 24, 2024
  67. 1/8 finalize_object_file(): check for name collision before renamingTaylor Blau, Sep 24, 2024
  68. Junio C HamanoSep 25, 2024
  69. 2/8 finalize_object_file(): refactor unlink_or_warn() placementTaylor Blau, Sep 24, 2024
  70. 3/8 finalize_object_file(): implement collision checkTaylor Blau, Sep 24, 2024
  71. Jeff KingSep 24, 2024
  72. Taylor BlauSep 24, 2024
  73. Jeff KingSep 24, 2024
  74. Taylor BlauSep 25, 2024
  75. Junio C HamanoSep 24, 2024
  76. Taylor BlauSep 24, 2024
  77. 4/8 pack-objects: use finalize_object_file() to rename pack/idx/etcTaylor Blau, Sep 24, 2024
  78. Junio C HamanoSep 24, 2024
  79. 5/8 sha1: do not redefine `platform_SHA_CTX` and friendsTaylor Blau, Sep 24, 2024
  80. 6/8 hash.h: scaffolding for _unsafe hashing variantsTaylor Blau, Sep 24, 2024
  81. 7/8 Makefile: allow specifying a SHA-1 for non-cryptographic usesTaylor Blau, Sep 24, 2024
  82. 8/8 csum-file.c: use unsafe SHA-1 implementation when availableTaylor Blau, Sep 24, 2024
  83. Jeff KingSep 24, 2024
  84. Elijah NewrenSep 25, 2024
  85. Junio C HamanoSep 25, 2024
  86. Taylor BlauSep 25, 2024
  87. Taylor BlauSep 25, 2024
  88. 0/8 hash.h: support choosing a separate SHA-1 for non-cryptographic usesTaylor Blau, Sep 26, 2024
  89. 1/8 finalize_object_file(): check for name collision before renamingTaylor Blau, Sep 26, 2024
  90. 2/8 finalize_object_file(): refactor unlink_or_warn() placementTaylor Blau, Sep 26, 2024
  91. 3/8 finalize_object_file(): implement collision checkTaylor Blau, Sep 26, 2024
  92. 4/8 pack-objects: use finalize_object_file() to rename pack/idx/etcTaylor Blau, Sep 26, 2024
  93. 5/8 sha1: do not redefine `platform_SHA_CTX` and friendsTaylor Blau, Sep 26, 2024
  94. 6/8 hash.h: scaffolding for _unsafe hashing variantsTaylor Blau, Sep 26, 2024
  95. 7/8 Makefile: allow specifying a SHA-1 for non-cryptographic usesTaylor Blau, Sep 26, 2024
  96. 8/8 csum-file.c: use unsafe SHA-1 implementation when availableTaylor Blau, Sep 26, 2024
  97. Elijah NewrenSep 26, 2024
  98. Junio C HamanoSep 27, 2024
  99. Jeff KingSep 27, 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.