Jul 2020

2 commits
Use RAND_poll() for seeding randomness after fork().
25 Jul 2020Noah Mischmaster14beta1REL_19_STABLE14beta1REL_18_STABLE14beta1REL_17_STABLE14beta1REL_16_STABLE14beta1REL_15_STABLE14beta1REL_14_STABLE14beta1
LLM analysis, 99% surepart of a posted patchsetlight changesfrom patchset #13 of 3 reviewers saw the final version

This commit takes the first patch of the two-patch series #1, with an identical diff in src/backend/postmaster/fork_process.c replacing RAND_cleanup() with RAND_poll() and documenting OpenSSL version behavior. The separate retry cleanup landed in 15e4419722, so together the commits cover the whole submission.

  • David Steelesaw the final versionLLM analysis, 100% sureThe commit explicitly credits David as a reviewer ce4939ff70, and his response assessed the proposal against Daniel's explanation and the OpenSSL wiki #2.
  • Michael Paquiersaw the final versionLLM analysis, 100% sureMichael evaluated OpenSSL version-specific seeding behavior and supported RAND_poll() after fork on HEAD only #6, and the commit explicitly credits his review ce4939ff70.
  • Noah Mischsaw the final versioncommittercommitted the patch
Remove optimization for RAND_poll() failing.
25 Jul 2020Noah Mischmaster14beta1REL_19_STABLE14beta1REL_18_STABLE14beta1REL_17_STABLE14beta1REL_16_STABLE14beta1REL_15_STABLE14beta1REL_14_STABLE14beta1
LLM analysis, 99% surepart of a posted patchsetlight changesfrom patchset #13 of 3 reviewers saw the final version

This commit takes the second patch of the two-patch series #1: src/port/pg_strong_random.c stops breaking immediately on RAND_poll() failure. Its executable changes match exactly; committer edits only reflow and polish the explanatory comment, while the first patch landed separately in ce4939ff70.

  • David Steelesaw the final versionLLM analysis, 100% sureDavid examined callers and questioned whether immediate retries could delay expected failure #2 #4, and the commit explicitly credits his review 15e4419722.
  • Michael Paquiersaw the final versionLLM analysis, 100% sureMichael checked the retry loop's history and supported retaining retries while relying on RAND_status() for validation #6, and the commit explicitly credits his review 15e4419722.
  • Noah Mischsaw the final versioncommittercommitted the patch

Credits across 2 commits

Author
Daniel GustafssonLLM2
Reviewer
Noah Mischcommitter2
Committer
Noah Misch2

N LLM means N of the credits come from LLM analysis, not from commit trailers.