Mar 2020

4 commits
Allow page lock to conflict among parallel group members.
21 Mar 2020Amit Kapilamaster13beta1REL_19_STABLE13beta1REL_18_STABLE13beta1REL_17_STABLE13beta1REL_16_STABLE13beta1REL_15_STABLE13beta1REL_14_STABLE13beta1REL_13_STABLE13beta1
LLM analysis, 100% surelanded as postedfrom patchset #1883 of 3 reviewers saw the final version

Matches the complete fourth patch in #188, including page-lock conflicts, deadlock exclusions, and planner/lock-manager comments; its normalized diff is identical. The attachment and commit credit Dilip Kumar and Amit Kapila, although Amit sent the series.

  • Amit Kapilasaw the final versiontrailerAmit revised and tested the conflict implementation in #162, requested page-lock tests in #178, and reviewed the final version as committer.
  • Kuntal Ghoshsaw the final versiontrailerKuntal examined deadlock handling in #163 and tested the series with the group-lock extension in #186, after which only light changes followed.
  • Sawada Masahikosaw the final versiontrailerSawada explicitly reviewed the series in #183, finding it worked as expected and suggesting assertion-only bookkeeping before subsequent light cleanup.
Allow relation extension lock to conflict among parallel group members.
20 Mar 2020Amit Kapilamaster13beta1REL_19_STABLE13beta1REL_18_STABLE13beta1REL_17_STABLE13beta1REL_16_STABLE13beta1REL_15_STABLE13beta1REL_14_STABLE13beta1REL_13_STABLE13beta1
LLM analysis, 100% surelanded as postedfrom patchset #1883 of 3 reviewers saw the final version

Matches the complete third patch in #188: extension-lock conflicts in lock.c, skipped deadlock edges, LOCK_LOCKTAG, and ProcSleep documentation; differing diff boundaries produce identical code. The patch credits Dilip Kumar and Amit Kapila.

  • Amit Kapilasaw the final versiontrailerAmit reviewed the initial implementation in #130, demonstrated correct intra-group blocking in #162, and reviewed the final version as committer.
  • Kuntal Ghoshsaw the final versiontrailerKuntal reviewed deadlock traversal in #163 and reported successful group-lock testing in #186, with only light changes afterward.
  • Sawada Masahikosaw the final versiontrailerSawada reviewed the current series in #183 and found it worked as expected, before only light changes to the final submission.
Add assert to ensure that page locks don't participate in deadlock cycle.
19 Mar 2020Amit Kapilamaster13beta1REL_19_STABLE13beta1REL_18_STABLE13beta1REL_17_STABLE13beta1REL_16_STABLE13beta1REL_15_STABLE13beta1REL_14_STABLE13beta1REL_13_STABLE13beta1
LLM analysis, 100% surelanded as postedfrom patchset #1883 of 3 reviewers saw the final version

Matches the complete second patch in #188, adding IsPageLockHeld and allowing only extension-lock acquisition while a page lock is held; applying the attachment reproduces the committed lock.c. The patch explicitly credits Dilip Kumar with changes by Amit Kapila.

  • Amit Kapilasaw the final versiontrailerAmit audited the page-lock acquisition path in #152, requested negative assertion tests in #182, and reviewed the final version as committer.
  • Kuntal Ghoshsaw the final versiontrailerKuntal reviewed the page-lock assertion comments and tested the series in #186, followed only by comment and conditional-compilation cleanup.
  • Sawada Masahikosaw the final versiontrailerSawada reviewed the page-lock flag and requested assertion-only bookkeeping in #183, which was adopted before light final cleanup.
Assert that we don't acquire a heavyweight lock on another object after relation extension lock.
18 Mar 2020Amit Kapilamaster13beta1REL_19_STABLE13beta1REL_18_STABLE13beta1REL_17_STABLE13beta1REL_16_STABLE13beta1REL_15_STABLE13beta1REL_14_STABLE13beta1REL_13_STABLE13beta1
LLM analysis, 100% surelanded as postedfrom patchset #1883 of 3 reviewers saw the final version

Matches the complete first patch in #188, including IsRelationExtensionLockHeld, CheckAndSetLockHeld, and LOCALLOCK_LOCKTAG; applying it reproduces the committed files despite different diff boundaries. Its author credit names Dilip Kumar with changes by Amit Kapila.

  • Amit Kapilasaw the final versiontrailerAmit reviewed recursive acquisition in #130, proposed centralized flag maintenance in #167, and reviewed the final version as committer.
  • Kuntal Ghoshsaw the final versiontrailerKuntal reviewed the extension-assertion comments and reported successful testing in #186, with only light changes afterward.
  • Sawada Masahikosaw the final versiontrailerSawada examined the earlier invariant in #144 and reviewed the later flag implementation in #183, requesting the adopted assertion-only bookkeeping.

Credits across 4 commits

Author
Dilip Kumar4
Amit Kapila2
with few changes by Amit Kapila2
Reviewer
Amit Kapila4
Kuntal Ghosh4
Masahiko Sawada4
Committer
Amit Kapila4