Don't run atexit callbacks in quickdie signal handlers.
exit() is not async-signal safe. Even if the libc implementation is, 3rd
party libraries might have installed unsafe atexit() callbacks. After
receiving SIGQUIT, we really just want to exit as quickly as possible, so
we don't really want to run the atexit() callbacks anyway.
The original report by Jimmy Yih was a self-deadlock in startup_die().
However, this patch doesn't address that scenario; the signal handling
while waiting for the startup packet is more complicated. But at least this
alleviates similar problems in the SIGQUIT handlers, like that reported
by Asim R P later in the same thread.
Backpatch to 9.3 (all supported versions).
Discussion:
https://www.postgresql.org/message-id/CAOMx_OAuRUHiAuCg2YgicZLzPVv5d9_H4KrL_OFsFP%3DVPekigA%40mail.gmail.com
Branch
------
REL_10_STABLE
Details
-------
https://git.postgresql.org/pg/commitdiff/2332020d6d1912bff4c03a8c2b37ad1c8b184281
Modified Files
--------------
src/backend/postmaster/bgworker.c | 25 +++++++++----------------
src/backend/postmaster/bgwriter.c | 24 +++++++++---------------
src/backend/postmaster/checkpointer.c | 24 +++++++++---------------
src/backend/postmaster/startup.c | 24 +++++++++---------------
src/backend/postmaster/walwriter.c | 24 +++++++++---------------
src/backend/replication/walreceiver.c | 32 +++++++++++++-------------------
src/backend/tcop/postgres.c | 32 +++++++++++++++++++-------------
7 files changed, 77 insertions(+), 108 deletions(-)