Horton

Re: [HACKERS] Weaker shmem interlock w/o postmaster.pid

Thomas Munro <thomas.munro@gmail.com>

View thread View on postgr.es
From: Thomas Munro <thomas.munro@gmail.com>
To: Noah Misch <noah@leadboat.com>
Cc: Daniel Gustafsson <daniel@yesql.se>, Kyotaro HORIGUCHI <horiguchi.kyotaro@lab.ntt.co.jp>, "9erthalion6@gmail.com" <9erthalion6@gmail.com>, "sfrost@snowman.net" <sfrost@snowman.net>, "pgsql-hackers@postgresql.org" <pgsql-hackers@postgresql.org>
Date: 2019-04-11T00:48:35Z
Lists: pgsql-hackers 
On Mon, Apr 8, 2019 at 6:42 PM Noah Misch <noah@leadboat.com> wrote:
> - AIX animals failed two ways.  First, I missed a "use" statement such that
>   poll_start() would fail if it needed more than one attempt.  Second, I
>   assumed $pid would be gone as soon as kill(9, $pid) returned[1].

> [1] POSIX says "sig or at least one pending unblocked signal shall be
> delivered to the sending thread before kill() returns."  I doubt the
> postmaster had another signal pending often enough to explain the failures, so
> AIX probably doesn't follow POSIX in this respect.

It looks like you fixed this, but I was curious about this obversation
as someone interested in learning more about kernel stuff and
portability... Maybe I misunderstood, but isn't POSIX referring to
kill(sig, $YOUR_OWN_PID) there?  That is, if you signal *yourself*,
and no other thread exists that could handle the signal, it will be
handled by the sending thread, and in the case of SIGKILL it will
therefore never return.  But here, you were talking about a perl
script that kills the postmaster, no?  If so, that passage doesn't
seem to apply.  In any case, regardless of whether the signal handler
has run to completion when kill() returns, doesn't the pid have to
continue to exist in the process table until it is reaped by its
parent (possibly in response to SIGCHLD), with one of the wait*()
family of system calls?

-- 
Thomas Munro
https://enterprisedb.com

Commits

  1. Probe only 127.0.0.1 when looking for ports on Unix.

  2. Don't write to stdin of a test process that could have already exited.

  3. Test both 0.0.0.0 and 127.0.0.x addresses to find a usable port.

  4. MSYS: Translate REGRESS_SHLIB to a Windows file name.

  5. MSYS: Skip src/test/recovery/t/017_shm.pl.

  6. When Perl "kill(9, ...)" fails, try "pg_ctl kill".

  7. Consistently test for in-use shared memory.

  8. Revert "Consistently test for in-use shared memory."

  9. Silence -Wimplicit-fallthrough in sysv_shmem.c.

  10. Make src/test/recovery/t/017_shm.pl safe for concurrent execution.

  11. Update HINT for pre-existing shared memory block.

  12. Add WL_EXIT_ON_PM_DEATH pseudo-event.

  13. The default values for shared_buffers and max_connections are now 1000

  14. XLOG (and related) changes:

  15. Significant cleanups in SysV IPC handling (shared mem and semaphores).