Date: Mon, 19 Nov 2012 12:58:04 +0100
From: Andres Freund <andres@2ndquadrant.com>
To: Alvaro Herrera <alvherre@2ndquadrant.com>
Cc: pgsql-hackers@postgresql.org
Subject: Re: foreign key locks
Message-ID: <20121119115804.GA28067@awork2.anarazel.de>
References: <4FEAB90A0200002500048B7D@gw.wicourts.gov>
	<1346389151-sup-8379@alvh.no-ip.org>
	<20121018195820.GO3763@alvh.no-ip.org>
	<201210270006.54728.andres@2ndquadrant.com>
	<20121114152347.GA5161@alvh.no-ip.org>
	<20121114162726.GE5161@alvh.no-ip.org>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20121114162726.GE5161@alvh.no-ip.org>


On 2012-11-14 13:27:26 -0300, Alvaro Herrera wrote:
> Alvaro Herrera wrote:
>
> > > * In heap_lock_tuple's  XMAX_IS_MULTI case
> > >
> > > [snip]
> > >
> > > why is it membermode > mode and not membermode >= mode?
> >
> > Uh, that's a bug.  Fixed.  As noticed in the comment above that snippet,
> > there was a deadlock possible here.  Maybe I should add a test to ensure
> > this doesn't happen.
>
> Done:
> https://github.com/alvherre/postgres/commit/df2847e38198e99f57e52490e1e9391ebb70d770

Some more review bits, based on ffd6250d1d393f2ecb9bfc55c2c6f715dcece557

- if oldestMultiXactId + db is set and then that database is dropped we seem to
  have a problem because MultiXactAdvanceOldest won't overwrite those
  values. Should probably use SetMultiXactIdLimit directly.

- what stop multixacts only being filled out (i.e RecordNewMultiXact()) *after*
  the XLogInsert() *and* after a MultiXactGetCheckptMulti()? Afaics
  MultiXactGenLock is not hold in CreateMultiXactId(). If we crash in that
  moment we loose the multixact data which now means potential data loss...

- multixact member group data crossing 512 sector boundaries makes me uneasy
  (as its 5 bytes). I don't really see a scenario where its dangerous, but
  ... Does anybody see a problem here?

- there are quite some places that do
	multiStopLimit = multiWrapLimit - 100;
	if (multiStopLimit < FirstMultiXactId)
		multiStopLimit -= FirstMultiXactId;

  perhaps MultiXactIdAdvance and MultiXactIdRetreat macros are in order?

- I find using a default: clause in switches with enum types where everything
  is expected to be handled like the following a bad idea, this way the
  compiler won't warn you if youve missed case's which makes changing/extending code harder:
		switch (rc->strength)
		{
			case LCS_FORNOKEYUPDATE:
				newrc->markType = ROW_MARK_EXCLUSIVE;
				break;
			case LCS_FORSHARE:
				newrc->markType = ROW_MARK_SHARE;
				break;
			case LCS_FORKEYSHARE:
				newrc->markType = ROW_MARK_KEYSHARE;
				break;
			case LCS_FORUPDATE:
				newrc->markType = ROW_MARK_KEYEXCLUSIVE;
				break;
			default:
				elog(ERROR, "unsupported rowmark type %d", rc->strength);
		}
-
#if 0
			/*
			 * The idea here is to remove the IS_MULTI bit, and replace the
			 * xmax with the updater's Xid.  However, we can't really do it:
			 * modifying the Xmax is not allowed under our buffer locking
			 * rules, unless we have an exclusive lock; but we don't know that
			 * we have it.  So the multi needs to remain in place :-(
			 */
			ResetMultiHintBit(tuple, buffer, xmax, true);
#endif

 Three things:
    - HeapTupleSatisfiesUpdate is actually always called exclusively locked ;)
    - Extending something like LWLockHeldByMe to also return the current
      lockmode doesn't sound hard
    - we seem to be using #ifdef NOT_YET for such cases

- Using a separate production for the lockmode seems to be nicer imo, not
  really important though
for_locking_item:
			FOR UPDATE locked_rels_list opt_nowait
...
			| FOR NO KEY UPDATE locked_rels_list opt_nowait
...
			| FOR SHARE locked_rels_list opt_nowait
...
			| FOR KEY SHARE locked_rels_list opt_nowait
		;

- not really padding, MultiXactStatus is 4bytes...
	/*
	 * XXX Note: there's a lot of padding space in MultiXactMember.  We could
	 * find a more compact representation of this Xlog record -- perhaps all the
	 * status flags in one XLogRecData, then all the xids in another one?  Not
	 * clear that it's worth the trouble though.
	 */
- why
#define SizeOfMultiXactCreate (offsetof(xl_multixact_create, nmembers) + sizeof(int32))
and not
#define SizeOfMultiXactCreate offsetof(xl_multixact_create, members)
- starting a critical section in GetNewMultiXactId but not ending it is ugly,
  but not new

Greetings,

Andres

-- 
 Andres Freund	                   http://www.2ndQuadrant.com/
 PostgreSQL Development, 24x7 Support, Training & Services