Horton

Re: BUG #15121: Multiple UBSAN errors

Tom Lane <tgl@sss.pgh.pa.us>

View thread View on postgr.es
From: Tom Lane <tgl@sss.pgh.pa.us>
To: Noah Misch <noah@leadboat.com>
Cc: Tomas Vondra <tomas.vondra@2ndquadrant.com>, marxin.liska@gmail.com, pgsql-bugs@lists.postgresql.org
Date: 2019-06-30T18:52:44Z
Lists: pgsql-bugs 
Noah Misch <noah@leadboat.com> writes:
> On Sun, Mar 18, 2018 at 09:02:59PM -0400, Tom Lane wrote:
>> What I think might be happening is that the compiler is taking the
>> fact that the pointer is declared as AnyArrayType *, where ...
>> to assume that the pointer must be aligned on an 8-byte boundary because
>> ExpandedArrayHeader would require that, even if we're only accessing the
>> "flt" member.

> Yep.

> Using different casts, as attached, silences these complaints without much
> change in code cosmetics.

I'd intentionally avoided using casts this way as an attempt to ensure
that these macros weren't accidentally applied to the wrong kind of
pointer.  However, avoiding UBSAN complaints (and perhaps wrong code
generation) seems like a more important goal.  This patch looks OK to me.

			regards, tom lane

Commits

  1. Don't read fields of a misaligned ExpandedObjectHeader or AnyArrayType.

  2. Doc: note that statement-level view triggers require an INSTEAD OF trigger.