From: Tom Lane <tgl@sss.pgh.pa.us>
To: Alexander Korotkov <aekorotkov@gmail.com>
cc: Thomas Munro <thomas.munro@gmail.com>,
        Heikki Linnakangas <hlinnaka@iki.fi>,
        Alexander Korotkov <akorotkov@postgresql.org>,
        PostgreSQL Hackers <pgsql-hackers@lists.postgresql.org>
Subject: Re: Detecting pointer misalignment (was Re: pgsql: Implementation of
 subscripting for jsonb)
In-reply-to: 
 <CAPpHfduDEYtWEB+xhM3PcP4bFwJ3aqZw7R7xFto8+pVDV5w-MA@mail.gmail.com>
References: <E1l6Jjn-0000kp-75@gemulon.postgresql.org>
 <37bb56a0-391d-759d-771b-9c1d8c15a84b@iki.fi>
 <3293925.1612163110@sss.pgh.pa.us>
 <CAPpHfdsne3=T=fMNU45PtxdhSL_J2PjLTeS8rwKnJzUR4YNd4w@mail.gmail.com>
 <CAPpHfdu8Jh0tySa50Zhkyg6gMvi2pQFdHAMiF+UkZNqsDLwTQw@mail.gmail.com>
 <475514.1612745257@sss.pgh.pa.us>
 <CAPpHfdvh+QrbBCo=uY=_fHK8w7yKr=Q4KGAtp=3tzp62SjODrA@mail.gmail.com>
 <CA+hUKGLs9xMUDY=Oh0uOyJysaY=NGg3y53q_Tr2C+LLqstYf6A@mail.gmail.com>
 <CAPpHfduDEYtWEB+xhM3PcP4bFwJ3aqZw7R7xFto8+pVDV5w-MA@mail.gmail.com>
Comments: In-reply-to Alexander Korotkov <aekorotkov@gmail.com>
	message dated "Fri, 12 Feb 2021 17:29:48 +0300"
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-ID: <1557040.1613150383.1@sss.pgh.pa.us>
Content-Transfer-Encoding: quoted-printable
Date: Fri, 12 Feb 2021 12:19:43 -0500
Message-ID: <1557041.1613150383@sss.pgh.pa.us>
Precedence: bulk

I've updated buildfarm member longfin to use "-fsanitize=3Dalignment
-fsanitize-trap=3Dalignment", and it just got through a run successfully
with that.  It'd be good perhaps if some other buildfarm owners
followed suit (mumble JIT coverage mumble).

Looking around at other recent reports, it looks like we'll need to tweak
the compiler version cutoffs a bit.  I see for instance that spurfowl,
with gcc (Ubuntu 5.4.0-6ubuntu1~16.04.11) 5.4.0 20160609, is whining:

pg_crc32c_sse42.c:24:1: warning: \342\200\230no_sanitize\342\200\231 attri=
bute directive ignored [-Wattributes]

So maybe it'd better be __GNUC__ >=3D 6 not __GNUC__ >=3D 5.  I think
we can wait a little bit for more reports before messing with that,
though.

Once this does settle, should we consider back-patching so that it's
possible to run alignment checks in the back branches too?

			regards, tom lane