Re: Crash on backend exit w/ OpenLDAP [2.4.24, 2.4.31] - Mailing list pgsql-hackers

From Tom Lane
Subject Re: Crash on backend exit w/ OpenLDAP [2.4.24, 2.4.31]
Date
Msg-id 3496.1403197294@sss.pgh.pa.us
Whole thread Raw
In response to Re: Crash on backend exit w/ OpenLDAP [2.4.24, 2.4.31]  (Noah Misch <noah@leadboat.com>)
Responses Re: Crash on backend exit w/ OpenLDAP [2.4.24, 2.4.31]  (Noah Misch <noah@leadboat.com>)
List pgsql-hackers
Noah Misch <noah@leadboat.com> writes:
> On Thu, Jun 12, 2014 at 05:02:19PM -0400, Noah Misch wrote:
>> You can cause the at-exit crash by building PostgreSQL against OpenLDAP
>> 2.4.31, connecting with LDAP authentication, and issuing "LOAD 'dblink'".

>> 4. Detect older OpenLDAP versions at runtime, just before we would otherwise
>> initialize OpenLDAP, and raise an error.  Possibly make the same check at
>> compile time, for packager convenience.

> Having pondered this some more, I lean toward the following conservative fix.
> Add to all supported branches a test case that triggers the crash and a
> configure-time warning if the OpenLDAP version falls in the vulnerable range.
> So long as those who build from source monitor either "configure" output or
> test suite failures, they'll have the opportunity to head off the problem.

+1 for a configure warning, but I share your concern that it's likely to
go unnoticed (sometimes I wish autoconf were not so chatty...).

Keep in mind that some distros patch bugs without changing the reported
version number, so I'm afraid we couldn't adopt the easy solution of
making configure give a hard error when the version is suspicious; and
for the same reason your #4 above is unworkable.

I'm not sure about the practicality of adding a test case --- how will we
test that if no LDAP server is at hand?

I concur with not working much harder than this, in any case.  It's really
OpenLDAP's bug to fix.
        regards, tom lane



pgsql-hackers by date:

Previous
From: Noah Misch
Date:
Subject: Re: Crash on backend exit w/ OpenLDAP [2.4.24, 2.4.31]
Next
From: Tom Lane
Date:
Subject: Re: change alter user to be a true alias for alter role