[exim] Re: failure to transfer data from subprocess

Startseite
Nachricht löschen
Nachricht beantworten
Autor: Jeremy Harris
Datum:  
To: exim-users
Betreff: [exim] Re: failure to transfer data from subprocess
On 13/05/2023 19:51, Robert Nicholson via Exim-users wrote:
> 02:19:39 13517 writing filter log as euid 1043


That tells us it came after handling a "logwrite" in a filter file...

> 02:19:39 13517   ╭considering: $header_from:
> 02:19:39 13517   ├──expanding: $header_from:
> 02:19:39 13517   ╰─────result: Firstname Lastname <user@???  <mailto:bbarlow@matlensilver.com>>
> 02:19:39 13517              ╰──(tainted)


This (in fact, expanding anything the results in a tainted value)
isn't necessarily a problem. Using that tainted value in
certain other contexts (basically, expanding *it* or the moral
equivalent, such as using it for a filename) would be a problem.
But even if you tried to do that it should *not* result in
a null-pointer-follow. It's a bug in Exim, even if you've managed
to trigger it with something in your config.

The first debug snippet you showed doesn't have that expansion,
so I'm slightly confused as to the time-sequence.
It has

> 02:19:39 13517 writing filter log as euid 1043

(as above)
> 02:19:39 13517 Filter: pipe message to: nice -10 $home/perlscripts/filter.pl


That was dealing with a "pipe" command in a filter file, after
the "logwrite" above.

Again, we don't really have a location for the null-pointer-follow.
A coredump would be the best way of debugging. If you want to
avoid leaking info, then the stacktrace from a coredump ("bt" in gdb)
would be useful (though, admittedly, a binary with debug-info would be
best - "-ggdb" in CFLAGS).

--
Cheers,
Jeremy


--
## subscription configuration (requires account):
## https://lists.exim.org/mailman3/postorius/lists/exim-users.lists.exim.org/
## unsubscribe (doesn't require an account):
## exim-users-unsubscribe@???
## Exim details at http://www.exim.org/
## Please use the Wiki with this list - http://wiki.exim.org/