Mailing List Archive: SIGSEGV in panic log

SIGSEGV in panic log

exim-users at lists

Feb 22, 2024, 7:11 AM

Post #1 of 10 (153 views)

I created a new VM for exim . I've loaded many VMs .. this chassis has
several active VMs actually. vmWare.
I have never seen the following .. it seems like a chassis memory error
.. however the other two active VMs on this chassis run without issue.
So I wonder ..
is there anything on exim that could -possibly- be triggering this ? SSL
? (this is on FreeBSD 14.0)

these errors show up in the exim panic log. stopping the exim instance
and restarting it, gets it up and running, but eventually it happens again.

2023-12-25 10:34:15 SIGSEGV (fault address: 0xffffffffffffff7c)
2023-12-25 10:34:15 SIGSEGV (maybe attempt to write to immutable memory)
2023-12-25 10:34:15 SIGSEGV (44383 handling incoming connection from
emailh.ca [23.235.65.100]
)
2023-12-25 10:34:15 SIGSEGV (fault address: 0xffffffffffffff7c)
2023-12-25 10:34:15 SIGSEGV (maybe attempt to write to immutable memory)
2023-12-25 10:34:15 SIGSEGV (44384 handling incoming connection from
emailh.ca [23.235.65.100]
)
2023-12-25 10:34:15 SIGSEGV (fault address: 0xffffffffffffff7c)
2023-12-25 10:34:15 SIGSEGV (maybe attempt to write to immutable memory)
2023-12-25 10:34:15 SIGSEGV (44385 handling incoming connection from
emailh.ca [23.235.65.100]
)
2023-12-25 10:34:15 SIGSEGV (fault address: 0xffffffffffffff7c)
2023-12-25 10:34:15 SIGSEGV (maybe attempt to write to immutable memory)
2023-12-25 10:34:15 SIGSEGV (44386 handling incoming connection from
emailh.ca [23.235.65.100]
)

I deleted the vm, reloaded it, and eventually the same issue resurfaced:

2024-02-21 05:09:07 SIGSEGV (fault address: 0xffffffffffffff74)
2024-02-21 05:09:07 SIGSEGV (maybe attempt to write to immutable memory)
2024-02-21 05:09:07 SIGSEGV ( 3685 handling incoming connection from
emailh.ca [23.235.65.100]
)
2024-02-21 05:09:07 SIGSEGV (fault address: 0xffffffffffffff74)
2024-02-21 05:09:07 SIGSEGV (maybe attempt to write to immutable memory)
2024-02-21 05:09:07 SIGSEGV ( 3686 handling incoming connection from
emailh.ca [23.235.65.100]
)
2024-02-21 05:09:07 SIGSEGV (fault address: 0xffffffffffffff74)
2024-02-21 05:09:07 SIGSEGV (maybe attempt to write to immutable memory)
2024-02-21 05:09:07 SIGSEGV ( 3687 handling incoming connection from
emailh.ca [23.235.65.100]
)
2024-02-21 05:09:07 SIGSEGV (fault address: 0xffffffffffffff74)
2024-02-21 05:09:07 SIGSEGV (maybe attempt to write to immutable memory)
2024-02-21 05:09:07 SIGSEGV ( 3688 handling incoming connection from
emailh.ca [23.235.65.100]
)

--
Jim Pazarena exim@paz.bz
Haida Gwaii - British Columbia - Canada

--
## 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@lists.exim.org
## Exim details at http://www.exim.org/
## Please use the Wiki with this list - http://wiki.exim.org/

Re: SIGSEGV in panic log [ In reply to ]

exim-users at lists

Feb 22, 2024, 9:47 AM

Post #2 of 10 (153 views)

On Thu, 22 Feb 2024, Jim Pazarena via Exim-users wrote:

> I created a new VM for exim . I've loaded many VMs .. this chassis has
> several active VMs actually. vmWare.
> I have never seen the following .. it seems like a chassis memory error ..
> however the other two active VMs on this chassis run without issue. So I
> wonder ..
> is there anything on exim that could -possibly- be triggering this ? SSL ?
> (this is on FreeBSD 14.0)
>
> these errors show up in the exim panic log. stopping the exim instance and
> restarting it, gets it up and running, but eventually it happens again.

Which version of Exim ?
What is in the error log at that time ?

Recent crash bugs include:
https://bugs.exim.org/show_bug.cgi?id=3059
SIGSEGV on SMTP delivery if remote badly responds
while PIPELINING and CHUNKED is active

https://bugs.exim.org/show_bug.cgi?id=3047
doesn't crash on known systems, but attempts to use a bad address
(though nearer -2G than -100).
What version of PCRE2 are you using?

Are you up to building a patched version of Exim -
both of these have patches which may help *if* you have that problem ?

> 2023-12-25 10:34:15 SIGSEGV (fault address: 0xffffffffffffff7c)
> 2023-12-25 10:34:15 SIGSEGV (maybe attempt to write to immutable memory)
> 2023-12-25 10:34:15 SIGSEGV (44383 handling incoming connection from
> emailh.ca [23.235.65.100]
> )
> 2023-12-25 10:34:15 SIGSEGV (fault address: 0xffffffffffffff7c)
> 2023-12-25 10:34:15 SIGSEGV (maybe attempt to write to immutable memory)
> 2023-12-25 10:34:15 SIGSEGV (44384 handling incoming connection from
> emailh.ca [23.235.65.100]
> )
> 2023-12-25 10:34:15 SIGSEGV (fault address: 0xffffffffffffff7c)
> 2023-12-25 10:34:15 SIGSEGV (maybe attempt to write to immutable memory)
> 2023-12-25 10:34:15 SIGSEGV (44385 handling incoming connection from
> emailh.ca [23.235.65.100]
> )
> 2023-12-25 10:34:15 SIGSEGV (fault address: 0xffffffffffffff7c)
> 2023-12-25 10:34:15 SIGSEGV (maybe attempt to write to immutable memory)
> 2023-12-25 10:34:15 SIGSEGV (44386 handling incoming connection from
> emailh.ca [23.235.65.100]
> )
>
> I deleted the vm, reloaded it, and eventually the same issue resurfaced:
>
> 2024-02-21 05:09:07 SIGSEGV (fault address: 0xffffffffffffff74)
> 2024-02-21 05:09:07 SIGSEGV (maybe attempt to write to immutable memory)
> 2024-02-21 05:09:07 SIGSEGV ( 3685 handling incoming connection from
> emailh.ca [23.235.65.100]
> )
> 2024-02-21 05:09:07 SIGSEGV (fault address: 0xffffffffffffff74)
> 2024-02-21 05:09:07 SIGSEGV (maybe attempt to write to immutable memory)
> 2024-02-21 05:09:07 SIGSEGV ( 3686 handling incoming connection from
> emailh.ca [23.235.65.100]
> )
> 2024-02-21 05:09:07 SIGSEGV (fault address: 0xffffffffffffff74)
> 2024-02-21 05:09:07 SIGSEGV (maybe attempt to write to immutable memory)
> 2024-02-21 05:09:07 SIGSEGV ( 3687 handling incoming connection from
> emailh.ca [23.235.65.100]
> )
> 2024-02-21 05:09:07 SIGSEGV (fault address: 0xffffffffffffff74)
> 2024-02-21 05:09:07 SIGSEGV (maybe attempt to write to immutable memory)
> 2024-02-21 05:09:07 SIGSEGV ( 3688 handling incoming connection from
> emailh.ca [23.235.65.100]
> )
>
>
> --
> Jim Pazarena exim@paz.bz
> Haida Gwaii - British Columbia - Canada

--
Andrew C. Aitchison Kendal, UK
andrew@aitchison.me.uk

--
## 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@lists.exim.org
## Exim details at http://www.exim.org/
## Please use the Wiki with this list - http://wiki.exim.org/

Re: SIGSEGV in panic log [ In reply to ]

exim-users at lists

Feb 22, 2024, 12:26 PM

Post #3 of 10 (153 views)

On 22/02/2024 15:11, Jim Pazarena via Exim-users wrote:
> is there anything on exim that could -possibly- be triggering this ? SSL ? (this is on FreeBSD 14.0)

Since you didn't say what version and patch level, that's a bit hard to answer.
--
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@lists.exim.org
## Exim details at http://www.exim.org/
## Please use the Wiki with this list - http://wiki.exim.org/

Re: SIGSEGV in panic log [ In reply to ]

exim-users at lists

Feb 22, 2024, 12:54 PM

Post #4 of 10 (153 views)

apologies for not including v.

this server, just loaded back in December and then again a week ago. The
same issue each time.
v 4.97

an identical "chassis" which also has same vm size, identical
assignments on the vm (memory / drive spc), running v 4.94.2, no problem
ever.

I really assumed that the response would have been "run a memtest on
your chassis".

Yes I could run a custom install for testing. This is a hot backup
chassis & vm. It really only exchanges email with the main exim server.
It is when I get notices of undeliverables from the main server that the
issue comes to my attention.

Thank you !

Jim

On 2024-02-22 12:26 PM, Jeremy Harris via Exim-users wrote:
> On 22/02/2024 15:11, Jim Pazarena via Exim-users wrote:
>> is there anything on exim that could -possibly- be triggering this ?
>> SSL ? (this is on FreeBSD 14.0)
>
> Since you didn't say what version and patch level, that's a bit hard to
> answer.

--
Jim Pazarena exim@paz.bz
Haida Gwaii - British Columbia - Canada

--
## 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@lists.exim.org
## Exim details at http://www.exim.org/
## Please use the Wiki with this list - http://wiki.exim.org/

Re: SIGSEGV in panic log [ In reply to ]

exim-users at lists

Feb 22, 2024, 3:11 PM

Post #5 of 10 (153 views)

On Thu, 22 Feb 2024, Jim Pazarena via Exim-users wrote:

> I really assumed that the response would have been
> "run a memtest on your chassis".

>> On 22/02/2024 15:11, Jim Pazarena via Exim-users
>> wrote:
>>> is there anything on exim that could -possibly-
>>> be triggering this ?

That question sort of rules out that answer.
Since this is a hot backup, scheduling memtest
will be straight forward, so I *would* start there.

--
Andrew C. Aitchison Kendal, UK
andrew@aitchison.me.uk

--
## 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@lists.exim.org
## Exim details at http://www.exim.org/
## Please use the Wiki with this list - http://wiki.exim.org/

Re: SIGSEGV in panic log [ In reply to ]

exim-users at lists

Feb 22, 2024, 3:38 PM

Post #6 of 10 (153 views)

On 22/02/2024 23:11, Andrew C Aitchison via Exim-users wrote:
> On Thu, 22 Feb 2024, Jim Pazarena via Exim-users wrote:
>
>> I really assumed that the response would have been "run a memtest on your chassis".
>
>>> On 22/02/2024 15:11, Jim Pazarena via Exim-users wrote:
>>>> is there anything on exim that could -possibly- be triggering this ?
>
> That question sort of rules out that answer.
> Since this is a hot backup, scheduling memtest
> will be straight forward, so I *would* start there.

I'm a pessimist, and would suggest you build using a git-head version
if you can. But the two thing are not exclusive.

--
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@lists.exim.org
## Exim details at http://www.exim.org/
## Please use the Wiki with this list - http://wiki.exim.org/

Re: SIGSEGV in panic log [ In reply to ]

exim-users at lists

Apr 6, 2024, 8:32 PM

Post #7 of 10 (90 views)

I am now experiencing the same error on multiple vms across multiple
chassis.

4.97.1
on FreeBSD 14.0p5

It is unclear to me how to find a newer version such as a testing version?
Is there a 4.97.2 forthcoming any time soon ? even a beta ?

On 2024-02-22 9:47 AM, Andrew C Aitchison via Exim-users wrote:
> On Thu, 22 Feb 2024, Jim Pazarena via Exim-users wrote:
>
>> I created a new VM for exim . I've loaded many VMs .. this chassis has
>> several active VMs actually. vmWare.
>> I have never seen the following .. it seems like a chassis memory
>> error .. however the other two active VMs on this chassis run without
>> issue. So I wonder ..
>> is there anything on exim that could -possibly- be triggering this ?
>> SSL ? (this is on FreeBSD 14.0)
>>
>> these errors show up in the exim panic log. stopping the exim instance
>> and restarting it, gets it up and running, but eventually it happens
>> again.
>
> Which version of Exim ?
> What is in the error log at that time ?
>
> Recent crash bugs include:
> https://bugs.exim.org/show_bug.cgi?id=3059
> SIGSEGV on SMTP delivery if remote badly responds
> while PIPELINING and CHUNKED is active
>
>
> https://bugs.exim.org/show_bug.cgi?id=3047
> doesn't crash on known systems, but attempts to use a bad address
> (though nearer -2G than -100).
> What version of PCRE2 are you using?
>
> Are you up to building a patched version of Exim -
> both of these have patches which may help *if* you have that problem ?
>
>
>> 2023-12-25 10:34:15 SIGSEGV (fault address: 0xffffffffffffff7c)
>> 2023-12-25 10:34:15 SIGSEGV (maybe attempt to write to immutable memory)
>> 2023-12-25 10:34:15 SIGSEGV (44383 handling incoming connection from
>> emailh.ca [23.235.65.100]
>> )
>> 2023-12-25 10:34:15 SIGSEGV (fault address: 0xffffffffffffff7c)
>> 2023-12-25 10:34:15 SIGSEGV (maybe attempt to write to immutable memory)
>> 2023-12-25 10:34:15 SIGSEGV (44384 handling incoming connection from
>> emailh.ca [23.235.65.100]
>> )
>> 2023-12-25 10:34:15 SIGSEGV (fault address: 0xffffffffffffff7c)
>> 2023-12-25 10:34:15 SIGSEGV (maybe attempt to write to immutable memory)
>> 2023-12-25 10:34:15 SIGSEGV (44385 handling incoming connection from
>> emailh.ca [23.235.65.100]
>> )
>> 2023-12-25 10:34:15 SIGSEGV (fault address: 0xffffffffffffff7c)
>> 2023-12-25 10:34:15 SIGSEGV (maybe attempt to write to immutable memory)
>> 2023-12-25 10:34:15 SIGSEGV (44386 handling incoming connection from
>> emailh.ca [23.235.65.100]
>> )
>>
>> I deleted the vm, reloaded it, and eventually the same issue resurfaced:
>>
>> 2024-02-21 05:09:07 SIGSEGV (fault address: 0xffffffffffffff74)
>> 2024-02-21 05:09:07 SIGSEGV (maybe attempt to write to immutable memory)
>> 2024-02-21 05:09:07 SIGSEGV ( 3685 handling incoming connection from
>> emailh.ca [23.235.65.100]
>> )
>> 2024-02-21 05:09:07 SIGSEGV (fault address: 0xffffffffffffff74)
>> 2024-02-21 05:09:07 SIGSEGV (maybe attempt to write to immutable memory)
>> 2024-02-21 05:09:07 SIGSEGV ( 3686 handling incoming connection from
>> emailh.ca [23.235.65.100]
>> )
>> 2024-02-21 05:09:07 SIGSEGV (fault address: 0xffffffffffffff74)
>> 2024-02-21 05:09:07 SIGSEGV (maybe attempt to write to immutable memory)
>> 2024-02-21 05:09:07 SIGSEGV ( 3687 handling incoming connection from
>> emailh.ca [23.235.65.100]
>> )
>> 2024-02-21 05:09:07 SIGSEGV (fault address: 0xffffffffffffff74)
>> 2024-02-21 05:09:07 SIGSEGV (maybe attempt to write to immutable memory)
>> 2024-02-21 05:09:07 SIGSEGV ( 3688 handling incoming connection from
>> emailh.ca [23.235.65.100]
>> )
>>
>>
>> --
>> Jim Pazarena exim@paz.bz
>> Haida Gwaii - British Columbia - Canada
>

--
Jim Pazarena exim@paz.bz
Haida Gwaii - British Columbia - Canada

--
## 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@lists.exim.org
## Exim details at http://www.exim.org/
## Please use the Wiki with this list - http://wiki.exim.org/

Re: SIGSEGV in panic log [ In reply to ]

exim-users at lists

Apr 7, 2024, 3:14 AM

Post #8 of 10 (90 views)

On 07/04/2024 04:32, Jim Pazarena via Exim-users wrote:
> 4.97.1
> on FreeBSD 14.0p5
>
> It is unclear to me how to find a newer version such as a testing version?
> Is there a 4.97.2 forthcoming any time soon ? even a beta ?

I can't speak to FreeBSD release schedules, but
- can you set up for a core dump, and get a stack trace?
- are you able to compile from source?
--
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@lists.exim.org
## Exim details at http://www.exim.org/
## Please use the Wiki with this list - http://wiki.exim.org/

Re: SIGSEGV in panic log [ In reply to ]

exim-users at lists

Apr 7, 2024, 6:18 AM

Post #9 of 10 (90 views)

> - can you set up for a core dump,

Under FreeBSD:
sysctl kern.sugid_coredump=1

> and get a stack trace?

One of ways:

pkg install gdb

Add to /etc/make.conf :

.if ${.CURDIR} == ${PORTSDIR}/mail/exim
CFLAGS+= -g -O0 -ggdb
.endif

> - are you able to compile from source?

One of ways to compile under FreeBSD:

cd /usr/ports/mail/exim
make clean patch
cd work/exim-4.97.1/src
...(change the source)
cd /usr/ports/mail/exim
make all deinstall reinstall

(when the segfault happens)
gdb -c /.../exim.core /usr/local/sbin/exim

I'm not sure, please correct me if I'm wrong.

--
## 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@lists.exim.org
## Exim details at http://www.exim.org/
## Please use the Wiki with this list - http://wiki.exim.org/

Re: SIGSEGV in panic log [ In reply to ]

exim-users at lists

Apr 7, 2024, 7:23 AM

Post #10 of 10 (90 views)

On 07/04/2024 14:18, Lena--- via Exim-users wrote:
> (when the segfault happens)
> gdb -c /.../exim.core /usr/local/sbin/exim

"bt" is the command to give to the gdb prompt for a stacktrace.

Gdb might offer to install all sorts of debug support packages for you
before it gives a command-prompt; you can decline these (we only care
about the parts that are within exim, not libraries, for the moment).
--
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@lists.exim.org
## Exim details at http://www.exim.org/
## Please use the Wiki with this list - http://wiki.exim.org/