Heiko Schlittermann (HS12-RIPE) [Sun, 28 Mar 2021 09:06:27 +0000 (11:06 +0200)]
Heiko Schlittermann (HS12-RIPE) [Sun, 28 Mar 2021 08:59:46 +0000 (10:59 +0200)]
Heiko Schlittermann (HS12-RIPE) [Sun, 28 Mar 2021 08:58:46 +0000 (10:58 +0200)]
Heiko Schlittermann (HS12-RIPE) [Sun, 28 Mar 2021 08:50:14 +0000 (10:50 +0200)]
Heiko Schlittermann (HS12-RIPE) [Sun, 28 Mar 2021 08:49:49 +0000 (10:49 +0200)]
Heiko Schlittermann (HS12-RIPE) [Thu, 1 Apr 2021 20:45:03 +0000 (22:45 +0200)]
Heiko Schlittermann (HS12-RIPE) [Thu, 1 Apr 2021 20:44:31 +0000 (22:44 +0200)]
Introduce main config option allow_insecure_tainted_data
This option is deprecated already now.
(cherry picked from commit
ec06d64532e4952fc36429f73e0222d26997ef7c)
Jeremy Harris [Tue, 22 Jun 2021 22:42:24 +0000 (23:42 +0100)]
GnuTLS: fix build with older GnuTLS
The ALPN handling we need requires later features than the basic functions.
Broken-byu:
f50a063dc0
Jeremy Harris [Tue, 22 Jun 2021 22:04:59 +0000 (23:04 +0100)]
TLS: as server, reject connections with ALPN indicating non-smtp use
Jeremy Harris [Mon, 21 Jun 2021 19:39:37 +0000 (20:39 +0100)]
Testsuite: fix testcases for non-TLS build
Jeremy Harris [Mon, 21 Jun 2021 19:22:23 +0000 (20:22 +0100)]
Testsuite: fix munging for no-TLS build
Broken-by: da40b1ec6b
Jeremy Harris [Sun, 20 Jun 2021 13:20:32 +0000 (14:20 +0100)]
Compiler quietening
Stupid static analysis failing to track crontrol dependencies
Jeremy Harris [Sat, 19 Jun 2021 19:12:09 +0000 (20:12 +0100)]
OpenSSL: on library versions too old to support session tickets
client-side limit the valid lifetime of resumable sessions
Jeremy Harris [Sat, 19 Jun 2021 18:11:43 +0000 (19:11 +0100)]
Testsuite: split out OpenSSL TLS1.3 resume tests
Older library versions do not support 1.3 so a separate numbered
testcase is needed
Jeremy Harris [Sat, 19 Jun 2021 18:10:26 +0000 (19:10 +0100)]
Testsuite: allow time for daemon to listen before terminating
Jeremy Harris [Thu, 17 Jun 2021 19:45:32 +0000 (20:45 +0100)]
OpenSSL: fix verify-certs stack initialization
Jeremy Harris [Thu, 17 Jun 2021 18:50:08 +0000 (19:50 +0100)]
Testsuite: output changes for OpenSSL library variants
Broken-by: 2f8e0a5f6b
Jeremy Harris [Thu, 17 Jun 2021 18:44:19 +0000 (19:44 +0100)]
Docs: typo
Jeremy Harris [Tue, 15 Jun 2021 18:27:04 +0000 (19:27 +0100)]
hosts_require_helo
Jeremy Harris [Sun, 13 Jun 2021 13:47:25 +0000 (14:47 +0100)]
Testsuite: EC cert
Jeremy Harris [Tue, 8 Jun 2021 20:42:23 +0000 (21:42 +0100)]
Fix server creds cache invalidation
Broken-by: 5fd673807d
Jeremy Harris [Mon, 7 Jun 2021 18:13:09 +0000 (19:13 +0100)]
compiler quietening
Jeremy Harris [Mon, 7 Jun 2021 17:47:14 +0000 (18:47 +0100)]
Re-fix non-Linux build
Jeremy Harris [Sun, 6 Jun 2021 21:23:03 +0000 (22:23 +0100)]
tidying
Vroken-by: ef77ddc923
Jeremy Harris [Sun, 6 Jun 2021 21:03:35 +0000 (22:03 +0100)]
Fix non-Linux build
Jeremy Harris [Sun, 6 Jun 2021 18:58:48 +0000 (19:58 +0100)]
Observability: listen queue backlog
Jeremy Harris [Sun, 6 Jun 2021 16:01:02 +0000 (17:01 +0100)]
Testsuite: testcase for multiple listener sockets ready
Jeremy Harris [Sun, 6 Jun 2021 13:05:02 +0000 (14:05 +0100)]
Avoid rescanning listen select set
Jeremy Harris [Sun, 6 Jun 2021 10:29:56 +0000 (11:29 +0100)]
Compute select fd_set outside daemon loop
Jeremy Harris [Sat, 5 Jun 2021 20:30:38 +0000 (21:30 +0100)]
Testsuite: fix OCSP/OpenSSL/1.3 testcase
Jeremy Harris [Sat, 5 Jun 2021 19:47:12 +0000 (20:47 +0100)]
Fix SSL creds file watching on kevent platforms (BSDs) for symlinks
Jeremy Harris [Fri, 4 Jun 2021 10:35:52 +0000 (11:35 +0100)]
DMARC: note unsupported library versions issue
Jeremy Harris [Tue, 1 Jun 2021 19:51:42 +0000 (20:51 +0100)]
debug: fix openssl output
Jeremy Harris [Tue, 1 Jun 2021 20:20:38 +0000 (21:20 +0100)]
Testsuite: regen certificates suite with fixed Authority Identifier
Jeremy Harris [Fri, 28 May 2021 19:04:44 +0000 (20:04 +0100)]
DKIM: under GnuTLS, permit weak algorithms
Recent versions of GnuTLS by default disallow use of some methods now regarded as
weak. This probably mean sha1, which is deprecated per DKIM standards.
Jeremy Harris [Fri, 28 May 2021 16:33:13 +0000 (17:33 +0100)]
Testsuite: use higher-spec certs, for more-recent GnuTLS versions which deprecate weaker ones
Needed for GnuTLS 3.6.15 (on Fedora 33)
Jeremy Harris [Fri, 28 May 2021 14:13:29 +0000 (15:13 +0100)]
tidying
Jeremy Harris [Fri, 28 May 2021 13:55:43 +0000 (14:55 +0100)]
Update testcase output to match newly applied default config limit
Broken-by: f07847e436
Jeremy Harris [Fri, 28 May 2021 13:41:00 +0000 (14:41 +0100)]
Fix testsuite output for DB cases
Broken-by: 186e99bafc
Jeremy Harris [Fri, 28 May 2021 13:09:45 +0000 (14:09 +0100)]
tidying
Jeremy Harris [Fri, 28 May 2021 12:33:49 +0000 (13:33 +0100)]
Logging: avoid pause during log-open under testsuite
It results in rearranged logging output, causing testsuite case failures
The downside is that we lose debug visbility of the extra process startup
Broken-by: b6c1434e47
Jeremy Harris [Fri, 28 May 2021 08:37:15 +0000 (09:37 +0100)]
Fix dmarc build
Broken-by: b6c1434e47
Jeremy Harris [Wed, 26 May 2021 12:41:13 +0000 (13:41 +0100)]
Docs: enhance section on redirect router :defer: & :fail:
Heiko Schlittermann (HS12-RIPE) [Thu, 27 May 2021 21:18:04 +0000 (23:18 +0200)]
Merge branch 'qualys-2020'
- all Qualys patches from 4.94.2
- all fixes from 4.94.2+fixes if not applied yet
Heiko Schlittermann (HS12-RIPE) [Fri, 30 Apr 2021 08:47:45 +0000 (10:47 +0200)]
Fix BDAT issue for body w/o trailing CRLF (again Bug 1974)
(cherry picked from commit
919111edac911ba9c15422eafd7c5bf14d416d26)
Heiko Schlittermann (HS12-RIPE) [Thu, 29 Apr 2021 22:37:53 +0000 (00:37 +0200)]
testsuite: reproduce BDAT with missing eol (Bug 1974)
(cherry picked from commit
e9cecc465a570c1a4f34b199eae6bdd0a52ee2b0)
Heiko Schlittermann (HS12-RIPE) [Mon, 26 Apr 2021 16:54:28 +0000 (18:54 +0200)]
Cleanup docs on cve-2020-qualys, point to the Exim website
(cherry picked from commit
6429b0fc79595f120703c022ae99aa10d698f909)
Heiko Schlittermann (HS12-RIPE) [Mon, 26 Apr 2021 14:16:49 +0000 (16:16 +0200)]
rewrite: revert to unchecked result of parse_extract_address()
Now it breaks 471, and overlong addresses won't make it into the rewrite
process, as they are handled as empty.
(cherry picked from commit
506286c62b8786a926dafb5bb05d3103492b86bc)
Heiko Schlittermann (HS12-RIPE) [Mon, 19 Apr 2021 20:23:14 +0000 (22:23 +0200)]
Honour the outcome of parse_extract_address(), testsuite 471
(cherry picked from commit
39d83bf19fc0c4364e0a665360b14194c62e4ab4)
Heiko Schlittermann (HS12-RIPE) [Wed, 21 Apr 2021 05:52:39 +0000 (07:52 +0200)]
Update upgrade notes and source about use of seteuid()
(cherry picked from commit
bc13bbca6e07267dfe0c4d275bb0a2e9aabf1dfb)
(cherry picked from commit
fee1a06ec05e58e0cda8cf04f28240688736f945)
Qualys Security Advisory [Tue, 23 Feb 2021 16:33:03 +0000 (08:33 -0800)]
CVE-2020-28007: Link attack in Exim's log directory
We patch this vulnerability by opening (instead of just creating) the
log file in an unprivileged (exim) child process, and by passing this
file descriptor back to the privileged (root) parent process. The two
functions log_send_fd() and log_recv_fd() are inspired by OpenSSH's
functions mm_send_fd() and mm_receive_fd(); thanks!
This patch also fixes:
- a NULL-pointer dereference in usr1_handler() (this signal handler is
installed before process_log_path is initialized);
- a file-descriptor leak in dmarc_write_history_file() (two return paths
did not close history_file_fd).
Note: the use of log_open_as_exim() in dmarc_write_history_file() should
be fine because the documentation explicitly states "Make sure the
directory of this file is writable by the user exim runs as."
(cherry picked from commit
2502cc41d1d92c1413eca6a4ba035c21162662bd)
(cherry picked from commit
93e9a18fbf09deb59bd133986f4c89aeb2d2d86a)
Heiko Schlittermann (HS12-RIPE) [Mon, 12 Apr 2021 21:05:44 +0000 (23:05 +0200)]
CVE-2020-28016: Heap out-of-bounds write in parse_fix_phrase()
Based on Phil Pennock's commit
76a1ce77.
Modified by Qualys.
(cherry picked from commit
f218fef171cbe9e61d10f15399aab8fa6956535b)
(cherry picked from commit
8b1e9bc2cac17ee24d595c97dcf97d9b016f8a46)
Heiko Schlittermann (HS12-RIPE) [Tue, 30 Mar 2021 20:59:25 +0000 (22:59 +0200)]
SECURITY: Avoid modification of constant data in dkim handling
Based on Heiko Schlittermann's commits
f880c7f3 and
c118c7f4. This
fixes:
6/ In src/pdkim/pdkim.c, pdkim_update_ctx_bodyhash() is sometimes called
with a global orig_data and hence canon_data, and the following line can
therefore modify data that should be constant:
773 canon_data->len = b->bodylength - b->signed_body_bytes;
For example, the following proof of concept sets lineending.len to 0
(this should not be possible):
(sleep 10; echo 'EHLO test'; sleep 3; echo 'MAIL FROM:<>'; sleep 3; echo 'RCPT TO:postmaster'; sleep 3; echo 'DATA'; date >&2; sleep 30; printf 'DKIM-Signature:a=rsa-sha1;c=simple/simple;l=0\r\n\r\n\r\nXXX\r\n.\r\n'; sleep 30) | nc -n -v 192.168.56.102 25
(gdb) print lineending
$1 = {data = 0x55e18035b2ad "\r\n", len = 2}
(gdb) print &lineending.len
$3 = (size_t *) 0x55e180385948 <lineending+8>
(gdb) watch *(size_t *) 0x55e180385948
Hardware watchpoint 1: *(size_t *) 0x55e180385948
Old value = 2
New value = 0
(gdb) print lineending
$5 = {data = 0x55e18035b2ad "\r\n", len = 0}
(cherry picked from commit
92359a62a0e31734ad8069c66f64b37f9eaaccbe)
(cherry picked from commit
c5f2f5cf2a6b45ae7ba0ed15e04fbe014727b210)
Heiko Schlittermann (HS12-RIPE) [Tue, 30 Mar 2021 20:48:06 +0000 (22:48 +0200)]
SECURITY: Leave a clean smtp_out input buffer even in case of read error
Based on Heiko Schlittermann's commit
54895bc3. This fixes:
7/ In src/smtp_out.c, read_response_line(), inblock->ptr is not updated
when -1 is returned. This does not seem to have bad consequences, but is
maybe not the intended behavior.
(cherry picked from commit
30f5d98786fb4e6ccfdd112fe65c153f0ee34c5f)
(cherry picked from commit
d600f6c4d0c5d33e3988dfbfee248ff6a1536673)
Qualys Security Advisory [Mon, 22 Feb 2021 06:09:06 +0000 (22:09 -0800)]
SECURITY: Always exit when LOG_PANIC_DIE is set
(cherry picked from commit
e20aa895b37f449d5c81c3e7b102fc534b5d23ba)
(cherry picked from commit
3b8c0ceb7339329188e19efb907da950dbe691d1)
Qualys Security Advisory [Mon, 22 Feb 2021 05:53:55 +0000 (21:53 -0800)]
CVE-2020-28012: Missing close-on-exec flag for privileged pipe
(cherry picked from commit
72dad1e64bb3d1ff387938f59678098cab1f60a3)
(cherry picked from commit
645a31d16195bb6b73f0a0d0c04b2251e5b28421)
Qualys Security Advisory [Mon, 22 Feb 2021 05:49:30 +0000 (21:49 -0800)]
CVE-2020-28024: Heap buffer underflow in smtp_ungetc()
(cherry picked from commit
998e5a9db121c3eff15cac16859bdffd7adcbe57)
(cherry picked from commit
638f7ca75694bcbb70cfbe7db2ef52af4aca5c83)
Qualys Security Advisory [Mon, 22 Feb 2021 05:45:19 +0000 (21:45 -0800)]
CVE-2020-28009: Integer overflow in get_stdinput()
(cherry picked from commit
bbf1bb10bee5a1d7cbcc97f178b348189219eb7d)
(cherry picked from commit
1241deaefb71c40436320af7d0bd04c7c9e54241)
Qualys Security Advisory [Mon, 22 Feb 2021 05:26:53 +0000 (21:26 -0800)]
CVE-2020-28015+28021: New-line injection into spool header file
(cherry picked from commit
31b1a42d0bd29cb05f85e56d3343b13bef20a2bd)
(cherry picked from commit
fcddccd650178ceeec3655c6c40f420164a8706e)
Heiko Schlittermann (HS12-RIPE) [Tue, 30 Mar 2021 20:03:49 +0000 (22:03 +0200)]
CVE-2020-28026: Line truncation and injection in spool_read_header()
This also fixes:
2/ In src/spool_in.c:
462 while ( (len = Ustrlen(big_buffer)) == big_buffer_size-1
463 && big_buffer[len-1] != '\n'
464 )
465 { /* buffer not big enough for line; certs make this possible */
466 uschar * buf;
467 if (big_buffer_size >= BIG_BUFFER_SIZE*4) goto SPOOL_READ_ERROR;
468 buf = store_get_perm(big_buffer_size *= 2, FALSE);
469 memcpy(buf, big_buffer, --len);
The --len in memcpy() chops off a useful byte (we know for sure that
big_buffer[len-1] is not a '\n' because we entered the while loop).
Based on a patch done by Qualys.
(cherry picked from commit
f0c307458e1ee81abbe7ed2d4a8d16b5cbd8a799)
(cherry picked from commit
4daba4bec729a57fb0863af786a1395e70794c76)
Heiko Schlittermann (HS12-RIPE) [Mon, 29 Mar 2021 21:12:02 +0000 (23:12 +0200)]
CVE-2020-28022: Heap out-of-bounds read and write in extract_option()
Based on Phil Pennock's commit
c5017adf.
(cherry picked from commit
9e941e1807b624b255c9ec0f41a0b3a89e144de3)
(cherry picked from commit
33d4c87653ddbbea9fd8cb8eb2ff78c149850006)
Heiko Schlittermann (HS12-RIPE) [Mon, 29 Mar 2021 21:05:58 +0000 (23:05 +0200)]
CVE-2020-28017: Integer overflow in receive_add_recipient()
Based on Phil Pennock's commit
e3b441f7.
(cherry picked from commit
18a19e18242edc5ab2082fa9c41cd6210d1b6087)
(cherry picked from commit
605716b999a4ca6c7d5777ab7463058e9b055dc2)
Heiko Schlittermann (HS12-RIPE) [Mon, 29 Mar 2021 21:02:34 +0000 (23:02 +0200)]
SECURITY: Refuse negative and large store allocations
Based on Phil Pennock's commits
b34d3046 and
e6c1606a. Done by Qualys.
(cherry picked from commit
09d36bd64fc5bf71d8882af35c41ac4e8599acc1)
(cherry picked from commit
f9c58fb385343b8e3fa13988efcbd30ae3285ea7)
Heiko Schlittermann (HS12-RIPE) [Mon, 29 Mar 2021 20:44:47 +0000 (22:44 +0200)]
CVE-2020-28013: Heap buffer overflow in parse_fix_phrase()
Based on Phil Pennock's
8a50c88a, done by Qualys
(cherry picked from commit
8161c16ec7320ac6164954bade23179a0ed095eb)
(cherry picked from commit
71585e8fcb8704a9f431f5a8d019280cccaad069)
Qualys Security Advisory [Mon, 22 Feb 2021 03:22:33 +0000 (19:22 -0800)]
CVE-2020-28011: Heap buffer overflow in queue_run()
(cherry picked from commit
6e1fb878e95f8e6f838ffde5258c7a969c981865)
(cherry picked from commit
08102cbe8102f99b31655aa0e926c45b427efe6d)
Heiko Schlittermann (HS12-RIPE) [Mon, 29 Mar 2021 20:16:28 +0000 (22:16 +0200)]
CVE-2020-28010: Heap out-of-bounds write in main()
Based on Phil Pennock's
0f57feb4. Done by Qualys, modified by me.
(cherry picked from commit
b0982c2776048948ebae48574b70fa487684cb8c)
(cherry picked from commit
dbc3ab675c2e5e2a07ed13dc5ede4daa018600e7)
Qualys Security Advisory [Mon, 22 Feb 2021 03:05:56 +0000 (19:05 -0800)]
CVE-2020-28018: Use-after-free in tls-openssl.c
(cherry picked from commit
6290686dd59d8158d100c67e8f96df27158a6fc5)
(cherry picked from commit
a53a7fcfb8216764e4420d8d263356b4ed7d5cef)
Qualys Security Advisory [Mon, 22 Feb 2021 02:54:16 +0000 (18:54 -0800)]
CVE-2020-28025: Heap out-of-bounds read in pdkim_finish_bodyhash()
(cherry picked from commit
cad30cd3fb96196e908e0d66b1b45fdf377c850c)
(cherry picked from commit
1c261b90f627f0489f7dfcf1e66b46cce67f477d)
Heiko Schlittermann (HS12-RIPE) [Thu, 25 Mar 2021 21:48:09 +0000 (22:48 +0100)]
CVE-2020-28014, CVE-2021-27216: PID file handling
Arbitrary PID file creation, clobbering, and deletion.
Patch provided by Qualys.
(cherry picked from commit
974f32939a922512b27d9f0a8a1cb5dec60e7d37)
(cherry picked from commit
43c6f0b83200b7082353c50187ef75de3704580a)
Heiko Schlittermann (HS12-RIPE) [Wed, 10 Mar 2021 22:37:29 +0000 (23:37 +0100)]
Add priv.c: reworked version of priv dropping code
(cherry picked from commit
82b545236e6dc82b7af34528c532811bfc74ea19)
(cherry picked from commit
be31ef213f118abe5fc68732f5492b6b16d28b87)
Heiko Schlittermann (HS12-RIPE) [Sun, 14 Mar 2021 11:16:57 +0000 (12:16 +0100)]
CVE-2020-28008: Assorted attacks in Exim's spool directory
We patch dbfn_open() by introducing two functions priv_drop_temp() and
priv_restore() (inspired by OpenSSH's functions temporarily_use_uid()
and restore_uid()), which temporarily drop and restore root privileges
thanks to seteuid(). This goes against Exim's developers' wishes ("Exim
(the project) doesn't trust seteuid to work reliably") but, to the best
of our knowledge, seteuid() works everywhere and is the only way to
securely fix dbfn_open().
(cherry picked from commit
18da59151dbafa89be61c63580bdb295db36e374)
(cherry picked from commit
b05dc3573f4cd476482374b0ac0393153d344338)
Jeremy Harris [Thu, 4 Mar 2021 21:19:08 +0000 (22:19 +0100)]
CVE-2020-28019: Failure to reset function pointer after BDAT error
Based on Phil Pennock's commits
4715403e and
151ffd72, and Jeremy
Harris's commits
aa171254 and
9aceb5c2.
(cherry picked from commit
0a3fbb7e3be375bc93b8e359c6aff333c7c2d76f)
(cherry picked from commit
99d057fad97a2def9f000ebccda83e4008112819)
Heiko Schlittermann (HS12-RIPE) [Wed, 2 Dec 2020 21:28:02 +0000 (22:28 +0100)]
SECURITY: smtp_out: Leave a clean input buffer, even in case of read error
Credits: Qualys
7/ In src/smtp_out.c, read_response_line(), inblock->ptr is not updated
when -1 is returned. This does not seem to have bad consequences, but is
maybe not the intended behavior.
(cherry picked from commit
f7ac5a7d1e817bf60f161e7a1d40b65d66da607f)
(cherry picked from commit
13f9998ebb937970d1d9d18f205a6e03e14105b4)
Heiko Schlittermann (HS12-RIPE) [Thu, 26 Nov 2020 17:16:59 +0000 (18:16 +0100)]
SECURITY: Avoid modification of constant data
Credits: Qualys
6/ In src/pdkim/pdkim.c, pdkim_update_ctx_bodyhash() is sometimes called
with a global orig_data and hence canon_data, and the following line can
therefore modify data that should be constant:
773 canon_data->len = b->bodylength - b->signed_body_bytes;
For example, the following proof of concept sets lineending.len to 0
(this should not be possible):
(sleep 10; echo 'EHLO test'; sleep 3; echo 'MAIL FROM:<>'; sleep 3; echo 'RCPT TO:postmaster'; sleep 3; echo 'DATA'; date >&2; sleep 30; printf 'DKIM-Signature:a=rsa-sha1;c=simple/simple;l=0\r\n\r\n\r\nXXX\r\n.\r\n'; sleep 30) | nc -n -v 192.168.56.102 25
(gdb) print lineending
$1 = {data = 0x55e18035b2ad "\r\n", len = 2}
(gdb) print &lineending.len
$3 = (size_t *) 0x55e180385948 <lineending+8>
(gdb) watch *(size_t *) 0x55e180385948
Hardware watchpoint 1: *(size_t *) 0x55e180385948
Old value = 2
New value = 0
(gdb) print lineending
$5 = {data = 0x55e18035b2ad "\r\n", len = 0}
(cherry picked from commit
9fce76f56459dde7489eb21ce1ff822e04e10f43)
(cherry picked from commit
667fb25b8f0dc3fbac57bce4051e345555fa776a)
Heiko Schlittermann (HS12-RIPE) [Wed, 25 Nov 2020 22:19:57 +0000 (23:19 +0100)]
SECURITY: Avoid memory corruption in dkim handling
Credits: Qualys
6/ In src/pdkim/pdkim.c, pdkim_update_ctx_bodyhash() is sometimes called
with a global orig_data and hence canon_data, and the following line can
therefore modify data that should be constant:
773 canon_data->len = b->bodylength - b->signed_body_bytes;
For example, the following proof of concept sets lineending.len to 0
(this should not be possible):
(sleep 10; echo 'EHLO test'; sleep 3; echo 'MAIL FROM:<>'; sleep 3; echo 'RCPT TO:postmaster'; sleep 3; echo 'DATA'; date >&2; sleep 30; printf 'DKIM-Signature:a=rsa-sha1;c=simple/simple;l=0\r\n\r\n\r\nXXX\r\n.\r\n'; sleep 30) | nc -n -v 192.168.56.102 25
(gdb) print lineending
$1 = {data = 0x55e18035b2ad "\r\n", len = 2}
(gdb) print &lineending.len
$3 = (size_t *) 0x55e180385948 <lineending+8>
(gdb) watch *(size_t *) 0x55e180385948
Hardware watchpoint 1: *(size_t *) 0x55e180385948
Old value = 2
New value = 0
(gdb) print lineending
$5 = {data = 0x55e18035b2ad "\r\n", len = 0}
(cherry picked from commit
ea850e27714ccda2090d781ebe89b410bc38c2c6)
(cherry picked from commit
4e784efa28b25683c81a857b464777df593cabee)
Heiko Schlittermann (HS12-RIPE) [Wed, 25 Nov 2020 21:58:58 +0000 (22:58 +0100)]
SECURITY: Avoid decrement of dkim_collect_input if already at 0
Credits: Qualys
5/ receive_msg() calls dkim_exim_verify_finish(), which sets
dkim_collect_input to 0 and calls pdkim_feed_finish(), which calls
pdkim_header_complete(), which decreases dkim_collect_input to UINT_MAX,
which reactivates the DKIM code.
As a result, pdkim_feed() is called again (through receive_getc at the
end of receive_msg()), but functions like pdkim_finish_bodyhash() and
exim_sha_finish() have already been called (in pdkim_feed_finish()).
This suggests a use-after-free.
But it seems that a use-after-free would happen only with
EVP_DigestFinal() (in exim_sha_finish()), which does not seem to be
reachable via DKIM (no SHA3). But we checked OpenSSL only, not GnuTLS.
Here is a proof of concept that triggers the bug (which came very close
to a security vulnerability):
(sleep 10; echo 'EHLO test'; sleep 3; echo 'MAIL FROM:<>'; sleep 3; echo 'RCPT TO:postmaster'; sleep 3; echo 'BDAT 42 LAST'; date >&2; sleep 30; printf 'not a valid header line\r\nDKIM-Signature:\r\nXXX'; sleep 30) | nc -n -v 192.168.56.102 25
(gdb) print &dkim_collect_input
$2 = (unsigned int *) 0x55e180386d90 <dkim_collect_input>
(gdb) watch *(unsigned int *) 0x55e180386d90
Hardware watchpoint 1: *(unsigned int *) 0x55e180386d90
Old value = 0
New value =
4294967295
#0 0x000055e18031f805 in pdkim_header_complete (ctx=ctx@entry=0x55e181b9e8e0) at pdkim.c:1006
#1 0x000055e18032106c in pdkim_feed_finish (ctx=0x55e181b9e8e0, return_signatures=0x55e180386d78 <dkim_signatures>, err=err@entry=0x7ffe443e1d00) at pdkim.c:1490
#2 0x000055e1802a3280 in dkim_exim_verify_finish () at dkim.c:328
#3 0x000055e1802c9d1d in receive_msg (extract_recip=extract_recip@entry=0) at receive.c:3409
(cherry picked from commit
e3674091056ac05eb7ef1c504accce790c434bd7)
(cherry picked from commit
8b39dd074e3ec70cbda70a52cef5b71ecbf69499)
Heiko Schlittermann (HS12-RIPE) [Wed, 25 Nov 2020 21:26:53 +0000 (22:26 +0100)]
SECURITY: Check overrun rcpt_count integer
Credits: Qualys
4/ In src/smtp_in.c:
4966 case RCPT_CMD:
4967 HAD(SCH_RCPT);
4968 rcpt_count++;
....
5123 if (rcpt_count > recipients_max && recipients_max > 0)
In theory this recipients_max check can be bypassed, because the int
rcpt_count can overflow (become negative). In practice this would either
consume too much memory or generate too much network traffic, but maybe
it should be fixed anyway.
(cherry picked from commit
04139ca809fbe56d8fe9c55a77640ea9fa93b8f1)
(cherry picked from commit
db96ca55137d7684a9afdf9d118feed9116906b7)
Heiko Schlittermann (HS12-RIPE) [Sat, 21 Nov 2020 21:41:28 +0000 (22:41 +0100)]
SECURITY: Fix safeguard against upward traversal in msglog files.
Credits: Qualys
3/ In src/deliver.c:
333 static int
334 open_msglog_file(uschar *filename, int mode, uschar **error)
335 {
336 if (Ustrstr(filename, US"/../"))
337 log_write(0, LOG_MAIN|LOG_PANIC,
338 "Attempt to open msglog file path with upward-traversal: '%s'\n", filename);
Should this be LOG_PANIC_DIE instead of LOG_PANIC? Right now it will log
the /../ attempt but will open the file anyway.
(cherry picked from commit
742c27f02d83792937dcb1719b380d3dde6228bf)
(cherry picked from commit
1e9a340c05d7233969637095a8a6378b14de2976)
Heiko Schlittermann (HS12-RIPE) [Sat, 21 Nov 2020 21:18:56 +0000 (22:18 +0100)]
SECURITY: Don't miss the very last byte when reading long lines from -H
Credits: Qualys
2/ In src/spool_in.c:
462 while ( (len = Ustrlen(big_buffer)) == big_buffer_size-1
463 && big_buffer[len-1] != '\n'
464 )
465 { /* buffer not big enough for line; certs make this possible */
466 uschar * buf;
467 if (big_buffer_size >= BIG_BUFFER_SIZE*4) goto SPOOL_READ_ERROR;
468 buf = store_get_perm(big_buffer_size *= 2, FALSE);
469 memcpy(buf, big_buffer, --len);
The --len in memcpy() chops off a useful byte (we know for sure that
big_buffer[len-1] is not a '\n' because we entered the while loop).
(cherry picked from commit
58454ea01c2e817481770954edf09ad82f3cd417)
(cherry picked from commit
2d9f1837bdd6c5946cb9cd997544eefc8cc14fc4)
Heiko Schlittermann (HS12-RIPE) [Sat, 21 Nov 2020 21:03:03 +0000 (22:03 +0100)]
SECURITY: off-by-one in smtp transport (read response)
Credits: Qualys
1/ In src/transports/smtp.c:
2281 int n = sizeof(sx->buffer);
2282 uschar * rsp = sx->buffer;
2283
2284 if (sx->esmtp_sent && (n = Ustrlen(sx->buffer)) < sizeof(sx->buffer)/2)
2285 { rsp = sx->buffer + n + 1; n = sizeof(sx->buffer) - n; }
This should probably be either:
rsp = sx->buffer + n + 1; n = sizeof(sx->buffer) - n - 1;
or:
rsp = sx->buffer + n; n = sizeof(sx->buffer) - n;
(not sure which) to avoid an off-by-one.
(cherry picked from commit
d2c44ef5dd94f1f43ba1d1a02bc4594f4fba5e38)
(cherry picked from commit
4045cb01a590ec480f45f80967cd9c59fe23a5d0)
Phil Pennock [Mon, 2 Nov 2020 07:39:38 +0000 (02:39 -0500)]
Start documenting the things we changed incompatibly.
(cherry picked from commit
8dad4da53bad2ed3b29fa6a3b9ef59bfec73dc0e)
(cherry picked from commit
125f0d4afbc858cf514c29326a3016c2d9d7bdc1)
Phil Pennock [Mon, 2 Nov 2020 07:23:14 +0000 (02:23 -0500)]
Inline four often-called new functions
The BDAT state switchers should happen so often during SMTP reception that a
compiler hint to inline seems wise.
The length filter checks happen on every start-up, which for Exim is often
enough that I think an inline these is warranted too.
(cherry picked from commit
6e3d0e3f1c8228ef19a3d1ba61f131cef3172ceb)
(cherry picked from commit
6ca5fcba34070f4495a0188f16eb2e4d78f3430a)
Jeremy Harris [Sat, 31 Oct 2020 14:36:55 +0000 (14:36 +0000)]
Phil Pennock [Fri, 30 Oct 2020 03:21:36 +0000 (23:21 -0400)]
SECURITY: rework BDAT receive function handling
(cherry picked from commit
dd1b9b753bb7c42df2b8f48d726b82928b67940b)
(cherry picked from commit
96fb195ebc2eb6790e6ad6dde46d478aee62198d)
Phil Pennock [Fri, 30 Oct 2020 02:40:59 +0000 (22:40 -0400)]
SECURITY: fix SMTP verb option parsing
A boundary case in looking for an opening quote before the closing quote could
walk off the front of the buffer.
(cherry picked from commit
515d8d43a18481d23d7cf410b8dc71b4e254ebb8)
(cherry picked from commit
467948de0c407bd2bbc2e84abbbf09f35b035538)
Phil Pennock [Fri, 30 Oct 2020 01:48:05 +0000 (21:48 -0400)]
SECURITY: Avoid integer overflow on too many recipients
(cherry picked from commit
323ff55e67b44e95f9d3cfaba155e385aa33c4bd)
(cherry picked from commit
3a54fcd1e303bf1cc49beca7ceac35d7448860a9)
Phil Pennock [Fri, 30 Oct 2020 01:38:25 +0000 (21:38 -0400)]
SECURITY: default recipients_max to 50,000
A default of "unlimited" can have unfortunate consequences when people start
putting many millions of recipients on a message.
(cherry picked from commit
1d7780722a66cea8da5fa4ae0775e85d185fbf7e)
(cherry picked from commit
a6e1f69d82adcfd3caab8f228d96750dfddc8f07)
Phil Pennock [Fri, 30 Oct 2020 01:30:04 +0000 (21:30 -0400)]
SECURITY: a second negative store guard
(cherry picked from commit
706864e934c70941ce7a327f97b7649a1e5f5556)
(cherry picked from commit
9f06dcd6848052f2524658bf871c60a8d48c7dbe)
Phil Pennock [Fri, 30 Oct 2020 00:49:49 +0000 (20:49 -0400)]
SECURITY: refuse too small store allocations
Negative sizes are definitely bad.
Optimistically, I'm saying that zero is bad too. But perhaps we have something
doing that, expecting to be able to grow. In which case we'll have to amend
this.
(cherry picked from commit
1c9afcec0043e2fb72607b2addb0613763705549)
(cherry picked from commit
6f5d7e5af8eff688c36f81334e4f063689561963)
Phil Pennock [Fri, 30 Oct 2020 00:42:40 +0000 (20:42 -0400)]
SECURITY: fix Qualys CVE-2020-PFPZA
(cherry picked from commit
29d7a8c25f182c91d5d30f124f9e296dce5c018e)
(cherry picked from commit
0a6a7a3fd8464bae9ce0cf889e8eeb0bf0bab756)
Phil Pennock [Thu, 29 Oct 2020 23:00:51 +0000 (19:00 -0400)]
SECURITY: fix Qualys CVE-2020-PFPSN
(cherry picked from commit
93b6044e1636404f3463f3e1113098742e295542)
(cherry picked from commit
4e59a5d5c448e1fcdcbead268ffe6561adf0224d)
Phil Pennock [Thu, 29 Oct 2020 15:47:58 +0000 (11:47 -0400)]
SECURITY: fix Qualys CVE-2020-SLCWD
(cherry picked from commit
bf5f9d56fadf9be8d947f141d31f7e0e8fa63762)
(cherry picked from commit
6d2cfb575c95c1b81597d6b9eb2904cd695d7e4a)
Phil Pennock [Thu, 29 Oct 2020 22:40:37 +0000 (18:40 -0400)]
SECURITY: pick up more argv length checks
(cherry picked from commit
f28a6a502c7973d8844d11d4b0990d4b0359fb3f)
(cherry picked from commit
7a7136ba7f5c2db33c7e320ffd4675335c4557e5)
Phil Pennock [Thu, 29 Oct 2020 22:11:35 +0000 (18:11 -0400)]
SECURITY: length limits on many cmdline options
We'll also now abort upon, rather than silently truncate, a driver name
(router, transport, ACL, etc) encountered in the config which is longer than
the 64-char limit.
(cherry picked from commit
ff8bef9ae2370db4a7873fe2ce573a607fe6999f)
(cherry picked from commit
a8bd24b96c2027fd839f95a9e6b3282453ae288e)
Phil Pennock [Fri, 18 Sep 2020 14:25:42 +0000 (10:25 -0400)]
Re-ran the conversion of all DH parameters
I get different results now to those I got before.
Now, using gen_pkcs3 linked against OpenSSL 1.1.1f-1ubuntu2 on Focal Fossa, I
get the results below. The ffdhe2048 value now matches that at
<https://ssl-config.mozilla.org/ffdhe2048.txt>.
I ran the same code yesterday for just the ffdhe2048 item and got code which
seemed to me then to match what was already in the C file. Something hinky is
going on, perhaps with my sanity.
(the commit IDs changee because of heavy rebasing (heiko))
(cherry picked from commit
76ed8115182e2daaadb437ec9655df8000796ec5)
(cherry picked from commit
0aafa26a5d3d528e79476c91537c28936154fe04)
Simon Arlott [Thu, 24 Sep 2020 22:03:14 +0000 (23:03 +0100)]
gen_pkcs3: Terminate string before calling BH_hex2bn()
Signed-off-by: Phil Pennock <pdp@exim.org>
(cherry picked from commit
1cf66e5872d517b620c308af634e4e26e3547f06)
(cherry picked from commit
48d8c54ecf9493c709d4305850877b6062f285a7)
Phil Pennock [Fri, 30 Oct 2020 02:14:45 +0000 (22:14 -0400)]
Default config: reject on too many bad RCPT
An example exploit failed against my system, because I had this sanity guard in
place; it's not a real security fix since a careful attacker could find enough
valid recipients to hit that problem, but it highlights that this is a useful
enough pattern that we should encourage its wider use.
(cherry picked from commit
2a636a39fff29b7c3da1798767a510dfed982a62)
(cherry picked from commit
346f96bad326893f9c1fa772a5b8ac35b1f8f7bd)
Heiko Schlittermann (HS12-RIPE) [Thu, 18 Mar 2021 06:56:59 +0000 (07:56 +0100)]
Handle SIGINT as we do with SIGTERM
(cherry picked from commit
cdc5c672e1c309294626cd5ed90acdccb05baaa1)
(cherry picked from commit
f9c8211fb0ad0dd362f471978a5e0abc5dfa71b4)
Heiko Schlittermann (HS12-RIPE) [Thu, 18 Mar 2021 06:59:21 +0000 (07:59 +0100)]
Enforce pid_file_path start at "/"
(cherry picked from commit
60f2a8e797d9ebaea1e3eac4ad28ff64e11bab40)
(cherry picked from commit
6b3d553c733475a1033c8b7a241e6506d7ed73b1)
Heiko Schlittermann (HS12-RIPE) [Mon, 29 Mar 2021 16:17:55 +0000 (18:17 +0200)]
testsuite: tidy logs/4520 and confs/4520
This fixed 4520 failure en-passant, but I'm sure it's a timing issue
here (the order of the mainlog output lines didn't exactly match the
logs/4520)
(cherry picked from commit
95306ca61531d9d79c5dac808a5a571158acd29c)
(cherry picked from commit
0439d2e0566d64c84feaf1434e0e4a3fd8ce29b3)