Commit Graph

10074 Commits

Author SHA1 Message Date
Damien Miller 13b3369830 avoid "return (value)" in void-declared function
spotted by Tim Rice; ok dtucker
2019-10-08 15:32:02 +11:00
Darren Tucker 0c7f8d2326 Make DEF_WEAK more likely to be correct.
Completely nop-ing out DEF_WEAK leaves an empty statemment which some
compilers don't like.  Replace with a no-op function template.  ok djm@
2019-10-08 14:48:32 +11:00
dtucker@openbsd.org b1e79ea8fa upstream: Instead of running sed over the whole log to remove CRs,
remove them only where it's needed (and confuses test(1) on at least OS X in
portable).

OpenBSD-Regress-ID: a6ab9b4bd1d33770feaf01b2dfb96f9e4189d2d0
2019-10-07 13:08:57 +11:00
Eduardo Barretto 8dc7d6b75a Enable specific ioctl call for EP11 crypto card (s390)
The EP11 crypto card needs to make an ioctl call, which receives an
specific argument. This crypto card is for s390 only.

Signed-off-by: Eduardo Barretto <ebarretto@linux.vnet.ibm.com>
2019-10-05 18:30:40 +10:00
djm@openbsd.org 07f2c7f349 upstream: fix memory leak in error path; bz#3074 patch from
krishnaiah.bommu@intel.com, ok dtucker

OpenBSD-Commit-ID: d031853f3ecf47b35a0669588f4d9d8e3b307b3c
2019-10-04 14:34:05 +10:00
djm@openbsd.org b7fbc75e11 upstream: space
OpenBSD-Commit-ID: 350648bcf00a2454e7ef998b7d88e42552b348ac
2019-10-04 14:34:05 +10:00
djm@openbsd.org 643ab68c79 upstream: more sshsig regress tests: check key revocation, the
check-novalidate signature test mode and signing keys in ssh-agent.

From Sebastian Kinne (slightly tweaked)

OpenBSD-Regress-ID: b39566f5cec70140674658cdcedf38752a52e2e2
2019-10-04 13:41:03 +10:00
dtucker@openbsd.org 714031a10b upstream: Check for gmtime failure in moduli generation. Based on
patch from krishnaiah.bommu@intel.com, ok djm@

OpenBSD-Commit-ID: 4c6a4cde0022188ac83737de08da0e875704eeaa
2019-10-04 13:40:57 +10:00
jmc@openbsd.org 6918974405 upstream: use a more common options order in SYNOPSIS and sync
usage(); while here, no need for Bk/Ek;

ok dtucker

OpenBSD-Commit-ID: 38715c3f10b166f599a2283eb7bc14860211bb90
2019-10-04 13:40:57 +10:00
djm@openbsd.org feff96b7d4 upstream: thinko in previous; spotted by Mantas
=?UTF-8?q?=20Mikul=C4=97nas?=
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

OpenBSD-Commit-ID: ffa3f5a45e09752fc47d9041e2203ee2ec15b24d
2019-10-02 19:53:40 +10:00
djm@openbsd.org b5a89eec41 upstream: make signature format match PROTOCO
=?UTF-8?q?=20as=20a=20string,=20not=20raw=20bytes.=20Spotted=20by=20Manta?=
=?UTF-8?q?s=20Mikul=C4=97nas?=
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

OpenBSD-Commit-ID: 80fcc6d52893f80c6de2bedd65353cebfebcfa8f
2019-10-02 18:08:17 +10:00
djm@openbsd.org dc6f81ee94 upstream: ban empty namespace strings for s
=?UTF-8?q?shsig;=20spotted=20by=20Mantas=20Mikul=C4=97nas?=
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

OpenBSD-Commit-ID: 7c5bcf40bed8f4e826230176f4aa353c52aeb698
2019-10-02 18:08:17 +10:00
Darren Tucker fa5bd8107e Put ssherr.h back as it's actually needed. 2019-10-02 14:30:55 +10:00
Lonnie Abelbeck 3ef92a6574 Deny (non-fatal) shmget/shmat/shmdt in preauth privsep child.
New wait_random_seeded() function on OpenSSL 1.1.1d uses shmget, shmat, and shmdt
in the preauth codepath, deny (non-fatal) in seccomp_filter sandbox.
2019-10-02 12:24:38 +10:00
Damien Miller edd1d3a626 remove duplicate #includes
Prompted by Jakub Jelen
2019-10-02 10:54:28 +10:00
Damien Miller 13c508dfed typo in comment 2019-10-02 10:51:15 +10:00
djm@openbsd.org d0c3ac427f upstream: remove some duplicate #includes
OpenBSD-Commit-ID: ed6827ab921eff8027669848ef4f70dc1da4098c
2019-10-02 10:43:47 +10:00
djm@openbsd.org 084682786d upstream: revert unconditional forced login implemented in r1.41 of
ssh-pkcs11.c; r1.45 added a forced login as a fallback for cases where the
token returns no objects and this is less disruptive for users of tokens
directly in ssh (rather than via ssh-agent) and in ssh-keygen

bz3006, patch from Jakub Jelen; ok markus

OpenBSD-Commit-ID: 33d6df589b072094384631ff93b1030103b3d02e
2019-10-01 20:24:07 +10:00
jmc@openbsd.org 6c91d42cce upstream: group and sort single letter options; ok deraadt
OpenBSD-Commit-ID: e1480e760a2b582f79696cdcff70098e23fc603f
2019-10-01 20:24:07 +10:00
jmc@openbsd.org 3b44bf39ff upstream: fix the DH-GEX text in -a; because this required a comma,
i added a comma to the first part, for balance...

OpenBSD-Commit-ID: 2c3464e9e82a41e8cdfe8f0a16d94266e43dbb58
2019-10-01 20:24:07 +10:00
deraadt@openbsd.org 3e53ef28fa upstream: identity_file[] should be PATH_MAX, not the arbitrary
number 1024

OpenBSD-Commit-ID: e775f94ad47ce9ab37bd1410d7cf3b7ea98b11b7
2019-10-01 20:24:07 +10:00
jmc@openbsd.org 90d4b2541e upstream: new sentence, new line;
OpenBSD-Commit-ID: c35ca5ec07be460e95e7406af12eee04a77b6698
2019-10-01 20:24:07 +10:00
Darren Tucker fbec7dba01 Include stdio.h for snprintf.
Patch from vapier@gentoo.org.
2019-09-30 18:01:12 +10:00
Darren Tucker 0a403bfde7 Add SKIP_LTESTS for skipping specific tests. 2019-09-30 14:11:42 +10:00
dtucker@openbsd.org 4d59f7a516 upstream: Test for empty result in expected bits. Remove CRs from log
as they confuse tools on some platforms.  Re-enable the 3des-cbc test.

OpenBSD-Regress-ID: edf536d4f29fc1ba412889b37247a47f1b49d250
2019-09-27 15:36:22 +10:00
Darren Tucker 7c817d129e Re-enable dhgex test.
Since we've added larger fallback groups to dh.c this test will pass
even if there is no moduli file installed on the system.
2019-09-27 15:26:22 +10:00
Darren Tucker c1e0a32fa8 Add more ToS bits, currently only used by netcat. 2019-09-24 21:17:20 +10:00
Darren Tucker 5a273a33ca Privsep is now required. 2019-09-19 15:41:23 +10:00
djm@openbsd.org 8aa2aa3cd4 upstream: Allow testing signature syntax and validity without verifying
that a signature came from a trusted signer. To discourage accidental or
unintentional use, this is invoked by the deliberately ugly option name
"check-novalidate"

from Sebastian Kinne

OpenBSD-Commit-ID: cea42c36ab7d6b70890e2d8635c1b5b943adcc0b
2019-09-16 13:25:53 +10:00
djm@openbsd.org 7047d5afe3 upstream: clarify that IdentitiesOnly also applies to the default
~/.ssh/id_* keys; bz#3062

OpenBSD-Commit-ID: 604be570e04646f0f4a17026f8b2aada6a585dfa
2019-09-13 14:53:45 +10:00
dtucker@openbsd.org b36ee3fcb2 upstream: Plug mem leaks on error paths, based in part on github
pr#120 from David Carlier.  ok djm@.

OpenBSD-Commit-ID: c57adeb1022a8148fc86e5a88837b3b156dbdb7e
2019-09-13 14:53:45 +10:00
djm@openbsd.org 2aefdf1aef upstream: whitespace
OpenBSD-Commit-ID: 57a71dd5f4cae8d61e0ac631a862589fb2bfd700
2019-09-13 14:53:45 +10:00
djm@openbsd.org fbe24b1429 upstream: allow %n to be expanded in ProxyCommand strings
From Zachary Harmany via github.com/openssh/openssh-portable/pull/118
ok dtucker@

OpenBSD-Commit-ID: 7eebf1b7695f50c66d42053d352a4db9e8fb84b6
2019-09-13 14:28:44 +10:00
djm@openbsd.org 2ce1d11600 upstream: clarify that ConnectTimeout applies both to the TCP
connection and to the protocol handshake/KEX. From Jean-Charles Longuet via
Github PR140

OpenBSD-Commit-ID: ce1766abc6da080f0d88c09c2c5585a32b2256bf
2019-09-13 14:09:21 +10:00
dtucker@openbsd.org df78011427 upstream: Fix potential truncation warning. ok deraadt.
OpenBSD-Commit-ID: d87b7e3a94ec935e8194e7fce41815e22804c3ff
2019-09-13 14:09:20 +10:00
Damien Miller ec0e624366 memleak of buffer in sshpam_query
coverity report via Ed Maste; ok dtucker@
2019-09-13 13:15:19 +10:00
Damien Miller c17e4638e5 explicitly test set[ug]id() return values
Legacy !_POSIX_SAVED_IDS path only; coverity report via Ed Maste
ok dtucker@
2019-09-13 13:15:14 +10:00
naddy@openbsd.org 91a2135f32 upstream: Allow prepending a list of algorithms to the default set
by starting the list with the '^' character, e.g.

HostKeyAlgorithms ^ssh-ed25519
Ciphers ^aes128-gcm@openssh.com,aes256-gcm@openssh.com

ok djm@ dtucker@

OpenBSD-Commit-ID: 1e1996fac0dc8a4b0d0ff58395135848287f6f97
2019-09-08 14:49:04 +10:00
djm@openbsd.org c8bdd2db77 upstream: key conversion should fail for !openssl builds, not fall
through to the key generation code

OpenBSD-Commit-ID: b957436adc43c4941e61d61958a193a708bc83c9
2019-09-08 14:49:04 +10:00
djm@openbsd.org 823f6c37eb upstream: typo in previous
OpenBSD-Commit-ID: 7c3b94110864771a6b80a0d8acaca34037c3c96e
2019-09-08 14:49:04 +10:00
Damien Miller 6a710d3e06 needs time.h for --without-openssl 2019-09-08 14:48:11 +10:00
Damien Miller f61f29afda make unittests pass for no-openssl case 2019-09-08 10:37:17 +10:00
djm@openbsd.org 105e1c9218 upstream: avoid compiling certain files that deeply depend on
libcrypto when WITH_OPENSSL isn't set

OpenBSD-Commit-ID: 569f08445c27124ec7c7f6c0268d844ec56ac061
2019-09-06 17:54:21 +10:00
djm@openbsd.org 670104b923 upstream: fixes for !WITH_OPENSSL compilation; ok dtucker@
OpenBSD-Commit-ID: 7fd68eaa9e0f7482b5d4c7e8d740aed4770a839f
2019-09-06 17:54:21 +10:00
djm@openbsd.org be02d7cbde upstream: lots of things were relying on libcrypto headers to
transitively include various system headers (mostly stdlib.h); include them
explicitly

OpenBSD-Commit-ID: 5b522f4f2d844f78bf1cc4f3f4cc392e177b2080
2019-09-06 17:54:21 +10:00
djm@openbsd.org d05aaaaadc upstream: remove leakmalloc reference; we used this early when
refactoring but not since

OpenBSD-Commit-ID: bb28ebda8f7c490b87b37954044a6cdd43a7eb2c
2019-09-06 16:06:22 +10:00
dtucker@openbsd.org 1268f0bcd8 upstream: Check for RSA support before using it for the user key,
otherwise use ed25519 which is supported when built without OpenSSL.

OpenBSD-Regress-ID: 3d23ddfe83c5062f00ac845d463f19a2ec78c0f7
2019-09-06 14:37:23 +10:00
Darren Tucker fd7a2dec65 Provide explicit path to configure-check.
On some platforms (at least OpenBSD) make won't search VPATH for target
files, so building out-of-tree will fail at configure-check.  Provide
explicit path.  ok djm@
2019-09-06 14:09:41 +10:00
djm@openbsd.org 00865c2969 upstream: better error code for bad arguments; inspired by
OpenBSD-Commit-ID: dfc263b6041de7f0ed921a1de0b81ddebfab1e0a
2019-09-06 12:01:45 +10:00
Damien Miller afdf27f5ac revert config.h/config.h.in freshness checks
turns out autoreconf and configure don't touch some files if their content
doesn't change, so the mtime can't be relied upon in a makefile rule
2019-09-05 21:38:40 +10:00