]>
git.ipfire.org Git - thirdparty/openssh-portable.git/log
djm@openbsd.org [Tue, 24 Jun 2025 09:22:03 +0000 (09:22 +0000)]
upstream: make "Match !final" not trigger a 2nd pass ssh_config
parsing pass (unless hostname canonicalisation or a separate "Match final"
does). bz3843
ok dtucker@
OpenBSD-Commit-ID:
ce82b6034828888f0f3f1c812e08f5e87400d802
djm@openbsd.org [Thu, 19 Jun 2025 05:49:05 +0000 (05:49 +0000)]
upstream: better debug diagnostics when loading keys. Will now list
key fingerprint and algorithm (not just algorithm number) as well as making
it explicit which keys didn't load.
OpenBSD-Commit-ID:
ee3e77a0271ab502e653922c6d161b1e091f8fee
djm@openbsd.org [Tue, 17 Jun 2025 01:20:17 +0000 (01:20 +0000)]
upstream: whitespace
OpenBSD-Commit-ID:
6e96814bcf70d0edbb0749ec61cc4fd8707f286d
djm@openbsd.org [Tue, 17 Jun 2025 01:19:27 +0000 (01:19 +0000)]
upstream: fix leak on error path; Coverity CID 481976
OpenBSD-Commit-ID:
963dba2c804e2fd8efea2256092899874d0dbc7b
Darren Tucker [Tue, 17 Jun 2025 11:46:37 +0000 (21:46 +1000)]
Update obsd tests to use current images.
Damien Miller [Mon, 16 Jun 2025 23:48:47 +0000 (09:48 +1000)]
add sshd-auth to RPM spec files
dtucker@openbsd.org [Mon, 16 Jun 2025 09:09:42 +0000 (09:09 +0000)]
upstream: Limit each moduli size to a max of 100 entries.
OpenBSD-Commit-ID:
747219d54565030ff7c45298b9f5e971801f6cb2
dtucker@openbsd.org [Mon, 16 Jun 2025 09:07:08 +0000 (09:07 +0000)]
upstream: Now that ssh-keygen defaults to the maximum memory for
moduli generation we no longer need to run it twice to get enough. Use mkdir
-p instead of a conditional, which allows "make -jN" to work without error.
OpenBSD-Commit-ID:
c2eb57285424f819f9520fa33e0d6d3c4a361a5e
dtucker@openbsd.org [Mon, 16 Jun 2025 09:02:19 +0000 (09:02 +0000)]
upstream: Fix overflow check in sshbuf_dup_string. It's already
constrained by SSHBUF_SIZE_MAX, but still worth fixing the check. Patch from
afonot via github PR#573, with & ok djm@
OpenBSD-Commit-ID:
438888498e66472fc6a48133196d6538d27bff18
dtucker@openbsd.org [Mon, 16 Jun 2025 08:53:04 +0000 (08:53 +0000)]
upstream: Plug mem leak. Patch from afonot via github PR#574, ok djm@
OpenBSD-Commit-ID:
65619f14ef206028ce39bc31f704b832a0609688
dtucker@openbsd.org [Mon, 16 Jun 2025 08:49:27 +0000 (08:49 +0000)]
upstream: Save return value from sshbuf_len instead of calling it
multiple times. Fixes Coverity CID 470521.
OpenBSD-Regress-ID:
356b8b43c8a232deaf445c1ff7526577b177a8e9
dtucker@openbsd.org [Fri, 13 Jun 2025 07:35:14 +0000 (07:35 +0000)]
upstream: Plug leak. Coverity CID 405058.
OpenBSD-Regress-ID:
7fb2fce68d2cb063cdb94d5d66f84fa3a2902792
dtucker@openbsd.org [Fri, 13 Jun 2025 07:23:07 +0000 (07:23 +0000)]
upstream: Remove dead code flagged by Coverity CID 307783. ok djm@
OpenBSD-Regress-ID:
e579f5ec2fd2eb2fe2bad654d16f2ba655a3e035
dtucker@openbsd.org [Thu, 12 Jun 2025 10:09:39 +0000 (10:09 +0000)]
upstream: Set user, host and path to NULL immediately before calling
parse_user_host_path in tests. This ensures that we don't accidentally use
the previous value if the function under test doesn't set them Also fixes
Coverity CIDs 405056 405065 405066.
OpenBSD-Regress-ID:
43678ff59001712f32214fe303b1c21c163c2960
dtucker@openbsd.org [Thu, 12 Jun 2025 09:26:57 +0000 (09:26 +0000)]
upstream: Plug mem leak on error path here too.
Coverity CID 307781.
OpenBSD-Regress-ID:
18e053d9b661fbb4227d3db03172077c1216bb2e
dtucker@openbsd.org [Thu, 12 Jun 2025 09:19:43 +0000 (09:19 +0000)]
upstream: Plug mem leak on error path.
Coverity CID 307776.
OpenBSD-Regress-ID:
c44246690973e1b8643e51079a2faa7ace26490c
dtucker@openbsd.org [Wed, 11 Jun 2025 13:27:11 +0000 (13:27 +0000)]
upstream: Remove dead code ternary. We always report at least
KB/s, so B/s is never used. Coverity CID 291809, ok djm@
OpenBSD-Commit-ID:
a67c5bcc9e19c8965bfeace0e337b13660efa058
dtucker@openbsd.org [Wed, 11 Jun 2025 13:24:05 +0000 (13:24 +0000)]
upstream: Improve termination condition of while loop to compare
size_t's. Assuming read() does what it's supposed to this shouldn't matter,
but should be more robust. Flagged by Coverity CID 470514, ok djm@
OpenBSD-Commit-ID:
d7b5ad60feb797b3464964b9ea67fd78fb9d6cc6
Darren Tucker [Tue, 10 Jun 2025 08:40:56 +0000 (18:40 +1000)]
Replace Windows 2019 runners with 2025 ones.
The windows-2019 runners are being decomissioned.
Darren Tucker [Wed, 4 Jun 2025 08:33:52 +0000 (18:33 +1000)]
Disable _FORTIFY_SOURCE during snprintf test.
Prevents mistakenly detecting snprintf as broken on FreeBSD 15 with
_FORTIFY_SOURCE enabled. bz#3809, patch from jlduran at gmail.com
dtucker@openbsd.org [Mon, 2 Jun 2025 14:09:34 +0000 (14:09 +0000)]
upstream: Fix x11_channel_used_recently() to return true when channel
has been used within the last second, instead of more than a second ago.
Should fix ~5s delay on X client startup when ObscureKeystrokeTiming is
enabled. bz#3820, ok (& sigh) djm@
OpenBSD-Commit-ID:
b741011e81fb3e3d42711d9bd3ed8a959924dee4
dtucker@openbsd.org [Thu, 29 May 2025 13:27:27 +0000 (13:27 +0000)]
upstream: When there's more than one x11 channel in use, return
lastused of most recently used x11 channel instead of the last one found. ok
djm@
OpenBSD-Commit-ID:
94a72bf988d40a5bae2e38608f4e117f712569fe
djm@openbsd.org [Sat, 24 May 2025 11:41:51 +0000 (11:41 +0000)]
upstream: replace xmalloc+memset(0) with xcalloc(); from AZero13 via
GHPR417
OpenBSD-Commit-ID:
921079436a4900325d22bd3b6a90c8d0d54f62f8
djm@openbsd.org [Sat, 24 May 2025 09:46:16 +0000 (09:46 +0000)]
upstream: fix punctuation around host key fingerprints to make them
easier to copy and paste.
Patch from Till Maas via GHPR556; ok dtucker@
OpenBSD-Commit-ID:
c0100182a30b6925c8cdb2225b18140264594b7b
dtucker@openbsd.org [Sat, 24 May 2025 08:13:29 +0000 (08:13 +0000)]
upstream: Replace strncmp + byte count with strprefix in Penalty
config parsing. ok kn@, djm@
OpenBSD-Commit-ID:
34a41bb1b9ba37fb6c7eb29a7ea909547bf02a5a
dtucker@openbsd.org [Sat, 24 May 2025 08:09:32 +0000 (08:09 +0000)]
upstream: Make the display number check relative to
X11DisplayOffset.
This will allows people to use X11DisplayOffset to configure much higher
port ranges if they really want, while not changing the default behaviour.
Patch from Roman Gubarev via github PR#559, ok djm@
OpenBSD-Commit-ID:
e0926af5dc0c11e364452b624c3ad0cda88550b5
Darren Tucker [Sat, 24 May 2025 07:20:57 +0000 (17:20 +1000)]
Remove progressmeter.o from libssh.a.
It's now explicitly included by the binaries that need it (scp & sftp).
bz#3810, patch from jlduran at gmail.com
dtucker@openbsd.org [Sat, 24 May 2025 06:50:28 +0000 (06:50 +0000)]
upstream: Null out keys between test runs.
BENCH_START and BENCH_FINISH are actually a while() loop in disguise,
so if sshkey_generate does not reset the key pointer on failure the test
may incorrectly pass. It also confuses Coverity (CID 551234).
OpenBSD-Regress-ID:
bf4d32079fc6df6dce1f26c2025f4ed492f13936
djm@openbsd.org [Sat, 24 May 2025 04:41:12 +0000 (04:41 +0000)]
upstream: add some verbosity
OpenBSD-Regress-ID:
11c86cda4435b5f9ab6172c4742b95899666c977
djm@openbsd.org [Sat, 24 May 2025 04:41:03 +0000 (04:41 +0000)]
upstream: use start_ssh_agent() to ensure we get logging
add some verbosity
OpenBSD-Regress-ID:
a89bf64696b9fb1b91be318e6b8940c9ab21c616
djm@openbsd.org [Sat, 24 May 2025 04:40:37 +0000 (04:40 +0000)]
upstream: add a start_ssh_agent() function that sets up an agent
with logging
OpenBSD-Regress-ID:
7f9f30f9c64acbd4b418a5e1a19140cc988071a8
dtucker@openbsd.org [Sat, 24 May 2025 06:43:37 +0000 (06:43 +0000)]
upstream: Plug leak of startup_pollfd in debug and child paths.
Coverity CID 405024, ok djm@
OpenBSD-Commit-ID:
db46047229253e9c4470c8bbf5f82706ac021377
Darren Tucker [Sat, 24 May 2025 07:11:38 +0000 (17:11 +1000)]
ssh-keygen changes were fixup'ed into single commit.
Marco Trevisan (Treviño) [Mon, 30 Sep 2024 11:14:11 +0000 (13:14 +0200)]
auth-pam: Check the user didn't change during PAM transaction
PAM modules can change the user during their execution, in such case ssh
would still use the user that has been provided giving potentially
access to another user with the credentials of another one.
So prevent this to happen, by ensuring that the final PAM user is
matching the one that initiated the transaction.
dtucker@openbsd.org [Sat, 24 May 2025 03:37:40 +0000 (03:37 +0000)]
upstream: Remove ssh-keygen's moduli screen -Omemory option.
This vaguely made sense 20 years ago, but these days you'd be hard
pressed to *find* a machine small enough to not support the maximum
(127MB), and no one is screening moduli on such machines anyway,
so just use the max. This also fixes Coverity CID 470522 by deleting
code in question. "kill it with fire" djm@.
OpenBSD-Commit-ID:
39036aa406a99f0a91923aa3a96afff1205558e6
dtucker@openbsd.org [Sat, 24 May 2025 02:33:33 +0000 (02:33 +0000)]
upstream: Fix compile error on 32bit platforms.
Spotted by & ok tb@
OpenBSD-Commit-ID:
cbcf518247886f3c7518fc54cb3bd911ffc69db7
dtucker@openbsd.org [Sat, 24 May 2025 02:01:28 +0000 (02:01 +0000)]
upstream: Use pointer from strprefix in error message,
missed in previous.
OpenBSD-Commit-ID:
d2cdec6cf0fcd4b0ee25e4e3fad8bc8cf0ee657d
dtucker@openbsd.org [Fri, 23 May 2025 12:52:45 +0000 (12:52 +0000)]
upstream: Replace strncmp and strncasecmp with hand-counting bytes
with strprefix. nits lucas@, ok lucas@ djm@
OpenBSD-Commit-ID:
f0888807f151ea2bdaf6fed36303ae81f259d1d4
dtucker@openbsd.org [Fri, 23 May 2025 11:54:50 +0000 (11:54 +0000)]
upstream: Include stdint.h for UINT32_MAX.
OpenBSD-Commit-ID:
edc29ed67e8bd03bac729d9b4849066d1d3a8cb9
dtucker@openbsd.org [Fri, 23 May 2025 11:25:35 +0000 (11:25 +0000)]
upstream: Ensure args to nh_update() fit within uint32, which it
should always anyway. Placates Coverity CID 470520. While there, fix the
upstream URL. ok djm@
OpenBSD-Commit-ID:
2478e89fde089a49fa02f9faf6287d35959c9f92
dtucker@openbsd.org [Fri, 23 May 2025 09:26:25 +0000 (09:26 +0000)]
upstream: Don't leak the args list. Coverity CIDs 481569 & 481570,
ok job@ tb@.
OpenBSD-Commit-ID:
becabcd00513d13d1435b68b7ccffa7151b72393
dtucker@openbsd.org [Fri, 23 May 2025 08:40:13 +0000 (08:40 +0000)]
upstream: Explictly set LC_ALL=C on each sort invocation.
Remove it from sshd_config (where it could be overridden by shell startup
scripts, eg on macos-15) causing random test failures. with & ok djm@
OpenBSD-Regress-ID:
ad0a6678964784096e9a9e6d15ead36beed92f18
Darren Tucker [Fri, 23 May 2025 06:39:18 +0000 (16:39 +1000)]
Allow setting LTESTS in repo variables.
Darren Tucker [Fri, 23 May 2025 06:26:20 +0000 (16:26 +1000)]
Rename debugging variable RUN_ONLY_TEST.
to RUN_ONLY_TARGET_CONFIG to make it more obvious what it matches.
Darren Tucker [Fri, 23 May 2025 06:11:48 +0000 (16:11 +1000)]
chown regress logs before uploading.
dtucker@openbsd.org [Fri, 23 May 2025 01:14:35 +0000 (01:14 +0000)]
upstream: Import regenerated moduli.
OpenBSD-Commit-ID:
07e29dc891e29b31e03e2e5493658b4a9ac19431
deraadt@openbsd.org [Fri, 23 May 2025 00:40:45 +0000 (00:40 +0000)]
upstream: use "const char * const" for malloc_options here also
OpenBSD-Commit-ID:
869715b9c7e1dd5b85efd07814e7e53f0286eea2
dtucker@openbsd.org [Thu, 22 May 2025 12:14:19 +0000 (12:14 +0000)]
upstream: Adjust debug message to prevent (unsigned) integer overflow.
Fixes Coverity CID 481110, ok djm@
OpenBSD-Commit-ID:
26178bf3b812707fb498ea85d076cadd1f2eb686
bluhm@openbsd.org [Thu, 22 May 2025 04:34:18 +0000 (04:34 +0000)]
upstream: Fix OpenBSD RCS ID typos. from Andrius V
OpenBSD-Regress-ID:
5c03a2ef5323969fc4978f2eec4f1a25c48c572a
Darren Tucker [Thu, 22 May 2025 12:09:48 +0000 (22:09 +1000)]
Remove debug change accidentally commited.
Fixes Coverity CID 481160.
Darren Tucker [Thu, 22 May 2025 11:16:37 +0000 (21:16 +1000)]
Collect all of regress dir on failure.
This may allow us to sort through its entrails and determine the cause
of some types of failures.
Damien Miller [Thu, 22 May 2025 08:42:44 +0000 (18:42 +1000)]
minimal shims for fstatat(2)/unlinkat(2) in agent
Add some very minimal and task-specific replacements for
fstatat(2) and unlinkat(2) in the ssh-agent socket cleanup
loop, for platforms that lack these functions. ok dtucker@
dtucker@openbsd.org [Thu, 22 May 2025 04:22:03 +0000 (04:22 +0000)]
upstream: Output the current name for PermitRootLogin's
"prohibit-password" in sshd -T instead of its deprecated alias
"without-password". bz#3788, patch from cjwatson at debian.org.
OpenBSD-Commit-ID:
2d5df18d5ad33a9b6c7547ec78a8e6ea13813df9
dtucker@openbsd.org [Thu, 22 May 2025 03:53:46 +0000 (03:53 +0000)]
upstream: Copy arg to be passed to dirname().
POSIX allows dirname() to modify its args and return a pointer into it,
so this prevents an overlapping strlcpy. bz#3819, patch from cjwatson
at debian.org
OpenBSD-Commit-ID:
c32e496e6a1618aba31c8b7a9d4e1376c5ea6aa1
dtucker@openbsd.org [Thu, 22 May 2025 03:41:10 +0000 (03:41 +0000)]
upstream: Add $OpenBSD$ marker for easier syncing.
OpenBSD-Commit-ID:
27ff3e1e2e6610d9981ebe43ae9b783236800035
djm@openbsd.org [Wed, 21 May 2025 12:12:20 +0000 (12:12 +0000)]
upstream: Correct FILES section to mention new default path to
agent sockets. Spotted by / ok jmc@
OpenBSD-Commit-ID:
91d736d78d71a4276c9cbb075b1462bbc3df55a6
Darren Tucker [Thu, 22 May 2025 02:25:35 +0000 (12:25 +1000)]
Fix nc install some more.
Darren Tucker [Thu, 22 May 2025 02:21:11 +0000 (12:21 +1000)]
Fix cvs up of nc.
Darren Tucker [Thu, 22 May 2025 01:34:04 +0000 (11:34 +1000)]
Install nc during upstream test.
This ensures that the installed nc matches the expectations of the
regress tests.
Darren Tucker [Thu, 22 May 2025 00:52:31 +0000 (10:52 +1000)]
Remove 9.7 branch from CI status page.
It's been obsolete long enough that github no longer reports its
status.
Damien Miller [Wed, 21 May 2025 09:14:47 +0000 (19:14 +1000)]
pull a small netcat SOCKS4A fix from upstream
djm@openbsd.org [Wed, 21 May 2025 08:41:52 +0000 (08:41 +0000)]
upstream: test SOCKS4A; ok tb
OpenBSD-Regress-ID:
d880b75280295cd581a86e39bb0996d347f122d2
djm@openbsd.org [Wed, 21 May 2025 08:36:39 +0000 (08:36 +0000)]
upstream: remove log tarballing "it seemed like a good idea at the
time" - dtucker@
ensure that log files have correct perms when running under sudo/doas
ok dtucker@
OpenBSD-Regress-ID:
20588c14b05de9519f85d638b374b66ae0678c89
djm@openbsd.org [Wed, 21 May 2025 06:44:24 +0000 (06:44 +0000)]
upstream: use logit_f("...") instead of logit("func: ...")
OpenBSD-Commit-ID:
c8d49eb39a9abff3cbcaeaf7df9d48468a5a0695
djm@openbsd.org [Wed, 21 May 2025 06:43:48 +0000 (06:43 +0000)]
upstream: function to make a sshbuf from a hex string; useful in
tests
also constify some arguments
OpenBSD-Commit-ID:
00f9c25b256be0efd73f2d8268ff041bc45ffb2c
Damien Miller [Wed, 21 May 2025 08:47:46 +0000 (18:47 +1000)]
merge netcat SOCKS4A support from OpenBSD
Not a full sync of this file as we have diverged substantially
from upstream (it has libtls support, etc.)
Darren Tucker [Tue, 20 May 2025 12:17:02 +0000 (22:17 +1000)]
Include OpenSSL compat shim where needed.
Darren Tucker [Tue, 20 May 2025 09:28:55 +0000 (19:28 +1000)]
Run all tests on Cygwin again.
... now that we've fixed ci-setup on Cygwin.
Darren Tucker [Tue, 20 May 2025 08:48:23 +0000 (18:48 +1000)]
Use USERNAME rather than LOGNAME on Cygwin.
LOGNAME is specified by POSIX, but Windows (or at least, github's
Windows images) don't set it.
Darren Tucker [Tue, 20 May 2025 08:28:52 +0000 (18:28 +1000)]
Add debug output when setting up CI environment.
Darren Tucker [Tue, 20 May 2025 09:16:38 +0000 (19:16 +1000)]
Include openssl compat shims in test.
Fixes tests on platforms using older LibreSSL releases prior to 3.4.
Darren Tucker [Tue, 20 May 2025 08:14:06 +0000 (18:14 +1000)]
Add compat shims for EC_POINT affine_coordinates
LibreSSL <3.4 does not have EC_POINT_[gs]et_affine_coordinates
but does have the now-deprecated _GFp variantes. We still support
LibreSSL back as far as 3.2.x so add a compat shim.
tb@openbsd.org [Mon, 12 May 2025 05:42:02 +0000 (05:42 +0000)]
upstream: Use EC_POINT_[sg]et_affine_coordinates()
It is available in all supported OpenSSL flavors/versions and the _GFp
variants will be removed from LibreSSL.
ok hshoexer jsing
OpenBSD-Regress-ID:
66cf1561e7b6c49002978f2d6720956f33a882f0
tb@openbsd.org [Mon, 12 May 2025 05:41:20 +0000 (05:41 +0000)]
upstream: Use EC_POINT_[sg]et_affine_coordinates()
It is available in all supported OpenSSL flavors/versions and the _GFp
variants will be removed from LibreSSL.
ok hshoexer jsing
OpenBSD-Commit-ID:
ecedca0e1ffa80e0c9ef7c787bc6a972882c596b
djm@openbsd.org [Fri, 9 May 2025 02:42:03 +0000 (02:42 +0000)]
upstream: make the progress-meter code safe against being called
when not initialised; spotted by tb@ feedback/ok tb@ deraadt@
OpenBSD-Commit-ID:
a9fda1ee08a24c62e0981ff6d15ca93b63467038
tedu@openbsd.org [Thu, 8 May 2025 17:32:53 +0000 (17:32 +0000)]
upstream: convert a last quad_t to int64_t. ok deraadt djm
OpenBSD-Commit-ID:
1c9e01ba1a9ccf442a9cdf10f222077f66885f1f
Darren Tucker [Tue, 20 May 2025 05:01:29 +0000 (15:01 +1000)]
Set runner pasword to random string.
The most recent version of the Github ubuntu-latest image sets the
password field to "!" which sshd considers to be a locked account,
breaking most of the tests.
Darren Tucker [Sun, 11 May 2025 12:54:13 +0000 (22:54 +1000)]
Debug log for why an account is considered locked.
Darren Tucker [Sun, 11 May 2025 06:35:31 +0000 (16:35 +1000)]
Move debug log output into separate workflow step.
Should reduce the need to scroll back to find out which test actually
failed.
Darren Tucker [Sat, 10 May 2025 11:48:06 +0000 (21:48 +1000)]
Skip sftp-perm on Cygwin too.
Darren Tucker [Sat, 10 May 2025 00:23:30 +0000 (10:23 +1000)]
Remove CYGWIN binmode as it's now obsolete.
Darren Tucker [Fri, 9 May 2025 23:25:18 +0000 (09:25 +1000)]
Also skip sftp-cmds test on Cygwin.
Fails at the hardlink step.
Darren Tucker [Fri, 9 May 2025 22:52:11 +0000 (08:52 +1000)]
Tell Cygwin to use native symlinks.
Darren Tucker [Fri, 9 May 2025 22:26:37 +0000 (08:26 +1000)]
Skip keygen-knownhost test on Cygwin.
It fails but at this time it's not clear why.
Darren Tucker [Fri, 9 May 2025 21:59:44 +0000 (07:59 +1000)]
Pass Cygwin setup location to CI setup.
(instead of hard coding it, wrongly).
Darren Tucker [Fri, 9 May 2025 20:37:24 +0000 (06:37 +1000)]
Add RUN_ONLY_TEST to limit which tests are run.
For testing, you can set the repo variable RUN_ONLY_TEST in your repo
(Repo -> Settings -> Security -> Actions -> Variables) to run only that test.
Darren Tucker [Fri, 9 May 2025 09:32:06 +0000 (19:32 +1000)]
Move misc-agent.o to LIBSSH_OBJS.
It's needed by the fuzzer.
Darren Tucker [Fri, 9 May 2025 09:08:36 +0000 (19:08 +1000)]
Put PRIV_ECDSA back, it's still used.
Should fix oss-fuzz test.
Darren Tucker [Thu, 8 May 2025 08:56:39 +0000 (18:56 +1000)]
Since it's unused, make dirfd() take void *.
Some platforms (eg Old BSDs) in some configurations define DIR to "void
*", which causes compile errors in the no-op implementation.
Darren Tucker [Thu, 8 May 2025 01:38:24 +0000 (11:38 +1000)]
Add no-op implmentation of dirfd().
Fixes build on pre-POSIX.1 2008 systems.
Daniel Kahn Gillmor [Wed, 16 Apr 2025 00:18:34 +0000 (10:18 +1000)]
ssh-agent: exit 0 from SIGTERM under systemd socket-activation
When the ssh-agent service is configured to be launched under systemd
socket-activation, the user can inspect the status of the agent with
something like:
systemctl --user status ssh-agent.service
If the user does:
systemctl --user stop ssh-agent.service
it causes the `systemd --user` supervisor to send a SIGTERM to the
agent, which terminates while leaving the systemd-managed socket in
place. That's good, and as expected. (If the user wants to close the
socket, they can do "systemctl --user stop ssh-agent.socket" instead)
But because ssh-agent exits with code 2 in response to a SIGTERM, the
supervisor marks the service as "failed", even though the state of the
supervised service is exactly the same as during session startup (not
running, ready to launch when a client connects to the socket).
This change makes ssh-agent exit cleanly (code 0) in response to a
SIGTERM when launched under socket activation. This aligns the systemd
supervisor's understanding of the state of supervised ssh-agent with
reality.
Signed-off-by: Daniel Kahn Gillmor <dkg@fifthhorseman.net>
Darren Tucker [Wed, 7 May 2025 11:05:06 +0000 (21:05 +1000)]
Skip d_type check on platforms that don't have it.
On those, the subsequent stat() should catch the sockets.
dtucker@openbsd.org [Wed, 7 May 2025 10:44:26 +0000 (10:44 +0000)]
upstream: Rename sockaddr_un sun -> sunaddr.
This makes things easier in -portable, where on Solaris an derivatives
"sun" is defined to "1", causing compilation errors. ok deraadt@.
OpenBSD-Commit-ID:
0669043afb49856b57b382f0489221bd98305d3b
djm@openbsd.org [Tue, 6 May 2025 06:05:48 +0000 (06:05 +0000)]
upstream: remove DSA from the regression/unit test suite too.
OpenBSD-Regress-ID:
4424d2eaf0bce3887318ef6d18de6c06f3617d6e
djm@openbsd.org [Tue, 15 Apr 2025 05:31:24 +0000 (05:31 +0000)]
upstream: another missing ifdef
OpenBSD-Regress-ID:
4f71f8f122eac4cbf7f1d2088a9be45317dd3e4a
djm@openbsd.org [Tue, 15 Apr 2025 05:00:13 +0000 (05:00 +0000)]
upstream: missing ifdef
OpenBSD-Regress-ID:
7260fb672de5738c17dec06c71a5be0186bb2b09
djm@openbsd.org [Wed, 7 May 2025 04:10:21 +0000 (04:10 +0000)]
upstream: memory leak on error path; bz3821
OpenBSD-Commit-ID:
65577596a15ad6dd9a1ab3fc24c1c31303ee6e2b
deraadt@openbsd.org [Tue, 6 May 2025 15:15:05 +0000 (15:15 +0000)]
upstream: test ssh-agent with the -T flag to force the old /tmp
location rather than inside the homedir. During relink operation,
/.ssh/agent was created which is surprising. This test sequence could use
some improvement so this is a temporary fix. observed by florian, change ok
semarie
OpenBSD-Commit-ID:
c7246a6b519ac390ca550719f91acfdaef1fa0f0
djm@openbsd.org [Tue, 6 May 2025 05:40:56 +0000 (05:40 +0000)]
upstream: finally remove DSA signature support from OpenSSH.
feedback/ok tb@, ok deraadt@
OpenBSD-Commit-ID:
bfe6ee73c1b676c81a2901030c791f8ec888228f
djm@openbsd.org [Mon, 5 May 2025 05:51:11 +0000 (05:51 +0000)]
upstream: Now that there's an I-D for certificate keys, refer to
that instead of the much more basic format description we had previously.
OpenBSD-Commit-ID:
cf01e0727a813fee8626ad7b3aa240621cc92014