Damien Miller [Wed, 17 Aug 2022 06:04:16 +0000 (16:04 +1000)]
on Cygwin, prefer WinHello FIDO device
If no FIDO device was explictly specified, then prefer the
windows://hello FIDO device. An exception to this is when
probing resident FIDO keys, in which case hardware FIDO
devices are preferred.
djm@openbsd.org [Fri, 12 Aug 2022 05:20:28 +0000 (05:20 +0000)]
upstream: sftp-server: support home-directory request
Add support to the sftp-server for the home-directory extension defined
in draft-ietf-secsh-filexfer-extensions-00. This overlaps a bit with the
existing expand-path@openssh.com, but uses a more official protocol name,
and so is a bit more likely to be implemented by non-OpenSSH clients.
Darren Tucker [Thu, 11 Aug 2022 03:33:51 +0000 (13:33 +1000)]
Skip hostbased during Valgrind tests.
Valgrind doesn't let ssh exec ssh-keysign (because it's setuid) so skip
it during the Valgrind based tests.
See https://bugs.kde.org/show_bug.cgi?id=119404 for a discussion of this
(ironically there the problematic binary was ssh(1) back when it could
still be setuid).
times and authorized_keys expiry-time options to accept dates in the UTC time
zone in addition to the default of interpreting them in the system time zone.
YYYYMMDD and YYMMDDHHMM[SS] dates/times will be interpreted as UTC if
suffixed with a 'Z' character.
Also allow certificate validity intervals to be specified in raw
seconds-since-epoch as hex value, e.g. -V 0x1234:0x4567890. This
is intended for use by regress tests and other tools that call
ssh-keygen as part of a CA workflow.
Darren Tucker [Wed, 10 Aug 2022 07:25:24 +0000 (17:25 +1000)]
Rename our getentropy to prevent possible loops.
Since arc4random seeds from getentropy, and we use OpenSSL for that
if enabled, there's the possibility that if we build on a system that
does not have getentropy then run on a system that does have it, then
OpenSSL could end up calling our getentropy and getting stuck in a loop.
Pointed out by deraadt@, ok djm@
Corinna Vinschen [Fri, 11 Feb 2022 13:33:41 +0000 (14:33 +0100)]
sk_sign: set FIDO2 uv attribute explicitely for WinHello
WinHello via libfido2 performs user verification by default.
However, if we stick to that, there's no way to differentiate
between keys created with or without "-O verify-required".
Set FIDO2 uv attribute explicitely to FIDO_OPT_FALSE, then check
if user verification has been requested.
Darren Tucker [Fri, 5 Aug 2022 03:12:27 +0000 (13:12 +1000)]
Factor out getrnd() and rename to getentropy().
Factor out the arc4random seeding into its own file and change the
interface to match getentropy. Use native getentropy if available.
This will make it easier to resync OpenBSD changes to arc4random.
Prompted by bz#3467, ok djm@.
Darren Tucker [Wed, 27 Jul 2022 08:31:14 +0000 (18:31 +1000)]
Move stale-configure check as early as possible.
We added a check in Makefile to catch the case where configure needs to
be rebuilt, however this did not happen until a build was attempted in
which case all of the work done by configure was wasted. Move this check
to the start of configure to catch it as early as possible. ok djm@
Darren Tucker [Thu, 21 Jul 2022 23:24:45 +0000 (09:24 +1000)]
Do not link scp, sftp and sftp-server w/ zlib.
Some of our binaries (eg sftp, sftp-server, scp) do not interact with
the channels code and thus do use libraries such as zlib and libcrypto
although they are linked with them. This adds a CHANNELLIBS and starts
by moving zlib into it, which means the aformentioned binaries are no
longer linked against zlib. ok djm@
Darren Tucker [Mon, 25 Jul 2022 11:49:04 +0000 (21:49 +1000)]
Remove workarounds for OpenSSL missing AES-CTR.
We have some compatibility hacks that were added to support OpenSSL
versions that do not support AES CTR mode. Since that time, however,
the minimum OpenSSL version that we support has moved to 1.0.1 which
*does* have CTR, so this is no longer needed. ok djm@
Darren Tucker [Mon, 25 Jul 2022 11:43:00 +0000 (21:43 +1000)]
Remove workarounds for OpenSSL missing AES-GCM.
We have some compatibility hacks that were added to support OpenSSL
versions that do not support AES GCM mode. Since that time, however,
the minimum OpenSSL version that we support has moved to 1.0.1 which
*does* have GCM, so this is no longer needed. ok djm@
Darren Tucker [Sat, 23 Jul 2022 04:38:22 +0000 (14:38 +1000)]
Convert "have_prog" function into "which".
"which" and its behaviour is not standardized, so convert the existing
have_prog function into "which" so we can rely on it being available
and what its semantics are. Add a have_prog wrapper that maintains the
existing behaviour.
upstream: when enrolling a resident key on a security token, check
if a credential with matching application and user ID strings already exists.
if so, prompt the user for confirmation before overwriting the credential.
patch from Pedro Martelletto via GHPR329
NB. cranks SSH_SK_VERSION_MAJOR, so any third-party FIDO middleware
implementations will need to adjust
Darren Tucker [Fri, 15 Jul 2022 11:31:48 +0000 (21:31 +1000)]
Move vmshutdown to first step.
If a previous run on a physical runner has failed to clean up, the next
run will fail because it'll try to check out the code to a broken
directory mount. Make cleanup the first step.
Darren Tucker [Thu, 14 Jul 2022 01:22:08 +0000 (11:22 +1000)]
Return ERANGE from getcwd() if buffer size is 1.
If getcwd() is supplied a buffer size of exactly 1 and a path of "/", it
could result in a nul byte being written out of array bounds. POSIX says
it should return ERANGE if the path will not fit in the available buffer
(with terminating nul). 1 byte cannot fit any possible path with its nul,
so immediately return ERANGE in that case.
OpenSSH never uses getcwd() with this buffer size, and all current
(and even quite old) platforms that we are currently known to work
on have a native getcwd() so this code is not used on those anyway.
Reported by Qualys, ok djm@
Darren Tucker [Thu, 14 Jul 2022 00:02:35 +0000 (10:02 +1000)]
Split README.platform into its own line.
README.platform has general platform-specific information, having it
following text about FIDO2 on the same line could imply that it only
has information about FIDO2.
Darren Tucker [Wed, 13 Jul 2022 23:56:01 +0000 (09:56 +1000)]
Clarify README.md text.
Clarify the text about the implications of building without OpenSSL, and
prefix the "configure --help" example command with a "./" so it's likely
to work as-is in more shells. From bz#3461.
Darren Tucker [Tue, 12 Jul 2022 09:48:44 +0000 (19:48 +1000)]
Remove special casing of crypt().
Configure goes to some lengths to pick crypt() from either libcrypt
or OpenSSL's libcrypto because they can more or less featureful (eg
supporting md5-style passwords).
OpenSSL removed its crypt() interface in 2002:
https://github.com/openssl/openssl/commit/69deec58 so these hijinks
should no longer be necessary. This also only links sshd with libcrypt
which is the only thing that needs it. ok djm@
Darren Tucker [Sun, 3 Jul 2022 07:54:49 +0000 (17:54 +1000)]
Move checks for pollfd.fd and nfds_t.
Move the checks for struct pollfd.fd and nfds_t to before the sandboxing
checks. This groups all the sandbox checks together so we can skip them
all when sandboxing is disabled.
djm@openbsd.org [Fri, 3 Jun 2022 04:30:46 +0000 (04:30 +0000)]
upstream: Make SetEnv directives first-match-wins in both
sshd_config and sshd_config; previously if the same name was reused then the
last would win (which is the opposite to how the config is supposed to work).
While there, make the ssh_config parsing more like sshd_config.
djm@openbsd.org [Fri, 27 May 2022 05:02:46 +0000 (05:02 +0000)]
upstream: split the low-level file handling functions out from
auth2-pubkey.c
Put them in a new auth2-pubkeyfile.c to make it easier to refer to them
(e.g. in unit/fuzz tests) without having to refer to everything else
pubkey auth brings in.
remove "struct ssh *" from arguments - this was only used to pass the
remote host/address. These can be passed in instead and the resulting
code is less tightly coupled to ssh_api.[ch]