]> git.ipfire.org Git - thirdparty/openssh-portable.git/log
thirdparty/openssh-portable.git
4 years agoupstream: Allow argv_split() to optionally terminate tokenisation
djm@openbsd.org [Tue, 8 Jun 2021 06:54:40 +0000 (06:54 +0000)] 
upstream: Allow argv_split() to optionally terminate tokenisation

when it encounters an unquoted comment.

Add some additional utility function for working with argument
vectors, since we'll be switching to using them to parse
ssh/sshd_config shortly.

ok markus@ as part of a larger diff; tested in snaps

OpenBSD-Commit-ID: fd9c108cef2f713f24e3bc5848861d221bb3a1ac

4 years agoSave logs on failure for upstream test
Darren Tucker [Mon, 7 Jun 2021 09:19:23 +0000 (19:19 +1000)] 
Save logs on failure for upstream test

4 years agoAdd obsdsnap-i386 upstream test target.
Darren Tucker [Mon, 7 Jun 2021 04:36:32 +0000 (14:36 +1000)] 
Add obsdsnap-i386 upstream test target.

4 years agoupstream: fix debug message when finding a private key to match a
djm@openbsd.org [Mon, 7 Jun 2021 03:38:38 +0000 (03:38 +0000)] 
upstream: fix debug message when finding a private key to match a

certificate being attempted for user authentication. Previously it would
print the certificate's path, whereas it was supposed to be showing the
private key's path. Patch from Alex Sherwin via GHPR247

OpenBSD-Commit-ID: d5af3be66d0f22c371dc1fe6195e774a18b2327b

4 years agoupstream: Match host certificates against host public keys, not private
djm@openbsd.org [Sun, 6 Jun 2021 11:34:16 +0000 (11:34 +0000)] 
upstream: Match host certificates against host public keys, not private

keys. Allows use of certificates with private keys held in a ssh-agent.
Reported by Miles Zhou in bz3524; ok dtucker@

OpenBSD-Commit-ID: 25f5bf70003126d19162862d9eb380bf34bac22a

4 years agoupstream: Client-side workaround for a bug in OpenSSH 7.4: this release
djm@openbsd.org [Sun, 6 Jun 2021 03:40:39 +0000 (03:40 +0000)] 
upstream: Client-side workaround for a bug in OpenSSH 7.4: this release

allows RSA/SHA2 signatures for public key authentication but fails to
advertise this correctly via SSH2_MSG_EXT_INFO. This causes clients of these
server to incorrectly match PubkeyAcceptedAlgorithms and potentially refuse
to offer valid keys.

Reported by and based on patch from Gordon Messmer via bz3213, thanks
also for additional analysis by Jakub Jelen. ok dtucker

OpenBSD-Commit-ID: d6d0b7351d5d44c45f3daaa26efac65847a564f7

4 years agoupstream: degrade gracefully if a sftp-server offers the
djm@openbsd.org [Sun, 6 Jun 2021 03:17:02 +0000 (03:17 +0000)] 
upstream: degrade gracefully if a sftp-server offers the

limits@openssh.com extension but fails when the client tries to invoke it.
Reported by Hector Martin via bz3318

OpenBSD-Commit-ID: bd9d1839c41811616ede4da467e25746fcd9b967

4 years agoupstream: the limits@openssh.com extension was incorrectly marked
djm@openbsd.org [Sun, 6 Jun 2021 03:15:39 +0000 (03:15 +0000)] 
upstream: the limits@openssh.com extension was incorrectly marked

as an operation that writes to the filesystem, which made it unavailable in
sftp-server read-only mode. Spotted by Hector Martin via bz3318

OpenBSD-Commit-ID: f054465230787e37516c4b57098fc7975e00f067

4 years agoupstream: PROTOCOL.certkeys: update reference from IETF draft to
naddy@openbsd.org [Sat, 5 Jun 2021 13:47:00 +0000 (13:47 +0000)] 
upstream: PROTOCOL.certkeys: update reference from IETF draft to

RFC

Also fix some typos.
ok djm@

OpenBSD-Commit-ID: 5e855b6c5a22b5b13f8ffa3897a868e40d349b44

4 years agoClear notify_pipe from readset if present.
Darren Tucker [Fri, 4 Jun 2021 13:41:29 +0000 (23:41 +1000)] 
Clear notify_pipe from readset if present.

Prevents leaking an implementation detail to the caller.

4 years agospace->tabs.
Darren Tucker [Fri, 4 Jun 2021 13:24:25 +0000 (23:24 +1000)] 
space->tabs.

4 years agoAdd pselect implementation for platforms without.
Darren Tucker [Fri, 4 Jun 2021 08:39:48 +0000 (18:39 +1000)] 
Add pselect implementation for platforms without.

This is basically the existing notify_pipe kludge from serverloop.c
moved behind a pselect interface.  It works by installing a signal
handler that writes to a pipe that the select is watching, then calls
the original handler.

The select call in serverloop will become pselect soon, at which point the
kludge will be removed from thereand will only exist in the compat layer.
Original code by markus, help from djm.

4 years agoauth_log: dont log partial successes as failures
Vincent Brillault [Sun, 24 May 2020 07:15:06 +0000 (09:15 +0200)] 
auth_log: dont log partial successes as failures

By design, 'partial' logins are successful logins, so initially with
authenticated set to 1, for which another authentication is required. As
a result, authenticated is always reset to 0 when partial is set to 1.
However, even if authenticated is 0, those are not failed login
attempts, similarly to attempts with authctxt->postponed set to 1.

4 years agoupstream: The RB_GENERATE_STATIC(3) macro expands to a series of
djm@openbsd.org [Fri, 4 Jun 2021 06:19:07 +0000 (06:19 +0000)] 
upstream: The RB_GENERATE_STATIC(3) macro expands to a series of

function definitions and not a statement, so there should be no semicolon
following them. Patch from Michael Forney

OpenBSD-Commit-ID: c975dd180580f0bdc0a4d5b7d41ab1f5e9b7bedd

4 years agoupstream: rework authorized_keys example section, removing irrelevant
djm@openbsd.org [Fri, 4 Jun 2021 05:59:18 +0000 (05:59 +0000)] 
upstream: rework authorized_keys example section, removing irrelevant

stuff, de-wrapping the example lines and better aligning the examples with
common usage and FAQs; ok jmc

OpenBSD-Commit-ID: d59f1c9281f828148e2a2e49eb9629266803b75c

4 years agoupstream: adjust SetEnv description to clarify $TERM handling
djm@openbsd.org [Fri, 4 Jun 2021 05:10:03 +0000 (05:10 +0000)] 
upstream: adjust SetEnv description to clarify $TERM handling

OpenBSD-Commit-ID: 8b8cc0124856bc1094949d55615e5c44390bcb22

4 years agoupstream: Switch the listening select loop from select() to
dtucker@openbsd.org [Fri, 4 Jun 2021 05:09:08 +0000 (05:09 +0000)] 
upstream: Switch the listening select loop from select() to

pselect() and mask signals while checking signal flags, umasking for pselect
and restoring afterwards. Also restore signals before sighup_restart so they
don't remain blocked after restart.

This prevents a race where a SIGTERM or SIGHUP can arrive between
checking the flag and calling select (eg if sshd is processing a
new connection) resulting in sshd not shutting down until the next
time it receives a new connection.  bz#2158, with & ok djm@

OpenBSD-Commit-ID: bf85bf880fd78e00d7478657644fcda97b9a936f

4 years agoupstream: allow ssh_config SetEnv to override $TERM, which is otherwise
djm@openbsd.org [Fri, 4 Jun 2021 05:02:40 +0000 (05:02 +0000)] 
upstream: allow ssh_config SetEnv to override $TERM, which is otherwise

handled specially by the protocol. Useful in ~/.ssh/config to set TERM to
something generic (e.g. "xterm" instead of "xterm-256color") for destinations
that lack terminfo entries. feedback and ok dtucker@

OpenBSD-Commit-ID: 38b1ef4d5bc159c7d9d589d05e3017433e2d5758

4 years agoupstream: correct extension name "no-presence-required" =>
djm@openbsd.org [Fri, 4 Jun 2021 04:02:21 +0000 (04:02 +0000)] 
upstream: correct extension name "no-presence-required" =>

"no-touch-required"

document "verify-required" option

OpenBSD-Commit-ID: 1879ff4062cf61d79b515e433aff0bf49a6c55c5

4 years agoRetire fbsd7 test target.
Darren Tucker [Wed, 2 Jun 2021 03:54:11 +0000 (13:54 +1000)] 
Retire fbsd7 test target.

It's the slowest of the selfhosted targets (since it's 32bit but has
most of the crypto algos). We still have coverage for 32bit i386.

4 years agoCheck for $OPENSSL in md5 fallback too.
Darren Tucker [Wed, 2 Jun 2021 01:21:40 +0000 (11:21 +1000)] 
Check for $OPENSSL in md5 fallback too.

4 years agoAdd dfly60 target.
Darren Tucker [Wed, 2 Jun 2021 01:17:54 +0000 (11:17 +1000)] 
Add dfly60 target.

4 years agoupstream: Merge back shell portability changes
dtucker@openbsd.org [Wed, 2 Jun 2021 00:17:45 +0000 (00:17 +0000)] 
upstream: Merge back shell portability changes

bringing it back in sync with -portable.

OpenBSD-Regress-ID: c07905ba931e66ad7d849b87b7d19648007175d1

4 years agoupstream: Use a default value for $OPENSSL,
dtucker@openbsd.org [Tue, 1 Jun 2021 23:56:20 +0000 (23:56 +0000)] 
upstream: Use a default value for $OPENSSL,

allowing it to be overridden. Do the same in the PuTTY tests since it's
needed there and not exported by test-exec.sh.

OpenBSD-Regress-ID: c49dcd6aa7602a8606b7afa192196ca1fa65de16

4 years agoupstream: Find openssl binary via environment variable. This
dtucker@openbsd.org [Mon, 24 May 2021 10:25:18 +0000 (10:25 +0000)] 
upstream: Find openssl binary via environment variable. This

allows overriding if necessary (eg in -portable where we're testing against a
specific version of OpenSSL).

OpenBSD-Regress-ID: 491f39cae9e762c71aa4bf045803d077139815c5

4 years agoupstream: fix memleak in test
djm@openbsd.org [Fri, 21 May 2021 04:03:47 +0000 (04:03 +0000)] 
upstream: fix memleak in test

OpenBSD-Regress-ID: 5e529d0982aa04666604936df43242e97a7a6f81

4 years agoupstream: also check contents of remaining string
djm@openbsd.org [Fri, 21 May 2021 03:59:01 +0000 (03:59 +0000)] 
upstream: also check contents of remaining string

OpenBSD-Regress-ID: d526fa07253f4eebbc7d6205a0ab3d491ec71a28

4 years agoupstream: unit test for misc.c:strdelim() that mostly servces to
djm@openbsd.org [Fri, 21 May 2021 03:48:07 +0000 (03:48 +0000)] 
upstream: unit test for misc.c:strdelim() that mostly servces to

highlight its inconsistencies

OpenBSD-Regress-ID: 8d2bf970fcc01ccc6e36a5065f89b9c7fa934195

4 years agoPut minix3 config in the host-specific block.
Darren Tucker [Thu, 27 May 2021 11:23:15 +0000 (21:23 +1000)] 
Put minix3 config in the host-specific block.

4 years agoupstream: Hash challenge supplied by client during FIDO key enrollment
djm@openbsd.org [Mon, 31 May 2021 06:48:42 +0000 (06:48 +0000)] 
upstream: Hash challenge supplied by client during FIDO key enrollment

prior to passing it to libfido2, which does expect a hash.

There is no effect for users who are simply generating FIDO keys using
ssh-keygen - by default we generate a random 256 bit challenge, but
people building attestation workflows around our tools should now have
a more consistent experience (esp. fewer failures when they fail to
guess the magic 32-byte challenge length requirement).

ok markus@

OpenBSD-Commit-ID: b8d5363a6a7ca3b23dc28f3ca69470472959f2b5

4 years agoInclude login_cap.h for login_getpwclass override.
Darren Tucker [Thu, 27 May 2021 11:14:15 +0000 (21:14 +1000)] 
Include login_cap.h for login_getpwclass override.

On minix3, login_getpwclass is __RENAME'ed to __login_getpwclass50 so
without this the include overriding login_getpwclass causes a compile
error.

4 years agoAdd minix3 test target.
Darren Tucker [Thu, 27 May 2021 11:13:38 +0000 (21:13 +1000)] 
Add minix3 test target.

4 years agoupstream: fix SEGV in UpdateHostkeys debug() message, triggered
djm@openbsd.org [Wed, 26 May 2021 01:47:24 +0000 (01:47 +0000)] 
upstream: fix SEGV in UpdateHostkeys debug() message, triggered

when the update removed more host keys than remain present. Fix tested by
reporter James Cook, via bugs@

OpenBSD-Commit-ID: 44f641f6ee02bb957f0c1d150495b60cf7b869d3

4 years agoupstream: ssh: The client configuration keyword is
naddy@openbsd.org [Sun, 23 May 2021 18:22:57 +0000 (18:22 +0000)] 
upstream: ssh: The client configuration keyword is

"hostbasedacceptedalgorithms"

This fixes a mistake that slipped in when "HostbasedKeyTypes" was
renamed to "HostbasedAcceptedAlgorithms".

Bug report by zack@philomathiclife.com

OpenBSD-Commit-ID: d745a7e8e50b2589fc56877f322ea204bc784f38

4 years agoRename README.md to ci-status.md.
Darren Tucker [Tue, 25 May 2021 01:40:47 +0000 (11:40 +1000)] 
Rename README.md to ci-status.md.

The original intent was to provide a status page for the CIs configured
in that directory, but it had the side effect of replacing the top-level
README.md.

4 years agoupstream: restore blocking status on stdio fds before close
djm@openbsd.org [Wed, 19 May 2021 01:24:05 +0000 (01:24 +0000)] 
upstream: restore blocking status on stdio fds before close

ssh(1) needs to set file descriptors to non-blocking mode to operate
but it was not restoring the original state on exit. This could cause
problems with fds shared with other programs via the shell, e.g.

> $ cat > test.sh << _EOF
> #!/bin/sh
> {
>         ssh -Fnone -oLogLevel=verbose ::1 hostname
>         cat /usr/share/dict/words
> } | sleep 10
> _EOF
> $ ./test.sh
> Authenticated to ::1 ([::1]:22).
> Transferred: sent 2352, received 2928 bytes, in 0.1 seconds
> Bytes per second: sent 44338.9, received 55197.4
> cat: stdout: Resource temporarily unavailable

This restores the blocking status for fds 0,1,2 (stdio) before ssh(1)
abandons/closes them.

This was reported as bz3280 and GHPR246; ok dtucker@

OpenBSD-Commit-ID: 8cc67346f05aa85a598bddf2383fcfcc3aae61ce

4 years agoupstream: fix breakage of -W forwaring introduced in 1.554; reported by
djm@openbsd.org [Mon, 17 May 2021 11:43:16 +0000 (11:43 +0000)] 
upstream: fix breakage of -W forwaring introduced in 1.554; reported by

naddy@ and sthen@, ok sthen@

OpenBSD-Commit-ID: f72558e643a26dc4150cff6e5097b5502f6c85fd

4 years agoupstream: Regenerate moduli.
dtucker@openbsd.org [Mon, 17 May 2021 07:22:45 +0000 (07:22 +0000)] 
upstream: Regenerate moduli.

OpenBSD-Commit-ID: 83c93a2a07c584c347ac6114d6329b18ce515557

4 years agoHandle Android libc returning NULL pw->pw_passwd
Damien Miller [Sun, 16 May 2021 23:40:23 +0000 (09:40 +1000)] 
Handle Android libc returning NULL pw->pw_passwd

Reported by Luke Dashjr

4 years agoupstream: fix previous: test saved no_shell_flag, not the one that just
djm@openbsd.org [Fri, 14 May 2021 05:20:32 +0000 (05:20 +0000)] 
upstream: fix previous: test saved no_shell_flag, not the one that just

got clobbered

OpenBSD-Commit-ID: b8deace085d9d941b2d02f810243b9c302e5355d

4 years agoupstream: Fix ssh started with ControlPersist incorrectly executing a
djm@openbsd.org [Fri, 14 May 2021 03:09:48 +0000 (03:09 +0000)] 
upstream: Fix ssh started with ControlPersist incorrectly executing a

shell when the -N (no shell) option was specified. bz3290 reported by Richard
Schwab; patch from markus@ ok me

OpenBSD-Commit-ID: ea1ea4af16a95687302f7690bdbe36a6aabf87e1

4 years agoupstream: Clarify language about moduli. While both ends of the
dtucker@openbsd.org [Wed, 12 May 2021 11:34:30 +0000 (11:34 +0000)] 
upstream: Clarify language about moduli. While both ends of the

connection do need to use the same parameters (ie groups), the DH-GEX
protocol takes care of that and both ends do not need the same contents in
the moduli file, which is what the previous text suggested.  ok djm@ jmc@

OpenBSD-Commit-ID: f0c18cc8e79c2fbf537a432a9070ed94e96a622a

4 years agoupstream: include pid in LogVerbose spam
djm@openbsd.org [Fri, 7 May 2021 04:11:51 +0000 (04:11 +0000)] 
upstream: include pid in LogVerbose spam

OpenBSD-Commit-ID: aacb86f96ee90c7cb84ec27452374285f89a7f00

4 years agoupstream: don't sigdie() in signal handler in privsep child process;
djm@openbsd.org [Fri, 7 May 2021 03:09:38 +0000 (03:09 +0000)] 
upstream: don't sigdie() in signal handler in privsep child process;

this can end up causing sandbox violations per bz3286; ok dtucker@

OpenBSD-Commit-ID: a7f40b2141dca4287920da68ede812bff7ccfdda

4 years agoupstream: Increase ConnectionAttempts from 4 to 10 as the tests
dtucker@openbsd.org [Fri, 7 May 2021 09:23:40 +0000 (09:23 +0000)] 
upstream: Increase ConnectionAttempts from 4 to 10 as the tests

occasionally time out on heavily loaded hosts.

OpenBSD-Regress-ID: 29a8cdef354fc9da471a301f7f65184770434f3a

4 years agoupstream: dump out a usable private key string too; inspired by Tyson
djm@openbsd.org [Fri, 7 May 2021 02:26:55 +0000 (02:26 +0000)] 
upstream: dump out a usable private key string too; inspired by Tyson

Whitehead

OpenBSD-Regress-ID: 65572d5333801cb2f650ebc778cbdc955e372058

4 years agoupstream: correct mistake in spec - the private key blobs are encoded
djm@openbsd.org [Fri, 7 May 2021 02:29:40 +0000 (02:29 +0000)] 
upstream: correct mistake in spec - the private key blobs are encoded

verbatim and not as strings (i.e. no 4-byte length header)

OpenBSD-Commit-ID: 3606b5d443d72118c5b76c4af6dd87a5d5a4f837

4 years agoupstream: Don't pass NULL as a string in debugging as it does not work
dtucker@openbsd.org [Tue, 4 May 2021 22:53:52 +0000 (22:53 +0000)] 
upstream: Don't pass NULL as a string in debugging as it does not work

on some platforms in -portable.  ok djm@

OpenBSD-Commit-ID: 937c892c99aa3c9c272a8ed78fa7c2aba3a44fc9

4 years agoupstream: more debugging for UpdateHostKeys signature failures
djm@openbsd.org [Mon, 3 May 2021 00:16:45 +0000 (00:16 +0000)] 
upstream: more debugging for UpdateHostKeys signature failures

OpenBSD-Commit-ID: 1ee95f03875e1725df15d5e4bea3e73493d57d36

4 years agoAdd obsd69 test target.
Darren Tucker [Sat, 1 May 2021 10:07:47 +0000 (20:07 +1000)] 
Add obsd69 test target.

4 years agoupstream: a little debugging in the main mux process for status
djm@openbsd.org [Fri, 30 Apr 2021 04:29:53 +0000 (04:29 +0000)] 
upstream: a little debugging in the main mux process for status

confirmation failures in multiplexed sessions

OpenBSD-Commit-ID: 6e27b87c95176107597035424e1439c3232bcb49

4 years agoupstream: Remove now-unused skey function prototypes leftover from
dtucker@openbsd.org [Fri, 30 Apr 2021 04:02:52 +0000 (04:02 +0000)] 
upstream: Remove now-unused skey function prototypes leftover from

skey removal.

OpenBSD-Commit-ID: 2fc36d519fd37c6f10ce74854c628561555a94c3

4 years agoWrap sntrup761x25519 inside ifdef.
Darren Tucker [Thu, 29 Apr 2021 03:01:50 +0000 (13:01 +1000)] 
Wrap sntrup761x25519 inside ifdef.

From balu.gajjala at gmail.com via bz#3306.

4 years agoAdd status badges for Actions-based tests.
Darren Tucker [Wed, 28 Apr 2021 04:44:07 +0000 (14:44 +1000)] 
Add status badges for Actions-based tests.

4 years agoAdd obsdsnap (OpenBSD snapshot) test target.
Darren Tucker [Wed, 28 Apr 2021 02:22:11 +0000 (12:22 +1000)] 
Add obsdsnap (OpenBSD snapshot) test target.

4 years agoAdd test building upstream OpenBSD source.
Darren Tucker [Wed, 28 Apr 2021 01:35:28 +0000 (11:35 +1000)] 
Add test building upstream OpenBSD source.

4 years agoTest against OpenSSL 1.1.0h instead of 1.1.0g.
Darren Tucker [Tue, 27 Apr 2021 04:22:20 +0000 (14:22 +1000)] 
Test against OpenSSL 1.1.0h instead of 1.1.0g.

1.1.0g requires a perl glob module that's not installed by default.

4 years agoUse the default VM type for libcrypto ver tests.
Darren Tucker [Tue, 27 Apr 2021 02:37:59 +0000 (12:37 +1000)] 
Use the default VM type for libcrypto ver tests.

4 years agoAlways build OpenSSL shared.
Darren Tucker [Tue, 27 Apr 2021 02:24:10 +0000 (12:24 +1000)] 
Always build OpenSSL shared.

This is the default for current versions but we need it to test against
earlier versions.

4 years agoFix custom OpenSSL tests.
Darren Tucker [Mon, 26 Apr 2021 23:18:02 +0000 (09:18 +1000)] 
Fix custom OpenSSL tests.

Check out specified OpenSSL version.  Install custom libcrypto where
configure expects to find it.  Remove unneeded OpenSSL config time
options.  Older OpenSSL versions were not make -j safe so remove it.

4 years agoExport CC and CFLAGS for c89 test.
Darren Tucker [Mon, 26 Apr 2021 07:18:25 +0000 (17:18 +1000)] 
Export CC and CFLAGS for c89 test.

4 years agoAdd c89 here too.
Darren Tucker [Mon, 26 Apr 2021 07:13:44 +0000 (17:13 +1000)] 
Add c89 here too.

4 years agoAdd test against OpenSSL w/out ECC.
Darren Tucker [Mon, 26 Apr 2021 05:34:23 +0000 (15:34 +1000)] 
Add test against OpenSSL w/out ECC.

4 years agoEnsure we can still build with C89.
Darren Tucker [Mon, 26 Apr 2021 04:49:59 +0000 (14:49 +1000)] 
Ensure we can still build with C89.

4 years agoInterop test agains PuTTY.
Darren Tucker [Mon, 26 Apr 2021 04:29:03 +0000 (14:29 +1000)] 
Interop test agains PuTTY.

4 years agoSupport testing against arbitary libcrytpo vers.
Darren Tucker [Mon, 26 Apr 2021 04:02:03 +0000 (14:02 +1000)] 
Support testing against arbitary libcrytpo vers.

Add tests against various LibreSSL and OpenSSL versions.

4 years agoAdd fbsd10 test target.
Darren Tucker [Mon, 26 Apr 2021 03:35:44 +0000 (13:35 +1000)] 
Add fbsd10 test target.

4 years agoDisable compiler hardening on nbsd4.
Darren Tucker [Sun, 25 Apr 2021 04:15:02 +0000 (14:15 +1000)] 
Disable compiler hardening on nbsd4.

The system compiler supports -fstack-protector-all, but using it will
result in an internal compiler error on some files.

4 years agoAdd nbsd3, nbsd4 and nbsd9 test targets.
Darren Tucker [Sun, 25 Apr 2021 03:01:34 +0000 (13:01 +1000)] 
Add nbsd3, nbsd4 and nbsd9 test targets.

4 years agoComment out nbsd2 test target for now.
Darren Tucker [Sat, 24 Apr 2021 12:03:46 +0000 (22:03 +1000)] 
Comment out nbsd2 test target for now.

4 years agoAdd OPENBSD ORIGINAL marker.
Darren Tucker [Sat, 24 Apr 2021 07:52:24 +0000 (17:52 +1000)] 
Add OPENBSD ORIGINAL marker.

4 years agoReplace "==" (a bashism) with "=".
Darren Tucker [Fri, 23 Apr 2021 09:49:46 +0000 (19:49 +1000)] 
Replace "==" (a bashism) with "=".

4 years agoAdd nbsd2 test target.
Darren Tucker [Fri, 23 Apr 2021 06:34:48 +0000 (16:34 +1000)] 
Add nbsd2 test target.

4 years agoAdd obsd68 test target.
Darren Tucker [Fri, 23 Apr 2021 04:54:10 +0000 (14:54 +1000)] 
Add obsd68 test target.

4 years agoRemove dependency on bash.
Darren Tucker [Fri, 23 Apr 2021 04:53:32 +0000 (14:53 +1000)] 
Remove dependency on bash.

4 years agoAdd obsd67 test target.
Darren Tucker [Fri, 23 Apr 2021 04:41:13 +0000 (14:41 +1000)] 
Add obsd67 test target.

4 years agoRe-add macos-11.0 test target.
Darren Tucker [Fri, 23 Apr 2021 01:08:23 +0000 (11:08 +1000)] 
Re-add macos-11.0 test target.

4 years agoAdd openindiana test target.
Darren Tucker [Fri, 23 Apr 2021 00:28:28 +0000 (10:28 +1000)] 
Add openindiana test target.

4 years agoTest krb5 on Solaris 11 too.
Darren Tucker [Fri, 23 Apr 2021 00:26:35 +0000 (10:26 +1000)] 
Test krb5 on Solaris 11 too.

4 years agoDon't always set SUDO.
Darren Tucker [Thu, 22 Apr 2021 12:27:26 +0000 (22:27 +1000)] 
Don't always set SUDO.

Rely on sourcing configs to set as appropriate.

4 years agoRemove now-unused 2nd arg to configs.
Darren Tucker [Thu, 22 Apr 2021 12:26:08 +0000 (22:26 +1000)] 
Remove now-unused 2nd arg to configs.

4 years agoAdd win10 test target.
Darren Tucker [Tue, 20 Apr 2021 15:08:04 +0000 (01:08 +1000)] 
Add win10 test target.

4 years agoAdd nbsd8 test target.
Darren Tucker [Tue, 20 Apr 2021 13:31:29 +0000 (23:31 +1000)] 
Add nbsd8 test target.

4 years agoAdd obsd51 target.
Darren Tucker [Fri, 16 Apr 2021 23:55:47 +0000 (09:55 +1000)] 
Add obsd51 target.

4 years agoAdd fbsd13 target.
Darren Tucker [Fri, 16 Apr 2021 08:14:25 +0000 (18:14 +1000)] 
Add fbsd13 target.

4 years agodepend V_8_6 anongit/V_8_6 github-selfhosted/V_8_6 github/V_8_6 V_8_6_P1
Damien Miller [Fri, 16 Apr 2021 03:55:25 +0000 (13:55 +1000)] 
depend

4 years agocrank version in README and RPM spec files
Damien Miller [Fri, 16 Apr 2021 03:53:02 +0000 (13:53 +1000)] 
crank version in README and RPM spec files

4 years agoupstream: openssh-8.6
djm@openbsd.org [Fri, 16 Apr 2021 03:42:00 +0000 (03:42 +0000)] 
upstream: openssh-8.6

OpenBSD-Commit-ID: b5f3e133c846127ec114812248bc17eff07c3e19

4 years agoupstream: do not pass file/func to monitor; noted by Ilja van Sprundel;
markus@openbsd.org [Thu, 15 Apr 2021 16:24:31 +0000 (16:24 +0000)] 
upstream: do not pass file/func to monitor; noted by Ilja van Sprundel;

ok djm@

OpenBSD-Commit-ID: 85ae5c063845c410283cbdce685515dcd19479fa

4 years agosshd don't exit on transient read errors
Damien Miller [Wed, 14 Apr 2021 01:42:55 +0000 (11:42 +1000)] 
sshd don't exit on transient read errors

openssh-8.5 introduced a regression that would cause sshd to exit
because of transient read errors on the network socket (e.g. EINTR,
EAGAIN). Reported by balu.gajjala AT gmail.com via bz3297

4 years agoperform report_failed_grab() inline
Damien Miller [Sat, 10 Apr 2021 08:45:00 +0000 (18:45 +1000)] 
perform report_failed_grab() inline

4 years agodedicated gnome-ssk-askpass3 source
Damien Miller [Sat, 10 Apr 2021 08:22:57 +0000 (18:22 +1000)] 
dedicated gnome-ssk-askpass3 source

Compatibility with Wayland requires that we use the gdk_seat_grab()
API for grabbing mouse/keyboard, however these API don't exist in
Gtk+2.

This branches gnome-ssk-askpass2.c => gnome-ssk-askpass3.c and
makes the changes to use the gdk_seat_grab() instead of grabbing
mouse/focus separately via GDK.

In the future, we can also use the branched file to avoid some
API that has been soft-deprecated in GTK+3, e.g. gtk_widget_modify_fg

4 years agoEnsure valgrind-out exists.
Darren Tucker [Thu, 8 Apr 2021 05:18:15 +0000 (15:18 +1000)] 
Ensure valgrind-out exists.

Normally the regress tests would create it, but running the unit tests
on their own would fail because the directory did not exist.

4 years agoPass OBJ to unit test make invocation.
Darren Tucker [Thu, 8 Apr 2021 05:17:19 +0000 (15:17 +1000)] 
Pass OBJ to unit test make invocation.

At least the Valgrind unit tests uses $OBJ.

4 years agoAdd pattern for valgrind-unit.
Darren Tucker [Thu, 8 Apr 2021 04:20:12 +0000 (14:20 +1000)] 
Add pattern for valgrind-unit.

4 years agoRun unit tests under valgrind.
Darren Tucker [Thu, 8 Apr 2021 03:31:08 +0000 (13:31 +1000)] 
Run unit tests under valgrind.

Run a separate build for the unit tests under Valgrind.  They take long
enough that running in parallel with the other Valgrind tests helps.

4 years agoifdef out MIN and MAX.
Darren Tucker [Thu, 8 Apr 2021 03:25:57 +0000 (13:25 +1000)] 
ifdef out MIN and MAX.

In -portable, defines.h ensures that these are defined, so redefining
potentially causes a warning.  We don't just delete it to make any
future code syncs a little but easier.  bz#3293.

4 years agoRemove only use of warn().
Darren Tucker [Wed, 7 Apr 2021 00:23:51 +0000 (10:23 +1000)] 
Remove only use of warn().

The warn() function is only used in one place in portable and does not
exist upstream.  Upgrade the only instance it's used to fail()
(the privsep/sandbox+proxyconnect, from back when that was new) and
remove the now-unused function.

4 years agoMove make_tmpdir() into portable-specific area.
Darren Tucker [Wed, 7 Apr 2021 00:18:32 +0000 (10:18 +1000)] 
Move make_tmpdir() into portable-specific area.

Reduces diff vs OpenBSD and makes it more likely diffs will apply
cleanly.

4 years agoupstream: Add TEST_SSH_ELAPSED_TIMES environment variable to print the
dtucker@openbsd.org [Tue, 6 Apr 2021 23:57:56 +0000 (23:57 +0000)] 
upstream: Add TEST_SSH_ELAPSED_TIMES environment variable to print the

elapsed time in seconds of each test.  This depends on "date +%s" which is
not specified by POSIX but is commonly implemented.

OpenBSD-Regress-ID: ec3c8c19ff49b2192116a0a646ee7c9b944e8a9c