binaries. This step splits sshd into a listener and a session binary. More
splits are planned.
After this changes, the listener binary will validate the configuration,
load the hostkeys, listen on port 22 and manage MaxStartups only. All
session handling will be performed by a new sshd-session binary that the
listener fork+execs.
This reduces the listener process to the minimum necessary and sets us
up for future work on the sshd-session binary.
feedback/ok markus@ deraadt@
NB. if you're updating via source, please restart sshd after installing,
otherwise you run the risk of locking yourself out.
OpenBSD-Commit-ID: 43c04a1ab96cdbdeb53d2df0125a6d42c5f19934
We've previously removed a lot of the really old compatibility code,
and with it went the need to include compat.h in most of the files that
have it.
OpenBSD-Commit-ID: 5af8baa194be00a3092d17598e88a5b29f7ea2b4
* upstream: attemp FIDO key signing without PIN and use the error
code returned to fall back only if necessary. Avoids PIN prompts for FIDO
tokens that don't require them; part of GHPR#302
OpenBSD-Commit-ID: 4f752aaf9f2e7c28bcaaf3d4f8fc290131bd038e
* Install Cygwin packages based on OS not config.
* initial list of allowed signers
* upstream: whitespace
OpenBSD-Commit-ID: d297e4387935d4aef091c5e9432578c2e513f538
* upstream: whitespace
OpenBSD-Commit-ID: a5d015efbfd228dc598ffdef612d2da3a579e5d8
* Add cygwin-release test target.
This also moves the cygwin package install from the workflow file to
setup_ci.sh so that we can install different sets of Cygwin packages
for different test configs.
* Add Windows 2022 test targets.
* Add libcrypt-devel to cygwin-release deps.
Based on feedback from vinschen at redhat.com.
* cross-sign allowed_signers with PGP key
Provides continuity of trust from legacy PGP release key to
the SSHSIG signing keys that we will use henceforth for git
signing.
* additional keys
* upstream: whitespace
OpenBSD-Commit-ID: c2bcbf93610d3d62ed206cdf9bf9ff98c6aaf232
* Move sftp from valgrind-2 to 3 to rebalance.
* upstream: sk-usbhid: fix key_lookup() on tokens with built-in UV
explicitly test whether the token performs built-in UV (e.g. biometric
tokens) and enable UV in that case. From Pedro Martelletto via GHPR#388
OpenBSD-Commit-ID: 007eb7e387d27cf3029ab06b88224e03eca62ccd
* Remove arc4random_uniform from arc4random.c
This was previously moved into its own file (matching OpenBSD) which
prematurely committed in commit 73541f2.
* Move OPENBSD ORIGINAL marker.
Putting this after the copyright statement (which doesn't change)
instead of before the version identifier (which does) prevents merge
conflicts when resyncing changes.
* Resync arc4random with OpenBSD.
This brings us up to current, including djm's random-reseeding change,
as prompted by logan at cyberstorm.mu in bz#3467. It brings the
platform-specific hooks from LibreSSL Portable, simplified to match our
use case. ok djm@.
* Remove DEF_WEAK, it's already in defines.h.
* openbsd-compat/bsd-asprintf: add <stdio.h> include for vsnprintf
Fixes the following build failure with Clang 15 on musl:
```
bsd-asprintf.c:51:8: error: call to undeclared library function 'vsnprintf' with type 'int (char *, unsigned long, const char *, struct __va_list_tag *)'; ISO C99 and laterclang -O2 -pipe -fdiagnostics-color=always -frecord-gcc-switches -pipe -Wunknown-warning-option -Qunused-arguments -Wall -Wpointer-arith -Wuninitialized -Wsign-compare -Wformat-security -Wsizeof-pointer-memaccess -Wno-pointer-sign -Wno-unused-result -Wmisleading-indentation -Wbitwise-instead-of-logical -fno-strict-aliasing -mretpoline -ftrapv -fzero-call-used-regs=all -fno-builtin-memset -fstack-protector-strong -fPIE -I. -I. -D_XOPEN_SOURCE=600 -D_BSD_SOURCE -D_DEFAULT_SOURCE -DSSHDIR=\"/etc/ssh\" -D_PATH_SSH_PROGRAM=\"/usr/bin/ssh\" -D_PATH_SSH_ASKPASS_DEFAULT=\"/usr/lib/misc/ssh-askpass\" -D_PATH_SFTP_SERVER=\"/usr/lib/misc/sftp-server\" -D_PATH_SSH_KEY_SIGN=\"/usr/lib/misc/ssh-keysign\" -D_PATH_SSH_PKCS11_HELPER=\"/usr/lib/misc/ssh-pkcs11-helper\" -D_PATH_SSH_SK_HELPER=\"/usr/lib/misc/ssh-sk-helper\" -D_PATH_SSH_PIDDIR=\"/run\" -D_PATH_PRIVSEP_CHROOT_DIR=\"/var/empty\" -DHAVE_CONFIG_H -c cipher-aes.c -o cipher-aes.o
do not support
implicit function declarations [-Wimplicit-function-declaration]
ret = vsnprintf(string, INIT_SZ, fmt, ap2);
^
bsd-asprintf.c:51:8: note: include the header <stdio.h> or explicitly provide a declaration for 'vsnprintf'
1 error generated.
```
* upstream: notifier_complete(NULL, ...) is a noop, so no need to test
that ctx!=NULL; from Corinna Vinschen
OpenBSD-Commit-ID: ade2f2e9cc519d01a586800c25621d910bce384a
* upstream: fix repeated words ok miod@ jmc@
OpenBSD-Commit-ID: 6765daefe26a6b648cc15cadbbe337596af709b7
* upstream: .Li -> .Vt where appropriate; from josiah frentsos,
tweaked by schwarze
ok schwarze
OpenBSD-Commit-ID: 565046e3ce68b46c2f440a93d67c2a92726de8ed
* upstream: ssh-agent: attempt FIDO key signing without PIN and use
the error to determine whether a PIN is required and prompt only if
necessary. from Corinna Vinschen
OpenBSD-Commit-ID: dd6be6a0b7148608e834ee737c3479b3270b00dd
* upstream: a little extra debugging
OpenBSD-Commit-ID: edf1601c1d0905f6da4c713f4d9cecc7d1c0295a
* upstream: sk_enroll: never drop SSH_SK_USER_VERIFICATION_REQD flag
from response
Now that all FIDO signing calls attempt first without PIN and then
fall back to trying PIN only if that attempt fails, we can remove the
hack^wtrick that removed the UV flag from the keys returned during
enroll.
By Corinna Vinschen
OpenBSD-Commit-ID: 684517608c8491503bf80cd175425f0178d91d7f
* upstream: sftp: Don't attempt to complete arguments for
non-existent commands
If user entered a non-existent command (e.g. because they made a
typo) there is no point in trying to complete its arguments. Skip
calling complete_match() if that's the case.
From Michal Privoznik
OpenBSD-Commit-ID: cf39c811a68cde2aeb98fc85addea4000ef6b07a
* upstream: sftp: Be a bit more clever about completions
There are commands (e.g. "get" or "put") that accept two
arguments, a local path and a remote path. However, the way
current completion is written doesn't take this distinction into
account and always completes remote or local paths.
By expanding CMD struct and "cmds" array this distinction can be
reflected and with small adjustment to completer code the correct
path can be completed.
By Michal Privoznik, ok dtucker@
OpenBSD-Commit-ID: 1396d921c4eb1befd531f5c4a8ab47e7a74b610b
* upstream: correct error value
OpenBSD-Commit-ID: 780efcbad76281f11f14b2a5ff04eb6db3dfdad4
* upstream: actually hook up restrict_websafe; the command-line flag
was never actually used. Spotted by Matthew Garrett
OpenBSD-Commit-ID: 0b363518ac4c2819dbaa3dfad4028633ab9cdff1
* upstream: Add a sshkey_check_rsa_length() call for checking the
length of an RSA key; ok markus@
OpenBSD-Commit-ID: de77cd5b11594297eda82edc594b0d32b8535134
* upstream: add a RequiredRSASize for checking RSA key length in
ssh(1). User authentication keys that fall beneath this limit will be
ignored. If a host presents a host key beneath this limit then the connection
will be terminated (unfortunately there are no fallbacks in the protocol for
host authentication).
feedback deraadt, Dmitry Belyavskiy; ok markus@
OpenBSD-Commit-ID: 430e339b2a79fa9ecc63f2837b06fdd88a7da13a
* upstream: Add RequiredRSASize for sshd(8); RSA keys that fall
beneath this limit will be ignored for user and host-based authentication.
Feedback deraadt@ ok markus@
OpenBSD-Commit-ID: 187931dfc19d51873df5930a04f2d972adf1f7f1
* upstream: better debugging for connect_next()
OpenBSD-Commit-ID: d16a307a0711499c971807f324484ed3a6036640
* upstream: sftp-server(8): add a "users-groups-by-id@openssh.com"
extension request that allows the client to obtain user/group names that
correspond to a set of uids/gids.
Will be used to make directory listings more useful and consistent
in sftp(1).
ok markus@
OpenBSD-Commit-ID: 7ebabde0bcb95ef949c4840fe89e697e30df47d3
* upstream: extend sftp-common.c:extend ls_file() to support supplied
user/group names; ok markus@
OpenBSD-Commit-ID: c70c70498b1fdcf158531117e405b6245863bfb0
* upstream: sftp client library support for
users-groups-by-id@openssh.com; ok markus@
OpenBSD-Commit-ID: ddb2f33a2da6349a9a89a8b5bcb9ca7c999394de
* upstream: use users-groups-by-id@openssh.com sftp-server extension
(when available) to fill in user/group names for directory listings.
Implement a client-side cache of see uid/gid=>user/group names. ok markus@
OpenBSD-Commit-ID: f239aeeadfa925a37ceee36ee8b256b8ccf4466e
* avoid Wuninitialized false positive in gcc-12ish
* no need for glob.h here
it also causes portability problems
* upstream: add RequiredRSASize to the list of keywords accepted by
-o; spotted by jmc@
OpenBSD-Commit-ID: fe871408cf6f9d3699afeda876f8adbac86a035e
* upstream: Fix typo. From AlexanderStohr via github PR#343.
OpenBSD-Commit-ID: a134c9b4039e48803fc6a87f955b0f4a03181497
* upstream: openssh-9.1
OpenBSD-Commit-ID: 5a467b2ee81da01a86adf1ad93b62b1728494e56
* crank versions in RPM spec files
* update release notes URL
* update .depend
* remove mention of --with-security-key-builtin
it is enabled by default when libfido2 is installed
* mention libfido2 autodetection
* whitespace at EOL
* Test commits to all branches of portable.
Only test OpenBSD upstream on commits to master since that's what it
tracks.
* Add 9.1 branch to CI status page.
* Add LibreSSL 3.6.0 to test suite.
While there, bump OpenSSL to latest 1.1.1q release.
* upstream: honour user's umask if it is more restrictive then the ssh
default (022); based on patch from Alex Henrie, ok dtucker@ deraadt@
OpenBSD-Commit-ID: fe1b9e15fc9a4f49fc338e848ce14d8727abe82d
* skip bsd-poll.h if poll.h found; ok dtucker
* Fix snprintf configure test for clang 15
Clang 15 -Wimplicit-int defaults to an error in C99 mode and above.
A handful of tests have "main(..." and not "int main(..." which caused
the tests to produce incorrect results.
* undef _get{short,long} before redefining
* revert c64b62338b and guard POLL* defines instead
c64b62338b broke OSX builds, which do have poll.h but lack ppoll(2)
Spotted by dtucker
* OpenSSL dev branch now identifies as 3.2.0.
* upstream: document "-O no-restrict-websafe"; spotted by Ross L
Richardson
OpenBSD-Commit-ID: fe9eaa50237693a14ebe5b5614bf32a02145fe8b
* upstream: ssh-agent.1: - use Nm not Xr for self-ref - while here,
wrap a long line
ssh-agent.c:
- add -O to usage()
OpenBSD-Commit-ID: 855dac4695cef22e96d69c53436496bc408ca389
* upstream: use correct type with sizeof ok djm@
OpenBSD-Commit-ID: d6c882c2e8a42ff831a5b3cbc2c961ecb2dd6143
* upstream: when scp(1) is using the SFTP protocol for transport (the
default), better match scp/rcp's handling of globs that don't match the
globbed characters but do match literally (e.g. trying to transfer
"foo.[1]").
Previously scp(1) in SFTP mode would not match these pathnames but
legacy scp/rcp mode would.
Reported by Michael Yagliyan in bz3488; ok dtucker@
OpenBSD-Commit-ID: d8a3773f53015ba811fddba7473769a2fd343e11
* upstream: regress test for unmatched glob characters; fails before
previous commit but passes now. bz3488; prodded by dtucker@
OpenBSD-Regress-ID: 0cc5cc9ea4a6fd170dc61b9212f15badaafb3bbd
* upstream: Be more paranoid with host/domain names coming from the
never write a name with bad characters to a known_hosts file.
reported by David Leadbeater, ok deraadt@
OpenBSD-Commit-ID: ba9b25fa8b5490b49398471e0c9657b0cbc7a5ad
* upstream: begin big refactor of sshkey
Move keytype data and some of the type-specific code (allocation,
cleanup, etc) out into each key type's implementation. Subsequent
commits will move more, with the goal of having each key-*.c file
owning as much of its keytype's implementation as possible.
lots of feedback + ok markus@
OpenBSD-Commit-ID: 0f2b4334f73914344e9e5b3d33522d41762a57ec
* upstream: factor out sshkey_equal_public()
feedback/ok markus@
OpenBSD-Commit-ID: 1368ba114cb37732fe6ec3d89c7e6d27ea6fdc94
* upstream: factor out public key serialization
feedback/ok markus@
OpenBSD-Commit-ID: a3570c4b97290c5662890aea7328d87f55939033
* upstream: refactor and simplify sshkey_read()
feedback/ok markus@
OpenBSD-Commit-ID: 0d93b7a56e31cd06a8bb0d2191d084ce254b0971
* upstream: factor out key generation
feedback/ok markus@
OpenBSD-Commit-ID: 5b4211bff4de8d9adb84bc72857a8c42c44e7ceb
* upstream: refactor sshkey_from_private()
feedback/ok markus@
OpenBSD-Commit-ID: e5dbe7a3545930c50f70ee75c867a1e08b382b53
* upstream: refactor sshkey_from_blob_internal()
feedback/ok markus@
OpenBSD-Commit-ID: 1f46c0cbb8060ee9666a02749594ad6658c8e283
* upstream: refactor sshkey_sign() and sshkey_verify()
feedback/ok markus@
OpenBSD-Commit-ID: 368e662c128c99d05cc043b1308d2b6c71a4d3cc
* upstream: refactor certify
feedback/ok markus@
OpenBSD-Commit-ID: 35d742992e223eaca3537e6fb3d3002c08eed4f6
* upstream: refactor sshkey_private_serialize_opt()
feedback/ok markus@
OpenBSD-Commit-ID: 61e0fe989897901294efe7c3b6d670cefaf44cbd
* upstream: refactor sshkey_private_deserialize
feedback/ok markus@
OpenBSD-Commit-ID: f5ca6932fdaf840a5e8250becb38315a29b5fc9f
* fix merge botch
* upstream: allow ssh-keyscan(1) to accept CIDR address ranges, e.g.
ssh-keyscan 192.168.0.0/24
If a CIDR range is passed, then it will be expanded to all possible
addresses in the range including the all-0s and all-1s addresses.
bz#976 feedback/ok markus@
OpenBSD-Commit-ID: ce6c5211f936ac0053fd4a2ddb415277931e6c4b
* upstream: put sshkey_check_rsa_length() back in sshkey.c to unbreak
OPENSSL=no builds
OpenBSD-Commit-ID: 99eec58abe382ecd14b14043b195ee1babb9cf6e
* OpenSSL dev branch is 302 not 320.
While there, also accept 301 which it shat it was previously.
* upstream: Use variable for diff options
instead of unconditionally specifying "-rN". This will make life easier
in -portable where not all diff's understand -N.
OpenBSD-Regress-ID: 8b8a407115546be1c6d72d350b1e4f1f960d3cd3
* Check for sockaddr_in.sin_len.
If found, set SOCK_HAS_LEN which is used in addr.c. Should fix keyscan
tests on platforms with this (eg old NetBSD).
* Always use compat getentropy.
Have it call native getentropy and fall back as required. Should fix
issues of platforms where libc has getentropy but it is not implemented
in the kernel. Based on github PR#354 from simsergey.
* Include time.h when defining timegm.
Fixes build on some platforms eg recent AIX.
* Compat tests need libcrypto.
This was moved to CHANNELLIBS during the libs refactor. Spotted by
rapier at psc.edu.
* Run compat regress tests too.
* Add tests for OpenSSL 3.0.7 and LibreSSL 3.6.1.
* Only run opensslver tests if built with OpenSSL.
* Increase selfhosted job timeout.
The default job timeout of 360 (6h) is not enough to complete the
regress tests for some of the slow VMs depending on the load on the host.
Increase to 600 (10h).
* Fix compat regress to work with non-GNU make.
* Link libssh into compat tests.
The cygwin compat code uses xmalloc, so add libssh.a so pick up that.
* Rerun tests on changes to Makefile.in in any dir.
* upstream: replace recently-added valid_domain() check for hostnames
going to known_hosts with a more relaxed check for bad characters; previous
commit broke address literals. Reported by/feedback from florian@
OpenBSD-Commit-ID: 10b86dc6a4b206adaa0c11b58b6d5933898d43e0
* Don't run openbsd-compat tests on Cygwin.
Add "compat-tests" to the default TEST_TARGET so we can override as
necessary. Override TEST_TARGET for Cygwin as the tests don't currently
compile there.
* Fix broken zlib link.
* configure.ac: Add <pty.h> include for openpty
Another Clang 16ish fix (which makes -Wimplicit-function-declaration
an error by default). github PR#355.
See: 2efd71da49b9cfeab7987058cf5919e473ff466b
See: be19763532
* configure.ac: Fix -Wstrict-prototypes
Clang 16 now warns on this and it'll be removed in C23, so let's
just be future proof. It also reduces noise when doing general
Clang 16 porting work (which is a big job as it is). github PR#355.
Signed-off-by: Sam James <sam@gentoo.org>
* Fix setres*id checks to work with clang-16.
glibc has the prototypes for setresuid and setresgid behind _GNU_SOURCE,
and clang 16 will error out on implicit function definitions, so add
_GNU_SOURCE and the required headers to the configure checks. From
sam at @gentoo.org via bz#3497.
* Fix tracing disable on FreeBSD.
Some versions of FreeBSD do not support using id 0 to refer to the
current pid for procctl, so pass getpid() explicitly. From
emaste at freebsd.org.
* Use "prohibit-password" in -portable comments.
"without-password" is the deprecated alias for "prohibit-password",
so we should reference the latter. From emaste at freebsd.org.
* Link to branch-specific queries for V_9_1 status.
* upstream: Fix typo. From pablomh via -portable github PR#344.
OpenBSD-Commit-ID: d056ee2e73691dc3ecdb44a6de68e6b88cd93827
* upstream: Import regenerated moduli.
OpenBSD-Commit-ID: b0e54ee4d703bd6929bbc624068666a7a42ecb1f
* Add CIFuzz integration
* Run cifuzz workflow on the actions as regular CI.
* Whitespace change to trigger CIFuzz workflow.
* Do not run CIFuzz on selfhosted tree.
We already run it on the regular tree, no need to double up.
* Add CIFuzz status badge.
* Branch-specific links for master status badges.
* Fix merge conflict.
* upstream: fix parsing of hex cert expiry time; was checking whether the
start time began with "0x", not the expiry time.
from Ed Maste
OpenBSD-Commit-ID: 6269242c3e1a130b47c92cfca4d661df15f05739
* upstream: Check for and disallow MaxStartups values less than or
equal to zero during config parsing, rather than faling later at runtime.
bz#3489, ok djm@
OpenBSD-Commit-ID: d79c2b7a8601eb9be493629a91245d761154308b
* upstream: Remove some set but otherwise unused variables, spotted
in -portable by clang 16's -Wunused-but-set-variable. ok djm@
OpenBSD-Commit-ID: 3d943ddf2369b38fbf89f5f19728e7dc1daf3982
* upstream: The IdentityFile option in ssh_config can also be used to
specify a public key file, as documented in ssh.1 for the -i option. Document
this also for IdentityFile in ssh_config.5, for documentation completeness.
From laalsaas at systemli.org via portable github PR#352, ok jmc@ djm@
OpenBSD-Commit-ID: 2f943be9f96e60ef81a9a4faa25b009999f9883b
* Split out rekey test since it runs the longest.
* Update checkout and upload actions.
Update actions/checkout and actions/upload-artifact to main branch for
compatibility with node.js v16.
* Add valrind-5 test here too.
* Run vm startup and shutdown from runner temp dir.
Should work even if the github workspace dir is on a stale sshfs mount.
* Shutdown any VM before trying to check out repo.
In the case where the previous run did not clean up, the checkout will
fail as it'll leave a stale mount.
* Avoid assuming layout of fd_set
POSIX doesn't specify the internal layout of the fd_set object, so let's
not assume it is just a bit mask. This increases compatibility with
systems that have a different layout.
The assumption is also worthless as we already refuse to use file
descriptors over FD_SETSIZE anyway. Meaning that the default size of
fd_set is quite sufficient.
* Fix comment text. From emaste at freebsd.org.
* Defer seed_rng until after closefrom call.
seed_rng will initialize OpenSSL, and some engine providers (eg Intel's
QAT) will open descriptors for their own use. bz#3483, patch from
joel.d.schuetze at intel.com, ok djm@
* upstream: typo in comment
OpenBSD-Commit-ID: 39c58f41e0f32d1ff31731fa6f5bbbc3ad25084a
* upstream: rename client_global_hostkeys_private_confirm() to
client_global_hostkeys_prove_confirm(), as it handles the
"hostkeys-prove00@openssh.com" message; no functional change
OpenBSD-Commit-ID: 31e09bd3cca6eed26855b88fb8beed18e9bd026d
* upstream: Remove errant colon and simplify format
string in error messages. Patch from vapier at chromium.org.
OpenBSD-Commit-ID: fc28466ebc7b74e0072331947a89bdd239c160d3
* upstream: Fix typo in fatal error message.
Patch from vapier at chromium.org.
OpenBSD-Commit-ID: 8a0c164a6a25eef0eedfc30df95bfa27644e35cf
* Skip reexec test on OpenSSL 1.1.1 specifically.
OpenSSL 1.1.1 has a bug in its RNG that breaks reexec fallback, so skip
that test. See bz#3483 for details.
* Remove seed passing over reexec.
This was added for the benefit of platforms using ssh-rand-helper to
prevent a delay on each connection as sshd reseeded itself.
ssh-random-helper is long gone, and since the re-exec happens before the
chroot the re-execed sshd can reseed itself normally. ok djm@
* upstream: Handle dynamic remote port forwarding in escape commandline's
-R processing. bz#3499, ok djm@
OpenBSD-Commit-ID: 194ee4cfe7ed0e2b8ad0727f493c798a50454208
* Add dfly62 test target.
* If we haven't found it yet, recheck for sys/stat.h.
On some very old platforms, sys/stat.h needs sys/types.h, however
autoconf 2.71's AC_CHECK_INCLUDES_DEFAULT checks for them in the
opposite order, which in combination with modern autoconf's
"present but cannot be compiled" behaviour causes it to not be
detected.
* Add fallback for old platforms w/out MAP_ANON.
* Remove explicit "default" test config argument.
Not specifying the test config implicitly selects default args.
* Remove unused self-hosted test targets.
* Rename "os" in matrix to "target".
This is in preparation to distinguish this from the host that the runner
runs on in case where they are separate (eg VMs).
* Add "libvirt" label to dfly30.
* Make "config" in matrix singular and pass in env.
This will allow the startup scripts to adapt their behaviour based on
the type and config.
* Run vmstartup from temp dir.
This will allow us to create ephemeral disk images per-runner.
* Rework how selfhosted tests interact with runners.
Previously there was one runner per test target (mostly VMs). This had
a few limitations:
- multiple tests that ran on the same target (eg multiple build
configs) were serialized on availability or that runner.
- it needed manual balancing of VMs over host machines.
To address this, make VMs that use ephemeral disks (ie most of them)
all use a pool of runners with the "libvirt" label. This requires that
we distinguish between "host" and "target" for those. Native runners
and VMs with persistent disks (eg the constantly-updated snapshot ones)
specify the same host and target.
This should improve test throughput.
* Skip unit tests on slow riscv64 hardware.
* Use -fzero-call-used-regs=used on clang 15.
clang 15 seems to have a problem with -fzero-call-used-reg=all which
causes spurious "incorrect signature" failures with ED25519. On those
versions, use -fzero-call-used-regs=used instead. (We may add exceptions
later if specific versions prove to be OK). Also move the GCC version
check to match.
Initial investigation by Daniel Pouzzner (douzzer at mega nu), workaround
suggested by Bill Wendling (morbo at google com). bz#3475, ok djm@
* upstream: In channel_request_remote_forwarding the parameters for
permission_set_add are leaked as they are also duplicated in the call. Found
by CodeChecker. ok djm
OpenBSD-Commit-ID: 4aef50fa9be7c0b138188814c8fe3dccc196f61e
* upstream: New EnableEscapeCommandline ssh_config(5) option
This option (default "no") controls whether the ~C escape is available.
Turning it off by default means we will soon be able to use a stricter
default pledge(2) in the client.
feedback deraadt@ dtucker@; tested in snaps for a while
OpenBSD-Commit-ID: 7e277595d60acb8263118dcb66554472257b387a
* upstream: tighten pledge(2) after session establishment
feedback, ok & testing in snaps deraadt@
OpenBSD-Commit-ID: aecf4d49d28586dfbcc74328d9333398fef9eb58
* upstream: Add void to client_repledge args to fix compiler warning. ok djm@
OpenBSD-Commit-ID: 7e964a641ce4a0a0a11f047953b29929d7a4b866
* upstream: Log output of ssh-agent and ssh-add
This should make debugging easier.
OpenBSD-Regress-ID: 5974b02651f428d7e1079b41304c498ca7e306c8
* upstream: Clean up ssh-add and ssh-agent logs.
OpenBSD-Regress-ID: 9eda8e4c3714d7f943ab2e73ed58a233bd29cd2c
* Restore ssh-agent permissions on exit.
...enough that subsequent builds can overwrite ssh-agent if necessary.
* upstream: make struct sshbuf private
and remove an unused field; ok dtucker
OpenBSD-Commit-ID: c7a3d77c0b8c153d463398606a8d57569186a0c3
* upstream: Remove duplicate includes.
Patch from AtariDreams via github PR#364.
OpenBSD-Commit-ID: b9186638a05cb8b56ef7c0de521922b6723644ea
* Fix typo in comment. Spotted by tim@
* Update autotools
Regenerate config files using latest autotools
* disable SANDBOX_SECCOMP_FILTER_DEBUG
It was mistakenly enabled in 2580916e48
Reported by Peter sec-openssh-com.22.fichtner AT 0sg.net
* Add SANDBOX_DEBUG to the kitchensink test build.
* upstream: Fix comment typo.
OpenBSD-Regress-ID: 3b04faced6511bb5e74648c6a4ef4bf2c4decf03
* upstream: remove '?' from getopt(3) loops
userspace: remove vestigial '?' cases from top-level getopt(3) loops
getopt(3) returns '?' when it encounters a flag not present in the in
the optstring or if a flag is missing its option argument. We can
handle this case with the "default" failure case with no loss of
legibility. Hence, remove all the redundant "case '?':" lines.
Prompted by dlg@. With help from dlg@ and millert@.
Link: https://marc.info/?l=openbsd-tech&m=167011979726449&w=2
ok naddy@ millert@ dlg@
OpenBSD-Commit-ID: b2f89346538ce4f5b33ab8011a23e0626a67e66e
* upstream: Add server debugging for hostbased auth.
auth_debug_add queues messages about the auth process which is sent to
the client after successful authentication. This also sends those to
the server debug log to aid in debugging. From bz#3507, ok djm@
OpenBSD-Commit-ID: 46ff67518cccf9caf47e06393e2a121ee5aa258a
* upstream: Warn if no host keys for hostbased auth can be loaded.
OpenBSD-Commit-ID: 2a0a13132000cf8d3593133c1b49768aa3c95977
* use calloc for allocating arc4random structs
ok dtucker
* Move obsdsnap test VMs to ephemeral runners.
* Run upstream obsdsnap tests on ephemeral runners.
* obsdsnap test VMs runs-on libvirt too.
* Fetch regress logs from obj dir.
* Set group perms on regress dir.
This ensures that the tests don't fail due to StrictMode checks.
* Use sudo when resetting perms on directories.
* Add tests for LibreSSL 3.7.0 and OpenSSL 1.1.1s.
* Simply handling of SSH_CONNECTION PAM env var.
Prompted by bz#3508: there's no need to cache the value of
sshpam_conninfo so remove the global. While there, add check of
return value from pam_putenv. ok djm@
* upstream: The idiomatic way of coping with signed char vs unsigned
char (which did not come from stdio read functions) in the presence of
ctype macros, is to always cast to (unsigned char). casting to (int)
for a "macro" which is documented to take int, is weird. And sadly wrong,
because of the sing extension risk.. same diff from florian
OpenBSD-Commit-ID: 65b9a49a68e22ff3a0ebd593f363e9f22dd73fea
* upstream: add a -X option to both scp(1) and sftp(1) to allow
control over some SFTP protocol knobs: the copy buffer length and
the number of inflight requests, both of which are used during
upload/download.
Previously these could be controlled in sftp(1) using the -b/-R options.
This makes them available in both SFTP protocol clients using the same
option character sequence.
ok dtucker@
OpenBSD-Commit-ID: 27502bffc589776f5da1f31df8cb51abe9a15f1c
* upstream: add -X to usage();
OpenBSD-Commit-ID: 1bdc3df7de11d766587b0428318336dbffe4a9d0
* upstream: Clear signal mask early in main(); sshd may have been
started with one or more signals masked (sigprocmask(2) is not cleared
on fork/exec) and this could interfere with various things, e.g. the
login grace timer.
Execution environments that fail to clear the signal mask before running
sshd are clearly broken, but apparently they do exist.
Reported by Sreedhar Balasubramanian; ok dtucker@
OpenBSD-Commit-ID: 77078c0b1c53c780269fc0c416f121d05e3010ae
* upstream: Mention that scp uses the SFTP protocol and remove
reference to legacy flag. Spotted by, feedback and ok jmc@
OpenBSD-Commit-ID: 9dfe04966f52e941966b46c7a2972147f95281b3
* upstream: spelling fixes; from paul tagliamonte amendments to his
diff are noted on tech
OpenBSD-Commit-ID: d776dd03d0b882ca9c83b84f6b384f6f9bd7de4a
* upstream: fix bug in PermitRemoteOpen which caused it to ignore its
first argument unless it was one of the special keywords "any" or "none".
Reported by Georges Chaudy in bz3515; ok dtucker@
OpenBSD-Commit-ID: c5678a39f1ff79993d5ae3cfac5746a4ae148ea5
* upstream: regression test for PermitRemoteOpen
OpenBSD-Regress-ID: 8271aafbf5c21950cd5bf966f08e585cebfe630c
* upstream: suppress "Connection closed" message when in quiet mode
OpenBSD-Commit-ID: 8a3ab7176764da55f60bfacfeae9b82d84e3908f
* upstream: add ptimeout API for keeping track of poll/ppoll
timeouts; ok dtucker markus
OpenBSD-Commit-ID: 3335268ca135b3ec15a947547d7cfbb8ff929ead
* upstream: replace manual poll/ppoll timeout math with ptimeout API
feedback markus / ok markus dtucker
OpenBSD-Commit-ID: c5ec4f2d52684cdb788cd9cbc1bcf89464014be2
* upstream: Add channel_force_close()
This will forcibly close an open channel by simulating read/write errors,
draining the IO buffers and calling the detach function.
Previously the detach function was only ever called during channel garbage
collection, but there was no way to signal the user of a channel (e.g.
session.c) that its channel was being closed deliberately (vs. by the
usual state-machine logic). So this adds an extra "force" argument to the
channel cleanup callback to indicate this condition.
ok markus dtucker
OpenBSD-Commit-ID: 23052707a42bdc62fda2508636e624afd466324b
* upstream: tweak channel ctype names
These are now used by sshd_config:ChannelTimeouts to specify timeouts by
channel type, so force them all to use a similar format without whitespace.
ok dtucker markus
OpenBSD-Commit-ID: 66834765bb4ae14f96d2bb981ac98a7dae361b65
* upstream: Add channel_set_xtype()
This sets an "extended" channel type after channel creation (e.g.
"session:subsystem:sftp") that will be used for setting channel inactivity
timeouts.
ok markus dtucker
OpenBSD-Commit-ID: 42564aa92345045b4a74300528f960416a15d4ca
* upstream: Implement channel inactivity timeouts
This adds a sshd_config ChannelTimeouts directive that allows channels that
have not seen traffic in a configurable interval to be automatically closed.
Different timeouts may be applied to session, X11, agent and TCP forwarding
channels.
Note: this only affects channels over an opened SSH connection and not
the connection itself. Most clients close the connection when their channels
go away, with a notable exception being ssh(1) in multiplexing mode.
ok markus dtucker
OpenBSD-Commit-ID: ae8bba3ed9d9f95ff2e2dc8dcadfa36b48e6c0b8
* unbreak scp on NetBSD 4.x
e555d5cad5 effectively increased the default copy buffer size for SFTP
transfers. This caused NetBSD 4.x to hang during the "copy local file to
remote file in place" scp.sh regression test.
This puts back the original 32KB copy buffer size until we can properly
figure out why.
lots of debugging assistance from dtucker@
* upstream: Copy bytes from the_banana[] rather than banana()
Fixes test failure due to segfault seen on arm64 with xonly snap.
ok djm
OpenBSD-Regress-ID: 86e2aa4bbd1dff1bc4ebb2969c0d6474485be046
* upstream: unit tests for misc.c:ptimeout_* API
OpenBSD-Regress-ID: 01f8fb12d08e5aaadd4bd4e71f456b6588be9a94
* upstream: fix typo in verbose logging
OpenBSD-Regress-ID: 0497cdb66e003b2f50ed77291a9104fba2e017e9
* upstream: regression test for ChannelTimeout
OpenBSD-Regress-ID: 280bfbefcfa415428ad744e43f69a8dede8ad685
* upstream: Save debug logs from ssh for debugging purposes.
OpenBSD-Regress-ID: 109e40b06de1c006a3b8e0d8745b790b2c5870a0
* Set OPENSSL_BIN from OpenSSL directory.
* Check openssl_bin path is executable before using.
* Use autoconf to find openssl binary.
It's possible to install an OpenSSL in a path not in the system's
default library search path. OpenSSH can still use this (eg if you
specify an rpath) but the openssl binary there may not work. If one is
available on the system path just use that.
* Use our own netcat for dynamic-forward test.
That way we can be surer about its behaviour rather than trying to
second-guess the behaviour of various netcat implementations.
* upstream: When OpenSSL is not available, skip parts of percent test
that require it. Based on github pr#368 from ren mingshuai.
OpenBSD-Regress-ID: 49a375b2cf61ccb95b52e75e2e025cd10988ebb2
* don't test IPv6 addresses if platform lacks support
* Skip dynamic-forward test on minix3.
This test relies on loopback addresses which minix does not have.
Previously the test would not run at all since it also doesn't have
netcat, but now we use our own netcat it tries and fails.
* try to improve logging for dynamic-forward test
previously the logs from the ssh used to exercise the forwarding
channel would clobber the logs from the ssh actually doing the
forwarding
* upstream: tweak previous; ok djm
OpenBSD-Commit-ID: 229c493452766d70a78b0f02f6ff9894f9028858
* upstream: Switch scp from using pipes to a socketpair for
communication with it's ssh sub-processes. We no longer need to reserve two
descriptors to ensure that we don't end up using fd 0-2 unexpectedly, that is
handled by sanitise_stdfd() in main(). Based on an original diff from djm@.
OK deraadt@ djm@
OpenBSD-Commit-ID: b80c372faac462471e955ddeab9480d668a2e48d
* add back use of pipes in scp.c under USE_PIPES
This matches sftp.c which prefers socketpair but uses pipes on
some older platforms.
* remove buffer len workaround for NetBSD 4.x
Switching to from pipes to a socketpair for communicating with the
ssh process avoids the (kernel bug?) problem.
* upstream: rewrite this test to use a multiplexed ssh session so we can
control its lifecycle without risk of race conditions; fixes some of the
Github integration tests for openssh-portable
OpenBSD-Regress-ID: 5451cad59ba0d43ae9eeda48ec80f54405fee969
* upstream: remove whitespace at EOL from code extracted from SUPERCOP
OpenBSD-Commit-ID: 1ec524ff2fbb9387d731601437c82008f35a60f4
* upstream: ignore bogus upload/download buffer lengths in the limits
extension
OpenBSD-Commit-ID: c5b023e0954693ba9a5376e4280c739b5db575f8
* upstream: clamp the minimum buffer lengths and number of inflight
requests too
OpenBSD-Commit-ID: c4965f62fa0ba850940fd66ae3f60cf516bbcd56
* upstream: avoid printf("%s", NULL) if using ssh
-oUserKnownHostsFile=none and a hostkey in one of the system known hosts file
changes; ok dtucker@
OpenBSD-Commit-ID: 7ca87614bfc6da491315536a7f2301434a9fe614
* upstream: Add a "Host" line to the output of ssh -G showing the
original host arg. Inspired by patch from vincent at bernat.ch via bz#3343,
ok djm@
OpenBSD-Commit-ID: 59c0f60a222113a44d0650cd394376e3beecc883
* Remove skipping test when scp not in path.
An upcoming change renders this obsolete by adding scp's path to the
test sshd's PATH, and removing this first will make the subsequent sync
easier.
* upstream: Add scp's path to test sshd's PATH.
If the scp we're testing is fully qualified (eg it's not in the system
PATH) then add its path to the under-test sshd's PATH so we can find
it. Prompted by bz#3518.
OpenBSD-Regress-ID: 7df4f5a0be3aa135495b7e5a6719d3cbc26cc4c0
* upstream: Move scp path setting to a helper function. The previous
commit to add scp to the test sshd's path causes the t-envpass test to fail
when the test scp is given using a fully qualified path. Put this in a
helper function and only call it from the scp tests.
OpenBSD-Regress-ID: 7533dc1c4265c1de716abb062957994195b36df4
* Retry package installation 3 times.
When setting up the CI environment, retry package installation 3 times
before going up. Should help prevent spurious failures during
infrastructure issues.
* upstream: Document "UserKnownHostsFile none". ok djm@
OpenBSD-Commit-ID: f695742d39e34ecdcc3c861c3739a84648a4bce5
* upstream: fix double phrase in previous;
OpenBSD-Commit-ID: 671e6c8dc5e9230518b2bbfa143daaa88adc66c2
* upstream: Instead of skipping the all-tokens test if we don't have
OpenSSL (since we use it to compute the hash), put the hash at the end and
just omit it if we don't have it. Prompted by bz#3521.
OpenBSD-Regress-ID: c79ecba64250ed3b6417294b6c965e6b12ca5eea
* upstream: Shell syntax fix. From ren mingshuai vi github PR#369.
OpenBSD-Regress-ID: 6696b2eeefe128099fc3d7ea9f23252cc35156f9
* Allow writev is seccomp sandbox.
This seems to be used by recent glibcs at least in some configurations.
From bz#3512, ok djm@
* upstream: update OpenSSH's Ed25519 code to the last version of SUPERCOP
(20221122) and change the import approach to the same one we use for
Streamlined NTRUPrime: use a shell script to extract the bits we need from
SUPERCOP, make some minor adjustments and squish them all into a single file.
ok tb@ tobhe@
OpenBSD-Commit-ID: 1bc0fd624cb6af440905b8ba74ac7c03311b8e3b
* upstream: adapt to ed25519 changes in src/usr.bin/ssh
OpenBSD-Regress-ID: 4b3e7ba7ee486ae8a0b4790f8112eded2bb7dcd5
* upstream: Add a sshd_config UnusedConnectionTimeout option to terminate
client connections that have no open channels for some length of time. This
complements the recently-added ChannelTimeout option that terminates inactive
channels after a timeout.
ok markus@
OpenBSD-Commit-ID: ca983be74c0350364c11f8ba3bd692f6f24f5da9
* upstream: unbreak test: cannot access shell positional parameters
past $9 without wrapping the position in braces (i.e. need ${10}, etc.)
OpenBSD-Regress-ID: 3750ec98d5d409ce6a93406fedde6f220d2ea2ac
* upstream: regression test for UnusedConnectionTimeout
OpenBSD-Regress-ID: 7f29001374a68e71e5e078f69e4520cf4bcca084
* upstream: also check that an active session inhibits
UnusedConnectionTimeout idea markus@
OpenBSD-Regress-ID: 55c0fb61f3bf9e092b0a53f9041d3d2012f14003
* upstream: For "ssh -V" always exit 0, there is no need to check opt
again. This was missed when the fallthrough in the switch case above it was
removed. OK deraadt@
OpenBSD-Commit-ID: 5583e5d8f6d62a8a4215cfa95a69932f344c8120
* upstream: Add a -V (version) option to sshd like the ssh client
has. OK markus@ deraadt@
OpenBSD-Commit-ID: abe990ec3e636fb040132aab8cbbede98f0c413e
* upstream: when restoring non-blocking mode to stdio fds, restore
exactly the flags that ssh started with and don't just clobber them with
zero, as this could also remove the append flag from the set;
bz3523; ok dtucker@
OpenBSD-Commit-ID: 1336b03e881db7564a4b66014eb24c5230e9a0c0
* Skip connection-timeout when missing FD passing.
This tests uses multiplexing which uses file descriptor passing, so
skip it if we don't have that. Fixes test failures on Cygwin.
* Skip connection-timeout test under Valgrind.
Valgrind slows things down so much that the timeout test fails. Skip
this test until we figure out if we can make it work.
* upstream: tweak previous; ok djm
OpenBSD-Commit-ID: df71ce4180c58202dfdc1d92626cfe900b91b7c3
* upstream: Create and install sshd random relink kit.
../Makefile.inc and Makfile are concatenated for reuse, which hopefully won't
be too fragile, we'll see if we need a different approach. The resulting sshd
binary is tested with the new sshd -V option before installation. As the
binary layout is now semi-unknown (meaning relative, fixed, and gadget
offsets are not precisely known), change the filesystem permissions to 511 to
prevent what I call "logged in BROP". I have ideas for improving this further
but this is a first step ok djm
OpenBSD-Commit-ID: 1e0a2692b7e20b126dda60bf04999d1d30d959d8
* upstream: delete useless dependency
OpenBSD-Commit-ID: e1dc11143f83082e3154d6094f9136d0dc2637ad
* fix libfido2 detection without pkg-config
Place libfido2 before additional libraries (that it may depend upon)
and not after. bz3530 from James Zhang; ok dtucker@
* Skip connection-timeout test on minix3.
Minix 3's Unix domain sockets don't seem to work the way we expect, so
skip connection-timeout test on that platform. While there, group
together all similarly skipped tests and explicitly comment.
* upstream: fix double-free caused by compat_kex_proposal(); bz3522
by dtucker@, ok me
OpenBSD-Commit-ID: 2bfc37cd2d41f67dad64c17a64cf2cd3806a5c80
* upstream: openssh-9.2
OpenBSD-Commit-ID: f7389f32413c74d6e2055f05cf65e7082de03923
* upstream: Check if we can copy sshd or need to use sudo to do so
during reexec test. Skip test if neither can work. Patch from anton@, tweaks
from me.
OpenBSD-Regress-ID: 731b96ae74d02d5744e1f1a8e51d09877ffd9b6d
* upstream: test compat_kex_proposal(); by dtucker@
OpenBSD-Regress-ID: 0e404ee264db546f9fdbf53390689ab5f8d38bf2
* adapt compat_kex_proposal() test to portable
* update version in README
* crank versions in RPM specs
* remove files from libssh project
* re-merge arc4random.c
* re-merge misc.c
* remove unused files from libssh.vcxproj
* fix outstanding merge conflicts
* fix build errors
* modify upstream workflows to trigger on workflow dispatch instead of all PRs
* fix scp client hanging with pipes
* fix some failing bash tests
* make bash test compatible with Windows
* address scp's sftp mode buf len limitations
* address review feedback
* address review feedback
* update comment
---------
Signed-off-by: Sam James <sam@gentoo.org>
Co-authored-by: djm@openbsd.org <djm@openbsd.org>
Co-authored-by: Darren Tucker <dtucker@dtucker.net>
Co-authored-by: Damien Miller <djm@mindrot.org>
Co-authored-by: Sam James <sam@gentoo.org>
Co-authored-by: jsg@openbsd.org <jsg@openbsd.org>
Co-authored-by: jmc@openbsd.org <jmc@openbsd.org>
Co-authored-by: dtucker@openbsd.org <dtucker@openbsd.org>
Co-authored-by: Harmen Stoppels <harmenstoppels@gmail.com>
Co-authored-by: Rochdi Nassah <rochdinassah.1998@gmail.com>
Co-authored-by: David Korczynski <david@adalogics.com>
Co-authored-by: Pierre Ossman <ossman@cendio.se>
Co-authored-by: mbuhl@openbsd.org <mbuhl@openbsd.org>
Co-authored-by: Rose <83477269+AtariDreams@users.noreply.github.com>
Co-authored-by: cheloha@openbsd.org <cheloha@openbsd.org>
Co-authored-by: deraadt@openbsd.org <deraadt@openbsd.org>
Co-authored-by: tb@openbsd.org <tb@openbsd.org>
Co-authored-by: millert@openbsd.org <millert@openbsd.org>
auth_debug_add queues messages about the auth process which is sent to
the client after successful authentication. This also sends those to
the server debug log to aid in debugging. From bz#3507, ok djm@
OpenBSD-Commit-ID: 46ff67518cccf9caf47e06393e2a121ee5aa258a
* upstream: fix poll() spin when a channel's output fd closes without
data in the channel buffer. Introduce more exact packing of channel fds into
the pollfd array. fixes bz3405 and bz3411; ok deraadt@ markus@
OpenBSD-Commit-ID: 06740737849c9047785622ad5d472cb6a3907d10
* upstream: select post-quantum KEX
sntrup761x25519-sha512@openssh.com as the default; ok markus@
OpenBSD-Commit-ID: f02d99cbfce22dffec2e2ab1b60905fbddf48fb9
* upstream: add support for the "corp-data" protocol extension to
allow server-side copies to be performed without having to go via the client.
Patch by Mike Frysinger, ok dtucker@
OpenBSD-Commit-ID: 00aa510940fedd66dab1843b58682de4eb7156d5
* upstream: add a sftp client "cp" command that supports server-side
copying of files. Useful for this task and for testing the copy-data
extension. Patch from Mike Frysinger; ok dtucker@
OpenBSD-Commit-ID: 1bb1b950af0d49f0d5425b1f267e197aa1b57444
* depend
* Skip slow tests on (very) slow test targets.
* Set Makefile SHELL as determined by configure.
This should improve compatibility for users with non-POSIX shells. If
using Makefile.in directly (eg make -f Makefile.in distprep) then SHELL
will need to be specified on the command line (along with MANFMT in that
particular case). ok djm@
* Use bash or ksh if available for SH in Makefile.
* Increase test timeout to allow slow VMs to finish
* Only run regression tests on slow VMs.
* Only return events from ppoll that were requested.
If the underlying system's select() returns bits that were not in the
request set, our ppoll() implementation can return revents for events
not requested, which can apparently cause a hang. Only return revents
for activity in the requested event set. bz#3416, analysis and fix by
yaroslav.kuzmin at vmssoftware com, ok djm@
* Specify TEST_SHELL=bash on AIX.
The system shells cause the agent-restrict test to fail due to some
quoting so explicitly specify bash until we can get configure to
autmatically work around that.
* Disable security key on fbsd6 test host.
* upstream: man pages: add missing commas between subordinate and
main clauses
jmc@ dislikes a comma before "then" in a conditional, so leave those
untouched.
ok jmc@
OpenBSD-Commit-ID: 9520801729bebcb3c9fe43ad7f9776ab4dd05ea3
* upstream: ssh: document sntrup761x25519-sha512@openssh.com as
default KEX
OpenBSD-Commit-ID: 12545bfa10bcbf552d04d9d9520d0f4e98b0e171
* upstream: openssh-9.0
OpenBSD-Commit-ID: 0dfb461188f4513ec024c1534da8c1ce14c20b64
* update version numbers for release
* update build-aux files to match autoconf-2.71
i.e. config.guess, config.sub and install-sh
* Revert "update build-aux files to match autoconf-2.71"
This reverts commit 0a8ca39fac.
It turns out that the checked-in copies of these files are actually newer
than autoconf-2.71's copies, so this was effectively a downgrade.
Spotted by Bo Anderson via github
* upstream: two defensive changes from Tobias Stoeckmann via GHPR287
enforce stricter invarient for sshbuf_set_parent() - never allow
a buffer to have a previously-set parent changed.
In sshbuf_reset(), if the reallocation fails, then zero the entire
buffer and not the (potentially smaller) default initial alloc size.
OpenBSD-Commit-ID: 14583203aa5d50ad38d2e209ae10abaf8955e6a9
* upstream: Note that curve25519-sha256 was later published in
RFC8731. ok djm@
OpenBSD-Commit-ID: 2ac2b5d642d4cf5918eaec8653cad9a4460b2743
* upstream: clear io_want/io_ready flags at start of poll() cycle;
avoids plausible spin during rekeying if channel io_want flags are reused
across cycles. ok markus@ deraadt@
OpenBSD-Commit-ID: 91034f855b7c73cd2591657c49ac30f10322b967
* Retire fbsd6 test VM.
It's long since out of support, relatively slow (it's i686) and the
compiler has trouble with PIE.
* Resync moduli.5 with upstream.
1.18: remove duplicate publication year; carsten dot kunze at arcor dot de
1.19: ssh-keygen's -G/-T have been replaced with -M generate/screen.
* upstream: Correct path for system known hosts file in description
of IgnoreUserKnownHosts. Patch from Martin Vahlensieck via tech@
OpenBSD-Commit-ID: 9b7784f054fa5aa4d63cb36bd563889477127215
* upstream: list the correct version number
for when usage of the sftp protocol became default and fix a typo
from ed maste
OpenBSD-Commit-ID: 24e1795ed2283fdeacf16413c2f07503bcdebb31
* upstream: Import regenerated moduli
OpenBSD-Commit-ID: f9a0726d957cf10692a231996a1f34e7f9cdfeb0
* upstream: Try to continue running local I/O for channels in state
OPEN during SSH transport rekeying. The most visible benefit is that it
should make ~-escapes work in the client (e.g. to exit) if the connection
happened to have stalled during a rekey event. Based work by and ok dtucker@
OpenBSD-Commit-ID: a66e8f254e92edd4ce09c9f750883ec8f1ea5f45
* upstream: Import regenerated moduli
OpenBSD-Commit-ID: f9a0726d957cf10692a231996a1f34e7f9cdfeb0
* upstream: regression test for sftp cp command
OpenBSD-Regress-ID: c96bea9edde3a384b254785e7f9b2b24a81cdf82
* upstream: Simplify forward-control test.
Since we no longer need to support SSH1 we don't need to run shell
commands on the other end of the connection and can use ssh -N instead.
This also makes the test less racy.
OpenBSD-Regress-ID: 32e94ce272820cc398f30b848b2b0f080d10302c
* upstream: Use ssh -f and ControlPersist ..
to start up test forwards and ssh -O stop to shut them down intead of
sleep loops. This speeds up the test by an order of magnitude.
OpenBSD-Regress-ID: eb3db5f805100919b092a3b2579c611fba3e83e7
* upstream: It looks like we can't completely avoid
waiting for processes to exit so retrieve the pid via controlmaster and
use that.
OpenBSD-Regress-ID: 8246f00f22b14e49d2ff1744c94897ead33d457b
* Cache timezone data in capsicum sandbox.
From emaste at freebsd.org, originally part of FreeBSD commit r339216
/ fc3c19a9 with autoconf bits added by me.
* Include stdlib.h for free() prototype.
... which is used inside the CUSTOM_SYS_AUTH_GET_LASTLOGIN_MSG block.
* Update OpenSSL and LibreSSL versions in tests.
* Add debian-riscv64 test target.
* upstream: Avoid an unnecessary xstrdup in rm_env() when matching
patterns. Since match_pattern() doesn't modify its arguments (they are
const), there is no need to make an extra copy of the strings in
options->send_env. From Martin Vahlensieck
OpenBSD-Commit-ID: 2c9db31e3f4d3403b49642c64ee048b2a0a39351
* upstream: Add missing includes of stdlib.h and stdint.h. We need
stdlib.h for malloc(3) and stdint.h for SIZE_MAX. Unlike the other xmss
files, ssh-xmss.c does not include xmss_commons.h so ssh-xmss.c must include
those headers itself. From Martin Vahlensieck
OpenBSD-Commit-ID: 70e28a9818cee3da1be2ef6503d4b396dd421e6b
* upstream: Remove unnecessary includes: openssl/hmac.h and
openssl/evp.h. From Martin Vahlensieck.
OpenBSD-Commit-ID: a6debb5fb0c8a44e43e8d5ca7cc70ad2f3ea31c3
* upstream: Check sshauthopt_new() for NULL. bz#3425, from
tessgauthier at microsoft.com. ok djm@
OpenBSD-Commit-ID: af0315bc3e44aa406daa7e0ae7c2d719a974483f
* upstream: Add authfd path to debug output. ok markus@
OpenBSD-Commit-ID: f735a17d1a6f2bee63bfc609d76ef8db8c090890
* upstream: avoid printing hash algorithm twice; from lucas AT sexy.is
OpenBSD-Commit-ID: 9d24671e10a84141b7c504396cabad600e47a941
* upstream: fix memleak on session-bind path; from Pedro Martelletto, ok
dtucker@
OpenBSD-Commit-ID: e85899a26ba402b4c0717b531317e8fc258f0a7e
* upstream: Don't leak SK device. Patch from Pedro Martelletto via
github PR#316. ok djm@
OpenBSD-Commit-ID: 17d11327545022e727d95fd08b213171c5a4585d
* upstream: mention that the helpers are used by ssh(1), ssh-agent(1)
and ssh-keygen(1). Previously only ssh(1) was mentioned. From Pedro
Martelletto
OpenBSD-Commit-ID: 30f880f989d4b329589c1c404315685960a5f153
* Remove now-empty int32_minmax.inc.
* Only run tests when source files change.
Also run tests on changes to V_9_0 branch.
* Add Mac OS X 12 test target.
* upstream: be stricter in which characters will be accepted in
specifying a mask length; allow only 0-9. From khaleesicodes via GHPR#278; ok
dtucker@
OpenBSD-Commit-ID: e267746c047ea86665cdeccef795a8a56082eeb2
* upstream: fix some integer overflows in sieve_large() that show up when
trying to generate modp groups > 16k bits. Reported via GHPR#306 by Bertram
Felgenhauer, but fixed in a different way. feedback/ok tb@
OpenBSD-Commit-ID: 81cbc6dd3a21c57bd6fadea10e44afe37bca558e
* upstream: remove an obsolete rsa1 format example from an example;
from megan batty
ok djm
OpenBSD-Commit-ID: db2c89879c29bf083df996bd830abfb1e70d62bf
* upstream: Add FIDO AUTHENTICATOR section and explain a bit how FIDO
works. The wording came mostly from the 8.2 OpenSSH release notes, addapted
to fit the man page. Then move the -O bits into the new section as is already
done for CERTIFICATES and MODULI GENERATION. Finally we can explain the
trade-offs of resident keys. While here, consistently refer to the FIDO
thingies as "FIDO authenticators", not "FIDO tokens".
input & OK jmc, naddy
OpenBSD-Commit-ID: dd98748d7644df048f78dcf793b3b63db9ab1d25
* upstream: make sure stdout is non-blocking; ok djm@
OpenBSD-Commit-ID: 64940fffbd1b882eda2d7c8c7a43c79368309c0d
* upstream: mux.c: mark argument as const; from Martin Vahlensieck
OpenBSD-Commit-ID: 69a1a93a55986c7c2ad9f733c093b46a47184341
* upstream: channel_new no longer frees remote_name. So update the
comment accordingly. As remote_name is not modified, it can be const as
well. From Martin Vahlensieck
OpenBSD-Commit-ID: e4e10dc8dc9f40c166ea5a8e991942bedc75a76a
* upstream: sshkey_unshield_private() contains a exact duplicate of
the code in private2_check_padding(). Pull private2_check_padding() up so the
code can be reused. From Martin Vahlensieck, ok deraadt@
OpenBSD-Commit-ID: 876884c3f0e62e8fd8d1594bab06900f971c9c85
* Add ubsan minimal testcase on OpenBSD.
As suggested by djm@.
* Note that, for now, we need variadic macros.
* Also retest OpenBSD upstream on .yml changes.
* upstream: When performing operations that glob(3) a remote path, ensure
that the implicit working directory used to construct that path escapes
glob(3) characters.
This prevents glob characters from being processed in places they
shouldn't, e.g. "cd /tmp/a*/", "get *.txt" should have the get operation
treat the path "/tmp/a*" literally and not attempt to expand it.
Reported by Lusia Kundel; ok markus@
OpenBSD-Commit-ID: 4f647f58482cbad3d58b1eab7f6a1691433deeef
* Remove duplicate bcrypt_pbkdf.o from Makefile
bcrypt_pbkdf.o is duplicated in the openbsd-compat Makefile's object
file list.
* upstream: improve error message when 'ssh-keygen -Y sign' is unable to
load a private key; bz3429, reported by Adam Szkoda ok dtucker@
OpenBSD-Commit-ID: bb57b285e67bea536ef81b1055467be2fc380e74
* upstream: Allow existing -U (use agent) flag to work with "-Y sign"
operations, where it will be interpreted to require that the private keys is
hosted in an agent; bz3429, suggested by Adam Szkoda; ok dtucker@
OpenBSD-Commit-ID: a7bc69873b99c32c42c7628ed9ea91565ba08c2f
* upstream: Remove errant apostrophe. From haruyama at queen-ml org.
OpenBSD-Commit-ID: dc6b294567cb84b384ad6ced9ca469f2bbf0bd10
* upstream: arrange for scp, when in sftp mode, to not ftruncate(3) files
early
previous behavious of unconditionally truncating the destination file
would cause "scp ~/foo localhost:" and "scp localhost:foo ~/" to
delete all the contents of their destination.
spotted by solene@ sthen@, also bz3431; ok dtucker@
OpenBSD-Commit-ID: ca39fdd39e0ec1466b9666f15cbcfddea6aaa179
* upstream: fix in-place copies; r1.163 incorrectly skipped truncation in
all cases, not just at the start of a transfer. This could cause overwrites
of larger files to leave junk at the end. Spotted by tb@
OpenBSD-Commit-ID: b189f19cd68119548c8e24e39c79f61e115bf92c
* upstream: Only run agent-ptrace.sh if gdb is available as all
architectures do not ship with gdb.
OpenBSD-Regress-ID: ec53e928803e6b87f9ac142d38888ca79a45348d
* upstream: regress test for in-place transfers and clobbering larger
files with smaller ones; would have caught last regression in scp(1)
OpenBSD-Regress-ID: 19de4e88dd3a4f7e5c1618c9be3c32415bd93bc2
* configure.ac: Add missing AC_DEFINE for caph_cache_tzdata test causing
HAVE_CAPH_CACHE_TZDATA to be missing from config.h.in.
Spotted by Bryan Drewery
* upstream: make SSHBUF_DBG/SSHBUF_TELL (off by default and only enabled
via #define) dump to stderr rather than stdout
OpenBSD-Commit-ID: 10298513ee32db8390aecb0397d782d68cb14318
* upstream: revert previous; it was broken (spotted by Theo)
OpenBSD-Commit-ID: 457c79afaca2f89ec2606405c1059b98b30d8b0d
* upstream: Note that ProxyJump also accepts the same tokens as
ProxyCommand. From pallxk via github PR#305.
OpenBSD-Commit-ID: 7115ac351b129205f1f1ffa6bbfd62abd76be7c5
* upstream: Avoid kill with -1 argument. The out_ctx label can be
reached before fork has been called. If this happens, then kill -1 would be
called, sending SIGTERM to all processes reachable by the current process.
From tobias@ and c3h2_ctf via github PR#286, ok djm@
OpenBSD-Commit-ID: 6277af1207d81202f5daffdccfeeaed4c763b1a8
* upstream: f sshpkt functions fail, then password is not cleared
with freezero. Unconditionally call freezero to guarantee that password is
removed from RAM.
From tobias@ and c3h2_ctf via github PR#286, ok djm@
OpenBSD-Commit-ID: 6b093619c9515328e25b0f8093779c52402c89cd
* upstream: refactor authorized_keys/principals handling
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]
ok dtucker@
OpenBSD-Commit-ID: 9d4373d013edc4cc4b5c21a599e1837ac31dda0d
* 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.
ok dtucker@
OpenBSD-Commit-ID: 3fdca2c61ad97dc1b8d4a7346816f83dc4ce2217
* fuzzer for authorized_keys parsing
mostly redundant to authopt_fuzz, but it's sensitive code so IMO it
makes sense to test this layer too
* Test against LibreSSL 3.5.3.
* Test against OpenSSL 1.1.1o and 3.0.3.
* fix some bugs in the fuzzer
* upstream: keywords ref ssh_config.5;
from caspar schutijser
OpenBSD-Commit-ID: f146a19d7d5c9374c3b9c520da43b2732d7d1a4e
* upstream: ssh-keygen: implement "verify-required" certificate option.
This was already documented when support for user-verified FIDO
keys was added, but the ssh-keygen(1) code was missing.
ok djm@
OpenBSD-Commit-ID: f660f973391b593fea4b7b25913c9a15c3eb8a06
* upstream: ssh-keygen -A: do not generate DSA keys by default.
Based on github PR#303 from jsegitz with man page text from jmc@, ok markus@
djm@
OpenBSD-Commit-ID: 5c4c57bdd7063ff03381cfb6696659dd3f9f5b9f
* upstream: Add period at end of "not known by any other names"
message. github PR#320 from jschauma, ok djm@
OpenBSD-Commit-ID: bd60809803c4bfd3ebb7c5c4d918b10e275266f2
* upstream: Add missing *-sk types to ssh-keyscan manpage. From
skazi0 via github PR#294.
OpenBSD-Commit-ID: fda2c869cdb871f3c90a89fb3f985370bb5d25c0
* 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.
bz3438, ok dtucker
OpenBSD-Commit-ID: 797909c1e0262c0d00e09280459d7ab00f18273b
* upstream: test setenv in both client and server, test first-match-wins
too
OpenBSD-Regress-ID: 4c8804f9db38a02db480b9923317457b377fe34b
* upstream: move auth_openprincipals() and auth_openkeyfile() over to
auth2-pubkeyfile.c too; they make more sense there.
OpenBSD-Commit-ID: 9970d99f900e1117fdaab13e9e910a621b7c60ee
* upstream: make sure that UseDNS hostname lookup happens in the monitor
and not in the pledge(2)'d unprivileged process; fixes regression caused by
recent refactoring spotted by henning@
OpenBSD-Commit-ID: a089870b95101cd8881a2dff65b2f1627d13e88d
* fix possible NULL deref when built without FIDO
Analysis/fix from kircher in bz3443; ok dtucker@
* automatically enable built-in FIDO support
If libfido2 is found and usable, then enable the built-in
security key support unless --without-security-key-builtin
was requested.
ok dtucker@
* upstream: Log an error if pipe() fails while accepting a
connection. bz#3447, from vincent-openssh at vinc17 net, ok djm@
OpenBSD-Commit-ID: 9d59f19872b94900a5c79da2d57850241ac5df94
* upstream: Don't attempt to fprintf a null identity comment. From
Martin Vahlensieck via tech@.
OpenBSD-Commit-ID: 4c54d20a8e8e4e9912c38a7b4ef5bfc5ca2e05c2
* upstream: Make sure not to fclose() the same fd twice in case of an
error.
ok dtucker@
OpenBSD-Commit-ID: e384c4e05d5521e7866b3d53ca59acd2a86eef99
* upstream: make it clear that RekeyLimit applies to both transmitted
and received data. GHPR#328 from Jan Pazdziora
OpenBSD-Commit-ID: d180a905fec9ff418a75c07bb96ea41c9308c3f9
* request 1.1x API compatibility for OpenSSL >=3.x
idea/patch from Pedro Martelletto via GHPR#322; ok dtucker@
* fix broken case statement in previous
* Disable SK support if FIDO libs not found.
* Zero out LIBFIDO2 when SK support not usable.
Prevents us from trying to link them into ssh-sk-helper and failing to
build.
* upstream: Don't leak the strings allocated by order_hostkeyalgs()
and list_hostkey_types() that are passed to compat_pkalg_proposal(). Part of
github PR#324 from ZoltanFridrich, ok djm@
OpenBSD-Commit-ID: b2f6e5f60f2bba293b831654328a8a0035ef4a1b
* upstream: Roll back previous KEX changes as they aren't safe until
compat_pkalg_proposal and friends always allocate their returned strings.
Reported by Qualys.
OpenBSD-Commit-ID: 1c7a88a0d5033f42f88ab9bec58ef1cf72c81ad0
* upstream: allow arguments to sftp -D option, e.g. sftp -D
"/usr/libexec/sftp-server -el debug3"
ok markus@
OpenBSD-Commit-ID: 5a002b9f3a7aef2731fc0ffa9c921cf15f38ecce
* Update OpenSSL tests to the most recent releases.
* upstream: reflect the update to -D arg name in usage();
OpenBSD-Commit-ID: abdcde4f92b1ef094ae44210ee99d3b0155aad9c
* upstream: ignore SIGPIPE earlier in main(), specifically before
muxclient() which performs operations that could cause one; Reported by Noam
Lewis via bz3454, ok dtucker@
OpenBSD-Commit-ID: 63d8e13276869eebac6d7a05d5a96307f9026e47
* upstream: Always return allocated strings from the kex filtering so
that we can free them later. Fix one leak in compat_kex_proposal. Based on
github PR#324 from ZoltanFridrich with some simplications by me. ok djm@
OpenBSD-Commit-ID: 9171616da3307612d0ede086fd511142f91246e4
* upstream: Don't leak the strings allocated by order_hostkeyalgs()
and list_hostkey_types() that are passed to compat_pkalg_proposal(). Part of
github PR#324 from ZoltanFridrich, ok djm@
This is a roll-forward of the previous rollback now that the required
changes in compat.c have been done.
OpenBSD-Commit-ID: c7cd93730b3b9f53cdad3ae32462922834ef73eb
* upstream: bump up loglevel from debug to info when unable to open
authorized keys/principals file for errno != ENOENT; bz2042 ok dtucker
OpenBSD-Commit-ID: e79aa550d91ade6a80f081bda689da24c086d66b
* Skip select+rlimit check if sandboxing is disabled
It's not needed in that case, and the test can fail when being built
with some compiler memory sanitizer flags. bz#3441
* upstream: use consistent field names (s/char/byte)
in format description
OpenBSD-Commit-ID: 3de33572733ee7fcfd7db33d37db23d2280254f0
* upstream: Remove leftover line.
Remove extra line leftover from merge conflict. ok djm@
OpenBSD-Commit-ID: 460e2290875d7ae64971a7e669c244b1d1c0ae2e
* 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.
* Skip all rlimit tests when sandboxing disabled.
The rlimit tests can hang when being run with some compiler sanitizers
so skip all of them if sandbox=no.
* Add clang sanitizer tests.
* upstream: Add TEST_REGRESS_CACHE_DIR.
If set, it is used to cache regress test names that have succeeded and
skip those on a re-run.
OpenBSD-Regress-ID: a7570dd29a58df59f2cca647c3c2ec989b49f247
* Move sanitizer logs into regress for collection.
* Add GCC address sanitizer build/test.
* Update sanitizer test targets:
- remove clang-sanitize-memory for now. It takes so long that the test
times out.
- add gcc sanitize-address and sanitize-undefined test targets.
* Test against openssl-3.0.5.
* Move unset to before we set anything.
* Refuse to use OpenSSL 3.0.4 due to potential RCE.
OpenSSL has a potential RCE in its RSA implementation (CVE-2022-2274)
so refuse to use that specific version.
* Capture stderr output from configure.
* Only refuse to use OpenSSL 3.0.4 on x86_64.
The potential RCE only impacts x86_64, so only refuse to use it if we're
targetting a potentially impacted architecture. ok djm@
* 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@
* 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.
* 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.
* 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@
* Remove unintended changes.
I inadvertently included a couple of local changes with the OpenSSL
3.0.4 change. Revert, anything that should be there will be committed
separately.
* Add AUDIT_ARCH_PPC to supported seccomp arches.
Patch from dries.deschout at dodeco.eu.
* Rename bbone test target to ARM.
* 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.
* upstream: pull passphrase reading and confirmation into a separate
function so it can be used for FIDO2 PINs; no functional change
OpenBSD-Commit-ID: bf34f76b8283cc1d3f54633e0d4f13613d87bb2f
* 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
OpenBSD-Commit-ID: e45e9f1bf2b2f32d9850669e7a8dbd64acc5fca4
* upstream: sk-usbhid: preserve error code returned by key_lookup()
it conveys useful information, such as the supplied pin being wrong.
Part of GHPR329 from Pedro Martelletto
OpenBSD-Commit-ID: c0647eb9290f793add363d81378439b273756c1b
* upstream: ssh-keygen: fix touch prompt, pin retries;
part of GHPR329 from Pedro Martelletto
OpenBSD-Commit-ID: 75d1005bd2ef8f29fa834c90d2684e73556fffe8
* crank SSH_SK_VERSION_MAJOR in sk-dummy.so
* Skip scp3 test if there's no scp on remote path.
scp -3 ends up using the scp that's in the remote path and will fail if
one is not available. Based on a patch from rapier at psc.edu.
* 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: Test TEST_SSH_ELAPSED_TIMES for empty string not
executable. No-op on most platforms but should prevent warnings in -portable
on systems that don't have 'date %s'.
OpenBSD-Regress-ID: e39d79867b8065e33d0c5926fa1a31f85659d2a4
* upstream: Restore missing "!" in TEST_SSH_ELAPSED_TIMES test.
OpenBSD-Regress-ID: 38783f9676ec348c5a792caecee9a16e354b37b0
* 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@
* 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@
* 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@
* Group libcrypto and PRNGD checks together.
They're related more than the libcrypt or libiaf checks which are
currently between them. ok djm@
* Remove seed_rng calls from scp, sftp, sftp-server.
These binaries don't use OpenSSL's random functions. The next step
will be to stop linking them against libcrypto. ok djm@
* Move libcrypto into CHANNELLIBS.
This will result in sftp, sftp-server and scp no longer being linked
against libcrypto. ok djm@
* 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@
* Remove deprecated MacOS 10.15 runners.
* upstream: avoid double-free in error path introduced in r1.70; report
and fix based on GHPR#332 by v-rzh ok dtucker@
OpenBSD-Commit-ID: 3d21aa127b1f37cfc5bdc21461db369a663a951f
* Include CHANNEL and FIDO2 libs in configure output
* 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@.
* compat code for fido_dev_is_winhello()
Signed-off-by: Corinna Vinschen <vinschen@redhat.com>
* check_sk_options: add temporary WinHello workaround
Up to libfido 1.10.0, WinHello advertises "clientPin" rather
than "uv" capability. This is fixed in 1.11.0. For the time
being, workaround it here.
Signed-off-by: Corinna Vinschen <vinschen@redhat.com>
* 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.
Signed-off-by: Corinna Vinschen <vinschen@redhat.com>
* upstream: don't prompt for FIDO passphrase before attempting to enroll
the credential, just let the enroll operating fail and we'll attempt to get a
PIN anyway. Might avoid some unneccessary PIN prompts.
Part of GHPR#302 from Corinna Vinschen; ok dtucker@
OpenBSD-Commit-ID: bd5342ffc353ee37d39617906867c305564d1ce2
* Give unused param a name.
Fixes builds on platforms that do have fido2 but don't have
fido_dev_is_winhello.
* Actually put HAVE_STDINT_H around the stdint.h.
* 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@
* Test hostbased auth on github runners.
* fix SANDBOX_SECCOMP_FILTER_DEBUG
* Fix conditional for running hostbased tests.
* upstream: allow certificate validity intervals, sshsig verification
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.
bz3468 ok dtucker
OpenBSD-Commit-ID: 454db1cdffa9fa346aea5211223a2ce0588dfe13
* upstream: add some tests for parse_absolute_time(), including cases
where it is forced to the UTC timezone. bz3468 ok dtucker
OpenBSD-Regress-ID: ea07ca31c2f3847a38df028ca632763ae44e8759
* 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).
* Rerun tests if any .github config file changes.
* Add a timegm implementation from Heimdal via Samba.
Fixes build on (at least Solaris 10).
* Replace deprecated ubuntu-18.04 runners with 22.04
* 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.
From Mike Frysinger, ok dtucker@
OpenBSD-Commit-ID: bfc580d05cc0c817831ae7ecbac4a481c23566ab
* fido_dev_is_winhello: return 0, not "false"
"false" is not used anywhere in OpenSSH, so return 0 like
everywhere else.
Signed-off-by: Corinna Vinschen <vinschen@redhat.com>
* Revert "check_sk_options: add temporary WinHello workaround"
Cygwin now comes with libfido2 1.11.0, so this workaround
isn't required anymore.
This reverts commit 242c044ab1.
Signed-off-by: Corinna Vinschen <vinschen@redhat.com>
* upstream: use .Cm for "sign"; from josiah frentsos
OpenBSD-Commit-ID: 7f80a53d54857ac6ae49ea6ad93c5bd12231d1e4
* upstream: add an extra flag to sk_probe() to indicate whether we're
probing for a FIDO resident key or not. Unused here, but will make like
easier for portable
OpenBSD-Commit-ID: 432c8ff70e270378df9dbceb9bdeaa5b43b5a832
* 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.
* Check for perms to run agent-getpeereid test.
Ubuntu 22.04 defaults to private home dirs which prevents "nobody"
running ssh-add during the agent-getpeereid test. Check for this and
add the necessary permissions.
* upstream: double free() in error path; from Eusgor via GHPR333
OpenBSD-Commit-ID: 39f35e16ba878c8d02b4d01d8826d9b321be26d4
* Add Cygwin (on windows-2019) test target.
In addition to installing the requisite Cygwin packages, we also need to
explicitly invoke "sh" for steps that run other scripts since the runner
environment doesn't understand #! paths.
* Add a bit more debug output.
* Fix cygwin conditional steps.
* upstream: Strictly enforce the maximum allowed SSH2 banner size in
ssh-keyscan and prevent a one-byte buffer overflow. Patch from Qualys, ok
djm@
OpenBSD-Commit-ID: 6ae664f9f4db6e8a0589425f74cd0bbf3aeef4e4
* upstream: remove incorrect check that can break enrolling a
resident key (introduced in r1.40)
OpenBSD-Commit-ID: 4cab364d518470e29e624af3d3f9ffa9c92b6f01
* upstream: attemp FIDO key signing without PIN and use the error
code returned to fall back only if necessary. Avoids PIN prompts for FIDO
tokens that don't require them; part of GHPR#302
OpenBSD-Commit-ID: 4f752aaf9f2e7c28bcaaf3d4f8fc290131bd038e
* Install Cygwin packages based on OS not config.
* initial list of allowed signers
* upstream: whitespace
OpenBSD-Commit-ID: d297e4387935d4aef091c5e9432578c2e513f538
* upstream: whitespace
OpenBSD-Commit-ID: a5d015efbfd228dc598ffdef612d2da3a579e5d8
* Add cygwin-release test target.
This also moves the cygwin package install from the workflow file to
setup_ci.sh so that we can install different sets of Cygwin packages
for different test configs.
* Add Windows 2022 test targets.
* Add libcrypt-devel to cygwin-release deps.
Based on feedback from vinschen at redhat.com.
* cross-sign allowed_signers with PGP key
Provides continuity of trust from legacy PGP release key to
the SSHSIG signing keys that we will use henceforth for git
signing.
* additional keys
* upstream: whitespace
OpenBSD-Commit-ID: c2bcbf93610d3d62ed206cdf9bf9ff98c6aaf232
* Move sftp from valgrind-2 to 3 to rebalance.
* upstream: sk-usbhid: fix key_lookup() on tokens with built-in UV
explicitly test whether the token performs built-in UV (e.g. biometric
tokens) and enable UV in that case. From Pedro Martelletto via GHPR#388
OpenBSD-Commit-ID: 007eb7e387d27cf3029ab06b88224e03eca62ccd
* Remove arc4random_uniform from arc4random.c
This was previously moved into its own file (matching OpenBSD) which
prematurely committed in commit 73541f2.
* Move OPENBSD ORIGINAL marker.
Putting this after the copyright statement (which doesn't change)
instead of before the version identifier (which does) prevents merge
conflicts when resyncing changes.
* Resync arc4random with OpenBSD.
This brings us up to current, including djm's random-reseeding change,
as prompted by logan at cyberstorm.mu in bz#3467. It brings the
platform-specific hooks from LibreSSL Portable, simplified to match our
use case. ok djm@.
* Remove DEF_WEAK, it's already in defines.h.
* openbsd-compat/bsd-asprintf: add <stdio.h> include for vsnprintf
Fixes the following build failure with Clang 15 on musl:
```
bsd-asprintf.c:51:8: error: call to undeclared library function 'vsnprintf' with type 'int (char *, unsigned long, const char *, struct __va_list_tag *)'; ISO C99 and laterclang -O2 -pipe -fdiagnostics-color=always -frecord-gcc-switches -pipe -Wunknown-warning-option -Qunused-arguments -Wall -Wpointer-arith -Wuninitialized -Wsign-compare -Wformat-security -Wsizeof-pointer-memaccess -Wno-pointer-sign -Wno-unused-result -Wmisleading-indentation -Wbitwise-instead-of-logical -fno-strict-aliasing -mretpoline -ftrapv -fzero-call-used-regs=all -fno-builtin-memset -fstack-protector-strong -fPIE -I. -I. -D_XOPEN_SOURCE=600 -D_BSD_SOURCE -D_DEFAULT_SOURCE -DSSHDIR=\"/etc/ssh\" -D_PATH_SSH_PROGRAM=\"/usr/bin/ssh\" -D_PATH_SSH_ASKPASS_DEFAULT=\"/usr/lib/misc/ssh-askpass\" -D_PATH_SFTP_SERVER=\"/usr/lib/misc/sftp-server\" -D_PATH_SSH_KEY_SIGN=\"/usr/lib/misc/ssh-keysign\" -D_PATH_SSH_PKCS11_HELPER=\"/usr/lib/misc/ssh-pkcs11-helper\" -D_PATH_SSH_SK_HELPER=\"/usr/lib/misc/ssh-sk-helper\" -D_PATH_SSH_PIDDIR=\"/run\" -D_PATH_PRIVSEP_CHROOT_DIR=\"/var/empty\" -DHAVE_CONFIG_H -c cipher-aes.c -o cipher-aes.o
do not support
implicit function declarations [-Wimplicit-function-declaration]
ret = vsnprintf(string, INIT_SZ, fmt, ap2);
^
bsd-asprintf.c:51:8: note: include the header <stdio.h> or explicitly provide a declaration for 'vsnprintf'
1 error generated.
```
* upstream: notifier_complete(NULL, ...) is a noop, so no need to test
that ctx!=NULL; from Corinna Vinschen
OpenBSD-Commit-ID: ade2f2e9cc519d01a586800c25621d910bce384a
* fix pester test failures
* upstream: fix repeated words ok miod@ jmc@
OpenBSD-Commit-ID: 6765daefe26a6b648cc15cadbbe337596af709b7
* upstream: .Li -> .Vt where appropriate; from josiah frentsos,
tweaked by schwarze
ok schwarze
OpenBSD-Commit-ID: 565046e3ce68b46c2f440a93d67c2a92726de8ed
* upstream: ssh-agent: attempt FIDO key signing without PIN and use
the error to determine whether a PIN is required and prompt only if
necessary. from Corinna Vinschen
OpenBSD-Commit-ID: dd6be6a0b7148608e834ee737c3479b3270b00dd
* upstream: a little extra debugging
OpenBSD-Commit-ID: edf1601c1d0905f6da4c713f4d9cecc7d1c0295a
* upstream: sk_enroll: never drop SSH_SK_USER_VERIFICATION_REQD flag
from response
Now that all FIDO signing calls attempt first without PIN and then
fall back to trying PIN only if that attempt fails, we can remove the
hack^wtrick that removed the UV flag from the keys returned during
enroll.
By Corinna Vinschen
OpenBSD-Commit-ID: 684517608c8491503bf80cd175425f0178d91d7f
* define HAVE_KILLPG
* upstream: sftp: Don't attempt to complete arguments for
non-existent commands
If user entered a non-existent command (e.g. because they made a
typo) there is no point in trying to complete its arguments. Skip
calling complete_match() if that's the case.
From Michal Privoznik
OpenBSD-Commit-ID: cf39c811a68cde2aeb98fc85addea4000ef6b07a
* upstream: sftp: Be a bit more clever about completions
There are commands (e.g. "get" or "put") that accept two
arguments, a local path and a remote path. However, the way
current completion is written doesn't take this distinction into
account and always completes remote or local paths.
By expanding CMD struct and "cmds" array this distinction can be
reflected and with small adjustment to completer code the correct
path can be completed.
By Michal Privoznik, ok dtucker@
OpenBSD-Commit-ID: 1396d921c4eb1befd531f5c4a8ab47e7a74b610b
* upstream: correct error value
OpenBSD-Commit-ID: 780efcbad76281f11f14b2a5ff04eb6db3dfdad4
* upstream: actually hook up restrict_websafe; the command-line flag
was never actually used. Spotted by Matthew Garrett
OpenBSD-Commit-ID: 0b363518ac4c2819dbaa3dfad4028633ab9cdff1
* upstream: Add a sshkey_check_rsa_length() call for checking the
length of an RSA key; ok markus@
OpenBSD-Commit-ID: de77cd5b11594297eda82edc594b0d32b8535134
* upstream: add a RequiredRSASize for checking RSA key length in
ssh(1). User authentication keys that fall beneath this limit will be
ignored. If a host presents a host key beneath this limit then the connection
will be terminated (unfortunately there are no fallbacks in the protocol for
host authentication).
feedback deraadt, Dmitry Belyavskiy; ok markus@
OpenBSD-Commit-ID: 430e339b2a79fa9ecc63f2837b06fdd88a7da13a
* upstream: Add RequiredRSASize for sshd(8); RSA keys that fall
beneath this limit will be ignored for user and host-based authentication.
Feedback deraadt@ ok markus@
OpenBSD-Commit-ID: 187931dfc19d51873df5930a04f2d972adf1f7f1
* upstream: better debugging for connect_next()
OpenBSD-Commit-ID: d16a307a0711499c971807f324484ed3a6036640
* upstream: sftp-server(8): add a "users-groups-by-id@openssh.com"
extension request that allows the client to obtain user/group names that
correspond to a set of uids/gids.
Will be used to make directory listings more useful and consistent
in sftp(1).
ok markus@
OpenBSD-Commit-ID: 7ebabde0bcb95ef949c4840fe89e697e30df47d3
* upstream: extend sftp-common.c:extend ls_file() to support supplied
user/group names; ok markus@
OpenBSD-Commit-ID: c70c70498b1fdcf158531117e405b6245863bfb0
* upstream: sftp client library support for
users-groups-by-id@openssh.com; ok markus@
OpenBSD-Commit-ID: ddb2f33a2da6349a9a89a8b5bcb9ca7c999394de
* upstream: use users-groups-by-id@openssh.com sftp-server extension
(when available) to fill in user/group names for directory listings.
Implement a client-side cache of see uid/gid=>user/group names. ok markus@
OpenBSD-Commit-ID: f239aeeadfa925a37ceee36ee8b256b8ccf4466e
* avoid Wuninitialized false positive in gcc-12ish
* no need for glob.h here
it also causes portability problems
* add debug on appveyor
* add sleep to pester test
* upstream: add RequiredRSASize to the list of keywords accepted by
-o; spotted by jmc@
OpenBSD-Commit-ID: fe871408cf6f9d3699afeda876f8adbac86a035e
* upstream: Fix typo. From AlexanderStohr via github PR#343.
OpenBSD-Commit-ID: a134c9b4039e48803fc6a87f955b0f4a03181497
* upstream: openssh-9.1
OpenBSD-Commit-ID: 5a467b2ee81da01a86adf1ad93b62b1728494e56
* crank versions in RPM spec files
* update release notes URL
* update .depend
* fix 9.1 compilation errors
* disable -p pester tests due to unreliability on older Windows versions
* remove extra sleep time from debugging scp pester tests
* modify -p tests to only run for Windows OS version 10 and above
* add windows specific code back into method moved from auth.c to auth2-pubkeyfile.c
* add preprocessor for WinHello
* revert preprocessor definition for winhello
* add windows preprocessor definition in key_lookup
* remove rdp block from appveyor since we are no longer debugging
* add ifdef to sftp-server.c
* make key_lookup compatible with winhello
* appveyor.yml
* increase debug of failing pester test
* add #ifdef SUPPORT_CRLF back into auth_check_principals_line method that was moved/renamed
* modify new scp.sh tests for windows
* remove in place tests from scp.sh
* remove rdp debug from appveyor
* retrigger appveyor
* change check of OS version in scp test
Signed-off-by: Corinna Vinschen <vinschen@redhat.com>
Co-authored-by: djm@openbsd.org <djm@openbsd.org>
Co-authored-by: Damien Miller <djm@mindrot.org>
Co-authored-by: Darren Tucker <dtucker@dtucker.net>
Co-authored-by: naddy@openbsd.org <naddy@openbsd.org>
Co-authored-by: dtucker@openbsd.org <dtucker@openbsd.org>
Co-authored-by: tj@openbsd.org <tj@openbsd.org>
Co-authored-by: millert@openbsd.org <millert@openbsd.org>
Co-authored-by: jmc@openbsd.org <jmc@openbsd.org>
Co-authored-by: florian@openbsd.org <florian@openbsd.org>
Co-authored-by: markus@openbsd.org <markus@openbsd.org>
Co-authored-by: Tobias Heider <me@tobhe.de>
Co-authored-by: anton@openbsd.org <anton@openbsd.org>
Co-authored-by: Tim Rice <tim@multitalents.net>
Co-authored-by: tobhe@openbsd.org <tobhe@openbsd.org>
Co-authored-by: Corinna Vinschen <vinschen@redhat.com>
Co-authored-by: Sam James <sam@gentoo.org>
Co-authored-by: jsg@openbsd.org <jsg@openbsd.org>
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.
ok dtucker@
OpenBSD-Commit-ID: 3fdca2c61ad97dc1b8d4a7346816f83dc4ce2217
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]
ok dtucker@
OpenBSD-Commit-ID: 9d4373d013edc4cc4b5c21a599e1837ac31dda0d
Also fixes an incorrect free on platforms with both libiaf and shadow
passwords (probably only Unixware). Prompted by github PR#284,
originally from @c3h2_ctf and stoeckmann@.
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.
FreeBSD has login_getpwclass() that does some special magic for
UID=0. Prefer this to login_getclass() as its easier to emulate
the former with the latter.
Based on FreeBSD PR 37416 via Ed Maste; ok dtucker@
make privilege dropping optional but allow it via callbacks (to avoid
need to link uidswap.c everywhere)
add some other flags (keep environment, disable strict path safety check)
that make this more useful for client-side use.
feedback & ok markus@
OpenBSD-Commit-ID: a80ea9fdcc156f1a18e9c166122c759fae1637bf
Add load_hostkeys_file() and hostkeys_foreach_file() that accept a
FILE* argument instead of opening the file directly.
Original load_hostkeys() and hostkeys_foreach() are implemented using
these new interfaces.
Add a u_int note field to the hostkey_entry and hostkey_foreach_line
structs that is passed directly from the load_hostkeys() and
hostkeys_foreach() call. This is a lightweight way to annotate results
between different invocations of load_hostkeys().
ok markus@
OpenBSD-Commit-ID: 6ff6db13ec9ee4edfa658b2c38baad0f505d8c20
This adds a "verify-required" authorized_keys flag and a corresponding
sshd_config option that tells sshd to require that FIDO keys verify the
user identity before completing the signing/authentication attempt.
Whether or not user verification was performed is already baked into the
signature made on the FIDO token, so this is just plumbing that flag
through and adding ways to require it.
feedback and ok markus@
OpenBSD-Commit-ID: 3a2313aae153e043d57763d766bb6d55c4e276e6
of files. This has sensible semantics wrt Match blocks and accepts glob(3)
patterns to specify the included files. Based on patch by Jakub Jelen in
bz2468; feedback and ok markus@
OpenBSD-Commit-ID: 36ed0e845b872e33f03355b936a4fff02d5794ff
sigaction(2). This wrapper blocks all other signals during the handler
preventing races between handlers, and sets SA_RESTART which should reduce
the potential for short read/write operations.
OpenBSD-Commit-ID: 5e047663fd77a40d7b07bdabe68529df51fd2519
a similar extension for certificates. This option disables the default
requirement that security key signatures attest that the user touched their
key to authorize them.
feedback deraadt, ok markus
OpenBSD-Commit-ID: f1fb56151ba68d55d554d0f6d3d4dba0cf1a452e
some arbitrary value < 0. errno is only updated in this case. Change all
(most?) callers of syscalls to follow this better, and let's see if this
strictness helps us in the future.
OpenBSD-Commit-ID: 48081f00db7518e3b712a49dca06efc2a5428075
Tess Gauthier
djm@openbsd.org
Tess Gauthier
dtucker@openbsd.org
Darren Tucker
Balu Gajjala
Vincent Brillault
Brian Katyl
Damien Miller
tobhe@openbsd.org
Manoj Ampalam
Manoj Ampalam
deraadt@openbsd.org
Bryan Berns
Corinna Vinschen