thr3ads.net - openssh bugs - [Bug 3848] New: regress/connect-bigconf.sh fails in the github action check for 4 of macos-14 and macos-15 [Jul 2025]

If this information is useful, please help other people find it:
Share via:

bugzilla-daemon at mindrot.org

2025-Jul-11 20:19 UTC

[Bug 3848] New: regress/connect-bigconf.sh fails in the github action check for 4 of macos-14 and macos-15

https://bugzilla.mindrot.org/show_bug.cgi?id=3848

            Bug ID: 3848
           Summary: regress/connect-bigconf.sh fails in the github action
                    check for 4 of macos-14 and macos-15
           Product: Portable OpenSSH
           Version: -current
          Hardware: ARM64
                OS: Mac OS X
            Status: NEW
          Severity: normal
          Priority: P5
         Component: Regression tests
          Assignee: unassigned-bugs at mindrot.org
          Reporter: jgilab at gmail.com

Since on or about July 4 with the introduction of
regress/connect-bigconf.sh these tests fail in the github actions
check.

Example:
https://github.com/openssh/openssh-portable/actions/runs/16087394976/job/45400204229

A link to the specific error message is:
https://github.com/openssh/openssh-portable/actions/runs/16087394976/job/45400204229#step:13:266

An extract from the sshd log:

```
debug3: mm_request_receive: entering [preauth]
debug3: mm_request_receive: entering
debug3: monitor_read: checking request 8
debug3: mm_answer_pwnamallow: entering
debug2: parse_server_config_depth: config reprocess config len 670977
debug3: checking match for 'group NONEXIST' user runner host 127.0.0.1
addr 127.0.0.1 laddr 127.0.0.1 lport 4242 on line 20
getgrouplist: groups list too small
debug1: do_cleanup
debug1: Killing privsep child 62310
FAIL: ssh direct connect with large sshd_config failed
```

The specific build artifact with all logs is here:
https://github.com/openssh/openssh-portable/actions/runs/16087394976/artifacts/3469614646

I don't have an Mac OS X machine available, so I can't debug further.

-- 
You are receiving this mail because:
You are watching the assignee of the bug.

bugzilla-daemon at mindrot.org

2025-Jul-11 23:58 UTC

head link

[Bug 3848] regress/connect-bigconf.sh fails in the github action check for 4 of macos-14 and macos-15

https://bugzilla.mindrot.org/show_bug.cgi?id=3848

Damien Miller <djm at mindrot.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |djm at mindrot.org

--- Comment #1 from Damien Miller <djm at mindrot.org> ---
Created attachment 3891
  --> https://bugzilla.mindrot.org/attachment.cgi?id=3891&action=edit
getgrouplist() test program

This seems to be macos getgrouplist() returning -1 when it can't find a
group. I added some more diagnostics and this is the error that's
causing the failure:

getgrouplist("runner", 20): groups list too big (have 16, need 16)

This isn't at all specific to connect-bigconf.sh - I added a "match
group nosuchgroup" to cfgmatch.sh and now it fails there too.

If you have a macos system then running the attached program might be
instructive. I don't have one at hand right now.

-- 
You are receiving this mail because:
You are watching the assignee of the bug.
You are watching someone on the CC list of the bug.

bugzilla-daemon at mindrot.org

2025-Jul-12 00:00 UTC

head link

[Bug 3848] regress/connect-bigconf.sh fails in the github action check for 4 of macos-14 and macos-15

https://bugzilla.mindrot.org/show_bug.cgi?id=3848

Damien Miller <djm at mindrot.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |dtucker at dtucker.net
   Attachment #3892|                            |ok?(dtucker at dtucker.net)
              Flags|                            |

--- Comment #2 from Damien Miller <djm at mindrot.org> ---
Created attachment 3892
  --> https://bugzilla.mindrot.org/attachment.cgi?id=3892&action=edit
fail ga_init() gracefully when getgroups fails

We could probably deal with this situation by having ga_init()
gracefully deal with the situation where getgrouplist() fails for a
reason other than a too-small groups array.

-- 
You are receiving this mail because:
You are watching the assignee of the bug.
You are watching someone on the CC list of the bug.

bugzilla-daemon at mindrot.org

2025-Jul-12 00:01 UTC

head link

[Bug 3848] regress/connect-bigconf.sh fails in the github action check for 4 of macos-14 and macos-15

https://bugzilla.mindrot.org/show_bug.cgi?id=3848

--- Comment #3 from Damien Miller <djm at mindrot.org> ---
FWIW, the test program produces identical results on Linux and OpenBSD:
> 1 groups
>  66878
-- 
You are receiving this mail because:
You are watching the assignee of the bug.
You are watching someone on the CC list of the bug.

bugzilla-daemon at mindrot.org

2025-Jul-12 00:04 UTC

head link

[Bug 3848] regress/connect-bigconf.sh fails in the github action check for 4 of macos-14 and macos-15

https://bugzilla.mindrot.org/show_bug.cgi?id=3848

Damien Miller <djm at mindrot.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Attachment #3892|ok?(dtucker at dtucker.net)    |
              Flags|                            |
   Attachment #3892|0                           |1
        is obsolete|                            |
   Attachment #3893|                            |ok?(dtucker at dtucker.net)
              Flags|                            |

--- Comment #4 from Damien Miller <djm at mindrot.org> ---
Created attachment 3893
  --> https://bugzilla.mindrot.org/attachment.cgi?id=3893&action=edit
better diff

improved diff - move the failure case check ahead of the retry check

-- 
You are receiving this mail because:
You are watching someone on the CC list of the bug.
You are watching the assignee of the bug.

bugzilla-daemon at mindrot.org

2025-Jul-12 00:15 UTC

head link

[Bug 3848] regress/connect-bigconf.sh fails in the github action check for 4 of macos-14 and macos-15

https://bugzilla.mindrot.org/show_bug.cgi?id=3848

Darren Tucker <dtucker at dtucker.net> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Attachment #3893|ok?(dtucker at dtucker.net)    |ok+
              Flags|                            |

-- 
You are receiving this mail because:
You are watching the assignee of the bug.
You are watching someone on the CC list of the bug.

bugzilla-daemon at mindrot.org

2025-Jul-12 00:16 UTC

head link

[Bug 3848] regress/connect-bigconf.sh fails in the github action check for 4 of macos-14 and macos-15

https://bugzilla.mindrot.org/show_bug.cgi?id=3848

Darren Tucker <dtucker at dtucker.net> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Attachment #3893|                            |ok+
              Flags|                            |

-- 
You are receiving this mail because:
You are watching the assignee of the bug.
You are watching someone on the CC list of the bug.

bugzilla-daemon at mindrot.org

2025-Jul-12 06:54 UTC

head link

[Bug 3848] ga_init can fail on for reasons other than insufficient space on macos-14 and macos-15

https://bugzilla.mindrot.org/show_bug.cgi?id=3848

Darren Tucker <dtucker at dtucker.net> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
         Resolution|---                         |FIXED
             Blocks|                            |3808
            Summary|regress/connect-bigconf.sh  |ga_init can fail on for
                   |fails in the github action  |reasons other than
                   |check for 4 of macos-14 and |insufficient space on
                   |macos-15                    |macos-14 and macos-15
             Status|NEW                         |RESOLVED

--- Comment #5 from Darren Tucker <dtucker at dtucker.net> ---
Damien committed this as
https://github.com/openssh/openssh-portable/commit/f9dc519259804702cab0fa0ca8b193a360e3ec38
and the tests did indeed pass
(https://github.com/openssh/openssh-portable/actions/runs/16232160892).
 Thanks for the report.


Referenced Bugs:

https://bugzilla.mindrot.org/show_bug.cgi?id=3808
[Bug 3808] Tracking bug for OpenSSH 10.1
-- 
You are receiving this mail because:
You are watching someone on the CC list of the bug.
You are watching the assignee of the bug.

openssh bugs - Jul 2025 - [Bug 3848] New: regress/connect-bigconf.sh fails in the github action check for 4 of macos-14 and macos-15

[Bug 3848] New: regress/connect-bigconf.sh fails in the github action check for 4 of macos-14 and macos-15

[Bug 3848] regress/connect-bigconf.sh fails in the github action check for 4 of macos-14 and macos-15

[Bug 3848] regress/connect-bigconf.sh fails in the github action check for 4 of macos-14 and macos-15

[Bug 3848] regress/connect-bigconf.sh fails in the github action check for 4 of macos-14 and macos-15

[Bug 3848] regress/connect-bigconf.sh fails in the github action check for 4 of macos-14 and macos-15

[Bug 3848] regress/connect-bigconf.sh fails in the github action check for 4 of macos-14 and macos-15

[Bug 3848] regress/connect-bigconf.sh fails in the github action check for 4 of macos-14 and macos-15

[Bug 3848] ga_init can fail on for reasons other than insufficient space on macos-14 and macos-15