[PATCH] Cygwin: Change service name to cygsshd

Nico Kadel-Garcia nkadel at gmail.com
Sun Jan 27 14:00:05 AEDT 2019

On Sat, Jan 26, 2019 at 6:30 PM Corinna Vinschen <vinschen at redhat.com> wrote:
> On Jan 26 18:12, Nico Kadel-Garcia wrote:
> > On Sat, Jan 26, 2019 at 5:07 PM Corinna Vinschen <vinschen at redhat.com> wrote:
> > >
> > > Microsoft hijacked the sshd service name without asking.
> >
> > How many people use the SSH daemon from Cygwin, versus using only the
> > client? I did some work with the daemon with rsync and tar, trying to
> > link it to Linux backup systems. It was unreliable due to anti-virus
> > interfering with warning when upgraded, and the workarounds being
> > eliminated from the anti-virus software. As a server, it also had
> > problsm doing scp, sftp, or rsync reliably due to Windows file
> > locking, so it wasn't as effective or useful as one might nope.
> In other words, you made wrong assumptions and blame the Cygwin server
> for that?  Do you honestly expect the Win32 OpenSSH from Microsoft
> doesn't have the exact same problems in the exact same situation?  After
> all, Windows file locking is at is is (it's not used by Cygwin
> processes, btw., which use POSIX or BSD advisory file locking).

What? No. I wasn't "making assumptions", I was "making observations of
reality" and explaining to the very nice people that an SSHD on a
server with dynamic content was not a reliable backup tool due to
those observations. They had to dump the project, and I gave them back
the consulting money out of ethical concern that I could not do what
they'd expected. I'm asking a purely practical question: how much does
Cygwin OpenSSH get used, due to the other limitations of the NTFS
filesystem causing locks and timeouts of transfer processes? How much
do we really lose if the service is renamed for CygWin?

> > In other words, is this a big loss if Microsoft accidentally kills
> > Cygwin sshd ?
> Yes.  Cygwin is a complete POSIX environment and the sshd server is
> a part of it for more than 16 years.  I'm not opposed to the Win32
> OpenSSH from Microsoft, it's a nice thing in fact.  But it doesn't
> invalidate the usefulness of Cygwin's sshd.

Is it, in fact, that useful? The client certainly is, and I've even
used the daemon myself in the last few years to do some port
forwarding for clients. But how many folks are actually using the
Cygwin sshd these days?

> > Is Microsoft going to use port 22? If there is a
> > conflict, will CygWin defer to Microsoft as the vendor and use an
> > alternate port?
> The people I know from the Cygwin ML disabled Win32 sshd and started the
> Cygwin sshd service instead.  And ultimately port 22 is just a default
> port.  There's sshd_config.

Well, yes. I'm a bit concerned that Cygwin users will muck with the
Microsoft sshd to enable the Cygwin daemon they personally expect, and
cause their own IT departments to scream bloody murder when they
realize some developer replaced the approved management daemon on
their laptop. I'm bringing popcorn for that one.

> Corinna
> --
> Corinna Vinschen
> Cygwin Maintainer
> Red Hat
> _______________________________________________
> openssh-unix-dev mailing list
> openssh-unix-dev at mindrot.org
> https://lists.mindrot.org/mailman/listinfo/openssh-unix-dev

More information about the openssh-unix-dev mailing list