https://bugzilla.mindrot.org/show_bug.cgi?id=2641
--- Comment #34 from Luca Boccassi ---
(In reply to Damien Miller from comment #33)
> Committed as 08f579231cd38 and will be in OpenSSH-9.8, due around
> June/July.
Thank you!
--
You are receiving this mail because:
You are watching the
https://bugzilla.mindrot.org/show_bug.cgi?id=2641
Damien Miller changed:
What|Removed |Added
Resolution|--- |FIXED
Status|NEW
https://bugzilla.mindrot.org/show_bug.cgi?id=2641
--- Comment #32 from Damien Miller ---
Comment on attachment 3809
--> https://bugzilla.mindrot.org/attachment.cgi?id=3809
standalone notify and timestamp patch
This looks fine to me. I'll commit it. Thanks for you help!
--
You are receiving
https://bugzilla.mindrot.org/show_bug.cgi?id=2641
Damien Miller changed:
What|Removed |Added
Attachment #3805|ok?(dtuc...@dtucker.net)|
Flags|
https://bugzilla.mindrot.org/show_bug.cgi?id=2641
--- Comment #31 from Luca Boccassi ---
Created attachment 3809
--> https://bugzilla.mindrot.org/attachment.cgi?id=3809=edit
standalone notify and timestamp patch
One more change, to support abstract namespace sockets (for containers)
as per
https://bugzilla.mindrot.org/show_bug.cgi?id=2641
--- Comment #30 from Michal Koutný ---
(In reply to Damien Miller from comment #28)
> Good catch about the sighup restart no longer running in a signal
> handler.
(In reply to Damien Miller from comment #13)
> ...
> It it also signal-handler
https://bugzilla.mindrot.org/show_bug.cgi?id=2641
Damien Miller changed:
What|Removed |Added
Attachment #3805||ok?(dtuc...@dtucker.net)
https://bugzilla.mindrot.org/show_bug.cgi?id=2641
--- Comment #29 from Luca Boccassi ---
(In reply to Damien Miller from comment #28)
> Created attachment 3805 [details]
> simplified further
>
> Good catch about the sighup restart no longer running in a signal
> handler.
>
> We can simplify
https://bugzilla.mindrot.org/show_bug.cgi?id=2641
Damien Miller changed:
What|Removed |Added
Attachment #3798|0 |1
is obsolete|
https://bugzilla.mindrot.org/show_bug.cgi?id=2641
Luca Boccassi changed:
What|Removed |Added
Attachment #3802|0 |1
is obsolete|
https://bugzilla.mindrot.org/show_bug.cgi?id=2641
--- Comment #26 from Colin Watson ---
Either version of Luca's patch looks fine to me.
--
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.
https://bugzilla.mindrot.org/show_bug.cgi?id=2641
Luca Boccassi changed:
What|Removed |Added
Attachment #3801|0 |1
is obsolete|
https://bugzilla.mindrot.org/show_bug.cgi?id=2641
--- Comment #24 from Luca Boccassi ---
Created attachment 3801
--> https://bugzilla.mindrot.org/attachment.cgi?id=3801=edit
standalone notify patch
The attached patch fixes the issue by creating a platform_post_listen()
hook, as suggested by
https://bugzilla.mindrot.org/show_bug.cgi?id=2641
--- Comment #23 from Luca Boccassi ---
(In reply to Colin Watson from comment #20)
> Actually, I noticed a slight race here. You're sending the
> readiness notification from platform_pre_listen; but, as the name
> implies, this is called
https://bugzilla.mindrot.org/show_bug.cgi?id=2641
--- Comment #22 from Luca Boccassi ---
(In reply to Colin Watson from comment #21)
> (In reply to Luca Boccassi from comment #19)
> > Mmmh hang on I don't think that should be the case. The
> > MONOTONIC_USEC is for the Type=notify-reload
https://bugzilla.mindrot.org/show_bug.cgi?id=2641
--- Comment #21 from Colin Watson ---
(In reply to Luca Boccassi from comment #19)
> Mmmh hang on I don't think that should be the case. The
> MONOTONIC_USEC is for the Type=notify-reload workflow, that
> automatically hooks sighup to the
https://bugzilla.mindrot.org/show_bug.cgi?id=2641
--- Comment #20 from Colin Watson ---
Actually, I noticed a slight race here. You're sending the readiness
notification from platform_pre_listen; but, as the name implies, this
is called _before_ the server has started listening. The point of
https://bugzilla.mindrot.org/show_bug.cgi?id=2641
--- Comment #19 from Luca Boccassi ---
(In reply to Colin Watson from comment #18)
> I've done some testing and this does seem to basically work.
>
> The one thing I'd point out is following on from Luca's comment:
> RELOADING=1 is ignored if
https://bugzilla.mindrot.org/show_bug.cgi?id=2641
--- Comment #18 from Colin Watson ---
I've done some testing and this does seem to basically work.
The one thing I'd point out is following on from Luca's comment:
RELOADING=1 is ignored if you don't also send MONOTONIC_USEC=. So if
you're not
https://bugzilla.mindrot.org/show_bug.cgi?id=2641
--- Comment #17 from Colin Watson ---
I don't see any problems from eyeballing the patch. I've pushed a
version of the Debian packaging with this (and consequent
modifications; we also have a socket activation patch from Ubuntu, but
reworking
https://bugzilla.mindrot.org/show_bug.cgi?id=2641
Richard W.M. Jones changed:
What|Removed |Added
CC||rjo...@redhat.com
--
You are
https://bugzilla.mindrot.org/show_bug.cgi?id=2641
Arkadiusz Miśkiewicz changed:
What|Removed |Added
CC||ar...@maven.pl
--
You are
https://bugzilla.mindrot.org/show_bug.cgi?id=2641
--- Comment #16 from Damien Miller ---
(In reply to Luca Boccassi from comment #14)
> While there, it would be really nice if the RELOADING=1 message also
> included MONOTONIC_USEC= (CLOCK_MONOTONIC in usec as a
> decimal string), which is used
https://bugzilla.mindrot.org/show_bug.cgi?id=2641
--- Comment #15 from Damien Miller ---
I think the READY=1 will be sent implicitly after sshd restarts
--
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.
https://bugzilla.mindrot.org/show_bug.cgi?id=2641
Sam James changed:
What|Removed |Added
CC||s...@gentoo.org
--
You are receiving this
https://bugzilla.mindrot.org/show_bug.cgi?id=2641
Luca Boccassi changed:
What|Removed |Added
CC||luca.bocca...@gmail.com
--- Comment #14
https://bugzilla.mindrot.org/show_bug.cgi?id=2641
Damien Miller changed:
What|Removed |Added
Attachment #2896|0 |1
is obsolete|
https://bugzilla.mindrot.org/show_bug.cgi?id=2641
git...@kalvdans.no-ip.org changed:
What|Removed |Added
CC||git...@kalvdans.no-ip.org
--
https://bugzilla.mindrot.org/show_bug.cgi?id=2641
Petr Cerny [:hrosik] changed:
What|Removed |Added
CC||pce...@suse.cz
https://bugzilla.mindrot.org/show_bug.cgi?id=2641
Damien Miller changed:
What|Removed |Added
CC||d...@mindrot.org
---
https://bugzilla.mindrot.org/show_bug.cgi?id=2641
Michal Koutný changed:
What|Removed |Added
CC||mkou...@suse.com
---
https://bugzilla.mindrot.org/show_bug.cgi?id=2641
Mario Trangoni changed:
What|Removed |Added
CC|
https://bugzilla.mindrot.org/show_bug.cgi?id=2641
--- Comment #9 from daniel.bl...@au.ibm.com ---
In a non-forking type=notify case systemd keeps track of the sshd
service as it is a child process. Any termination of the child process
is a signal to the system launching process and all operations
https://bugzilla.mindrot.org/show_bug.cgi?id=2641
--- Comment #8 from Jakub Jelen ---
(In reply to Darren Tucker from comment #7)
> (In reply to Jakub Jelen from comment #6)
> > Created attachment 2950 [details]
> > fixed patch
> >
> > Never mind. Nothing from above resolves
https://bugzilla.mindrot.org/show_bug.cgi?id=2641
--- Comment #7 from Darren Tucker ---
(In reply to Jakub Jelen from comment #6)
> Created attachment 2950 [details]
> fixed patch
>
> Never mind. Nothing from above resolves the race condition between
> systemd reading PID
https://bugzilla.mindrot.org/show_bug.cgi?id=2641
Jakub Jelen changed:
What|Removed |Added
Attachment #2893|0 |1
is obsolete|
https://bugzilla.mindrot.org/show_bug.cgi?id=2641
daniel.bl...@au.ibm.com changed:
What|Removed |Added
CC||daniel.bl...@au.ibm.com
--
You
https://bugzilla.mindrot.org/show_bug.cgi?id=2641
--- Comment #5 from Jakub Jelen ---
The test if the process is daemon passes for every process that is run
as a systemd service so it does not solve the problem for us. As the
result with the pushed change the 7.4p1 will not
https://bugzilla.mindrot.org/show_bug.cgi?id=2641
--- Comment #4 from Jakub Jelen ---
Colin, the -D was omitted (and Type=Forking added) because systemd was
unable to report failures (when configuration file had an error), as
described in the Red Hat bugzilla [1].
Certainly,
https://bugzilla.mindrot.org/show_bug.cgi?id=2641
--- Comment #3 from Colin Watson ---
Jakub: That's only a problem because for some reason Fedora doesn't use
the -D option when running from systemd (if you did, then
no_daemon_flag would be true so the value of
https://bugzilla.mindrot.org/show_bug.cgi?id=2641
--- Comment #2 from Jakub Jelen ---
Sorry for a quite late answer.
I got bitten by this while trying to rebase to the OpenSSH 7.4. Your
patch got into this release in 7fc4766a, but was partially reverted in
f2398eb7.
But to
https://bugzilla.mindrot.org/show_bug.cgi?id=2641
Darren Tucker changed:
What|Removed |Added
CC||dtuc...@zip.com.au
42 matches
Mail list logo