#914819 cinnamon: Cinnamon start is delayed

Package:
cinnamon
Source:
cinnamon
Description:
Innovative and comfortable desktop
Submitter:
Kinky Nekoboi
Date:
2026-08-16 11:49:01 UTC
Severity:
normal
#914819#5
Date:
2018-11-27 17:21:55 UTC
From:
To:
Dear Maintainer,

This behaviour is shown in 2 Debian Buster installations:

Startup of Cinnamon Environment is Delayed, only a Blank Black screen is shown.

Sometimes it starts than after like 2 Minutes

OR
Switching to ttyv1 and back
or moving the mouse, clicking mouse buttons ends the Delay.

#914819#10
Date:
2018-11-27 17:34:22 UTC
From:
To:
Install rng-tools5

There is a bugreport about it in connection with entropy. The bug number
does not know on the fast one.


Holger...

#914819#15
Date:
2026-08-16 11:47:10 UTC
From:
To:
Closing this bug, which was reported almost eight years ago against
cinnamon 3.8.8-1 in buster and was never reproduced since.

The symptoms described here - a black screen for up to two minutes after
login, which ends as soon as the mouse is moved, a key is pressed or the
user switches VT - are the classic signature of an early boot blocking on
getrandom() while the kernel CRNG is still uninitialised: user input feeds
the entropy pool and unblocks the session. That is also why the follow-up
in this report suggested installing rng-tools5. Nothing in it points at a
defect in Cinnamon itself.

This has been addressed in the kernel: as the description of the haveged
package puts it, "with linux 5.6, haveged is generally not needed anymore
as the kernel gained enough entropy functionality on its own". Buster
shipped 4.19; every suite since bullseye has 5.10 or newer (6.x in trixie
and unstable), so the condition no longer occurs, and no comparable report
has been filed against the more recent Cinnamon versions.

If a slow or black-screen session start still happens on a current version,
please file a new bug rather than reopening this one, including the exact
cinnamon and kernel versions and the output of "journalctl -b" for the
affected session, so the delay can be attributed to a specific component.

Thanks for the original report.