On Sun, May 25, 2025 at 08:32:20PM +0200, Salvatore Bonaccorso wrote:
Control: severity -1 serious
Hi Roland, hi Cyril,
On Sun, May 25, 2025 at 05:44:42PM +0200, Roland Clobus wrote:
Hello,
On 25/05/2025 17:00, Cyril Brulebois wrote:
To be honest, I almost asked what the plans were for 6.12.29-1 regarding this issue, and whether Severity: important was appropriate. But I got busy with other things and let this fly…
Salvatore Bonaccorso <carnil@debian.org> (2025-05-25):
I see your point and having though the ITS mitigations in my opinion
is more important. But stepping a bit back I fully understand as well your point of view.
Cyril and Debian boot people, would you actually want to not get 6.12.29-1 into trixie?
Regarding the installer itself, I'm fine with either having .29 then .30 or just waiting for .30. I definitely can see how live people might want to avoid a known broken kernel's reaching testing though.
FWIW, the next upload *will* include a fix for the loop issue. I have already imported 6.12.30 for the packaging and marked the loop fix as to be backported to 6.12.
If 6.12.30 is nearly ready to be uploaded, I would propose to skip 6.12.29 and have 6.12.30 in (perhaps even faster than with a 10 day delay)
If we were to have 6.12.29, the live images based on trixie will be untestable and we would fly blind regarding trixie until 6.12.30 lands there. We are currently unable to test the live images based on sid, which is already a bit uncomfortable to me, I have even considered writing a hack to use the trixie kernel instead.
So 6.12.27 in trixie and 6.12.30 in sid seems the better option to me.
While I think the ITS mitigations would be important, I do not want to
have your life harder working on the life images. Let's raise the
severity of this bug to RC and it will be fixed with the next upload
to unstable.
With my Grml hat on, I would also like to avoid having broken images
until 6.12.30 shows up in trixie.
Thanks,
Chris
--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)