Greg K-H
Posts
273Following
88Followers
2846
@agraf @kernellogger 1/2 year to rebase to a new version? You are doing something really really really wrong there.
The Android kernel team has been keeping the android kernel tree up to date with Linus's -rc releases within a few days at most, usually within a few hours (note, it is not up to date with 6.3-rc1 right now for "gpio reasons"). Keep your tree constantly up to date, that way it just does not matter when a new LTS comes out, AND you catch any rebase/conflict issues way way early when you can get upstream's help with it.
So if it takes you 6+ months to rebase, you really have the wrong development and maintenance model and need to fix it now And yes, Google prodkernel team, I'm looking at you...
If anyone wants me to come and yell at internal managers about all of this, please let me know, that's one of the funnest parts of my job.
The Android kernel team has been keeping the android kernel tree up to date with Linus's -rc releases within a few days at most, usually within a few hours (note, it is not up to date with 6.3-rc1 right now for "gpio reasons"). Keep your tree constantly up to date, that way it just does not matter when a new LTS comes out, AND you catch any rebase/conflict issues way way early when you can get upstream's help with it.
So if it takes you 6+ months to rebase, you really have the wrong development and maintenance model and need to fix it now And yes, Google prodkernel team, I'm looking at you...
If anyone wants me to come and yell at internal managers about all of this, please let me know, that's one of the funnest parts of my job.
0
12
23
@kernellogger No one ever asked, that sounds like a good idea.
Maybe we can do that in the future as we wean people off of the 6 year model which has obviously failed to work well, thanks!
Maybe we can do that in the future as we wean people off of the 6 year model which has obviously failed to work well, thanks!
2
3
16
@fenruspdx @Aissen @kernellogger We already do this today, but no one seems to notice: https://github.com/cloudsecurityalliance/gsd-database
0
0
2
@Aissen @kernellogger Based on what I see today, in the wild, almost no one is actually using the old LTS kernels we provide today. So I really doubt this is going to change anything.
0
0
1
@Aissen @kernellogger The LTS kernels are not going away, please use them. It's just that you can not expect them to live for "forever".
The simplest solution is for you to demand support for your SoC and devices upstream, that gets rid of the issue of what kernel you are forced to use immediately. We did that decades ago for the "Enterprise" Linux market, and all of the problems of "we are stuck on this old kernel because the vendor never forward ported their code" instantly went away.
The simplest solution is for you to demand support for your SoC and devices upstream, that gets rid of the issue of what kernel you are forced to use immediately. We did that decades ago for the "Enterprise" Linux market, and all of the problems of "we are stuck on this old kernel because the vendor never forward ported their code" instantly went away.
2
8
20
@bert_hubert My big problem is, who defines "known"?
With our without "exploitable" the main issue seems to be ignored whereby the EU will have to spin up a European version of what China and the US attempt to do with their vulnerability tracking efforts, despite them failing horribly for open source software.
I think the phrase I'm looking for is, "Ik zie beren op de weg", right?
With our without "exploitable" the main issue seems to be ignored whereby the EU will have to spin up a European version of what China and the US attempt to do with their vulnerability tracking efforts, despite them failing horribly for open source software.
I think the phrase I'm looking for is, "Ik zie beren op de weg", right?
1
0
1
Greg K-H
repeated
repeated
K. Ryabitsev
monsieuricon
Me: we have 50TB on our backend storage system that stores kernel tarballs, so this should be plenty for the next 5+ years.
@gregkh: challenge accepted!
@gregkh: challenge accepted!
1
4
32
Greg K-H
gregkh
You get a stable kernel release, and you get a stable kernel release, and you get a stable kernel release!
Hopefully things now settle down to the normal constant crazy pace we are used to (1-2 releases a week), instead of the mass of releases we had in the past few days.
Hopefully things now settle down to the normal constant crazy pace we are used to (1-2 releases a week), instead of the mass of releases we had in the past few days.
0
6
17
@kernellogger Famous last words, yet-another 5.15.y release is now out as well.
1
0
4
@kernellogger Now finished with the rest, it's been an "interesting" Saturday... Everyone go update!
1
3
10
Greg K-H
gregkh
@killyourfm Just a short note, your recent #mixtape playlists have been the soundtrack for my recent kernel development and release work for the past few weeks, great work! Many thanks for sharing them!
1
2
15
Greg K-H
gregkh
"The magic to remember is type [ESC]ZZ to save and exit."
"Yes that is an odd set of things to remember."
"I don't know, historical reasons."
"Yes, graphical editors are prettier, but sometimes you will have to use this."
Parents, don't forget to have the uncomfortable conversation with your children about vim _before_ they leave for college and are exposed to the siren-call of vscode.
"Yes that is an odd set of things to remember."
"I don't know, historical reasons."
"Yes, graphical editors are prettier, but sometimes you will have to use this."
Parents, don't forget to have the uncomfortable conversation with your children about vim _before_ they leave for college and are exposed to the siren-call of vscode.
1
9
21
@kees See https://social.kernel.org/notice/ARWvggnOvXny0CUCIa for more details.
0
1
5