Profile

Cover photo
Greg Kroah-Hartman
Works at Linux
16,449,731 views
AboutPosts

Stream

Greg Kroah-Hartman

Shared publicly  - 
 
Yes, 4.9 is the next longterm kernel. I've been saying that for a while, but somehow if it wasn't on the kernel.org website, no one believed me :)
105
11
Greg Kroah-Hartman's profile photoJustin Zobel's profile photoIkey Doherty's profile photoJoseph Chen's profile photo
8 comments
 
Good to know that :D
Add a comment...

Greg Kroah-Hartman

Shared publicly  - 
 
Best speaker gift ever, 6 pack of homebrew. It's as if they want me to come back and talk to them again!
91
Thomas Ruecker (TBR)'s profile photoBob Pianka's profile photoRob Herring's profile photoSimon B's profile photo
7 comments
Simon B
 
Come to Düsseldorf... we also offer homebrew...
Add a comment...

Greg Kroah-Hartman

Shared publicly  - 
 
Trying out a new build test server, thanks to +Konstantin Ryabitsev and packet.net

76
5
Konstantin Ryabitsev's profile photoDavid Alan Gilbert's profile photoTony Luck's profile photoJohn Lifsey's profile photo
18 comments
 
Can't make use of 96 core system. Too busy watching htop.
Add a comment...

Greg Kroah-Hartman

Shared publicly  - 
 
How not to write a commit message...

Tells me nothing about why we should warn about this, who should be setting this value, what should be done if the warning happens, or really, anything else at all.

{sigh}

Ugh, I do hate gnome-terminal, this is yet another example of something "new" happening that breaks or warns when updating to the latest version of the software.

People, don't break existing stuff! Or if you do, provide a way to solve the issue that you are trying to point out! Don't just blindly print a warning message and expect people to pay attention to it.

Time to go find a different terminal program, terminology is my normal choice, but for scripts that start up terminals running other scripts, gnome-terminal has been my go-to choice. Time to change that...
[gnome-terminal/gnome-3-22] client: legacy: Warn if no startup ID provided. From: Christian Persch ; To: commits-list gnome org; Cc: Subject: [gnome-terminal/gnome-3-22] client: legacy: Warn if no startup ID provided; Date: Sun, 16 Oct 2016 09:28:02 +0000 (UTC) ...
27
Martin Kühne's profile photoTheodore Ts'o's profile photoDren Kajm's profile photoMarc MERLIN's profile photo
23 comments
 
What you wrote is true of most of gnome too...
Add a comment...

Greg Kroah-Hartman

Shared publicly  - 
 
+Theodore Ts'o and I spoke last week at the SFLC's annual conference about Linux kernel related things, and it was a lot of fun. The transcript and video of that talk will hopefully be up soon, but what was really the best thing there, was Eben Moglen's closing remarks, the transcript of which is linked to below.

Go read it, or even better yet, watch the video at https://moglen.law.columbia.edu/sflc2016/09_closing.webm

Yes, it's long, but very very worth it, seriously good stuff, and something that anyone interested in Linux or any open source communities should pay attention to.
The Software Freedom Law Center provides legal representation and other law related services to protect and advance Free and Open Source Software.
35
8
Andreas Tunek's profile photoMarc MERLIN's profile photoRobert Berger's profile photoJ.O. Aho's profile photo
13 comments
 
Message a bit different depending browser, but I think in the old days the warning was more clear what was wrong with the certificate, now it's just some scare tactics without any proper information.
Add a comment...

Greg Kroah-Hartman

Shared publicly  - 
 
Have you ever wondered just how many lines of code your running kernel is using? Well, time for a dumb script that parses .c and .h files with sed!

Yeah, it's a horrid hack, but it's fun to get an idea of just how many lines of the 22 million included in the kernel tree that you rely on.

Build a kernel, and then go into the directory and run the script. Yeah, I should add an option to handle cross-builds better where the .o files are in another directory, feel free to send me patches...

For an example, my laptop, with a slimmed down configuration to only include the hardware I use, plus some drivers I test-build for:

~/linux/gregkh $ ~/linux/scripts/count_lines
Finding files for arch x86
Finding all .c files used based on the .o files...
Searching for what .h files the .c code pulls in...
Searching for what .h files the .h files pull in...
C: 2910 files 1716784 lines
H: 2232 files 419425 lines
Totals: 5142 files 2136209 lines

For my 4.9-stable allmodconfig test builds:
~/linux/stable/linux-4.9.y $ ~/linux/scripts/count_lines
Finding files for arch x86
Finding all .c files used based on the .o files...
Searching for what .h files the .c code pulls in...
Searching for what .h files the .h files pull in...
C: 23584 files 12485681 lines
H: 7948 files 820944 lines
Totals: 31532 files 13306625 lines

And finally, the defconfig for a Pixel phone:
~/linux/tmp/android_common $ ~/linux/scripts/count_lines arm64
Finding files for arch arm64
Finding all .c files used based on the .o files...
Searching for what .h files the .c code pulls in...
Searching for what .h files the .h files pull in...
C: 3312 files 2779211 lines
H: 2658 files 472081 lines
Totals: 5970 files 3251292 lines

95
7
Krishnan Nagarajan's profile photoLoïc Blot's profile photoGreg Kroah-Hartman's profile photoVishal Verma's profile photo
8 comments
 
Next step, how much of your code is running using git blame? :)
Add a comment...

Greg Kroah-Hartman

Shared publicly  - 
 
Saving here primarily as I keep sending this link to others in private email threads, but everyone in the Computer Science "industry" should look at this list, it's really good.

I think the best thing here is the "N-system" problem, where people somehow think that if you give a specification to N numbers of different groups, they will come up with N different implementations that can be then compared against each other in order to provide a "safe" system (i.e. run all 3, compare the results of the output, and if they differ, you know something went wrong somewhere.)

Turns out us programmers all end up making almost the same mistakes, so those "independent" systems almost always end up with the same bugs!

It also turns out that people really don't like being proven wrong about their ideas, so they keep saying that the "N-system" design is a secure one and should be used in new devices (cars, airplanes, etc.) The rebuttal to the critics is one of the best "stop going around and lying about our work" academic response I've read in a long time.

Lots of good reading for a winter weekend here, highly recommended.
The other day a non-CS friend remarked to me that since computer science is a quantitative, technical discipline, most issues probably have an obvious objective truth. Of course this is not at all the case, and it is not uncommon to find major disagreements even when all parties are apparently ...
58
11
Dave Wiltshire's profile photoAlan Cox's profile photoMatt Fleming's profile photoDarren Hart's profile photo
10 comments
 
+Matt Fleming​​, sure, I'll pull that together. After more discussion, Levenson has dropped to the bottom of the stack. The paper above makes the mistake of testing the effectiveness of n systems with an overly specific problem that 61508 discourages as not offering sufficient freedom of implementation to benefit from a diverse implementation. Response to her book was also fairly negative as a survey of solutions and why they won't actually work. There may be some good stuff there, but I have more promising material for the time being.
Add a comment...

Greg Kroah-Hartman

Shared publicly  - 
 
Instead of just posting to instragram stuff like this, the family says I should put it on g+ as I have more than just 5 followers here...

And yes, I'm ignoring all kernel patches until the new year, you should too.
172
1
Nik Th (NickTh)'s profile photoZamri M. S's profile photoTim Bird's profile photoRussell Nelson's profile photo
10 comments
 
Pixel camera is awesome.
Add a comment...

Greg Kroah-Hartman

Shared publicly  - 
 
For those late-night coding sessions when you are trying to remember exactly what 'const char * const *path[];' really means...
99
17
Max Eliaser's profile photoAlison Chaiken's profile photoCare Free Jatt's profile photo
11 comments
 
Plz give me a visa
Add a comment...

Greg Kroah-Hartman

Shared publicly  - 
 
Interesting summary of Thunderbolt 3.0 (i.e. "we aren't dead yet!")

The real fun is when you realize that almost all of this logic is done by your BIOS, not the operating system. Oh this is going to be fun...
15
2
Ashton Holmes's profile photoGreg Kroah-Hartman's profile photoMarc MERLIN's profile photoJason Clinton's profile photo
8 comments
 
+Greg Kroah-Hartman​ if I disable security, the external PCI dock and all of the devices attached to it work fine. If I enable User Auth, nothing works in Linux but it works beautifully it Windows 10: a dialog pops up asking if it trust the plugged in device. 
Add a comment...

Greg Kroah-Hartman

Shared publicly  - 
 
Here is the video of the discussion that +Theodore Ts'o and I had at the SFLC's annual conference about the Linux kernel community and the "norms" by which we work with companies to bring them into our community.
19
1
Add a comment...

Greg Kroah-Hartman

Shared publicly  - 
 
Oooh, someone looking for 4-9 years of experience with the old, obsolete, buggy, and insecure 2.6 Linux kernel!

I guess if you are looking for a job where you know your work will never go upstream...

Thanks to Sudip for the morning humor...
Posted By - Brainstorm Technologies. Keywords - . Job Description - Kernel Engineer with development experience in software design, debug, and verification of drivers and kernel code. You will use your nb pdeep understanding of kernel development, scheduling, locking and performance; to contribute to and interoperate within in a highly scalable multi- protocol network attached file system. Development and maintenance of drivers and Linu...
33
3
Johan Myréen's profile photoWilliam Leemans's profile photoPhilip Balister's profile photoAngus Grayling's profile photo
9 comments
 
Sounds like a job at +Ubiquiti Networks
Add a comment...
Work
Employment
  • Linux
    present
Links
Basic Information
Gender
Male