Posts
5584
Following
351
Followers
550
.
jarkko

Jarkko Sakkinen

Awesome: https://www.vim.org/vim-9.2-released.php
0
1
3
jarkko

Jarkko Sakkinen

The update I 'm still fine-tuning for dhowell's patch set is not just respin but is actually conceptually different.

I.e. from being task_struct centered into nsproxy centered, and repeated spawning that Eric Biedermann complained about (for legit reasons) is being addressed.

Containers are 1st class namespace now members managed by nsproxy to address the need of not wanting to use namespaces ;-)
0
0
0
jarkko

Jarkko Sakkinen

Reply to @jarkko
Tests try to address in coverage at least major part of situations related to review comments and it should be easy to populate more based on feedback. It's a rudimentary container manager.
0
0
0
jarkko

Jarkko Sakkinen

Edited 3 months ago
Getting there: https://git.kernel.org/pub/scm/linux/kernel/git/jarkko/linux-tpmdd.git/log/?h=container

# /usr/lib/kselftests/run_kselftest.sh -t containers:run.sh
TAP version 13
1..1
# timeout set to 45
# selftests: containers: run.sh
# create: alpha: pass
# create: beta: pass
# list/json: fetch: pass
# list/json: has alpha: pass
# list/json: has beta: pass
# list/json: has ready state: pass
# ps/json: fetch: pass
# ps/json: has alpha: pass
# ps/json: has beta: pass
# create: reject invalid name: pass
# run: no procfs without --proc: pass
# run: procfs with --proc: pass
# wait: observes termination: pass
# stop: SIGTERM exits init: pass
# stop: fails when TERM ignored: pass
# kill: SIGKILL exits init when TERM ignored: pass
# run: non-zero exit fails: pass
ok 1 selftests: containers: run.sh
1
0
0
jarkko

Jarkko Sakkinen

Reply to @jarkko
I've wanted to do something about this since I talked about this topic with a Netflix engineer at Linux Plumbers 2016 so here we go...
0
0
0
jarkko

Jarkko Sakkinen

Opaque container objects can be sealed better or that can be tuned very far at least.

This has the advantage when you actually want bare metal + containers instead of bare metal + vm wrapping the containers.

I mean when you want that in *production* and don't want to worry too much of security aspects.

One example use case is to maximize the throughput in video streaming service for each node. This is the "Netflix use case" (they use FreeBSD).
1
0
0
jarkko

Jarkko Sakkinen

A new git tidbit learned: git branch --edit-description

Read:

git config --get branch.container.description

Now git-format-patch will import it as the body of the cover-letter:

git format-patch --cover-letter master
0
2
8
jarkko

Jarkko Sakkinen

I'm worried that it will be a murder to send container object patch set tbh. We'll see. Everytime I think it is ready I see something that makes me unhappy.

And even if it was right technical sense it is not high odds patch set by definition.

Especially trying to nail Al Viro's and Eric W. Biedermann's feedback to previous iteration from 2019 but I'm sure I'll miss some detail, ugh :-)
0
0
0
jarkko

Jarkko Sakkinen

Reply to @andrew@this.wplr.rocks
@andrew i decided to call it kpodman.c for now and mimic distantly podman cli :-)
1
0
0
jarkko

Jarkko Sakkinen

Reply to @jarkko
@jani And I think both are great support functions for kernel dev. Right and I use for both of these a cli. My text editor does not and will not have an AI :-)
0
0
1
jarkko

Jarkko Sakkinen

Reply to @jarkko
@jani Right and second application: mechanical but tedious git history rewrites, namely "split a commit" :-)
1
0
1
jarkko

Jarkko Sakkinen

Reply to @jani@floss.social
@jani I use LLM a lot like this:

1. Scope: find issues. Don't write code. Don't propose a fix.
2. Then I go through of findings and drop false positives.
3. For the remaining, I figure out something :-)
1
0
1
jarkko

Jarkko Sakkinen

I'm trying to figure out a name for C file containing minimal container manager (or distantly a container manager). All I can make up is kontainer.c :-/ I guess I have to fix to that then. This is for kselftest.
2
0
0
jarkko

Jarkko Sakkinen

Reply to @jani@floss.social
@jani Yeah, I watched this at Github lol. Someone linked it to me.
0
0
1
jarkko

Jarkko Sakkinen

Reply to @jarkko
@kernellogger I was thinking first to give a shot with SGX because I know the opcodes really well but since I have already this crate [1] and TPM2 affects my life more than SGX ATM, let's try that at first :-)

[1] https://crates.io/crates/tpm2-protocol
0
0
0
jarkko

Jarkko Sakkinen

Reply to @kernellogger@hachyderm.io
Edited 3 months ago
@kernellogger For TPM2 type of devices having Rust to manage emulation is quite essential given the combination of structural complexity and cryptography in the binary protocol.

I could imagine that Rust could be also a great angle to deliver confidential computing opcodes for QEMU.
1
0
1
jarkko

Jarkko Sakkinen

Reply to @kernellogger@hachyderm.io
Edited 3 months ago
@kernellogger Not as many may know this but QEMU also has a nice Rust integration these days.

I implemented a feature for integrated TPM chip emulations in QEMU, and have a profile for Infienon SLB 9672 in progress. It's a half-C-half-Rust feature and so far progress has been smooth. VM sees my chip etc. and the developer experience has been surprisingly smooth :-)
1
0
7
jarkko
repeated
kernellogger@hachyderm.io

Thorsten Leemhuis (acct. 1/4)

The support in the is now officially a first class citizen and not considered experimental any more:

https://git.kernel.org/torvalds/c/9fa7153c31a3e5fe578b83d23bc9f185fde115da; for more details, see also: https://lwn.net/Articles/1050174/

This is one of the highlights from the main for 7.0 that was merged a few hours ago ; for others, see https://git.kernel.org/torvalds/c/a9aabb3b839aba094ed80861054993785c61462c

2
15
1
jarkko

Jarkko Sakkinen

Edited 3 months ago
I'll replace the test-container.c from the original patch set with kcontainer.c, which is a stripped down container manager, and a shell script containing the test cases.

It's really just a wrapper for container_* but should demonstrate with reasonable realism the capabilities of the kernel feature.
0
0
0
jarkko

Jarkko Sakkinen

A new Git subcommand I was not aware of: git range-diff. It compares two versions of branch.

This came up now that I forked dhowell's container object patch set.

E.g.,

git range-diff refs/remotes/fs/container...container
0
1
2
Show older
About social.kernel.org

Terms of service

  1. Please do not use this service in violation of the Linux Kernel Code of Conduct. Doing so will result in your account suspension with the referral of the matter to the CoC committee.

  2. "Repeating"/"boosting" someone else's status on this platform will be treated as endorsement and will fall under rule #1.

  3. You are encouraged to use this platform to promote your work on the Linux Kernel, but there is no restriction on permitted topics (with the exception of anything covered by #1 above).

  4. There is no requirement to post in English, but it should be considered the primary language of communication on this platform.

Privacy notice

The admins of this service have access to all posted statuses. They aren't looking, but if it's something they shouldn't know about, then you should not post it on this platform.

Please see the Linux Foundation Privacy Policy, which applies to this platform as well.

Getting your own account

If you would like an account on this instance, please check that the following applies to you:

  • You are listed in MAINTAINERS or CREDITS

  • OR: You have a kernel.org account or email address

  • OR: You have a long and established history of involvement with the Linux Kernel

If the above is true and you agree with the Terms of Service and Privacy Notice listed above, please use these instructions to request an account: