Posts
4690
Following
319
Followers
484
Linux kernel hacker and maintainer etc.

OpenPGP: 3AB05486C7752FE1

George Takei verified 🏳️‍🌈🖖🏽

Taiwan has a message for Elon Musk, who just compared Taiwan to (checks notes) Hawaii.

1
1
0

Jarkko Sakkinen

Edited 1 year ago
You should alwaysbe a beginner in something, and I'm a beginner with FPGA chips :-) If you are not a beginner in anything, you're not *really* learning.
0
1
2

We are gushing from all the love, dear users!

1.3 has surpassed the one million mark!

We ❤ you and your support for our free and software!

0
2
2

Jarkko Sakkinen

0
0
1

Jarkko Sakkinen

pretty rare these days how #ckermit is laid out: https://github.com/KermitProject/ckermit. Just a #makefile and simple set of #c-files.
0
0
1
@liw This reflects me on cover letters of large patch sets. Putting weird acronyms and reference to spec is only good for details. You still need to write the story and let the reviewer use the acronyms and spec to check the minor details.
0
1
1
@hyang yeah it seems that the number of followers in this network is relative to the viewer :-)
0
0
1

Jarkko Sakkinen

to all my 💯 #followers: thank you
1
0
1

Jarkko Sakkinen

Have had to use serial port a lot lately because working with #FPGA’s and the #lowRISC #ethernet driver is not that stable. #kermit is still the ultimate choice for this type of job IMHO :-)

Whereas in some other options you have surf through menus, this is all I need with kermit:

$ cat ~/.kermrc
set line /dev/ttyUSB0
set speed 115200
set carrier-watch off
set flow-control none
0
0
1
@panda ya, I learned that rough way but at least it is hard to forget it now :-) thanks anyway for good commentary!
0
0
1
@panda I get the features in the sense that they exist but not in the sense that they are defaults :-) I mean it is less of a harm to tune the performance later on than accidentally brick your computer. Especially this is somewhat nasty experience for someone like me who does not otherwise care about containers but has to drive-by-use given the project requirements or whatever.
0
0
1
@panda it was no longer possible to run that command because there was no enough disk space to run it, so the whole mess ended up into a race condition with the disk usage. Super bad software design Docker has IMHO.
1
0
0

Jarkko Sakkinen

Edited 1 year ago
@panda tbh, making system a brick for the sake of performance optimization is malicious behavior by definition. OK it was almost a brick, i could luckily still access SSH and did not have to start burning a USB stick just to boot up my system :-) Full workday wasted on total nonsense.

This would only make sense if you could optionally enable whatever it was doing for BTRFS because then you would know. Now it comes as a surprise, and not nice one for someone like me who had to use Docker for the 1st time in their life.
0
1
0
@ikkeT I also have one use case where Docker is the only thing that I know works. I've spun off TPM2 TSS test suite with Docker on RISC-V environment which is bootstrapped with #Buildroot image. If I recall correctly for RISC-V LXC does not work at all and for a reason unknown to me BR2_PACKAGE_PODMAN does not exist at all.
0
0
1
@ikkeT did the trick:

```
$ podman images -qa
de5f96374006
6df894023726
```

Two is expected: docker.io/library/ubuntu:20.04 + R&D image. I don't really have opinions on container languages tbh :-) I use docker format because it is used by the keystone project. podman-unshare is just useful tool for what I do sometimes.
1
0
1

Jarkko Sakkinen

Edited 1 year ago

@ikkeT After searching I think I found a solution: export BUILDAH_LAYERS=false

1
0
1
@ikkeT i like `podman-unshare` command
0
0
1
Show older