Posts
473
Following
87
Followers
92
n00b Kernel Hacker
- Ex-Intern @ NVIDIA Korea (Security System Software) (2024.06-2024.11)
- Ex-Intern @ Panmneisa (CXL emulation stuff) (~2023.12)
- Undergraduate majoring CSE (estimated graduation: Feb. 2025)
- Working as reviewer at Linux Slab subsystem
- Born in August 6, 2000

Opinions are my own.

My interests are:
Memory Management,
Computer Architecture,
Circuit Design,
Virtualization
Edited 1 year ago
damn it's too hot
0
0
1
Beef tartare
3
0
2
Edited 1 year ago
I've just built my first 2 socket server (36C/72T, DDR4 256GB ECC/REG RAM) for experimenting with NUMA.
It's an old xeon server (Broadwell), but enough to play with!
1
0
1
Sometimes my schedule goes out of control, and I find myself buried in the work I started. It can be overwhelming, and this is one of those moments.

However, there is some good news: it's likely that I will be participating in a field exercise program next semester at a startup that designs CXL devices.

It's a very nice opportunity, but at the same time, it's a pity that I'll be missing some classes!
1
0
3
How am I supposed to take it out
2
1
3
poor photographer
0
0
1
Since when has kernel compilation required this much RAM? My laptop with 16GiB of RAM uses tens of GBs of swap space during linkage step
1
0
0
I was removing the CPU cooler when the CPU became stuck to it and came off together.

ahh why this happened to me :(

tried to straighten bent pins but
not sure if it will work again.
2
0
0
someone took a bite out of my slipper.
1
0
1

Krzysztof Kozlowski

Edited 1 year ago
It was great to see at the EOSS 2023 commitments from some ARM64 SoC manufacturers to upstreaming their SoCs. It was visible in the conference talks but also I got such clear vibes during the talks behind the scenes.
What was a bit missing is a story of Qualcomm which recently is heavily visible upstream - both directly and through the Qualcomm Landing Team in Linaro. November last year Qualcomm released its newest, shiniest mobile SoC: the Snapdragon 8 Gen 2 (SM8550).
And guess what happened? The next day you could find all the Linux kernel support for this new SoC on the mailing list. The newest SoC was released with upstream support at the same time. There is also story from @superna9999: https://www.linaro.org/blog/upstream-linux-support-now-available-for-the-the-qualcomm-snapdragon-8-gen-2-mobile-platform/
#eoss2023 #elce2023 #EmbeddedOSSummit
1
15
39
https://2023.openinfradays.kr/schedule/day2

OCP APAC Techday is being held today in Seoul. Most of sessions are about CXL, AI, SSD. Still not sure if sessions are in english or korean.
1
0
2
Edited 1 year ago
Something strange is always found while running and experimenting new development kernels
1
0
1
1
0
1
The traditional Korean age system has been retired in South Korea, and now we use the international standard aging system.

I was 24 in the Korean age system, but now I'm legally 22 :P

https://www.bbc.com/news/world-asia-63903771
2
0
2
still haven't recovered from jet lag (?) from the final exam period. today went to bed at 11 PM, woke up at 2 AM, and it's now 6 AM.
1
0
1
Edited 1 year ago
sometimes wanna stop posting normally and just shitpost all the time
1
0
2
Edited 1 year ago
started wondering where it's documented that x86 processors deal with the synonym problem, as they use VIPT data caches.

on ARMv8 processors there is an architectural guarantee that data caches acts as PIPT cache even if it's actually VIPT, so it's guaranteed not to suffer from D-cache aliasing.

I've been heard that there is a similar guarantee on x86 processors but haven't found any official references about it - errr, I guess it's time to read some part of the intel manual.

btw this is quite old but is very nice introduction:
https://www.linuxjournal.com/article/7105
1
1
1
Edited 1 year ago
I bought a new graphics card yesterday, and it came with a freebie. now my machine hangs whenever I boot a kernel with lockdep enabled.

Here is the link to the issue I posted on GitLab: https://gitlab.freedesktop.org/drm/nouveau/-/issues/237

BTW does DRM folks prefer using GitLab or the mailing list?
0
0
2
I want NVIDIA's proprietary GPU drivers 🥲
nouveau does not support virtual GPUs...
0
0
0
Show older