Posts
502
Following
97
Followers
113
A relatively new professional kernel hacker, born in August 6, 2000, and living in Korea (South!).

- Linux Kernel Developer @ Oracle (Linux Kernel MM) (2025.02 ~ Present)
- Reviewer for the Linux Slab & Reverse Mapping subsystem
- Former Intern @ NVIDIA, SK Hynix, Panmnesia (Security, MM and CXL)
- B.Sc. in Computer Science & Engineering, Chungnam National University (Class of 2025)

Opinions are my own.

My interests are:
Memory Management,
Computer Architecture,
Circuit Design,
Virtualization

All microconferences (MCs) at LPC 2025 have been accepted! It is time to submit topics to your favorite MCs.

Please check out our latest blog post for the list of MCs, and how to create a ideal MC topic.

https://lpc.events/blog/current/index.php/2025/07/25/all-microconferences-have-been-accepted/

0
6
3

I've been playing with the LLM code assistants, trying to stress them out with the Linux kernel code. So far, I've had success with them writing reasonable unit tests:
https://lore.kernel.org/lkml/20250717085156.work.363-kees@kernel.org/
https://lore.kernel.org/lkml/20250724030233.work.486-kees@kernel.org/
These both saved me some time since it emitted quite a lot of good boundary testing code, but I had to massage them both a bit until I was happy with the coverage. But it was a net win on time spent.

And then I walked it through fixing a buffer overflow. This one didn't save me any time because I had to tell it how to look at the problem. Since it was a shorter/simpler session, I included my exact prompts just for anyone interested in what I did:
https://lore.kernel.org/lkml/20250724080756.work.741-kees@kernel.org/

4
3
3
@ljs Glad you could help her! Hopefully she’ll be fine….🙏
0
0
2
@ljs how can a human being can do that to their pets 😡
1
0
2
Edited 29 days ago

This graph is the one I'm most excited about: the lifetime of security flaws in Linux is finally starting to get shorter (and the number of fixed flaws continues to rise).

https://hachyderm.io/@LinuxSecSummit@social.kernel.org/114750428620118674

1
13
3

Harry (Hyeonggon) Yoo

Edited 3 days ago
@jmorris
Personally I use it to rephrase/proofread text or quickly look something up (as a first resort before grepping through git history or mailing list archive). I heard AUTOSEL uses LLMs to determine whether a patch should be backported to -stable or not.

But at this point, I really hope no one sends any code written or guided by AI to the mailing list. It might be fine to get shallow reviews of human-written code from LLMs, but I doubt they're capable of writing sane kernel code or offering deep reviews.
0
0
1
@ljs wise. that's what reclaim does to you!
0
0
2

Oh good, some undocumented (in fact - man page implies you CANNOT DO IT) mremap() behaviour that even I, who have rewritten the entirety of mremap() code missed... + then accidentally removed...

Tests and documentation incoming...

1
1
1

Harry (Hyeonggon) Yoo

Edited 12 days ago
@ljs reviews are always appreciated :)
feeling bad to give you even more work to do :(
0
0
2

Harry (Hyeonggon) Yoo

Recently, I was debugging some intermittent boot failures and discovered that, on x86, forgetting to synchronize kernel page tables when installing new PGD entries can lead to kernel crashes.

Most of the time, PGD entries are populated during the boot process, and new tasks inherit page tables derived from the swapper process, so everything works fine.

However, if you're adding memory at runtime via memory hotplugging and populating new PGD entries for the vmemmap (i.e., the struct page array) and the direct mapping area, you need to iterate over all page tables in the system and update them accordingly to make it visible to all tasks.

Failing to handle this properly can lead to a situation where you think the page tables are set up correctly, but in reality, only init_mm.pgd (page table of the swapper process) was updated, not the page table of current task. This leads to kernel crashes since the PGD entry is not set up properly.

x86 code has mechanisms to handle this kind of synchronization, but it’s easy to be overlooked and introduce kernel crashes as the code evolves. Here’s my patch series that addresses the issue and aims to make it more robust: https://lore.kernel.org/linux-mm/20250709131657.5660-1-harry.yoo@oracle.com/
0
1
7
@ljs thought it’s a naughty spam bot before translating it
0
0
1

Harry (Hyeonggon) Yoo

Edited 15 days ago
@vbabka @wagi

Haha, chatting over a beer while sitting there is always fun!
0
0
2
@oleksandr @vitaut
$ porn add
$ porn commit -s
$ porn push
1
1
1
@wagi I can relate to this with cafe and convenience store distribution in South Korea :)
1
0
2
@vbabka @ljs
Saturday is Churnday!
0
0
3

I got ask how to get into kernel development. Basically asking for a starting point to work (simple bugs etc). I don't know a good answer and kernelknewbies seems also pretty outdated. Any ideas?

4
1
1

GNU Butt

8
3
0

Jonathan Corbet

It took a long time and over 60 articles but, at @lwn, we have finally managed to complete our reporting from the 2025 Linux Storage, Filesystem, Memory Management, and BPF Summit. If you want to know what is going on in those core parts of the kernel, this is the place to look.

We've put together an EPUB version of the whole set as well — good bedtime reading!

https://lwn.net/Articles/1026338/
1
29
57

Harry (Hyeonggon) Yoo

Edited 1 month ago
How does the kernel manage LRU lists and perform memory reclaim when memory cgroup feature is enabled? I wrote an article to answer that question.

Memory Cgroup Naturalization in the Linux Kernel v3.3
https://vmscan.org/memory-cgroup-naturalization-in-the-linux-kernel-v33

In the early days, memory cgroup was “bolted onto” the memory management subsystem, using a double-LRU scheme: each page was linked to both global and per-memcg lists—leading to a moderate memory overhead and complicating reclaim logic.

Linux v3.3 changed that. Memory cgroups were fully integrated into the MM core, global LRU lists were removed on memcg-enabled systems, and global reclaim began walking memcg hierarchy and scanning the per-memcg LRU lists, scanning each group in proportion to its memory usage.

This reduced memory overhead and simplified the reclaim code. More details in the article!
0
4
5
Show older