Hyeonggon Yoo
Posts
399Following
84Followers
85- Pre-Intern @ NVIDIA Korea (Security System Software) (2024.06 ~)
- 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
Hyeonggon Yoo
repeated
repeated
Nathan Chancellor
nathanchance@fosstodon.org
For the 2023 ClangBuiltLinux meetup, I decided to revisit my benchmarks comparing how building LLVM with different optimizations such as LTO, PGO, and BOLT can impact how fast it can build Linux kernels:
TL;DR: PGO and BOLT can give huge wins, LTO not so much.
It was pretty wild to see how much PGO made a difference, almost twice as fast in some cases... The kernel.org LLVM toolchains I provide are built with PGO and BOLT for this reason:
2
7
2
Hyeonggon Yoo
hyeyoo
Edited 5 months ago
Hmm, I'm having difficulty understanding how interrupts from PCIe devices (with MSI-X support) can be distributed to different processors.
Only an address (to an Local APIC) can be assigned to an interrupt vector?
Only an address (to an Local APIC) can be assigned to an interrupt vector?
1
0
0
Hyeonggon Yoo
repeated
repeated
K. Ryabitsev
monsieuricon
Yes, you heard it right, we're moving away from patches sent via email.
The fax number you should use from now on is 1-900-555-1212.
The fax number you should use from now on is 1-900-555-1212.
18
173
359
Hyeonggon Yoo
repeated
repeated
Vlastimil Babka
vbabka
I reduce maintainer burnout by deleting as much of the subsystem as possible.
2
13
31
Hyeonggon Yoo
repeated
repeated
sj
sj
I found yet another insightful arXiv paper exploring DAMON and finding grateful area to improve. My holiday would never be boring :)
https://arxiv.org/pdf/2311.10275.pdf
#linux #kernel #damon
https://arxiv.org/pdf/2311.10275.pdf
#linux #kernel #damon
0
2
2
Hyeonggon Yoo
hyeyoo
Edited 5 months ago
Oh, I shouldn't tag regzbot when a commit hasn't landed in the mainline yet :)
It isn't really a "regression" yet
It isn't really a "regression" yet
0
0
2
Hyeonggon Yoo
repeated
repeated
K. Ryabitsev
monsieuricon
When patchwork integration is configured, b4 will now retrieve the CI status of each patch as well.
5
9
34
Hyeonggon Yoo
repeated
repeated
Hyeonggon Yoo
repeated
repeated
Vlastimil Babka
vbabka
Edited 6 months ago
OK, enough joking, time to get serious now https://lore.kernel.org/all/20231113191340.17482-22-vbabka@suse.cz/T/
2
7
17
Hyeonggon Yoo
hyeyoo
Edited 6 months ago
there are so many batteries I should charge everyday... laptop, smartphone, smart watch and bluetooth earphones.
start to feel nervous when one of them is gone.
start to feel nervous when one of them is gone.
1
0
2
Hyeonggon Yoo
repeated
repeated
K. Ryabitsev
monsieuricon
So, you want to read LKML with Gmail (experimental, testers needed)
https://lore.kernel.org/workflows/20231115-black-partridge-of-growth-54bf2e@nitro/
https://lore.kernel.org/workflows/20231115-black-partridge-of-growth-54bf2e@nitro/
2
15
18
Hyeonggon Yoo
repeated
repeated
LWN.net
LWN@fosstodon.org[$] Faster kernel testing with virtme-ng https://lwn.net/Articles/951313/ #LWN
0
3
5
Hyeonggon Yoo
repeated
repeated
Hyeonggon Yoo
hyeyoo
Edited 6 months ago
A few diagrams about PCIe. 1) An example PCIe topology 2) PCIe Switch and 3) Root Complex
1
0
2
Hyeonggon Yoo
hyeyoo
what, there's a script to import linux kernel headers to QEMU.
From commit 412a8245 ("pci_regs.h: import from linux"):
```
It seems to make sense to import pci_regs.h from linux: why maintain our own?
As a first step, move the header to standard-headers, and add it to the update script.
```
From commit 412a8245 ("pci_regs.h: import from linux"):
```
It seems to make sense to import pci_regs.h from linux: why maintain our own?
As a first step, move the header to standard-headers, and add it to the update script.
```
0
0
1