sj
Posts
294Following
44Followers
140
sj
repeated
repeated
Marco Elver
melver@fosstodon.org
Finally, with Clang 22's release, here's v1 of Linux kernel slab cache partitioning with help of Clang's AllocToken feature: https://lore.kernel.org/all/20260331111240.153913-1-elver@google.com/
"Separating pointer-containing objects from pointerless objects and data allocations can help mitigate certain classes of memory corruption exploits: attackers who gains a buffer overflow on a primitive buffer cannot use it to directly corrupt pointers or other critical metadata in an object residing in a different, isolated heap region."
0
2
2
0
0
2
@ffmancera @ljs FWIW, I use 'hkml patch format' [1] for all my patches.
[1] https://github.com/sjp38/hackermail/blob/master/USAGE.md#formatting-patches
[1] https://github.com/sjp38/hackermail/blob/master/USAGE.md#formatting-patches
1
0
2
sj
sj
Edited 1 month ago
Being able to read kernel patch AI review comments that provided by sashiko.dev [1] on terminal was nice. Showing the entire review status and sharing the outputs together with my review of the review were bit tedious for me. Particularly for the sharing of my review of the review, I wanted to just continue the mail based workflow.
So I implemented new hkml features for forwarding sashiko review status and comments. It feels much comfy. The documentation [2] is also updated.
Of course, once sashiko.dev provides the email service, it will be much more easier to use. Then, the hkml features could be deprecated and make my life easier. But, meanwhile, the new hkml features will help me.
[1] https://social.kernel.org/notice/B4G2aqLIKD9iUEF1Fo
[2] https://github.com/sjp38/hackermail/blob/master/USAGE.md#sashikodev
#linux #kernel #hkml #sashiko
So I implemented new hkml features for forwarding sashiko review status and comments. It feels much comfy. The documentation [2] is also updated.
Of course, once sashiko.dev provides the email service, it will be much more easier to use. Then, the hkml features could be deprecated and make my life easier. But, meanwhile, the new hkml features will help me.
[1] https://social.kernel.org/notice/B4G2aqLIKD9iUEF1Fo
[2] https://github.com/sjp38/hackermail/blob/master/USAGE.md#sashikodev
#linux #kernel #hkml #sashiko
0
0
1
sj
sj
And sashiko.dev, the kernel patch AI review service, is now officially announced! https://lore.kernel.org/all/7ia4o6kmpj5s.fsf@castle.c.googlers.com/
RE: https://social.kernel.org/objects/c3ee0fda-3165-4fd4-a3bb-2fce05da98b8
RE: https://social.kernel.org/objects/c3ee0fda-3165-4fd4-a3bb-2fce05da98b8
0
1
4
sj
sj
Edited 2 months ago
Yesterday I learned [1] existence of sashiko.dev, which provides AI reviews for kernel patches, thanks to @monsieuricon 's awesome b4 reivew feature.
Today I added new hkml features [2] for showing the AI review on hkml. It can be used interactively (hkml list -> open menu -> 'handle as patches' -> 'show sashiko.dev review') or from command line (hkml patch sashiko_dev <msgid>).
It was possible only owing to the awesome b4 source code. Huge appreciation to @monsieuricon for b4 development! Of course, great thanks to sashiko.dev developers!
[1] https://social.kernel.org/notice/B4EZlkuDmrLUyvT9Zw
[2] https://github.com/sjp38/hackermail/blob/master/USAGE.md#reading-sashikodev-ai-review
#linux #kernel #hkml #sashiko_dev #b4
Today I added new hkml features [2] for showing the AI review on hkml. It can be used interactively (hkml list -> open menu -> 'handle as patches' -> 'show sashiko.dev review') or from command line (hkml patch sashiko_dev <msgid>).
It was possible only owing to the awesome b4 source code. Huge appreciation to @monsieuricon for b4 development! Of course, great thanks to sashiko.dev developers!
[1] https://social.kernel.org/notice/B4EZlkuDmrLUyvT9Zw
[2] https://github.com/sjp38/hackermail/blob/master/USAGE.md#reading-sashikodev-ai-review
#linux #kernel #hkml #sashiko_dev #b4
0
2
7
0
0
3
sj
sj
@monsieuricon Today I learned the existence of sashiko.dev. I don't know who made it for what purpose, but it allowed me finding a bug in a DAMON patch. I'm already a fan of it.
RE: https://social.kernel.org/objects/75e42566-867a-4589-a5db-c1b1c4713e62
RE: https://social.kernel.org/objects/75e42566-867a-4589-a5db-c1b1c4713e62
1
0
1
sj
repeated
repeated
Linux Plumbers Conference
linuxplumbersconf@mastodon.socialCFP for LPC 2026 is open!
Important dates:
Thursday, April 23, 2026: Deadline to submit proposals to host a microconference
Sunday, June 28, 2026: Deadline to submit LPC Refereed Track Presentations Proposals and Kernel Summit Presentations Proposals.
Please use the following to access the full CFP and submit your proposal!
0
19
5
sj
sj
Edited 2 months ago
I just posted [1] my third LSF/MM/BPF 2026 [2] topic proposal. Hopefully this is my last topic proposal for the conference.
Subject of the proposal is: "DAMON Updates: Tiering, Pagel Level Monitoring and DAMON-X"
[1] https://lore.kernel.org/20260307210250.204245-1-sj@kernel.org
[2] https://events.linuxfoundation.org/lsfmmbpf/
#linux #kernel #damon #lsfmmbpf
Subject of the proposal is: "DAMON Updates: Tiering, Pagel Level Monitoring and DAMON-X"
[1] https://lore.kernel.org/20260307210250.204245-1-sj@kernel.org
[2] https://events.linuxfoundation.org/lsfmmbpf/
#linux #kernel #damon #lsfmmbpf
0
0
0
@mpdesouza @monsieuricon I guess you want to send all replies at once? Indeed 'hkml' has no good feature for such things. I do such kind of batched replying with below workaround.
After writing a reply, 'hkml' asks you if you really want to send it, or not. If you say no, 'hkml' will ask you if you want to save and tag it as 'drafts'. In the case, if you try to reply to the mail again, 'hkml' finds the draft reply and ask you if you want to continue writing the draft. You can also list all drafts using 'hkml list drafts' and continue writing it, using the menu.
So when I want to review all patches and then send all comments at once, I save my drafts for each patch, and iterate multiple times on each draft for adding more comments on the drafts. Once I think all replies are ready to go, I send those one by one, but with no more draft updates.
Maybe I can add a new hkml feature for sending all drafts for a given thread at once.
#hkml
After writing a reply, 'hkml' asks you if you really want to send it, or not. If you say no, 'hkml' will ask you if you want to save and tag it as 'drafts'. In the case, if you try to reply to the mail again, 'hkml' finds the draft reply and ask you if you want to continue writing the draft. You can also list all drafts using 'hkml list drafts' and continue writing it, using the menu.
So when I want to review all patches and then send all comments at once, I save my drafts for each patch, and iterate multiple times on each draft for adding more comments on the drafts. Once I think all replies are ready to go, I send those one by one, but with no more draft updates.
Maybe I can add a new hkml feature for sending all drafts for a given thread at once.
#hkml
0
0
2
sj
sj
DAMON release news for v6.19-rc1..v7.0-rc1 is just posted [1]. To quote the "tl;dr:"
'''
Interesting DAMON development ideas including PMU-based extension, automated
THP application, and new DAMOS quota auto-tune metrics for CXL-based tiered
memory systems have been proposed.
Two DAMON topics for LSF/MM/BPF are proposed.
In the time period, 12 grateful contributors landed 75 commits on the linux
mainline.
'''
[1] https://lore.kernel.org/20260228194805.97228-1-sj@kernel.org
#linux #kernel #damon #release_news #7.0-rc1
'''
Interesting DAMON development ideas including PMU-based extension, automated
THP application, and new DAMOS quota auto-tune metrics for CXL-based tiered
memory systems have been proposed.
Two DAMON topics for LSF/MM/BPF are proposed.
In the time period, 12 grateful contributors landed 75 commits on the linux
mainline.
'''
[1] https://lore.kernel.org/20260228194805.97228-1-sj@kernel.org
#linux #kernel #damon #release_news #7.0-rc1
0
0
1
sj
sj
I'm honored to get the LSF/MM/BPF invitation [1], and just booked the hotel. Looking forward to meet my friends and upstream community colleagues in Zagreb!
[1] https://events.linuxfoundation.org/lsfmmbpf/
#linux #kernel #lsfmmbpf
[1] https://events.linuxfoundation.org/lsfmmbpf/
#linux #kernel #lsfmmbpf
0
0
2
sj
sj
While working on Linux kernel, I wanted to be able to easily see what patches are in what stage on the road to the maintainers and the mainline, and their review status. For the reason, I made some scripts and started backing up the outputs, for the mm subsystem and its sub-subsystems including DAMON. After a few weeks of experimental use, I find it is useful for me at least. So I just shared it publicly [1], namely "mm.git dashboard".
[1] https://lore.kernel.org/all/20260224153622.60028-1-sj@kernel.org/
#linux #kernel #mm #damon #mm_git_dashboard
[1] https://lore.kernel.org/all/20260224153622.60028-1-sj@kernel.org/
#linux #kernel #mm #damon #mm_git_dashboard
0
0
1
sj
sj
Because the Lunar new year for 2026 is also passed, I shared my DAMON development yearly retrospect for 2025. To quote the conclusion,
"2025 was the busiest year of DAMON development in its history. That was
motivated by automation, observability, and memory tiering needs. Meta,
Micron, Huawei, SK hynix and individuals made the major selfish ;) and grateful
contributions."
[1] https://lore.kernel.org/all/20260222210102.153347-1-sj@kernel.org/
#linux #kernel #damon #yearly_retro #2025
"2025 was the busiest year of DAMON development in its history. That was
motivated by automation, observability, and memory tiering needs. Meta,
Micron, Huawei, SK hynix and individuals made the major selfish ;) and grateful
contributions."
[1] https://lore.kernel.org/all/20260222210102.153347-1-sj@kernel.org/
#linux #kernel #damon #yearly_retro #2025
0
0
0