Single RunQueue Proxy Execution Appears Ready For Linux 6.17
17 July 2025 - Proxy Execution

The long in development work around proxy execution for the Linux kernel appears to be ready for the upcoming Linux 6.17 merge window with the Single RunQueue Proxy Execution patches queued into a TIP branch after going through 19 rounds of patch review/revisions.

Comments

#Phoronix #LinuxKernel

LLVM Update
2025-07-18
#LLVM

Release tag:
LLVM 21.1.0-rc1 (Pre-release)

Assets:
There are too many, see here

Slim LLVM toolchains:
Here

Moto Widget (ported by @meoify) Update
2025-07-19
#MotoWidget

Build version:
4.13.39

Build date:
2025-07-18 18:46:36

MD5:
e2ac9677cafc853e07f0808dcd0298ad

Download:
Moto-Widget-4.13.39.apk
Pling | Direct

Size:
48.90 MB

Linux Kernel Patches Speed-Up CRC32 Performance For CPUs With "Good" AVX-512
20 July 2025 - crc32c + AVX-512 VPCLMULQDQ

Google engineer Eric Biggers who has been responsible for many great Linux cryptography subsystem performance optimizations in recent years has another exciting patch series. Biggers has done some great work for optimizing various functions for modern Intel/AMD CPUs especially around AVX-512 implementations and now he has another big optimization coming for the CRC32 checksum performance.

Comments

#Phoronix #LinuxKernel

Linux 6.16-rc7 Bringing Fix For Possible Bogus/Miscalculated Load Averages
20 July 2025 - sched/urgent

Among the fixes merged today ahead of Linus Torvalds releasing the Linux 6.16-rc7 test kernel release is a lone patch on the "sched/urgent" side to fix possible bogus load average values. Reported system load averages within the kernel's scheduler code could potentially be off going back to May of 2021.

Comments

#Phoronix #LinuxKernel

Linux 6.16-rc7 Released: "I Think We're In Good Shape"
20 July 2025 - Linux 6.16

The seventh weekly release candidate of Linux 6.16 is now availablr for testing with the stable release debuting hopefully next Sunday otherwise the following week.

Comments

#Phoronix #LinuxKernel

Google already merged 5.10.240 into android12-5.10-lts

Xiaomi.eu inject Module Update
2025-07-22
#XiaomiEU

Build date:
Mon, 21 Jul 2025 14:57:33 UT

MD5:
14c4b5504c41bc97af6a964abb6f01b6

Download:
XiaomiEUModule_2025.07.21.apk

Size:
12.8 KB

Clash Verge Rev Update
2025-07-22
#Clash #ClashMeta

Release tag:
Clash Verge Rev v2.3.2

Assets:
There are too many, see here

Beyond Compare 5 Update
2025-07-23
#BeyondCompare5

Changelog:
https://www.scootersoftware.com/download/v5changelog

Download:
5.1.1, build 31157

The Linux Kernel Seeing Rare Code Activity Around SPARC64
24 July 2025 - SPARC

One of the CPU architectures continuing to be supported by the mainline Linux kernel but rarely ever seeing any code activity is the SPARC64 architecture port for the once-interesting processors from Sun Microsystems.

Comments

#Phoronix #LinuxKernel

Beyond Compare 5 Update
2025-07-25
#BeyondCompare5

Changelog:
https://www.scootersoftware.com/download/v5changelog

Download:
5.1.2, build 31185

Google Clang Prebuilt Update
#clang

Commit

Download tar.gz:
clang 21.0.0 (based on r563880b) from build 13818152

KernelSU-Next has increased the minimum supported KernelSU version number in v1.0.9, so you will find that even if you install this patch package on Glow Kernel v4.0, you cannot use the KernelSU-Next v1.0.9 app.

Currently I have two options:
1. Completely abandon support for KernelSU-Next (I prefer this).
2. Make Glow Kernel officially support KernelSU-Next by modifying the kernel source code. But if one day KernelSU-Next makes kernel-level changes that I find difficult to accept, I will withdraw support for it.

Linux 6.16 Is Exciting For Open-Source NVIDIA, OpenVPN DCO & More Performance
25 July 2025 - Linux 6.16

The Linux 6.16 kernel is expected to be released as stable this coming Sunday, 27 July, barring any last minute issues that cause Linus Torvalds to have reservations over issuing v6.16 stable and to instead do a v6.16-rc8 test release. With Linux 6.16 imminent, here's a reminder about some of the most interesting features in this next Linux kernel version.

Comments

#Phoronix #LinuxKernel

碎碎念：

KernelSU目前最大的问题是碎片化，各个非官方版你搞你的我搞我的，各自用各自的签名密钥。

这种碎片化给内核开发者带来了很大的困扰，用户们想要使用的KernelSU版本都不尽相同，有人喜欢用MKSU，有人喜欢用KernelSU-Next，有人喜欢用SukiSU-Ultra，还有人喜欢用原版的KernelSU，内核开发者实在没法同时满足所有人的需求。

当然，内核开发者也可以把一大堆签名密钥信息加到kernel使其能够“兼容”足够多的各种非官方版KernelSU，但这又牵扯出了另一个问题：某些非官方KernelSU项目已经开始魔改kernel部分的源代码，特别是和管理器app识别相关的代码，你确定真的能兼容？

LKM安装方式的出现一定程度上缓解了碎片化，用户可以想用哪个KernelSU就选择装哪个LKM，但随着susfs的流行，开发者和用户们对闭源init（用于加载LKM）的不信任，以及某些厂商和rom维护者开始魔改GKI，LKM也变得越来越没存在感。

并且，你越是深入了解KernelSU，你就越能发现它工作在内核层所带来的局限性。

我很早以前就已经在我的 blog文章 里提到了我对KernelSU的看法：

kernel属于很接近硬件底层的东西，kernel已经很累了，请善待它。

真的，有些事情或许不应该交给内核来做。

用Magisk有什么不好？

看完了这些内容，或许你就理解了为什么Melt Kernel和Glow Kernel仅支持原版KernelSU和MKSU了。

（这里原本有一大段我对KernelSU-Next和SukiSU-Ultra的吐槽，但为了避免争议我还是删了。。。）

写到这里了我就再多吐槽几句：

为了让Glow Kernel“兼容”这些非官方KernelSU我还特地做了一个补丁包，但长远来看，这种强行兼容的做法实在是不可取。

我做的补丁包的实现原理是：通过修改KernelSU的源代码，给KernelSU新增了两个模块参数，用于保存一组额外的apk签名信息（size和sha256），以该签名信息签名的apk将能够成为KernelSU的管理器app。为安全起见，这两个模块参数是只读的，在用户空间不能修改，只能通过cmdline进行初始化，试想一下，一个恶意app拿到了root权限，然后它修改了这组可自定义的签名信息，然后以该签名信息签名的另一个恶意app就悄悄地成为了新的KernelSU管理器app，这个恶意的管理器app不需要向你申请root权限，因为它是KernelSU的管理器app所以它天生就有root权限，甚至有与KernelSU内核层交互的权限。

作为一个内核开发者，我这样做是为了兼容更多的非官方版KernelSU，并且已经在有限范围内尽可能做得安全，无可厚非，但令我疑惑的是后来SukiSU-Ultra和KernelSU-Next都先后借鉴了我的这个想法，允许用户不受限制地自定义签名信息。不是，哥们，你自己搞非官方KernelSU老老实实只支持自己签名的管理器就行了，你主动去支持其他的管理器app是想干嘛啊？

Linux Kernel Proposal Documents Rules For Using AI Coding Assistants
25 July 2025 - Linux Kernel + AI

Longtime Linux developer Sasha Levin of NVIDIA (and formerly of Google and Microsoft) as well as being the Linux LTS kernel co-maintainer today proposed a Linux kernel AI coding assistant configuration and documentation/rules for contributing to the Linux kernel with patches that are (co)authored by AI coding utilities.

Comments

#Phoronix #LinuxKernel