虽然我说自己停止更新了，并且归档了repo，但我以前在blog里说过“用户不用我自己还得用呢😅”。因此，Melt Kernel的开发不会停止，我仍然会在频道分享一些大家感兴趣的知识，但我大概率是不再会公开发布新版本的Melt Kernel了。

最后再补充一点吐槽：

我刚开始维护Melt Kernel的时候，也曾想过把所有能换成开源的内核模块都换成开源的，但试了很多次，都不太可行。
那时候我就这一台marble，还是主力机，没法总是拿来调试。
那时候的AOSP roms也都在使用小米预编译的内核。
那时候display没开源、camera没开源、触屏驱动也没开源。
那时候。。。不对，直到现在我都还在使用xiaomi.eu的修改版MIUI14/HyperOS，如果我直接把不能从源码编译的小米秘密内核模块全部剔除掉（比如：millet，migt这些），那是要出大问题的。
因此我很快就放弃了这个想法，闭源就闭源吧，能用就行，想那么多干嘛，小米还能在内核模块里植入远程操控我设备的恶意代码不成？

OSS kernel确实很好，我同意，但我不想用，因为没必要，我也不能用，因为我在使用MIUI14/HyperOS。

有心思认为我的Melt Kernel内置了小米预编译的内核模块而违反了GPL，那我还是建议你去Micode的issues骂街要求小米开源所有内核模块然后造福全人类。

Linux 6.15-rc4 To Fix The Kernel Crashing For 32-bit Systems With Too Much RAM
26 April 2025 - Linux x86 32-bit

With the Linux kernel now limiting 32-bit systems to 4GB of memory even with the "HIGHMEM" Kconfig option, an issue was uncovered where if the system was still populated with more memory than addressable by 32-bit systems, the kernel would crash. With the Linux 6.15-rc4 kernel due out on Sunday, this issue will be addressed.

#Phoronix #LinuxKernel

Fair DRM Scheduler v4 Running Well On Steam Deck, "Looks Solid"Tvrtko
26 April 2025 - Fair DRM Scheduler

Ursulin of Igalia has been pursuing the Fair DRM Scheduler as a "fair" scheduling policy to help with multiple applications/processes aiming to make use of the GPU concurrently. With this week's v4 patch-set to the DRM Fair Scheduler there are some big code changes but overall looking well as a nice scheduling policy for multiple apps/games/processes wanting equal access to GPU resources.

#Phoronix #LinuxKernel

Zblock Compressed Slab Memory Allocator Looks Like It Could Be Coming In Linux 6.16
26 April 2025 - Zblock Allocator

Ahead of the Linux 6.16 merge window opening up in just one month, the new Zblock allocator was queued up into Andrew Morton's "MM" tree of memory management material likely destined for the next kernel merge window. Zblock is showing much potential as a compressed slab memory allocator.

#Phoronix #LinuxKernel

The Linux Kernel's SHA-256 Code Being Improved Upon For Easier & Performant Use
27 April 2025 - SHA256 Refactoring

The SHA-256 code within the Linux kernel's cryptography subsystem is in the process of being refactoring so that it's available via the crypto's library API and also opening it up to support architecture-optimized implementations.

#Phoronix #LinuxKernel

Phoronix 最近似乎开启了Cloudflare防护，我的爬虫程序没法绕过它🙃
在Phoronix撤掉Cloudflare之前，我将手动转发Linux Kernel新闻。

Deferred THP Insertion Nearing The Linux Kernel To Help Avoid Memory Waste
27 April 2025 - Transparent Huge Pages Defer Mode

A patch currently residing within Andrew Morton's "MM" memory management branch ahead of the upcoming Linux 6.16 merge window is an addition by Red Hat for introducing deferred THP insertion to khugepaged. This deferred Transparent Huge Page (THP) insertion aims to help reduce memory waste on Linux with some workloads.

#Phoronix #LinuxKernel

Linux 6.15-rc4 Released With Performance Regression Fix, Corrected Bcachefs Case Folding
27 April 2025 - Linux 6.15-rc4

Linux 6.15-rc4 is now available after a rather eventful week and about one month to go until the stable Linux 6.15.

#Phoronix #LinuxKernel

New Linux Patches Aim To Customize Out-Of-Memory Behavior Using BPF
28 April 2025 - OOM + BPF

A fresh take is being pursued around allowing the Linux kernel's out-of-memory (OOM) behavior to be customized using BPF programs.

#Phoronix #LinuxKernel

SHA1: 87B1D8A410E92298FDAF502C1887BB2221765EA9

此安装包将会将设备当前的内核（内核镜像、内核模块、dtb、dtbo）全部替换为小米预编译的。
它们提取自 V14.0.27.0.TMRCNXM 和 OS2.0.5.0.VMRCNXM。

注：
不要问我这个包是谁做的，我也不知道。一切有关开源协议的法律纠纷与我无关。🤷‍♂️

This package will replace the current kernel (kernel image, kernel modules, dtb, dtbo) of the device with the pre-compiled ones by Xiaomi.
They are extracted from V14.0.27.0.TMRCNXM and OS2.0.5.0.VMRCNXM.

Note:
Do not ask me who made this package, I don't know either. Any legal disputes regarding open source agreements have nothing to do with me. 🤷‍♂️

Magisk Canary Update
2025-04-30
#MagiskCanary

Build version:
28104

Changelog:
https://github.com/topjohnwu/Magisk/releases/download/canary-28104/notes.md

Download:
app-release.apk

LLVM Update
2025-04-30
#LLVM

Release tag:
LLVM 20.1.4

Slim LLVM toolchains:
Here

DRM Code For Linux 6.16 Hides The "Disgusting Turds" & Adds TI AM68 GPU Support
29 April 2025 - DRM-Misc-Next

Sent out today was a batch of drm-misc-next patches for queuing ahead of the Linux 6.16 merge window. There are a few notable changes here as part of the Direct Rendering Manager updates for the core code and smaller kernel drivers.

#Phoronix #LinuxKernel

Bytedance Proposes Faster Linux Inter-Process Communication With "Run Process As Library"
30 April 2025 - Linux RPAL

Bytedance engineers are exploring faster inter-process communication (IPC) on Linux via a new approach they call Run Process As Library (RPAL). Their initial benchmarks of RPAL are very promising for faster Linux IPC performance.

#Phoronix #LinuxKernel

现在回过头来想了想前几天发生的事，让我决定用更客观的角度去审视自己在marble开发者社区中的状态。

站在用户的角度，我很少有看到有关Melt Kernel的正面评价。我曾看到用户评价安装Melt Kernel的行为是“上当”，我也曾看到用户因为Melt Kernel不能完美兼容E**X rom而骂Melt Kernel是垃圾。

站在开发者的角度，自打OSS kernel出现之后，Melt Kernel便越来越不受rom维护者们待见，他们明确建议用户不要安装Melt Kernel。OSS kernel团队似乎也不待见Melt Kernel，他们指责Melt Kernel内置的小米闭源内核模块中存在恶意代码（他们没有提供证据，但他们也不需要提供证据，因为他们的道德站位就是比我更高）。然后他们指责Melt Kernel内置小米闭源内核模块的行为违反了GPL。嗯，很棒，现在我在道德和法律两个层面都输了。

思来想去，我做错了什么。

后来我终于想明白了，Melt Kernel的存在本身就是个错误，他们的目的很明确：排除异己。他们就是想让Melt Kernel死。

因此，值此五一劳动佳节之际，我正式宣布：Melt Kernel的复活计划胎死腹中。Melt Kernel v3.9将永远不会发布（明明3.9是一个很棒的数字...唉）。

当你看到这条消息的时候，Melt Kernel在GitHub的repo已经取消归档，并且源代码将在未来恢复更新。

我不再指望什么正反馈和什么情绪价值，未来我将抛弃这些，只注重于自我实现。

Now I think back to what happened a few days ago, which made me decide to look at my status in the marble developer community from a more objective perspective.

From the user's perspective, I rarely see positive comments about Melt Kernel. I have seen users comment that installing Melt Kernel is "cheated", and I have also seen users scold Melt Kernel for not being perfectly compatible with E**X rom.

From the developer's perspective, since the emergence of OSS kernel, Melt Kernel has become less and less popular with rom maintainers, and they explicitly recommend users not to install Melt Kernel. The OSS kernel team also seems to dislike Melt Kernel. They accuse the Xiaomi closed-source kernel module built into Melt Kernel of malicious code (they did not provide evidence, but they don't need to provide evidence because their moral position is higher than mine). Then they accused Melt Kernel of violating the GPL by including Xiaomi's closed-source kernel module. Well, great, now I have lost both morally and legally.

After thinking about it, what did I do wrong?

Later, I finally figured it out. The existence of Melt Kernel itself is a mistake. Their purpose is very clear: to exclude dissidents. They just want Melt Kernel to DIE.

Therefore, on the occasion of May Day, I officially announce that the resurrection plan of Melt Kernel is stillborn. Melt Kernel v3.9 will never be released (3.9 is a great number...sigh).

When you see this message, the Melt Kernel repository on GitHub has been unarchived and the source code will continue to be updated in the future.

I no longer expect any positive feedback and any emotional value. In the future, I will abandon these and focus only on self-realization.

Melt Kernel是我的骄傲，339个star（截至2025年4月30日）可能是GitHub上star数最多的第三方Android内核项目。

但我始终没能想到，热爱开源的我，被同样热爱开源的人搞死。

Melt Kernel is my pride. With 339 stars (as of April 30, 2025), it may be the custom Android kernel project with the most stars on GitHub.

But I never thought that I, who loves open source, would be killed by people who love open source as well.