dantmnf

实际 hook 的函数是 `NCSI_INTERFACE_ATTRIBUTES::SetCapability`，无论检测结果如何都修改为固定值。 注入方式是 DLL 劫持，即目标进程载入 DLL 时安装 hook，同时转发原 DLL 中的功能。修改注册表中的服务入口 DLL 是一种侵入性比较低的劫持方式，同时也不会被小更新重置。 NCSI 运行期间会输出 event log (事件查看器中 Applications and Services Logs\Microsoft\Windows\NCSI\Operational），其中有记录来源进程的 PID，可以通过 PID 反查 NCSI 在哪个服务中运行，从而劫持此服务注入 hook。

另外正如 README 所说，我卸载 npcap 之后就没有遇到 NCSI 不灵的问题了，所以这个东西我自己也没在用了（

> I am using windows 10 home version 22H2 The offset value in install.reg is likely outdated. You can follow the instructions in README to update it.

Your dhclient relies on systemd to apply some settings (most likely resolver), which is not officially supported by WSL now.

It only adds L2 connectivity to the VM, further setup (static IP/DHCP/IPv6/DNS/routing/etc.) is up to the user.

Can you reproduce the buggy reset behavior by setting both source and destination primaries to "Default", or "sRGB"?

I have reproduced the issue. Seems AMD finally added reboot persistence for regamma LUT but broke the parametric curve option.

Further analysis: * Applying regamma options other than LUT on boot (done by the driver) is broken. * Setting a regamma option identical to the current one will be a...

I think it doesn't make much sense to use source gamut other than sRGB. Almost all applications that need wide-gamut output are ICC-aware, assigning the device profile with Windows Color...

Short answer: I don't know. AMD doesn't give such clarification, nor any word that describes the behavior, for APIs we used. Even worse, the regamma API is revealed from leaked...