Upstream kernel?

[amstan]

I was looking at issue MiSTerArch/PKGBUILDs#12 and thought: wouldn't it be cool if we could just use the linux-armv7 arch package that's already tracking upstream?

Sigh, what am I getting into?

https://github.com/MiSTer-devel/Linux-Kernel_MiSTer/commits/MiSTer-v5.15

Doesn't seem like there's much to it. I can boil it down to:

• DTS for de10 with lots of mister additions
  • I wonder if this can be upstreamed and the mister additions be had with just a dt overlay (that can even be loaded later via userspace)
• Mister Framebuffer: 1 file, looks pretty self contained, could probably be a module
• Mister Alsa driver: 1 file, looks pretty self contained, could probably be a module
• A ton of random device (that people use with their misters) driver fixes that I'm baffled why they have to be in this kernel tree and not upstream. If I were to maintain my own kernel I would drop this.
• A few intrusive core driver changes: usb overcurrent, vty count, input device ioctls.
• EXFAT + loop device hacks, really unnecessary with something like Arch.

[amstan]

I did a rebase of all mister kernel comits on top of the real v5.15.1 tag, no issues there. So looks like 5.15.1 was really that.

Looks like EXFAT is the first thing that fails as soon as I attempt a rebase on 5.15.25. Dropping it now.

At some point in the future I'll try starting from the other direction: keep it pure and add bare minimum patches instead.

[amstan]

Looks like EXFAT is the first thing that fails as soon as I attempt a rebase on 5.15.25. Dropping it now.

Eh, that was just a "this file was already deleted, i'm confused" kind of conflict. It seems happy if I rebase it manually.