Skip to contentSkip to conversation

Getting the kvm-riscv tree in next

DISCUSSION Oliver Upton 4 participants 4 messages Apr 21, 2023 – Apr 22, 2023 KVM

lore
Share & tools
lore
#0 lore raw

Hi Anup,

I was recently poking around on linux-next and noticed that kvm-riscv isn't included in it.

Having all of the KVM ports represented in -next is quite beneficial, as it gives us an early signal to any conflicts that may arise between our trees. Additionally, Linus likes to see that patches have been sitting in -next for a while, and is generally suspicious of any content applied immediately before the merge window. I've also noticed that for the past few kernel release cycles you've used an extremely late rc (i.e. -rc7 or -rc8), which I fear only draws more scrutiny.

So, in the interest of greasing the wheels of KVM maintenance, could you consider doing the following:

- Apply patches well in advance of the upcoming merge window on an early -rc. At least for KVM/arm64 we tend to base things on -rc3, allowing for a few weeks of soak time in -next.

- Ask Stephen to include your tree in linux-next.

As I said, I just care about reducing friction in KVM, so hopefully this doesn't come off as though I'm policing your tree.

Signature · 2 lines
Thanks,
Oliver
maintainer
#1 lore raw
Quoted 3 lines · Oliver Upton

I've also noticed that for the past few kernel release cycles you've used an extremely late rc (i.e. -rc7 or -rc8), which I fear only draws more scrutiny.

Heh, I just wrote the same thing to Anup. In particular, having kvm-riscv in next (either directly or by sending early pull requests to me) would have helped me understand the conflicts between the core and KVM trees for RISC-V, because Stephen Rothwell would have alerted me about them.

Paolo

#2 lore raw
Quoted 9 lines · Paolo Bonzini
Quoted 3 lines · Oliver Upton

I've also noticed that for the past few kernel release cycles you've used an extremely late rc (i.e. -rc7 or -rc8), which I fear only draws more scrutiny.

Heh, I just wrote the same thing to Anup. In particular, having kvm-riscv in next (either directly or by sending early pull requests to me) would have helped me understand the conflicts between the core and KVM trees for RISC-V, because Stephen Rothwell would have alerted me about them.

Speaking of not-early pull requests, I'm not going to get the x86 pull requests sent until Monday. Everything has been in place for a few weeks now, but I buried myself too deep in UPM/restrictedmem stuff and ran out of time today (and I don't trust my brain to not make stupid mistakes at this point).

#3 lore raw
In reply to Paolo Bonzini
Quoted 11 lines · Paolo Bonzini
Quoted 3 lines · Oliver Upton

I've also noticed that for the past few kernel release cycles you've used an extremely late rc (i.e. -rc7 or -rc8), which I fear only draws more scrutiny.

Heh, I just wrote the same thing to Anup. In particular, having kvm-riscv in next (either directly or by sending early pull requests to me) would have helped me understand the conflicts between the core and KVM trees for RISC-V, because Stephen Rothwell would have alerted me about them.

I think it is better to have patches sit in linux-next for at least a week before sending pull requests. I will request Stephen to include KVM RISC-V tree in linux-next.

Thanks,
Anup

Quoted 3 lines

Paolo