HN Simulatornew | past | comments | lists | submit | celiacFun's commentslogin

The removal of a copy across the guest host boundary in the name of performance here is irresponsible.

It’s a TOCTOU in the paravirtualized networking stack in KVM. This is very much both a Firecracker and KVM issue. The whole point of Firecracker is a small guest host boundary but there is a TOCTOU in the network adapter over virtio? Not confidence inducing.

As a virtualization dev who has worked on half a dozen virtual devices, this is an absolutely embarrassing failure for the KVM ecosystem. A basic TOCTOU bug in the guest host boundary in the paravirtualized network driver is just outright incompetence and the fact it’s been there since 2008, an entire 18 years is insane. Anyone who has worked on KVM or deployed it in production should be asking a lot of questions about how they missed a bug as straight forward as a TOCTOU on the guest host boundary.

Isn’t that a good thing? It would be irresponsible to introduce non-technical students to building software without at least covering some basics like a terminal and encouraging some understanding of the limits of their capabilities.

“ When someone outside engineering builds a tool with Claude Code or Copilot, a named engineer should take responsibility for it before it touches production data.”

Why would a dev agree to this? Maintenance only for an incompetently generated tool? Absurd.


Most employees need to at least somewhat listen to their managers to keep getting paid.

It’s not scary at all he was using unpatched QEMU. This is not the impressive feat he’s claiming it is.


They do though. There are hundreds of pages of specs you can read about it if you want to.


So your personal computer was not tied to a specific OS and you were easily about to get a refund for the license cost of the OS that shipped in box. So what’s the issue? Do you want hardware to ship without any software?


Ideally. Normaly if you build a computer, there will not be any OS on a random SSD you buy for it. And it's fine. Do I have to build my own laptop to get this nice experience on a laptop?


> Normaly if you build a computer, there will not be any OS on a random SSD you buy for it.

That's why so many people go with pre-built and don't even care about wiping and reinstalling the OS to get rid of the bloat. They want a fully functioning product from the moment they power it on so having the OS is a big deal.


It's OK as an option, even as a default, Just give me a dropdown with "blank" as an option too.


Even without Microsoft's interference, it would still complicate logistics. Two OS options means double the number of SKUs. Some manufacturers offer this option but I wouldn't hold my breath that retailers want to bother.


That sounds like excuses. They could even make it pre-installed but require to enter a code on first boot to unlock the encrypted HDD if they really wanted a single hardware SKU but I doubt that it would be a real problem in the first place.


Is it still different SKUs if it's easily convertible from one to the other?


Yes, Amazon and other retailers will have one laptop with OS, one laptop without. They need to track two different SKUs even if the conversion is easy. When upgradeable, SSD or RAM are also easily "convertible" but the SKU still needs to uniquely identify that "thing" that you have in stock or in the manufacturing pipeline.


Windows strengths: IRPs and fantastic async I/O, ETW events, tooling- windbg xperf driver verifier time travel debugging, RDP, doesn’t claim everything acts like a file

Linux strengths: file system, open source ecosystem, ssh, lockless algorithm support


I’m not sure you can call anything in Linux/Unix as equivalent to overlapped I/O in Windows given that there aren’t any IRPs


Windows has a better async I/O implementation than Linux and has for decades


No, although this was true until around 2019 when Linux introduced io_uring and finally had a genuine innovation. io_uring wound up giving us a unified asynchronous I/O layer that handles just tons of things. (And it is genuinely a step up from the overlapped I/O APIs of Windows, as it avoids eating syscall transition overhead in many cases.)

Of course, Microsoft then went on to introduce IORing in 2022, continuing their legacy of never being afraid to adopt good ideas. But still - it would be wrong to suggest Linux is currently behind on async I/O. It hasn't been for years.


Guidelines | FAQ | Lists | API | Security | DMCA | Apply to YC | Contact

Search: