Peter

> Interesting, so that means the existing checks for GSO don't work. Can you log this error? > Actually, the fact that it fails on the comparison means we never...

> If we can find a way of reading the kernel version, we could also disable GSO based on that within quinn_udp Maybe the `release` field of [`utsname`](https://docs.rs/libc/latest/libc/struct.utsname.html) returned by...

On the affected device this code ```rust let mut n = unsafe { std::mem::zeroed() }; let r = unsafe { libc::uname(&mut n) }; assert_eq!(r, 0); let release = unsafe {...

Maybe something like this? https://github.com/inetic/quinn/commit/26e3950bfd37048777b9674edf45b05bb6e59abd > @djc @mxinden @Ralith @gretchenfrage opinions? Apologies for the noise, I got nerd sniped, I'll wait for the opinions from here on :-). In the...

> Wanna open a PR to discuss these details? 👍 https://github.com/quinn-rs/quinn/pull/2248

> I don't love parsing a string for it, but if that's the best option, so be it. One other option that comes to my mind is to use the...

> Superseded by https://github.com/JulianAssmann/flutter_background/pull/80 I agree, I'm closing this one

The problem is in how `file_stream` keeps track of the current position. All `win_iocp_file_service` async/non-async and read/write functions have the problem. It's easiest to see in [`write_some`](https://github.com/chriskohlhoff/asio/blob/4730c543ba2b8f9396ef1964a25ccc26eb1aea64/asio/include/asio/detail/win_iocp_file_service.hpp#L169) ```cpp template size_t...