yomimono

With `dream` wanting to target SQL databases via `caqti`, this has become even more relevant and desirable :)

That's great to hear! Please do let me know if there's any way I, as a PGX and MirageOS user who's also written against Caqti, can help. :)

@bobot, did you run into this problem when running a Crowbar test under `afl-fuzz`? I had a problem where `afl-fuzz` would report one unique crash after about 3,500 runs, but...

I've updated this issue to reflect the `MCLOCK`/`PCLOCK` split in `mirage-types`; the dependencies here are on `MCLOCK` rather than `PCLOCK`, which is lighter. We should still evaluate refactoring to remove...

Thanks very much for this thoughtful and well-written report, @mfp . I'll have a closer look at this later today. Do you have code for reproducing this, or did you...

Thanks again for this report and for the repro code. I've adapted your gist (crediting you with the initial commit) and mashed it into the test suite at https://github.com/yomimono/mirage-tcpip/tree/deadlock_test ....

> I settled with a simpler expression for the lower bound (I originally had max sz (Int32.of_int (((3 * t.tx_mss + 1 lsl t.rx_wnd_scale) lsr t.rx_wnd_scale) lsl t.rx_wnd_scale)), in an...

> What about trivial getter/setters then (leaving it up to the application to set the values)? Yeah, that's much friendlier, and what I was about to come back to suggest....

`xmit` will be some invocation of `WIRE.xmit` which is [this function](https://github.com/mirage/mirage-tcpip/blob/v3.7.1/src/tcp/wire.ml#L48). It should fail only if the buffer given is too small or if the underlying interface fails to write...

Dropping some logic into ipv4 that automatically resolves "our own" IP to the correct MAC reveals another problem, at least for tests: vnetif doesn't send traffic where the source is...