Re: [PATCH v6 4/7] rust: time: add Delta::to_jiffies_timeout() for timeout conversion
Miguel Ojeda <[email protected]>
| Newsgroups | org.kernel.vger.rust-for-linux |
|---|---|
| Message-ID | <CANiq72kHP1s7bGr4=xfJGR8KY=aUpjOB4OwLEG5E=TKULXg5Rg@mail.gmail.com> |
On Sat, Aug 8, 2026 at 9:48 AM FUJITA Tomonori <[email protected]> wrote: > > + // CAST: `__msecs_to_jiffies()` returns a value in `0..=MAX_JIFFY_OFFSET`, i.e. > + // `((LONG_MAX >> 1) - 1)`, which is non-negative and well within `isize`. I don't think I can pick up this one, because this comment isn't true for arm32 `HZ=1000`, i.e. `__msecs_to_jiffies()` can be higher than `MAX_JIFFY_OFFSET` and can actually reach `isize::MAX`. Which in turns means that we don't saturate to `MAX_JIFFY_OFFSET` like the commit message mentions. It also makes the method non-monotonic, i.e. going from: Delta::from_millis(i64::from(i32::MAX)).to_jiffies_timeout() to: Delta::from_millis(i64::from(i32::MAX) + 1).to_jiffies_timeout() means a big drop, which was not what I would naively expect from the method. Could the documentation clarify what the saturation is? And add some doctests? (Also to the previous patch's function if possible). Thanks! Cheers, Miguel