Message-ID: <25f16677-8633-8c06-cf67-8024fae7854f@asahilina.net>
Date:   Wed, 22 Feb 2023 01:37:44 +0900
MIME-Version: 1.0
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:91.0) Gecko/20100101
 Thunderbird/91.9.1
Subject: Re: [PATCH] rust: time: New module for timekeeping functions
Content-Language: en-US
From:   Asahi Lina <lina@asahilina.net>
To:     Boqun Feng <boqun.feng@gmail.com>,
        Thomas Gleixner <tglx@linutronix.de>
Cc:     Miguel Ojeda <ojeda@kernel.org>,
        Alex Gaynor <alex.gaynor@gmail.com>,
        Wedson Almeida Filho <wedsonaf@gmail.com>,
        Gary Guo <gary@garyguo.net>,
        =?UTF-8?Q?Bj=c3=b6rn_Roy_Baron?= <bjorn3_gh@protonmail.com>,
        John Stultz <jstultz@google.com>,
        Stephen Boyd <sboyd@kernel.org>, linux-kernel@vger.kernel.org,
        rust-for-linux@vger.kernel.org, asahi@lists.linux.dev
References: <20230221-gpu-up-time-v1-1-bf8fe74b7f55@asahilina.net>
 <87v8jvnqq4.ffs@tglx> <Y/TP6as7qqwfcI42@boqun-archlinux>
 <6aa15295-219b-225c-607d-e87e3d51d048@asahilina.net>
In-Reply-To: <6aa15295-219b-225c-607d-e87e3d51d048@asahilina.net>
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 7bit
Precedence: bulk

On 22/02/2023 01.27, Asahi Lina wrote:
> I think that's the better approach, but I was hoping to leave that for a
> future patch, especially since right now I'm the only user of this API
> in the upcoming Apple GPU driver and it only uses it to implement some
> really simple timeouts for polled operations, which isn't much API
> coverage... I figured we might get a better idea for what to do once a
> second user comes along. For example, do we want straight methods like
> that or std::ops trait impls? And do we make everything fallible or
> panic on overflow or just wrap?

Also, it's probably worth mentioning that this kind of refactor can be
done without rewriting all the user code. For example, here is how I use
the APIs:

    let timeout = time::ktime_get() + Duration::from_millis(...);
    while time::ktime_get() < timeout {
        [...]
    }

Even if ktime_get() starts returning another type, as long as it can
interoperate with core::time::Duration the same way, it will continue to
compile (and if it only interoperates with a new kernel-specific
Duration, you'd only have to change the `use` statement at the top).

~~ Lina