// Copyright © SixtyFPS GmbH <info@slint.dev>
// SPDX-License-Identifier: GPL-3.0-only OR LicenseRef-Slint-Royalty-free-2.0 OR LicenseRef-Slint-Software-3.0

//! Drives animation frames from the display's vsync using the Android Choreographer.
//! Optimized for ultra-low latency: self-re-entrant Choreographer callbacks wake
//! the main loop directly on vsync without thread round-trip serialization delay.

use std::sync::atomic::{AtomicBool, Ordering};
use std::sync::{Arc, Condvar, Mutex};
use std::thread::JoinHandle;

use crate::android_activity::AndroidAppWaker;
use crate::ndk::looper::{ForeignLooper, ThreadLooper};

mod ffi {
    use core::ffi::{c_long, c_void};

    #[repr(C)]
    pub struct AChoreographer {
        _unused: [u8; 0],
    }
    pub type FrameCallback = unsafe extern "C" fn(frame_time_nanos: c_long, data: *mut c_void);

    #[link(name = "android")]
    unsafe extern "C" {
        pub fn AChoreographer_getInstance() -> *mut AChoreographer;
        pub fn AChoreographer_postFrameCallback(
            choreographer: *mut AChoreographer,
            callback: FrameCallback,
            data: *mut c_void,
        );
    }
}

struct Shared {
    state: Mutex<State>,
    cond: Condvar,
    driving: AtomicBool,
    looper: Mutex<Option<ForeignLooper>>,
    main_waker: AndroidAppWaker,
    choreographer: *mut ffi::AChoreographer,
}

unsafe impl Send for Shared {}
unsafe impl Sync for Shared {}

#[derive(Default)]
struct State {
    animating: bool,
    quit: bool,
    callback_posted: bool,
}

pub struct VsyncDriver {
    shared: Arc<Shared>,
    handle: Option<JoinHandle<()>>,
    waker_handle: Option<JoinHandle<()>>,
}

impl VsyncDriver {
    pub fn new(main_waker: AndroidAppWaker) -> Self {
        let shared = Arc::new(Shared {
            state: Mutex::new(State::default()),
            cond: Condvar::new(),
            driving: AtomicBool::new(false),
            looper: Mutex::new(None),
            main_waker: main_waker.clone(),
            choreographer: std::ptr::null_mut(),
        });
        let handle = std::thread::Builder::new()
            .name("slint vsync".into())
            .spawn({
                let shared = shared.clone();
                move || vsync_thread(shared)
            })
            .ok();
        // EARLY WAKER (the real 120Hz fix, part 2): the Choreographer
        // callback only reaches the UI thread ~2.5-6 ms AFTER the vsync
        // edge — with a 2.5-5 ms Skia render plus present, that leaves
        // no phase margin inside the 8.3 ms slot (measured: 51% of frames
        // landed one slot late). This thread learns the vsync grid from
        // the callback's `frame_time_nanos` (vkprobe::note_vsync_edge) and
        // wakes the main loop slightly BEFORE each edge, so render start ≈
        // the edge and present lands at a constant ~edge+3 ms phase.
        // The Choreographer's late wake remains as fallback (the render
        // freshness guard in the event loop prevents double renders).
        let waker_handle = std::thread::Builder::new()
            .name("slint early-waker".into())
            .spawn({
                let shared = shared.clone();
                move || early_waker_thread(shared)
            })
            .ok();
        Self {
            shared,
            handle,
            waker_handle,
        }
    }

    pub fn is_driving(&self) -> bool {
        self.shared.driving.load(Ordering::Acquire)
    }

    pub fn set_animating(&self, animating: bool) {
        let mut state = self.shared.state.lock().unwrap();
        if state.animating != animating {
            state.animating = animating;
            drop(state);
            if animating {
                // Wake the vsync thread: it may be parked in `poll_once` (the
                // Choreographer queue drained once the previous animation's
                // self-sustaining callback chain ended). Waking the looper is
                // what makes it re-check `state.animating` and post the first
                // frame callback for this new animation.
                if let Some(looper) = self.shared.looper.lock().unwrap().as_ref() {
                    looper.wake();
                }
            }
            self.shared.cond.notify_one();
        }
    }
}

impl Drop for VsyncDriver {
    fn drop(&mut self) {
        let mut state = self.shared.state.lock().unwrap();
        state.quit = true;
        drop(state);
        self.shared.cond.notify_one();
        if let Some(looper) = self.shared.looper.lock().unwrap().as_ref() {
            looper.wake();
        }
        if let Some(handle) = self.handle.take() {
            let _ = handle.join();
        }
        if let Some(handle) = self.waker_handle.take() {
            let _ = handle.join();
        }
    }
}

/// Wakes the main loop a few hundred microseconds before each vsync edge
/// (see the `early_waker_handle` docs in `VsyncDriver::new`). Runs at 120 Hz
/// only while animating; otherwise parks in short sleeps.
fn early_waker_thread(shared: Arc<Shared>) {
    let waker = shared.main_waker.clone();
    // Wake slightly BEFORE the edge so poll-returns + render-start lands on
    // the edge; with a 2.5-5 ms Skia render the present then sits at a
    // constant ~edge+3 ms (middle of the slot, ~2.5 ms of deadline margin
    // both ways).
    const LEAD_NS: u64 = 800_000;
    loop {
        {
            let state = shared.state.lock().unwrap();
            if state.quit {
                return;
            }
            if !state.animating {
                drop(state);
                std::thread::sleep(std::time::Duration::from_millis(5));
                continue;
            }
        }
        let edge = i_slint_renderer_skia::wgpu_30_surface::vkprobe::VSYNC_EDGE_NS
            .load(Ordering::Relaxed);
        let period = i_slint_renderer_skia::wgpu_30_surface::vkprobe::VSYNC_PERIOD_NS
            .load(Ordering::Relaxed);
        let now = i_slint_renderer_skia::wgpu_30_surface::vkprobe::monotonic_ns();
        // Grid unknown or stale (no callbacks for ~100 ms, e.g. the app
        // went to the background): park briefly and re-check.
        if edge == 0 || period == 0 || now == 0 || now > edge + 100_000_000 {
            std::thread::sleep(std::time::Duration::from_millis(10));
            continue;
        }
        // Next edge at/after (now + LEAD).
        let mut target = edge + period;
        while target < now.saturating_add(LEAD_NS) {
            target += period;
        }
        let delay = target.saturating_sub(now + LEAD_NS);
        if delay > 0 {
            std::thread::sleep(std::time::Duration::from_nanos(delay));
        }
        let landed = i_slint_renderer_skia::wgpu_30_surface::vkprobe::monotonic_ns()
            .saturating_sub(target);
        // Wake only if we landed within ~1 ms of the edge. If we overshot
        // (CPU starvation), the Choreographer's late wake (~5 ms after the
        // edge) covers the frame instead.
        if landed <= 1_000_000 {
            waker.wake();
        }
    }
}

// Jank-hunt: frame-alignment probes. `frame_time_nanos` is the Choreographer
// vsync timestamp (System.nanoTime epoch = CLOCK_MONOTONIC on Android); we
// ALSO stamp CLOCK_MONOTONIC here so VSYNC and PRES lines share one clock and
// we can measure, per presented frame, how far past the vsync edge it landed
// (phase). If presents systematically land in the back half of the 8.3 ms
// window the compositor shows them one vsync late → effective 60 Hz + judder
// even though our frame-to-frame cadence looks like 120 Hz.
unsafe fn mon_ns() -> u64 {
    #[repr(C)]
    struct Ts {
        tv_sec: i64,
        tv_nsec: i64,
    }
    unsafe extern "C" {
        fn clock_gettime(clkid: i32, tp: *mut Ts) -> i32;
    }
    const CLOCK_MONOTONIC: i32 = 1;
    let mut ts = Ts { tv_sec: 0, tv_nsec: 0 };
    if unsafe { clock_gettime(CLOCK_MONOTONIC, &mut ts) } == 0 {
        ts.tv_sec as u64 * 1_000_000_000 + ts.tv_nsec as u64
    } else {
        0
    }
}

unsafe extern "C" fn frame_callback(
    frame_time_nanos: core::ffi::c_long,
    data: *mut core::ffi::c_void,
) {
    let shared = &*(data as *const Shared);
    // CRITICAL PATH: wake the main loop FIRST. The VSYNC log line below is
    // a file write (50-200us typical, worse under I/O pressure) — doing it
    // before the wake would push the main loop's render start (and thus the
    // present phase) later in the slot.
    shared.main_waker.wake();
    // PRESENT GATE: publish the exact vsync edge (CLOCK_MONOTONIC ns — the
    // same epoch as `mon_ns()` above) so the renderer can quantize its
    // `Queue::present` to the vsync grid (wgpu_30_surface::vkprobe).
    i_slint_renderer_skia::wgpu_30_surface::vkprobe::note_vsync_edge(
        frame_time_nanos as u64,
    );
    let mon = unsafe { mon_ns() };
    if mon > 0 {
        log::info!("VSYNC ftn={} mon={}", frame_time_nanos, mon);
    }

    let mut state = shared.state.lock().unwrap();
    if state.animating && !state.quit {
        unsafe {
            ffi::AChoreographer_postFrameCallback(
                shared.choreographer,
                frame_callback,
                data,
            );
        }
    } else {
        state.callback_posted = false;
    }
}

fn vsync_thread(shared: Arc<Shared>) {
    let looper = ThreadLooper::prepare();
    *shared.looper.lock().unwrap() = Some(looper.as_foreign().clone());

    let choreographer = unsafe { ffi::AChoreographer_getInstance() };
    if choreographer.is_null() {
        return;
    }

    let ptr = Arc::as_ptr(&shared) as *mut Shared;
    unsafe {
        (*ptr).choreographer = choreographer;
    }

    shared.driving.store(true, Ordering::Release);

    let raw_shared = Arc::into_raw(shared);

    loop {
        {
            let mut state = unsafe { &(*raw_shared).state }.lock().unwrap();
            if state.quit {
                break;
            }
            // Post the FIRST frame callback ONLY when we're animating AND no
            // callback is already pending. While animating, `frame_callback`
            // re-posts itself on every vsync (self-sustaining chain), so a
            // callback is always pending — we must NOT post another here, or
            // the pending-callback count grows by one every vsync (1, 2, 3,
            // 4…) and each vsync fires N callbacks → N main-loop wakes with N
            // unbounded. That was the ~1600 iters/s busy-spin. See
            // docs/BUGS/scroll-jank-120hz.md.
            if state.animating && !state.callback_posted {
                state.callback_posted = true;
                unsafe {
                    ffi::AChoreographer_postFrameCallback(
                        choreographer,
                        frame_callback,
                        raw_shared as *mut core::ffi::c_void,
                    );
                }
            }
        }

        // Block until the next Choreographer message (a vsync fires the frame
        // callback) or a wake (set_animating/quit). While animating the
        // self-sustaining chain keeps delivering one message per vsync; when
        // the animation stops the chain ends and this blocks until the next
        // animation's set_animating(true) wakes the looper.
        let _ = looper.poll_once();
    }

    let _ = unsafe { Arc::from_raw(raw_shared) };
}