Skip to content

Commit

Permalink
Change {frame,copy_lpf}_progress to AtomicI32
Browse files Browse the repository at this point in the history
  • Loading branch information
@rinon
rinon committed Jan 9, 2024
1 parent 5b444a7 commit 2f9c108
Show file tree
Hide file tree
Showing 3 changed files with 27 additions and 55 deletions.
6 changes: 3 additions & 3 deletions src/internal.rs
View file
Original file line number Diff line number Diff line change
Expand Up @@ -78,6 +78,7 @@ use libc::ptrdiff_t;
use std::ffi::c_int;
use std::ffi::c_uint;
use std::sync::atomic::AtomicI32;
use std::sync::atomic::AtomicU32;
use std::sync::Arc;

#[repr(C)]
Expand Down Expand Up @@ -361,8 +362,8 @@ pub struct Rav1dFrameContext_frame_thread {
pub next_tile_row: [c_int; 2], /* 0: reconstruction, 1: entropy */
pub entropy_progress: AtomicI32,
pub deblock_progress: AtomicI32, // in sby units
pub frame_progress: *mut atomic_uint,
pub copy_lpf_progress: *mut atomic_uint,
pub frame_progress: Vec<AtomicU32>,
pub copy_lpf_progress: Vec<AtomicU32>,
// indexed using t->by * f->b4_stride + t->bx
pub b: *mut Av1Block,
pub cbi: *mut CodedBlockInfo,
Expand All @@ -371,7 +372,6 @@ pub struct Rav1dFrameContext_frame_thread {
// iterated over inside tile state
pub pal_idx: *mut u8,
pub cf: *mut DynCoef,
pub prog_sz: c_int,
pub pal_sz: c_int,
pub pal_idx_sz: c_int,
pub cf_sz: c_int,
Expand Down
3 changes: 2 additions & 1 deletion src/lib.rs
View file
Original file line number Diff line number Diff line change
Expand Up @@ -1029,7 +1029,8 @@ unsafe fn close_internal(c_out: &mut *mut Rav1dContext, flush: c_int) {
pthread_cond_destroy(&mut (*f).task_thread.cond);
pthread_mutex_destroy(&mut (*f).task_thread.lock);
}
freep(&mut (*f).frame_thread.frame_progress as *mut *mut atomic_uint as *mut c_void);
mem::take(&mut (*f).frame_thread.frame_progress); // TODO: remove when context is owned
mem::take(&mut (*f).frame_thread.copy_lpf_progress); // TODO: remove when context is owned
freep(&mut (*f).task_thread.tasks as *mut *mut Rav1dTask as *mut c_void);
freep(
&mut *((*f).task_thread.tile_tasks).as_mut_ptr().offset(0) as *mut *mut Rav1dTask
Expand Down
73 changes: 22 additions & 51 deletions src/thread_task.rs
View file
Original file line number Diff line number Diff line change
Expand Up @@ -53,6 +53,7 @@ use std::ffi::c_uint;
use std::ffi::c_void;
use std::process::abort;
use std::sync::atomic::AtomicI32;
use std::sync::atomic::AtomicU32;
use std::sync::atomic::Ordering;

#[cfg(target_os = "linux")]
Expand Down Expand Up @@ -381,29 +382,15 @@ unsafe fn create_filter_sbrow(
if pass & 1 != 0 {
(*f).frame_thread.entropy_progress = AtomicI32::new(0);
} else {
let prog_sz = ((*f).sbh + 31 & !(31 as c_int)) >> 5;
if prog_sz > (*f).frame_thread.prog_sz {
let prog: *mut atomic_uint = realloc(
(*f).frame_thread.frame_progress as *mut c_void,
((2 * prog_sz) as usize).wrapping_mul(::core::mem::size_of::<atomic_uint>()),
) as *mut atomic_uint;
if prog.is_null() {
return -(1 as c_int);
}
(*f).frame_thread.frame_progress = prog;
(*f).frame_thread.copy_lpf_progress = prog.offset(prog_sz as isize);
}
(*f).frame_thread.prog_sz = prog_sz;
memset(
(*f).frame_thread.frame_progress as *mut c_void,
0 as c_int,
(prog_sz as usize).wrapping_mul(::core::mem::size_of::<atomic_uint>()),
);
memset(
(*f).frame_thread.copy_lpf_progress as *mut c_void,
0 as c_int,
(prog_sz as usize).wrapping_mul(::core::mem::size_of::<atomic_uint>() as usize),
);
let prog_sz = (((*f).sbh + 31 & !(31 as c_int)) >> 5) as usize;
(*f).frame_thread.frame_progress.clear();
(*f).frame_thread
.frame_progress
.resize_with(prog_sz, || AtomicU32::new(0));
(*f).frame_thread.copy_lpf_progress.clear();
(*f).frame_thread
.copy_lpf_progress
.resize_with(prog_sz, || AtomicU32::new(0));
(*f).frame_thread
.deblock_progress
.store(0, Ordering::SeqCst);
Expand Down Expand Up @@ -659,16 +646,14 @@ unsafe fn get_frame_progress(c: *const Rav1dContext, f: *const Rav1dFrameContext
let mut idx = (frame_prog >> (*f).sb_shift + 7) as c_int;
let mut prog;
loop {
let state: *mut atomic_uint =
&mut *((*f).frame_thread.frame_progress).offset(idx as isize) as *mut atomic_uint;
let val: c_uint = !::core::intrinsics::atomic_load_seqcst(state);
let val: c_uint = !((*f).frame_thread.frame_progress)[idx as usize].load(Ordering::SeqCst);
prog = if val != 0 { ctz(val) } else { 32 as c_int };
if prog != 32 as c_int {
break;
}
prog = 0 as c_int;
idx += 1;
if !(idx < (*f).frame_thread.prog_sz) {
if !((idx as usize) < (*f).frame_thread.frame_progress.len()) {
break;
}
}
Expand Down Expand Up @@ -1037,13 +1022,9 @@ pub unsafe extern "C" fn rav1d_worker_task(data: *mut c_void) -> *mut c_void {
} else if (*t).type_0 as c_uint
== RAV1D_TASK_TYPE_CDEF as c_int as c_uint
{
let prog_0: *mut atomic_uint =
(*f).frame_thread.copy_lpf_progress;
let p1_1 = ::core::intrinsics::atomic_load_seqcst(
&mut *prog_0.offset(((*t).sby - 1 >> 5) as isize)
as *mut atomic_uint,
)
as c_int;
let p1_1 = (*f).frame_thread.copy_lpf_progress
[((*t).sby - 1 >> 5) as usize]
.load(Ordering::SeqCst);
if p1_1 as c_uint & (1 as c_uint) << ((*t).sby - 1 & 31)
!= 0
{
Expand Down Expand Up @@ -1476,19 +1457,12 @@ pub unsafe extern "C" fn rav1d_worker_task(data: *mut c_void) -> *mut c_void {
pthread_cond_signal(&mut (*ttd).cond);
}
} else if (*(*f).seq_hdr).cdef != 0 || (*f).lf.restore_planes != 0 {
::core::intrinsics::atomic_or_seqcst(
&mut *((*f).frame_thread.copy_lpf_progress)
.offset((sby >> 5) as isize)
as *mut atomic_uint,
(1 as c_uint) << (sby & 31),
);
(*f).frame_thread.copy_lpf_progress[(sby >> 5) as usize]
.fetch_or((1 as c_uint) << (sby & 31), Ordering::SeqCst);
if sby != 0 {
let prog_1 = ::core::intrinsics::atomic_load_seqcst(
&mut *((*f).frame_thread.copy_lpf_progress)
.offset((sby - 1 >> 5) as isize)
as *mut atomic_uint,
)
as c_int;
let prog_1 = (*f).frame_thread.copy_lpf_progress
[(sby - 1 >> 5) as usize]
.load(Ordering::SeqCst);
if !prog_1 as c_uint & (1 as c_uint) << (sby - 1 & 31) != 0
{
(*t).type_0 = RAV1D_TASK_TYPE_CDEF;
Expand Down Expand Up @@ -1630,11 +1604,8 @@ pub unsafe extern "C" fn rav1d_worker_task(data: *mut c_void) -> *mut c_void {
continue;
}
} else {
::core::intrinsics::atomic_or_seqcst(
&mut *((*f).frame_thread.frame_progress).offset((sby >> 5) as isize)
as *mut atomic_uint,
(1 as c_uint) << (sby & 31),
);
(*f).frame_thread.frame_progress[(sby >> 5) as usize]
.fetch_or((1 as c_uint) << (sby & 31), Ordering::SeqCst);
pthread_mutex_lock(&mut (*f).task_thread.lock);
sby = get_frame_progress(c, f);
error_0 =
Expand Down

0 comments on commit 2f9c108

Please sign in to comment.