1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354
use crate::def_id::{DefIndex, LocalDefId};
use crate::hygiene::SyntaxContext;
use crate::SPAN_TRACK;
use crate::{BytePos, SpanData};
use rustc_data_structures::fx::FxIndexSet;
// This code is very hot and uses lots of arithmetic, avoid overflow checks for performance.
// See https://github.com/rust-lang/rust/pull/119440#issuecomment-1874255727
use rustc_serialize::int_overflow::DebugStrictAdd;
/// A compressed span.
///
/// [`SpanData`] is 16 bytes, which is too big to stick everywhere. `Span` only
/// takes up 8 bytes, with less space for the length, parent and context. The
/// vast majority (99.9%+) of `SpanData` instances can be made to fit within
/// those 8 bytes. Any `SpanData` whose fields don't fit into a `Span` are
/// stored in a separate interner table, and the `Span` will index into that
/// table. Interning is rare enough that the cost is low, but common enough
/// that the code is exercised regularly.
///
/// An earlier version of this code used only 4 bytes for `Span`, but that was
/// slower because only 80--90% of spans could be stored inline (even less in
/// very large crates) and so the interner was used a lot more. That version of
/// the code also predated the storage of parents.
///
/// There are four different span forms.
///
/// Inline-context format (requires non-huge length, non-huge context, and no parent):
/// - `span.lo_or_index == span_data.lo`
/// - `span.len_with_tag_or_marker == len == span_data.hi - span_data.lo` (must be `<= MAX_LEN`)
/// - `span.ctxt_or_parent_or_marker == span_data.ctxt` (must be `<= MAX_CTXT`)
///
/// Inline-parent format (requires non-huge length, root context, and non-huge parent):
/// - `span.lo_or_index == span_data.lo`
/// - `span.len_with_tag_or_marker & !PARENT_TAG == len == span_data.hi - span_data.lo`
/// (must be `<= MAX_LEN`)
/// - `span.len_with_tag_or_marker` has top bit (`PARENT_TAG`) set
/// - `span.ctxt_or_parent_or_marker == span_data.parent` (must be `<= MAX_CTXT`)
///
/// Partially-interned format (requires non-huge context):
/// - `span.lo_or_index == index` (indexes into the interner table)
/// - `span.len_with_tag_or_marker == BASE_LEN_INTERNED_MARKER`
/// - `span.ctxt_or_parent_or_marker == span_data.ctxt` (must be `<= MAX_CTXT`)
///
/// Fully-interned format (all cases not covered above):
/// - `span.lo_or_index == index` (indexes into the interner table)
/// - `span.len_with_tag_or_marker == BASE_LEN_INTERNED_MARKER`
/// - `span.ctxt_or_parent_or_marker == CTXT_INTERNED_MARKER`
///
/// The partially-interned form requires looking in the interning table for
/// lo and length, but the context is stored inline as well as interned.
/// This is useful because context lookups are often done in isolation, and
/// inline lookups are quicker.
///
/// Notes about the choice of field sizes:
/// - `lo` is 32 bits in both `Span` and `SpanData`, which means that `lo`
/// values never cause interning. The number of bits needed for `lo`
/// depends on the crate size. 32 bits allows up to 4 GiB of code in a crate.
/// Having no compression on this field means there is no performance cliff
/// if a crate exceeds a particular size.
/// - `len` is ~15 bits in `Span` (a u16, minus 1 bit for PARENT_TAG) and 32
/// bits in `SpanData`, which means that large `len` values will cause
/// interning. The number of bits needed for `len` does not depend on the
/// crate size. The most common numbers of bits for `len` are from 0 to 7,
/// with a peak usually at 3 or 4, and then it drops off quickly from 8
/// onwards. 15 bits is enough for 99.99%+ of cases, but larger values
/// (sometimes 20+ bits) might occur dozens of times in a typical crate.
/// - `ctxt_or_parent_or_marker` is 16 bits in `Span` and two 32 bit fields in
/// `SpanData`, which means intering will happen if `ctxt` is large, if
/// `parent` is large, or if both values are non-zero. The number of bits
/// needed for `ctxt` values depend partly on the crate size and partly on
/// the form of the code. No crates in `rustc-perf` need more than 15 bits
/// for `ctxt_or_parent_or_marker`, but larger crates might need more than 16
/// bits. The number of bits needed for `parent` hasn't been measured,
/// because `parent` isn't currently used by default.
///
/// In order to reliably use parented spans in incremental compilation,
/// the dependency to the parent definition's span. This is performed
/// using the callback `SPAN_TRACK` to access the query engine.
///
#[derive(Clone, Copy, Eq, PartialEq, Hash)]
#[rustc_pass_by_value]
pub struct Span {
lo_or_index: u32,
len_with_tag_or_marker: u16,
ctxt_or_parent_or_marker: u16,
}
impl Span {
#[inline]
fn data_inline_ctxt(self) -> SpanData {
let len = self.len_with_tag_or_marker as u32;
debug_assert!(len <= MAX_LEN);
SpanData {
lo: BytePos(self.lo_or_index),
hi: BytePos(self.lo_or_index.debug_strict_add(len)),
ctxt: SyntaxContext::from_u32(self.ctxt_or_parent_or_marker as u32),
parent: None,
}
}
#[inline]
fn data_inline_parent(self) -> SpanData {
let len = (self.len_with_tag_or_marker & !PARENT_TAG) as u32;
debug_assert!(len <= MAX_LEN);
let parent = LocalDefId {
local_def_index: DefIndex::from_u32(self.ctxt_or_parent_or_marker as u32),
};
SpanData {
lo: BytePos(self.lo_or_index),
hi: BytePos(self.lo_or_index.debug_strict_add(len)),
ctxt: SyntaxContext::root(),
parent: Some(parent),
}
}
#[inline]
fn data_partially_interned(self) -> SpanData {
SpanData {
ctxt: SyntaxContext::from_u32(self.ctxt_or_parent_or_marker as u32),
..with_span_interner(|interner| interner.spans[self.lo_or_index as usize])
}
}
#[inline]
fn data_interned(self) -> SpanData {
with_span_interner(|interner| interner.spans[self.lo_or_index as usize])
}
}
// `MAX_LEN` is chosen so that `PARENT_TAG | MAX_LEN` is distinct from
// `BASE_LEN_INTERNED_MARKER`. (If `MAX_LEN` was 1 higher, this wouldn't be true.)
const MAX_LEN: u32 = 0b0111_1111_1111_1110;
const MAX_CTXT: u32 = 0b0111_1111_1111_1110;
const PARENT_TAG: u16 = 0b1000_0000_0000_0000;
const BASE_LEN_INTERNED_MARKER: u16 = 0b1111_1111_1111_1111;
const CTXT_INTERNED_MARKER: u16 = 0b1111_1111_1111_1111;
/// The dummy span has zero position, length, and context, and no parent.
pub const DUMMY_SP: Span =
Span { lo_or_index: 0, len_with_tag_or_marker: 0, ctxt_or_parent_or_marker: 0 };
impl Span {
#[inline]
pub fn new(
mut lo: BytePos,
mut hi: BytePos,
ctxt: SyntaxContext,
parent: Option<LocalDefId>,
) -> Self {
if lo > hi {
std::mem::swap(&mut lo, &mut hi);
}
// Small len may enable one of fully inline formats (or may not).
let (len, ctxt32) = (hi.0 - lo.0, ctxt.as_u32());
if len <= MAX_LEN {
if ctxt32 <= MAX_CTXT && parent.is_none() {
// Inline-context format.
return Span {
lo_or_index: lo.0,
len_with_tag_or_marker: len as u16,
ctxt_or_parent_or_marker: ctxt32 as u16,
};
} else if ctxt32 == 0
&& let Some(parent) = parent
&& let parent32 = parent.local_def_index.as_u32()
&& parent32 <= MAX_CTXT
{
// Inline-parent format.
return Span {
lo_or_index: lo.0,
len_with_tag_or_marker: PARENT_TAG | len as u16,
ctxt_or_parent_or_marker: parent32 as u16,
};
}
}
// Otherwise small ctxt may enable the partially inline format.
let index = |ctxt| {
with_span_interner(|interner| interner.intern(&SpanData { lo, hi, ctxt, parent }))
};
if ctxt32 <= MAX_CTXT {
// Partially-interned format.
Span {
// Interned ctxt should never be read, so it can use any value.
lo_or_index: index(SyntaxContext::from_u32(u32::MAX)),
len_with_tag_or_marker: BASE_LEN_INTERNED_MARKER,
ctxt_or_parent_or_marker: ctxt32 as u16,
}
} else {
// Interned format.
Span {
lo_or_index: index(ctxt),
len_with_tag_or_marker: BASE_LEN_INTERNED_MARKER,
ctxt_or_parent_or_marker: CTXT_INTERNED_MARKER,
}
}
}
#[inline]
pub fn data(self) -> SpanData {
let data = self.data_untracked();
if let Some(parent) = data.parent {
(*SPAN_TRACK)(parent);
}
data
}
/// Internal function to translate between an encoded span and the expanded representation.
/// This function must not be used outside the incremental engine.
#[inline]
pub fn data_untracked(self) -> SpanData {
if self.len_with_tag_or_marker != BASE_LEN_INTERNED_MARKER {
if self.len_with_tag_or_marker & PARENT_TAG == 0 {
// Inline-context format.
self.data_inline_ctxt()
} else {
// Inline-parent format.
self.data_inline_parent()
}
} else if self.ctxt_or_parent_or_marker != CTXT_INTERNED_MARKER {
// Partially-interned format.
self.data_partially_interned()
} else {
// Interned format.
self.data_interned()
}
}
/// Returns `true` if this is a dummy span with any hygienic context.
#[inline]
pub fn is_dummy(self) -> bool {
if self.len_with_tag_or_marker != BASE_LEN_INTERNED_MARKER {
// Inline-context or inline-parent format.
let lo = self.lo_or_index;
let len = (self.len_with_tag_or_marker & !PARENT_TAG) as u32;
debug_assert!(len <= MAX_LEN);
lo == 0 && len == 0
} else {
// Fully-interned or partially-interned format.
let index = self.lo_or_index;
let data = with_span_interner(|interner| interner.spans[index as usize]);
data.lo == BytePos(0) && data.hi == BytePos(0)
}
}
// For optimization we are interested in cases in which the context is inline and the context
// update doesn't change format. All non-inline or format changing scenarios require accessing
// interner and can fall back to `Span::new`.
#[inline]
pub fn update_ctxt(&mut self, update: impl FnOnce(SyntaxContext) -> SyntaxContext) {
let (updated_ctxt32, data);
if self.len_with_tag_or_marker != BASE_LEN_INTERNED_MARKER {
if self.len_with_tag_or_marker & PARENT_TAG == 0 {
// Inline-context format.
updated_ctxt32 =
update(SyntaxContext::from_u32(self.ctxt_or_parent_or_marker as u32)).as_u32();
// Any small new context including zero will preserve the format.
if updated_ctxt32 <= MAX_CTXT {
self.ctxt_or_parent_or_marker = updated_ctxt32 as u16;
return;
}
data = self.data_inline_ctxt();
} else {
// Inline-parent format.
updated_ctxt32 = update(SyntaxContext::root()).as_u32();
// Only if the new context is zero the format will be preserved.
if updated_ctxt32 == 0 {
// Do nothing.
return;
}
data = self.data_inline_parent();
}
} else if self.ctxt_or_parent_or_marker != CTXT_INTERNED_MARKER {
// Partially-interned format.
updated_ctxt32 =
update(SyntaxContext::from_u32(self.ctxt_or_parent_or_marker as u32)).as_u32();
// Any small new context excluding zero will preserve the format.
// Zero may change the format to `InlineParent` if parent and len are small enough.
if updated_ctxt32 <= MAX_CTXT && updated_ctxt32 != 0 {
self.ctxt_or_parent_or_marker = updated_ctxt32 as u16;
return;
}
data = self.data_partially_interned();
} else {
// Interned format.
data = self.data_interned();
updated_ctxt32 = update(data.ctxt).as_u32();
}
// We could not keep the span in the same inline format, fall back to the complete logic.
*self = data.with_ctxt(SyntaxContext::from_u32(updated_ctxt32));
}
// Returns either syntactic context, if it can be retrieved without taking the interner lock,
// or an index into the interner if it cannot.
#[inline]
fn inline_ctxt(self) -> Result<SyntaxContext, usize> {
if self.len_with_tag_or_marker != BASE_LEN_INTERNED_MARKER {
if self.len_with_tag_or_marker & PARENT_TAG == 0 {
// Inline-context format.
Ok(SyntaxContext::from_u32(self.ctxt_or_parent_or_marker as u32))
} else {
// Inline-parent format.
Ok(SyntaxContext::root())
}
} else if self.ctxt_or_parent_or_marker != CTXT_INTERNED_MARKER {
// Partially-interned format.
Ok(SyntaxContext::from_u32(self.ctxt_or_parent_or_marker as u32))
} else {
// Interned format.
Err(self.lo_or_index as usize)
}
}
/// This function is used as a fast path when decoding the full `SpanData` is not necessary.
/// It's a cut-down version of `data_untracked`.
#[cfg_attr(not(test), rustc_diagnostic_item = "SpanCtxt")]
#[inline]
pub fn ctxt(self) -> SyntaxContext {
self.inline_ctxt()
.unwrap_or_else(|index| with_span_interner(|interner| interner.spans[index].ctxt))
}
#[inline]
pub fn eq_ctxt(self, other: Span) -> bool {
match (self.inline_ctxt(), other.inline_ctxt()) {
(Ok(ctxt1), Ok(ctxt2)) => ctxt1 == ctxt2,
(Ok(ctxt), Err(index)) | (Err(index), Ok(ctxt)) => {
with_span_interner(|interner| ctxt == interner.spans[index].ctxt)
}
(Err(index1), Err(index2)) => with_span_interner(|interner| {
interner.spans[index1].ctxt == interner.spans[index2].ctxt
}),
}
}
}
#[derive(Default)]
pub struct SpanInterner {
spans: FxIndexSet<SpanData>,
}
impl SpanInterner {
fn intern(&mut self, span_data: &SpanData) -> u32 {
let (index, _) = self.spans.insert_full(*span_data);
index as u32
}
}
// If an interner exists, return it. Otherwise, prepare a fresh one.
#[inline]
fn with_span_interner<T, F: FnOnce(&mut SpanInterner) -> T>(f: F) -> T {
crate::with_session_globals(|session_globals| f(&mut session_globals.span_interner.lock()))
}