From 3f80529c643be8449fc755adc6bb37e2ea92114b Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?R=C3=A9my=20Rakic?= Date: Mon, 27 Mar 2023 15:52:17 +0000 Subject: [PATCH] make InitMask lazy for fully init/uninit cases Avoid materializing bits in the InitMask bitset when a single value would be enough: when the mask represents a fully initialized or fully uninitialized const allocation. --- .../src/mir/interpret/allocation/init_mask.rs | 283 ++++++++++++++---- 1 file changed, 225 insertions(+), 58 deletions(-) diff --git a/compiler/rustc_middle/src/mir/interpret/allocation/init_mask.rs b/compiler/rustc_middle/src/mir/interpret/allocation/init_mask.rs index 82e9a961a2b..3f7e2523106 100644 --- a/compiler/rustc_middle/src/mir/interpret/allocation/init_mask.rs +++ b/compiler/rustc_middle/src/mir/interpret/allocation/init_mask.rs @@ -10,20 +10,185 @@ type Block = u64; /// A bitmask where each bit refers to the byte with the same index. If the bit is `true`, the byte /// is initialized. If it is `false` the byte is uninitialized. -// Note: for performance reasons when interning, some of the `InitMask` fields can be partially -// hashed. (see the `Hash` impl below for more details), so the impl is not derived. -#[derive(Clone, Debug, Eq, PartialEq, TyEncodable, TyDecodable)] -#[derive(HashStable)] +/// The actual bits are only materialized when needed, and we try to keep this data lazy as long as +/// possible. Currently, if all the blocks have the same value, then the mask represents either a +/// fully initialized or fully uninitialized const allocation, so we can only store that single +/// value. +#[derive(Clone, Debug, Eq, PartialEq, TyEncodable, TyDecodable, Hash, HashStable)] pub struct InitMask { - blocks: Vec, + pub(crate) blocks: InitMaskBlocks, len: Size, } +#[derive(Clone, Debug, Eq, PartialEq, TyEncodable, TyDecodable, Hash, HashStable)] +pub(crate) enum InitMaskBlocks { + Lazy { + /// Whether the lazy init mask is fully initialized or uninitialized. + state: bool, + }, + Materialized(InitMaskMaterialized), +} + +impl InitMask { + pub fn new(size: Size, state: bool) -> Self { + // Blocks start lazily allocated, until we have to materialize them. + let blocks = InitMaskBlocks::Lazy { state }; + InitMask { len: size, blocks } + } + + /// Checks whether the `range` is entirely initialized. + /// + /// Returns `Ok(())` if it's initialized. Otherwise returns a range of byte + /// indexes for the first contiguous span of the uninitialized access. + #[inline] + pub fn is_range_initialized(&self, range: AllocRange) -> Result<(), AllocRange> { + let end = range.end(); + if end > self.len { + return Err(AllocRange::from(self.len..end)); + } + + match self.blocks { + InitMaskBlocks::Lazy { state } => { + // Lazily allocated blocks represent the full mask, and cover the requested range by + // definition. + if state { Ok(()) } else { Err(range) } + } + InitMaskBlocks::Materialized(ref blocks) => { + blocks.is_range_initialized(range.start, end) + } + } + } + + /// Sets a specified range to a value. If the range is out-of-bounds, the mask will grow to + /// accomodate it entirely. + pub fn set_range(&mut self, range: AllocRange, new_state: bool) { + let start = range.start; + let end = range.end(); + + let is_full_overwrite = start == Size::ZERO && end >= self.len; + + // Optimize the cases of a full init/uninit state, while handling growth if needed. + match self.blocks { + InitMaskBlocks::Lazy { ref mut state } if is_full_overwrite => { + // This is fully overwriting the mask, and we'll still have a single initialization + // state: the blocks can stay lazy. + *state = new_state; + self.len = end; + } + InitMaskBlocks::Materialized(_) if is_full_overwrite => { + // This is also fully overwriting materialized blocks with a single initialization + // state: we'll have no need for these blocks anymore and can make them lazy. + self.blocks = InitMaskBlocks::Lazy { state: new_state }; + self.len = end; + } + InitMaskBlocks::Lazy { state } if state == new_state => { + // Here we're partially overwriting the mask but the initialization state doesn't + // change: the blocks can stay lazy. + if end > self.len { + self.len = end; + } + } + _ => { + // Otherwise, we have a partial overwrite that can result in a mix of initialization + // states, so we'll need materialized blocks. + let len = self.len; + let blocks = self.materialize_blocks(); + + // There are 3 cases of interest here, if we have: + // + // [--------] + // ^ ^ + // 0 len + // + // 1) the range to set can be in-bounds: + // + // xxxx = [start, end] + // [--------] + // ^ ^ + // 0 len + // + // Here, we'll simply set the single `start` to `end` range. + // + // 2) the range to set can be partially out-of-bounds: + // + // xxxx = [start, end] + // [--------] + // ^ ^ + // 0 len + // + // We have 2 subranges to handle: + // - we'll set the existing `start` to `len` range. + // - we'll grow and set the `len` to `end` range. + // + // 3) the range to set can be fully out-of-bounds: + // + // ---xxxx = [start, end] + // [--------] + // ^ ^ + // 0 len + // + // Since we're growing the mask to a single `new_state` value, we consider the gap + // from `len` to `start` to be part of the range, and have a single subrange to + // handle: we'll grow and set the `len` to `end` range. + // + // Note that we have to materialize, set blocks, and grow the mask. We could + // therefore slightly optimize things in situations where these writes overlap. + // However, as of writing this, growing the mask doesn't happen in practice yet, so + // we don't do this micro-optimization. + + if end <= len { + // Handle case 1. + blocks.set_range_inbounds(start, end, new_state); + } else { + if start < len { + // Handle the first subrange of case 2. + blocks.set_range_inbounds(start, len, new_state); + } + + // Handle the second subrange of case 2, and case 3. + blocks.grow(len, end - len, new_state); // `Size` operation + self.len = end; + } + } + } + } + + /// Materializes this mask's blocks when the mask is lazy. + #[inline] + fn materialize_blocks(&mut self) -> &mut InitMaskMaterialized { + if let InitMaskBlocks::Lazy { state } = self.blocks { + self.blocks = InitMaskBlocks::Materialized(InitMaskMaterialized::new(self.len, state)); + } + + let InitMaskBlocks::Materialized(ref mut blocks) = self.blocks else { + bug!("initmask blocks must be materialized here") + }; + blocks + } + + /// Returns the initialization state at the specified in-bounds index. + #[inline] + pub fn get(&self, idx: Size) -> bool { + match self.blocks { + InitMaskBlocks::Lazy { state } => state, + InitMaskBlocks::Materialized(ref blocks) => blocks.get(idx), + } + } +} + +/// The actual materialized blocks of the bitmask, when we can't keep the `InitMask` lazy. +// Note: for performance reasons when interning, some of the fields can be partially +// hashed. (see the `Hash` impl below for more details), so the impl is not derived. +#[derive(Clone, Debug, Eq, PartialEq, TyEncodable, TyDecodable, HashStable)] +pub(crate) struct InitMaskMaterialized { + pub(crate) blocks: Vec, +} + // Const allocations are only hashed for interning. However, they can be large, making the hashing // expensive especially since it uses `FxHash`: it's better suited to short keys, not potentially // big buffers like the allocation's init mask. We can partially hash some fields when they're // large. -impl hash::Hash for InitMask { +impl hash::Hash for InitMaskMaterialized { fn hash(&self, state: &mut H) { const MAX_BLOCKS_TO_HASH: usize = super::MAX_BYTES_TO_HASH / std::mem::size_of::(); const MAX_BLOCKS_LEN: usize = super::MAX_HASHED_BUFFER_LEN / std::mem::size_of::(); @@ -41,18 +206,15 @@ impl hash::Hash for InitMask { } else { self.blocks.hash(state); } - - // Hash the other fields as usual. - self.len.hash(state); } } -impl InitMask { +impl InitMaskMaterialized { pub const BLOCK_SIZE: u64 = 64; - pub fn new(size: Size, state: bool) -> Self { - let mut m = InitMask { blocks: vec![], len: Size::ZERO }; - m.grow(size, state); + fn new(size: Size, state: bool) -> Self { + let mut m = InitMaskMaterialized { blocks: vec![] }; + m.grow(Size::ZERO, size, state); m } @@ -62,8 +224,8 @@ impl InitMask { // Each bit in a `Block` represents the initialization state of one byte of an allocation, // so we use `.bytes()` here. let bits = bits.bytes(); - let a = bits / InitMask::BLOCK_SIZE; - let b = bits % InitMask::BLOCK_SIZE; + let a = bits / Self::BLOCK_SIZE; + let b = bits % Self::BLOCK_SIZE; (usize::try_from(a).unwrap(), usize::try_from(b).unwrap()) } @@ -71,7 +233,7 @@ impl InitMask { fn size_from_bit_index(block: impl TryInto, bit: impl TryInto) -> Size { let block = block.try_into().ok().unwrap(); let bit = bit.try_into().ok().unwrap(); - Size::from_bytes(block * InitMask::BLOCK_SIZE + bit) + Size::from_bytes(block * Self::BLOCK_SIZE + bit) } /// Checks whether the `range` is entirely initialized. @@ -79,13 +241,8 @@ impl InitMask { /// Returns `Ok(())` if it's initialized. Otherwise returns a range of byte /// indexes for the first contiguous span of the uninitialized access. #[inline] - pub fn is_range_initialized(&self, range: AllocRange) -> Result<(), AllocRange> { - let end = range.end(); - if end > self.len { - return Err(AllocRange::from(self.len..end)); - } - - let uninit_start = self.find_bit(range.start, end, false); + fn is_range_initialized(&self, start: Size, end: Size) -> Result<(), AllocRange> { + let uninit_start = self.find_bit(start, end, false); match uninit_start { Some(uninit_start) => { @@ -96,15 +253,6 @@ impl InitMask { } } - pub fn set_range(&mut self, range: AllocRange, new_state: bool) { - let end = range.end(); - let len = self.len; - if end > len { - self.grow(end - len, new_state); - } - self.set_range_inbounds(range.start, end, new_state); - } - fn set_range_inbounds(&mut self, start: Size, end: Size, new_state: bool) { let (blocka, bita) = Self::bit_index(start); let (blockb, bitb) = Self::bit_index(end); @@ -150,27 +298,35 @@ impl InitMask { } #[inline] - pub fn get(&self, i: Size) -> bool { + fn get(&self, i: Size) -> bool { let (block, bit) = Self::bit_index(i); (self.blocks[block] & (1 << bit)) != 0 } - fn grow(&mut self, amount: Size, new_state: bool) { + fn grow(&mut self, len: Size, amount: Size, new_state: bool) { if amount.bytes() == 0 { return; } let unused_trailing_bits = - u64::try_from(self.blocks.len()).unwrap() * Self::BLOCK_SIZE - self.len.bytes(); + u64::try_from(self.blocks.len()).unwrap() * Self::BLOCK_SIZE - len.bytes(); + + // If there's not enough capacity in the currently allocated blocks, allocate some more. if amount.bytes() > unused_trailing_bits { let additional_blocks = amount.bytes() / Self::BLOCK_SIZE + 1; - self.blocks.extend( - // FIXME(oli-obk): optimize this by repeating `new_state as Block`. - iter::repeat(0).take(usize::try_from(additional_blocks).unwrap()), - ); + + // We allocate the blocks to the correct value for the requested init state, so we won't + // have to manually set them with another write. + let block = if new_state { u64::MAX } else { 0 }; + self.blocks + .extend(iter::repeat(block).take(usize::try_from(additional_blocks).unwrap())); + } + + // New blocks have already been set here, so we only need to set the unused trailing bits, + // if any. + if unused_trailing_bits > 0 { + let in_bounds_tail = Size::from_bytes(unused_trailing_bits); + self.set_range_inbounds(len, len + in_bounds_tail, new_state); // `Size` operation } - let start = self.len; - self.len += amount; - self.set_range_inbounds(start, start + amount, new_state); // `Size` operation } /// Returns the index of the first bit in `start..end` (end-exclusive) that is equal to is_init. @@ -188,7 +344,7 @@ impl InitMask { /// ``` /// Also, if not stated, assume that `is_init = true`, that is, we are searching for the first 1 bit. fn find_bit_fast( - init_mask: &InitMask, + init_mask: &InitMaskMaterialized, start: Size, end: Size, is_init: bool, @@ -223,7 +379,7 @@ impl InitMask { None } else { let bit = bits.trailing_zeros(); - Some(InitMask::size_from_bit_index(block, bit)) + Some(InitMaskMaterialized::size_from_bit_index(block, bit)) } } @@ -253,9 +409,9 @@ impl InitMask { // This provides the desired behavior of searching blocks 0 and 1 for (a), // and searching only block 0 for (b). // There is no concern of overflows since we checked for `start >= end` above. - let (start_block, start_bit) = InitMask::bit_index(start); + let (start_block, start_bit) = InitMaskMaterialized::bit_index(start); let end_inclusive = Size::from_bytes(end.bytes() - 1); - let (end_block_inclusive, _) = InitMask::bit_index(end_inclusive); + let (end_block_inclusive, _) = InitMaskMaterialized::bit_index(end_inclusive); // Handle first block: need to skip `start_bit` bits. // @@ -340,7 +496,7 @@ impl InitMask { #[cfg_attr(not(debug_assertions), allow(dead_code))] fn find_bit_slow( - init_mask: &InitMask, + init_mask: &InitMaskMaterialized, start: Size, end: Size, is_init: bool, @@ -436,10 +592,19 @@ impl<'a> Iterator for InitChunkIter<'a> { return None; } - let end_of_chunk = - self.init_mask.find_bit(self.start, self.end, !self.is_init).unwrap_or(self.end); + let end_of_chunk = match self.init_mask.blocks { + InitMaskBlocks::Lazy { .. } => { + // If we're iterating over the chunks of lazy blocks, we just emit a single + // full-size chunk. + self.end + } + InitMaskBlocks::Materialized(ref blocks) => { + let end_of_chunk = + blocks.find_bit(self.start, self.end, !self.is_init).unwrap_or(self.end); + end_of_chunk + } + }; let range = self.start..end_of_chunk; - let ret = Some(if self.is_init { InitChunk::Init(range) } else { InitChunk::Uninit(range) }); @@ -504,17 +669,19 @@ impl InitMask { /// Applies multiple instances of the run-length encoding to the initialization mask. pub fn apply_copy(&mut self, defined: InitCopy, range: AllocRange, repeat: u64) { - // An optimization where we can just overwrite an entire range of initialization - // bits if they are going to be uniformly `1` or `0`. + // An optimization where we can just overwrite an entire range of initialization bits if + // they are going to be uniformly `1` or `0`. If this happens to be a full-range overwrite, + // we won't need materialized blocks either. if defined.ranges.len() <= 1 { - self.set_range_inbounds( - range.start, - range.start + range.size * repeat, // `Size` operations - defined.initial, - ); + let start = range.start; + let end = range.start + range.size * repeat; // `Size` operations + self.set_range(AllocRange::from(start..end), defined.initial); return; } + // We're about to do one or more partial writes, so we ensure the blocks are materialized. + let blocks = self.materialize_blocks(); + for mut j in 0..repeat { j *= range.size.bytes(); j += range.start.bytes(); @@ -522,7 +689,7 @@ impl InitMask { for range in &defined.ranges { let old_j = j; j += range; - self.set_range_inbounds(Size::from_bytes(old_j), Size::from_bytes(j), cur); + blocks.set_range_inbounds(Size::from_bytes(old_j), Size::from_bytes(j), cur); cur = !cur; } }