513 lines
19 KiB
Rust
Raw Normal View History

Revised mir-dataflow. Incorporates many fixes contributed by arielb1. ---- revise borrowck::mir::dataflow code to allow varying domain for bitvectors. This particular code implements the `BitDenotation` trait for three analyses: * `MovingOutStatements`, which, like `borrowck::move_data`, maps each bit-index to a move instruction, and a 1 means "the effect of this move reaches this point" (and the assigned l-value, if a scoped declaration, is still in scope). * `MaybeInitializedLvals`, which maps each bit-index to an l-value. A 1 means "there exists a control flow path to this point that initializes the associated l-value." * `MaybeUninitializedLvals`, which maps each bit-index to an l-value A 1 means "there exists a control flow path to this point that de-initializes the associated l-value." ---- Revised `graphviz` dataflow-rendering support in `borrowck::mir`. One big difference is that this code is now parameterized over the `BitDenotation`, so that it can be used to render dataflow results independent of how the dataflow bitvectors are interpreted; see where reference to `MoveOut` is replaced by the type parameter `D`. ---- Factor out routine to query subattributes in `#[rustc_mir(..)]`. (Later commits build upon this for some unit testing and instrumentation.) ---- thread through a tcx so that I can query types of lvalues as part of analysis. ---- Revised `BitDenotation::Ctxt`, allowing variation beyond `MoveData`. The main motivation is to ease threading through a `TyCtxt`. (In hindsight it might have been better to instead attach the `TyCtxt` to each of the different dataflow implementations, but that would require e.g. switching away from having a `Default` impl, so I am leaving that experiment for another time.)
2016-05-02 15:50:27 +02:00
// Copyright 2012-2016 The Rust Project Developers. See the COPYRIGHT
// file at the top-level directory of this distribution and at
// http://rust-lang.org/COPYRIGHT.
//
// Licensed under the Apache License, Version 2.0 <LICENSE-APACHE or
// http://www.apache.org/licenses/LICENSE-2.0> or the MIT license
// <LICENSE-MIT or http://opensource.org/licenses/MIT>, at your
// option. This file may not be copied, modified, or distributed
// except according to those terms.
use rustc::ty::TyCtxt;
use rustc::mir::repr::{self, Mir};
use std::fmt::Debug;
use std::io;
use std::mem;
use std::path::PathBuf;
use std::usize;
use super::MirBorrowckCtxtPreDataflow;
use super::gather_moves::{MoveData};
Revised mir-dataflow. Incorporates many fixes contributed by arielb1. ---- revise borrowck::mir::dataflow code to allow varying domain for bitvectors. This particular code implements the `BitDenotation` trait for three analyses: * `MovingOutStatements`, which, like `borrowck::move_data`, maps each bit-index to a move instruction, and a 1 means "the effect of this move reaches this point" (and the assigned l-value, if a scoped declaration, is still in scope). * `MaybeInitializedLvals`, which maps each bit-index to an l-value. A 1 means "there exists a control flow path to this point that initializes the associated l-value." * `MaybeUninitializedLvals`, which maps each bit-index to an l-value A 1 means "there exists a control flow path to this point that de-initializes the associated l-value." ---- Revised `graphviz` dataflow-rendering support in `borrowck::mir`. One big difference is that this code is now parameterized over the `BitDenotation`, so that it can be used to render dataflow results independent of how the dataflow bitvectors are interpreted; see where reference to `MoveOut` is replaced by the type parameter `D`. ---- Factor out routine to query subattributes in `#[rustc_mir(..)]`. (Later commits build upon this for some unit testing and instrumentation.) ---- thread through a tcx so that I can query types of lvalues as part of analysis. ---- Revised `BitDenotation::Ctxt`, allowing variation beyond `MoveData`. The main motivation is to ease threading through a `TyCtxt`. (In hindsight it might have been better to instead attach the `TyCtxt` to each of the different dataflow implementations, but that would require e.g. switching away from having a `Default` impl, so I am leaving that experiment for another time.)
2016-05-02 15:50:27 +02:00
use bitslice::{bitwise, BitwiseOperator};
use indexed_set::{Idx, IdxSet, OwnIdxSet};
Revised mir-dataflow. Incorporates many fixes contributed by arielb1. ---- revise borrowck::mir::dataflow code to allow varying domain for bitvectors. This particular code implements the `BitDenotation` trait for three analyses: * `MovingOutStatements`, which, like `borrowck::move_data`, maps each bit-index to a move instruction, and a 1 means "the effect of this move reaches this point" (and the assigned l-value, if a scoped declaration, is still in scope). * `MaybeInitializedLvals`, which maps each bit-index to an l-value. A 1 means "there exists a control flow path to this point that initializes the associated l-value." * `MaybeUninitializedLvals`, which maps each bit-index to an l-value A 1 means "there exists a control flow path to this point that de-initializes the associated l-value." ---- Revised `graphviz` dataflow-rendering support in `borrowck::mir`. One big difference is that this code is now parameterized over the `BitDenotation`, so that it can be used to render dataflow results independent of how the dataflow bitvectors are interpreted; see where reference to `MoveOut` is replaced by the type parameter `D`. ---- Factor out routine to query subattributes in `#[rustc_mir(..)]`. (Later commits build upon this for some unit testing and instrumentation.) ---- thread through a tcx so that I can query types of lvalues as part of analysis. ---- Revised `BitDenotation::Ctxt`, allowing variation beyond `MoveData`. The main motivation is to ease threading through a `TyCtxt`. (In hindsight it might have been better to instead attach the `TyCtxt` to each of the different dataflow implementations, but that would require e.g. switching away from having a `Default` impl, so I am leaving that experiment for another time.)
2016-05-02 15:50:27 +02:00
pub use self::sanity_check::sanity_check_via_rustc_peek;
pub use self::impls::{MaybeInitializedLvals, MaybeUninitializedLvals};
pub use self::impls::{DefinitelyInitializedLvals, MovingOutStatements};
Revised mir-dataflow. Incorporates many fixes contributed by arielb1. ---- revise borrowck::mir::dataflow code to allow varying domain for bitvectors. This particular code implements the `BitDenotation` trait for three analyses: * `MovingOutStatements`, which, like `borrowck::move_data`, maps each bit-index to a move instruction, and a 1 means "the effect of this move reaches this point" (and the assigned l-value, if a scoped declaration, is still in scope). * `MaybeInitializedLvals`, which maps each bit-index to an l-value. A 1 means "there exists a control flow path to this point that initializes the associated l-value." * `MaybeUninitializedLvals`, which maps each bit-index to an l-value A 1 means "there exists a control flow path to this point that de-initializes the associated l-value." ---- Revised `graphviz` dataflow-rendering support in `borrowck::mir`. One big difference is that this code is now parameterized over the `BitDenotation`, so that it can be used to render dataflow results independent of how the dataflow bitvectors are interpreted; see where reference to `MoveOut` is replaced by the type parameter `D`. ---- Factor out routine to query subattributes in `#[rustc_mir(..)]`. (Later commits build upon this for some unit testing and instrumentation.) ---- thread through a tcx so that I can query types of lvalues as part of analysis. ---- Revised `BitDenotation::Ctxt`, allowing variation beyond `MoveData`. The main motivation is to ease threading through a `TyCtxt`. (In hindsight it might have been better to instead attach the `TyCtxt` to each of the different dataflow implementations, but that would require e.g. switching away from having a `Default` impl, so I am leaving that experiment for another time.)
2016-05-02 15:50:27 +02:00
mod graphviz;
mod sanity_check;
mod impls;
Revised mir-dataflow. Incorporates many fixes contributed by arielb1. ---- revise borrowck::mir::dataflow code to allow varying domain for bitvectors. This particular code implements the `BitDenotation` trait for three analyses: * `MovingOutStatements`, which, like `borrowck::move_data`, maps each bit-index to a move instruction, and a 1 means "the effect of this move reaches this point" (and the assigned l-value, if a scoped declaration, is still in scope). * `MaybeInitializedLvals`, which maps each bit-index to an l-value. A 1 means "there exists a control flow path to this point that initializes the associated l-value." * `MaybeUninitializedLvals`, which maps each bit-index to an l-value A 1 means "there exists a control flow path to this point that de-initializes the associated l-value." ---- Revised `graphviz` dataflow-rendering support in `borrowck::mir`. One big difference is that this code is now parameterized over the `BitDenotation`, so that it can be used to render dataflow results independent of how the dataflow bitvectors are interpreted; see where reference to `MoveOut` is replaced by the type parameter `D`. ---- Factor out routine to query subattributes in `#[rustc_mir(..)]`. (Later commits build upon this for some unit testing and instrumentation.) ---- thread through a tcx so that I can query types of lvalues as part of analysis. ---- Revised `BitDenotation::Ctxt`, allowing variation beyond `MoveData`. The main motivation is to ease threading through a `TyCtxt`. (In hindsight it might have been better to instead attach the `TyCtxt` to each of the different dataflow implementations, but that would require e.g. switching away from having a `Default` impl, so I am leaving that experiment for another time.)
2016-05-02 15:50:27 +02:00
pub trait Dataflow<BD: BitDenotation> {
fn dataflow<P>(&mut self, p: P) where P: Fn(&BD::Ctxt, BD::Idx) -> &Debug;
Revised mir-dataflow. Incorporates many fixes contributed by arielb1. ---- revise borrowck::mir::dataflow code to allow varying domain for bitvectors. This particular code implements the `BitDenotation` trait for three analyses: * `MovingOutStatements`, which, like `borrowck::move_data`, maps each bit-index to a move instruction, and a 1 means "the effect of this move reaches this point" (and the assigned l-value, if a scoped declaration, is still in scope). * `MaybeInitializedLvals`, which maps each bit-index to an l-value. A 1 means "there exists a control flow path to this point that initializes the associated l-value." * `MaybeUninitializedLvals`, which maps each bit-index to an l-value A 1 means "there exists a control flow path to this point that de-initializes the associated l-value." ---- Revised `graphviz` dataflow-rendering support in `borrowck::mir`. One big difference is that this code is now parameterized over the `BitDenotation`, so that it can be used to render dataflow results independent of how the dataflow bitvectors are interpreted; see where reference to `MoveOut` is replaced by the type parameter `D`. ---- Factor out routine to query subattributes in `#[rustc_mir(..)]`. (Later commits build upon this for some unit testing and instrumentation.) ---- thread through a tcx so that I can query types of lvalues as part of analysis. ---- Revised `BitDenotation::Ctxt`, allowing variation beyond `MoveData`. The main motivation is to ease threading through a `TyCtxt`. (In hindsight it might have been better to instead attach the `TyCtxt` to each of the different dataflow implementations, but that would require e.g. switching away from having a `Default` impl, so I am leaving that experiment for another time.)
2016-05-02 15:50:27 +02:00
}
impl<'a, 'tcx: 'a, BD> Dataflow<BD> for MirBorrowckCtxtPreDataflow<'a, 'tcx, BD>
where BD: BitDenotation<Ctxt=MoveData<'tcx>> + DataflowOperator,
BD::Bit: Debug,
Revised mir-dataflow. Incorporates many fixes contributed by arielb1. ---- revise borrowck::mir::dataflow code to allow varying domain for bitvectors. This particular code implements the `BitDenotation` trait for three analyses: * `MovingOutStatements`, which, like `borrowck::move_data`, maps each bit-index to a move instruction, and a 1 means "the effect of this move reaches this point" (and the assigned l-value, if a scoped declaration, is still in scope). * `MaybeInitializedLvals`, which maps each bit-index to an l-value. A 1 means "there exists a control flow path to this point that initializes the associated l-value." * `MaybeUninitializedLvals`, which maps each bit-index to an l-value A 1 means "there exists a control flow path to this point that de-initializes the associated l-value." ---- Revised `graphviz` dataflow-rendering support in `borrowck::mir`. One big difference is that this code is now parameterized over the `BitDenotation`, so that it can be used to render dataflow results independent of how the dataflow bitvectors are interpreted; see where reference to `MoveOut` is replaced by the type parameter `D`. ---- Factor out routine to query subattributes in `#[rustc_mir(..)]`. (Later commits build upon this for some unit testing and instrumentation.) ---- thread through a tcx so that I can query types of lvalues as part of analysis. ---- Revised `BitDenotation::Ctxt`, allowing variation beyond `MoveData`. The main motivation is to ease threading through a `TyCtxt`. (In hindsight it might have been better to instead attach the `TyCtxt` to each of the different dataflow implementations, but that would require e.g. switching away from having a `Default` impl, so I am leaving that experiment for another time.)
2016-05-02 15:50:27 +02:00
{
fn dataflow<P>(&mut self, p: P) where P: Fn(&BD::Ctxt, BD::Idx) -> &Debug {
Revised mir-dataflow. Incorporates many fixes contributed by arielb1. ---- revise borrowck::mir::dataflow code to allow varying domain for bitvectors. This particular code implements the `BitDenotation` trait for three analyses: * `MovingOutStatements`, which, like `borrowck::move_data`, maps each bit-index to a move instruction, and a 1 means "the effect of this move reaches this point" (and the assigned l-value, if a scoped declaration, is still in scope). * `MaybeInitializedLvals`, which maps each bit-index to an l-value. A 1 means "there exists a control flow path to this point that initializes the associated l-value." * `MaybeUninitializedLvals`, which maps each bit-index to an l-value A 1 means "there exists a control flow path to this point that de-initializes the associated l-value." ---- Revised `graphviz` dataflow-rendering support in `borrowck::mir`. One big difference is that this code is now parameterized over the `BitDenotation`, so that it can be used to render dataflow results independent of how the dataflow bitvectors are interpreted; see where reference to `MoveOut` is replaced by the type parameter `D`. ---- Factor out routine to query subattributes in `#[rustc_mir(..)]`. (Later commits build upon this for some unit testing and instrumentation.) ---- thread through a tcx so that I can query types of lvalues as part of analysis. ---- Revised `BitDenotation::Ctxt`, allowing variation beyond `MoveData`. The main motivation is to ease threading through a `TyCtxt`. (In hindsight it might have been better to instead attach the `TyCtxt` to each of the different dataflow implementations, but that would require e.g. switching away from having a `Default` impl, so I am leaving that experiment for another time.)
2016-05-02 15:50:27 +02:00
self.flow_state.build_sets();
self.pre_dataflow_instrumentation(|c,i| p(c,i)).unwrap();
Revised mir-dataflow. Incorporates many fixes contributed by arielb1. ---- revise borrowck::mir::dataflow code to allow varying domain for bitvectors. This particular code implements the `BitDenotation` trait for three analyses: * `MovingOutStatements`, which, like `borrowck::move_data`, maps each bit-index to a move instruction, and a 1 means "the effect of this move reaches this point" (and the assigned l-value, if a scoped declaration, is still in scope). * `MaybeInitializedLvals`, which maps each bit-index to an l-value. A 1 means "there exists a control flow path to this point that initializes the associated l-value." * `MaybeUninitializedLvals`, which maps each bit-index to an l-value A 1 means "there exists a control flow path to this point that de-initializes the associated l-value." ---- Revised `graphviz` dataflow-rendering support in `borrowck::mir`. One big difference is that this code is now parameterized over the `BitDenotation`, so that it can be used to render dataflow results independent of how the dataflow bitvectors are interpreted; see where reference to `MoveOut` is replaced by the type parameter `D`. ---- Factor out routine to query subattributes in `#[rustc_mir(..)]`. (Later commits build upon this for some unit testing and instrumentation.) ---- thread through a tcx so that I can query types of lvalues as part of analysis. ---- Revised `BitDenotation::Ctxt`, allowing variation beyond `MoveData`. The main motivation is to ease threading through a `TyCtxt`. (In hindsight it might have been better to instead attach the `TyCtxt` to each of the different dataflow implementations, but that would require e.g. switching away from having a `Default` impl, so I am leaving that experiment for another time.)
2016-05-02 15:50:27 +02:00
self.flow_state.propagate();
self.post_dataflow_instrumentation(|c,i| p(c,i)).unwrap();
Revised mir-dataflow. Incorporates many fixes contributed by arielb1. ---- revise borrowck::mir::dataflow code to allow varying domain for bitvectors. This particular code implements the `BitDenotation` trait for three analyses: * `MovingOutStatements`, which, like `borrowck::move_data`, maps each bit-index to a move instruction, and a 1 means "the effect of this move reaches this point" (and the assigned l-value, if a scoped declaration, is still in scope). * `MaybeInitializedLvals`, which maps each bit-index to an l-value. A 1 means "there exists a control flow path to this point that initializes the associated l-value." * `MaybeUninitializedLvals`, which maps each bit-index to an l-value A 1 means "there exists a control flow path to this point that de-initializes the associated l-value." ---- Revised `graphviz` dataflow-rendering support in `borrowck::mir`. One big difference is that this code is now parameterized over the `BitDenotation`, so that it can be used to render dataflow results independent of how the dataflow bitvectors are interpreted; see where reference to `MoveOut` is replaced by the type parameter `D`. ---- Factor out routine to query subattributes in `#[rustc_mir(..)]`. (Later commits build upon this for some unit testing and instrumentation.) ---- thread through a tcx so that I can query types of lvalues as part of analysis. ---- Revised `BitDenotation::Ctxt`, allowing variation beyond `MoveData`. The main motivation is to ease threading through a `TyCtxt`. (In hindsight it might have been better to instead attach the `TyCtxt` to each of the different dataflow implementations, but that would require e.g. switching away from having a `Default` impl, so I am leaving that experiment for another time.)
2016-05-02 15:50:27 +02:00
}
}
struct PropagationContext<'b, 'a: 'b, 'tcx: 'a, O>
where O: 'b + BitDenotation, O::Ctxt: 'a+HasMoveData<'tcx>
Revised mir-dataflow. Incorporates many fixes contributed by arielb1. ---- revise borrowck::mir::dataflow code to allow varying domain for bitvectors. This particular code implements the `BitDenotation` trait for three analyses: * `MovingOutStatements`, which, like `borrowck::move_data`, maps each bit-index to a move instruction, and a 1 means "the effect of this move reaches this point" (and the assigned l-value, if a scoped declaration, is still in scope). * `MaybeInitializedLvals`, which maps each bit-index to an l-value. A 1 means "there exists a control flow path to this point that initializes the associated l-value." * `MaybeUninitializedLvals`, which maps each bit-index to an l-value A 1 means "there exists a control flow path to this point that de-initializes the associated l-value." ---- Revised `graphviz` dataflow-rendering support in `borrowck::mir`. One big difference is that this code is now parameterized over the `BitDenotation`, so that it can be used to render dataflow results independent of how the dataflow bitvectors are interpreted; see where reference to `MoveOut` is replaced by the type parameter `D`. ---- Factor out routine to query subattributes in `#[rustc_mir(..)]`. (Later commits build upon this for some unit testing and instrumentation.) ---- thread through a tcx so that I can query types of lvalues as part of analysis. ---- Revised `BitDenotation::Ctxt`, allowing variation beyond `MoveData`. The main motivation is to ease threading through a `TyCtxt`. (In hindsight it might have been better to instead attach the `TyCtxt` to each of the different dataflow implementations, but that would require e.g. switching away from having a `Default` impl, so I am leaving that experiment for another time.)
2016-05-02 15:50:27 +02:00
{
builder: &'b mut DataflowAnalysis<'a, 'tcx, O>,
changed: bool,
}
impl<'a, 'tcx: 'a, BD> DataflowAnalysis<'a, 'tcx, BD>
where BD: BitDenotation + DataflowOperator, BD::Ctxt: HasMoveData<'tcx>
Revised mir-dataflow. Incorporates many fixes contributed by arielb1. ---- revise borrowck::mir::dataflow code to allow varying domain for bitvectors. This particular code implements the `BitDenotation` trait for three analyses: * `MovingOutStatements`, which, like `borrowck::move_data`, maps each bit-index to a move instruction, and a 1 means "the effect of this move reaches this point" (and the assigned l-value, if a scoped declaration, is still in scope). * `MaybeInitializedLvals`, which maps each bit-index to an l-value. A 1 means "there exists a control flow path to this point that initializes the associated l-value." * `MaybeUninitializedLvals`, which maps each bit-index to an l-value A 1 means "there exists a control flow path to this point that de-initializes the associated l-value." ---- Revised `graphviz` dataflow-rendering support in `borrowck::mir`. One big difference is that this code is now parameterized over the `BitDenotation`, so that it can be used to render dataflow results independent of how the dataflow bitvectors are interpreted; see where reference to `MoveOut` is replaced by the type parameter `D`. ---- Factor out routine to query subattributes in `#[rustc_mir(..)]`. (Later commits build upon this for some unit testing and instrumentation.) ---- thread through a tcx so that I can query types of lvalues as part of analysis. ---- Revised `BitDenotation::Ctxt`, allowing variation beyond `MoveData`. The main motivation is to ease threading through a `TyCtxt`. (In hindsight it might have been better to instead attach the `TyCtxt` to each of the different dataflow implementations, but that would require e.g. switching away from having a `Default` impl, so I am leaving that experiment for another time.)
2016-05-02 15:50:27 +02:00
{
fn propagate(&mut self) {
let mut temp = OwnIdxSet::new_empty(self.flow_state.sets.bits_per_block);
Revised mir-dataflow. Incorporates many fixes contributed by arielb1. ---- revise borrowck::mir::dataflow code to allow varying domain for bitvectors. This particular code implements the `BitDenotation` trait for three analyses: * `MovingOutStatements`, which, like `borrowck::move_data`, maps each bit-index to a move instruction, and a 1 means "the effect of this move reaches this point" (and the assigned l-value, if a scoped declaration, is still in scope). * `MaybeInitializedLvals`, which maps each bit-index to an l-value. A 1 means "there exists a control flow path to this point that initializes the associated l-value." * `MaybeUninitializedLvals`, which maps each bit-index to an l-value A 1 means "there exists a control flow path to this point that de-initializes the associated l-value." ---- Revised `graphviz` dataflow-rendering support in `borrowck::mir`. One big difference is that this code is now parameterized over the `BitDenotation`, so that it can be used to render dataflow results independent of how the dataflow bitvectors are interpreted; see where reference to `MoveOut` is replaced by the type parameter `D`. ---- Factor out routine to query subattributes in `#[rustc_mir(..)]`. (Later commits build upon this for some unit testing and instrumentation.) ---- thread through a tcx so that I can query types of lvalues as part of analysis. ---- Revised `BitDenotation::Ctxt`, allowing variation beyond `MoveData`. The main motivation is to ease threading through a `TyCtxt`. (In hindsight it might have been better to instead attach the `TyCtxt` to each of the different dataflow implementations, but that would require e.g. switching away from having a `Default` impl, so I am leaving that experiment for another time.)
2016-05-02 15:50:27 +02:00
let mut propcx = PropagationContext {
builder: self,
changed: true,
};
while propcx.changed {
propcx.changed = false;
propcx.reset(&mut temp);
propcx.walk_cfg(&mut temp);
}
}
fn build_sets(&mut self) {
// First we need to build the entry-, gen- and kill-sets. The
// gather_moves information provides a high-level mapping from
// mir-locations to the MoveOuts (and those correspond
// directly to gen-sets here). But we still need to figure out
// the kill-sets.
{
let sets = &mut self.flow_state.sets.for_block(repr::START_BLOCK.index());
self.flow_state.operator.start_block_effect(&self.ctxt, sets);
}
for bb in self.mir.all_basic_blocks() {
let &repr::BasicBlockData { ref statements,
ref terminator,
is_cleanup: _ } =
self.mir.basic_block_data(bb);
let sets = &mut self.flow_state.sets.for_block(bb.index());
for j_stmt in 0..statements.len() {
self.flow_state.operator.statement_effect(&self.ctxt, sets, bb, j_stmt);
}
if terminator.is_some() {
let stmts_len = statements.len();
self.flow_state.operator.terminator_effect(&self.ctxt, sets, bb, stmts_len);
}
}
}
}
impl<'b, 'a: 'b, 'tcx: 'a, BD> PropagationContext<'b, 'a, 'tcx, BD>
where BD: BitDenotation + DataflowOperator, BD::Ctxt: HasMoveData<'tcx>
Revised mir-dataflow. Incorporates many fixes contributed by arielb1. ---- revise borrowck::mir::dataflow code to allow varying domain for bitvectors. This particular code implements the `BitDenotation` trait for three analyses: * `MovingOutStatements`, which, like `borrowck::move_data`, maps each bit-index to a move instruction, and a 1 means "the effect of this move reaches this point" (and the assigned l-value, if a scoped declaration, is still in scope). * `MaybeInitializedLvals`, which maps each bit-index to an l-value. A 1 means "there exists a control flow path to this point that initializes the associated l-value." * `MaybeUninitializedLvals`, which maps each bit-index to an l-value A 1 means "there exists a control flow path to this point that de-initializes the associated l-value." ---- Revised `graphviz` dataflow-rendering support in `borrowck::mir`. One big difference is that this code is now parameterized over the `BitDenotation`, so that it can be used to render dataflow results independent of how the dataflow bitvectors are interpreted; see where reference to `MoveOut` is replaced by the type parameter `D`. ---- Factor out routine to query subattributes in `#[rustc_mir(..)]`. (Later commits build upon this for some unit testing and instrumentation.) ---- thread through a tcx so that I can query types of lvalues as part of analysis. ---- Revised `BitDenotation::Ctxt`, allowing variation beyond `MoveData`. The main motivation is to ease threading through a `TyCtxt`. (In hindsight it might have been better to instead attach the `TyCtxt` to each of the different dataflow implementations, but that would require e.g. switching away from having a `Default` impl, so I am leaving that experiment for another time.)
2016-05-02 15:50:27 +02:00
{
fn reset(&mut self, bits: &mut IdxSet<BD::Idx>) {
let e = if BD::bottom_value() {!0} else {0};
for b in bits.words_mut() {
Revised mir-dataflow. Incorporates many fixes contributed by arielb1. ---- revise borrowck::mir::dataflow code to allow varying domain for bitvectors. This particular code implements the `BitDenotation` trait for three analyses: * `MovingOutStatements`, which, like `borrowck::move_data`, maps each bit-index to a move instruction, and a 1 means "the effect of this move reaches this point" (and the assigned l-value, if a scoped declaration, is still in scope). * `MaybeInitializedLvals`, which maps each bit-index to an l-value. A 1 means "there exists a control flow path to this point that initializes the associated l-value." * `MaybeUninitializedLvals`, which maps each bit-index to an l-value A 1 means "there exists a control flow path to this point that de-initializes the associated l-value." ---- Revised `graphviz` dataflow-rendering support in `borrowck::mir`. One big difference is that this code is now parameterized over the `BitDenotation`, so that it can be used to render dataflow results independent of how the dataflow bitvectors are interpreted; see where reference to `MoveOut` is replaced by the type parameter `D`. ---- Factor out routine to query subattributes in `#[rustc_mir(..)]`. (Later commits build upon this for some unit testing and instrumentation.) ---- thread through a tcx so that I can query types of lvalues as part of analysis. ---- Revised `BitDenotation::Ctxt`, allowing variation beyond `MoveData`. The main motivation is to ease threading through a `TyCtxt`. (In hindsight it might have been better to instead attach the `TyCtxt` to each of the different dataflow implementations, but that would require e.g. switching away from having a `Default` impl, so I am leaving that experiment for another time.)
2016-05-02 15:50:27 +02:00
*b = e;
}
}
fn walk_cfg(&mut self, in_out: &mut IdxSet<BD::Idx>) {
Revised mir-dataflow. Incorporates many fixes contributed by arielb1. ---- revise borrowck::mir::dataflow code to allow varying domain for bitvectors. This particular code implements the `BitDenotation` trait for three analyses: * `MovingOutStatements`, which, like `borrowck::move_data`, maps each bit-index to a move instruction, and a 1 means "the effect of this move reaches this point" (and the assigned l-value, if a scoped declaration, is still in scope). * `MaybeInitializedLvals`, which maps each bit-index to an l-value. A 1 means "there exists a control flow path to this point that initializes the associated l-value." * `MaybeUninitializedLvals`, which maps each bit-index to an l-value A 1 means "there exists a control flow path to this point that de-initializes the associated l-value." ---- Revised `graphviz` dataflow-rendering support in `borrowck::mir`. One big difference is that this code is now parameterized over the `BitDenotation`, so that it can be used to render dataflow results independent of how the dataflow bitvectors are interpreted; see where reference to `MoveOut` is replaced by the type parameter `D`. ---- Factor out routine to query subattributes in `#[rustc_mir(..)]`. (Later commits build upon this for some unit testing and instrumentation.) ---- thread through a tcx so that I can query types of lvalues as part of analysis. ---- Revised `BitDenotation::Ctxt`, allowing variation beyond `MoveData`. The main motivation is to ease threading through a `TyCtxt`. (In hindsight it might have been better to instead attach the `TyCtxt` to each of the different dataflow implementations, but that would require e.g. switching away from having a `Default` impl, so I am leaving that experiment for another time.)
2016-05-02 15:50:27 +02:00
let mir = self.builder.mir;
for (bb_idx, bb_data) in mir.basic_blocks.iter().enumerate() {
let builder = &mut self.builder;
{
let sets = builder.flow_state.sets.for_block(bb_idx);
debug_assert!(in_out.words().len() == sets.on_entry.words().len());
in_out.clone_from(sets.on_entry);
in_out.union(sets.gen_set);
in_out.subtract(sets.kill_set);
Revised mir-dataflow. Incorporates many fixes contributed by arielb1. ---- revise borrowck::mir::dataflow code to allow varying domain for bitvectors. This particular code implements the `BitDenotation` trait for three analyses: * `MovingOutStatements`, which, like `borrowck::move_data`, maps each bit-index to a move instruction, and a 1 means "the effect of this move reaches this point" (and the assigned l-value, if a scoped declaration, is still in scope). * `MaybeInitializedLvals`, which maps each bit-index to an l-value. A 1 means "there exists a control flow path to this point that initializes the associated l-value." * `MaybeUninitializedLvals`, which maps each bit-index to an l-value A 1 means "there exists a control flow path to this point that de-initializes the associated l-value." ---- Revised `graphviz` dataflow-rendering support in `borrowck::mir`. One big difference is that this code is now parameterized over the `BitDenotation`, so that it can be used to render dataflow results independent of how the dataflow bitvectors are interpreted; see where reference to `MoveOut` is replaced by the type parameter `D`. ---- Factor out routine to query subattributes in `#[rustc_mir(..)]`. (Later commits build upon this for some unit testing and instrumentation.) ---- thread through a tcx so that I can query types of lvalues as part of analysis. ---- Revised `BitDenotation::Ctxt`, allowing variation beyond `MoveData`. The main motivation is to ease threading through a `TyCtxt`. (In hindsight it might have been better to instead attach the `TyCtxt` to each of the different dataflow implementations, but that would require e.g. switching away from having a `Default` impl, so I am leaving that experiment for another time.)
2016-05-02 15:50:27 +02:00
}
builder.propagate_bits_into_graph_successors_of(in_out,
&mut self.changed,
(repr::BasicBlock::new(bb_idx), bb_data));
}
}
}
fn dataflow_path(context: &str, prepost: &str, path: &str) -> PathBuf {
format!("{}_{}", context, prepost);
let mut path = PathBuf::from(path);
let new_file_name = {
let orig_file_name = path.file_name().unwrap().to_str().unwrap();
format!("{}_{}", context, orig_file_name)
};
path.set_file_name(new_file_name);
path
}
Revised mir-dataflow. Incorporates many fixes contributed by arielb1. ---- revise borrowck::mir::dataflow code to allow varying domain for bitvectors. This particular code implements the `BitDenotation` trait for three analyses: * `MovingOutStatements`, which, like `borrowck::move_data`, maps each bit-index to a move instruction, and a 1 means "the effect of this move reaches this point" (and the assigned l-value, if a scoped declaration, is still in scope). * `MaybeInitializedLvals`, which maps each bit-index to an l-value. A 1 means "there exists a control flow path to this point that initializes the associated l-value." * `MaybeUninitializedLvals`, which maps each bit-index to an l-value A 1 means "there exists a control flow path to this point that de-initializes the associated l-value." ---- Revised `graphviz` dataflow-rendering support in `borrowck::mir`. One big difference is that this code is now parameterized over the `BitDenotation`, so that it can be used to render dataflow results independent of how the dataflow bitvectors are interpreted; see where reference to `MoveOut` is replaced by the type parameter `D`. ---- Factor out routine to query subattributes in `#[rustc_mir(..)]`. (Later commits build upon this for some unit testing and instrumentation.) ---- thread through a tcx so that I can query types of lvalues as part of analysis. ---- Revised `BitDenotation::Ctxt`, allowing variation beyond `MoveData`. The main motivation is to ease threading through a `TyCtxt`. (In hindsight it might have been better to instead attach the `TyCtxt` to each of the different dataflow implementations, but that would require e.g. switching away from having a `Default` impl, so I am leaving that experiment for another time.)
2016-05-02 15:50:27 +02:00
impl<'a, 'tcx: 'a, BD> MirBorrowckCtxtPreDataflow<'a, 'tcx, BD>
where BD: BitDenotation<Ctxt=MoveData<'tcx>>, BD::Bit: Debug
Revised mir-dataflow. Incorporates many fixes contributed by arielb1. ---- revise borrowck::mir::dataflow code to allow varying domain for bitvectors. This particular code implements the `BitDenotation` trait for three analyses: * `MovingOutStatements`, which, like `borrowck::move_data`, maps each bit-index to a move instruction, and a 1 means "the effect of this move reaches this point" (and the assigned l-value, if a scoped declaration, is still in scope). * `MaybeInitializedLvals`, which maps each bit-index to an l-value. A 1 means "there exists a control flow path to this point that initializes the associated l-value." * `MaybeUninitializedLvals`, which maps each bit-index to an l-value A 1 means "there exists a control flow path to this point that de-initializes the associated l-value." ---- Revised `graphviz` dataflow-rendering support in `borrowck::mir`. One big difference is that this code is now parameterized over the `BitDenotation`, so that it can be used to render dataflow results independent of how the dataflow bitvectors are interpreted; see where reference to `MoveOut` is replaced by the type parameter `D`. ---- Factor out routine to query subattributes in `#[rustc_mir(..)]`. (Later commits build upon this for some unit testing and instrumentation.) ---- thread through a tcx so that I can query types of lvalues as part of analysis. ---- Revised `BitDenotation::Ctxt`, allowing variation beyond `MoveData`. The main motivation is to ease threading through a `TyCtxt`. (In hindsight it might have been better to instead attach the `TyCtxt` to each of the different dataflow implementations, but that would require e.g. switching away from having a `Default` impl, so I am leaving that experiment for another time.)
2016-05-02 15:50:27 +02:00
{
fn pre_dataflow_instrumentation<P>(&self, p: P) -> io::Result<()>
where P: Fn(&BD::Ctxt, BD::Idx) -> &Debug
{
Revised mir-dataflow. Incorporates many fixes contributed by arielb1. ---- revise borrowck::mir::dataflow code to allow varying domain for bitvectors. This particular code implements the `BitDenotation` trait for three analyses: * `MovingOutStatements`, which, like `borrowck::move_data`, maps each bit-index to a move instruction, and a 1 means "the effect of this move reaches this point" (and the assigned l-value, if a scoped declaration, is still in scope). * `MaybeInitializedLvals`, which maps each bit-index to an l-value. A 1 means "there exists a control flow path to this point that initializes the associated l-value." * `MaybeUninitializedLvals`, which maps each bit-index to an l-value A 1 means "there exists a control flow path to this point that de-initializes the associated l-value." ---- Revised `graphviz` dataflow-rendering support in `borrowck::mir`. One big difference is that this code is now parameterized over the `BitDenotation`, so that it can be used to render dataflow results independent of how the dataflow bitvectors are interpreted; see where reference to `MoveOut` is replaced by the type parameter `D`. ---- Factor out routine to query subattributes in `#[rustc_mir(..)]`. (Later commits build upon this for some unit testing and instrumentation.) ---- thread through a tcx so that I can query types of lvalues as part of analysis. ---- Revised `BitDenotation::Ctxt`, allowing variation beyond `MoveData`. The main motivation is to ease threading through a `TyCtxt`. (In hindsight it might have been better to instead attach the `TyCtxt` to each of the different dataflow implementations, but that would require e.g. switching away from having a `Default` impl, so I am leaving that experiment for another time.)
2016-05-02 15:50:27 +02:00
if let Some(ref path_str) = self.print_preflow_to {
let path = dataflow_path(BD::name(), "preflow", path_str);
graphviz::print_borrowck_graph_to(self, &path, p)
Revised mir-dataflow. Incorporates many fixes contributed by arielb1. ---- revise borrowck::mir::dataflow code to allow varying domain for bitvectors. This particular code implements the `BitDenotation` trait for three analyses: * `MovingOutStatements`, which, like `borrowck::move_data`, maps each bit-index to a move instruction, and a 1 means "the effect of this move reaches this point" (and the assigned l-value, if a scoped declaration, is still in scope). * `MaybeInitializedLvals`, which maps each bit-index to an l-value. A 1 means "there exists a control flow path to this point that initializes the associated l-value." * `MaybeUninitializedLvals`, which maps each bit-index to an l-value A 1 means "there exists a control flow path to this point that de-initializes the associated l-value." ---- Revised `graphviz` dataflow-rendering support in `borrowck::mir`. One big difference is that this code is now parameterized over the `BitDenotation`, so that it can be used to render dataflow results independent of how the dataflow bitvectors are interpreted; see where reference to `MoveOut` is replaced by the type parameter `D`. ---- Factor out routine to query subattributes in `#[rustc_mir(..)]`. (Later commits build upon this for some unit testing and instrumentation.) ---- thread through a tcx so that I can query types of lvalues as part of analysis. ---- Revised `BitDenotation::Ctxt`, allowing variation beyond `MoveData`. The main motivation is to ease threading through a `TyCtxt`. (In hindsight it might have been better to instead attach the `TyCtxt` to each of the different dataflow implementations, but that would require e.g. switching away from having a `Default` impl, so I am leaving that experiment for another time.)
2016-05-02 15:50:27 +02:00
} else {
Ok(())
}
}
fn post_dataflow_instrumentation<P>(&self, p: P) -> io::Result<()>
where P: Fn(&BD::Ctxt, BD::Idx) -> &Debug
{
Revised mir-dataflow. Incorporates many fixes contributed by arielb1. ---- revise borrowck::mir::dataflow code to allow varying domain for bitvectors. This particular code implements the `BitDenotation` trait for three analyses: * `MovingOutStatements`, which, like `borrowck::move_data`, maps each bit-index to a move instruction, and a 1 means "the effect of this move reaches this point" (and the assigned l-value, if a scoped declaration, is still in scope). * `MaybeInitializedLvals`, which maps each bit-index to an l-value. A 1 means "there exists a control flow path to this point that initializes the associated l-value." * `MaybeUninitializedLvals`, which maps each bit-index to an l-value A 1 means "there exists a control flow path to this point that de-initializes the associated l-value." ---- Revised `graphviz` dataflow-rendering support in `borrowck::mir`. One big difference is that this code is now parameterized over the `BitDenotation`, so that it can be used to render dataflow results independent of how the dataflow bitvectors are interpreted; see where reference to `MoveOut` is replaced by the type parameter `D`. ---- Factor out routine to query subattributes in `#[rustc_mir(..)]`. (Later commits build upon this for some unit testing and instrumentation.) ---- thread through a tcx so that I can query types of lvalues as part of analysis. ---- Revised `BitDenotation::Ctxt`, allowing variation beyond `MoveData`. The main motivation is to ease threading through a `TyCtxt`. (In hindsight it might have been better to instead attach the `TyCtxt` to each of the different dataflow implementations, but that would require e.g. switching away from having a `Default` impl, so I am leaving that experiment for another time.)
2016-05-02 15:50:27 +02:00
if let Some(ref path_str) = self.print_postflow_to {
let path = dataflow_path(BD::name(), "postflow", path_str);
graphviz::print_borrowck_graph_to(self, &path, p)
Revised mir-dataflow. Incorporates many fixes contributed by arielb1. ---- revise borrowck::mir::dataflow code to allow varying domain for bitvectors. This particular code implements the `BitDenotation` trait for three analyses: * `MovingOutStatements`, which, like `borrowck::move_data`, maps each bit-index to a move instruction, and a 1 means "the effect of this move reaches this point" (and the assigned l-value, if a scoped declaration, is still in scope). * `MaybeInitializedLvals`, which maps each bit-index to an l-value. A 1 means "there exists a control flow path to this point that initializes the associated l-value." * `MaybeUninitializedLvals`, which maps each bit-index to an l-value A 1 means "there exists a control flow path to this point that de-initializes the associated l-value." ---- Revised `graphviz` dataflow-rendering support in `borrowck::mir`. One big difference is that this code is now parameterized over the `BitDenotation`, so that it can be used to render dataflow results independent of how the dataflow bitvectors are interpreted; see where reference to `MoveOut` is replaced by the type parameter `D`. ---- Factor out routine to query subattributes in `#[rustc_mir(..)]`. (Later commits build upon this for some unit testing and instrumentation.) ---- thread through a tcx so that I can query types of lvalues as part of analysis. ---- Revised `BitDenotation::Ctxt`, allowing variation beyond `MoveData`. The main motivation is to ease threading through a `TyCtxt`. (In hindsight it might have been better to instead attach the `TyCtxt` to each of the different dataflow implementations, but that would require e.g. switching away from having a `Default` impl, so I am leaving that experiment for another time.)
2016-05-02 15:50:27 +02:00
} else{
Ok(())
}
}
}
/// Maps each block to a set of bits
#[derive(Debug)]
struct Bits<E:Idx> {
bits: OwnIdxSet<E>,
}
impl<E:Idx> Clone for Bits<E> {
fn clone(&self) -> Self { Bits { bits: self.bits.clone() } }
Revised mir-dataflow. Incorporates many fixes contributed by arielb1. ---- revise borrowck::mir::dataflow code to allow varying domain for bitvectors. This particular code implements the `BitDenotation` trait for three analyses: * `MovingOutStatements`, which, like `borrowck::move_data`, maps each bit-index to a move instruction, and a 1 means "the effect of this move reaches this point" (and the assigned l-value, if a scoped declaration, is still in scope). * `MaybeInitializedLvals`, which maps each bit-index to an l-value. A 1 means "there exists a control flow path to this point that initializes the associated l-value." * `MaybeUninitializedLvals`, which maps each bit-index to an l-value A 1 means "there exists a control flow path to this point that de-initializes the associated l-value." ---- Revised `graphviz` dataflow-rendering support in `borrowck::mir`. One big difference is that this code is now parameterized over the `BitDenotation`, so that it can be used to render dataflow results independent of how the dataflow bitvectors are interpreted; see where reference to `MoveOut` is replaced by the type parameter `D`. ---- Factor out routine to query subattributes in `#[rustc_mir(..)]`. (Later commits build upon this for some unit testing and instrumentation.) ---- thread through a tcx so that I can query types of lvalues as part of analysis. ---- Revised `BitDenotation::Ctxt`, allowing variation beyond `MoveData`. The main motivation is to ease threading through a `TyCtxt`. (In hindsight it might have been better to instead attach the `TyCtxt` to each of the different dataflow implementations, but that would require e.g. switching away from having a `Default` impl, so I am leaving that experiment for another time.)
2016-05-02 15:50:27 +02:00
}
impl<E:Idx> Bits<E> {
fn new(bits: OwnIdxSet<E>) -> Self {
Bits { bits: bits }
Revised mir-dataflow. Incorporates many fixes contributed by arielb1. ---- revise borrowck::mir::dataflow code to allow varying domain for bitvectors. This particular code implements the `BitDenotation` trait for three analyses: * `MovingOutStatements`, which, like `borrowck::move_data`, maps each bit-index to a move instruction, and a 1 means "the effect of this move reaches this point" (and the assigned l-value, if a scoped declaration, is still in scope). * `MaybeInitializedLvals`, which maps each bit-index to an l-value. A 1 means "there exists a control flow path to this point that initializes the associated l-value." * `MaybeUninitializedLvals`, which maps each bit-index to an l-value A 1 means "there exists a control flow path to this point that de-initializes the associated l-value." ---- Revised `graphviz` dataflow-rendering support in `borrowck::mir`. One big difference is that this code is now parameterized over the `BitDenotation`, so that it can be used to render dataflow results independent of how the dataflow bitvectors are interpreted; see where reference to `MoveOut` is replaced by the type parameter `D`. ---- Factor out routine to query subattributes in `#[rustc_mir(..)]`. (Later commits build upon this for some unit testing and instrumentation.) ---- thread through a tcx so that I can query types of lvalues as part of analysis. ---- Revised `BitDenotation::Ctxt`, allowing variation beyond `MoveData`. The main motivation is to ease threading through a `TyCtxt`. (In hindsight it might have been better to instead attach the `TyCtxt` to each of the different dataflow implementations, but that would require e.g. switching away from having a `Default` impl, so I am leaving that experiment for another time.)
2016-05-02 15:50:27 +02:00
}
}
pub trait HasMoveData<'tcx> {
fn move_data(&self) -> &MoveData<'tcx>;
}
impl<'tcx> HasMoveData<'tcx> for MoveData<'tcx> {
fn move_data(&self) -> &MoveData<'tcx> { self }
}
impl<'tcx, A, B> HasMoveData<'tcx> for (A, B, MoveData<'tcx>) {
fn move_data(&self) -> &MoveData<'tcx> { &self.2 }
}
pub struct DataflowAnalysis<'a, 'tcx: 'a, O>
where O: BitDenotation, O::Ctxt: 'a+HasMoveData<'tcx>
Revised mir-dataflow. Incorporates many fixes contributed by arielb1. ---- revise borrowck::mir::dataflow code to allow varying domain for bitvectors. This particular code implements the `BitDenotation` trait for three analyses: * `MovingOutStatements`, which, like `borrowck::move_data`, maps each bit-index to a move instruction, and a 1 means "the effect of this move reaches this point" (and the assigned l-value, if a scoped declaration, is still in scope). * `MaybeInitializedLvals`, which maps each bit-index to an l-value. A 1 means "there exists a control flow path to this point that initializes the associated l-value." * `MaybeUninitializedLvals`, which maps each bit-index to an l-value A 1 means "there exists a control flow path to this point that de-initializes the associated l-value." ---- Revised `graphviz` dataflow-rendering support in `borrowck::mir`. One big difference is that this code is now parameterized over the `BitDenotation`, so that it can be used to render dataflow results independent of how the dataflow bitvectors are interpreted; see where reference to `MoveOut` is replaced by the type parameter `D`. ---- Factor out routine to query subattributes in `#[rustc_mir(..)]`. (Later commits build upon this for some unit testing and instrumentation.) ---- thread through a tcx so that I can query types of lvalues as part of analysis. ---- Revised `BitDenotation::Ctxt`, allowing variation beyond `MoveData`. The main motivation is to ease threading through a `TyCtxt`. (In hindsight it might have been better to instead attach the `TyCtxt` to each of the different dataflow implementations, but that would require e.g. switching away from having a `Default` impl, so I am leaving that experiment for another time.)
2016-05-02 15:50:27 +02:00
{
flow_state: DataflowState<O>,
mir: &'a Mir<'tcx>,
ctxt: &'a O::Ctxt,
Revised mir-dataflow. Incorporates many fixes contributed by arielb1. ---- revise borrowck::mir::dataflow code to allow varying domain for bitvectors. This particular code implements the `BitDenotation` trait for three analyses: * `MovingOutStatements`, which, like `borrowck::move_data`, maps each bit-index to a move instruction, and a 1 means "the effect of this move reaches this point" (and the assigned l-value, if a scoped declaration, is still in scope). * `MaybeInitializedLvals`, which maps each bit-index to an l-value. A 1 means "there exists a control flow path to this point that initializes the associated l-value." * `MaybeUninitializedLvals`, which maps each bit-index to an l-value A 1 means "there exists a control flow path to this point that de-initializes the associated l-value." ---- Revised `graphviz` dataflow-rendering support in `borrowck::mir`. One big difference is that this code is now parameterized over the `BitDenotation`, so that it can be used to render dataflow results independent of how the dataflow bitvectors are interpreted; see where reference to `MoveOut` is replaced by the type parameter `D`. ---- Factor out routine to query subattributes in `#[rustc_mir(..)]`. (Later commits build upon this for some unit testing and instrumentation.) ---- thread through a tcx so that I can query types of lvalues as part of analysis. ---- Revised `BitDenotation::Ctxt`, allowing variation beyond `MoveData`. The main motivation is to ease threading through a `TyCtxt`. (In hindsight it might have been better to instead attach the `TyCtxt` to each of the different dataflow implementations, but that would require e.g. switching away from having a `Default` impl, so I am leaving that experiment for another time.)
2016-05-02 15:50:27 +02:00
}
impl<'a, 'tcx: 'a, O> DataflowAnalysis<'a, 'tcx, O>
where O: BitDenotation, O::Ctxt: HasMoveData<'tcx>
{
pub fn results(self) -> DataflowResults<O> {
DataflowResults(self.flow_state)
Revised mir-dataflow. Incorporates many fixes contributed by arielb1. ---- revise borrowck::mir::dataflow code to allow varying domain for bitvectors. This particular code implements the `BitDenotation` trait for three analyses: * `MovingOutStatements`, which, like `borrowck::move_data`, maps each bit-index to a move instruction, and a 1 means "the effect of this move reaches this point" (and the assigned l-value, if a scoped declaration, is still in scope). * `MaybeInitializedLvals`, which maps each bit-index to an l-value. A 1 means "there exists a control flow path to this point that initializes the associated l-value." * `MaybeUninitializedLvals`, which maps each bit-index to an l-value A 1 means "there exists a control flow path to this point that de-initializes the associated l-value." ---- Revised `graphviz` dataflow-rendering support in `borrowck::mir`. One big difference is that this code is now parameterized over the `BitDenotation`, so that it can be used to render dataflow results independent of how the dataflow bitvectors are interpreted; see where reference to `MoveOut` is replaced by the type parameter `D`. ---- Factor out routine to query subattributes in `#[rustc_mir(..)]`. (Later commits build upon this for some unit testing and instrumentation.) ---- thread through a tcx so that I can query types of lvalues as part of analysis. ---- Revised `BitDenotation::Ctxt`, allowing variation beyond `MoveData`. The main motivation is to ease threading through a `TyCtxt`. (In hindsight it might have been better to instead attach the `TyCtxt` to each of the different dataflow implementations, but that would require e.g. switching away from having a `Default` impl, so I am leaving that experiment for another time.)
2016-05-02 15:50:27 +02:00
}
pub fn mir(&self) -> &'a Mir<'tcx> { self.mir }
}
pub struct DataflowResults<O>(DataflowState<O>) where O: BitDenotation;
Revised mir-dataflow. Incorporates many fixes contributed by arielb1. ---- revise borrowck::mir::dataflow code to allow varying domain for bitvectors. This particular code implements the `BitDenotation` trait for three analyses: * `MovingOutStatements`, which, like `borrowck::move_data`, maps each bit-index to a move instruction, and a 1 means "the effect of this move reaches this point" (and the assigned l-value, if a scoped declaration, is still in scope). * `MaybeInitializedLvals`, which maps each bit-index to an l-value. A 1 means "there exists a control flow path to this point that initializes the associated l-value." * `MaybeUninitializedLvals`, which maps each bit-index to an l-value A 1 means "there exists a control flow path to this point that de-initializes the associated l-value." ---- Revised `graphviz` dataflow-rendering support in `borrowck::mir`. One big difference is that this code is now parameterized over the `BitDenotation`, so that it can be used to render dataflow results independent of how the dataflow bitvectors are interpreted; see where reference to `MoveOut` is replaced by the type parameter `D`. ---- Factor out routine to query subattributes in `#[rustc_mir(..)]`. (Later commits build upon this for some unit testing and instrumentation.) ---- thread through a tcx so that I can query types of lvalues as part of analysis. ---- Revised `BitDenotation::Ctxt`, allowing variation beyond `MoveData`. The main motivation is to ease threading through a `TyCtxt`. (In hindsight it might have been better to instead attach the `TyCtxt` to each of the different dataflow implementations, but that would require e.g. switching away from having a `Default` impl, so I am leaving that experiment for another time.)
2016-05-02 15:50:27 +02:00
// FIXME: This type shouldn't be public, but the graphviz::MirWithFlowState trait
// references it in a method signature. Look into using `pub(crate)` to address this.
pub struct DataflowState<O: BitDenotation>
{
/// All the sets for the analysis. (Factored into its
/// own structure so that we can borrow it mutably
/// on its own separate from other fields.)
pub sets: AllSets<O::Idx>,
Revised mir-dataflow. Incorporates many fixes contributed by arielb1. ---- revise borrowck::mir::dataflow code to allow varying domain for bitvectors. This particular code implements the `BitDenotation` trait for three analyses: * `MovingOutStatements`, which, like `borrowck::move_data`, maps each bit-index to a move instruction, and a 1 means "the effect of this move reaches this point" (and the assigned l-value, if a scoped declaration, is still in scope). * `MaybeInitializedLvals`, which maps each bit-index to an l-value. A 1 means "there exists a control flow path to this point that initializes the associated l-value." * `MaybeUninitializedLvals`, which maps each bit-index to an l-value A 1 means "there exists a control flow path to this point that de-initializes the associated l-value." ---- Revised `graphviz` dataflow-rendering support in `borrowck::mir`. One big difference is that this code is now parameterized over the `BitDenotation`, so that it can be used to render dataflow results independent of how the dataflow bitvectors are interpreted; see where reference to `MoveOut` is replaced by the type parameter `D`. ---- Factor out routine to query subattributes in `#[rustc_mir(..)]`. (Later commits build upon this for some unit testing and instrumentation.) ---- thread through a tcx so that I can query types of lvalues as part of analysis. ---- Revised `BitDenotation::Ctxt`, allowing variation beyond `MoveData`. The main motivation is to ease threading through a `TyCtxt`. (In hindsight it might have been better to instead attach the `TyCtxt` to each of the different dataflow implementations, but that would require e.g. switching away from having a `Default` impl, so I am leaving that experiment for another time.)
2016-05-02 15:50:27 +02:00
/// operator used to initialize, combine, and interpret bits.
operator: O,
}
#[derive(Debug)]
pub struct AllSets<E: Idx> {
Revised mir-dataflow. Incorporates many fixes contributed by arielb1. ---- revise borrowck::mir::dataflow code to allow varying domain for bitvectors. This particular code implements the `BitDenotation` trait for three analyses: * `MovingOutStatements`, which, like `borrowck::move_data`, maps each bit-index to a move instruction, and a 1 means "the effect of this move reaches this point" (and the assigned l-value, if a scoped declaration, is still in scope). * `MaybeInitializedLvals`, which maps each bit-index to an l-value. A 1 means "there exists a control flow path to this point that initializes the associated l-value." * `MaybeUninitializedLvals`, which maps each bit-index to an l-value A 1 means "there exists a control flow path to this point that de-initializes the associated l-value." ---- Revised `graphviz` dataflow-rendering support in `borrowck::mir`. One big difference is that this code is now parameterized over the `BitDenotation`, so that it can be used to render dataflow results independent of how the dataflow bitvectors are interpreted; see where reference to `MoveOut` is replaced by the type parameter `D`. ---- Factor out routine to query subattributes in `#[rustc_mir(..)]`. (Later commits build upon this for some unit testing and instrumentation.) ---- thread through a tcx so that I can query types of lvalues as part of analysis. ---- Revised `BitDenotation::Ctxt`, allowing variation beyond `MoveData`. The main motivation is to ease threading through a `TyCtxt`. (In hindsight it might have been better to instead attach the `TyCtxt` to each of the different dataflow implementations, but that would require e.g. switching away from having a `Default` impl, so I am leaving that experiment for another time.)
2016-05-02 15:50:27 +02:00
/// Analysis bitwidth for each block.
bits_per_block: usize,
/// Number of words associated with each block entry
/// equal to bits_per_block / usize::BITS, rounded up.
words_per_block: usize,
/// For each block, bits generated by executing the statements in
/// the block. (For comparison, the Terminator for each block is
/// handled in a flow-specific manner during propagation.)
gen_sets: Bits<E>,
Revised mir-dataflow. Incorporates many fixes contributed by arielb1. ---- revise borrowck::mir::dataflow code to allow varying domain for bitvectors. This particular code implements the `BitDenotation` trait for three analyses: * `MovingOutStatements`, which, like `borrowck::move_data`, maps each bit-index to a move instruction, and a 1 means "the effect of this move reaches this point" (and the assigned l-value, if a scoped declaration, is still in scope). * `MaybeInitializedLvals`, which maps each bit-index to an l-value. A 1 means "there exists a control flow path to this point that initializes the associated l-value." * `MaybeUninitializedLvals`, which maps each bit-index to an l-value A 1 means "there exists a control flow path to this point that de-initializes the associated l-value." ---- Revised `graphviz` dataflow-rendering support in `borrowck::mir`. One big difference is that this code is now parameterized over the `BitDenotation`, so that it can be used to render dataflow results independent of how the dataflow bitvectors are interpreted; see where reference to `MoveOut` is replaced by the type parameter `D`. ---- Factor out routine to query subattributes in `#[rustc_mir(..)]`. (Later commits build upon this for some unit testing and instrumentation.) ---- thread through a tcx so that I can query types of lvalues as part of analysis. ---- Revised `BitDenotation::Ctxt`, allowing variation beyond `MoveData`. The main motivation is to ease threading through a `TyCtxt`. (In hindsight it might have been better to instead attach the `TyCtxt` to each of the different dataflow implementations, but that would require e.g. switching away from having a `Default` impl, so I am leaving that experiment for another time.)
2016-05-02 15:50:27 +02:00
/// For each block, bits killed by executing the statements in the
/// block. (For comparison, the Terminator for each block is
/// handled in a flow-specific manner during propagation.)
kill_sets: Bits<E>,
Revised mir-dataflow. Incorporates many fixes contributed by arielb1. ---- revise borrowck::mir::dataflow code to allow varying domain for bitvectors. This particular code implements the `BitDenotation` trait for three analyses: * `MovingOutStatements`, which, like `borrowck::move_data`, maps each bit-index to a move instruction, and a 1 means "the effect of this move reaches this point" (and the assigned l-value, if a scoped declaration, is still in scope). * `MaybeInitializedLvals`, which maps each bit-index to an l-value. A 1 means "there exists a control flow path to this point that initializes the associated l-value." * `MaybeUninitializedLvals`, which maps each bit-index to an l-value A 1 means "there exists a control flow path to this point that de-initializes the associated l-value." ---- Revised `graphviz` dataflow-rendering support in `borrowck::mir`. One big difference is that this code is now parameterized over the `BitDenotation`, so that it can be used to render dataflow results independent of how the dataflow bitvectors are interpreted; see where reference to `MoveOut` is replaced by the type parameter `D`. ---- Factor out routine to query subattributes in `#[rustc_mir(..)]`. (Later commits build upon this for some unit testing and instrumentation.) ---- thread through a tcx so that I can query types of lvalues as part of analysis. ---- Revised `BitDenotation::Ctxt`, allowing variation beyond `MoveData`. The main motivation is to ease threading through a `TyCtxt`. (In hindsight it might have been better to instead attach the `TyCtxt` to each of the different dataflow implementations, but that would require e.g. switching away from having a `Default` impl, so I am leaving that experiment for another time.)
2016-05-02 15:50:27 +02:00
/// For each block, bits valid on entry to the block.
on_entry_sets: Bits<E>,
Revised mir-dataflow. Incorporates many fixes contributed by arielb1. ---- revise borrowck::mir::dataflow code to allow varying domain for bitvectors. This particular code implements the `BitDenotation` trait for three analyses: * `MovingOutStatements`, which, like `borrowck::move_data`, maps each bit-index to a move instruction, and a 1 means "the effect of this move reaches this point" (and the assigned l-value, if a scoped declaration, is still in scope). * `MaybeInitializedLvals`, which maps each bit-index to an l-value. A 1 means "there exists a control flow path to this point that initializes the associated l-value." * `MaybeUninitializedLvals`, which maps each bit-index to an l-value A 1 means "there exists a control flow path to this point that de-initializes the associated l-value." ---- Revised `graphviz` dataflow-rendering support in `borrowck::mir`. One big difference is that this code is now parameterized over the `BitDenotation`, so that it can be used to render dataflow results independent of how the dataflow bitvectors are interpreted; see where reference to `MoveOut` is replaced by the type parameter `D`. ---- Factor out routine to query subattributes in `#[rustc_mir(..)]`. (Later commits build upon this for some unit testing and instrumentation.) ---- thread through a tcx so that I can query types of lvalues as part of analysis. ---- Revised `BitDenotation::Ctxt`, allowing variation beyond `MoveData`. The main motivation is to ease threading through a `TyCtxt`. (In hindsight it might have been better to instead attach the `TyCtxt` to each of the different dataflow implementations, but that would require e.g. switching away from having a `Default` impl, so I am leaving that experiment for another time.)
2016-05-02 15:50:27 +02:00
}
pub struct BlockSets<'a, E: Idx> {
on_entry: &'a mut IdxSet<E>,
gen_set: &'a mut IdxSet<E>,
kill_set: &'a mut IdxSet<E>,
Revised mir-dataflow. Incorporates many fixes contributed by arielb1. ---- revise borrowck::mir::dataflow code to allow varying domain for bitvectors. This particular code implements the `BitDenotation` trait for three analyses: * `MovingOutStatements`, which, like `borrowck::move_data`, maps each bit-index to a move instruction, and a 1 means "the effect of this move reaches this point" (and the assigned l-value, if a scoped declaration, is still in scope). * `MaybeInitializedLvals`, which maps each bit-index to an l-value. A 1 means "there exists a control flow path to this point that initializes the associated l-value." * `MaybeUninitializedLvals`, which maps each bit-index to an l-value A 1 means "there exists a control flow path to this point that de-initializes the associated l-value." ---- Revised `graphviz` dataflow-rendering support in `borrowck::mir`. One big difference is that this code is now parameterized over the `BitDenotation`, so that it can be used to render dataflow results independent of how the dataflow bitvectors are interpreted; see where reference to `MoveOut` is replaced by the type parameter `D`. ---- Factor out routine to query subattributes in `#[rustc_mir(..)]`. (Later commits build upon this for some unit testing and instrumentation.) ---- thread through a tcx so that I can query types of lvalues as part of analysis. ---- Revised `BitDenotation::Ctxt`, allowing variation beyond `MoveData`. The main motivation is to ease threading through a `TyCtxt`. (In hindsight it might have been better to instead attach the `TyCtxt` to each of the different dataflow implementations, but that would require e.g. switching away from having a `Default` impl, so I am leaving that experiment for another time.)
2016-05-02 15:50:27 +02:00
}
impl<'a, E:Idx> BlockSets<'a, E> {
fn gen(&mut self, e: &E) {
self.gen_set.add(e);
self.kill_set.remove(e);
}
fn kill(&mut self, e: &E) {
self.gen_set.remove(e);
self.kill_set.add(e);
}
}
impl<E:Idx> AllSets<E> {
Revised mir-dataflow. Incorporates many fixes contributed by arielb1. ---- revise borrowck::mir::dataflow code to allow varying domain for bitvectors. This particular code implements the `BitDenotation` trait for three analyses: * `MovingOutStatements`, which, like `borrowck::move_data`, maps each bit-index to a move instruction, and a 1 means "the effect of this move reaches this point" (and the assigned l-value, if a scoped declaration, is still in scope). * `MaybeInitializedLvals`, which maps each bit-index to an l-value. A 1 means "there exists a control flow path to this point that initializes the associated l-value." * `MaybeUninitializedLvals`, which maps each bit-index to an l-value A 1 means "there exists a control flow path to this point that de-initializes the associated l-value." ---- Revised `graphviz` dataflow-rendering support in `borrowck::mir`. One big difference is that this code is now parameterized over the `BitDenotation`, so that it can be used to render dataflow results independent of how the dataflow bitvectors are interpreted; see where reference to `MoveOut` is replaced by the type parameter `D`. ---- Factor out routine to query subattributes in `#[rustc_mir(..)]`. (Later commits build upon this for some unit testing and instrumentation.) ---- thread through a tcx so that I can query types of lvalues as part of analysis. ---- Revised `BitDenotation::Ctxt`, allowing variation beyond `MoveData`. The main motivation is to ease threading through a `TyCtxt`. (In hindsight it might have been better to instead attach the `TyCtxt` to each of the different dataflow implementations, but that would require e.g. switching away from having a `Default` impl, so I am leaving that experiment for another time.)
2016-05-02 15:50:27 +02:00
pub fn bits_per_block(&self) -> usize { self.bits_per_block }
pub fn for_block(&mut self, block_idx: usize) -> BlockSets<E> {
Revised mir-dataflow. Incorporates many fixes contributed by arielb1. ---- revise borrowck::mir::dataflow code to allow varying domain for bitvectors. This particular code implements the `BitDenotation` trait for three analyses: * `MovingOutStatements`, which, like `borrowck::move_data`, maps each bit-index to a move instruction, and a 1 means "the effect of this move reaches this point" (and the assigned l-value, if a scoped declaration, is still in scope). * `MaybeInitializedLvals`, which maps each bit-index to an l-value. A 1 means "there exists a control flow path to this point that initializes the associated l-value." * `MaybeUninitializedLvals`, which maps each bit-index to an l-value A 1 means "there exists a control flow path to this point that de-initializes the associated l-value." ---- Revised `graphviz` dataflow-rendering support in `borrowck::mir`. One big difference is that this code is now parameterized over the `BitDenotation`, so that it can be used to render dataflow results independent of how the dataflow bitvectors are interpreted; see where reference to `MoveOut` is replaced by the type parameter `D`. ---- Factor out routine to query subattributes in `#[rustc_mir(..)]`. (Later commits build upon this for some unit testing and instrumentation.) ---- thread through a tcx so that I can query types of lvalues as part of analysis. ---- Revised `BitDenotation::Ctxt`, allowing variation beyond `MoveData`. The main motivation is to ease threading through a `TyCtxt`. (In hindsight it might have been better to instead attach the `TyCtxt` to each of the different dataflow implementations, but that would require e.g. switching away from having a `Default` impl, so I am leaving that experiment for another time.)
2016-05-02 15:50:27 +02:00
let offset = self.words_per_block * block_idx;
let range = E::new(offset)..E::new(offset + self.words_per_block);
Revised mir-dataflow. Incorporates many fixes contributed by arielb1. ---- revise borrowck::mir::dataflow code to allow varying domain for bitvectors. This particular code implements the `BitDenotation` trait for three analyses: * `MovingOutStatements`, which, like `borrowck::move_data`, maps each bit-index to a move instruction, and a 1 means "the effect of this move reaches this point" (and the assigned l-value, if a scoped declaration, is still in scope). * `MaybeInitializedLvals`, which maps each bit-index to an l-value. A 1 means "there exists a control flow path to this point that initializes the associated l-value." * `MaybeUninitializedLvals`, which maps each bit-index to an l-value A 1 means "there exists a control flow path to this point that de-initializes the associated l-value." ---- Revised `graphviz` dataflow-rendering support in `borrowck::mir`. One big difference is that this code is now parameterized over the `BitDenotation`, so that it can be used to render dataflow results independent of how the dataflow bitvectors are interpreted; see where reference to `MoveOut` is replaced by the type parameter `D`. ---- Factor out routine to query subattributes in `#[rustc_mir(..)]`. (Later commits build upon this for some unit testing and instrumentation.) ---- thread through a tcx so that I can query types of lvalues as part of analysis. ---- Revised `BitDenotation::Ctxt`, allowing variation beyond `MoveData`. The main motivation is to ease threading through a `TyCtxt`. (In hindsight it might have been better to instead attach the `TyCtxt` to each of the different dataflow implementations, but that would require e.g. switching away from having a `Default` impl, so I am leaving that experiment for another time.)
2016-05-02 15:50:27 +02:00
BlockSets {
on_entry: self.on_entry_sets.bits.range_mut(&range),
gen_set: self.gen_sets.bits.range_mut(&range),
kill_set: self.kill_sets.bits.range_mut(&range),
Revised mir-dataflow. Incorporates many fixes contributed by arielb1. ---- revise borrowck::mir::dataflow code to allow varying domain for bitvectors. This particular code implements the `BitDenotation` trait for three analyses: * `MovingOutStatements`, which, like `borrowck::move_data`, maps each bit-index to a move instruction, and a 1 means "the effect of this move reaches this point" (and the assigned l-value, if a scoped declaration, is still in scope). * `MaybeInitializedLvals`, which maps each bit-index to an l-value. A 1 means "there exists a control flow path to this point that initializes the associated l-value." * `MaybeUninitializedLvals`, which maps each bit-index to an l-value A 1 means "there exists a control flow path to this point that de-initializes the associated l-value." ---- Revised `graphviz` dataflow-rendering support in `borrowck::mir`. One big difference is that this code is now parameterized over the `BitDenotation`, so that it can be used to render dataflow results independent of how the dataflow bitvectors are interpreted; see where reference to `MoveOut` is replaced by the type parameter `D`. ---- Factor out routine to query subattributes in `#[rustc_mir(..)]`. (Later commits build upon this for some unit testing and instrumentation.) ---- thread through a tcx so that I can query types of lvalues as part of analysis. ---- Revised `BitDenotation::Ctxt`, allowing variation beyond `MoveData`. The main motivation is to ease threading through a `TyCtxt`. (In hindsight it might have been better to instead attach the `TyCtxt` to each of the different dataflow implementations, but that would require e.g. switching away from having a `Default` impl, so I am leaving that experiment for another time.)
2016-05-02 15:50:27 +02:00
}
}
fn lookup_set_for<'a>(&self, sets: &'a Bits<E>, block_idx: usize) -> &'a IdxSet<E> {
Revised mir-dataflow. Incorporates many fixes contributed by arielb1. ---- revise borrowck::mir::dataflow code to allow varying domain for bitvectors. This particular code implements the `BitDenotation` trait for three analyses: * `MovingOutStatements`, which, like `borrowck::move_data`, maps each bit-index to a move instruction, and a 1 means "the effect of this move reaches this point" (and the assigned l-value, if a scoped declaration, is still in scope). * `MaybeInitializedLvals`, which maps each bit-index to an l-value. A 1 means "there exists a control flow path to this point that initializes the associated l-value." * `MaybeUninitializedLvals`, which maps each bit-index to an l-value A 1 means "there exists a control flow path to this point that de-initializes the associated l-value." ---- Revised `graphviz` dataflow-rendering support in `borrowck::mir`. One big difference is that this code is now parameterized over the `BitDenotation`, so that it can be used to render dataflow results independent of how the dataflow bitvectors are interpreted; see where reference to `MoveOut` is replaced by the type parameter `D`. ---- Factor out routine to query subattributes in `#[rustc_mir(..)]`. (Later commits build upon this for some unit testing and instrumentation.) ---- thread through a tcx so that I can query types of lvalues as part of analysis. ---- Revised `BitDenotation::Ctxt`, allowing variation beyond `MoveData`. The main motivation is to ease threading through a `TyCtxt`. (In hindsight it might have been better to instead attach the `TyCtxt` to each of the different dataflow implementations, but that would require e.g. switching away from having a `Default` impl, so I am leaving that experiment for another time.)
2016-05-02 15:50:27 +02:00
let offset = self.words_per_block * block_idx;
let range = E::new(offset)..E::new(offset + self.words_per_block);
sets.bits.range(&range)
Revised mir-dataflow. Incorporates many fixes contributed by arielb1. ---- revise borrowck::mir::dataflow code to allow varying domain for bitvectors. This particular code implements the `BitDenotation` trait for three analyses: * `MovingOutStatements`, which, like `borrowck::move_data`, maps each bit-index to a move instruction, and a 1 means "the effect of this move reaches this point" (and the assigned l-value, if a scoped declaration, is still in scope). * `MaybeInitializedLvals`, which maps each bit-index to an l-value. A 1 means "there exists a control flow path to this point that initializes the associated l-value." * `MaybeUninitializedLvals`, which maps each bit-index to an l-value A 1 means "there exists a control flow path to this point that de-initializes the associated l-value." ---- Revised `graphviz` dataflow-rendering support in `borrowck::mir`. One big difference is that this code is now parameterized over the `BitDenotation`, so that it can be used to render dataflow results independent of how the dataflow bitvectors are interpreted; see where reference to `MoveOut` is replaced by the type parameter `D`. ---- Factor out routine to query subattributes in `#[rustc_mir(..)]`. (Later commits build upon this for some unit testing and instrumentation.) ---- thread through a tcx so that I can query types of lvalues as part of analysis. ---- Revised `BitDenotation::Ctxt`, allowing variation beyond `MoveData`. The main motivation is to ease threading through a `TyCtxt`. (In hindsight it might have been better to instead attach the `TyCtxt` to each of the different dataflow implementations, but that would require e.g. switching away from having a `Default` impl, so I am leaving that experiment for another time.)
2016-05-02 15:50:27 +02:00
}
pub fn gen_set_for(&self, block_idx: usize) -> &IdxSet<E> {
Revised mir-dataflow. Incorporates many fixes contributed by arielb1. ---- revise borrowck::mir::dataflow code to allow varying domain for bitvectors. This particular code implements the `BitDenotation` trait for three analyses: * `MovingOutStatements`, which, like `borrowck::move_data`, maps each bit-index to a move instruction, and a 1 means "the effect of this move reaches this point" (and the assigned l-value, if a scoped declaration, is still in scope). * `MaybeInitializedLvals`, which maps each bit-index to an l-value. A 1 means "there exists a control flow path to this point that initializes the associated l-value." * `MaybeUninitializedLvals`, which maps each bit-index to an l-value A 1 means "there exists a control flow path to this point that de-initializes the associated l-value." ---- Revised `graphviz` dataflow-rendering support in `borrowck::mir`. One big difference is that this code is now parameterized over the `BitDenotation`, so that it can be used to render dataflow results independent of how the dataflow bitvectors are interpreted; see where reference to `MoveOut` is replaced by the type parameter `D`. ---- Factor out routine to query subattributes in `#[rustc_mir(..)]`. (Later commits build upon this for some unit testing and instrumentation.) ---- thread through a tcx so that I can query types of lvalues as part of analysis. ---- Revised `BitDenotation::Ctxt`, allowing variation beyond `MoveData`. The main motivation is to ease threading through a `TyCtxt`. (In hindsight it might have been better to instead attach the `TyCtxt` to each of the different dataflow implementations, but that would require e.g. switching away from having a `Default` impl, so I am leaving that experiment for another time.)
2016-05-02 15:50:27 +02:00
self.lookup_set_for(&self.gen_sets, block_idx)
}
pub fn kill_set_for(&self, block_idx: usize) -> &IdxSet<E> {
Revised mir-dataflow. Incorporates many fixes contributed by arielb1. ---- revise borrowck::mir::dataflow code to allow varying domain for bitvectors. This particular code implements the `BitDenotation` trait for three analyses: * `MovingOutStatements`, which, like `borrowck::move_data`, maps each bit-index to a move instruction, and a 1 means "the effect of this move reaches this point" (and the assigned l-value, if a scoped declaration, is still in scope). * `MaybeInitializedLvals`, which maps each bit-index to an l-value. A 1 means "there exists a control flow path to this point that initializes the associated l-value." * `MaybeUninitializedLvals`, which maps each bit-index to an l-value A 1 means "there exists a control flow path to this point that de-initializes the associated l-value." ---- Revised `graphviz` dataflow-rendering support in `borrowck::mir`. One big difference is that this code is now parameterized over the `BitDenotation`, so that it can be used to render dataflow results independent of how the dataflow bitvectors are interpreted; see where reference to `MoveOut` is replaced by the type parameter `D`. ---- Factor out routine to query subattributes in `#[rustc_mir(..)]`. (Later commits build upon this for some unit testing and instrumentation.) ---- thread through a tcx so that I can query types of lvalues as part of analysis. ---- Revised `BitDenotation::Ctxt`, allowing variation beyond `MoveData`. The main motivation is to ease threading through a `TyCtxt`. (In hindsight it might have been better to instead attach the `TyCtxt` to each of the different dataflow implementations, but that would require e.g. switching away from having a `Default` impl, so I am leaving that experiment for another time.)
2016-05-02 15:50:27 +02:00
self.lookup_set_for(&self.kill_sets, block_idx)
}
pub fn on_entry_set_for(&self, block_idx: usize) -> &IdxSet<E> {
Revised mir-dataflow. Incorporates many fixes contributed by arielb1. ---- revise borrowck::mir::dataflow code to allow varying domain for bitvectors. This particular code implements the `BitDenotation` trait for three analyses: * `MovingOutStatements`, which, like `borrowck::move_data`, maps each bit-index to a move instruction, and a 1 means "the effect of this move reaches this point" (and the assigned l-value, if a scoped declaration, is still in scope). * `MaybeInitializedLvals`, which maps each bit-index to an l-value. A 1 means "there exists a control flow path to this point that initializes the associated l-value." * `MaybeUninitializedLvals`, which maps each bit-index to an l-value A 1 means "there exists a control flow path to this point that de-initializes the associated l-value." ---- Revised `graphviz` dataflow-rendering support in `borrowck::mir`. One big difference is that this code is now parameterized over the `BitDenotation`, so that it can be used to render dataflow results independent of how the dataflow bitvectors are interpreted; see where reference to `MoveOut` is replaced by the type parameter `D`. ---- Factor out routine to query subattributes in `#[rustc_mir(..)]`. (Later commits build upon this for some unit testing and instrumentation.) ---- thread through a tcx so that I can query types of lvalues as part of analysis. ---- Revised `BitDenotation::Ctxt`, allowing variation beyond `MoveData`. The main motivation is to ease threading through a `TyCtxt`. (In hindsight it might have been better to instead attach the `TyCtxt` to each of the different dataflow implementations, but that would require e.g. switching away from having a `Default` impl, so I am leaving that experiment for another time.)
2016-05-02 15:50:27 +02:00
self.lookup_set_for(&self.on_entry_sets, block_idx)
}
}
/// Parameterization for the precise form of data flow that is used.
pub trait DataflowOperator: BitwiseOperator {
Revised mir-dataflow. Incorporates many fixes contributed by arielb1. ---- revise borrowck::mir::dataflow code to allow varying domain for bitvectors. This particular code implements the `BitDenotation` trait for three analyses: * `MovingOutStatements`, which, like `borrowck::move_data`, maps each bit-index to a move instruction, and a 1 means "the effect of this move reaches this point" (and the assigned l-value, if a scoped declaration, is still in scope). * `MaybeInitializedLvals`, which maps each bit-index to an l-value. A 1 means "there exists a control flow path to this point that initializes the associated l-value." * `MaybeUninitializedLvals`, which maps each bit-index to an l-value A 1 means "there exists a control flow path to this point that de-initializes the associated l-value." ---- Revised `graphviz` dataflow-rendering support in `borrowck::mir`. One big difference is that this code is now parameterized over the `BitDenotation`, so that it can be used to render dataflow results independent of how the dataflow bitvectors are interpreted; see where reference to `MoveOut` is replaced by the type parameter `D`. ---- Factor out routine to query subattributes in `#[rustc_mir(..)]`. (Later commits build upon this for some unit testing and instrumentation.) ---- thread through a tcx so that I can query types of lvalues as part of analysis. ---- Revised `BitDenotation::Ctxt`, allowing variation beyond `MoveData`. The main motivation is to ease threading through a `TyCtxt`. (In hindsight it might have been better to instead attach the `TyCtxt` to each of the different dataflow implementations, but that would require e.g. switching away from having a `Default` impl, so I am leaving that experiment for another time.)
2016-05-02 15:50:27 +02:00
/// Specifies the initial value for each bit in the `on_entry` set
fn bottom_value() -> bool;
Revised mir-dataflow. Incorporates many fixes contributed by arielb1. ---- revise borrowck::mir::dataflow code to allow varying domain for bitvectors. This particular code implements the `BitDenotation` trait for three analyses: * `MovingOutStatements`, which, like `borrowck::move_data`, maps each bit-index to a move instruction, and a 1 means "the effect of this move reaches this point" (and the assigned l-value, if a scoped declaration, is still in scope). * `MaybeInitializedLvals`, which maps each bit-index to an l-value. A 1 means "there exists a control flow path to this point that initializes the associated l-value." * `MaybeUninitializedLvals`, which maps each bit-index to an l-value A 1 means "there exists a control flow path to this point that de-initializes the associated l-value." ---- Revised `graphviz` dataflow-rendering support in `borrowck::mir`. One big difference is that this code is now parameterized over the `BitDenotation`, so that it can be used to render dataflow results independent of how the dataflow bitvectors are interpreted; see where reference to `MoveOut` is replaced by the type parameter `D`. ---- Factor out routine to query subattributes in `#[rustc_mir(..)]`. (Later commits build upon this for some unit testing and instrumentation.) ---- thread through a tcx so that I can query types of lvalues as part of analysis. ---- Revised `BitDenotation::Ctxt`, allowing variation beyond `MoveData`. The main motivation is to ease threading through a `TyCtxt`. (In hindsight it might have been better to instead attach the `TyCtxt` to each of the different dataflow implementations, but that would require e.g. switching away from having a `Default` impl, so I am leaving that experiment for another time.)
2016-05-02 15:50:27 +02:00
}
pub trait BitDenotation {
Revised mir-dataflow. Incorporates many fixes contributed by arielb1. ---- revise borrowck::mir::dataflow code to allow varying domain for bitvectors. This particular code implements the `BitDenotation` trait for three analyses: * `MovingOutStatements`, which, like `borrowck::move_data`, maps each bit-index to a move instruction, and a 1 means "the effect of this move reaches this point" (and the assigned l-value, if a scoped declaration, is still in scope). * `MaybeInitializedLvals`, which maps each bit-index to an l-value. A 1 means "there exists a control flow path to this point that initializes the associated l-value." * `MaybeUninitializedLvals`, which maps each bit-index to an l-value A 1 means "there exists a control flow path to this point that de-initializes the associated l-value." ---- Revised `graphviz` dataflow-rendering support in `borrowck::mir`. One big difference is that this code is now parameterized over the `BitDenotation`, so that it can be used to render dataflow results independent of how the dataflow bitvectors are interpreted; see where reference to `MoveOut` is replaced by the type parameter `D`. ---- Factor out routine to query subattributes in `#[rustc_mir(..)]`. (Later commits build upon this for some unit testing and instrumentation.) ---- thread through a tcx so that I can query types of lvalues as part of analysis. ---- Revised `BitDenotation::Ctxt`, allowing variation beyond `MoveData`. The main motivation is to ease threading through a `TyCtxt`. (In hindsight it might have been better to instead attach the `TyCtxt` to each of the different dataflow implementations, but that would require e.g. switching away from having a `Default` impl, so I am leaving that experiment for another time.)
2016-05-02 15:50:27 +02:00
/// Specifies what is represented by each bit in the dataflow bitvector.
type Bit;
/// Specifies what index type is used to access the bitvector.
type Idx: Idx;
Revised mir-dataflow. Incorporates many fixes contributed by arielb1. ---- revise borrowck::mir::dataflow code to allow varying domain for bitvectors. This particular code implements the `BitDenotation` trait for three analyses: * `MovingOutStatements`, which, like `borrowck::move_data`, maps each bit-index to a move instruction, and a 1 means "the effect of this move reaches this point" (and the assigned l-value, if a scoped declaration, is still in scope). * `MaybeInitializedLvals`, which maps each bit-index to an l-value. A 1 means "there exists a control flow path to this point that initializes the associated l-value." * `MaybeUninitializedLvals`, which maps each bit-index to an l-value A 1 means "there exists a control flow path to this point that de-initializes the associated l-value." ---- Revised `graphviz` dataflow-rendering support in `borrowck::mir`. One big difference is that this code is now parameterized over the `BitDenotation`, so that it can be used to render dataflow results independent of how the dataflow bitvectors are interpreted; see where reference to `MoveOut` is replaced by the type parameter `D`. ---- Factor out routine to query subattributes in `#[rustc_mir(..)]`. (Later commits build upon this for some unit testing and instrumentation.) ---- thread through a tcx so that I can query types of lvalues as part of analysis. ---- Revised `BitDenotation::Ctxt`, allowing variation beyond `MoveData`. The main motivation is to ease threading through a `TyCtxt`. (In hindsight it might have been better to instead attach the `TyCtxt` to each of the different dataflow implementations, but that would require e.g. switching away from having a `Default` impl, so I am leaving that experiment for another time.)
2016-05-02 15:50:27 +02:00
/// Specifies what, if any, separate context needs to be supplied for methods below.
type Ctxt;
/// A name describing the dataflow analysis that this
/// BitDenotation is supporting. The name should be something
/// suitable for plugging in as part of a filename e.g. avoid
/// space-characters or other things that tend to look bad on a
/// file system, like slashes or periods. It is also better for
/// the name to be reasonably short, again because it will be
/// plugged into a filename.
fn name() -> &'static str;
/// Size of each bitvector allocated for each block in the analysis.
fn bits_per_block(&self, &Self::Ctxt) -> usize;
/// Provides the meaning of each entry in the dataflow bitvector.
/// (Mostly intended for use for better debug instrumentation.)
fn interpret<'a>(&self, &'a Self::Ctxt, idx: usize) -> &'a Self::Bit;
/// Mutates the block-sets (the flow sets for the given
/// basic block) according to the effects that have been
/// established *prior* to entering the start block.
///
/// (For example, establishing the call arguments.)
///
/// (Typically this should only modify `sets.on_entry`, since the
/// gen and kill sets should reflect the effects of *executing*
/// the start block itself.)
fn start_block_effect(&self, ctxt: &Self::Ctxt, sets: &mut BlockSets<Self::Idx>);
Revised mir-dataflow. Incorporates many fixes contributed by arielb1. ---- revise borrowck::mir::dataflow code to allow varying domain for bitvectors. This particular code implements the `BitDenotation` trait for three analyses: * `MovingOutStatements`, which, like `borrowck::move_data`, maps each bit-index to a move instruction, and a 1 means "the effect of this move reaches this point" (and the assigned l-value, if a scoped declaration, is still in scope). * `MaybeInitializedLvals`, which maps each bit-index to an l-value. A 1 means "there exists a control flow path to this point that initializes the associated l-value." * `MaybeUninitializedLvals`, which maps each bit-index to an l-value A 1 means "there exists a control flow path to this point that de-initializes the associated l-value." ---- Revised `graphviz` dataflow-rendering support in `borrowck::mir`. One big difference is that this code is now parameterized over the `BitDenotation`, so that it can be used to render dataflow results independent of how the dataflow bitvectors are interpreted; see where reference to `MoveOut` is replaced by the type parameter `D`. ---- Factor out routine to query subattributes in `#[rustc_mir(..)]`. (Later commits build upon this for some unit testing and instrumentation.) ---- thread through a tcx so that I can query types of lvalues as part of analysis. ---- Revised `BitDenotation::Ctxt`, allowing variation beyond `MoveData`. The main motivation is to ease threading through a `TyCtxt`. (In hindsight it might have been better to instead attach the `TyCtxt` to each of the different dataflow implementations, but that would require e.g. switching away from having a `Default` impl, so I am leaving that experiment for another time.)
2016-05-02 15:50:27 +02:00
/// Mutates the block-sets (the flow sets for the given
/// basic block) according to the effects of evaluating statement.
///
/// This is used, in particular, for building up the
/// "transfer-function" represnting the overall-effect of the
/// block, represented via GEN and KILL sets.
///
/// The statement here is `idx_stmt.1`; `idx_stmt.0` is just
/// an identifying index: namely, the index of the statement
/// in the basic block.
fn statement_effect(&self,
ctxt: &Self::Ctxt,
sets: &mut BlockSets<Self::Idx>,
Revised mir-dataflow. Incorporates many fixes contributed by arielb1. ---- revise borrowck::mir::dataflow code to allow varying domain for bitvectors. This particular code implements the `BitDenotation` trait for three analyses: * `MovingOutStatements`, which, like `borrowck::move_data`, maps each bit-index to a move instruction, and a 1 means "the effect of this move reaches this point" (and the assigned l-value, if a scoped declaration, is still in scope). * `MaybeInitializedLvals`, which maps each bit-index to an l-value. A 1 means "there exists a control flow path to this point that initializes the associated l-value." * `MaybeUninitializedLvals`, which maps each bit-index to an l-value A 1 means "there exists a control flow path to this point that de-initializes the associated l-value." ---- Revised `graphviz` dataflow-rendering support in `borrowck::mir`. One big difference is that this code is now parameterized over the `BitDenotation`, so that it can be used to render dataflow results independent of how the dataflow bitvectors are interpreted; see where reference to `MoveOut` is replaced by the type parameter `D`. ---- Factor out routine to query subattributes in `#[rustc_mir(..)]`. (Later commits build upon this for some unit testing and instrumentation.) ---- thread through a tcx so that I can query types of lvalues as part of analysis. ---- Revised `BitDenotation::Ctxt`, allowing variation beyond `MoveData`. The main motivation is to ease threading through a `TyCtxt`. (In hindsight it might have been better to instead attach the `TyCtxt` to each of the different dataflow implementations, but that would require e.g. switching away from having a `Default` impl, so I am leaving that experiment for another time.)
2016-05-02 15:50:27 +02:00
bb: repr::BasicBlock,
idx_stmt: usize);
/// Mutates the block-sets (the flow sets for the given
/// basic block) according to the effects of evaluating
/// the terminator.
///
/// This is used, in particular, for building up the
/// "transfer-function" represnting the overall-effect of the
/// block, represented via GEN and KILL sets.
///
/// The terminator here is `idx_term.1`; `idx_term.0` is just an
/// identifying index: namely, the number of statements in `bb`
/// itself.
///
/// The effects applied here cannot depend on which branch the
/// terminator took.
fn terminator_effect(&self,
ctxt: &Self::Ctxt,
sets: &mut BlockSets<Self::Idx>,
Revised mir-dataflow. Incorporates many fixes contributed by arielb1. ---- revise borrowck::mir::dataflow code to allow varying domain for bitvectors. This particular code implements the `BitDenotation` trait for three analyses: * `MovingOutStatements`, which, like `borrowck::move_data`, maps each bit-index to a move instruction, and a 1 means "the effect of this move reaches this point" (and the assigned l-value, if a scoped declaration, is still in scope). * `MaybeInitializedLvals`, which maps each bit-index to an l-value. A 1 means "there exists a control flow path to this point that initializes the associated l-value." * `MaybeUninitializedLvals`, which maps each bit-index to an l-value A 1 means "there exists a control flow path to this point that de-initializes the associated l-value." ---- Revised `graphviz` dataflow-rendering support in `borrowck::mir`. One big difference is that this code is now parameterized over the `BitDenotation`, so that it can be used to render dataflow results independent of how the dataflow bitvectors are interpreted; see where reference to `MoveOut` is replaced by the type parameter `D`. ---- Factor out routine to query subattributes in `#[rustc_mir(..)]`. (Later commits build upon this for some unit testing and instrumentation.) ---- thread through a tcx so that I can query types of lvalues as part of analysis. ---- Revised `BitDenotation::Ctxt`, allowing variation beyond `MoveData`. The main motivation is to ease threading through a `TyCtxt`. (In hindsight it might have been better to instead attach the `TyCtxt` to each of the different dataflow implementations, but that would require e.g. switching away from having a `Default` impl, so I am leaving that experiment for another time.)
2016-05-02 15:50:27 +02:00
bb: repr::BasicBlock,
idx_term: usize);
/// Mutates the block-sets according to the (flow-dependent)
/// effect of a successful return from a Call terminator.
///
/// If basic-block BB_x ends with a call-instruction that, upon
/// successful return, flows to BB_y, then this method will be
/// called on the exit flow-state of BB_x in order to set up the
/// entry flow-state of BB_y.
///
/// This is used, in particular, as a special case during the
/// "propagate" loop where all of the basic blocks are repeatedly
/// visited. Since the effects of a Call terminator are
/// flow-dependent, the current MIR cannot encode them via just
/// GEN and KILL sets attached to the block, and so instead we add
/// this extra machinery to represent the flow-dependent effect.
fn propagate_call_return(&self,
ctxt: &Self::Ctxt,
in_out: &mut IdxSet<Self::Idx>,
Revised mir-dataflow. Incorporates many fixes contributed by arielb1. ---- revise borrowck::mir::dataflow code to allow varying domain for bitvectors. This particular code implements the `BitDenotation` trait for three analyses: * `MovingOutStatements`, which, like `borrowck::move_data`, maps each bit-index to a move instruction, and a 1 means "the effect of this move reaches this point" (and the assigned l-value, if a scoped declaration, is still in scope). * `MaybeInitializedLvals`, which maps each bit-index to an l-value. A 1 means "there exists a control flow path to this point that initializes the associated l-value." * `MaybeUninitializedLvals`, which maps each bit-index to an l-value A 1 means "there exists a control flow path to this point that de-initializes the associated l-value." ---- Revised `graphviz` dataflow-rendering support in `borrowck::mir`. One big difference is that this code is now parameterized over the `BitDenotation`, so that it can be used to render dataflow results independent of how the dataflow bitvectors are interpreted; see where reference to `MoveOut` is replaced by the type parameter `D`. ---- Factor out routine to query subattributes in `#[rustc_mir(..)]`. (Later commits build upon this for some unit testing and instrumentation.) ---- thread through a tcx so that I can query types of lvalues as part of analysis. ---- Revised `BitDenotation::Ctxt`, allowing variation beyond `MoveData`. The main motivation is to ease threading through a `TyCtxt`. (In hindsight it might have been better to instead attach the `TyCtxt` to each of the different dataflow implementations, but that would require e.g. switching away from having a `Default` impl, so I am leaving that experiment for another time.)
2016-05-02 15:50:27 +02:00
call_bb: repr::BasicBlock,
dest_bb: repr::BasicBlock,
dest_lval: &repr::Lvalue);
}
impl<'a, 'tcx: 'a, D> DataflowAnalysis<'a, 'tcx, D>
where D: BitDenotation + DataflowOperator,
D::Ctxt: HasMoveData<'tcx>
Revised mir-dataflow. Incorporates many fixes contributed by arielb1. ---- revise borrowck::mir::dataflow code to allow varying domain for bitvectors. This particular code implements the `BitDenotation` trait for three analyses: * `MovingOutStatements`, which, like `borrowck::move_data`, maps each bit-index to a move instruction, and a 1 means "the effect of this move reaches this point" (and the assigned l-value, if a scoped declaration, is still in scope). * `MaybeInitializedLvals`, which maps each bit-index to an l-value. A 1 means "there exists a control flow path to this point that initializes the associated l-value." * `MaybeUninitializedLvals`, which maps each bit-index to an l-value A 1 means "there exists a control flow path to this point that de-initializes the associated l-value." ---- Revised `graphviz` dataflow-rendering support in `borrowck::mir`. One big difference is that this code is now parameterized over the `BitDenotation`, so that it can be used to render dataflow results independent of how the dataflow bitvectors are interpreted; see where reference to `MoveOut` is replaced by the type parameter `D`. ---- Factor out routine to query subattributes in `#[rustc_mir(..)]`. (Later commits build upon this for some unit testing and instrumentation.) ---- thread through a tcx so that I can query types of lvalues as part of analysis. ---- Revised `BitDenotation::Ctxt`, allowing variation beyond `MoveData`. The main motivation is to ease threading through a `TyCtxt`. (In hindsight it might have been better to instead attach the `TyCtxt` to each of the different dataflow implementations, but that would require e.g. switching away from having a `Default` impl, so I am leaving that experiment for another time.)
2016-05-02 15:50:27 +02:00
{
pub fn new(_tcx: TyCtxt<'a, 'tcx, 'tcx>,
mir: &'a Mir<'tcx>,
ctxt: &'a D::Ctxt,
Revised mir-dataflow. Incorporates many fixes contributed by arielb1. ---- revise borrowck::mir::dataflow code to allow varying domain for bitvectors. This particular code implements the `BitDenotation` trait for three analyses: * `MovingOutStatements`, which, like `borrowck::move_data`, maps each bit-index to a move instruction, and a 1 means "the effect of this move reaches this point" (and the assigned l-value, if a scoped declaration, is still in scope). * `MaybeInitializedLvals`, which maps each bit-index to an l-value. A 1 means "there exists a control flow path to this point that initializes the associated l-value." * `MaybeUninitializedLvals`, which maps each bit-index to an l-value A 1 means "there exists a control flow path to this point that de-initializes the associated l-value." ---- Revised `graphviz` dataflow-rendering support in `borrowck::mir`. One big difference is that this code is now parameterized over the `BitDenotation`, so that it can be used to render dataflow results independent of how the dataflow bitvectors are interpreted; see where reference to `MoveOut` is replaced by the type parameter `D`. ---- Factor out routine to query subattributes in `#[rustc_mir(..)]`. (Later commits build upon this for some unit testing and instrumentation.) ---- thread through a tcx so that I can query types of lvalues as part of analysis. ---- Revised `BitDenotation::Ctxt`, allowing variation beyond `MoveData`. The main motivation is to ease threading through a `TyCtxt`. (In hindsight it might have been better to instead attach the `TyCtxt` to each of the different dataflow implementations, but that would require e.g. switching away from having a `Default` impl, so I am leaving that experiment for another time.)
2016-05-02 15:50:27 +02:00
denotation: D) -> Self {
let bits_per_block = denotation.bits_per_block(&ctxt);
let usize_bits = mem::size_of::<usize>() * 8;
let words_per_block = (bits_per_block + usize_bits - 1) / usize_bits;
// (now rounded up to multiple of word size)
let bits_per_block = words_per_block * usize_bits;
Revised mir-dataflow. Incorporates many fixes contributed by arielb1. ---- revise borrowck::mir::dataflow code to allow varying domain for bitvectors. This particular code implements the `BitDenotation` trait for three analyses: * `MovingOutStatements`, which, like `borrowck::move_data`, maps each bit-index to a move instruction, and a 1 means "the effect of this move reaches this point" (and the assigned l-value, if a scoped declaration, is still in scope). * `MaybeInitializedLvals`, which maps each bit-index to an l-value. A 1 means "there exists a control flow path to this point that initializes the associated l-value." * `MaybeUninitializedLvals`, which maps each bit-index to an l-value A 1 means "there exists a control flow path to this point that de-initializes the associated l-value." ---- Revised `graphviz` dataflow-rendering support in `borrowck::mir`. One big difference is that this code is now parameterized over the `BitDenotation`, so that it can be used to render dataflow results independent of how the dataflow bitvectors are interpreted; see where reference to `MoveOut` is replaced by the type parameter `D`. ---- Factor out routine to query subattributes in `#[rustc_mir(..)]`. (Later commits build upon this for some unit testing and instrumentation.) ---- thread through a tcx so that I can query types of lvalues as part of analysis. ---- Revised `BitDenotation::Ctxt`, allowing variation beyond `MoveData`. The main motivation is to ease threading through a `TyCtxt`. (In hindsight it might have been better to instead attach the `TyCtxt` to each of the different dataflow implementations, but that would require e.g. switching away from having a `Default` impl, so I am leaving that experiment for another time.)
2016-05-02 15:50:27 +02:00
let num_blocks = mir.basic_blocks.len();
let num_overall = num_blocks * bits_per_block;
Revised mir-dataflow. Incorporates many fixes contributed by arielb1. ---- revise borrowck::mir::dataflow code to allow varying domain for bitvectors. This particular code implements the `BitDenotation` trait for three analyses: * `MovingOutStatements`, which, like `borrowck::move_data`, maps each bit-index to a move instruction, and a 1 means "the effect of this move reaches this point" (and the assigned l-value, if a scoped declaration, is still in scope). * `MaybeInitializedLvals`, which maps each bit-index to an l-value. A 1 means "there exists a control flow path to this point that initializes the associated l-value." * `MaybeUninitializedLvals`, which maps each bit-index to an l-value A 1 means "there exists a control flow path to this point that de-initializes the associated l-value." ---- Revised `graphviz` dataflow-rendering support in `borrowck::mir`. One big difference is that this code is now parameterized over the `BitDenotation`, so that it can be used to render dataflow results independent of how the dataflow bitvectors are interpreted; see where reference to `MoveOut` is replaced by the type parameter `D`. ---- Factor out routine to query subattributes in `#[rustc_mir(..)]`. (Later commits build upon this for some unit testing and instrumentation.) ---- thread through a tcx so that I can query types of lvalues as part of analysis. ---- Revised `BitDenotation::Ctxt`, allowing variation beyond `MoveData`. The main motivation is to ease threading through a `TyCtxt`. (In hindsight it might have been better to instead attach the `TyCtxt` to each of the different dataflow implementations, but that would require e.g. switching away from having a `Default` impl, so I am leaving that experiment for another time.)
2016-05-02 15:50:27 +02:00
let zeroes = Bits::new(OwnIdxSet::new_empty(num_overall));
let on_entry = Bits::new(if D::bottom_value() {
OwnIdxSet::new_filled(num_overall)
} else {
OwnIdxSet::new_empty(num_overall)
});
DataflowAnalysis {
ctxt: ctxt,
mir: mir,
flow_state: DataflowState {
sets: AllSets {
bits_per_block: bits_per_block,
words_per_block: words_per_block,
gen_sets: zeroes.clone(),
kill_sets: zeroes,
on_entry_sets: on_entry,
},
operator: denotation,
Revised mir-dataflow. Incorporates many fixes contributed by arielb1. ---- revise borrowck::mir::dataflow code to allow varying domain for bitvectors. This particular code implements the `BitDenotation` trait for three analyses: * `MovingOutStatements`, which, like `borrowck::move_data`, maps each bit-index to a move instruction, and a 1 means "the effect of this move reaches this point" (and the assigned l-value, if a scoped declaration, is still in scope). * `MaybeInitializedLvals`, which maps each bit-index to an l-value. A 1 means "there exists a control flow path to this point that initializes the associated l-value." * `MaybeUninitializedLvals`, which maps each bit-index to an l-value A 1 means "there exists a control flow path to this point that de-initializes the associated l-value." ---- Revised `graphviz` dataflow-rendering support in `borrowck::mir`. One big difference is that this code is now parameterized over the `BitDenotation`, so that it can be used to render dataflow results independent of how the dataflow bitvectors are interpreted; see where reference to `MoveOut` is replaced by the type parameter `D`. ---- Factor out routine to query subattributes in `#[rustc_mir(..)]`. (Later commits build upon this for some unit testing and instrumentation.) ---- thread through a tcx so that I can query types of lvalues as part of analysis. ---- Revised `BitDenotation::Ctxt`, allowing variation beyond `MoveData`. The main motivation is to ease threading through a `TyCtxt`. (In hindsight it might have been better to instead attach the `TyCtxt` to each of the different dataflow implementations, but that would require e.g. switching away from having a `Default` impl, so I am leaving that experiment for another time.)
2016-05-02 15:50:27 +02:00
},
}
}
}
impl<'a, 'tcx: 'a, D> DataflowAnalysis<'a, 'tcx, D>
where D: BitDenotation + DataflowOperator,
D::Ctxt: HasMoveData<'tcx>,
Revised mir-dataflow. Incorporates many fixes contributed by arielb1. ---- revise borrowck::mir::dataflow code to allow varying domain for bitvectors. This particular code implements the `BitDenotation` trait for three analyses: * `MovingOutStatements`, which, like `borrowck::move_data`, maps each bit-index to a move instruction, and a 1 means "the effect of this move reaches this point" (and the assigned l-value, if a scoped declaration, is still in scope). * `MaybeInitializedLvals`, which maps each bit-index to an l-value. A 1 means "there exists a control flow path to this point that initializes the associated l-value." * `MaybeUninitializedLvals`, which maps each bit-index to an l-value A 1 means "there exists a control flow path to this point that de-initializes the associated l-value." ---- Revised `graphviz` dataflow-rendering support in `borrowck::mir`. One big difference is that this code is now parameterized over the `BitDenotation`, so that it can be used to render dataflow results independent of how the dataflow bitvectors are interpreted; see where reference to `MoveOut` is replaced by the type parameter `D`. ---- Factor out routine to query subattributes in `#[rustc_mir(..)]`. (Later commits build upon this for some unit testing and instrumentation.) ---- thread through a tcx so that I can query types of lvalues as part of analysis. ---- Revised `BitDenotation::Ctxt`, allowing variation beyond `MoveData`. The main motivation is to ease threading through a `TyCtxt`. (In hindsight it might have been better to instead attach the `TyCtxt` to each of the different dataflow implementations, but that would require e.g. switching away from having a `Default` impl, so I am leaving that experiment for another time.)
2016-05-02 15:50:27 +02:00
{
/// Propagates the bits of `in_out` into all the successors of `bb`,
/// using bitwise operator denoted by `self.operator`.
///
/// For most blocks, this is entirely uniform. However, for blocks
/// that end with a call terminator, the effect of the call on the
/// dataflow state may depend on whether the call returned
/// successfully or unwound.
///
/// To reflect this, the `propagate_call_return` method of the
/// `BitDenotation` mutates `in_out` when propagating `in_out` via
/// a call terminator; such mutation is performed *last*, to
/// ensure its side-effects do not leak elsewhere (e.g. into
/// unwind target).
fn propagate_bits_into_graph_successors_of(
&mut self,
in_out: &mut IdxSet<D::Idx>,
Revised mir-dataflow. Incorporates many fixes contributed by arielb1. ---- revise borrowck::mir::dataflow code to allow varying domain for bitvectors. This particular code implements the `BitDenotation` trait for three analyses: * `MovingOutStatements`, which, like `borrowck::move_data`, maps each bit-index to a move instruction, and a 1 means "the effect of this move reaches this point" (and the assigned l-value, if a scoped declaration, is still in scope). * `MaybeInitializedLvals`, which maps each bit-index to an l-value. A 1 means "there exists a control flow path to this point that initializes the associated l-value." * `MaybeUninitializedLvals`, which maps each bit-index to an l-value A 1 means "there exists a control flow path to this point that de-initializes the associated l-value." ---- Revised `graphviz` dataflow-rendering support in `borrowck::mir`. One big difference is that this code is now parameterized over the `BitDenotation`, so that it can be used to render dataflow results independent of how the dataflow bitvectors are interpreted; see where reference to `MoveOut` is replaced by the type parameter `D`. ---- Factor out routine to query subattributes in `#[rustc_mir(..)]`. (Later commits build upon this for some unit testing and instrumentation.) ---- thread through a tcx so that I can query types of lvalues as part of analysis. ---- Revised `BitDenotation::Ctxt`, allowing variation beyond `MoveData`. The main motivation is to ease threading through a `TyCtxt`. (In hindsight it might have been better to instead attach the `TyCtxt` to each of the different dataflow implementations, but that would require e.g. switching away from having a `Default` impl, so I am leaving that experiment for another time.)
2016-05-02 15:50:27 +02:00
changed: &mut bool,
(bb, bb_data): (repr::BasicBlock, &repr::BasicBlockData))
{
match bb_data.terminator().kind {
repr::TerminatorKind::Return |
repr::TerminatorKind::Resume => {}
repr::TerminatorKind::Goto { ref target } |
repr::TerminatorKind::Drop { ref target, value: _, unwind: None } => {
self.propagate_bits_into_entry_set_for(in_out, changed, target);
}
repr::TerminatorKind::Drop { ref target, value: _, unwind: Some(ref unwind) } => {
self.propagate_bits_into_entry_set_for(in_out, changed, target);
self.propagate_bits_into_entry_set_for(in_out, changed, unwind);
}
repr::TerminatorKind::If { ref targets, .. } => {
self.propagate_bits_into_entry_set_for(in_out, changed, &targets.0);
self.propagate_bits_into_entry_set_for(in_out, changed, &targets.1);
}
repr::TerminatorKind::Switch { ref targets, .. } |
repr::TerminatorKind::SwitchInt { ref targets, .. } => {
for target in targets {
self.propagate_bits_into_entry_set_for(in_out, changed, target);
}
}
repr::TerminatorKind::Call { ref cleanup, ref destination, func: _, args: _ } => {
if let Some(ref unwind) = *cleanup {
self.propagate_bits_into_entry_set_for(in_out, changed, unwind);
}
if let Some((ref dest_lval, ref dest_bb)) = *destination {
// N.B.: This must be done *last*, after all other
// propagation, as documented in comment above.
self.flow_state.operator.propagate_call_return(
&self.ctxt, in_out, bb, *dest_bb, dest_lval);
self.propagate_bits_into_entry_set_for(in_out, changed, dest_bb);
}
}
}
}
fn propagate_bits_into_entry_set_for(&mut self,
in_out: &IdxSet<D::Idx>,
Revised mir-dataflow. Incorporates many fixes contributed by arielb1. ---- revise borrowck::mir::dataflow code to allow varying domain for bitvectors. This particular code implements the `BitDenotation` trait for three analyses: * `MovingOutStatements`, which, like `borrowck::move_data`, maps each bit-index to a move instruction, and a 1 means "the effect of this move reaches this point" (and the assigned l-value, if a scoped declaration, is still in scope). * `MaybeInitializedLvals`, which maps each bit-index to an l-value. A 1 means "there exists a control flow path to this point that initializes the associated l-value." * `MaybeUninitializedLvals`, which maps each bit-index to an l-value A 1 means "there exists a control flow path to this point that de-initializes the associated l-value." ---- Revised `graphviz` dataflow-rendering support in `borrowck::mir`. One big difference is that this code is now parameterized over the `BitDenotation`, so that it can be used to render dataflow results independent of how the dataflow bitvectors are interpreted; see where reference to `MoveOut` is replaced by the type parameter `D`. ---- Factor out routine to query subattributes in `#[rustc_mir(..)]`. (Later commits build upon this for some unit testing and instrumentation.) ---- thread through a tcx so that I can query types of lvalues as part of analysis. ---- Revised `BitDenotation::Ctxt`, allowing variation beyond `MoveData`. The main motivation is to ease threading through a `TyCtxt`. (In hindsight it might have been better to instead attach the `TyCtxt` to each of the different dataflow implementations, but that would require e.g. switching away from having a `Default` impl, so I am leaving that experiment for another time.)
2016-05-02 15:50:27 +02:00
changed: &mut bool,
bb: &repr::BasicBlock) {
let entry_set = self.flow_state.sets.for_block(bb.index()).on_entry;
let set_changed = bitwise(entry_set.words_mut(),
in_out.words(),
&self.flow_state.operator);
Revised mir-dataflow. Incorporates many fixes contributed by arielb1. ---- revise borrowck::mir::dataflow code to allow varying domain for bitvectors. This particular code implements the `BitDenotation` trait for three analyses: * `MovingOutStatements`, which, like `borrowck::move_data`, maps each bit-index to a move instruction, and a 1 means "the effect of this move reaches this point" (and the assigned l-value, if a scoped declaration, is still in scope). * `MaybeInitializedLvals`, which maps each bit-index to an l-value. A 1 means "there exists a control flow path to this point that initializes the associated l-value." * `MaybeUninitializedLvals`, which maps each bit-index to an l-value A 1 means "there exists a control flow path to this point that de-initializes the associated l-value." ---- Revised `graphviz` dataflow-rendering support in `borrowck::mir`. One big difference is that this code is now parameterized over the `BitDenotation`, so that it can be used to render dataflow results independent of how the dataflow bitvectors are interpreted; see where reference to `MoveOut` is replaced by the type parameter `D`. ---- Factor out routine to query subattributes in `#[rustc_mir(..)]`. (Later commits build upon this for some unit testing and instrumentation.) ---- thread through a tcx so that I can query types of lvalues as part of analysis. ---- Revised `BitDenotation::Ctxt`, allowing variation beyond `MoveData`. The main motivation is to ease threading through a `TyCtxt`. (In hindsight it might have been better to instead attach the `TyCtxt` to each of the different dataflow implementations, but that would require e.g. switching away from having a `Default` impl, so I am leaving that experiment for another time.)
2016-05-02 15:50:27 +02:00
if set_changed {
*changed = true;
}
}
}