rust/compiler/rustc_codegen_llvm/src/coverageinfo/mod.rs

Ignoring revisions in .git-blame-ignore-revs. Click here to bypass and see the normal blame view.

226 lines
10 KiB
Rust
Raw Normal View History

use std::cell::{OnceCell, RefCell};
use std::ffi::{CStr, CString};
use rustc_abi::Size;
use rustc_codegen_ssa::traits::{
BuilderMethods, ConstCodegenMethods, CoverageInfoBuilderMethods, MiscCodegenMethods,
};
use rustc_data_structures::fx::{FxHashMap, FxIndexMap};
use rustc_middle::mir::coverage::CoverageKind;
use rustc_middle::ty::Instance;
use rustc_middle::ty::layout::HasTyCtxt;
use tracing::{debug, instrument};
use crate::builder::Builder;
use crate::common::CodegenCx;
use crate::coverageinfo::map_data::FunctionCoverageCollector;
use crate::llvm;
pub(crate) mod ffi;
mod llvm_cov;
pub(crate) mod map_data;
mod mapgen;
/// A context object for maintaining all state needed by the coverageinfo module.
pub(crate) struct CrateCoverageContext<'ll, 'tcx> {
/// Coverage data for each instrumented function identified by DefId.
pub(crate) function_coverage_map:
RefCell<FxIndexMap<Instance<'tcx>, FunctionCoverageCollector<'tcx>>>,
coverage bug fixes and optimization support Adjusted LLVM codegen for code compiled with `-Zinstrument-coverage` to address multiple, somewhat related issues. Fixed a significant flaw in prior coverage solution: Every counter generated a new counter variable, but there should have only been one counter variable per function. This appears to have bloated .profraw files significantly. (For a small program, it increased the size by about 40%. I have not tested large programs, but there is anecdotal evidence that profraw files were way too large. This is a good fix, regardless, but hopefully it also addresses related issues. Fixes: #82144 Invalid LLVM coverage data produced when compiled with -C opt-level=1 Existing tests now work up to at least `opt-level=3`. This required a detailed analysis of the LLVM IR, comparisons with Clang C++ LLVM IR when compiled with coverage, and a lot of trial and error with codegen adjustments. The biggest hurdle was figuring out how to continue to support coverage results for unused functions and generics. Rust's coverage results have three advantages over Clang's coverage results: 1. Rust's coverage map does not include any overlapping code regions, making coverage counting unambiguous. 2. Rust generates coverage results (showing zero counts) for all unused functions, including generics. (Clang does not generate coverage for uninstantiated template functions.) 3. Rust's unused functions produce minimal stubbed functions in LLVM IR, sufficient for including in the coverage results; while Clang must generate the complete LLVM IR for each unused function, even though it will never be called. This PR removes the previous hack of attempting to inject coverage into some other existing function instance, and generates dedicated instances for each unused function. This change, and a few other adjustments (similar to what is required for `-C link-dead-code`, but with lower impact), makes it possible to support LLVM optimizations. Fixes: #79651 Coverage report: "Unexecuted instantiation:..." for a generic function from multiple crates Fixed by removing the aforementioned hack. Some "Unexecuted instantiation" notices are unavoidable, as explained in the `used_crate.rs` test, but `-Zinstrument-coverage` has new options to back off support for either unused generics, or all unused functions, which avoids the notice, at the cost of less coverage of unused functions. Fixes: #82875 Invalid LLVM coverage data produced with crate brotli_decompressor Fixed by disabling the LLVM function attribute that forces inlining, if `-Z instrument-coverage` is enabled. This attribute is applied to Rust functions with `#[inline(always)], and in some cases, the forced inlining breaks coverage instrumentation and reports.
2021-03-15 23:32:45 +00:00
pub(crate) pgo_func_name_var_map: RefCell<FxHashMap<Instance<'tcx>, &'ll llvm::Value>>,
pub(crate) mcdc_condition_bitmap_map: RefCell<FxHashMap<Instance<'tcx>, Vec<&'ll llvm::Value>>>,
covfun_section_name: OnceCell<CString>,
}
coverage bug fixes and optimization support Adjusted LLVM codegen for code compiled with `-Zinstrument-coverage` to address multiple, somewhat related issues. Fixed a significant flaw in prior coverage solution: Every counter generated a new counter variable, but there should have only been one counter variable per function. This appears to have bloated .profraw files significantly. (For a small program, it increased the size by about 40%. I have not tested large programs, but there is anecdotal evidence that profraw files were way too large. This is a good fix, regardless, but hopefully it also addresses related issues. Fixes: #82144 Invalid LLVM coverage data produced when compiled with -C opt-level=1 Existing tests now work up to at least `opt-level=3`. This required a detailed analysis of the LLVM IR, comparisons with Clang C++ LLVM IR when compiled with coverage, and a lot of trial and error with codegen adjustments. The biggest hurdle was figuring out how to continue to support coverage results for unused functions and generics. Rust's coverage results have three advantages over Clang's coverage results: 1. Rust's coverage map does not include any overlapping code regions, making coverage counting unambiguous. 2. Rust generates coverage results (showing zero counts) for all unused functions, including generics. (Clang does not generate coverage for uninstantiated template functions.) 3. Rust's unused functions produce minimal stubbed functions in LLVM IR, sufficient for including in the coverage results; while Clang must generate the complete LLVM IR for each unused function, even though it will never be called. This PR removes the previous hack of attempting to inject coverage into some other existing function instance, and generates dedicated instances for each unused function. This change, and a few other adjustments (similar to what is required for `-C link-dead-code`, but with lower impact), makes it possible to support LLVM optimizations. Fixes: #79651 Coverage report: "Unexecuted instantiation:..." for a generic function from multiple crates Fixed by removing the aforementioned hack. Some "Unexecuted instantiation" notices are unavoidable, as explained in the `used_crate.rs` test, but `-Zinstrument-coverage` has new options to back off support for either unused generics, or all unused functions, which avoids the notice, at the cost of less coverage of unused functions. Fixes: #82875 Invalid LLVM coverage data produced with crate brotli_decompressor Fixed by disabling the LLVM function attribute that forces inlining, if `-Z instrument-coverage` is enabled. This attribute is applied to Rust functions with `#[inline(always)], and in some cases, the forced inlining breaks coverage instrumentation and reports.
2021-03-15 23:32:45 +00:00
impl<'ll, 'tcx> CrateCoverageContext<'ll, 'tcx> {
pub(crate) fn new() -> Self {
coverage bug fixes and optimization support Adjusted LLVM codegen for code compiled with `-Zinstrument-coverage` to address multiple, somewhat related issues. Fixed a significant flaw in prior coverage solution: Every counter generated a new counter variable, but there should have only been one counter variable per function. This appears to have bloated .profraw files significantly. (For a small program, it increased the size by about 40%. I have not tested large programs, but there is anecdotal evidence that profraw files were way too large. This is a good fix, regardless, but hopefully it also addresses related issues. Fixes: #82144 Invalid LLVM coverage data produced when compiled with -C opt-level=1 Existing tests now work up to at least `opt-level=3`. This required a detailed analysis of the LLVM IR, comparisons with Clang C++ LLVM IR when compiled with coverage, and a lot of trial and error with codegen adjustments. The biggest hurdle was figuring out how to continue to support coverage results for unused functions and generics. Rust's coverage results have three advantages over Clang's coverage results: 1. Rust's coverage map does not include any overlapping code regions, making coverage counting unambiguous. 2. Rust generates coverage results (showing zero counts) for all unused functions, including generics. (Clang does not generate coverage for uninstantiated template functions.) 3. Rust's unused functions produce minimal stubbed functions in LLVM IR, sufficient for including in the coverage results; while Clang must generate the complete LLVM IR for each unused function, even though it will never be called. This PR removes the previous hack of attempting to inject coverage into some other existing function instance, and generates dedicated instances for each unused function. This change, and a few other adjustments (similar to what is required for `-C link-dead-code`, but with lower impact), makes it possible to support LLVM optimizations. Fixes: #79651 Coverage report: "Unexecuted instantiation:..." for a generic function from multiple crates Fixed by removing the aforementioned hack. Some "Unexecuted instantiation" notices are unavoidable, as explained in the `used_crate.rs` test, but `-Zinstrument-coverage` has new options to back off support for either unused generics, or all unused functions, which avoids the notice, at the cost of less coverage of unused functions. Fixes: #82875 Invalid LLVM coverage data produced with crate brotli_decompressor Fixed by disabling the LLVM function attribute that forces inlining, if `-Z instrument-coverage` is enabled. This attribute is applied to Rust functions with `#[inline(always)], and in some cases, the forced inlining breaks coverage instrumentation and reports.
2021-03-15 23:32:45 +00:00
Self {
function_coverage_map: Default::default(),
pgo_func_name_var_map: Default::default(),
mcdc_condition_bitmap_map: Default::default(),
covfun_section_name: Default::default(),
coverage bug fixes and optimization support Adjusted LLVM codegen for code compiled with `-Zinstrument-coverage` to address multiple, somewhat related issues. Fixed a significant flaw in prior coverage solution: Every counter generated a new counter variable, but there should have only been one counter variable per function. This appears to have bloated .profraw files significantly. (For a small program, it increased the size by about 40%. I have not tested large programs, but there is anecdotal evidence that profraw files were way too large. This is a good fix, regardless, but hopefully it also addresses related issues. Fixes: #82144 Invalid LLVM coverage data produced when compiled with -C opt-level=1 Existing tests now work up to at least `opt-level=3`. This required a detailed analysis of the LLVM IR, comparisons with Clang C++ LLVM IR when compiled with coverage, and a lot of trial and error with codegen adjustments. The biggest hurdle was figuring out how to continue to support coverage results for unused functions and generics. Rust's coverage results have three advantages over Clang's coverage results: 1. Rust's coverage map does not include any overlapping code regions, making coverage counting unambiguous. 2. Rust generates coverage results (showing zero counts) for all unused functions, including generics. (Clang does not generate coverage for uninstantiated template functions.) 3. Rust's unused functions produce minimal stubbed functions in LLVM IR, sufficient for including in the coverage results; while Clang must generate the complete LLVM IR for each unused function, even though it will never be called. This PR removes the previous hack of attempting to inject coverage into some other existing function instance, and generates dedicated instances for each unused function. This change, and a few other adjustments (similar to what is required for `-C link-dead-code`, but with lower impact), makes it possible to support LLVM optimizations. Fixes: #79651 Coverage report: "Unexecuted instantiation:..." for a generic function from multiple crates Fixed by removing the aforementioned hack. Some "Unexecuted instantiation" notices are unavoidable, as explained in the `used_crate.rs` test, but `-Zinstrument-coverage` has new options to back off support for either unused generics, or all unused functions, which avoids the notice, at the cost of less coverage of unused functions. Fixes: #82875 Invalid LLVM coverage data produced with crate brotli_decompressor Fixed by disabling the LLVM function attribute that forces inlining, if `-Z instrument-coverage` is enabled. This attribute is applied to Rust functions with `#[inline(always)], and in some cases, the forced inlining breaks coverage instrumentation and reports.
2021-03-15 23:32:45 +00:00
}
}
fn take_function_coverage_map(
&self,
) -> FxIndexMap<Instance<'tcx>, FunctionCoverageCollector<'tcx>> {
self.function_coverage_map.replace(FxIndexMap::default())
}
/// LLVM use a temp value to record evaluated mcdc test vector of each decision, which is
/// called condition bitmap. In order to handle nested decisions, several condition bitmaps can
/// be allocated for a function body. These values are named `mcdc.addr.{i}` and are a 32-bit
/// integers. They respectively hold the condition bitmaps for decisions with a depth of `i`.
fn try_get_mcdc_condition_bitmap(
&self,
instance: &Instance<'tcx>,
decision_depth: u16,
) -> Option<&'ll llvm::Value> {
self.mcdc_condition_bitmap_map
.borrow()
.get(instance)
.and_then(|bitmap_map| bitmap_map.get(decision_depth as usize))
.copied() // Dereference Option<&&Value> to Option<&Value>
}
}
impl<'ll, 'tcx> CodegenCx<'ll, 'tcx> {
pub(crate) fn coverageinfo_finalize(&self) {
mapgen::finalize(self)
}
/// Returns the section name to use when embedding per-function coverage information
/// in the object file, according to the target's object file format. LLVM's coverage
/// tools use information from this section when producing coverage reports.
///
/// Typical values are:
/// - `__llvm_covfun` on Linux
/// - `__LLVM_COV,__llvm_covfun` on macOS (includes `__LLVM_COV,` segment prefix)
/// - `.lcovfun$M` on Windows (includes `$M` sorting suffix)
fn covfun_section_name(&self) -> &CStr {
self.coverage_cx()
.covfun_section_name
.get_or_init(|| llvm_cov::covfun_section_name(self.llmod))
}
/// For LLVM codegen, returns a function-specific `Value` for a global
/// string, to hold the function name passed to LLVM intrinsic
/// `instrprof.increment()`. The `Value` is only created once per instance.
/// Multiple invocations with the same instance return the same `Value`.
coverage bug fixes and optimization support Adjusted LLVM codegen for code compiled with `-Zinstrument-coverage` to address multiple, somewhat related issues. Fixed a significant flaw in prior coverage solution: Every counter generated a new counter variable, but there should have only been one counter variable per function. This appears to have bloated .profraw files significantly. (For a small program, it increased the size by about 40%. I have not tested large programs, but there is anecdotal evidence that profraw files were way too large. This is a good fix, regardless, but hopefully it also addresses related issues. Fixes: #82144 Invalid LLVM coverage data produced when compiled with -C opt-level=1 Existing tests now work up to at least `opt-level=3`. This required a detailed analysis of the LLVM IR, comparisons with Clang C++ LLVM IR when compiled with coverage, and a lot of trial and error with codegen adjustments. The biggest hurdle was figuring out how to continue to support coverage results for unused functions and generics. Rust's coverage results have three advantages over Clang's coverage results: 1. Rust's coverage map does not include any overlapping code regions, making coverage counting unambiguous. 2. Rust generates coverage results (showing zero counts) for all unused functions, including generics. (Clang does not generate coverage for uninstantiated template functions.) 3. Rust's unused functions produce minimal stubbed functions in LLVM IR, sufficient for including in the coverage results; while Clang must generate the complete LLVM IR for each unused function, even though it will never be called. This PR removes the previous hack of attempting to inject coverage into some other existing function instance, and generates dedicated instances for each unused function. This change, and a few other adjustments (similar to what is required for `-C link-dead-code`, but with lower impact), makes it possible to support LLVM optimizations. Fixes: #79651 Coverage report: "Unexecuted instantiation:..." for a generic function from multiple crates Fixed by removing the aforementioned hack. Some "Unexecuted instantiation" notices are unavoidable, as explained in the `used_crate.rs` test, but `-Zinstrument-coverage` has new options to back off support for either unused generics, or all unused functions, which avoids the notice, at the cost of less coverage of unused functions. Fixes: #82875 Invalid LLVM coverage data produced with crate brotli_decompressor Fixed by disabling the LLVM function attribute that forces inlining, if `-Z instrument-coverage` is enabled. This attribute is applied to Rust functions with `#[inline(always)], and in some cases, the forced inlining breaks coverage instrumentation and reports.
2021-03-15 23:32:45 +00:00
fn get_pgo_func_name_var(&self, instance: Instance<'tcx>) -> &'ll llvm::Value {
debug!("getting pgo_func_name_var for instance={:?}", instance);
let mut pgo_func_name_var_map = self.coverage_cx().pgo_func_name_var_map.borrow_mut();
pgo_func_name_var_map.entry(instance).or_insert_with(|| {
let llfn = self.get_fn(instance);
let mangled_fn_name: &str = self.tcx.symbol_name(instance).name;
llvm_cov::create_pgo_func_name_var(llfn, mangled_fn_name)
})
coverage bug fixes and optimization support Adjusted LLVM codegen for code compiled with `-Zinstrument-coverage` to address multiple, somewhat related issues. Fixed a significant flaw in prior coverage solution: Every counter generated a new counter variable, but there should have only been one counter variable per function. This appears to have bloated .profraw files significantly. (For a small program, it increased the size by about 40%. I have not tested large programs, but there is anecdotal evidence that profraw files were way too large. This is a good fix, regardless, but hopefully it also addresses related issues. Fixes: #82144 Invalid LLVM coverage data produced when compiled with -C opt-level=1 Existing tests now work up to at least `opt-level=3`. This required a detailed analysis of the LLVM IR, comparisons with Clang C++ LLVM IR when compiled with coverage, and a lot of trial and error with codegen adjustments. The biggest hurdle was figuring out how to continue to support coverage results for unused functions and generics. Rust's coverage results have three advantages over Clang's coverage results: 1. Rust's coverage map does not include any overlapping code regions, making coverage counting unambiguous. 2. Rust generates coverage results (showing zero counts) for all unused functions, including generics. (Clang does not generate coverage for uninstantiated template functions.) 3. Rust's unused functions produce minimal stubbed functions in LLVM IR, sufficient for including in the coverage results; while Clang must generate the complete LLVM IR for each unused function, even though it will never be called. This PR removes the previous hack of attempting to inject coverage into some other existing function instance, and generates dedicated instances for each unused function. This change, and a few other adjustments (similar to what is required for `-C link-dead-code`, but with lower impact), makes it possible to support LLVM optimizations. Fixes: #79651 Coverage report: "Unexecuted instantiation:..." for a generic function from multiple crates Fixed by removing the aforementioned hack. Some "Unexecuted instantiation" notices are unavoidable, as explained in the `used_crate.rs` test, but `-Zinstrument-coverage` has new options to back off support for either unused generics, or all unused functions, which avoids the notice, at the cost of less coverage of unused functions. Fixes: #82875 Invalid LLVM coverage data produced with crate brotli_decompressor Fixed by disabling the LLVM function attribute that forces inlining, if `-Z instrument-coverage` is enabled. This attribute is applied to Rust functions with `#[inline(always)], and in some cases, the forced inlining breaks coverage instrumentation and reports.
2021-03-15 23:32:45 +00:00
}
}
impl<'tcx> CoverageInfoBuilderMethods<'tcx> for Builder<'_, '_, 'tcx> {
fn init_coverage(&mut self, instance: Instance<'tcx>) {
let Some(function_coverage_info) =
self.tcx.instance_mir(instance.def).function_coverage_info.as_deref()
else {
return;
};
// If there are no MC/DC bitmaps to set up, return immediately.
if function_coverage_info.mcdc_bitmap_bits == 0 {
return;
}
let fn_name = self.get_pgo_func_name_var(instance);
let hash = self.const_u64(function_coverage_info.function_source_hash);
let bitmap_bits = self.const_u32(function_coverage_info.mcdc_bitmap_bits as u32);
self.mcdc_parameters(fn_name, hash, bitmap_bits);
// Create pointers named `mcdc.addr.{i}` to stack-allocated condition bitmaps.
let mut cond_bitmaps = vec![];
for i in 0..function_coverage_info.mcdc_num_condition_bitmaps {
// MC/DC intrinsics will perform loads/stores that use the ABI default
// alignment for i32, so our variable declaration should match.
let align = self.tcx.data_layout.i32_align.abi;
let cond_bitmap = self.alloca(Size::from_bytes(4), align);
llvm::set_value_name(cond_bitmap, format!("mcdc.addr.{i}").as_bytes());
self.store(self.const_i32(0), cond_bitmap, align);
cond_bitmaps.push(cond_bitmap);
}
self.coverage_cx().mcdc_condition_bitmap_map.borrow_mut().insert(instance, cond_bitmaps);
}
#[instrument(level = "debug", skip(self))]
fn add_coverage(&mut self, instance: Instance<'tcx>, kind: &CoverageKind) {
// Our caller should have already taken care of inlining subtleties,
// so we can assume that counter/expression IDs in this coverage
// statement are meaningful for the given instance.
//
// (Either the statement was not inlined and directly belongs to this
// instance, or it was inlined *from* this instance.)
let bx = self;
let Some(function_coverage_info) =
bx.tcx.instance_mir(instance.def).function_coverage_info.as_deref()
else {
debug!("function has a coverage statement but no coverage info");
return;
};
// FIXME(#132395): Unwrapping `coverage_cx` here has led to ICEs in the
// wild, so keep this early-return until we understand why.
let mut coverage_map = match bx.coverage_cx {
Some(ref cx) => cx.function_coverage_map.borrow_mut(),
None => return,
};
let func_coverage = coverage_map
.entry(instance)
.or_insert_with(|| FunctionCoverageCollector::new(instance, function_coverage_info));
match *kind {
CoverageKind::SpanMarker | CoverageKind::BlockMarker { .. } => unreachable!(
"marker statement {kind:?} should have been removed by CleanupPostBorrowck"
),
CoverageKind::CounterIncrement { id } => {
func_coverage.mark_counter_id_seen(id);
// We need to explicitly drop the `RefMut` before calling into
// `instrprof_increment`, as that needs an exclusive borrow.
drop(coverage_map);
// The number of counters passed to `llvm.instrprof.increment` might
// be smaller than the number originally inserted by the instrumentor,
// if some high-numbered counters were removed by MIR optimizations.
// If so, LLVM's profiler runtime will use fewer physical counters.
let num_counters =
bx.tcx().coverage_ids_info(instance.def).max_counter_id.as_u32() + 1;
assert!(
num_counters as usize <= function_coverage_info.num_counters,
"num_counters disagreement: query says {num_counters} but function info only has {}",
function_coverage_info.num_counters
);
let fn_name = bx.get_pgo_func_name_var(instance);
let hash = bx.const_u64(function_coverage_info.function_source_hash);
let num_counters = bx.const_u32(num_counters);
let index = bx.const_u32(id.as_u32());
debug!(
"codegen intrinsic instrprof.increment(fn_name={:?}, hash={:?}, num_counters={:?}, index={:?})",
fn_name, hash, num_counters, index,
);
bx.instrprof_increment(fn_name, hash, num_counters, index);
}
CoverageKind::ExpressionUsed { id } => {
func_coverage.mark_expression_id_seen(id);
}
CoverageKind::CondBitmapUpdate { index, decision_depth } => {
drop(coverage_map);
let cond_bitmap = bx
.coverage_cx()
.try_get_mcdc_condition_bitmap(&instance, decision_depth)
.expect("mcdc cond bitmap should have been allocated for updating");
let cond_index = bx.const_i32(index as i32);
bx.mcdc_condbitmap_update(cond_index, cond_bitmap);
}
CoverageKind::TestVectorBitmapUpdate { bitmap_idx, decision_depth } => {
drop(coverage_map);
let cond_bitmap = bx.coverage_cx()
.try_get_mcdc_condition_bitmap(&instance, decision_depth)
.expect("mcdc cond bitmap should have been allocated for merging into the global bitmap");
assert!(
bitmap_idx as usize <= function_coverage_info.mcdc_bitmap_bits,
"bitmap index of the decision out of range"
);
let fn_name = bx.get_pgo_func_name_var(instance);
let hash = bx.const_u64(function_coverage_info.function_source_hash);
let bitmap_index = bx.const_u32(bitmap_idx);
bx.mcdc_tvbitmap_update(fn_name, hash, bitmap_index, cond_bitmap);
bx.mcdc_condbitmap_reset(cond_bitmap);
}
}
}
}