1
Fork 0
rust/compiler/rustc_codegen_llvm/src/debuginfo/create_scope_map.rs

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

131 lines
4.9 KiB
Rust
Raw Normal View History

use rustc_codegen_ssa::mir::debuginfo::{DebugScope, FunctionDebugContext};
use rustc_codegen_ssa::traits::*;
use rustc_index::Idx;
use rustc_index::bit_set::BitSet;
2020-03-29 16:41:09 +02:00
use rustc_middle::mir::{Body, SourceScope};
use rustc_middle::ty::layout::FnAbiOf;
use rustc_middle::ty::{self, Instance};
use rustc_session::config::DebugInfo;
2015-04-29 18:14:37 +12:00
use super::metadata::file_metadata;
use super::utils::DIB;
use crate::common::CodegenCx;
2019-02-18 03:58:58 +09:00
use crate::llvm;
use crate::llvm::debuginfo::{DILocation, DIScope};
2019-02-08 14:53:55 +01:00
/// Produces DIScope DIEs for each MIR Scope which has variables defined in it.
// FIXME(eddyb) almost all of this should be in `rustc_codegen_ssa::mir::debuginfo`.
pub(crate) fn compute_mir_scopes<'ll, 'tcx>(
cx: &CodegenCx<'ll, 'tcx>,
instance: Instance<'tcx>,
mir: &Body<'tcx>,
debug_context: &mut FunctionDebugContext<'tcx, &'ll DIScope, &'ll DILocation>,
) {
2022-03-25 03:50:41 -07:00
// Find all scopes with variables defined in them.
let variables = if cx.sess().opts.debuginfo == DebugInfo::Full {
let mut vars = BitSet::new_empty(mir.source_scopes.len());
// FIXME(eddyb) take into account that arguments always have debuginfo,
// irrespective of their name (assuming full debuginfo is enabled).
// NOTE(eddyb) actually, on second thought, those are always in the
// function scope, which always exists.
for var_debug_info in &mir.var_debug_info {
2022-03-25 03:50:41 -07:00
vars.insert(var_debug_info.source_info.scope);
}
2022-03-25 03:50:41 -07:00
Some(vars)
} else {
// Nothing to emit, of course.
None
};
2022-04-26 21:00:19 +08:00
let mut instantiated = BitSet::new_empty(mir.source_scopes.len());
// Instantiate all scopes.
for idx in 0..mir.source_scopes.len() {
let scope = SourceScope::new(idx);
2022-04-26 21:00:19 +08:00
make_mir_scope(cx, instance, mir, &variables, debug_context, &mut instantiated, scope);
}
2022-04-26 21:00:19 +08:00
assert!(instantiated.count() == mir.source_scopes.len());
}
fn make_mir_scope<'ll, 'tcx>(
cx: &CodegenCx<'ll, 'tcx>,
instance: Instance<'tcx>,
mir: &Body<'tcx>,
2022-03-25 03:50:41 -07:00
variables: &Option<BitSet<SourceScope>>,
debug_context: &mut FunctionDebugContext<'tcx, &'ll DIScope, &'ll DILocation>,
2022-04-26 21:00:19 +08:00
instantiated: &mut BitSet<SourceScope>,
scope: SourceScope,
) {
2022-04-26 21:00:19 +08:00
if instantiated.contains(scope) {
return;
}
let scope_data = &mir.source_scopes[scope];
let parent_scope = if let Some(parent) = scope_data.parent_scope {
2022-04-26 21:00:19 +08:00
make_mir_scope(cx, instance, mir, variables, debug_context, instantiated, parent);
debug_context.scopes[parent]
} else {
// The root is the function itself.
let file = cx.sess().source_map().lookup_source_file(mir.span.lo());
debug_context.scopes[scope] = DebugScope {
file_start_pos: file.start_pos,
file_end_pos: file.end_position(),
2022-04-26 21:00:19 +08:00
..debug_context.scopes[scope]
};
2022-04-26 21:00:19 +08:00
instantiated.insert(scope);
return;
};
2022-03-25 03:50:41 -07:00
if let Some(vars) = variables
&& !vars.contains(scope)
&& scope_data.inlined.is_none()
{
// Do not create a DIScope if there are no variables defined in this
// MIR `SourceScope`, and it's not `inlined`, to avoid debuginfo bloat.
debug_context.scopes[scope] = parent_scope;
2022-04-26 21:00:19 +08:00
instantiated.insert(scope);
return;
}
let loc = cx.lookup_debug_loc(scope_data.span.lo());
let file_metadata = file_metadata(cx, &loc.file);
Rework MIR inlining debuginfo so function parameters show up in debuggers. Line numbers of multiply-inlined functions were fixed in #114643 by using a single DISubprogram. That, however, triggered assertions because parameters weren't deduplicated. The "solution" to that in #115417 was to insert a DILexicalScope below the DISubprogram and parent all of the parameters to that scope. That fixed the assertion, but debuggers (including gdb and lldb) don't recognize variables that are not parented to the subprogram itself as parameters, even if they are emitted with DW_TAG_formal_parameter. Consider the program: use std::env; fn square(n: i32) -> i32 { n * n } fn square_no_inline(n: i32) -> i32 { n * n } fn main() { let x = square(env::vars().count() as i32); let y = square_no_inline(env::vars().count() as i32); println!("{x} == {y}"); } When making a release build with debug=2 and rustc 1.82.0-nightly (8b3870784 2024-08-07) (gdb) r Starting program: /ephemeral/tmp/target/release/tmp [Thread debugging using libthread_db enabled] Using host libthread_db library "/lib/x86_64-linux-gnu/libthread_db.so.1". Breakpoint 1, tmp::square () at src/main.rs:5 5 n * n (gdb) info args No arguments. (gdb) info locals n = 31 (gdb) c Continuing. Breakpoint 2, tmp::square_no_inline (n=31) at src/main.rs:10 10 n * n (gdb) info args n = 31 (gdb) info locals No locals. This issue is particularly annoying because it removes arguments from stack traces. The DWARF for the inlined function looks like this: < 2><0x00002132 GOFF=0x00002132> DW_TAG_subprogram DW_AT_linkage_name _ZN3tmp6square17hc507052ff3d2a488E DW_AT_name square DW_AT_decl_file 0x0000000f /ephemeral/tmp/src/main.rs DW_AT_decl_line 0x00000004 DW_AT_type 0x00001a56<.debug_info+0x00001a56> DW_AT_inline DW_INL_inlined < 3><0x00002142 GOFF=0x00002142> DW_TAG_lexical_block < 4><0x00002143 GOFF=0x00002143> DW_TAG_formal_parameter DW_AT_name n DW_AT_decl_file 0x0000000f /ephemeral/tmp/src/main.rs DW_AT_decl_line 0x00000004 DW_AT_type 0x00001a56<.debug_info+0x00001a56> < 4><0x0000214e GOFF=0x0000214e> DW_TAG_null < 3><0x0000214f GOFF=0x0000214f> DW_TAG_null That DW_TAG_lexical_block inhibits every debugger I've tested from recognizing 'n' as a parameter. This patch removes the additional lexical scope. Parameters can be easily deduplicated by a tuple of their scope and the argument index, at the trivial cost of taking a Hash + Eq bound on DIScope.
2024-08-12 19:20:00 -07:00
let dbg_scope = match scope_data.inlined {
Some((callee, _)) => {
// FIXME(eddyb) this would be `self.monomorphize(&callee)`
// if this is moved to `rustc_codegen_ssa::mir::debuginfo`.
2023-09-25 15:46:38 +02:00
let callee = cx.tcx.instantiate_and_normalize_erasing_regions(
instance.args,
ty::ParamEnv::reveal_all(),
2023-05-29 13:46:10 +02:00
ty::EarlyBinder::bind(callee),
);
debug_context.inlined_function_scopes.entry(callee).or_insert_with(|| {
let callee_fn_abi = cx.fn_abi_of_instance(callee, ty::List::empty());
cx.dbg_scope_fn(callee, callee_fn_abi, None)
})
}
Rework MIR inlining debuginfo so function parameters show up in debuggers. Line numbers of multiply-inlined functions were fixed in #114643 by using a single DISubprogram. That, however, triggered assertions because parameters weren't deduplicated. The "solution" to that in #115417 was to insert a DILexicalScope below the DISubprogram and parent all of the parameters to that scope. That fixed the assertion, but debuggers (including gdb and lldb) don't recognize variables that are not parented to the subprogram itself as parameters, even if they are emitted with DW_TAG_formal_parameter. Consider the program: use std::env; fn square(n: i32) -> i32 { n * n } fn square_no_inline(n: i32) -> i32 { n * n } fn main() { let x = square(env::vars().count() as i32); let y = square_no_inline(env::vars().count() as i32); println!("{x} == {y}"); } When making a release build with debug=2 and rustc 1.82.0-nightly (8b3870784 2024-08-07) (gdb) r Starting program: /ephemeral/tmp/target/release/tmp [Thread debugging using libthread_db enabled] Using host libthread_db library "/lib/x86_64-linux-gnu/libthread_db.so.1". Breakpoint 1, tmp::square () at src/main.rs:5 5 n * n (gdb) info args No arguments. (gdb) info locals n = 31 (gdb) c Continuing. Breakpoint 2, tmp::square_no_inline (n=31) at src/main.rs:10 10 n * n (gdb) info args n = 31 (gdb) info locals No locals. This issue is particularly annoying because it removes arguments from stack traces. The DWARF for the inlined function looks like this: < 2><0x00002132 GOFF=0x00002132> DW_TAG_subprogram DW_AT_linkage_name _ZN3tmp6square17hc507052ff3d2a488E DW_AT_name square DW_AT_decl_file 0x0000000f /ephemeral/tmp/src/main.rs DW_AT_decl_line 0x00000004 DW_AT_type 0x00001a56<.debug_info+0x00001a56> DW_AT_inline DW_INL_inlined < 3><0x00002142 GOFF=0x00002142> DW_TAG_lexical_block < 4><0x00002143 GOFF=0x00002143> DW_TAG_formal_parameter DW_AT_name n DW_AT_decl_file 0x0000000f /ephemeral/tmp/src/main.rs DW_AT_decl_line 0x00000004 DW_AT_type 0x00001a56<.debug_info+0x00001a56> < 4><0x0000214e GOFF=0x0000214e> DW_TAG_null < 3><0x0000214f GOFF=0x0000214f> DW_TAG_null That DW_TAG_lexical_block inhibits every debugger I've tested from recognizing 'n' as a parameter. This patch removes the additional lexical scope. Parameters can be easily deduplicated by a tuple of their scope and the argument index, at the trivial cost of taking a Hash + Eq bound on DIScope.
2024-08-12 19:20:00 -07:00
None => unsafe {
llvm::LLVMRustDIBuilderCreateLexicalBlock(
DIB(cx),
parent_scope.dbg_scope,
file_metadata,
loc.line,
loc.col,
)
},
};
let inlined_at = scope_data.inlined.map(|(_, callsite_span)| {
// FIXME(eddyb) this doesn't account for the macro-related
// `Span` fixups that `rustc_codegen_ssa::mir::debuginfo` does.
let callsite_scope = parent_scope.adjust_dbg_scope_for_span(cx, callsite_span);
cx.dbg_loc(callsite_scope, parent_scope.inlined_at, callsite_span)
});
debug_context.scopes[scope] = DebugScope {
2022-04-26 21:00:19 +08:00
dbg_scope,
inlined_at: inlined_at.or(parent_scope.inlined_at),
file_start_pos: loc.file.start_pos,
file_end_pos: loc.file.end_position(),
};
2022-04-26 21:00:19 +08:00
instantiated.insert(scope);
}