RPCS3/llvm-mirror
1
0
Fork 0
mirror of https://github.com/RPCS3/llvm-mirror.git synced 2024-11-24 19:52:54 +01:00
Code Issues Projects Releases Wiki Activity
afd51e50cf
llvm-mirror/test/DebugInfo/missing-abstract-variable.ll

183 lines
7.4 KiB
LLVM
Raw Normal View History

DebugInfo: Partial revert r209984 due to more cases where abstract DbgVariables do not have associated DIEs. Along with a test case to demonstrate that due to inlining order there are cases where abstract variable DIEs are not constructed since the abstract subprogram was built due to a previous inlining that optimized away those variables. This produces incorrect debug info (the 'missing' abstract variable causes the inlined instance of that variable to be emitted with a full description (name, line, file) rather than referencing the abstract origin), but this commit at least ensures that it doesn't crash... llvm-svn: 210143
2014-06-04 03:30:59 +02:00
; REQUIRES: object-emission
; RUN: %llc_dwarf -O0 -filetype=obj < %s | llvm-dwarfdump -debug-dump=info - | FileCheck %s
XFAIL: test/DebugInfo/missing-abstract-variable.ll on mips and ppc64 due to an inlined parameter that goes missing. llvm-svn: 210200
2014-06-04 19:05:23 +02:00
; The formal parameter 'b' for Function 'x' when inlined within 'a' is lost on
; mips and powerpc64 (and on x86_64 at at least -O2). Presumably this is a
; SelectionDAG issue (do mips/powerpc64 use FastISel?).
XFAIL: test/DebugInfo/missing-abstract-variable.ll on s390x as well llvm-svn: 210264
2014-06-05 16:20:54 +02:00
; XFAIL: mips, powerpc64, s390x
XFAIL: test/DebugInfo/missing-abstract-variable.ll on mips and ppc64 due to an inlined parameter that goes missing. llvm-svn: 210200
2014-06-04 19:05:23 +02:00
DebugInfo: Partial revert r209984 due to more cases where abstract DbgVariables do not have associated DIEs. Along with a test case to demonstrate that due to inlining order there are cases where abstract variable DIEs are not constructed since the abstract subprogram was built due to a previous inlining that optimized away those variables. This produces incorrect debug info (the 'missing' abstract variable causes the inlined instance of that variable to be emitted with a full description (name, line, file) rather than referencing the abstract origin), but this commit at least ensures that it doesn't crash... llvm-svn: 210143
2014-06-04 03:30:59 +02:00
; Build from the following source with clang -O2.
; The important details are that 'x's abstract definition is first built during
; the definition of 'b', where the parameter to 'x' is constant and so 'x's 's'
; variable is optimized away. No abstract definition DIE for 's' is constructed.
; Then, during 'a' emission, the abstract DbgVariable for 's' is created, but
; the abstract DIE isn't (since the abstract definition for 'b' is already
; built). This results in 's' inlined in 'a' being emitted with its name, line,
; file there, rather than referencing an abstract definition.
; extern int t;
;
; void f(int);
;
; inline void x(bool b) {
; if (b) {
; int s = t;
; f(s);
; }
; f(0);
; }
;
; void b() {
; x(false);
; }
;
; void a(bool u) {
; x(u);
; }
[dwarfdump] Print the name for referenced specification of abstract_origin DIEs. Reviewers: dblaikie, samsonov, echristo, aprantl Subscribers: llvm-commits Differential Revision: http://reviews.llvm.org/D5466 llvm-svn: 219099
2014-10-06 05:36:31 +02:00
; CHECK: DW_TAG_subprogram
DebugInfo: Partial revert r209984 due to more cases where abstract DbgVariables do not have associated DIEs. Along with a test case to demonstrate that due to inlining order there are cases where abstract variable DIEs are not constructed since the abstract subprogram was built due to a previous inlining that optimized away those variables. This produces incorrect debug info (the 'missing' abstract variable causes the inlined instance of that variable to be emitted with a full description (name, line, file) rather than referencing the abstract origin), but this commit at least ensures that it doesn't crash... llvm-svn: 210143
2014-06-04 03:30:59 +02:00
; CHECK-NOT: DW_TAG
; CHECK: DW_AT_name {{.*}} "x"
; CHECK-NOT: {{DW_TAG|NULL}}
[dwarfdump] Resolve also variable specifications/abstract_origins. DW_AT_specification and DW_AT_abstract_origin resolving was only performed on subroutine DIEs because it used the getSubroutineName method. Introduce a more generic getName() and use it to dump the reference attributes. Testcases have been updated to check the printed names instead of the offsets except when the name could be ambiguous. Reviewers: dblaikie, samsonov Subscribers: llvm-commits Differential Revision: http://reviews.llvm.org/D5625 llvm-svn: 219506
2014-10-10 17:51:02 +02:00
; CHECK: DW_TAG_formal_parameter
DebugInfo: Partial revert r209984 due to more cases where abstract DbgVariables do not have associated DIEs. Along with a test case to demonstrate that due to inlining order there are cases where abstract variable DIEs are not constructed since the abstract subprogram was built due to a previous inlining that optimized away those variables. This produces incorrect debug info (the 'missing' abstract variable causes the inlined instance of that variable to be emitted with a full description (name, line, file) rather than referencing the abstract origin), but this commit at least ensures that it doesn't crash... llvm-svn: 210143
2014-06-04 03:30:59 +02:00
; CHECK-NOT: DW_TAG
; CHECK: DW_AT_name {{.*}} "b"
DebugInfo: Reapply r209984 (reverted in r210143), asserting that abstract DbgVariables have DIEs. Abstract variables within abstract scopes that are entirely optimized away in their first inlining are omitted because their scope is not present so the variable is never created. Instead, we should ensure the scope is created so the variable can be added, even if it's been optimized away in its first inlining. This fixes the incorrect debug info in missing-abstract-variable.ll (added in r210143) and passes an asserts self-hosting build, so hopefully there's not more of these issues left behind... *fingers crossed*. llvm-svn: 210221
2014-06-05 01:50:52 +02:00
; CHECK-NOT: {{DW_TAG|NULL}}
; CHECK: DW_TAG_lexical_block
; CHECK-NOT: {{DW_TAG|NULL}}
[dwarfdump] Resolve also variable specifications/abstract_origins. DW_AT_specification and DW_AT_abstract_origin resolving was only performed on subroutine DIEs because it used the getSubroutineName method. Introduce a more generic getName() and use it to dump the reference attributes. Testcases have been updated to check the printed names instead of the offsets except when the name could be ambiguous. Reviewers: dblaikie, samsonov Subscribers: llvm-commits Differential Revision: http://reviews.llvm.org/D5625 llvm-svn: 219506
2014-10-10 17:51:02 +02:00
; CHECK: DW_TAG_variable
DebugInfo: Reapply r209984 (reverted in r210143), asserting that abstract DbgVariables have DIEs. Abstract variables within abstract scopes that are entirely optimized away in their first inlining are omitted because their scope is not present so the variable is never created. Instead, we should ensure the scope is created so the variable can be added, even if it's been optimized away in its first inlining. This fixes the incorrect debug info in missing-abstract-variable.ll (added in r210143) and passes an asserts self-hosting build, so hopefully there's not more of these issues left behind... *fingers crossed*. llvm-svn: 210221
2014-06-05 01:50:52 +02:00
; CHECK-NOT: DW_TAG
; CHECK: DW_AT_name {{.*}} "s"
DebugInfo: Partial revert r209984 due to more cases where abstract DbgVariables do not have associated DIEs. Along with a test case to demonstrate that due to inlining order there are cases where abstract variable DIEs are not constructed since the abstract subprogram was built due to a previous inlining that optimized away those variables. This produces incorrect debug info (the 'missing' abstract variable causes the inlined instance of that variable to be emitted with a full description (name, line, file) rather than referencing the abstract origin), but this commit at least ensures that it doesn't crash... llvm-svn: 210143
2014-06-04 03:30:59 +02:00
; CHECK: DW_TAG_subprogram
; CHECK-NOT: DW_TAG
; CHECK: DW_AT_name {{.*}} "b"
; CHECK-NOT: {{DW_TAG|NULL}}
; CHECK: DW_TAG_inlined_subroutine
; CHECK-NOT: DW_TAG
[dwarfdump] Print the name for referenced specification of abstract_origin DIEs. Reviewers: dblaikie, samsonov, echristo, aprantl Subscribers: llvm-commits Differential Revision: http://reviews.llvm.org/D5466 llvm-svn: 219099
2014-10-06 05:36:31 +02:00
; CHECK: DW_AT_abstract_origin {{.*}} "_Z1xb"
DebugInfo: Partial revert r209984 due to more cases where abstract DbgVariables do not have associated DIEs. Along with a test case to demonstrate that due to inlining order there are cases where abstract variable DIEs are not constructed since the abstract subprogram was built due to a previous inlining that optimized away those variables. This produces incorrect debug info (the 'missing' abstract variable causes the inlined instance of that variable to be emitted with a full description (name, line, file) rather than referencing the abstract origin), but this commit at least ensures that it doesn't crash... llvm-svn: 210143
2014-06-04 03:30:59 +02:00
; CHECK-NOT: {{DW_TAG|NULL}}
; CHECK: DW_TAG_formal_parameter
; CHECK-NOT: DW_TAG
[dwarfdump] Resolve also variable specifications/abstract_origins. DW_AT_specification and DW_AT_abstract_origin resolving was only performed on subroutine DIEs because it used the getSubroutineName method. Introduce a more generic getName() and use it to dump the reference attributes. Testcases have been updated to check the printed names instead of the offsets except when the name could be ambiguous. Reviewers: dblaikie, samsonov Subscribers: llvm-commits Differential Revision: http://reviews.llvm.org/D5625 llvm-svn: 219506
2014-10-10 17:51:02 +02:00
; CHECK: DW_AT_abstract_origin {{.*}} "b"
DebugInfo: Partial revert r209984 due to more cases where abstract DbgVariables do not have associated DIEs. Along with a test case to demonstrate that due to inlining order there are cases where abstract variable DIEs are not constructed since the abstract subprogram was built due to a previous inlining that optimized away those variables. This produces incorrect debug info (the 'missing' abstract variable causes the inlined instance of that variable to be emitted with a full description (name, line, file) rather than referencing the abstract origin), but this commit at least ensures that it doesn't crash... llvm-svn: 210143
2014-06-04 03:30:59 +02:00
; Notice 'x's local variable 's' is missing. Not necessarily a bug here,
; since it's been optimized entirely away and it should be described in
; abstract subprogram.
; CHECK-NOT: DW_TAG
; CHECK: NULL
; CHECK-NOT: DW_TAG
; CHECK: NULL
; CHECK: DW_TAG_subprogram
; CHECK-NOT: DW_TAG
; CHECK: DW_AT_name {{.*}} "a"
; CHECK-NOT: {{DW_TAG|NULL}}
; CHECK: DW_TAG_formal_parameter
; CHECK-NOT: {{DW_TAG|NULL}}
; CHECK: DW_TAG_inlined_subroutine
; CHECK-NOT: DW_TAG
[dwarfdump] Print the name for referenced specification of abstract_origin DIEs. Reviewers: dblaikie, samsonov, echristo, aprantl Subscribers: llvm-commits Differential Revision: http://reviews.llvm.org/D5466 llvm-svn: 219099
2014-10-06 05:36:31 +02:00
; CHECK: DW_AT_abstract_origin {{.*}} "_Z1xb"
DebugInfo: Partial revert r209984 due to more cases where abstract DbgVariables do not have associated DIEs. Along with a test case to demonstrate that due to inlining order there are cases where abstract variable DIEs are not constructed since the abstract subprogram was built due to a previous inlining that optimized away those variables. This produces incorrect debug info (the 'missing' abstract variable causes the inlined instance of that variable to be emitted with a full description (name, line, file) rather than referencing the abstract origin), but this commit at least ensures that it doesn't crash... llvm-svn: 210143
2014-06-04 03:30:59 +02:00
; CHECK-NOT: {{DW_TAG|NULL}}
DebugInfo: Reapply r209984 (reverted in r210143), asserting that abstract DbgVariables have DIEs. Abstract variables within abstract scopes that are entirely optimized away in their first inlining are omitted because their scope is not present so the variable is never created. Instead, we should ensure the scope is created so the variable can be added, even if it's been optimized away in its first inlining. This fixes the incorrect debug info in missing-abstract-variable.ll (added in r210143) and passes an asserts self-hosting build, so hopefully there's not more of these issues left behind... *fingers crossed*. llvm-svn: 210221
2014-06-05 01:50:52 +02:00
; FIXME: This formal parameter goes missing at least at -O2 (& on
; mips/powerpc), maybe before that. Perhaps SelectionDAG is to blame (and
; fastisel succeeds).
DebugInfo: Partial revert r209984 due to more cases where abstract DbgVariables do not have associated DIEs. Along with a test case to demonstrate that due to inlining order there are cases where abstract variable DIEs are not constructed since the abstract subprogram was built due to a previous inlining that optimized away those variables. This produces incorrect debug info (the 'missing' abstract variable causes the inlined instance of that variable to be emitted with a full description (name, line, file) rather than referencing the abstract origin), but this commit at least ensures that it doesn't crash... llvm-svn: 210143
2014-06-04 03:30:59 +02:00
; CHECK: DW_TAG_formal_parameter
; CHECK-NOT: DW_TAG
[dwarfdump] Resolve also variable specifications/abstract_origins. DW_AT_specification and DW_AT_abstract_origin resolving was only performed on subroutine DIEs because it used the getSubroutineName method. Introduce a more generic getName() and use it to dump the reference attributes. Testcases have been updated to check the printed names instead of the offsets except when the name could be ambiguous. Reviewers: dblaikie, samsonov Subscribers: llvm-commits Differential Revision: http://reviews.llvm.org/D5625 llvm-svn: 219506
2014-10-10 17:51:02 +02:00
; CHECK: DW_AT_abstract_origin {{.*}} "b"
DebugInfo: Partial revert r209984 due to more cases where abstract DbgVariables do not have associated DIEs. Along with a test case to demonstrate that due to inlining order there are cases where abstract variable DIEs are not constructed since the abstract subprogram was built due to a previous inlining that optimized away those variables. This produces incorrect debug info (the 'missing' abstract variable causes the inlined instance of that variable to be emitted with a full description (name, line, file) rather than referencing the abstract origin), but this commit at least ensures that it doesn't crash... llvm-svn: 210143
2014-06-04 03:30:59 +02:00
; CHECK-NOT: {{DW_TAG|NULL}}
; CHECK: DW_TAG_lexical_block
; CHECK-NOT: {{DW_TAG|NULL}}
DebugInfo: Elide lexical scopes which only contain other (inline or lexical) scopes. DW_TAG_lexical_scopes inform debuggers about the instruction range for which a given variable (or imported declaration/module/etc) is valid. If the scope doesn't itself contain any such entities, it's a waste of space and should be omitted. We were correctly doing this for entirely empty leaves, but not for intermediate nodes. Reduces total (not just debug sections) .o file size for a bootstrap -gmlt LLVM by 22% and bootstrap -gmlt clang executable by 13%. The wins for a full -g build will be less as a % (and in absolute terms), but should still be substantial - with some of that win being fewer relocations, thus more substantiall reducing link times than fewer bytes alone would have. llvm-svn: 216861
2014-08-31 23:26:22 +02:00
; CHECK: DW_TAG_variable
DebugInfo: Partial revert r209984 due to more cases where abstract DbgVariables do not have associated DIEs. Along with a test case to demonstrate that due to inlining order there are cases where abstract variable DIEs are not constructed since the abstract subprogram was built due to a previous inlining that optimized away those variables. This produces incorrect debug info (the 'missing' abstract variable causes the inlined instance of that variable to be emitted with a full description (name, line, file) rather than referencing the abstract origin), but this commit at least ensures that it doesn't crash... llvm-svn: 210143
2014-06-04 03:30:59 +02:00
; CHECK-NOT: DW_TAG
[dwarfdump] Resolve also variable specifications/abstract_origins. DW_AT_specification and DW_AT_abstract_origin resolving was only performed on subroutine DIEs because it used the getSubroutineName method. Introduce a more generic getName() and use it to dump the reference attributes. Testcases have been updated to check the printed names instead of the offsets except when the name could be ambiguous. Reviewers: dblaikie, samsonov Subscribers: llvm-commits Differential Revision: http://reviews.llvm.org/D5625 llvm-svn: 219506
2014-10-10 17:51:02 +02:00
; CHECK: DW_AT_abstract_origin {{.*}} "s"
DebugInfo: Partial revert r209984 due to more cases where abstract DbgVariables do not have associated DIEs. Along with a test case to demonstrate that due to inlining order there are cases where abstract variable DIEs are not constructed since the abstract subprogram was built due to a previous inlining that optimized away those variables. This produces incorrect debug info (the 'missing' abstract variable causes the inlined instance of that variable to be emitted with a full description (name, line, file) rather than referencing the abstract origin), but this commit at least ensures that it doesn't crash... llvm-svn: 210143
2014-06-04 03:30:59 +02:00
@t = external global i32
; Function Attrs: uwtable
define void @_Z1bv() #0 {
entry:
IR: Make metadata typeless in assembly Now that `Metadata` is typeless, reflect that in the assembly. These are the matching assembly changes for the metadata/value split in r223802. - Only use the `metadata` type when referencing metadata from a call intrinsic -- i.e., only when it's used as a `Value`. - Stop pretending that `ValueAsMetadata` is wrapped in an `MDNode` when referencing it from call intrinsics. So, assembly like this: define @foo(i32 %v) { call void @llvm.foo(metadata !{i32 %v}, metadata !0) call void @llvm.foo(metadata !{i32 7}, metadata !0) call void @llvm.foo(metadata !1, metadata !0) call void @llvm.foo(metadata !3, metadata !0) call void @llvm.foo(metadata !{metadata !3}, metadata !0) ret void, !bar !2 } !0 = metadata !{metadata !2} !1 = metadata !{i32* @global} !2 = metadata !{metadata !3} !3 = metadata !{} turns into this: define @foo(i32 %v) { call void @llvm.foo(metadata i32 %v, metadata !0) call void @llvm.foo(metadata i32 7, metadata !0) call void @llvm.foo(metadata i32* @global, metadata !0) call void @llvm.foo(metadata !3, metadata !0) call void @llvm.foo(metadata !{!3}, metadata !0) ret void, !bar !2 } !0 = !{!2} !1 = !{i32* @global} !2 = !{!3} !3 = !{} I wrote an upgrade script that handled almost all of the tests in llvm and many of the tests in cfe (even handling many `CHECK` lines). I've attached it (or will attach it in a moment if you're speedy) to PR21532 to help everyone update their out-of-tree testcases. This is part of PR21532. llvm-svn: 224257
2014-12-15 20:07:53 +01:00
tail call void @llvm.dbg.value(metadata i1 false, i64 0, metadata !25, metadata !{!"0x102"}), !dbg !27
DebugInfo: Partial revert r209984 due to more cases where abstract DbgVariables do not have associated DIEs. Along with a test case to demonstrate that due to inlining order there are cases where abstract variable DIEs are not constructed since the abstract subprogram was built due to a previous inlining that optimized away those variables. This produces incorrect debug info (the 'missing' abstract variable causes the inlined instance of that variable to be emitted with a full description (name, line, file) rather than referencing the abstract origin), but this commit at least ensures that it doesn't crash... llvm-svn: 210143
2014-06-04 03:30:59 +02:00
tail call void @_Z1fi(i32 0), !dbg !28
ret void, !dbg !29
}
; Function Attrs: uwtable
define void @_Z1ab(i1 zeroext %u) #0 {
entry:
IR: Make metadata typeless in assembly Now that `Metadata` is typeless, reflect that in the assembly. These are the matching assembly changes for the metadata/value split in r223802. - Only use the `metadata` type when referencing metadata from a call intrinsic -- i.e., only when it's used as a `Value`. - Stop pretending that `ValueAsMetadata` is wrapped in an `MDNode` when referencing it from call intrinsics. So, assembly like this: define @foo(i32 %v) { call void @llvm.foo(metadata !{i32 %v}, metadata !0) call void @llvm.foo(metadata !{i32 7}, metadata !0) call void @llvm.foo(metadata !1, metadata !0) call void @llvm.foo(metadata !3, metadata !0) call void @llvm.foo(metadata !{metadata !3}, metadata !0) ret void, !bar !2 } !0 = metadata !{metadata !2} !1 = metadata !{i32* @global} !2 = metadata !{metadata !3} !3 = metadata !{} turns into this: define @foo(i32 %v) { call void @llvm.foo(metadata i32 %v, metadata !0) call void @llvm.foo(metadata i32 7, metadata !0) call void @llvm.foo(metadata i32* @global, metadata !0) call void @llvm.foo(metadata !3, metadata !0) call void @llvm.foo(metadata !{!3}, metadata !0) ret void, !bar !2 } !0 = !{!2} !1 = !{i32* @global} !2 = !{!3} !3 = !{} I wrote an upgrade script that handled almost all of the tests in llvm and many of the tests in cfe (even handling many `CHECK` lines). I've attached it (or will attach it in a moment if you're speedy) to PR21532 to help everyone update their out-of-tree testcases. This is part of PR21532. llvm-svn: 224257
2014-12-15 20:07:53 +01:00
tail call void @llvm.dbg.value(metadata i1 %u, i64 0, metadata !13, metadata !{!"0x102"}), !dbg !30
tail call void @llvm.dbg.value(metadata i1 %u, i64 0, metadata !31, metadata !{!"0x102"}), !dbg !33
DebugInfo: Partial revert r209984 due to more cases where abstract DbgVariables do not have associated DIEs. Along with a test case to demonstrate that due to inlining order there are cases where abstract variable DIEs are not constructed since the abstract subprogram was built due to a previous inlining that optimized away those variables. This produces incorrect debug info (the 'missing' abstract variable causes the inlined instance of that variable to be emitted with a full description (name, line, file) rather than referencing the abstract origin), but this commit at least ensures that it doesn't crash... llvm-svn: 210143
2014-06-04 03:30:59 +02:00
br i1 %u, label %if.then.i, label %_Z1xb.exit, !dbg !34
if.then.i: ; preds = %entry
%0 = load i32* @t, align 4, !dbg !35, !tbaa !36
IR: Make metadata typeless in assembly Now that `Metadata` is typeless, reflect that in the assembly. These are the matching assembly changes for the metadata/value split in r223802. - Only use the `metadata` type when referencing metadata from a call intrinsic -- i.e., only when it's used as a `Value`. - Stop pretending that `ValueAsMetadata` is wrapped in an `MDNode` when referencing it from call intrinsics. So, assembly like this: define @foo(i32 %v) { call void @llvm.foo(metadata !{i32 %v}, metadata !0) call void @llvm.foo(metadata !{i32 7}, metadata !0) call void @llvm.foo(metadata !1, metadata !0) call void @llvm.foo(metadata !3, metadata !0) call void @llvm.foo(metadata !{metadata !3}, metadata !0) ret void, !bar !2 } !0 = metadata !{metadata !2} !1 = metadata !{i32* @global} !2 = metadata !{metadata !3} !3 = metadata !{} turns into this: define @foo(i32 %v) { call void @llvm.foo(metadata i32 %v, metadata !0) call void @llvm.foo(metadata i32 7, metadata !0) call void @llvm.foo(metadata i32* @global, metadata !0) call void @llvm.foo(metadata !3, metadata !0) call void @llvm.foo(metadata !{!3}, metadata !0) ret void, !bar !2 } !0 = !{!2} !1 = !{i32* @global} !2 = !{!3} !3 = !{} I wrote an upgrade script that handled almost all of the tests in llvm and many of the tests in cfe (even handling many `CHECK` lines). I've attached it (or will attach it in a moment if you're speedy) to PR21532 to help everyone update their out-of-tree testcases. This is part of PR21532. llvm-svn: 224257
2014-12-15 20:07:53 +01:00
tail call void @llvm.dbg.value(metadata i32 %0, i64 0, metadata !40, metadata !{!"0x102"}), !dbg !35
DebugInfo: Partial revert r209984 due to more cases where abstract DbgVariables do not have associated DIEs. Along with a test case to demonstrate that due to inlining order there are cases where abstract variable DIEs are not constructed since the abstract subprogram was built due to a previous inlining that optimized away those variables. This produces incorrect debug info (the 'missing' abstract variable causes the inlined instance of that variable to be emitted with a full description (name, line, file) rather than referencing the abstract origin), but this commit at least ensures that it doesn't crash... llvm-svn: 210143
2014-06-04 03:30:59 +02:00
tail call void @_Z1fi(i32 %0), !dbg !41
br label %_Z1xb.exit, !dbg !42
_Z1xb.exit: ; preds = %entry, %if.then.i
tail call void @_Z1fi(i32 0), !dbg !43
ret void, !dbg !44
}
declare void @_Z1fi(i32) #1
; Function Attrs: nounwind readnone
Move the complex address expression out of DIVariable and into an extra argument of the llvm.dbg.declare/llvm.dbg.value intrinsics. Previously, DIVariable was a variable-length field that has an optional reference to a Metadata array consisting of a variable number of complex address expressions. In the case of OpPiece expressions this is wasting a lot of storage in IR, because when an aggregate type is, e.g., SROA'd into all of its n individual members, the IR will contain n copies of the DIVariable, all alike, only differing in the complex address reference at the end. By making the complex address into an extra argument of the dbg.value/dbg.declare intrinsics, all of the pieces can reference the same variable and the complex address expressions can be uniqued across the CU, too. Down the road, this will allow us to move other flags, such as "indirection" out of the DIVariable, too. The new intrinsics look like this: declare void @llvm.dbg.declare(metadata %storage, metadata %var, metadata %expr) declare void @llvm.dbg.value(metadata %storage, i64 %offset, metadata %var, metadata %expr) This patch adds a new LLVM-local tag to DIExpressions, so we can detect and pretty-print DIExpression metadata nodes. What this patch doesn't do: This patch does not touch the "Indirect" field in DIVariable; but moving that into the expression would be a natural next step. http://reviews.llvm.org/D4919 rdar://problem/17994491 Thanks to dblaikie and dexonsmith for reviewing this patch! Note: I accidentally committed a bogus older version of this patch previously. llvm-svn: 218787
2014-10-01 20:55:02 +02:00
declare void @llvm.dbg.value(metadata, i64, metadata, metadata) #2
DebugInfo: Partial revert r209984 due to more cases where abstract DbgVariables do not have associated DIEs. Along with a test case to demonstrate that due to inlining order there are cases where abstract variable DIEs are not constructed since the abstract subprogram was built due to a previous inlining that optimized away those variables. This produces incorrect debug info (the 'missing' abstract variable causes the inlined instance of that variable to be emitted with a full description (name, line, file) rather than referencing the abstract origin), but this commit at least ensures that it doesn't crash... llvm-svn: 210143
2014-06-04 03:30:59 +02:00
attributes #0 = { uwtable "less-precise-fpmad"="false" "no-frame-pointer-elim"="false" "no-infs-fp-math"="false" "no-nans-fp-math"="false" "stack-protector-buffer-size"="8" "unsafe-fp-math"="false" "use-soft-float"="false" }
attributes #1 = { "less-precise-fpmad"="false" "no-frame-pointer-elim"="false" "no-infs-fp-math"="false" "no-nans-fp-math"="false" "stack-protector-buffer-size"="8" "unsafe-fp-math"="false" "use-soft-float"="false" }
attributes #2 = { nounwind readnone }
!llvm.dbg.cu = !{!0}
!llvm.module.flags = !{!21, !22}
!llvm.ident = !{!23}
IR: Make metadata typeless in assembly Now that `Metadata` is typeless, reflect that in the assembly. These are the matching assembly changes for the metadata/value split in r223802. - Only use the `metadata` type when referencing metadata from a call intrinsic -- i.e., only when it's used as a `Value`. - Stop pretending that `ValueAsMetadata` is wrapped in an `MDNode` when referencing it from call intrinsics. So, assembly like this: define @foo(i32 %v) { call void @llvm.foo(metadata !{i32 %v}, metadata !0) call void @llvm.foo(metadata !{i32 7}, metadata !0) call void @llvm.foo(metadata !1, metadata !0) call void @llvm.foo(metadata !3, metadata !0) call void @llvm.foo(metadata !{metadata !3}, metadata !0) ret void, !bar !2 } !0 = metadata !{metadata !2} !1 = metadata !{i32* @global} !2 = metadata !{metadata !3} !3 = metadata !{} turns into this: define @foo(i32 %v) { call void @llvm.foo(metadata i32 %v, metadata !0) call void @llvm.foo(metadata i32 7, metadata !0) call void @llvm.foo(metadata i32* @global, metadata !0) call void @llvm.foo(metadata !3, metadata !0) call void @llvm.foo(metadata !{!3}, metadata !0) ret void, !bar !2 } !0 = !{!2} !1 = !{i32* @global} !2 = !{!3} !3 = !{} I wrote an upgrade script that handled almost all of the tests in llvm and many of the tests in cfe (even handling many `CHECK` lines). I've attached it (or will attach it in a moment if you're speedy) to PR21532 to help everyone update their out-of-tree testcases. This is part of PR21532. llvm-svn: 224257
2014-12-15 20:07:53 +01:00
!0 = !{!"0x11\004\00clang version 3.5.0 \001\00\000\00\001", !1, !2, !2, !3, !2, !2} ; [ DW_TAG_compile_unit ] [/tmp/dbginfo/missing-abstract-variables.cc] [DW_LANG_C_plus_plus]
!1 = !{!"missing-abstract-variables.cc", !"/tmp/dbginfo"}
!2 = !{}
!3 = !{!4, !8, !14}
!4 = !{!"0x2e\00b\00b\00_Z1bv\0013\000\001\000\006\00256\001\0013", !1, !5, !6, null, void ()* @_Z1bv, null, null, !2} ; [ DW_TAG_subprogram ] [line 13] [def] [b]
!5 = !{!"0x29", !1} ; [ DW_TAG_file_type ] [/tmp/dbginfo/missing-abstract-variables.cc]
!6 = !{!"0x15\00\000\000\000\000\000\000", i32 0, null, null, !7, null, null, null} ; [ DW_TAG_subroutine_type ] [line 0, size 0, align 0, offset 0] [from ]
!7 = !{null}
!8 = !{!"0x2e\00a\00a\00_Z1ab\0017\000\001\000\006\00256\001\0017", !1, !5, !9, null, void (i1)* @_Z1ab, null, null, !12} ; [ DW_TAG_subprogram ] [line 17] [def] [a]
!9 = !{!"0x15\00\000\000\000\000\000\000", i32 0, null, null, !10, null, null, null} ; [ DW_TAG_subroutine_type ] [line 0, size 0, align 0, offset 0] [from ]
!10 = !{null, !11}
!11 = !{!"0x24\00bool\000\008\008\000\000\002", null, null} ; [ DW_TAG_base_type ] [bool] [line 0, size 8, align 8, offset 0, enc DW_ATE_boolean]
!12 = !{!13}
!13 = !{!"0x101\00u\0016777233\000", !8, !5, !11} ; [ DW_TAG_arg_variable ] [u] [line 17]
!14 = !{!"0x2e\00x\00x\00_Z1xb\005\000\001\000\006\00256\001\005", !1, !5, !9, null, null, null, null, !15} ; [ DW_TAG_subprogram ] [line 5] [def] [x]
!15 = !{!16, !17}
!16 = !{!"0x101\00b\0016777221\000", !14, !5, !11} ; [ DW_TAG_arg_variable ] [b] [line 5]
!17 = !{!"0x100\00s\007\000", !18, !5, !20} ; [ DW_TAG_auto_variable ] [s] [line 7]
!18 = !{!"0xb\006\000\000", !1, !19} ; [ DW_TAG_lexical_block ] [/tmp/dbginfo/missing-abstract-variables.cc]
!19 = !{!"0xb\006\000\000", !1, !14} ; [ DW_TAG_lexical_block ] [/tmp/dbginfo/missing-abstract-variables.cc]
!20 = !{!"0x24\00int\000\0032\0032\000\000\005", null, null} ; [ DW_TAG_base_type ] [int] [line 0, size 32, align 32, offset 0, enc DW_ATE_signed]
!21 = !{i32 2, !"Dwarf Version", i32 4}
!22 = !{i32 2, !"Debug Info Version", i32 2}
!23 = !{!"clang version 3.5.0 "}
!24 = !{i1 false}
!25 = !{!"0x101\00b\0016777221\000", !14, !5, !11, !26} ; [ DW_TAG_arg_variable ] [b] [line 5]
!26 = !{i32 14, i32 0, !4, null}
!27 = !{i32 5, i32 0, !14, !26}
!28 = !{i32 10, i32 0, !14, !26}
!29 = !{i32 15, i32 0, !4, null}
!30 = !{i32 17, i32 0, !8, null}
!31 = !{!"0x101\00b\0016777221\000", !14, !5, !11, !32} ; [ DW_TAG_arg_variable ] [b] [line 5]
!32 = !{i32 18, i32 0, !8, null}
!33 = !{i32 5, i32 0, !14, !32}
!34 = !{i32 6, i32 0, !19, !32}
!35 = !{i32 7, i32 0, !18, !32}
!36 = !{!37, !37, i64 0}
!37 = !{!"int", !38, i64 0}
!38 = !{!"omnipotent char", !39, i64 0}
!39 = !{!"Simple C/C++ TBAA"}
!40 = !{!"0x100\00s\007\000", !18, !5, !20, !32} ; [ DW_TAG_auto_variable ] [s] [line 7]
!41 = !{i32 8, i32 0, !18, !32}
!42 = !{i32 9, i32 0, !18, !32}
!43 = !{i32 10, i32 0, !14, !32}
!44 = !{i32 19, i32 0, !8, null}
Reference in New Issue Copy Permalink