<html>
<head>
<base href="https://llvm.org/bugs/" />
</head>
<body><table border="1" cellspacing="0" cellpadding="8">
<tr>
<th>Bug ID</th>
<td><a class="bz_bug_link
bz_status_NEW "
title="NEW --- - ThinLTO + switch table + O2 cause undefined symbols"
href="https://llvm.org/bugs/show_bug.cgi?id=31052">31052</a>
</td>
</tr>
<tr>
<th>Summary</th>
<td>ThinLTO + switch table + O2 cause undefined symbols
</td>
</tr>
<tr>
<th>Product</th>
<td>new-bugs
</td>
</tr>
<tr>
<th>Version</th>
<td>trunk
</td>
</tr>
<tr>
<th>Hardware</th>
<td>PC
</td>
</tr>
<tr>
<th>OS</th>
<td>Linux
</td>
</tr>
<tr>
<th>Status</th>
<td>NEW
</td>
</tr>
<tr>
<th>Severity</th>
<td>normal
</td>
</tr>
<tr>
<th>Priority</th>
<td>P
</td>
</tr>
<tr>
<th>Component</th>
<td>new bugs
</td>
</tr>
<tr>
<th>Assignee</th>
<td>unassignedbugs@nondot.org
</td>
</tr>
<tr>
<th>Reporter</th>
<td>krasin@google.com
</td>
</tr>
<tr>
<th>CC</th>
<td>llvm-bugs@lists.llvm.org
</td>
</tr>
<tr>
<th>Classification</th>
<td>Unclassified
</td>
</tr></table>
<p>
<div>
<pre>Created <span class=""><a href="attachment.cgi?id=17607" name="attach_17607" title="reproducer">attachment 17607</a> <a href="attachment.cgi?id=17607&action=edit" title="reproducer">[details]</a></span>
reproducer
Consider the following program that consists of two files (the archive with the
sources is attached):
main.cc:
#include <stdio.h>
typedef int (*CallFunc) ();
CallFunc GetCallN(int i, int j);
int main(void) {
printf("%d\n", GetCallN(1, 1)());
printf("%d\n", GetCallN(2, 1)());
}
call.cc:
typedef int (*CallFunc) ();
static int CallX() { return 0; }
static int CallY() { return 1; }
const CallFunc Funcs0[] = {CallX, CallY, CallY, CallY};
const CallFunc Funcs1[] = {CallY, CallX, CallY, CallY};
const CallFunc Funcs2[] = {CallY, CallY, CallX, CallY};
CallFunc GetCallN(int i, int j) {
switch (i) {
case 0:
return Funcs0[j];
case 1:
return Funcs1[j];
case 2:
return Funcs2[j];
default:
break;
}
return 0;
}
Let's build it with ThinLTO and LLD:
clang++ -o main main.cc call.cc -O2 -fuse-ld=lld -flto=thin
The following error will be displayed:
/usr/local/google/home/krasin/src/llvm.org/release-build/bin/ld.lld: error:
/tmp/main-fa77ab.o0 (function main): undefined symbol 'CallX()
(.llvm.5BF168B0)'
/usr/local/google/home/krasin/src/llvm.org/release-build/bin/ld.lld: error:
/tmp/main-fa77ab.o0 (function main): undefined symbol 'CallY()
(.llvm.5BF168B0)'
clang-3.9: error: linker command failed with exit code 1 (use -v to see
invocation)
- It works correctly with FullLTO:
$ clang++ -o main main.cc call.cc -O2 -fuse-ld=lld -flto && ./main
0
1
- It works with a regular build (with lld, as well as ld).
- It works with FullLTO + Gold
- It fails with ThinLTO + Gold:
$ clang++ -o main main.cc call.cc -O2 -fuse-ld=gold -flto=thin
/tmp/lto-llvm-22f26f.o:/tmp/main-155d01.o:function main: error: undefined
reference to '_ZL5CallXv.llvm.5BF168B0'
/tmp/lto-llvm-22f26f.o:/tmp/main-155d01.o:function main: error: undefined
reference to '_ZL5CallYv.llvm.5BF168B0'
clang-3.9: error: linker command failed with exit code 1 (use -v to see
invocation)
Looking at the bitcode, -O2 causes the switch statement be converted into a
switch table:
@_ZL6Funcs0 = internal unnamed_addr constant [4 x i32 ()*] [i32 ()*
@_ZL5CallXv, i32 ()* @_ZL5CallYv, i32 ()* @_ZL5CallYv, i32 ()* @_ZL5CallYv],
align 16
@_ZL6Funcs1 = internal unnamed_addr constant [4 x i32 ()*] [i32 ()*
@_ZL5CallYv, i32 ()* @_ZL5CallXv, i32 ()* @_ZL5CallYv, i32 ()* @_ZL5CallYv],
align 16
@_ZL6Funcs2 = internal unnamed_addr constant [4 x i32 ()*] [i32 ()*
@_ZL5CallYv, i32 ()* @_ZL5CallYv, i32 ()* @_ZL5CallXv, i32 ()* @_ZL5CallYv],
align 16
@switch.table = private unnamed_addr constant [3 x [4 x i32 ()*]*] [[4 x i32
()*]* @_ZL6Funcs0, [4 x i32 ()*]* @_ZL6Funcs1, [4 x i32 ()*]* @_ZL6Funcs2]
; Function Attrs: norecurse nounwind readonly uwtable
define i32 ()* @_Z8GetCallNii(i32 %i, i32 %j) local_unnamed_addr #0 {
entry:
%0 = icmp ult i32 %i, 3
br i1 %0, label %switch.lookup, label %return
switch.lookup: ; preds = %entry
%1 = sext i32 %i to i64
%switch.gep = getelementptr inbounds [3 x [4 x i32 ()*]*], [3 x [4 x i32
()*]*]* @switch.table, i64 0, i64 %1
%switch.load = load [4 x i32 ()*]*, [4 x i32 ()*]** %switch.gep, align 8
%idxprom5 = sext i32 %j to i64
%arrayidx6 = getelementptr inbounds [4 x i32 ()*], [4 x i32 ()*]*
%switch.load, i64 0, i64 %idxprom5
%2 = load i32 ()*, i32 ()** %arrayidx6, align 8, !tbaa !1
br label %return
return: ; preds = %switch.lookup,
%entry
%retval.0 = phi i32 ()* [ null, %entry ], [ %2, %switch.lookup ]
ret i32 ()* %retval.0
}
Whenever I change the code so that the switch table is not generated, it would
compile with ThinLTO just fine.</pre>
</div>
</p>
<hr>
<span>You are receiving this mail because:</span>
<ul>
<li>You are on the CC list for the bug.</li>
</ul>
</body>
</html>