Fix Bug 11740: Turn off -g when there are directive file debug info

[zhizhouy]

Summary

For the bug 11740:
bug 11740

If there is already debug info in the assembly file, and user hope to use -g option for
compiling, we think we should not directly report an error.

According to what GNU assembler did, it just reused the debug info in the assembly file,
and turned off the DEBUG_TYPE option so that there will be no new debug info emitted
by assembler. This fix is just as what GNU assembler did.

The concern is the situation that there are two .text sections in the assembly file, one with
debug info and the other one without. Currently with this fix, the assembler will no longer
generate any debug info for the second .text section. And this is what GNU assembler
exactly did for this situation. So I think this still make some sense.

[dblaikie]

What about other pieces of the debug info (when generating debug info for assembly we generate labels at least, maybe other things?) - will they be appropriately disabled by the presence of existing debug info in the assembly?

test/MC/AsmParser/directive_file-2.s

2

What's this second run line for?

[zhizhouy]

I think the turning off of the GenDwarfForAssembly at the parsing stage
will lead to emitting no new debug info by the assembler at emitting stage.
So this still works for other debug info.

test/MC/AsmParser/directive_file-2.s

2

Just to run twice with different command line options for stability. The same as what
is in the test/MC/AsmParser/directive_file.s

[zhizhouy]

test/MC/AsmParser/directive_file-2.s

2

In fact I am not quite familiar with the other test case (directive_file.s),
it seems it has these two lines to face two different situations.

So I just reused it since this test case should also work as the same as
the former one.

[dblaikie]

test/MC/AsmParser/directive_file-2.s

2

I'd rather not cargo cult the test in - if we don't know why it's there, could you leave it out?

3

Might be worth describing what the behavior/issue is that's being tested here.

[zhizhouy]

Updated the test file, removed the useless test command,
and added some description about the test case.

[dblaikie]

Sounds good - thanks

[zhizhouy]

Hi David,

Since I currently don't have access to submit, would you please submit
this patch for me?

Thanks.

lib/MC/MCParser/AsmParser.cpp

bool AsmParser::parseDirectiveFile(SMLoc DirectiveLoc) {
   }
 
   if (getLexer().isNot(AsmToken::EndOfStatement))
     return TokError("unexpected token in '.file' directive");
 
   if (FileNumber == -1)
     getStreamer().EmitFileDirective(Filename);
   else {
+    // If there is -g option as well as debug info from directive file,
+    // we turn off -g option, directly use the existing debug info instead.
     if (getContext().getGenDwarfForAssembly())
-      Error(DirectiveLoc,
-            "input can't have .file dwarf directives when -g is "
-            "used to generate dwarf debug info for assembly code");
-
-    if (getStreamer().EmitDwarfFileDirective(FileNumber, Directory, Filename) ==
+      getContext().setGenDwarfForAssembly(false);
+    else if (getStreamer().EmitDwarfFileDirective(FileNumber, Directory, Filename) ==
         0)
       Error(FileNumberLoc, "file number already allocated");
   }
 
   return false;
 }
 
 /// parseDirectiveLine

test/MC/AsmParser/directive_file-2.s

+# RUN: llvm-mc -triple i386-unknown-unknown %s | FileCheck %s
+# RUN: llvm-mc -triple i386-unknown-unknown %s -filetype=null
+// Test for Bug 11740
+
+        .file "hello"
+        .file 1 "world"
+
+// CHECK: .file "hello"
+// CHECK: .file 1 "world"

test/MC/AsmParser/directive_file-errors.s

-// RUN: not llvm-mc -g -triple i386-unknown-unknown %s 2> %t.err | FileCheck %s
-// RUN: FileCheck --check-prefix=CHECK-ERRORS %s < %t.err
-// Test for Bug 11740
-
-        .file "hello"
-        .file 1 "world"
-
-// CHECK: .file "hello"
-// CHECK-ERRORS:6:9: error: input can't have .file dwarf directives when -g is used to generate dwarf debug info for assembly code