Rename human-readable name for DW_LANG_Mips_Assembler

[aprantl]

Summary

The Mips in DW_LANG_Mips_Assembler is a vendor name not an architecture name and in lack of a proper generic DW_LANG_assembler, some assemblers emit DWARF using this tag. Due to a warning I recently introduced users will now be greeted with

This version of LLDB has no plugin for the mipsassem language. Inspection of frame variables will be limited.

What about renaming this to just assembler? Are there any risks in doing so?

rdar://77214764

[JDevlieghere]

LGTM

[clayborg]

If other architectures are emitting this language for assembly files, then this LGTM.

[dim]

Hmm so now on FreeBSD, basically *every* program will cause this warning, as our C startup objects are compiled from .S files, e.g.:

% objdump --dwarf /bin/sleep|grep -B6 "MIPS assembler"
 <0><619>: Abbrev Number: 1 (DW_TAG_compile_unit)
    <61a>   DW_AT_stmt_list   : 0x3f1
    <61e>   DW_AT_ranges      : 0x30
    <622>   DW_AT_name        : /usr/src/lib/csu/amd64/crti.S
    <640>   DW_AT_comp_dir    : /usr/obj/usr/src/amd64.amd64/lib/csu/amd64
    <66b>   DW_AT_producer    : FreeBSD clang version 12.0.1 ([e-mail]:llvm/llvm-project.git llvmorg-12.0.1-0-gfed41342a82f)
    <6ce>   DW_AT_language    : 32769   (MIPS assembler)
--
 <0><de8>: Abbrev Number: 1 (DW_TAG_compile_unit)
    <de9>   DW_AT_stmt_list   : 0x93a
    <ded>   DW_AT_ranges      : 0xb0
    <df1>   DW_AT_name        : /usr/src/lib/csu/amd64/crtn.S
    <e0f>   DW_AT_comp_dir    : /usr/obj/usr/src/amd64.amd64/lib/csu/amd64
    <e3a>   DW_AT_producer    : FreeBSD clang version 12.0.1 ([e-mail]:llvm/llvm-project.git llvmorg-12.0.1-0-gfed41342a82f)
    <e9d>   DW_AT_language    : 32769   (MIPS assembler)

% lldb /bin/sleep 10
(lldb) target create "/bin/sleep"
Current executable set to '/bin/sleep' (x86_64).
(lldb) settings set -- target.run-args  "10"
(lldb) r
Process 11879 launched: '/bin/sleep' (x86_64)
This version of LLDB has no plugin for the mipsassem language. Inspection of frame variables will be limited.
Process 11879 stopped
* thread #1, name = 'sleep', stop reason = signal SIGSTOP
    frame #0: 0x00000008011a610a libc.so.7`__sys_nanosleep at _nanosleep.S:4

This can't have been the intention of this commit? As far as I can see, there *is* no plugin for plain assembler?

[clayborg]

Everyone uses this for any assembly in their code. Should we get rid of this warning? I find it is useless and would like to see it go.

[dim]

In D101406#2948152, @clayborg wrote:

Everyone uses this for any assembly in their code. Should we get rid of this warning? I find it is useless and would like to see it go.

Well, for other languages it might be a helpful warning, but not for the special case of "mips" assembler. :)

[aprantl]

This can't have been the intention of this commit? As far as I can see, there *is* no plugin for plain assembler?

This commit changes the warning from

This version of LLDB has no plugin for the mipsassem language.

to

This version of LLDB has no plugin for the language "assembler".

Yes, there is no plugin for the language assembler. But that hasn't chnaged with this commit.

[dim]

In D101406#2950731, @aprantl wrote:

This can't have been the intention of this commit? As far as I can see, there *is* no plugin for plain assembler?

...

Yes, there is no plugin for the language assembler. But that hasn't chnaged with this commit.

Sorry, that was rG220c17ffd4e1b127bcc02b25980b7934184ee1da ("Print a warning when stopped in a frame LLDB has no plugin for"), which is probably a good idea in general, but maybe there could be a specific exception for the DW_LANG_Mips_Assembler case then? It doesn't make sense to warn about that. :)

[aprantl]

It doesn't make sense to warn about that. :)

Why not? LLDB doesn't have support for assembler as a source language, no assembler expression evaluator, etc..

I don't have much experience debugging assembler sources with LLDB, so I"m not sure about the expectations.

lldb/source/Target/Language.cpp

struct language_name_pair language_names[] = {
     {"c11", eLanguageTypeC11},
     {"swift", eLanguageTypeSwift},
     {"julia", eLanguageTypeJulia},
     {"dylan", eLanguageTypeDylan},
     {"c++14", eLanguageTypeC_plus_plus_14},
     {"fortran03", eLanguageTypeFortran03},
     {"fortran08", eLanguageTypeFortran08},
     // Vendor Extensions
-    {"mipsassem", eLanguageTypeMipsAssembler},
+    {"assembler", eLanguageTypeMipsAssembler},
     {"renderscript", eLanguageTypeExtRenderScript},
     // Now synonyms, in arbitrary order
     {"objc", eLanguageTypeObjC},
     {"objc++", eLanguageTypeObjC_plus_plus},
     {"pascal", eLanguageTypePascal83}};
 
 static uint32_t num_languages =
     sizeof(language_names) / sizeof(struct language_name_pair);

lldb/source/Target/Process.cpp

void Process::PrintWarningUnsupportedLanguage(const SymbolContext &sc) {
   LanguageType language = sc.GetLanguage();
   if (language == eLanguageTypeUnknown)
     return;
   LanguageSet plugins =
       PluginManager::GetAllTypeSystemSupportedLanguagesForTypes();
   if (!plugins[language]) {
     PrintWarning(Process::Warnings::eWarningsUnsupportedLanguage,
                  sc.module_sp.get(),
-                 "This version of LLDB has no plugin for the %s language. "
+                 "This version of LLDB has no plugin for the language \"%s\". "
                  "Inspection of frame variables will be limited.\n",
                  Language::GetNameForLanguageType(language));
   }
 }
 
 bool Process::GetProcessInfo(ProcessInstanceInfo &info) {
   info.Clear();
 

lldb/test/Shell/Process/UnsupportedLanguage.test

 Test warnings.
 REQUIRES: shell
 RUN: %clang_host %S/Inputs/true.c -std=c99 -g -c -S -emit-llvm -o - \
-RUN:   | sed -e 's/DW_LANG_C99/DW_LANG_PLI/g' >%t.ll
+RUN:   | sed -e 's/DW_LANG_C99/DW_LANG_Mips_Assembler/g' >%t.ll
 RUN: %clang_host %t.ll -g -o %t.exe
 RUN: %lldb -o "b main" -o r -o q -b %t.exe | FileCheck %s
 
-CHECK: This version of LLDB has no plugin for the pli language
+CHECK: This version of LLDB has no plugin for the language "assembler"