[llvm] Add support for the .base64 directive (fixes #165499) (PR #165549)

[Ben Kallus via llvm-commits]

https://github.com/kenballus updated https://github.com/llvm/llvm-project/pull/165549

>From 31a16a8da6d393a5f5683108f6dcaf7d64740e07 Mon Sep 17 00:00:00 2001
From: Ben Kallus <benjamin.p.kallus.gr at dartmouth.edu>
Date: Wed, 29 Oct 2025 08:14:38 -0400
Subject: [PATCH 1/2] Add support for the .base64 directive

---
 llvm/lib/MC/MCParser/AsmParser.cpp | 21 +++++++++++++++++++++
 1 file changed, 21 insertions(+)

diff --git a/llvm/lib/MC/MCParser/AsmParser.cpp b/llvm/lib/MC/MCParser/AsmParser.cpp
index dd1bc2be5feb4..54bb1451a5a73 100644
--- a/llvm/lib/MC/MCParser/AsmParser.cpp
+++ b/llvm/lib/MC/MCParser/AsmParser.cpp
@@ -46,6 +46,7 @@
 #include "llvm/MC/MCSymbolMachO.h"
 #include "llvm/MC/MCTargetOptions.h"
 #include "llvm/MC/MCValue.h"
+#include "llvm/Support/Base64.h"
 #include "llvm/Support/Casting.h"
 #include "llvm/Support/CommandLine.h"
 #include "llvm/Support/ErrorHandling.h"
@@ -530,6 +531,7 @@ class AsmParser : public MCAsmParser {
     DK_LTO_SET_CONDITIONAL,
     DK_CFI_MTE_TAGGED_FRAME,
     DK_MEMTAG,
+    DK_BASE64,
     DK_END
   };
 
@@ -552,6 +554,7 @@ class AsmParser : public MCAsmParser {
 
   // ".ascii", ".asciz", ".string"
   bool parseDirectiveAscii(StringRef IDVal, bool ZeroTerminated);
+  bool parseDirectiveBase64();                  // ".base64"
   bool parseDirectiveReloc(SMLoc DirectiveLoc); // ".reloc"
   bool parseDirectiveValue(StringRef IDVal,
                            unsigned Size);       // ".byte", ".long", ...
@@ -1953,6 +1956,8 @@ bool AsmParser::parseStatement(ParseStatementInfo &Info,
     case DK_ASCIZ:
     case DK_STRING:
       return parseDirectiveAscii(IDVal, true);
+    case DK_BASE64:
+      return parseDirectiveBase64();
     case DK_BYTE:
     case DK_DC_B:
       return parseDirectiveValue(IDVal, 1);
@@ -3076,6 +3081,21 @@ bool AsmParser::parseDirectiveAscii(StringRef IDVal, bool ZeroTerminated) {
   return parseMany(parseOp);
 }
 
+/// parseDirectiveBase64:
+//    ::= .base64 "string"
+bool AsmParser::parseDirectiveBase64() {
+  std::vector<char> Decoded;
+
+  std::string str;
+
+  if (parseEscapedString(str) || str.empty() || decodeBase64(str, Decoded)) {
+    return true;
+  }
+
+  getStreamer().emitBytes(std::string(Decoded.begin(), Decoded.end()));
+  return false;
+}
+
 /// parseDirectiveReloc
 ///  ::= .reloc expression , identifier [ , expression ]
 bool AsmParser::parseDirectiveReloc(SMLoc DirectiveLoc) {
@@ -5345,6 +5365,7 @@ void AsmParser::initializeDirectiveKindMap() {
   DirectiveKindMap[".asciz"] = DK_ASCIZ;
   DirectiveKindMap[".string"] = DK_STRING;
   DirectiveKindMap[".byte"] = DK_BYTE;
+  DirectiveKindMap[".base64"] = DK_BASE64;
   DirectiveKindMap[".short"] = DK_SHORT;
   DirectiveKindMap[".value"] = DK_VALUE;
   DirectiveKindMap[".2byte"] = DK_2BYTE;

>From 2001eb2876ba1a09e746cb02a53dfc9538520c21 Mon Sep 17 00:00:00 2001
From: Ben Kallus <benjamin.p.kallus.gr at dartmouth.edu>
Date: Wed, 29 Oct 2025 13:33:48 -0400
Subject: [PATCH 2/2] Add test, and rework implementation to no longer accept
 escape strings

---
 llvm/lib/MC/MCParser/AsmParser.cpp        | 11 +++++++----
 llvm/test/MC/AsmParser/directive_base64.s | 12 ++++++++++++
 2 files changed, 19 insertions(+), 4 deletions(-)
 create mode 100644 llvm/test/MC/AsmParser/directive_base64.s

diff --git a/llvm/lib/MC/MCParser/AsmParser.cpp b/llvm/lib/MC/MCParser/AsmParser.cpp
index 54bb1451a5a73..19abc952a22b3 100644
--- a/llvm/lib/MC/MCParser/AsmParser.cpp
+++ b/llvm/lib/MC/MCParser/AsmParser.cpp
@@ -3084,15 +3084,18 @@ bool AsmParser::parseDirectiveAscii(StringRef IDVal, bool ZeroTerminated) {
 /// parseDirectiveBase64:
 //    ::= .base64 "string"
 bool AsmParser::parseDirectiveBase64() {
-  std::vector<char> Decoded;
-
-  std::string str;
+  if (checkForValidSection() || check(getTok().isNot(AsmToken::String), "expected string")) {
+    return true;
+  }
 
-  if (parseEscapedString(str) || str.empty() || decodeBase64(str, Decoded)) {
+  std::vector<char> Decoded;
+  std::string const str = getTok().getStringContents().str();
+  if (str.empty() || decodeBase64(str, Decoded)) {
     return true;
   }
 
   getStreamer().emitBytes(std::string(Decoded.begin(), Decoded.end()));
+  Lex();
   return false;
 }
 
diff --git a/llvm/test/MC/AsmParser/directive_base64.s b/llvm/test/MC/AsmParser/directive_base64.s
new file mode 100644
index 0000000000000..7e5440ea65b68
--- /dev/null
+++ b/llvm/test/MC/AsmParser/directive_base64.s
@@ -0,0 +1,12 @@
+# RUN: llvm-mc -triple i386-unknown-unknown %s | FileCheck %s
+
+        .data
+# CHECK: TEST0:
+# CHECK: .byte 0
+TEST0:
+        .base64 "AA=="
+        
+# CHECK: TEST1:
+# CHECK: .ascii "abcxyz"
+TEST1:
+        .base64 "YWJjeHl6"