[ELF] - Linkerscript: add support for suffixes in numbers.

[grimar]

Summary

Both bfd and gold accept:
foo = 1K;
bar = 1M;
zed = 1H;

And lowercase forms: k, m, h.
Patch adds support for that.

[ruiu]

ELF/LinkerScript.cpp

1235 ↗

(On Diff #70165)

Something's not right with this function. You seems to be mixing two different concepts, base and multiplier in this function, which confused me while reading this code. You cannot use K/M suffixes for hex numbers, so hex numbers should be handled first.

The code should be something like this.

if (Tok.startswith_lower("0x"))
  return Tok.substr(2).getAsInteger(16, Result);
if (Tok.endswith_lower("H"))
  return Tok.drop_back().getAsInteger(16, Result);

int Suffix = 1;
if (Tok.endswith_lower("K")) {
  Suffix = 1024;
  Tok = Tok.drop_back();
} else if (Tok.endswith_lower("M")) {
  Suffix = 1024 * 1024;
  Tok = Tok.drop_back();
}
if (Tok.getAsInteger(10, Result))
  return false;
Result *= Suffix;
return true;

1236 ↗

(On Diff #70165)

Delete this enum and use these numbers directly. They will never change.

[grimar]

• Addressed review comments.

ELF/LinkerScript.cpp

1235 ↗

(On Diff #70165)

Done.

[ruiu]

LGTM with this change.

ELF/LinkerScript.cpp

1326–1331 ↗

(On Diff #70170)

Please try to simplify when you flip the meaning of the return value.

// Tok is a literal number.
uint64_t V;
if (readInteger(Tok, V))
  return [=](uint64_t Dot) { return V; };

// Tok is a symbol name.
if (Tok != "." && !isValidCIdentifier(Tok))
  setError("malformed number: " + Tok);
return [=](uint64_t Dot) { return getSymbolValue(Tok, Dot); };

lld/trunk/ELF/LinkerScript.cpp

uint64_t static getConstant(StringRef S) {
   if (S == "COMMONPAGESIZE")
     return Target->PageSize;
   if (S == "MAXPAGESIZE")
     return Target->MaxPageSize;
   error("unknown constant: " + S);
   return 0;
 }
 
+static bool readInteger(StringRef Tok, uint64_t &Result) {
+  if (Tok.startswith_lower("0x"))
+    return !Tok.substr(2).getAsInteger(16, Result);
+  if (Tok.endswith_lower("H"))
+    return !Tok.drop_back().getAsInteger(16, Result);
+
+  int Suffix = 1;
+  if (Tok.endswith_lower("K")) {
+    Suffix = 1024;
+    Tok = Tok.drop_back();
+  } else if (Tok.endswith_lower("M")) {
+    Suffix = 1024 * 1024;
+    Tok = Tok.drop_back();
+  }
+  if (Tok.getAsInteger(10, Result))
+    return false;
+  Result *= Suffix;
+  return true;
+}
+
 Expr ScriptParser::readPrimary() {
   if (peek() == "(")
     return readParenExpr();
 
   StringRef Tok = next();
 
   // Built-in functions are parsed here.
   // https://sourceware.org/binutils/docs/ld/Builtin-Functions.html.
if (Tok == "SIZEOF") {
     expect("(");
     StringRef Name = next();
     expect(")");
     return [=](uint64_t Dot) { return getSectionSize(Name); };
   }
   if (Tok == "SIZEOF_HEADERS")
     return [=](uint64_t Dot) { return getHeaderSize(); };
 
-  // Parse a symbol name or a number literal.
-  uint64_t V = 0;
-  if (Tok.getAsInteger(0, V)) {
-    if (Tok != "." && !isValidCIdentifier(Tok))
-      setError("malformed number: " + Tok);
-    return [=](uint64_t Dot) { return getSymbolValue(Tok, Dot); };
-  }
-  return [=](uint64_t Dot) { return V; };
-}
+  // Tok is a literal number.
+  uint64_t V;
+  if (readInteger(Tok, V))
+    return [=](uint64_t Dot) { return V; };
+
+  // Tok is a symbol name.
 
 Expr ScriptParser::readTernary(Expr Cond) {
   next();
   Expr L = readExpr();
   expect(":");
   Expr R = readExpr();
   return [=](uint64_t Dot) { return Cond(Dot) ? L(Dot) : R(Dot); };
 }

lld/trunk/test/ELF/linkerscript/linkerscript-numbers.s

+# REQUIRES: x86
+# RUN: llvm-mc -filetype=obj -triple=x86_64-unknown-linux %s -o %t
+# RUN: echo "SECTIONS { \
+# RUN:  . = 1000h;              \
+# RUN:  .hex1 : { *(.hex.1) }   \
+# RUN:  . = 1010H;              \
+# RUN:  .hex2 : { *(.hex.2) }   \
+# RUN:  . = 10k;                \
+# RUN:  .kilo1 : { *(.kilo.1) } \
+# RUN:  . = 11K;                \
+# RUN:  .kilo2 : { *(.kilo.2) } \
+# RUN:  . = 1m;                 \
+# RUN:  .mega1 : { *(.mega.1) } \
+# RUN:  . = 2M;                 \
+# RUN:  .mega2 : { *(.mega.2) } \
+# RUN: }" > %t.script
+# RUN: ld.lld %t --script %t.script -o %t2
+# RUN: llvm-objdump -section-headers %t2 | FileCheck %s
+
+# CHECK:     Sections:
+# CHECK-NEXT: Idx Name          Size      Address        
+# CHECK-NEXT:   0               00000000 0000000000000000
+# CHECK-NEXT:   1 .hex1         00000008 0000000000001000
+# CHECK-NEXT:   2 .hex2         00000008 0000000000001010
+# CHECK-NEXT:   3 .kilo1        00000008 0000000000002800
+# CHECK-NEXT:   4 .kilo2        00000008 0000000000002c00
+# CHECK-NEXT:   5 .mega1        00000008 0000000000100000
+# CHECK-NEXT:   6 .mega2        00000008 0000000000200000
+
+## Mailformed number errors.
+# RUN: echo "SECTIONS { \
+# RUN:  . = 0x11h; \
+# RUN: }" > %t2.script
+# RUN: not ld.lld %t --script %t2.script -o %t3 2>&1 | \
+# RUN:  FileCheck --check-prefix=ERR1 %s
+# ERR1: malformed number: 0x11h
+
+# RUN: echo "SECTIONS { \
+# RUN:  . = 0x11k; \
+# RUN: }" > %t3.script
+# RUN: not ld.lld %t --script %t3.script -o %t4 2>&1 | \
+# RUN:  FileCheck --check-prefix=ERR2 %s
+# ERR2: malformed number: 0x11k
+
+# RUN: echo "SECTIONS { \
+# RUN:  . = 0x11m; \
+# RUN: }" > %t4.script
+# RUN: not ld.lld %t --script %t4.script -o %t5 2>&1 | \
+# RUN:  FileCheck --check-prefix=ERR3 %s
+# ERR3: malformed number: 0x11m
+
+.globl _start
+_start:
+nop
+
+.section .hex.1, "a"
+.quad 0
+
+.section .kilo.1, "a"
+.quad 0
+
+.section .mega.1, "a"
+.quad 0
+
+.section .hex.2, "a"
+.quad 0
+
+.section .kilo.2, "a"
+.quad 0
+
+.section .mega.2, "a"
+.quad 0