[lld] Fix output section offset and contents when linker script uses memory region and data commands

[owenpshaw]

Summary

Advance the memory region offset when handling a linker script data command such as BYTE or LONG. Failure to advance the offset results in corrupted output with overlapping sections.

Update tests to check for this combination of both a) memory regions and b) data commands.

Fixes https://bugs.llvm.org/show_bug.cgi?id=35565

[grimar]

You need to add reviewers when posting the patches. It is at least Rui as code owner of LLD.
Logic of patch looks correct for me. I'll add him.

ELF/LinkerScript.cpp

696

You do not need bracers around single line.

test/ELF/linkerscript/data-commands.s

47

MEMORY {

51

SECTIONS {

[grimar]

"Group Reviewers lld" (https://reviews.llvm.org/project/members/2/)
you used is something new for me honestly, but looking at members I suspect it
is just an old group used for old LLD (different elf linker existed before this one, new LLD).

[owenpshaw]

Thanks for the help!

Updated the patch with suggested changes to braces.

[grimar]

I have no other comments, this looks good to me. You need approval from reviewers to commit.

[owenpshaw]

Closing. As already noted in mailing list, Rafael committed this change as r321418.

ELF/LinkerScript.cpp

if (auto *Cmd = dyn_cast<SymbolAssignment>(Base)) {
       assignSymbol(Cmd, true);
       continue;
     }
 
     // Handle BYTE(), SHORT(), LONG(), or QUAD().
     if (auto *Cmd = dyn_cast<ByteCommand>(Base)) {
       Cmd->Offset = Dot - Ctx->OutSec->Addr;
       Dot += Cmd->Size;
+      if (Ctx->MemRegion)
+        Ctx->MemRegionOffset[Ctx->MemRegion] += Cmd->Size;
       Ctx->OutSec->Size = Dot - Ctx->OutSec->Addr;
       continue;
     }
 
     // Handle ASSERT().
     if (auto *Cmd = dyn_cast<AssertCommand>(Base)) {
       Cmd->Expression();
       continue;

test/ELF/linkerscript/data-commands.s

 # RUN: llvm-objdump -s %t2 | FileCheck %s --check-prefix=BE
 # BE:      Contents of section .foo:
 # BE-NEXT:   ff11ff11 22ff1122 3344ff11 22334455
 # BE-NEXT:   667788
 # BE-NEXT: Contents of section .bar:
 # BE-NEXT:   ff12ff11 22ff1122 3346ff11 22334455
 # BE-NEXT:   667788
 
+# RUN: echo "MEMORY {                \
+# RUN:     rom (rwx) : ORIGIN = 0x00, LENGTH = 2K \
+# RUN:  }                            \
+# RUN:  SECTIONS {                   \
+# RUN:    .foo : {                   \
+# RUN:      *(.foo.1)                \
+# RUN:      BYTE(0x11)               \
+# RUN:      *(.foo.2)                \
+# RUN:      SHORT(0x1122)            \
+# RUN:      *(.foo.3)                \
+# RUN:      LONG(0x11223344)         \
+# RUN:      *(.foo.4)                \
+# RUN:      QUAD(0x1122334455667788) \
+# RUN:    } > rom                    \
+# RUN:    .bar : {                   \
+# RUN:      *(.bar.1)                \
+# RUN:      BYTE(a + 1)              \
+# RUN:      *(.bar.2)                \
+# RUN:      SHORT(b)                 \
+# RUN:      *(.bar.3)                \
+# RUN:      LONG(c + 2)              \
+# RUN:      *(.bar.4)                \
+# RUN:      QUAD(d)                  \
+# RUN:    } > rom                    \
+# RUN:  }" > %t-memory.script
+# RUN: ld.lld -o %t-memory %t.o --script %t-memory.script
+# RUN: llvm-objdump -s %t-memory | FileCheck %s --check-prefix=MEM
+
+# MEM:      Contents of section .foo:
+# MEM-NEXT:   0000 ff11ff22 11ff4433 2211ff88 77665544
+# MEM-NEXT:   0010 332211
+
+# MEM:      Contents of section .bar:
+# MEM-NEXT:   0013 ff12ff22 11ff4633 2211ff88 77665544
+# MEM-NEXT:   0023 332211
+
 .global a
 a = 0x11
 
 .global b
 b = 0x1122
 
 .global c
 c = 0x11223344