[ELF] -M/-Map: fix VMA/LMA/Size columns of symbol assignments when address/size>=2**32

[MaskRay]

Summary

SymbolAssignment::addr stores the location counter. The type should be
uint64_t instead of unsigned. The upper half of the address space is
commonly used by operating system kernels.

Similarly, SymbolAssignment::size should be an uint64_t. A kernel linker
script can move the location counter from 0 to the upper half of the
address space.

[grimar]

The code change LG, few comments about the test.

lld/test/ELF/map-file-kernel.s

1 ↗

(On Diff #254966)

map-file-kernel.s probably doesn't describe the intention of the test? Not sure why there is "kernel" mentioned.

7 ↗

(On Diff #254966)

The second check line looks broken/shifted because of the huge size it seems, isn't?
Are we going to fix it or was your intention just to document this behavior?
(If we want to fix this, I guess it can be done in another patch, but I'd probably add a comment regarding this line then).

11 ↗

(On Diff #254966)

Perhaps would be good to add the comment daying what we wanted to test here.
Without knowledge about the code change it is unclear what is tested perhaps.

[MaskRay]

Address comments

[MaskRay]

lld/test/ELF/map-file-kernel.s

1 ↗

(On Diff #254966)

Renamed to map-file-64bit.s

7 ↗

(On Diff #254966)

Just to document the behavior. This is very uncommon. In an arbitrary Linux vmlinux.lds, there is only one such line: there is just one far-distance symbol assignment.

[grimar]

LGTM

lld/ELF/LinkerScript.h

struct SymbolAssignment : BaseCommand {
 
   // Holds file name and line number for error reporting.
   std::string location;
 
   // A string representation of this command. We use this for -Map.
   std::string commandString;
 
   // Address of this assignment command.
-  unsigned addr;
+  uint64_t addr;
 
   // Size of this assignment command. This is usually 0, but if
   // you move '.' this may be greater than 0.
-  unsigned size;
+  uint64_t size;
 };
 
 // Linker scripts allow additional constraints to be put on output sections.
 // If an output section is marked as ONLY_IF_RO, the section is created
 // only if its input sections are read-only. Likewise, an output section
 // with ONLY_IF_RW is created if all input sections are RW.
 enum class ConstraintKind { NoConstraint, ReadOnly, ReadWrite };
 

lld/test/ELF/map-file-64bit.s

+# REQUIRES: x86
+## Test how we display the upper half of the address space, which is commonly
+## used by operating system kernels.
+
+# RUN: llvm-mc -filetype=obj -triple=x86_64 /dev/null -o %t.o
+# RUN: ld.lld -M -T %s %t.o -o /dev/null | FileCheck %s --match-full-lines --strict-whitespace
+
+## . = 0xffffffff80000000; has a misaligned Size column.
+#      CHECK:             VMA              LMA     Size Align Out     In      Symbol
+# CHECK-NEXT:               0                0 ffffffff80000000     1 . = 0xffffffff80000000
+# CHECK-NEXT:ffffffff80000000                0   100000     1 . += 0x100000
+# CHECK-NEXT:ffffffff80100000                0        0     1 _text = .
+# CHECK-NEXT:ffffffff80100000 ffffffff80100000        0     4 .text
+
+SECTIONS {
+  . = 0xffffffff80000000;
+  . += 0x100000;
+  _text = .;
+}