[lld][LoongArch] Support the R_LARCH_PCREL20_S2 relocation type

[SixWeining]

Summary

R_LARCH_PCREL20_S2 is a new added relocation type in LoongArch ELF
psABI v2.10 [1] which is not corvered by D138135 except R_LARCH_64_PCREL.

A motivation to support R_LARCH_PCREL20_S2 in lld is to build the
runtime of .NET core (a.k.a CoreCLR) in which strict PC-relative
semantics need to be guaranteed [2]. The normal pcalau12i + addi.d
approach doesn't work because the code will be copied to other places
with different "page" and offsets. To achieve this, we can use pcaddi
with explicit R_LARCH_PCREL20_S2 reloc to address +-2MB PC-relative
range with 4-bytes aligned.

[1]: https://github.com/loongson/la-abi-specs/releases/tag/v2.10
[2]: https://github.com/dotnet/runtime/blob/release/7.0/src/coreclr/vm/loongarch64/asmhelpers.S#L307

[xen0n]

Thanks for the patch, actually I had this one in my local repo for a while but hadn't submitted just yet ;-) The code change is trivially correct and I only have a minor suggestion for the test case. Let's wait for @MaskRay's opinions though...

lld/test/ELF/loongarch-pcrel20-s2.s

13–17

I remember MaskRay once mentioned the desire to keep lld invocations to a minimum? To me this looks like two pcaddi's pointing to two separate sections (one above .text and one below), then only one run for the positive cases would be enough. Let's see what he thinks.

[MaskRay]

lld/test/ELF/loongarch-pcrel20-s2.s

13–17

Yes, we should reduce the number of invocations. If a linker script can test multiple edge cases, we should use that.

[SixWeining]

Remove redundant lld invocations in tests.

[SixWeining]

rebase

[xen0n]

LGTM (LoongArch-wise); @MaskRay please take another look (we'll wait for your comment before landing of course), thanks!

lld/ELF/Arch/LoongArch.cpp

RelExpr LoongArch::getRelExpr(const RelType type, const Symbol &s,
   case R_LARCH_SUB16:
   case R_LARCH_SUB32:
   case R_LARCH_SUB64:
     // The LoongArch add/sub relocs behave like the RISCV counterparts; reuse
     // the RelExpr to avoid code duplication.
     return R_RISCV_ADD;
   case R_LARCH_32_PCREL:
   case R_LARCH_64_PCREL:
+  case R_LARCH_PCREL20_S2:
     return R_PC;
   case R_LARCH_B16:
   case R_LARCH_B21:
   case R_LARCH_B26:
     return R_PLT_PC;
   case R_LARCH_GOT_PC_HI20:
   case R_LARCH_GOT64_PC_LO20:
   case R_LARCH_GOT64_PC_HI12:
case R_LARCH_TLS_DTPREL32:
     write32le(loc, val);
     return;
   case R_LARCH_64:
   case R_LARCH_TLS_DTPREL64:
   case R_LARCH_64_PCREL:
     write64le(loc, val);
     return;
 
+  case R_LARCH_PCREL20_S2:
+    checkInt(loc, val, 22, rel);
+    checkAlignment(loc, val, 4, rel);
+    write32le(loc, setJ20(read32le(loc), val >> 2));
+    return;
+
   case R_LARCH_B16:
     checkInt(loc, val, 18, rel);
     checkAlignment(loc, val, 4, rel);
     write32le(loc, setK16(read32le(loc), val >> 2));
     return;
 
   case R_LARCH_B21:
     checkInt(loc, val, 23, rel);

lld/test/ELF/loongarch-pcrel20-s2.s

+# REQUIRES: loongarch
+
+# RUN: llvm-mc --filetype=obj --triple=loongarch32-unknown-elf %s -o %t.la32.o
+# RUN: llvm-mc --filetype=obj --triple=loongarch64-unknown-elf %s -o %t.la64.o
+
+# RUN: ld.lld %t.la32.o --section-start=.text=0x20000 --section-start=.data=0x20008 -o %t.la32.1
+# RUN: ld.lld %t.la64.o --section-start=.text=0x20000 --section-start=.data=0x20008 -o %t.la64.1
+# RUN: llvm-objdump --no-show-raw-insn -d %t.la32.1 | FileCheck --match-full-lines %s
+# RUN: llvm-objdump --no-show-raw-insn -d %t.la64.1 | FileCheck --match-full-lines %s
+# CHECK: 20000: pcaddi $t0, 2
+
+# RUN: not ld.lld %t.la32.o --section-start=.text=0x20000 --section-start=.data=0x220000 -o /dev/null 2>&1 | \
+# RUN:   FileCheck -DFILE=%t.la32.o --check-prefix=ERROR-RANGE %s
+# RUN: not ld.lld %t.la64.o --section-start=.text=0x20000 --section-start=.data=0x220000 -o /dev/null 2>&1 | \
+# RUN:   FileCheck -DFILE=%t.la64.o --check-prefix=ERROR-RANGE %s
+# ERROR-RANGE: error: [[FILE]]:(.text+0x0): relocation R_LARCH_PCREL20_S2 out of range: 2097152 is not in [-2097152, 2097151]; references section '.data'
+
+# RUN: not ld.lld %t.la32.o --section-start=.text=0x20000 --section-start=.data=0x40001 -o /dev/null 2>&1 | \
+# RUN:   FileCheck -DFILE=%t.la32.o --check-prefix=ERROR-ALIGN %s
+# RUN: not ld.lld %t.la64.o --section-start=.text=0x20000 --section-start=.data=0x40001 -o /dev/null 2>&1 | \
+# RUN:   FileCheck -DFILE=%t.la64.o --check-prefix=ERROR-ALIGN %s
+# ERROR-ALIGN: error: [[FILE]]:(.text+0x0): improper alignment for relocation R_LARCH_PCREL20_S2: 0x20001 is not aligned to 4 bytes
+
+.global _start
+
+_start:
+1:
+  pcaddi $t0, 0
+  .reloc 1b, R_LARCH_PCREL20_S2, .data
+
+.data
+  .word 0