[lld][RISCV] Add break to nested switch in mergeAtomic

[4vtomat]

This prevent the warnings from compiler.

[llvmbot]

@llvm/pr-subscribers-lld

@llvm/pr-subscribers-lld-elf

Author: Brandon Wu (4vtomat)

Changes

This prevent the warnings from compiler.

---

Full diff: https://github.com/llvm/llvm-project/pull/99762.diff

1 Files Affected:

• (modified) lld/ELF/Arch/RISCV.cpp (+3)

diff --git a/lld/ELF/Arch/RISCV.cpp b/lld/ELF/Arch/RISCV.cpp
index 6af89ce3517b7..56759c28dcf41 100644
--- a/lld/ELF/Arch/RISCV.cpp
+++ b/lld/ELF/Arch/RISCV.cpp
@@ -1131,6 +1131,7 @@ static void mergeAtomic(DenseMap<unsigned, unsigned>::iterator it,
     case RISCVAttrs::RISCVAtomicAbiTag::A6C:
       return;
     };
+    break;
 
   case RISCVAtomicAbiTag::A6S:
     switch (newTag) {
@@ -1144,6 +1145,7 @@ static void mergeAtomic(DenseMap<unsigned, unsigned>::iterator it,
     case RISCVAttrs::RISCVAtomicAbiTag::A6S:
       return;
     };
+    break;
 
   case RISCVAtomicAbiTag::A7:
     switch (newTag) {
@@ -1157,6 +1159,7 @@ static void mergeAtomic(DenseMap<unsigned, unsigned>::iterator it,
     case RISCVAttrs::RISCVAtomicAbiTag::A7:
       return;
     };
+    break;
   };
 
   // If we get here, then we have an invalid tag, so report it.

[ilovepi]

The change seems fine, but what compiler is warning about this? lld was compiling without warning for me recently with ToT clang.

[4vtomat]

The change seems fine, but what compiler is warning about this? lld was compiling without warning for me recently with ToT clang.

The warning is

llvm-project/lld/ELF/Arch/RISCV.cpp: In function ‘void mergeAtomic(llvm::DenseMapBase<llvm::DenseMap<unsigned int, unsigned int>, unsigned int, unsigned int, llvm
::DenseMapInfo<unsigned int>, llvm::detail::DenseMapPair<unsigned int, unsigned int> >::iterator, const lld::elf::InputSectionBase*, const lld::elf::InputSectionBase*, llvm::RISCVA
ttrs::RISCVAtomicAbiTag, llvm::RISCVAttrs::RISCVAtomicAbiTag)’:
llvm-project/lld/ELF/Arch/RISCV.cpp:1123:5: warning: this statement may fall through [-Wimplicit-fallthrough=]
     switch (newTag) {
     ^~~~~~
llvm-project/lld/ELF/Arch/RISCV.cpp:1135:3: note: here
   case RISCVAtomicAbiTag::A6S:
   ^~~~
llvm-project/lld/ELF/Arch/RISCV.cpp:1136:5: warning: this statement may fall through [-Wimplicit-fallthrough=]
     switch (newTag) {
     ^~~~~~
llvm-project/lld/ELF/Arch/RISCV.cpp:1148:3: note: here
   case RISCVAtomicAbiTag::A7:
   ^~~~

[ilovepi]

Right, but I don’t recall seeing that with a pretty recent clang, and I want to know if you’re using gcc or another compiler so I know if I should see if we should fix that in clang.

[4vtomat]

Right, but I don’t recall seeing that with a pretty recent clang, and I want to know if you’re using gcc or another compiler so I know if I should see if we should fix that in clang.

Actually it's during compiling the lld itself lol~

[ilovepi]

Right, but I don’t recall seeing that with a pretty recent clang, and I want to know if you’re using gcc or another compiler so I know if I should see if we should fix that in clang.

Actually it's during compiling the lld itself lol~

Yes, but are you compiling with clang as the compiler or something else? Because if it’s clang then I don’t need to file a bug and try to run down why we’re not handling that. If you’re using something else, then I do.

[4vtomat]

Right, but I don’t recall seeing that with a pretty recent clang, and I want to know if you’re using gcc or another compiler so I know if I should see if we should fix that in clang.

Actually it's during compiling the lld itself lol~

Yes, but are you compiling with clang as the compiler or something else? Because if it’s clang then I don’t need to file a bug and try to run down why we’re not handling that. If you’re using something else, then I do.

Oh, I got your point, I was using gcc.

[ilovepi]

Thanks I’ll take a look at top of tree clang again, and file a bug if we’re missing warnings in that case.