[clang] [llvm] [DebugInfo] Emit DW_AT_bit_size for _BitInt types and others (PR #164372)

[David Blaikie via cfe-commits]

================
@@ -766,8 +766,17 @@ void DwarfUnit::constructTypeDIE(DIE &Buffer, const DIBasicType *BTy) {
     addUInt(Buffer, dwarf::DW_AT_encoding, dwarf::DW_FORM_data1,
             BTy->getEncoding());
 
-  uint64_t Size = BTy->getSizeInBits() >> 3;
-  addUInt(Buffer, dwarf::DW_AT_byte_size, std::nullopt, Size);
+  uint64_t SizeInBytes = divideCeil(BTy->getSizeInBits(), 8);
----------------
dwblaikie wrote:

Hmm, I think the quoted parts you mentioned do refute/address my concerns - though perhaps the spec expects us to also emit a DW_AT_data_bit_offset of zero in this case? Again, I don't mind matching GCC if it's just a bit weird - though it wouldn't be hard or expensive to add DW_AT_data_bit_offset as well (especially if it's just always 0 and can be stored in the abbrev with implicit_const - but I guess that depends where the padding is and where the bits are stored (if they're LSB or MSB, etc))

https://github.com/llvm/llvm-project/pull/164372