[PATCH] D59348: [DebugInfo] Combine Trivial and NonTrivial flags

[Adrian Prantl via llvm-commits]

If it wasn't in the last LLVM release and nobody in the community released an LLVM-based product to users that use PDB debug info and LTO in the mean time, it will not be a big deal.

-- adrian

> On Apr 10, 2019, at 3:50 PM, Aaron Smith <aaron.lee.smith at gmail.com> wrote:
> 
> I know very little about bitcode compatibility in llvm. Since both of these flags were added in the last two or three months this is probably not a big deal? The whole reason for the change is to reclaim the bit.
> 
>> On Apr 10, 2019, at 4:43 PM, Adrian Prantl via Phabricator <reviews at reviews.llvm.org> wrote:
>> 
>> aprantl added inline comments.
>> 
>> 
>> ================
>> Comment at: include/llvm/IR/DebugInfoFlags.def:57-60
>> +HANDLE_DI_FLAG((1 << 26), NonTrivial)
>> HANDLE_DI_FLAG((1 << 27), BigEndian)
>> HANDLE_DI_FLAG((1 << 28), LittleEndian)
>> HANDLE_DI_FLAG((1 << 29), AllCallsDescribed)
>> ----------------
>> dblaikie wrote:
>>> @aprantl - could you review this for bitcode compatibility?
>>> 
>>> I  imagine reusing a bit for a different meaning (the exact opposite) would be problematic?
>>> 
>>> Would we have to settle for keeping the bitcode encoding as 1=trivial 0=nontrivial (only for structs/classes? Maybe enums too?) - or burn the 26th bit and keep the 30th bit as NonTrivial?
>> Since this only affects PDB debug info I personally don't worry about breaking bitcode compatibility, however other people may do so.
>> 
>> The correct way to update this is to bump the version of the records (DICompositeType) and parse the old or the new format accordingly.
>> 
>> If this isn't done any LLVM module with bitcode that currently uses bit 30 (which is now > Largest) will cause a verifier error and the debug info from that module will be dropped with a warning. LTO users may or may not care about this.
>> 
>> 
>> CHANGES SINCE LAST ACTION
>> https://reviews.llvm.org/D59348/new/
>> 
>> https://reviews.llvm.org/D59348
>> 
>> 
>>