bulbazord added a comment.

In D157041#4559255 <https://reviews.llvm.org/D157041#4559255>, @augusto2112 
wrote:

>> Do all of these need to be protected with a mutex? In your description 
>> you're saying TSan is detecting data races. What piece of data are you 
>> observing the data race on?
>
> Only SetAsBoolean/GetAsBoolean are being caught when running our test suite 
> at the moment.
>
>> Do we need a recursive mutex? I assume that these operations might call into 
>> each other, but if not it would be nice to just have it be a std::mutex.
>
> If we want to protect all accesses then yes, since these functions call each 
> other. It we decide to only lock against what's being caught by tsan then we 
> can get by with a regular mutex.

Only `SetAsBoolean` and `GetAsBoolean` are being caught right now, I see. The 
implementation of those methods really isn't any different (AFAICT) from the 
other methods so I would assume we would need to protect all the other `Get` 
and `Set` methods as well then, makes sense. I'd probably keep the recursive 
mutex for now then.


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D157041/new/

https://reviews.llvm.org/D157041

_______________________________________________
lldb-commits mailing list
lldb-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/lldb-commits

Reply via email to