C++2a synchronisation inefficient in GCC 11
Thiago Macieira
thiago.macieira@intel.com
Wed Mar 3 17:12:12 GMT 2021
On Wednesday, 3 March 2021 06:24:45 PST Jonathan Wakely wrote:
> But it's not a new problem for GCC 11, and not specific to the atomic
> wait feature. It's worth further consideration though.
It isn't, except that in this case we know, before release, that ABI will
silently break in ways that may cause heisenbugs and deadlocks. Given that we
know that, I'm asking for a conscious decision on one of three options:
1) make the functionality unavailable (by default or at all)
2) make the ABI break non-silent when it comes
3) accept the silent ABI break
--
Thiago Macieira - thiago.macieira (AT) intel.com
Software Architect - Intel DPG Cloud Engineering
More information about the Libstdc++
mailing list