On Wed Jan 31 16:59:45 2024 +0000, Rémi Bernon wrote:
Is this something we should be worried about already? Fwiw this mechanism has been in Proton for a while and I don't think there's been issues because of that. I don't think we can do much about it and if that only means pooling the mappings, I would prefer to only pay the price of the increased complexity if truly necessary.
The shared mapping allocation affects existing code path, so I'm inclined to believe that out-of-memory regression is inevitable, especially for 32bit apps.
Instead, how about calling `wine_server_handle_to_fd()` (include/wine/server.h) on the mapping and passing the returned fd to mmap(2)? No pooling necessary.