cudaMalloc & cudaMemcpy from different host threads Anything new concerning this issue

Hi all

Problem :
How to cudaMalloc a pointer from one host thread, and then cudaMemcpy or use it in a kernel from another host thread.

Searching the web from this issue gives me post dating from 2007-2008 saying that’s not possible, period.

Anything new concerning sharing cuda memory context between host threads ?

Best regards