Basically, using timestamp counter.
Rewritten vm::reservation_op with the same principle.
Rewritten another transaction helper.
Add two new settings for configuring fallbacks.
Two limits are specified in nanoseconds (first and second).
Fix PUTLLC reload logic (prevent reusing garbage).
Add prefetch hint list parameter.
Workloads may be executed by another thread on another CPU core.
It means they may benefit from directly prefetching the data as hinted.
Also implement mov_rdata_nt, for "streaming" data from such workloads.
The OS defaults for SO_RCVBUF / SO_SNDBUF may not match what the defaults for the PS3 are.
And the code may be happy with the PS3 defaults, so may not set this explicitly.
So we'll do it when we establish the socket.
It also looks like the Windows recv behaviour is different for the MSG_PEEK option (and possibly in all situations where a smaller buffer is provided). I believe other platforms will return the size of the data received into the socket buffer, even if the supplied user buffer is smaller. Windows returns WSAEMSGSIZE instead. For the length of '1' MSG_PEEK request that is currently seen, we just mask this and return the full size of the buffer. This might need to be the case in all situations, I assume there will be applications that will supply an artificially small buffer and resize it as necessary based on the size of the received packet.
Add some additional translations from native->PS3 error codes
Rewrote handling of MSG_PEEK to just be a copy of what GalCiv suggested
Added WSAESHUTDOWN, and WSASetLastError as done by Cygwin
Allow more in first-chance transactions.
Allow abandonment of PUTLLC as in original path.
Make PUTLLUC unconditionally shared-locked.
Give PUTLLC +1 priority (minor change).
Remove vm::reservation_lock from it.
Use lock bits to prevent memory clobbering in GETLLAR.
Improve u128 for MSVC since it's used for bitlocking.
Improve 128 bit atomics for the same reason.
Improve vm::reservation_op and friends.
Allocate "personal" range lock variable for each spu_thread.
Switch from reservation_lock to range lock for all stores.
Detect actual memory mirrors in shareable cache setup logic.
