Stability fixes

[elad335]

Adds 2 new bug fixes. (not regression fixes)

• Fix an LV2 bug that over time can randomly kill the PPUs at once. Timeline of the race condition:

1. The PPU is in SLEEP state. state = suspend.
2. lv2_obj::awake is called on the traced thread and is now in ONPROC state, state = signal.
3. lv2_obj::awake is called by another thread externally with a priority higher than our traced thread and appends it to g_pending. state = suspend + signal.
4. lv2_obj::sleep/set_priority (highering priority) is called on any thread which is in ONPROC. Causing it to enter SLEEP or RUNNING state, while the traced thread is back in queue in ONPROC. state = suspend + signal.
5. The traced thread finally calls lv2_obj::awake on itself, g_pending decrements to 0 and we a have a rescheduling event, after XOR state = no flags (no signal)
6. In cpu_thread::check_state: cpu_sleep_called is now true and remains this way.
7. Another thread with a higher prioty kicks in and appends the traced thread into g_pending. state = suspend.
8. The traced thread is at cpu_thread::cpu_wait() waiting on state flags, awake is not called again because cpu_sleep_called is true. And that function is where it's gonna spend the rest of its life.
   Fix this bug by not using XOR, instead properly add signal and substruct suspend flags.

• Fix a loose notification in lwcond wait: put awake under mutex lock so when an lwmutex waiter times out a notification won't be able to awake it from other syscalls.

For me it fixes freezes in Demon Souls after a while, but its effect is global on games some may experience the freeze much earlier than others.

[Nerboruto]

possible that after all these fixes demon souls run well even with only one thread ppu? before it crashed, without recompiling cache.
tested on intel quad core cpu, performances seem identical indeed perhaps even a bit more fluid.

[elad335]

Why would you run on a single PPU? PPU thread count is not a "performance hint", it has real consequences on game compatibility. Lots of games aren't compatible with 1 PPU thread and they will never be.

[Nerboruto]

yes I tried others and they crash ... I did the test because I saw that by limiting the spu threads to 3 the game runs better on quad core and so I wanted to try the ppu too ... anyway I remember first that ds crashed and I repeat it seems more fluid with 1 ppu.

other tricks that I have tried and improve the performance are patches to 60 but vblanlk to 30. much smoother than the default 30 and decompress the archives but compress the iso. maybe so I shift the compression load out of the emulator ..