Adds support for a --deterministic command-line option that makes
repeated runs the same:
- Keyboard and mouse input is ignored
- The sound server does a periodic pull from the DMA channel (so that
it gets drained), but only does so via a periodic timer (instead of
being driven by a cubeb callback, which could arrive at different
times)
- Disk image writes are disabled (reads of a modified area still
work via an in-memory copy)
- NVRAM writes are disabled
- The current time that ViaCuda initializes the guest OS is always the
same.
This makes execution exactly the same each time, which should
make debugging of more subtle issues easier.
To validate that the deterministic mode is working, I've added a
periodic log of the current "time" (measured in cycle count), PC
and opcode. When comparing two runs with --log-no-uptime, the generated
log files are identical.
Shutdown will enter the debugger or quit depending on the execution mode.
Quit is different from shutdown since it is triggered outside the guest by using the host Quit menu item.
There's no reason for it to be a global, we always set it and use it
in instruction implementations, and we never read it directly.
Perhaps the compiler could optimize this away, but it's better to be
simpler (and also be easier to read).
Mode 1 contains real addressing mode entries, which by definition cannot
be using segment registers. By skipping over them, we can shave off a
couple of seconds from the 10.2 boot time.
They happen surprisingly often, and flushing the TLB is expensive
because we need to walk over all entries.
Takes booting 10.2 on a Beige G3 from binary start to "Welcome to Macintosh"
from 58s to 38s on my machine.
Keeps track of instructions (including operands) that are executed,
to see if there are any hotspots that could be optimized or fastpaths
that should be added.
Also adds a mode where CPU profiler data is periodically output, to
make it easier to get at these instruction counts during startup.
cur_dma_rgn->end is the last byte of a region. It is not the byte after the region. Therefore, subtract 1 from size before doing compare.
Also add more detail to the abort messages.
None of the POWER opcodes uses it now, plus it is a duplicate of ppc_setsoov (though ppc_setsoov is inline so it would have to be moved to be able to use it in poweropcodes.cpp?
Use U instead of UL. U will use the smallest size that can fit all the unsigned bytes. Since 0xFFFFFFFF fits in 32 bits, the 0xFFFFFFFFU is a uint32_t.
Including bits of rot_sh in the rA and MQ calculations is nonsensical since it is a rotation count and not a source of bits to be extracted or rotated.
The mask is not complicated, so we don't need to use power_rot_mask.
Fix carry flag calculation. Anding with the rotation count (n = rB) is nonsensical.
(r & ~mask) is the rotated word ANDed with the complement of the generated mask of n zeros followed by 32 - n ones.
The manual says this 32-bit result is ORed together. This means all the bits are ORed together which is equivalent to saying 0 if all zeros and 1 if any ones. In other words: (r & ~mask) != 0.
This boolean is ANDed with bit 0 of rS to produce the carry. int32_t(rS) < 0 will test bit 0. The && operator will treat each side as a boolean so you can exclude "!= 0" tests.
