restore memcpy comment
delete all the things!
fix jshint issues
restore memcpy comment
remove duplicate fxsave assignment
Count cache drops
Use already available physical address instead of calling read_imm8
Remove useless assertion
Just move around to reduce later diff
Run jit paging test with assertions enabled
Run jit-paging test on CI
Extend jit-paging test
Fix deleting invalidated code across memory pages
Add jit-paging test to gitlab ci
Remove jit_in_progress
Clean up old comments, use bool for jit_jump
Fix state image not begin garbage collected
Add ENABLE_PROFILER_TIMES to configure slow profiling times
Move to jit_generate and jit_run_interpreted to separate function
Add missing struct field
Fix: Don't write jit cache entry until no more faults can happen
Download image for jit paging test
Add missing initialiser
Mark jit_{generate,run_interpreted} as static
Specify full path to profiler.h
Clean up duplicate/missing declaration after rebase
mmap error handling, line length and fix some warnings
remove further unused code
move js imports to single header file
After restore_state, the pit.counter_start_time can be larger
than the current timestamp, leading to negative diff values, and
as a result, when get_counter_value is called, it can return
values larger than 16 bit integers. Thus, when ISA ports 0x40, 0x41, or
0x42 are read, the high-byte that counter_write returns can be larger
than 8-bit. Ultimately, this error is caught by the asserts in io.js. By
always applying a positive modulo to the counter value, it is always
ensured that the counter value is between zero and the reload value.
Everything that is valid json plus arraybuffers and arraybuffer-views
can be restored automatically (without making changes to other code).
Properties that should not be part of the state buffer needed to be
annotated using a magic `_state_skip` property. Shared arraybuffer views
are broken during restoration and also need to be annotated. Adapters
are not (and should never be) saved.
Current limitations:
- Before restoring, the cpu needs to be initialised with the original
settings. This is not a hard requirement and can be fixed later.
- Pending asynchronous operations are lost. It's a good idea to pause
the cpu before saving the state.
This change was necessary for several reasons:
- Code is much clearer and more modular
- The CPU object is now serializeable and therefore can be safed and
restored
- A dynamic translator from instructions to (optimized) JavaScript is
planned. It would use `new Function` to create functions at runtime.
Functions created this way don't have access to any scope (except for
the global scope) and therefore were not have been able to modify
registers, etc. This is now possible through the cpu object
- Possibly, but not necessarily a speedgain can be achieved.
See http://mrale.ph/blog/2012/09/23/grokking-v8-closures-for-fun.html
Additionally, some code was moved around.
