| Mon, 23 Jun 2014 11:46:56 -0400 |
Mike Pavone |
Save and restore guest address in the write function for a code memory chunk in the "slow" path for inconvenient host addresses. This fixes an intermittent crash on OSX in the code that checks whether the memory written may contain code
|
| Mon, 23 Jun 2014 11:05:55 -0400 |
Mike Pavone |
Fix x86_rrindex_sizedir. Pass the correct scale to mov_rindexr in gen_mem_fun. BlastEm now sort of works on OS X. Runs reliably from lldb, but only intermittently from the shell
|
| Tue, 17 Jun 2014 01:50:29 -0400 |
Mike Pavone |
blastem builds and almost works on OS X now
|
| Mon, 24 Feb 2014 01:30:16 -0800 |
Mike Pavone |
Use a typedef code_ptr in place of uint8_t * in 68K core to better support host instruction sets with different instruction word sizes. Make x86_68k_options contain a cpu_options so that gen_mem_fun can eventually be shared with the Z80 core.
|
| Mon, 24 Feb 2014 00:50:15 -0800 |
Mike Pavone |
Rename x86_backend.h and x86_backend.c to backend.h and backend.c respectively
|
| Mon, 24 Feb 2014 00:39:02 -0800 |
Michael Pavone |
Start work on refactoring some of the backend code to allow more sharing between M68K and Z80 cores and possibly also between x86 and the ARM backend when it exists
|
| Wed, 19 Feb 2014 00:22:27 -0800 |
Michael Pavone |
Apart from the Z80 core, BlastEm now supports 32-bit x86
|
| Tue, 18 Feb 2014 21:37:31 -0800 |
Michael Pavone |
Initial work on the x86-32 target
|
| Tue, 18 Feb 2014 20:32:10 -0800 |
Michael Pavone |
Generate get_sr, set_sr and set_ccr at runtime so they can respect the flag_regs setting
|
| Mon, 17 Feb 2014 19:58:37 -0800 |
Michael Pavone |
Make references to flags in the M68K core respect the flag_regs options array so that flags can be moved out of registers for the 32-bit port. set/get ccr/sr still need to be updated to support this, but everything else should be done.
|
| Sun, 16 Feb 2014 19:54:27 -0800 |
Michael Pavone |
Generate handle_cycle_limit at runtime so it can use the generated save/load_context functions. Since the hand written versions of save/load are no longer used they have been removed.
|
| Sun, 16 Feb 2014 19:21:39 -0800 |
Michael Pavone |
Generate native_addr and native_addr_and_sync at runtime so they can use the generated save/load_context functions
|
| Sun, 16 Feb 2014 18:17:59 -0800 |
Michael Pavone |
Generate retrans_stub at runtime so it can use the generated save/load_context functions
|
| Sun, 16 Feb 2014 16:50:10 -0800 |
Michael Pavone |
Removed old debug print function that is no longer needed
|
| Sat, 15 Feb 2014 22:35:18 -0800 |
Michael Pavone |
Get rid of the native stack option the 68K core. Trying to make it work with code that messes with the stack is not worth the trouble.
|
