6c4f332705
Improve synchapi impls; fix heap races & 64-bit fixes
2025-11-10 18:15:33 -07:00
2494b541b7
x86_64 Linux: Add fallback for CPUs without RDFSBASE/WRFSBASE
2025-11-10 00:25:28 -07:00
76f97efe07
macOS: Add async_io_kqueue; alloc LDT entry per thread
2025-11-10 00:25:23 -07:00
4a7a6b9d72
Add CMake toolchains and update CI
2025-11-10 00:24:58 -07:00
d85671e3c6
Download and use embedded msvcrt.dll
2025-11-09 13:18:33 -07:00
cc0a887302
Initial macOS support (x86_64 with Rosetta 2)
2025-11-08 01:45:47 -07:00
3dd9fb77ff
Experimental 64-bit host support
2025-11-06 19:14:48 -07:00
73b1ffb8d6
Numerous msvcrt/crt fixes; support variable resolution
2025-11-03 21:46:31 -07:00
8c39e06cca
Use CDECL_NO_CONV for va_list thunks; remove .globl
2025-11-03 19:25:20 -07:00
bfba360d30
Implement stack switching in guest<->host trampolines
2025-11-03 19:21:35 -07:00
4dc599381f
Refactor memory management into wibo::heap
...
- Removes blockUpper2GB hack; we now start early in the process
and reserve all (available) space in the lower 2GB address
space, leaving the upper 2GB untouched for host code
- All virtual memory operations flow through wibo::heap for
bookkeeping
- All guest code uses a guest mimalloc area + thread-local heaps
reserved in the guest address space
2025-11-03 13:58:33 -07:00
52cdd7c811
Fully functional guest<->host trampolines
2025-10-31 00:45:32 -06:00
55a62a27c9
Initial guest-to-host trampoline generation
2025-10-30 02:23:09 -06:00
