I had verified on hardware for passing state. However, from Cxbx-Reloaded is failing for incorrect procedure but did not crash. Even after I add FS offset of 0x58 patch for writer support, it does not pass the test anyway due to host's KPCR is stored per thread.
One other thing I notice about Cxbx-Reloaded's FS patches, it does not offer full set of patches for each FS offset which can be a problematic. Cxbx-Reloaded currently has reader support but does not have writer support which may break Windows until a FS patch is implemented.
I made couple of TODOs that are in need a fix.
1) There's missing kernel defines in nxdk that are in need of implement. For now, I extend the define to avoid conflict if they are upstream into nxdk.
2) FIXED I tried my best to make AT&T syntax for read/write FS offset instructions. But kept having problems with compiler, even though it looks correct to me according to the documentations.
I had verified on hardware for passing state. However, from Cxbx-Reloaded is failing for incorrect procedure but did not crash. Even after I add FS offset of 0x58 patch for writer support, it does not pass the test anyway due to host's KPCR is stored per thread.
One other thing I notice about Cxbx-Reloaded's FS patches, it does not offer full set of patches for each FS offset which can be a problematic. Cxbx-Reloaded currently has reader support but does not have writer support which may break Windows until a FS patch is implemented.
I made couple of TODOs that are in need a fix. 1) There's missing kernel defines in nxdk that are in need of implement. For now, I extend the define to avoid conflict if they are upstream into nxdk. 2) FIXED
I tried my best to make AT&T syntax for read/write FS offset instructions. But kept having problems with compiler, even though it looks correct to me according to the documentations.