@pk: Does moving the Mask variables inside the function fix the problem then? This should be then also fixed in the cvs... Yup, moving these variables onto the stack frees up 4 general purpose registers, and solves the compilation issue on my 32bit system (I've attached the patch for you).

No luck with the const I'm afraid. That results in the following errors when processing the assembly: FormatConversion.cpp:468: error: memory input 3 is not directly addressable FormatConversion.cpp:468: error: memory input 4 is not directly addressable FormatConversion.cpp:468: error: memory input …

Ahhhh. Okay, now I see what you mean. :blush: However, that function isn't actually using 3 registers. It's using 7. Those masks (with the m class) still use a register each, because their address is placed into esi, ecx, edx and eax, and then they're referenced from these registers. Because the ebp …

Thanks for the solution. It does seem to solve the issue. I've changed it to: : /* No outputs */ : "m" ((FxU)NumberOfPixels), "r" (Src), "r" (Dst), /*Inputs */ "m" (Mask565A), "m" (Mask565B), "m" (Mask565G), "m" (Mask565R) : "%mm0", "%mm1", "%mm2", "%mm3", "%mm4", "%mm5", "%mm6", "%mm7", "memory" /* …

Firstly, let me just say that this is fantastic work from everyone responsible! Tomb Raider 1 runs really well on my Linux system using Dosbox and these patches. I'm having a problem compiling some of the inline assembly inside the Convert565to8888 function (Line 486 of openglide-mmx.diff). In …