Jools Wills c4076aea1a gzdoom - fix build with gcc 12 for armv8 on aarch64 kernel ...

When compiling with armv8 with gcc12 and an aarch64 kernel, the build fails with

    /usr/lib/gcc/arm-linux-gnueabihf/12/include/arm_neon.h: In function ‘float16x8_t vmulq_n_f16(float16x8_t, float16_t)’:
    /usr/lib/gcc/arm-linux-gnueabihf/12/include/arm_neon.h:17784:14: error: conversion of scalar ‘float’ to vector ‘float16x8_t’ involves truncation
    17784 |   return __a * __b;
          |          ~~~~^~~~~
    make[2]: *** [src/CMakeFiles/zdoom.dir/build.make:685: src/CMakeFiles/zdoom.dir/common/rendering/gles/gles_renderer.cpp.o] Error 1

Enabling -fmath-errno resolves the build error. This may be a gcc bug - not sure.

2025-02-19 23:13:30 +00:00

..

armv8_gcc12_fix.diff

gzdoom - fix build with gcc 12 for armv8 on aarch64 kernel

2025-02-19 23:13:30 +00:00

lzma_armv7_crc.diff

gzdoom - fix building 7zip/lzma on armv7 due to missing hardware crc

2024-03-19 23:10:32 +00:00