disable code which uses attribute(target_clones) to provide functions
optimized for different CPUs. apparently clang 15 adds supports for this,
but we don't have ifunc support in the toolchain, so it fails to link.

Index: src/common/darktable.h
--- src/common/darktable.h.orig
+++ src/common/darktable.h
@@ -127,18 +127,7 @@ typedef unsigned int u_int;
 /* Create cloned functions for various CPU SSE generations */
 /* See for instructions https://hannes.hauswedell.net/post/2017/12/09/fmv/ */
 /* TL;DR : use only on SIMD functions containing low-level paralellized/vectorized loops */
-#if __has_attribute(target_clones) && !defined(_WIN32) && !defined(NATIVE_ARCH)
-# if defined(__amd64__) || defined(__amd64) || defined(__x86_64__) || defined(__x86_64)
-#define __DT_CLONE_TARGETS__ __attribute__((target_clones("default", "sse2", "sse3", "sse4.1", "sse4.2", "popcnt", "avx", "avx2", "avx512f", "fma4")))
-# elif defined(__PPC64__)
-/* __PPC64__ is the only macro tested for in is_supported_platform.h, other macros would fail there anyway. */
-#define __DT_CLONE_TARGETS__ __attribute__((target_clones("default","cpu=power9")))
-# else
 #define __DT_CLONE_TARGETS__
-# endif
-#else
-#define __DT_CLONE_TARGETS__
-#endif
 
 /* Helper to force stack vectors to be aligned on 64 bits blocks to enable AVX2 */
 #define DT_IS_ALIGNED(x) __builtin_assume_aligned(x, 64)