On 2023 Sep 28 (Thu) at 17:25:57 +0200 (+0200), Omar Polo wrote: :On 2023/09/28 09:04:02 -0600, phess...@openbsd.org wrote: :> http://build-failures.rhaalovely.net/aarch64/2023-09-26/archivers/unarr.log : :on non-apple arm devices it assumes there is sys/auxv.h and :asm/hwcap.h and uses getauxval(AT_HWCAP) to see if some features :(CRC32, SHA1, SHA2 and AES) are supported by the CPU. : :This is probably the smallest fix, but it unconditionally disables the :use of the cpu instructions. I don't have an aarch64 machine atm and :can't test. :
This works for me on my Thinkpad X13s. OK :What can we use to detect the availability of these instructions? : :Index: patches/patch-lzmasdk_CpuArch_c :=================================================================== :RCS file: patches/patch-lzmasdk_CpuArch_c :diff -N patches/patch-lzmasdk_CpuArch_c :--- /dev/null 1 Jan 1970 00:00:00 -0000 :+++ patches/patch-lzmasdk_CpuArch_c 28 Sep 2023 15:21:40 -0000 :@@ -0,0 +1,14 @@ :+Index: lzmasdk/CpuArch.c :+--- lzmasdk/CpuArch.c.orig :++++ lzmasdk/CpuArch.c :+@@ -760,10 +760,6 @@ BoolInt CPU_IsSupported_AES (void) { return APPLE_CRYP :+ :+ #else // __APPLE__ :+ :+-#include <sys/auxv.h> :+- :+-#define USE_HWCAP :+- :+ #ifdef USE_HWCAP :+ :+ #include <asm/hwcap.h> -- It's easier to fight for one's principles than to live up to them.