Maamoun TK maamoun.tk@googlemail.com writes:
+dnl NETTLE_CHECK_POWER_CRYPTO_EXT +dnl --------------------- +dnl Check if POWER crypto extensions should be used. +dnl Obeys enable_power_crypto_ext, which should be set earlier. +AC_DEFUN([NETTLE_CHECK_POWER_CRYPTO_EXT], +[if test "$enable_power_crypto_ext" = auto ; then
- if test "$cross_compiling" = yes ; then
- dnl Check if compiler/assembler accepts it.
- AC_CACHE_CHECK([if assembler accepts crypto instructions],
nettle_cv_asm_power_vcrypto,[GMP_TRY_ASSEMBLE([+.text +foo:
- vcipher 0, 0, 1
+],
[nettle_cv_asm_power_vcrypto=yes],[nettle_cv_asm_power_vcrypto=no])])- enable_power_crypto_ext="$nettle_cv_asm_power_vcrypto"
- else
- AC_CACHE_CHECK([if crypto extensions supported],
nettle_cv_asm_power_vcrypto,[AC_RUN_IFELSE([AC_LANG_PROGRAM([[+#if defined(__FreeBSD__) && __FreeBSD__ < 12 +#include <sys/sysctl.h> +#else +#include <sys/auxv.h> +#endif
Do you expect that this "auto" logic does what that user wants? I'm thinking, maybe it's simpler to stick with just yes/no (no being the default), and then add support for --enable-fat later, to select code at run-time?
--- /dev/null +++ b/powerpc64/machine.m4 @@ -0,0 +1,24 @@ +define(<PROLOGUE>, +<ifelse(WORDS_BIGENDIAN,no, +<.align 5 +.globl C_NAME($1) +DECLARE_FUNC(C_NAME($1)) +C_NAME($1): +addis 2,12,(.TOC.-C_NAME($1))@ha +addi 2,2,(.TOC.-C_NAME($1))@l +.localentry C_NAME($1), .-C_NAME($1)>, +<.globl C_NAME($1) +DECLARE_FUNC(C_NAME($1)) +.section ".opd","aw" +.align 3 +C_NAME($1): +.quad .C_NAME($1),.TOC.@tocbase,0 +.previous +.align 5 +.C_NAME($1):>)>)
Overriding PROLOGUE here looks fine, but it would be nice with a comment explaining what's needed, and/or linking the some appropriate ABI specification.
Regards, /Niels