X-Git-Url: https://git.distorted.org.uk/~mdw/catacomb/blobdiff_plain/f48fb6a6b1e0f29a45b42beae638ef9886312579..a90d420cbe87490c844ae422c966e746d3134b07:/symm/salsa20-arm-neon.S diff --git a/symm/salsa20-arm-neon.S b/symm/salsa20-arm-neon.S index cea40192..3b6beb06 100644 --- a/symm/salsa20-arm-neon.S +++ b/symm/salsa20-arm-neon.S @@ -25,18 +25,19 @@ /// MA 02111-1307, USA. ///-------------------------------------------------------------------------- -/// External definitions. +/// Preliminaries. #include "config.h" #include "asm-common.h" -///-------------------------------------------------------------------------- -/// Main.code. - .arch armv7-a .fpu neon + .text +///-------------------------------------------------------------------------- +/// Main.code. + FUNC(salsa20_core_arm_neon) // Arguments are in registers. @@ -71,11 +72,6 @@ FUNC(salsa20_core_arm_neon) // [ 8 9 10 11] [ 8 13 2 7] (c, q10) // [12 13 14 15] [12 1 6 11] (d, q11) // - // [ 0 1 2 3] (a, q8) - // [ 4 5 6 7] (b, q9) - // [ 8 9 10 11] (c, q10) - // [12 13 14 15] (d, q11) - // // We need a copy for later. Rather than waste time copying them by // hand, we'll use the three-address nature of the instruction set. // But this means that the main loop is offset by a bit.