^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1) // SPDX-License-Identifier: GPL-2.0-only
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3) * linux/arch/arm/lib/xor-neon.c
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 4) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 5) * Copyright (C) 2013 Linaro Ltd <ard.biesheuvel@linaro.org>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 6) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 7)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 8) #include <linux/raid/xor.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 9) #include <linux/module.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 10)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 11) MODULE_LICENSE("GPL");
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 12)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 13) #ifndef __ARM_NEON__
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 14) #error You should compile this file with '-march=armv7-a -mfloat-abi=softfp -mfpu=neon'
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 15) #endif
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 16)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 17) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 18) * Pull in the reference implementations while instructing GCC (through
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 19) * -ftree-vectorize) to attempt to exploit implicit parallelism and emit
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 20) * NEON instructions.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 21) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 22) #if __GNUC__ > 4 || (__GNUC__ == 4 && __GNUC_MINOR__ >= 6)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 23) #pragma GCC optimize "tree-vectorize"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 24) #else
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 25) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 26) * While older versions of GCC do not generate incorrect code, they fail to
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 27) * recognize the parallel nature of these functions, and emit plain ARM code,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 28) * which is known to be slower than the optimized ARM code in asm-arm/xor.h.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 29) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 30) #warning This code requires at least version 4.6 of GCC
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 31) #endif
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 32)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 33) #pragma GCC diagnostic ignored "-Wunused-variable"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 34) #include <asm-generic/xor.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 35)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 36) struct xor_block_template const xor_block_neon_inner = {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 37) .name = "__inner_neon__",
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 38) .do_2 = xor_8regs_2,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 39) .do_3 = xor_8regs_3,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 40) .do_4 = xor_8regs_4,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 41) .do_5 = xor_8regs_5,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 42) };
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 43) EXPORT_SYMBOL(xor_block_neon_inner);