/* -*- linux-c -*- ------------------------------------------------------- * * * Copyright 2002-2004 H. Peter Anvin - All Rights Reserved * * This program is free software; you can redistribute it and/or modify * it under the terms of the GNU General Public License as published by * the Free Software Foundation, Inc., 53 Temple Place Ste 330, * Boston MA 02111-1307, USA; either version 2 of the License, or * (at your option) any later version; incorporated herein by reference. * * ----------------------------------------------------------------------- */ /* * int8.c * * 8-way unrolled portable integer math RAID-6 instruction set * * This file is postprocessed using unroll.awk */ #include /* * This is the C data type to use */ /* Change this from BITS_PER_LONG if there is something better... */ #if BITS_PER_LONG == 64 # define NBYTES(x) ((x) * 0x0101010101010101UL) # define NSIZE 8 # define NSHIFT 3 # define NSTRING "64" typedef u64 unative_t; #else # define NBYTES(x) ((x) * 0x01010101U) # define NSIZE 4 # define NSHIFT 2 # define NSTRING "32" typedef u32 unative_t; #endif /* * These sub-operations are separate inlines since they can sometimes be * specially optimized using architecture-specific hacks. */ /* * The SHLBYTE() operation shifts each byte left by 1, *not* * rolling over into the next byte */ static inline __attribute_const__ unative_t SHLBYTE(unative_t v) { unative_t vv; vv = (v << 1) & NBYTES(0xfe); return vv; } /* * The MASK() operation returns 0xFF in any byte for which the high * bit is 1, 0x00 for any byte for which the high bit is 0. */ static inline __attribute_const__ unative_t MASK(unative_t v) { unative_t vv; vv = v & NBYTES(0x80); vv = (vv << 1) - (vv >> 7); /* Overflow on the top bit is OK */ return vv; } static void raid6_int8_gen_syndrome(int disks, size_t bytes, void **ptrs) { u8 **dptr = (u8 **)ptrs; u8 *p, *q; int d, z, z0; unative_t wd0, wq0, wp0, w10, w20; unative_t wd1, wq1, wp1, w11, w21; unative_t wd2, wq2, wp2, w12, w22; unative_t wd3, wq3, wp3, w13, w23; unative_t wd4, wq4, wp4, w14, w24; unative_t wd5, wq5, wp5, w15, w25; unative_t wd6, wq6, wp6, w16, w26; unative_t wd7, wq7, wp7, w17, w27; z0 = disks - 3; /* Highest data disk */ p = dptr[z0+1]; /* XOR parity */ q = dptr[z0+2]; /* RS syndrome */ for ( d = 0 ; d < bytes ; d += NSIZE*8 ) { wq0 = wp0 = *(unative_t *)&dptr[z0][d+0*NSIZE]; wq1 = wp1 = *(unative_t *)&dptr[z0][d+1*NSIZE]; wq2 = wp2 = *(unative_t *)&dptr[z0][d+2*NSIZE]; wq3 = wp3 = *(unative_t *)&dptr[z0][d+3*NSIZE]; wq4 = wp4 = *(unative_t *)&dptr[z0][d+4*NSIZE]; wq5 = wp5 = *(unative_t *)&dptr[z0][d+5*NSIZE]; wq6 = wp6 = *(unative_t *)&dptr[z0][d+6*NSIZE]; wq7 = wp7 = *(unative_t *)&dptr[z0][d+7*NSIZE]; for ( z = z0-1 ; z >= 0 ; z-- ) { wd0 = *(unative_t *)&dptr[z][d+0*NSIZE]; wd1 = *(unative_t *)&dptr[z][d+1*NSIZE]; wd2 = *(unative_t *)&dptr[z][d+2*NSIZE]; wd3 = *(unative_t *)&dptr[z][d+3*NSIZE]; wd4 = *(unative_t *)&dptr[z][d+4*NSIZE]; wd5 = *(unative_t *)&dptr[z][d+5*NSIZE]; wd6 = *(unative_t *)&dptr[z][d+6*NSIZE]; wd7 = *(unative_t *)&dptr[z][d+7*NSIZE]; wp0 ^= wd0; wp1 ^= wd1; wp2 ^= wd2; wp3 ^= wd3; wp4 ^= wd4; wp5 ^= wd5; wp6 ^= wd6; wp7 ^= wd7; w20 = MASK(wq0); w21 = MASK(wq1); w22 = MASK(wq2); w23 = MASK(wq3); w24 = MASK(wq4); w25 = MASK(wq5); w26 = MASK(wq6); w27 = MASK(wq7); w10 = SHLBYTE(wq0); w11 = SHLBYTE(wq1); w12 = SHLBYTE(wq2); w13 = SHLBYTE(wq3); w14 = SHLBYTE(wq4); w15 = SHLBYTE(wq5); w16 = SHLBYTE(wq6); w17 = SHLBYTE(wq7); w20 &= NBYTES(0x1d); w21 &= NBYTES(0x1d); w22 &= NBYTES(0x1d); w23 &= NBYTES(0x1d); w24 &= NBYTES(0x1d); w25 &= NBYTES(0x1d); w26 &= NBYTES(0x1d); w27 &= NBYTES(0x1d); w10 ^= w20; w11 ^= w21; w12 ^= w22; w13 ^= w23; w14 ^= w24; w15 ^= w25; w16 ^= w26; w17 ^= w27; wq0 = w10 ^ wd0; wq1 = w11 ^ wd1; wq2 = w12 ^ wd2; wq3 = w13 ^ wd3; wq4 = w14 ^ wd4; wq5 = w15 ^ wd5; wq6 = w16 ^ wd6; wq7 = w17 ^ wd7; } *(unative_t *)&p[d+NSIZE*0] = wp0; *(unative_t *)&p[d+NSIZE*1] = wp1; *(unative_t *)&p[d+NSIZE*2] = wp2; *(unative_t *)&p[d+NSIZE*3] = wp3; *(unative_t *)&p[d+NSIZE*4] = wp4; *(unative_t *)&p[d+NSIZE*5] = wp5; *(unative_t *)&p[d+NSIZE*6] = wp6; *(unative_t *)&p[d+NSIZE*7] = wp7; *(unative_t *)&q[d+NSIZE*0] = wq0; *(unative_t *)&q[d+NSIZE*1] = wq1; *(unative_t *)&q[d+NSIZE*2] = wq2; *(unative_t *)&q[d+NSIZE*3] = wq3; *(unative_t *)&q[d+NSIZE*4] = wq4; *(unative_t *)&q[d+NSIZE*5] = wq5; *(unative_t *)&q[d+NSIZE*6] = wq6; *(unative_t *)&q[d+NSIZE*7] = wq7; } } static void raid6_int8_xor_syndrome(int disks, int start, int stop, size_t bytes, void **ptrs) { u8 **dptr = (u8 **)ptrs; u8 *p, *q; int d, z, z0; unative_t wd0, wq0, wp0, w10, w20; unative_t wd1, wq1, wp1, w11, w21; unative_t wd2, wq2, wp2, w12, w22; unative_t wd3, wq3, wp3, w13, w23; unative_t wd4, wq4, wp4, w14, w24; unative_t wd5, wq5, wp5, w15, w25; unative_t wd6, wq6, wp6, w16, w26; unative_t wd7, wq7, wp7, w17, w27; z0 = stop; /* P/Q right side optimization */ p = dptr[disks-2]; /* XOR parity */ q = dptr[disks-1]; /* RS syndrome */ for ( d = 0 ; d < bytes ; d += NSIZE*8 ) { /* P/Q data pages */ wq0 = wp0 = *(unative_t *)&dptr[z0][d+0*NSIZE]; wq1 = wp1 = *(unative_t *)&dptr[z0][d+1*NSIZE]; wq2 = wp2 = *(unative_t *)&dptr[z0][d+2*NSIZE]; wq3 = wp3 = *(unative_t *)&dptr[z0][d+3*NSIZE]; wq4 = wp4 = *(unative_t *)&dptr[z0][d+4*NSIZE]; wq5 = wp5 = *(unative_t *)&dptr[z0][d+5*NSIZE]; wq6 = wp6 = *(unative_t *)&dptr[z0][d+6*NSIZE]; wq7 = wp7 = *(unative_t *)&dptr[z0][d+7*NSIZE]; for ( z = z0-1 ; z >= start ; z-- ) { wd0 = *(unative_t *)&dptr[z][d+0*NSIZE]; wd1 = *(unative_t *)&dptr[z][d+1*NSIZE]; wd2 = *(unative_t *)&dptr[z][d+2*NSIZE]; wd3 = *(unative_t *)&dptr[z][d+3*NSIZE]; wd4 = *(unative_t *)&dptr[z][d+4*NSIZE]; wd5 = *(unative_t *)&dptr[z][d+5*NSIZE]; wd6 = *(unative_t *)&dptr[z][d+6*NSIZE]; wd7 = *(unative_t *)&dptr[z][d+7*NSIZE]; wp0 ^= wd0; wp1 ^= wd1; wp2 ^= wd2; wp3 ^= wd3; wp4 ^= wd4; wp5 ^= wd5; wp6 ^= wd6; wp7 ^= wd7; w20 = MASK(wq0); w21 = MASK(wq1); w22 = MASK(wq2); w23 = MASK(wq3); w24 = MASK(wq4); w25 = MASK(wq5); w26 = MASK(wq6); w27 = MASK(wq7); w10 = SHLBYTE(wq0); w11 = SHLBYTE(wq1); w12 = SHLBYTE(wq2); w13 = SHLBYTE(wq3); w14 = SHLBYTE(wq4); w15 = SHLBYTE(wq5); w16 = SHLBYTE(wq6); w17 = SHLBYTE(wq7); w20 &= NBYTES(0x1d); w21 &= NBYTES(0x1d); w22 &= NBYTES(0x1d); w23 &= NBYTES(0x1d); w24 &= NBYTES(0x1d); w25 &= NBYTES(0x1d); w26 &= NBYTES(0x1d); w27 &= NBYTES(0x1d); w10 ^= w20; w11 ^= w21; w12 ^= w22; w13 ^= w23; w14 ^= w24; w15 ^= w25; w16 ^= w26; w17 ^= w27; wq0 = w10 ^ wd0; wq1 = w11 ^ wd1; wq2 = w12 ^ wd2; wq3 = w13 ^ wd3; wq4 = w14 ^ wd4; wq5 = w15 ^ wd5; wq6 = w16 ^ wd6; wq7 = w17 ^ wd7; } /* P/Q left side optimization */ for ( z = start-1 ; z >= 0 ; z-- ) { w20 = MASK(wq0); w21 = MASK(wq1); w22 = MASK(wq2); w23 = MASK(wq3); w24 = MASK(wq4); w25 = MASK(wq5); w26 = MASK(wq6); w27 = MASK(wq7); w10 = SHLBYTE(wq0); w11 = SHLBYTE(wq1); w12 = SHLBYTE(wq2); w13 = SHLBYTE(wq3); w14 = SHLBYTE(wq4); w15 = SHLBYTE(wq5); w16 = SHLBYTE(wq6); w17 = SHLBYTE(wq7); w20 &= NBYTES(0x1d); w21 &= NBYTES(0x1d); w22 &= NBYTES(0x1d); w23 &= NBYTES(0x1d); w24 &= NBYTES(0x1d); w25 &= NBYTES(0x1d); w26 &= NBYTES(0x1d); w27 &= NBYTES(0x1d); wq0 = w10 ^ w20; wq1 = w11 ^ w21; wq2 = w12 ^ w22; wq3 = w13 ^ w23; wq4 = w14 ^ w24; wq5 = w15 ^ w25; wq6 = w16 ^ w26; wq7 = w17 ^ w27; } *(unative_t *)&p[d+NSIZE*0] ^= wp0; *(unative_t *)&p[d+NSIZE*1] ^= wp1; *(unative_t *)&p[d+NSIZE*2] ^= wp2; *(unative_t *)&p[d+NSIZE*3] ^= wp3; *(unative_t *)&p[d+NSIZE*4] ^= wp4; *(unative_t *)&p[d+NSIZE*5] ^= wp5; *(unative_t *)&p[d+NSIZE*6] ^= wp6; *(unative_t *)&p[d+NSIZE*7] ^= wp7; *(unative_t *)&q[d+NSIZE*0] ^= wq0; *(unative_t *)&q[d+NSIZE*1] ^= wq1; *(unative_t *)&q[d+NSIZE*2] ^= wq2; *(unative_t *)&q[d+NSIZE*3] ^= wq3; *(unative_t *)&q[d+NSIZE*4] ^= wq4; *(unative_t *)&q[d+NSIZE*5] ^= wq5; *(unative_t *)&q[d+NSIZE*6] ^= wq6; *(unative_t *)&q[d+NSIZE*7] ^= wq7; } } const struct raid6_calls raid6_intx8 = { raid6_int8_gen_syndrome, raid6_int8_xor_syndrome, NULL, /* always valid */ "int" NSTRING "x8", 0 };