mirror of
https://github.com/amadvance/snapraid.git
synced 2026-11-22 16:06:30 -05:00
Preliminary support for AVX512GFNI
This commit is contained in:
parent
16a8c90bd8
commit
c080c3aa91
15 changed files with 958 additions and 80 deletions
49
raid/gf.h
49
raid/gf.h
|
|
@ -71,7 +71,39 @@ static __always_inline uint8_t A(int p, int d)
|
|||
#define v_64(p) (*(uint64_t *)&(p))
|
||||
|
||||
/*
|
||||
* Multiply each byte of a uint32 by 2 in the GF(2^8).
|
||||
* Polynomial-dependent XOR masks for GF(2^8) multiply and divide by 2.
|
||||
*
|
||||
* RAID_POLY_32 and RAID_POLY_64 are RAID_POLY repeated in every byte
|
||||
* of a uint32/uint64, used as the conditional XOR in the multiply-by-2
|
||||
* path: when the MSB of a byte is set before the left shift, the
|
||||
* reducing polynomial is XORed in to reduce the result back into the
|
||||
* field.
|
||||
*
|
||||
* RAID_INV2_32 and RAID_INV2_64 are the GF(2^8) inverse of 2 (i.e.
|
||||
* 2^{-1} mod RAID_POLY) repeated in every byte, used in the
|
||||
* divide-by-2 path. The inverse depends on the polynomial:
|
||||
*
|
||||
* RAID_POLY = 0x1d -> 2^{-1} = 0x8e (standard RAID)
|
||||
* RAID_POLY = 0x1b -> 2^{-1} = 0x8d (AES polynomial)
|
||||
*
|
||||
* Both are computed from RAID_POLY via the macro RAID_INV2_BYTE:
|
||||
* in GF(2^8), 2^{-1} is the value x such that 2*x = 1, which for
|
||||
* any polynomial of the form x^8 + ... + 1 resolves to:
|
||||
*
|
||||
* 2^{-1} = (RAID_POLY >> 1) | 0x80
|
||||
*
|
||||
* since right-shifting the polynomial by 1 and setting the MSB gives
|
||||
* the unique solution in both cases.
|
||||
*/
|
||||
#define RAID_INV2_BYTE (((RAID_POLY) >> 1) | 0x80)
|
||||
|
||||
#define RAID_POLY_32 ((uint32_t)RAID_POLY * 0x01010101U)
|
||||
#define RAID_POLY_64 ((uint64_t)RAID_POLY * 0x0101010101010101ULL)
|
||||
#define RAID_INV2_32 ((uint32_t)RAID_INV2_BYTE * 0x01010101U)
|
||||
#define RAID_INV2_64 ((uint64_t)RAID_INV2_BYTE * 0x0101010101010101ULL)
|
||||
|
||||
/*
|
||||
* Multiply each byte of a uint32 by 2 in GF(2^8).
|
||||
*/
|
||||
static __always_inline uint32_t x2_32(uint32_t v)
|
||||
{
|
||||
|
|
@ -79,12 +111,12 @@ static __always_inline uint32_t x2_32(uint32_t v)
|
|||
|
||||
mask = (mask << 1) - (mask >> 7);
|
||||
v = (v << 1) & 0xfefefefeU;
|
||||
v ^= mask & 0x1d1d1d1dU;
|
||||
v ^= mask & RAID_POLY_32;
|
||||
return v;
|
||||
}
|
||||
|
||||
/*
|
||||
* Multiply each byte of a uint64 by 2 in the GF(2^8).
|
||||
* Multiply each byte of a uint64 by 2 in GF(2^8).
|
||||
*/
|
||||
static __always_inline uint64_t x2_64(uint64_t v)
|
||||
{
|
||||
|
|
@ -92,12 +124,12 @@ static __always_inline uint64_t x2_64(uint64_t v)
|
|||
|
||||
mask = (mask << 1) - (mask >> 7);
|
||||
v = (v << 1) & 0xfefefefefefefefeULL;
|
||||
v ^= mask & 0x1d1d1d1d1d1d1d1dULL;
|
||||
v ^= mask & RAID_POLY_64;
|
||||
return v;
|
||||
}
|
||||
|
||||
/*
|
||||
* Divide each byte of a uint32 by 2 in the GF(2^8).
|
||||
* Divide each byte of a uint32 by 2 in GF(2^8).
|
||||
*/
|
||||
static __always_inline uint32_t d2_32(uint32_t v)
|
||||
{
|
||||
|
|
@ -105,12 +137,12 @@ static __always_inline uint32_t d2_32(uint32_t v)
|
|||
|
||||
mask = (mask << 8) - mask;
|
||||
v = (v >> 1) & 0x7f7f7f7fU;
|
||||
v ^= mask & 0x8e8e8e8eU;
|
||||
v ^= mask & RAID_INV2_32;
|
||||
return v;
|
||||
}
|
||||
|
||||
/*
|
||||
* Divide each byte of a uint64 by 2 in the GF(2^8).
|
||||
* Divide each byte of a uint64 by 2 in GF(2^8).
|
||||
*/
|
||||
static __always_inline uint64_t d2_64(uint64_t v)
|
||||
{
|
||||
|
|
@ -118,9 +150,8 @@ static __always_inline uint64_t d2_64(uint64_t v)
|
|||
|
||||
mask = (mask << 8) - mask;
|
||||
v = (v >> 1) & 0x7f7f7f7f7f7f7f7fULL;
|
||||
v ^= mask & 0x8e8e8e8e8e8e8e8eULL;
|
||||
v ^= mask & RAID_INV2_64;
|
||||
return v;
|
||||
}
|
||||
|
||||
#endif
|
||||
|
||||
|
|
|
|||
Loading…
Add table
Add a link
Reference in a new issue