path: root/security/nss/lib/freebl/gcm.c

/* This Source Code Form is subject to the terms of the Mozilla Public
 * License, v. 2.0. If a copy of the MPL was not distributed with this
 * file, You can obtain one at http://mozilla.org/MPL/2.0/. */
/* Thanks to Thomas Pornin for the ideas how to implement the constat time
 * binary multiplication. */

#ifdef FREEBL_NO_DEPEND
#include "stubs.h"
#endif
#include "blapii.h"
#include "blapit.h"
#include "gcm.h"
#include "ctr.h"
#include "secerr.h"
#include "prtypes.h"
#include "pkcs11t.h"

#include <limits.h>

#ifdef NSS_X86_OR_X64
#include <wmmintrin.h> /* clmul */
#endif

/* Forward declarations */
SECStatus gcm_HashMult_hw(gcmHashContext *ghash, const unsigned char *buf,
                          unsigned int count);
SECStatus gcm_HashMult_sftw(gcmHashContext *ghash, const unsigned char *buf,
                            unsigned int count);
SECStatus gcm_HashMult_sftw32(gcmHashContext *ghash, const unsigned char *buf,
                              unsigned int count);

uint64_t
get64(const unsigned char *bytes)
{
    return ((uint64_t)bytes[0]) << 56 |
           ((uint64_t)bytes[1]) << 48 |
           ((uint64_t)bytes[2]) << 40 |
           ((uint64_t)bytes[3]) << 32 |
           ((uint64_t)bytes[4]) << 24 |
           ((uint64_t)bytes[5]) << 16 |
           ((uint64_t)bytes[6]) << 8 |
           ((uint64_t)bytes[7]);
}

/* Initialize a gcmHashContext */
SECStatus
gcmHash_InitContext(gcmHashContext *ghash, const unsigned char *H, PRBool sw)
{
    ghash->cLen = 0;
    ghash->bufLen = 0;
    PORT_Memset(ghash->counterBuf, 0, sizeof(ghash->counterBuf));

    ghash->h_low = get64(H + 8);
    ghash->h_high = get64(H);
    if (clmul_support() && !sw) {
#ifdef NSS_X86_OR_X64
        ghash->ghash_mul = gcm_HashMult_hw;
        ghash->x = _mm_setzero_si128();
        /* MSVC requires __m64 to load epi64. */
        ghash->h = _mm_set_epi32(ghash->h_high >> 32, (uint32_t)ghash->h_high,
                                 ghash->h_low >> 32, (uint32_t)ghash->h_low);
        ghash->hw = PR_TRUE;
#else
        PORT_SetError(SEC_ERROR_LIBRARY_FAILURE);
        return SECFailure;
#endif /* NSS_X86_OR_X64 */
    } else {
/* We fall back to the software implementation if we can't use / don't
         * want to use pclmul. */
#ifdef HAVE_INT128_SUPPORT
        ghash->ghash_mul = gcm_HashMult_sftw;
#else
        ghash->ghash_mul = gcm_HashMult_sftw32;
#endif
        ghash->x_high = ghash->x_low = 0;
        ghash->hw = PR_FALSE;
    }
    return SECSuccess;
}

#ifdef HAVE_INT128_SUPPORT
/* Binary multiplication x * y = r_high << 64 | r_low. */
void
bmul(uint64_t x, uint64_t y, uint64_t *r_high, uint64_t *r_low)
{
    uint128_t x1, x2, x3, x4, x5;
    uint128_t y1, y2, y3, y4, y5;
    uint128_t r, z;

    uint128_t m1 = (uint128_t)0x2108421084210842 << 64 | 0x1084210842108421;
    uint128_t m2 = (uint128_t)0x4210842108421084 << 64 | 0x2108421084210842;
    uint128_t m3 = (uint128_t)0x8421084210842108 << 64 | 0x4210842108421084;
    uint128_t m4 = (uint128_t)0x0842108421084210 << 64 | 0x8421084210842108;
    uint128_t m5 = (uint128_t)0x1084210842108421 << 64 | 0x0842108421084210;

    x1 = x & m1;
    y1 = y & m1;
    x2 = x & m2;
    y2 = y & m2;
    x3 = x & m3;
    y3 = y & m3;
    x4 = x & m4;
    y4 = y & m4;
    x5 = x & m5;
    y5 = y & m5;

    z = (x1 * y1) ^ (x2 * y5) ^ (x3 * y4) ^ (x4 * y3) ^ (x5 * y2);
    r = z & m1;
    z = (x1 * y2) ^ (x2 * y1) ^ (x3 * y5) ^ (x4 * y4) ^ (x5 * y3);
    r |= z & m2;
    z = (x1 * y3) ^ (x2 * y2) ^ (x3 * y1) ^ (x4 * y5) ^ (x5 * y4);
    r |= z & m3;
    z = (x1 * y4) ^ (x2 * y3) ^ (x3 * y2) ^ (x4 * y1) ^ (x5 * y5);
    r |= z & m4;
    z = (x1 * y5) ^ (x2 * y4) ^ (x3 * y3) ^ (x4 * y2) ^ (x5 * y1);
    r |= z & m5;

    *r_high = (uint64_t)(r >> 64);
    *r_low = (uint64_t)r;
}

SECStatus
gcm_HashMult_sftw(gcmHashContext *ghash, const unsigned char *buf,
                  unsigned int count)
{
    uint64_t ci_low, ci_high;
    size_t i;
    uint64_t z2_low, z2_high, z0_low, z0_high, z1a_low, z1a_high;
    uint128_t z_high = 0, z_low = 0;

    ci_low = ghash->x_low;
    ci_high = ghash->x_high;
    for (i = 0; i < count; i++, buf += 16) {
        ci_low ^= get64(buf + 8);
        ci_high ^= get64(buf);

        /* Do binary mult ghash->X = C * ghash->H (Karatsuba). */
        bmul(ci_high, ghash->h_high, &z2_high, &z2_low);
        bmul(ci_low, ghash->h_low, &z0_high, &z0_low);
        bmul(ci_high ^ ci_low, ghash->h_high ^ ghash->h_low, &z1a_high, &z1a_low);
        z1a_high ^= z2_high ^ z0_high;
        z1a_low ^= z2_low ^ z0_low;
        z_high = ((uint128_t)z2_high << 64) | (z2_low ^ z1a_high);
        z_low = (((uint128_t)z0_high << 64) | z0_low) ^ (((uint128_t)z1a_low) << 64);

        /* Shift one (multiply by x) as gcm spec is stupid. */
        z_high = (z_high << 1) | (z_low >> 127);
        z_low <<= 1;

        /* Reduce */
        z_low ^= (z_low << 127) ^ (z_low << 126) ^ (z_low << 121);
        z_high ^= z_low ^ (z_low >> 1) ^ (z_low >> 2) ^ (z_low >> 7);
        ci_low = (uint64_t)z_high;
        ci_high = (uint64_t)(z_high >> 64);
    }
    ghash->x_low = ci_low;
    ghash->x_high = ci_high;
    return SECSuccess;
}
#else
/* Binary multiplication x * y = r_high << 32 | r_low. */
void
bmul32(uint32_t x, uint32_t y, uint32_t *r_high, uint32_t *r_low)
{
    uint32_t x0, x1, x2, x3;
    uint32_t y0, y1, y2, y3;
    uint32_t m1 = (uint32_t)0x11111111;
    uint32_t m2 = (uint32_t)0x22222222;
    uint32_t m4 = (uint32_t)0x44444444;
    uint32_t m8 = (uint32_t)0x88888888;
    uint64_t z0, z1, z2, z3;
    uint64_t z;

    x0 = x & m1;
    x1 = x & m2;
    x2 = x & m4;
    x3 = x & m8;
    y0 = y & m1;
    y1 = y & m2;
    y2 = y & m4;
    y3 = y & m8;
    z0 = ((uint64_t)x0 * y0) ^ ((uint64_t)x1 * y3) ^
         ((uint64_t)x2 * y2) ^ ((uint64_t)x3 * y1);
    z1 = ((uint64_t)x0 * y1) ^ ((uint64_t)x1 * y0) ^
         ((uint64_t)x2 * y3) ^ ((uint64_t)x3 * y2);
    z2 = ((uint64_t)x0 * y2) ^ ((uint64_t)x1 * y1) ^
         ((uint64_t)x2 * y0) ^ ((uint64_t)x3 * y3);
    z3 = ((uint64_t)x0 * y3) ^ ((uint64_t)x1 * y2) ^
         ((uint64_t)x2 * y1) ^ ((uint64_t)x3 * y0);
    z0 &= ((uint64_t)m1 << 32) | m1;
    z1 &= ((uint64_t)m2 << 32) | m2;
    z2 &= ((uint64_t)m4 << 32) | m4;
    z3 &= ((uint64_t)m8 << 32) | m8;
    z = z0 | z1 | z2 | z3;
    *r_high = (uint32_t)(z >> 32);
    *r_low = (uint32_t)z;
}

SECStatus
gcm_HashMult_sftw32(gcmHashContext *ghash, const unsigned char *buf,
                    unsigned int count)
{
    size_t i;
    uint64_t ci_low, ci_high;
    uint64_t z_high_h, z_high_l, z_low_h, z_low_l;
    uint32_t ci_high_h, ci_high_l, ci_low_h, ci_low_l;
    uint32_t b_a_h, b_a_l, a_a_h, a_a_l, b_b_h, b_b_l;
    uint32_t a_b_h, a_b_l, b_c_h, b_c_l, a_c_h, a_c_l, c_c_h, c_c_l;
    uint32_t ci_highXlow_h, ci_highXlow_l, c_a_h, c_a_l, c_b_h, c_b_l;

    uint32_t h_high_h = (uint32_t)(ghash->h_high >> 32);
    uint32_t h_high_l = (uint32_t)ghash->h_high;
    uint32_t h_low_h = (uint32_t)(ghash->h_low >> 32);
    uint32_t h_low_l = (uint32_t)ghash->h_low;
    uint32_t h_highXlow_h = h_high_h ^ h_low_h;
    uint32_t h_highXlow_l = h_high_l ^ h_low_l;
    uint32_t h_highX_xored = h_highXlow_h ^ h_highXlow_l;

    for (i = 0; i < count; i++, buf += 16) {
        ci_low = ghash->x_low ^ get64(buf + 8);
        ci_high = ghash->x_high ^ get64(buf);
        ci_low_h = (uint32_t)(ci_low >> 32);
        ci_low_l = (uint32_t)ci_low;
        ci_high_h = (uint32_t)(ci_high >> 32);
        ci_high_l = (uint32_t)ci_high;
        ci_highXlow_h = ci_high_h ^ ci_low_h;
        ci_highXlow_l = ci_high_l ^ ci_low_l;

        /* Do binary mult ghash->X = C * ghash->H (recursive Karatsuba). */
        bmul32(ci_high_h, h_high_h, &a_a_h, &a_a_l);
        bmul32(ci_high_l, h_high_l, &a_b_h, &a_b_l);
        bmul32(ci_high_h ^ ci_high_l, h_high_h ^ h_high_l, &a_c_h, &a_c_l);
        a_c_h ^= a_a_h ^ a_b_h;
        a_c_l ^= a_a_l ^ a_b_l;
        a_a_l ^= a_c_h;
        a_b_h ^= a_c_l;
        /* ci_high * h_high = a_a_h:a_a_l:a_b_h:a_b_l */

        bmul32(ci_low_h, h_low_h, &b_a_h, &b_a_l);
        bmul32(ci_low_l, h_low_l, &b_b_h, &b_b_l);
        bmul32(ci_low_h ^ ci_low_l, h_low_h ^ h_low_l, &b_c_h, &b_c_l);
        b_c_h ^= b_a_h ^ b_b_h;
        b_c_l ^= b_a_l ^ b_b_l;
        b_a_l ^= b_c_h;
        b_b_h ^= b_c_l;
        /* ci_low * h_low = b_a_h:b_a_l:b_b_h:b_b_l */

        bmul32(ci_highXlow_h, h_highXlow_h, &c_a_h, &c_a_l);
        bmul32(ci_highXlow_l, h_highXlow_l, &c_b_h, &c_b_l);
        bmul32(ci_highXlow_h ^ ci_highXlow_l, h_highX_xored, &c_c_h, &c_c_l);
        c_c_h ^= c_a_h ^ c_b_h;
        c_c_l ^= c_a_l ^ c_b_l;
        c_a_l ^= c_c_h;
        c_b_h ^= c_c_l;
        /* (ci_high ^ ci_low) * (h_high ^ h_low) = c_a_h:c_a_l:c_b_h:c_b_l */

        c_a_h ^= b_a_h ^ a_a_h;
        c_a_l ^= b_a_l ^ a_a_l;
        c_b_h ^= b_b_h ^ a_b_h;
        c_b_l ^= b_b_l ^ a_b_l;
        z_high_h = ((uint64_t)a_a_h << 32) | a_a_l;
        z_high_l = (((uint64_t)a_b_h << 32) | a_b_l) ^
                   (((uint64_t)c_a_h << 32) | c_a_l);
        z_low_h = (((uint64_t)b_a_h << 32) | b_a_l) ^
                  (((uint64_t)c_b_h << 32) | c_b_l);
        z_low_l = ((uint64_t)b_b_h << 32) | b_b_l;

        /* Shift one (multiply by x) as gcm spec is stupid. */
        z_high_h = z_high_h << 1 | z_high_l >> 63;
        z_high_l = z_high_l << 1 | z_low_h >> 63;
        z_low_h = z_low_h << 1 | z_low_l >> 63;
        z_low_l <<= 1;

        /* Reduce */
        z_low_h ^= (z_low_l << 63) ^ (z_low_l << 62) ^ (z_low_l << 57);
        z_high_h ^= z_low_h ^ (z_low_h >> 1) ^ (z_low_h >> 2) ^ (z_low_h >> 7);
        z_high_l ^= z_low_l ^ (z_low_l >> 1) ^ (z_low_l >> 2) ^ (z_low_l >> 7) ^
                    (z_low_h << 63) ^ (z_low_h << 62) ^ (z_low_h << 57);
        ghash->x_high = z_high_h;
        ghash->x_low = z_high_l;
    }
    return SECSuccess;
}
#endif /* HAVE_INT128_SUPPORT */

SECStatus
gcm_HashMult_hw(gcmHashContext *ghash, const unsigned char *buf,
                unsigned int count)
{
#ifdef NSS_X86_OR_X64
    size_t i;
    pre_align __m128i z_high post_align;
    pre_align __m128i z_low post_align;
    pre_align __m128i C post_align;
    pre_align __m128i D post_align;
    pre_align __m128i E post_align;
    pre_align __m128i F post_align;
    pre_align __m128i bin post_align;
    pre_align __m128i Ci post_align;
    pre_align __m128i tmp post_align;

    for (i = 0; i < count; i++, buf += 16) {
        bin = _mm_set_epi16(((uint16_t)buf[0] << 8) | buf[1],
                            ((uint16_t)buf[2] << 8) | buf[3],
                            ((uint16_t)buf[4] << 8) | buf[5],
                            ((uint16_t)buf[6] << 8) | buf[7],
                            ((uint16_t)buf[8] << 8) | buf[9],
                            ((uint16_t)buf[10] << 8) | buf[11],
                            ((uint16_t)buf[12] << 8) | buf[13],
                            ((uint16_t)buf[14] << 8) | buf[15]);
        Ci = _mm_xor_si128(bin, ghash->x);

        /* Do binary mult ghash->X = Ci * ghash->H. */
        C = _mm_clmulepi64_si128(Ci, ghash->h, 0x00);
        D = _mm_clmulepi64_si128(Ci, ghash->h, 0x11);
        E = _mm_clmulepi64_si128(Ci, ghash->h, 0x01);
        F = _mm_clmulepi64_si128(Ci, ghash->h, 0x10);
        tmp = _mm_xor_si128(E, F);
        z_high = _mm_xor_si128(tmp, _mm_slli_si128(D, 8));
        z_high = _mm_unpackhi_epi64(z_high, D);
        z_low = _mm_xor_si128(_mm_slli_si128(tmp, 8), C);
        z_low = _mm_unpackhi_epi64(_mm_slli_si128(C, 8), z_low);

        /* Shift one to the left (multiply by x) as gcm spec is stupid. */
        C = _mm_slli_si128(z_low, 8);
        E = _mm_srli_epi64(C, 63);
        D = _mm_slli_si128(z_high, 8);
        F = _mm_srli_epi64(D, 63);
        /* Carry over */
        C = _mm_srli_si128(z_low, 8);
        D = _mm_srli_epi64(C, 63);
        z_low = _mm_or_si128(_mm_slli_epi64(z_low, 1), E);
        z_high = _mm_or_si128(_mm_or_si128(_mm_slli_epi64(z_high, 1), F), D);

        /* Reduce */
        C = _mm_slli_si128(z_low, 8);
        /* D = z_low << 127 */
        D = _mm_slli_epi64(C, 63);
        /* E = z_low << 126 */
        E = _mm_slli_epi64(C, 62);
        /* F = z_low << 121 */
        F = _mm_slli_epi64(C, 57);
        /* z_low ^= (z_low << 127) ^ (z_low << 126) ^ (z_low << 121); */
        z_low = _mm_xor_si128(_mm_xor_si128(_mm_xor_si128(z_low, D), E), F);
        C = _mm_srli_si128(z_low, 8);
        /* D = z_low >> 1 */
        D = _mm_slli_epi64(C, 63);
        D = _mm_or_si128(_mm_srli_epi64(z_low, 1), D);
        /* E = z_low >> 2 */
        E = _mm_slli_epi64(C, 62);
        E = _mm_or_si128(_mm_srli_epi64(z_low, 2), E);
        /* F = z_low >> 7 */
        F = _mm_slli_epi64(C, 57);
        F = _mm_or_si128(_mm_srli_epi64(z_low, 7), F);
        /* ghash->x ^= z_low ^ (z_low >> 1) ^ (z_low >> 2) ^ (z_low >> 7); */
        ghash->x = _mm_xor_si128(_mm_xor_si128(
                                     _mm_xor_si128(_mm_xor_si128(z_high, z_low), D), E),
                                 F);
    }
    return SECSuccess;
#else
    PORT_SetError(SEC_ERROR_LIBRARY_FAILURE);
    return SECFailure;
#endif /* NSS_X86_OR_X64 */
}

static SECStatus
gcm_zeroX(gcmHashContext *ghash)
{
    if (ghash->hw) {
#ifdef NSS_X86_OR_X64
        ghash->x = _mm_setzero_si128();
        return SECSuccess;
#else
        PORT_SetError(SEC_ERROR_LIBRARY_FAILURE);
        return SECFailure;
#endif /* NSS_X86_OR_X64 */
    }

    ghash->x_high = ghash->x_low = 0;
    return SECSuccess;
}

/*
 * implement GCM GHASH using the freebl GHASH function. The gcm_HashMult
 * function always takes AES_BLOCK_SIZE lengths of data. gcmHash_Update will
 * format the data properly.
 */
SECStatus
gcmHash_Update(gcmHashContext *ghash, const unsigned char *buf,
               unsigned int len)
{
    unsigned int blocks;
    SECStatus rv;

    ghash->cLen += (len * PR_BITS_PER_BYTE);

    /* first deal with the current buffer of data. Try to fill it out so
     * we can hash it */
    if (ghash->bufLen) {
        unsigned int needed = PR_MIN(len, AES_BLOCK_SIZE - ghash->bufLen);
        if (needed != 0) {
            PORT_Memcpy(ghash->buffer + ghash->bufLen, buf, needed);
        }
        buf += needed;
        len -= needed;
        ghash->bufLen += needed;
        if (len == 0) {
            /* didn't add enough to hash the data, nothing more do do */
            return SECSuccess;
        }
        PORT_Assert(ghash->bufLen == AES_BLOCK_SIZE);
        /* hash the buffer and clear it */
        rv = ghash->ghash_mul(ghash, ghash->buffer, 1);
        PORT_Memset(ghash->buffer, 0, AES_BLOCK_SIZE);
        ghash->bufLen = 0;
        if (rv != SECSuccess) {
            return SECFailure;
        }
    }
    /* now hash any full blocks remaining in the data stream */
    blocks = len / AES_BLOCK_SIZE;
    if (blocks) {
        rv = ghash->ghash_mul(ghash, buf, blocks);
        if (rv != SECSuccess) {
            return SECFailure;
        }
        buf += blocks * AES_BLOCK_SIZE;
        len -= blocks * AES_BLOCK_SIZE;
    }

    /* save any remainder in the buffer to be hashed with the next call */
    if (len != 0) {
        PORT_Memcpy(ghash->buffer, buf, len);
        ghash->bufLen = len;
    }
    return SECSuccess;
}

/*
 * write out any partial blocks zero padded through the GHASH engine,
 * save the lengths for the final completion of the hash
 */
static SECStatus
gcmHash_Sync(gcmHashContext *ghash)
{
    int i;
    SECStatus rv;

    /* copy the previous counter to the upper block */
    PORT_Memcpy(ghash->counterBuf, &ghash->counterBuf[GCM_HASH_LEN_LEN],
                GCM_HASH_LEN_LEN);
    /* copy the current counter in the lower block */
    for (i = 0; i < GCM_HASH_LEN_LEN; i++) {
        ghash->counterBuf[GCM_HASH_LEN_LEN + i] =
            (ghash->cLen >> ((GCM_HASH_LEN_LEN - 1 - i) * PR_BITS_PER_BYTE)) & 0xff;
    }
    ghash->cLen = 0;

    /* now zero fill the buffer and hash the last block */
    if (ghash->bufLen) {
        PORT_Memset(ghash->buffer + ghash->bufLen, 0, AES_BLOCK_SIZE - ghash->bufLen);
        rv = ghash->ghash_mul(ghash, ghash->buffer, 1);
        PORT_Memset(ghash->buffer, 0, AES_BLOCK_SIZE);
        ghash->bufLen = 0;
        if (rv != SECSuccess) {
            return SECFailure;
        }
    }
    return SECSuccess;
}

#define WRITE64(x, bytes)   \
    (bytes)[0] = (x) >> 56; \
    (bytes)[1] = (x) >> 48; \
    (bytes)[2] = (x) >> 40; \
    (bytes)[3] = (x) >> 32; \
    (bytes)[4] = (x) >> 24; \
    (bytes)[5] = (x) >> 16; \
    (bytes)[6] = (x) >> 8;  \
    (bytes)[7] = (x);

/*
 * This does the final sync, hashes the lengths, then returns
 * "T", the hashed output.
 */
SECStatus
gcmHash_Final(gcmHashContext *ghash, unsigned char *outbuf,
              unsigned int *outlen, unsigned int maxout)
{
    unsigned char T[MAX_BLOCK_SIZE];
    SECStatus rv;

    rv = gcmHash_Sync(ghash);
    if (rv != SECSuccess) {
        goto cleanup;
    }

    rv = ghash->ghash_mul(ghash, ghash->counterBuf,
                          (GCM_HASH_LEN_LEN * 2) / AES_BLOCK_SIZE);
    if (rv != SECSuccess) {
        goto cleanup;
    }

    if (ghash->hw) {
#ifdef NSS_X86_OR_X64
        uint64_t tmp_out[2];
        _mm_storeu_si128((__m128i *)tmp_out, ghash->x);
        WRITE64(tmp_out[0], T + 8);
        WRITE64(tmp_out[1], T);
#else
        PORT_SetError(SEC_ERROR_LIBRARY_FAILURE);
        return SECFailure;
#endif /* NSS_X86_OR_X64 */
    } else {
        WRITE64(ghash->x_low, T + 8);
        WRITE64(ghash->x_high, T);
    }

    if (maxout > AES_BLOCK_SIZE) {
        maxout = AES_BLOCK_SIZE;
    }
    PORT_Memcpy(outbuf, T, maxout);
    *outlen = maxout;
    rv = SECSuccess;

cleanup:
    PORT_Memset(T, 0, sizeof(T));
    return rv;
}

SECStatus
gcmHash_Reset(gcmHashContext *ghash, const unsigned char *AAD,
              unsigned int AADLen)
{
    SECStatus rv;

    ghash->cLen = 0;
    PORT_Memset(ghash->counterBuf, 0, GCM_HASH_LEN_LEN * 2);
    ghash->bufLen = 0;
    rv = gcm_zeroX(ghash);
    if (rv != SECSuccess) {
        return rv;
    }

    /* now kick things off by hashing the Additional Authenticated Data */
    if (AADLen != 0) {
        rv = gcmHash_Update(ghash, AAD, AADLen);
        if (rv != SECSuccess) {
            return SECFailure;
        }
        rv = gcmHash_Sync(ghash);
        if (rv != SECSuccess) {
            return SECFailure;
        }
    }
    return SECSuccess;
}

/**************************************************************************
 *           Now implement the GCM using gcmHash and CTR                  *
 **************************************************************************/

/* state to handle the full GCM operation (hash and counter) */
struct GCMContextStr {
    gcmHashContext *ghash_context;
    CTRContext ctr_context;
    unsigned long tagBits;
    unsigned char tagKey[MAX_BLOCK_SIZE];
};

GCMContext *
GCM_CreateContext(void *context, freeblCipherFunc cipher,
                  const unsigned char *params)
{
    GCMContext *gcm = NULL;
    gcmHashContext *ghash = NULL;
    unsigned char H[MAX_BLOCK_SIZE];
    unsigned int tmp;
    PRBool freeCtr = PR_FALSE;
    const CK_GCM_PARAMS *gcmParams = (const CK_GCM_PARAMS *)params;
    CK_AES_CTR_PARAMS ctrParams;
    SECStatus rv;
#ifdef DISABLE_HW_GCM
    const PRBool sw = PR_TRUE;
#else
    const PRBool sw = PR_FALSE;
#endif

    if (gcmParams->ulIvLen == 0) {
        PORT_SetError(SEC_ERROR_INVALID_ARGS);
        return NULL;
    }
    gcm = PORT_ZNew(GCMContext);
    if (gcm == NULL) {
        return NULL;
    }
    /* aligned_alloc is C11 so we have to do it the old way. */
    ghash = PORT_ZAlloc(sizeof(gcmHashContext) + 15);
    if (ghash == NULL) {
        PORT_SetError(SEC_ERROR_NO_MEMORY);
        goto loser;
    }
    ghash->mem = ghash;
    ghash = (gcmHashContext *)(((uintptr_t)ghash + 15) & ~(uintptr_t)0x0F);

    /* first plug in the ghash context */
    gcm->ghash_context = ghash;
    PORT_Memset(H, 0, AES_BLOCK_SIZE);
    rv = (*cipher)(context, H, &tmp, AES_BLOCK_SIZE, H, AES_BLOCK_SIZE, AES_BLOCK_SIZE);
    if (rv != SECSuccess) {
        goto loser;
    }
    rv = gcmHash_InitContext(ghash, H, sw);
    if (rv != SECSuccess) {
        goto loser;
    }

    /* fill in the Counter context */
    ctrParams.ulCounterBits = 32;
    PORT_Memset(ctrParams.cb, 0, sizeof(ctrParams.cb));
    if (gcmParams->ulIvLen == 12) {
        PORT_Memcpy(ctrParams.cb, gcmParams->pIv, gcmParams->ulIvLen);
        ctrParams.cb[AES_BLOCK_SIZE - 1] = 1;
    } else {
        rv = gcmHash_Update(ghash, gcmParams->pIv, gcmParams->ulIvLen);
        if (rv != SECSuccess) {
            goto loser;
        }
        rv = gcmHash_Final(ghash, ctrParams.cb, &tmp, AES_BLOCK_SIZE);
        if (rv != SECSuccess) {
            goto loser;
        }
    }
    rv = CTR_InitContext(&gcm->ctr_context, context, cipher,
                         (unsigned char *)&ctrParams);
    if (rv != SECSuccess) {
        goto loser;
    }
    freeCtr = PR_TRUE;

    /* fill in the gcm structure */
    gcm->tagBits = gcmParams->ulTagBits; /* save for final step */
    /* calculate the final tag key. NOTE: gcm->tagKey is zero to start with.
     * if this assumption changes, we would need to explicitly clear it here */
    rv = CTR_Update(&gcm->ctr_context, gcm->tagKey, &tmp, AES_BLOCK_SIZE,
                    gcm->tagKey, AES_BLOCK_SIZE, AES_BLOCK_SIZE);
    if (rv != SECSuccess) {
        goto loser;
    }

    /* finally mix in the AAD data */
    rv = gcmHash_Reset(ghash, gcmParams->pAAD, gcmParams->ulAADLen);
    if (rv != SECSuccess) {
        goto loser;
    }

    return gcm;

loser:
    if (freeCtr) {
        CTR_DestroyContext(&gcm->ctr_context, PR_FALSE);
    }
    if (ghash && ghash->mem) {
        PORT_Free(ghash->mem);
    }
    if (gcm) {
        PORT_Free(gcm);
    }
    return NULL;
}

void
GCM_DestroyContext(GCMContext *gcm, PRBool freeit)
{
    /* these two are statically allocated and will be freed when we free
     * gcm. call their destroy functions to free up any locally
     * allocated data (like mp_int's) */
    CTR_DestroyContext(&gcm->ctr_context, PR_FALSE);
    PORT_Free(gcm->ghash_context->mem);
    PORT_Memset(&gcm->tagBits, 0, sizeof(gcm->tagBits));
    PORT_Memset(gcm->tagKey, 0, sizeof(gcm->tagKey));
    if (freeit) {
        PORT_Free(gcm);
    }
}

static SECStatus
gcm_GetTag(GCMContext *gcm, unsigned char *outbuf,
           unsigned int *outlen, unsigned int maxout)
{
    unsigned int tagBytes;
    unsigned int extra;
    unsigned int i;
    SECStatus rv;

    tagBytes = (gcm->tagBits + (PR_BITS_PER_BYTE - 1)) / PR_BITS_PER_BYTE;
    extra = tagBytes * PR_BITS_PER_BYTE - gcm->tagBits;

    if (outbuf == NULL) {
        *outlen = tagBytes;
        PORT_SetError(SEC_ERROR_OUTPUT_LEN);
        return SECFailure;
    }

    if (maxout < tagBytes) {
        *outlen = tagBytes;
        PORT_SetError(SEC_ERROR_OUTPUT_LEN);
        return SECFailure;
    }
    maxout = tagBytes;
    rv = gcmHash_Final(gcm->ghash_context, outbuf, outlen, maxout);
    if (rv != SECSuccess) {
        return SECFailure;
    }

    for (i = 0; i < *outlen; i++) {
        outbuf[i] ^= gcm->tagKey[i];
    }
    /* mask off any extra bits we got */
    if (extra) {
        outbuf[tagBytes - 1] &= ~((1 << extra) - 1);
    }
    return SECSuccess;
}

/*
 * See The Galois/Counter Mode of Operation, McGrew and Viega.
 *  GCM is basically counter mode with a specific initialization and
 *  built in macing operation.
 */
SECStatus
GCM_EncryptUpdate(GCMContext *gcm, unsigned char *outbuf,
                  unsigned int *outlen, unsigned int maxout,
                  const unsigned char *inbuf, unsigned int inlen,
                  unsigned int blocksize)
{
    SECStatus rv;
    unsigned int tagBytes;
    unsigned int len;

    PORT_Assert(blocksize == AES_BLOCK_SIZE);
    if (blocksize != AES_BLOCK_SIZE) {
        PORT_SetError(SEC_ERROR_LIBRARY_FAILURE);
        return SECFailure;
    }

    tagBytes = (gcm->tagBits + (PR_BITS_PER_BYTE - 1)) / PR_BITS_PER_BYTE;
    if (UINT_MAX - inlen < tagBytes) {
        PORT_SetError(SEC_ERROR_INPUT_LEN);
        return SECFailure;
    }
    if (maxout < inlen + tagBytes) {
        *outlen = inlen + tagBytes;
        PORT_SetError(SEC_ERROR_OUTPUT_LEN);
        return SECFailure;
    }

    rv = CTR_Update(&gcm->ctr_context, outbuf, outlen, maxout,
                    inbuf, inlen, AES_BLOCK_SIZE);
    if (rv != SECSuccess) {
        return SECFailure;
    }
    rv = gcmHash_Update(gcm->ghash_context, outbuf, *outlen);
    if (rv != SECSuccess) {
        PORT_Memset(outbuf, 0, *outlen); /* clear the output buffer */
        *outlen = 0;
        return SECFailure;
    }
    rv = gcm_GetTag(gcm, outbuf + *outlen, &len, maxout - *outlen);
    if (rv != SECSuccess) {
        PORT_Memset(outbuf, 0, *outlen); /* clear the output buffer */
        *outlen = 0;
        return SECFailure;
    };
    *outlen += len;
    return SECSuccess;
}

/*
 * See The Galois/Counter Mode of Operation, McGrew and Viega.
 *  GCM is basically counter mode with a specific initialization and
 *  built in macing operation. NOTE: the only difference between Encrypt
 *  and Decrypt is when we calculate the mac. That is because the mac must
 *  always be calculated on the cipher text, not the plain text, so for
 *  encrypt, we do the CTR update first and for decrypt we do the mac first.
 */
SECStatus
GCM_DecryptUpdate(GCMContext *gcm, unsigned char *outbuf,
                  unsigned int *outlen, unsigned int maxout,
                  const unsigned char *inbuf, unsigned int inlen,
                  unsigned int blocksize)
{
    SECStatus rv;
    unsigned int tagBytes;
    unsigned char tag[MAX_BLOCK_SIZE];
    const unsigned char *intag;
    unsigned int len;

    PORT_Assert(blocksize == AES_BLOCK_SIZE);
    if (blocksize != AES_BLOCK_SIZE) {
        PORT_SetError(SEC_ERROR_LIBRARY_FAILURE);
        return SECFailure;
    }

    tagBytes = (gcm->tagBits + (PR_BITS_PER_BYTE - 1)) / PR_BITS_PER_BYTE;

    /* get the authentication block */
    if (inlen < tagBytes) {
        PORT_SetError(SEC_ERROR_INPUT_LEN);
        return SECFailure;
    }

    inlen -= tagBytes;
    intag = inbuf + inlen;

    /* verify the block */
    rv = gcmHash_Update(gcm->ghash_context, inbuf, inlen);
    if (rv != SECSuccess) {
        return SECFailure;
    }
    rv = gcm_GetTag(gcm, tag, &len, AES_BLOCK_SIZE);
    if (rv != SECSuccess) {
        return SECFailure;
    }
    /* Don't decrypt if we can't authenticate the encrypted data!
     * This assumes that if tagBits is not a multiple of 8, intag will
     * preserve the masked off missing bits.  */
    if (NSS_SecureMemcmp(tag, intag, tagBytes) != 0) {
        /* force a CKR_ENCRYPTED_DATA_INVALID error at in softoken */
        PORT_SetError(SEC_ERROR_BAD_DATA);
        PORT_Memset(tag, 0, sizeof(tag));
        return SECFailure;
    }
    PORT_Memset(tag, 0, sizeof(tag));
    /* finish the decryption */
    return CTR_Update(&gcm->ctr_context, outbuf, outlen, maxout,
                      inbuf, inlen, AES_BLOCK_SIZE);
}