/*
 * Licensed under the Apache License, Version 2.0 (the "License");
 * you may not use this file except in compliance with the License.
 * You may obtain a copy of the License at
 *
 *     http://www.apache.org/licenses/LICENSE-2.0
 *
 * Unless required by applicable law or agreed to in writing, software
 * distributed under the License is distributed on an "AS IS" BASIS,
 * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 * See the License for the specific language governing permissions and
 * limitations under the License.
 *
 * Copyright 2018 Danny Robson <danny@nerdcruft.net>
 */

#include "cpuid_x86.hpp"

#include "endian.hpp"
#include "view.hpp"
#include "format.hpp"
#include "bitwise.hpp"

#include <cstdint>
#include <ostream>

using util::cpu::x86;


///////////////////////////////////////////////////////////////////////////////
auto cpuid (int query, int param = 0)
{
    struct {
        uint32_t a, b, c, d;
    } res;

    asm (
        "cpuid"
        : "=a" (res.a), "=b" (res.b), "=c" (res.c), "=d" (res.d)
        :  "a" (query),  "c" (param)
    );

    return res;
}


///////////////////////////////////////////////////////////////////////////////
x86::x86 ()
{
    const auto &[largest_function, vendor0, vendor2, vendor1] = cpuid (0);
    (void)largest_function;

    memcpy (vendor_name.data () + 0, &vendor0, sizeof (vendor0));
    memcpy (vendor_name.data () + 4, &vendor1, sizeof (vendor1));
    memcpy (vendor_name.data () + 8, &vendor2, sizeof (vendor2));

    const bool is_amd = vendor0 == 0x68747541 && vendor1 == 0x69746e65 && vendor2 == 0x444d4163;

    {
        auto features = cpuid (1);

        simd.sse3  = features.c & (1u <<  0);
        simd.ssse3 = features.c & (1u <<  9);
        simd.sse41 = features.c & (1u << 19);
        simd.sse42 = features.c & (1u << 20);
        simd.avx   = features.c & (1u << 28);
        simd.sse   = features.d & (1u << 25);
        simd.sse2  = features.d & (1u << 26);

        cores.hyper_threading = features.d & (1u << 28);
        cores.logical = from_bits (features.b, 23, 16);
    }

    {
        const auto product0 = cpuid (0x8000'0002);
        const auto product1 = cpuid (0x8000'0003);
        const auto product2 = cpuid (0x8000'0004);

        memcpy (&product_name[0x00], &product0, sizeof (product0));
        memcpy (&product_name[0x10], &product1, sizeof (product1));
        memcpy (&product_name[0x20], &product2, sizeof (product2));
    }

    // Function 4 isn't implemented by AMD. Intel uses it for cache
    // descriptors. It has some useful information (like procesor counts) but
    // it's ugly as sin; try to avoid using it.
    // { (void)cpuid (0x0b, 4); }

    if (is_amd) {
        auto long_mode = cpuid (0x8000'0008);
        auto apic_id_size = from_bits (long_mode.c, 15, 12);
        (void)apic_id_size;
        cores.physical = 0; //from_bits (long_mode.c, 7, 0) + 1;
    }
}


///////////////////////////////////////////////////////////////////////////////
std::ostream&
util::cpu::operator<< (std::ostream &os, const x86 &val)
{
    return os << util::format::printf (
        "{ name: { vendor: '%!', product: '%!' }"
        ", cores: { logical: %!, physical: %!, hyper_threading: %! }"
        ", simd { sse: %!, sse2: %!, sse3: %!, ssse3: %!, sse41: %!, sse42: %! }"
        " }",
        util::view {val.vendor_name},
        util::view {val.product_name},
        val.cores.logical,
        val.cores.physical,
        val.cores.hyper_threading,
        val.simd.sse,
        val.simd.sse2,
        val.simd.sse3,
        val.simd.ssse3,
        val.simd.sse41,
        val.simd.sse42
    );
}