onednn-src 0.1.13

Source of oneAPI Deep Neural Network Library (oneDNN)
Documentation 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234

/*******************************************************************************
* Copyright 2024 Intel Corporation
*
* 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.
*******************************************************************************/

#include <cstdint>
#include <cstring>
#include <fstream>
#include <stdexcept>
#include <utility>
#include <vector>

#include "generator/microkernel/elf.hpp"
#include "gemmstone/microkernel/fuser.hpp"

GEMMSTONE_NAMESPACE_START
namespace microkernel {

static void fixupJumpTargets(uint8_t *start, size_t len, ptrdiff_t adjust);

void fuse(std::vector<uint8_t> &binary,
        const std::vector<uint8_t> &microkernel, long id) {
    auto base = binary.data();
    auto bytes = binary.size();

    auto fheaderPtr = reinterpret_cast<FileHeader *>(base);

    bool ok = bytes >= sizeof(fheaderPtr) && fheaderPtr->magic == ELFMagic
            && fheaderPtr->elfClass == ELFClass64
            && fheaderPtr->endian == ELFLittleEndian
            && fheaderPtr->sectionHeaderSize == sizeof(SectionHeader)
            && (fheaderPtr->version == 0 || fheaderPtr->version == ELFVersion1)
            && (fheaderPtr->type == ZebinExec
                    || fheaderPtr->type == ELFRelocatable)
            && bytes >= sizeof(fheaderPtr)
                            + sizeof(SectionHeader) * fheaderPtr->sectionCount;

    if (!ok)
        throw std::runtime_error(
                "IGC did not generate a valid zebin program binary");

    bool foundZeInfo = false;
    const char *snames = nullptr;
    std::vector<std::pair<SectionHeader *, int>> textSections;

    auto *sheaders = reinterpret_cast<SectionHeader *>(
            base + fheaderPtr->sectionTableOff);

    snames = reinterpret_cast<char *>(
            base + sheaders[fheaderPtr->strTableIndex].offset);

    for (int s = 0; s < fheaderPtr->sectionCount; s++) {
        switch (sheaders[s].type) {
            case SectionHeader::Type::ZeInfo: foundZeInfo = true; break;
            case SectionHeader::Type::Program: {
                if (snames) {
                    std::string sname(snames + sheaders[s].name);
                    if (sname == ".text.Intel_Symbol_Table_Void_Program")
                        continue;
                    if (sname.substr(0, 6) != ".text.") continue;
                }
                textSections.emplace_back(sheaders + s, s);
                break;
            }
            default: break;
        }
    }

    if (!foundZeInfo || textSections.empty())
        throw std::runtime_error(
                "IGC did not generate a valid zebin program binary");

    for (auto &entry : textSections) {
        auto *text = entry.first;
        int textSectionID = entry.second;
        if (text->offset + text->size > bytes) continue;

        auto *insn = reinterpret_cast<const uint32_t *>(base + text->offset);
        auto *iend = reinterpret_cast<const uint32_t *>(
                base + text->offset + text->size);

        const uint8_t *spliceStart = nullptr;
        const uint8_t *spliceEnd = nullptr;

        for (; insn < iend; insn += 4) {
            if (insn[0] & (1u << 29))
                insn -= 2;
            else if (insn[3] == (sigilStart ^ id))
                spliceStart = reinterpret_cast<const uint8_t *>(insn);
            else if (insn[3] == (sigilEnd ^ id)) {
                spliceEnd = reinterpret_cast<const uint8_t *>(insn);
                break;
            }
        }

        if (!spliceStart || !spliceEnd) continue;

        int relSectionID = -1;
        std::string rname = ".rel";
        rname += (snames + text->name);
        for (int s = 0; s < fheaderPtr->sectionCount; s++) {
            if (sheaders[s].type != SectionHeader::Type::Relocation) continue;
            if (rname != (snames + sheaders[s].name)) continue;
            if (relSectionID >= 0)
                throw std::runtime_error(
                        "Multiple relocation sections for kernel");
            relSectionID = s;
        }

        auto removeBytes = spliceEnd - spliceStart + 16;

        size_t before = spliceStart - base;
        auto after = bytes - before - removeBytes;
        ptrdiff_t sizeAdjust = microkernel.size() - removeBytes;

        auto kbefore = before - text->offset;
        auto kafter = text->size - kbefore - removeBytes;

        std::vector<uint8_t> newBinary(bytes + sizeAdjust);
        auto newBase = newBinary.data();

        memmove(newBase, base, before);
        memmove(newBase + before, microkernel.data(), microkernel.size());
        memmove(newBase + before + microkernel.size(),
                spliceStart + removeBytes, after);

        fixupJumpTargets(newBase + text->offset, kbefore, +sizeAdjust);
        fixupJumpTargets(
                newBase + before + microkernel.size(), kafter, -sizeAdjust);

        fheaderPtr = reinterpret_cast<FileHeader *>(newBase);

        if (fheaderPtr->sectionTableOff > before)
            fheaderPtr->sectionTableOff += sizeAdjust;

        sheaders = reinterpret_cast<SectionHeader *>(
                newBase + fheaderPtr->sectionTableOff);
        sheaders[textSectionID].size += sizeAdjust;
        for (int s = 0; s < fheaderPtr->sectionCount; s++)
            if (sheaders[s].offset > before) sheaders[s].offset += sizeAdjust;

        if (relSectionID >= 0) {
            auto relSection = sheaders + relSectionID;
            auto rel = reinterpret_cast<Relocation *>(
                    newBase + relSection->offset);
            auto relEnd = reinterpret_cast<Relocation *>(
                    newBase + relSection->offset + relSection->size);
            for (; rel < relEnd; rel++) {
                if (rel->offset >= kbefore) rel->offset += sizeAdjust;
            }
        }

#ifdef SPLICE_DEBUG
        std::ofstream dump0("original." + std::to_string(id) + ".bin");
        dump0.write((const char *)binary.data(), binary.size());

        std::ofstream dump("patched." + std::to_string(id) + ".bin");
        dump.write((const char *)newBinary.data(), newBinary.size());
#endif

        std::swap(binary, newBinary);

        // Tail-recurse to handle any further instances of this microkernel
        fuse(binary, microkernel, id);
        return;
    }
}

void fuse(std::vector<uint8_t> &binary, const char *source) {
    std::vector<uint8_t> microkernel;
    const auto sigilLen = strlen(sigilBinary);

    auto toNybble = [](char c) {
        return ((c >= 'A') ? (c - 'A' + 10) : (c - '0')) & 0xF;
    };

    for (const char *s = std::strstr(source, sigilBinary); s;
            s = std::strstr(s, sigilBinary)) {
        s += sigilLen;
        char *after;
        long id = strtol(s, &after, 10);
        microkernel.clear();
        for (s = after + 1; *s != '\n'; s += 2) {
            if (!s[0] || !s[1]) break;
            microkernel.push_back(static_cast<uint8_t>(
                    (toNybble(s[0]) << 4) | toNybble(s[1])));
        }
        fuse(binary, microkernel, id);
    }
}

static void fixupJumpTargets(uint8_t *start, size_t len, ptrdiff_t adjust) {
    auto istart = reinterpret_cast<int32_t *>(start);
    auto iend = reinterpret_cast<int32_t *>(start + len);

    for (auto insn = istart; insn < iend; insn += 4) {
        if (insn[0] & (1u << 29)) {
            insn -= 2; /* skip compacted instructions */
            continue;
        }
        uint8_t op = insn[0] & 0xFF;
        if ((op & 0xF0) != 0x20) continue; /* skip non-jumps */
        if (op == 0x2B || op == 0x2D) continue; /* skip ret/calla */
        bool hasUIP = (op == 0x22 || op == 0x23 || op == 0x24 || op == 0x28
                || op == 0x2A || op == 0x2E);

        auto jumpFixup = [=](int32_t &ip) {
            auto target = ((insn - istart) << 2) + ip;
            if (target < 0 || target >= ptrdiff_t(len))
                ip += static_cast<int32_t>(adjust);
        };

        if (hasUIP) jumpFixup(insn[2]);
        jumpFixup(insn[3]);
    }
}

bool hasMicrokernels(const char *source) {
    return std::strstr(source, sigilBinary);
}

}
GEMMSTONE_NAMESPACE_END