aboutsummaryrefslogtreecommitdiffstats
path: root/ffmpeg_cuvid.c
blob: 866f43b9d4401e6a9f70f50f9a25e46fd7d38d17 (plain) (blame)
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
235
236
/*
 * This file is part of FFmpeg.
 *
 * FFmpeg is free software; you can redistribute it and/or
 * modify it under the terms of the GNU Lesser General Public
 * License as published by the Free Software Foundation; either
 * version 2.1 of the License, or (at your option) any later version.
 *
 * FFmpeg is distributed in the hope that it will be useful,
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
 * Lesser General Public License for more details.
 *
 * You should have received a copy of the GNU Lesser General Public
 * License along with FFmpeg; if not, write to the Free Software
 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
 */

#include "libavutil/hwcontext.h"
#include "libavutil/hwcontext_cuda.h"

#include "ffmpeg.h"

#include <cuda.h>

typedef struct CUVIDContext {
    AVBufferRef *hw_frames_ctx;
} CUVIDContext;

static void cuvid_uninit(AVCodecContext *avctx)
{
    InputStream  *ist = avctx->opaque;
    CUVIDContext *ctx = ist->hwaccel_ctx;

    if (ctx) {
        av_buffer_unref(&ctx->hw_frames_ctx);
        av_freep(&ctx);
    }

    av_buffer_unref(&ist->hw_frames_ctx);

    ist->hwaccel_ctx = 0;
    ist->hwaccel_uninit = 0;
}

int cuvid_init(AVCodecContext *avctx)
{
    InputStream  *ist = avctx->opaque;
    CUVIDContext *ctx = ist->hwaccel_ctx;

    av_log(NULL, AV_LOG_TRACE, "Initializing cuvid hwaccel\n");

    if (!ctx) {
        av_log(NULL, AV_LOG_ERROR, "CUVID transcoding is not initialized. "
               "-hwaccel cuvid should only be used for one-to-one CUVID transcoding "
               "with no (software) filters.\n");
        return AVERROR(EINVAL);
    }

    return 0;
}

static void cuvid_ctx_free(AVHWDeviceContext *ctx)
{
    AVCUDADeviceContext *hwctx = ctx->hwctx;
    cuCtxDestroy(hwctx->cuda_ctx);
}

int cuvid_transcode_init(OutputStream *ost)
{
    InputStream *ist;
    const enum AVPixelFormat *pix_fmt;
    AVCUDADeviceContext *device_hwctx;
    AVHWDeviceContext *device_ctx;
    AVHWFramesContext *hwframe_ctx;
    CUVIDContext *ctx = NULL;
    CUdevice device;
    CUcontext cuda_ctx = NULL;
    CUcontext dummy;
    CUresult err;
    int ret = 0;

    av_log(NULL, AV_LOG_TRACE, "Initializing cuvid transcoding\n");

    if (ost->source_index < 0)
        return 0;

    ist = input_streams[ost->source_index];

    /* check if the encoder supports CUVID */
    if (!ost->enc->pix_fmts)
        goto cancel;
    for (pix_fmt = ost->enc->pix_fmts; *pix_fmt != AV_PIX_FMT_NONE; pix_fmt++)
        if (*pix_fmt == AV_PIX_FMT_CUDA)
            break;
    if (*pix_fmt == AV_PIX_FMT_NONE)
        goto cancel;

    /* check if the decoder supports CUVID */
    if (ist->hwaccel_id != HWACCEL_CUVID || !ist->dec || !ist->dec->pix_fmts)
        goto cancel;
    for (pix_fmt = ist->dec->pix_fmts; *pix_fmt != AV_PIX_FMT_NONE; pix_fmt++)
        if (*pix_fmt == AV_PIX_FMT_CUDA)
            break;
    if (*pix_fmt == AV_PIX_FMT_NONE)
        goto cancel;

    av_log(NULL, AV_LOG_VERBOSE, "Setting up CUVID transcoding\n");

    if (ist->hwaccel_ctx) {
        ctx = ist->hwaccel_ctx;
    } else {
        ctx = av_mallocz(sizeof(*ctx));
        if (!ctx) {
            ret = AVERROR(ENOMEM);
            goto error;
        }
    }

    if (!hw_device_ctx) {
        hw_device_ctx = av_hwdevice_ctx_alloc(AV_HWDEVICE_TYPE_CUDA);
        if (!hw_device_ctx) {
            av_log(NULL, AV_LOG_ERROR, "av_hwdevice_ctx_alloc(AV_HWDEVICE_TYPE_CUDA) failed\n");
            ret = AVERROR(ENOMEM);
            goto error;
        }

        err = cuInit(0);
        if (err != CUDA_SUCCESS) {
            av_log(NULL, AV_LOG_ERROR, "Could not initialize the CUDA driver API\n");
            ret = AVERROR_UNKNOWN;
            goto error;
        }

        err = cuDeviceGet(&device, 0); ///TODO: Make device index configurable
        if (err != CUDA_SUCCESS) {
            av_log(NULL, AV_LOG_ERROR, "Could not get the device number %d\n", 0);
            ret = AVERROR_UNKNOWN;
            goto error;
        }

        err = cuCtxCreate(&cuda_ctx, CU_CTX_SCHED_BLOCKING_SYNC, device);
        if (err != CUDA_SUCCESS) {
            av_log(NULL, AV_LOG_ERROR, "Error creating a CUDA context\n");
            ret = AVERROR_UNKNOWN;
            goto error;
        }

        device_ctx = (AVHWDeviceContext*)hw_device_ctx->data;
        device_ctx->free = cuvid_ctx_free;

        device_hwctx = device_ctx->hwctx;
        device_hwctx->cuda_ctx = cuda_ctx;

        err = cuCtxPopCurrent(&dummy);
        if (err != CUDA_SUCCESS) {
            av_log(NULL, AV_LOG_ERROR, "cuCtxPopCurrent failed\n");
            ret = AVERROR_UNKNOWN;
            goto error;
        }

        ret = av_hwdevice_ctx_init(hw_device_ctx);
        if (ret < 0) {
            av_log(NULL, AV_LOG_ERROR, "av_hwdevice_ctx_init failed\n");
            goto error;
        }
    } else {
        device_ctx = (AVHWDeviceContext*)hw_device_ctx->data;
        device_hwctx = device_ctx->hwctx;
        cuda_ctx = device_hwctx->cuda_ctx;
    }

    if (device_ctx->type != AV_HWDEVICE_TYPE_CUDA) {
        av_log(NULL, AV_LOG_ERROR, "Hardware device context is already initialized for a diffrent hwaccel.\n");
        ret = AVERROR(EINVAL);
        goto error;
    }

    if (!ctx->hw_frames_ctx) {
        ctx->hw_frames_ctx = av_hwframe_ctx_alloc(hw_device_ctx);
        if (!ctx->hw_frames_ctx) {
            av_log(NULL, AV_LOG_ERROR, "av_hwframe_ctx_alloc failed\n");
            ret = AVERROR(ENOMEM);
            goto error;
        }
    }

    /* This is a bit hacky, av_hwframe_ctx_init is called by the cuvid decoder
     * once it has probed the neccesary format information. But as filters/nvenc
     * need to know the format/sw_format, set them here so they are happy.
     * This is fine as long as CUVID doesn't add another supported pix_fmt.
     */
    hwframe_ctx = (AVHWFramesContext*)ctx->hw_frames_ctx->data;
    hwframe_ctx->format = AV_PIX_FMT_CUDA;
    hwframe_ctx->sw_format = AV_PIX_FMT_NV12;

    ost->hwaccel_ctx = ctx;
    ost->enc_ctx->hw_frames_ctx = av_buffer_ref(ctx->hw_frames_ctx);
    ost->enc_ctx->pix_fmt = AV_PIX_FMT_CUDA;

    if (!ost->enc_ctx->hw_frames_ctx) {
        av_log(NULL, AV_LOG_ERROR, "av_buffer_ref failed\n");
        ret = AVERROR(ENOMEM);
        goto error;
    }

    if (!ist->hwaccel_ctx) {
        ist->hwaccel_ctx = ctx;
        ist->hw_frames_ctx = av_buffer_ref(ctx->hw_frames_ctx);
        ist->dec_ctx->hw_frames_ctx = av_buffer_ref(ctx->hw_frames_ctx);
        ist->dec_ctx->pix_fmt = AV_PIX_FMT_CUDA;
        ist->resample_pix_fmt = AV_PIX_FMT_CUDA;

        ist->hwaccel_uninit = cuvid_uninit;

        if (!ist->hw_frames_ctx || !ist->dec_ctx->hw_frames_ctx) {
            av_log(NULL, AV_LOG_ERROR, "av_buffer_ref failed\n");
            ret = AVERROR(ENOMEM);
            goto error;
        }
    }

    return 0;

error:
    av_freep(&ctx);
    return ret;

cancel:
    if (ist->hwaccel_id == HWACCEL_CUVID) {
        av_log(NULL, AV_LOG_ERROR, "CUVID hwaccel requested, but impossible to achive.\n");
        return AVERROR(EINVAL);
    }

    return 0;
}