diff options
| -rw-r--r-- | alc/backends/pipewire.cpp | 26 |
1 files changed, 13 insertions, 13 deletions
diff --git a/alc/backends/pipewire.cpp b/alc/backends/pipewire.cpp index ec852020..bbab7204 100644 --- a/alc/backends/pipewire.cpp +++ b/alc/backends/pipewire.cpp @@ -1071,32 +1071,32 @@ void PipeWirePlayback::outputCallback() pw_buffer *pw_buf{pw_stream_dequeue_buffer(mStream.get())}; if(unlikely(!pw_buf)) return; - spa_buffer *spa_buf{pw_buf->buffer}; - uint length{mRateMatch ? mRateMatch->size : mDevice->UpdateSize}; /* For planar formats, each datas[] seems to contain one channel, so store * the pointers in an array. Limit the render length in case the available * buffer length in any one channel is smaller than we wanted (shouldn't * be, but just in case). */ - const size_t chancount{minu(mNumChannels, spa_buf->n_datas)}; + spa_data *datas{pw_buf->buffer->datas}; + const size_t chancount{minu(mNumChannels, pw_buf->buffer->n_datas)}; + /* TODO: How many samples should actually be written? 'maxsize' can be 16k + * samples, which is excessive (~341ms @ 48khz). SPA_IO_RateMatch contains + * a 'size' field that apparently indicates how many samples should be + * written per update, but it's not obviously right. + */ + uint length{mRateMatch ? mRateMatch->size : mDevice->UpdateSize}; for(size_t i{0};i < chancount;++i) { - length = minu(length, spa_buf->datas[i].maxsize/sizeof(float)); - mChannelPtrs[i] = static_cast<float*>(spa_buf->datas[i].data); + length = minu(length, datas[i].maxsize/sizeof(float)); + mChannelPtrs[i] = static_cast<float*>(datas[i].data); } - /* TODO: How many samples should actually be written? 'maxsize' can be 16k - * samples, which is excessive (~341ms @ 48khz), but aside from what gets - * specified with PW_KEY_NODE_LATENCY, there's nothing here saying how much - * is needed to keep the stream healthy. - */ mDevice->renderSamples({mChannelPtrs.get(), chancount}, length); for(size_t i{0};i < chancount;++i) { - spa_buf->datas[i].chunk->offset = 0; - spa_buf->datas[i].chunk->stride = sizeof(float); - spa_buf->datas[i].chunk->size = length * sizeof(float); + datas[i].chunk->offset = 0; + datas[i].chunk->stride = sizeof(float); + datas[i].chunk->size = length * sizeof(float); } pw_buf->size = length; pw_stream_queue_buffer(mStream.get(), pw_buf); |