Samuel Thibault, le Tue 24 Mar 2009 23:43:41 +0100, a écrit : > The alsa case is quite odd, we're getting > ioctl(5, SNDRV_PCM_IOCTL_READI_FRAMES, 0x7fc0937b1fc0) = -1 EPIPE (Broken > pipe) > I haven't found why yet.
Ok, I've resorted to putting printks in the kernel, which got me to snd_pcm_pre_start() which calls snd_pcm_playback_data where runtime->stop_threshold >= runtime->boundary is not true because stop_threshold is 0x7fffffff, while runtime->boundary is 0x4800000000000000. The attached patch fixes it. Samuel
--- chuck-1.2.0.8.dfsg/src/rtaudio.cpp.orig 2009-03-25 00:22:20.000000000 +0100 +++ chuck-1.2.0.8.dfsg/src/rtaudio.cpp 2009-03-25 00:22:26.000000000 +0100 @@ -4047,7 +4047,7 @@ snd_pcm_sw_params_alloca( &sw_params ); snd_pcm_sw_params_current( handle, sw_params ); snd_pcm_sw_params_set_start_threshold( handle, sw_params, *bufferSize ); - snd_pcm_sw_params_set_stop_threshold( handle, sw_params, 0x7fffffff ); + snd_pcm_sw_params_set_stop_threshold( handle, sw_params, LONG_MAX ); snd_pcm_sw_params_set_silence_threshold( handle, sw_params, 0 ); snd_pcm_sw_params_set_silence_size( handle, sw_params, INT_MAX ); err = snd_pcm_sw_params( handle, sw_params );
