On Sun, 25 Aug 2013, Eldad Zack wrote:
Hi,
This patch series attempts to fix a long standing problem with the concurrent usage of playback and capture on implicit feedback devices.
My crude regression tester caught a problem causing the streams to fail with this order (though it doesn't trip any WARNs or oops):
++ Test stream permutation 0000013f * [1/14] capture => pcm_open * [2/14] capture => pcm_hw_params * [3/14] capture => pcm_prepare * [4/14] capture => pcm_start * [5/14] capture => waitsleep * [6/14] capture => pcm_stop ++ Frames [capture]: 39360 * [7/14] playback => pcm_open * [8/14] playback => pcm_hw_params * [9/14] capture => pcm_close * [10/14] playback => pcm_prepare * [11/14] playback => pcm_start * [12/14] playback => waitsleep * [13/14] playback => pcm_stop ++ Frames [playback]: 101760 * [14/14] playback => pcm_close ++ Test stream permutation 0000015f * [1/14] capture => pcm_open * [2/14] capture => pcm_hw_params * [3/14] capture => pcm_prepare * [4/14] capture => pcm_start * [5/14] capture => waitsleep * [6/14] playback => pcm_open * [7/14] capture => pcm_stop ++ Frames [capture]: 0 !! Test stream permutation 0000015f failed !!
For reference, I ran the tool with my hda-intel card (all 3432 combinations) and there were no errors. So it's still doesn't cover all cases.
Cheers, Eldad