Bugzilla – Bug 13895
Audio cuts out, when driver is re-opened
Last modified: 2016-11-27 21:10:58 UTC
The audio driver has to be re-opened when changing sample rate or enabling line-in. If this is done repeatable at some stage the audio stops working until the system is rebooted. This is easiest to see by toggling the line-in option. Caleb and myself have already done some investigation. It appears the the dac bit block does not work when in this state. I have tried disabling the headphone detect, and performing a software reset on the dac with no effect.
== Auto-comment from SVN commit #7475 to the jive repo by ccrome == == https://svn.slimdevices.com/jive?view=revision&revision=7475 == Bug #13895. Maurice and I did many investigations into this problem. Long story short: no love. We can always reproduce the problem. Long story Long: We tried the following: * Slowing down the I2C bus -- no effect * (Slowed to 31khz from 100kHz). * Initialize the PLL properly. The current implementation uses out-of-range values for the PLL. This checkin fixes that (see comments in code). * Initialize the PLL in the correct sequence. PLL first -> then init DACs. No effect. * Checked thousands of I2C read/write cycles, no bit errors at all. * Added read-after-write to verify register contents. no errors. * Changed drive strength of MCLK to normal. It was strong. This had no effect.
== Auto-comment from SVN commit #7476 to the jive repo by ccrome == == https://svn.slimdevices.com/jive?view=revision&revision=7476 == Bug #13895. oops. accidentally enabled the patch by default. just disabling it in the series file.
== Auto-comment from SVN commit #7559 to the jive repo by richard == == https://svn.slimdevices.com/jive?view=revision&revision=7559 == Bug #13895 Integrate (most) of Caleb's audio driver patch.
In addition to Caleb's changes I have also tried a software and hardware resets on the DAC, and various delays. Adding delay or printk's does seem to make the system more stable, but it does always fail. I'm not sure what else to try, Caleb?
Ugh, I don't really know what to try either. We need to get TI involved. We'll call Yuan.
This guy will probably not get solved before PLD. We have a few methods in mind to work around the problem: Richard, does a HW reset clear the problem, or is a power-off required? Possible solutions: * Fix it: (i.e. get TI to tell us what we're doing wrong) * HW Reset (if it works to clear the issue). * Always use 44.1 kHz, and never reconfigure the DAC. * Use the i.MX25 as the master I2S. Should we keep this as a 7.4 P1?
Don't need to get this in for 7.4. Retargeting for 8.0.
I guess this is rather a WONTFIX...
I am experiencing a bug with the Logitech Radio that, I believe, is related to this bug. The problem is simple and reliably reproducible: * Set the power-saving mode to put the device into standby mode. * Plug an external audio device into the Line-In port. Play some sound - all is well. * Stop the sound and wait and allow the device to drop into standby mode. * Wake up the device somehow (e.g., touching a button). * Play some sound through the external audio device - no sound. * Unplug the Line-In connection. Plug it back in. Play sound - all is well again. Is the bug that I'm experiencing related to the bug described here? If so... well, please fix it. =)
Unassigned bugs cannot have a priority.