mbox series

[0/4] ASoC: ti: davinci: Fixes for daVinci boards with McBSP

Message ID 20190830103841.25128-1-peter.ujfalusi@ti.com (mailing list archive)
Headers show
Series ASoC: ti: davinci: Fixes for daVinci boards with McBSP | expand

Message

Peter Ujfalusi Aug. 30, 2019, 10:38 a.m. UTC
Hi,

Morimoto-san pointed out in a private thread that the davinci-i2s driver is
using snd_soc_rtdcom_lookup(), which is should not really.

It took a while to get access to a daVinci board which uses McBSP, but
eventually I got access to a remote dm644x-evm and dma365-evm.

Initially the audio was not even probing on them (fixes sent for mach-davinci),
but then the driver was still crashing and not behaving correctly.

With the 4 patch dm644x-evm have working audio and the offending
snd_soc_rtdcom_lookup() is gone from the driver.
dm365-evm still have some issues regarding to EDMA, but it looks like a system
issue which can be fixed later.

Regards,
Peter
---
Peter Ujfalusi (4):
  ASoC: ti: davinci-i2s: Add S32_LE as support format
  ASoC: ti: davinci-i2s: Move the XSYNCERR workaround to .prepare
    callback
  ASoC: ti: edma-pcm: Fix for legacy dma_slave_map based channel lookup
  ASoC: ti: davinci-evm: Do not fail if the dai_set_sysclk returns
    -ENOTSUPP

 sound/soc/ti/davinci-evm.c |  2 +-
 sound/soc/ti/davinci-i2s.c | 90 +++++++++++++++++---------------------
 sound/soc/ti/edma-pcm.c    | 17 ++++++-
 3 files changed, 57 insertions(+), 52 deletions(-)