On 5/9/2013 7:12 PM, Fabio Estevam wrote:
On Thu, May 9, 2013 at 10:17 PM, Troy Kisky troy.kisky@boundarydevices.com wrote:
Did you test a reset/reboot ? Since the fill_defaults now happens after the read of the device id, how does this fix the mentioned problem ?
Yes, tested several times :-)
The chip ID has to be read from the register and it always read correctly. In patch 1/2, I just moved the reading to a more standard location (in i2c_probe function), just like many other codecs do.
Patch 2/2 fixes the issue by ensuring that we start from sane values from power-on reset.
Let's look at the original error:
sgtl5000 0-000a: Device with ID register ffff is not a sgtl5000 sgtl5000 0-000a: ASoC: failed to probe CODEC -19 imx-sgtl5000 sound.12: ASoC: failed to instantiate card -19 imx-sgtl5000 sound.12: snd_soc_register_card failed (-19)
If I move the chip ID reading to i2c_probe (ie, only apply patch 1/2), the device ID would be read correctly, but the fail would happen at a later point:
sgtl5000 0-000a: sgtl5000 revision 0x11 sgtl5000 0-000a: Failed to get supply 'VDDD': -517 mmcblk0: p1 0-000a: 1200 mV normal sgtl5000 0-000a: Using internal LDO instead of VDDD usb 2-1: new high-speed USB device number 2 using ci_hdrc hub 2-1:1.0: USB hub found hub 2-1:1.0: 3 ports detected sgtl5000 0-000a: ASoC: failed to probe CODEC -110 imx-sgtl5000 sound.12: ASoC: failed to instantiate card -110 imx-sgtl5000 sound.12: snd_soc_register_card failed (-110) imx-sgtl5000: probe of sound.12 failed with error -110
Shouldn't you update your commit log with this then ?
So the original issue was not about reading the chip ID correctly.
The power related registers change from POR to reset (among others)
If the chip is not properly powered, then we are not able to read its ID and we get that original error.
Thanks for fixing, I guess I don't need to understand how the fix works.
Troy