Hi Mark,
On 02/26/2015 03:54 PM, Eric Nelson wrote:
This patch set addresses a structural problem in the handling of regulators for VDDIO, VDDA, and VDDD in the SGTL5000 driver.
The first two of these power rails must be powered on prior to any I2C communication, and yet the regulators were tied to the codec, which is instantiated only after a fair amount of I2C communication takes place.
In other words, these regulators could never have function, and we can surmise that no user of this driver has switched power supply rails connected to them.
The third power rail (VDDD) can be derived internally (by using I2C registers) though the data sheet says that if an external VDDD is used, it should be enabled before MCLK is started and I2C activity begins.
Patch 1 moves the regulators to the I2C device and initializes them before I2C activity. Patch 2 addresses an issue found during development of the patch and is somewhat unrelated to regulators. It is included for discussion Patch 3 removes the root cause of patch 2 Patch 4 adds return value checks when setting the LINREG and ANA_POWER registers Patch 5 addresses the cause of follow-on failures and the source of the error in patch 3. Patch 6 removes the naive use of the supplies to lower power consumption and replaces it with something more modest.
The patch set is being sent as an RFC for discussion, and should probably be squashed into two or three patches covering the regulator updates (patches 1, 3-5 and perhaps 6), initialization error handling (patch 2).
Eric Nelson (6): ASoC: sgtl5000: fix regulator support ASoC: sgtl5000: write all default registers ASoC: sgtl5000: initialize CHIP_ANA_POWER to power-on defaults ASoC: sgtl5000: check return values ASoC: sgtl5000: disable internal PLL early ASoC: sgtl5000: Don't disable regulators in SND_SOC_BIAS_OFF
Based on your review, it seems like the following is needed for a V2 patch (set):
- Move call to sgtl5000_fill_defaults() earlier in the sgtl5000_i2c_probe routine, before adjusting ANA_POWER. This will allow registers LINEREG_CTRL, CLK_CTRL, and even ANA_POWER back into the default register list. - Move regulators from codec to I2C device - switch to devm_regulator api - various code cleanups to simplify logic - Adjust regulator usage in SND_SOC_BIAS_OFF
I'll leave out the "write all default" registers for the moment.
Let me know if you'd like to see this in multiple patches or a single patch.
I'll wait to see if there's other feedback before prepping V2.
Regards,
Eric