On 2019-12-11 9:27 AM, Marco Felsch wrote: > On 19-12-11 01:06, Anson Huang wrote: >>> Subject: Re: [PATCH 1/3] ARM: dts: imx6ul-14x14-evk: Add sensors' GPIO >>> regulator >>> >>> On 24.10.2019 11:51, Anson Huang wrote: >>>> On i.MX6UL 14x14 EVK board, sensors' power are controlled by >>>> GPIO5_IO02, add GPIO regulator for sensors to manage their power. >>>> >>>> Signed-off-by: Anson Huang <Anson.Huang@xxxxxxx> >>> >>> For me this breaks network boot on imx6ul evk, relevant log snippet is this: >>> >>> fec 20b4000.ethernet eth0: Unable to connect to phy >>> IP-Config: Failed to open eth0 >>> >>> Looking at schematics (SPF-28616_C2.pdf) I see that SNVS_TAMPER2 pin is >>> connected to PERI_PWREN which controls VPERI_3V3 which is used across >>> the board: >>> * Sensors (VSENSOR_3V3) >>> * Ethernet (VENET_3V3) >>> * Bluetooth >>> * CAN >>> * Arduino header >>> * Camera >>> >>> Maybe there are board revision differences? As far as I can tell this regulator >>> is not specific to sensors so it should be always on. >> >> You are correct, this regulator controls many other peripherals, I should make it always ON for now >> to make sure NOT break other peripheral, and after all other peripherals controlled >> by this regulator have added this regulator management, then the always ON can be >> removed. > > IMHO marking the regulator as always on shouldn't be the fix. Is it to > much work to add all required regulators? At least please add a comment > which describes the need of the always-on property. I don't have the hardware to test all affected peripherals on hand and no familiarity with stuff like CAN. Renaming reg_sensor and adding a comment makes sense. -- Regards, Leonard