[U-Boot] A20-olinuxino-micro hangs in U-boot if powered from MINI-USB only.
Hi all,
I'm observing that the said A20-onlinuxino-micro board hangs in U-boot in probing USB2 (see example below) in case it is powered by connecting its mini-usb connector to a computer. The problem does not happen if 12V power connector is also empoyed. (Note: the example attached below is for RC1, but RC2 is no different) I can't say if such problem was also present in previous version of u-boot or not.
Thank you,
Regards, Nikolai
U-Boot 2019.01-rc1 (Dec 13 2018 - 15:11:11 +0300) Allwinner Technology
CPU: Allwinner A20 (SUN7I) Model: Olimex A20-Olinuxino Micro I2C: ready DRAM: 1 GiB MMC: SUNXI SD/MMC: 0, SUNXI SD/MMC: 1 Loading Environment from FAT... MMC: no card present In: serial Out: serial Err: serial Net: eth0: ethernet@1c50000 starting USB... USB0: USB EHCI 1.00 USB1: USB OHCI 1.0 USB2:
On Mon, 24 Dec 2018 01:04:04 +0300 Nikolai Zhubr n-a-zhubr@yandex.ru wrote:
Hi all,
I'm observing that the said A20-onlinuxino-micro board hangs in U-boot in probing USB2 (see example below) in case it is powered by connecting
I think it's normal. During the boot, u-boot initialize the usb otg, and for a fraction of second the power is disconnected. I solved it by attaching a battery (and I have a PSU as a side effect :) ).
Regards, Diego.
Hi, 24.12.2018 20:32, Diego Roversi:
I'm observing that the said A20-onlinuxino-micro board hangs in U-boot in probing USB2 (see example below) in case it is powered by connecting
I think it's normal. During the boot, u-boot initialize the usb otg, and for a fraction of second the power is disconnected. I solved it by attaching a battery (and I have a PSU as a side effect :) ).
Ah, got it. Thats a funny issue then. Its not the software to blame. Ok. (Ideally, this peculiarity should probably have been mentioned by Olimex in their docs, but maybe I've overlooked it.) Solved it for now by disabling USB altogether.
Thank you,
Regards, Nikolai
Regards, Diego.
On Monday, 24 December 2018, 19:10:24 GMT, Nikolai Zhubr n-a-zhubr@yandex.ru wrote:
Hi,24.12.2018 20:32, Diego Roversi:
I'm observing that the said A20-onlinuxino-micro board hangs in U-boot
in probing USB2 (see example below) in case it is powered by connecting
I think it's normal. During the boot, u-boot initialize the usb otg, and for a fraction of second the power is disconnected. I solved it by attaching a battery (and I have a PSU as a side effect :) ).
Ah, got it. Thats a funny issue then. Its not the software to blame. Ok. (Ideally, this peculiarity should probably have been mentioned by Olimex in their docs, but maybe I've overlooked it.) Solved it for now by disabling USB altogether.
Isn't it caused by the software as stated above, i.e. nothing to do with Olimex?
John
Hi, 24.12.2018 22:54, John S:
On Monday, 24 December 2018, 19:10:24 GMT, Nikolai Zhubrn-a-zhubr@yandex.ru wrote:
Hi,24.12.2018 20:32, Diego Roversi:
I'm observing that the said A20-onlinuxino-micro board hangs in U-boot
in probing USB2 (see example below) in case it is powered by connecting
I think it's normal. During the boot, u-boot initialize the usb otg, and for a fraction of second the power is disconnected. I solved it by attaching a battery (and I have a PSU as a side effect :) ).
Ah, got it. Thats a funny issue then. Its not the software to blame. Ok. (Ideally, this peculiarity should probably have been mentioned by Olimex in their docs, but maybe I've overlooked it.) Solved it for now by disabling USB altogether.
Isn't it caused by the software as stated above, i.e. nothing to do with Olimex?
Well, IMHO is has to do with Olimex because they designed this specific circuitry, and the circuitry is such that 5V voltage constantly applied to the micro-usb connector might essentially become disconnected (at the board side) by programming -- if I understood it correctly. I suppose this certainly could be prevented in hardware, by e.g. just wiring the 5V usb line to a power unit through a diode, whatever, unless it is prohibited by some sort of usb compatability requirements. I don't know if it would be good though. Of course the effect is caused by software in some sence, but the software can not know that there is no other power supply line present when it probes this usb port? Or could it indeed? I'm not sure if AXP state could be somehow probed to do a workaround, in theory, but that would be an extreme overcomplication anyway.
Thank you,
Regards, Nikolai
John
On Monday, 24 December 2018, 21:37:01 GMT, Nikolai Zhubr n-a-zhubr@yandex.ru wrote:
24.12.2018 22:54, John S: Isn't it caused by the software as stated above, i.e. nothing to do with Olimex?
Well, IMHO is has to do with Olimex because they designed this specific circuitry, and the circuitry is such that 5V voltage constantly applied to the micro-usb connector might essentially become disconnected (at the board side) by programming -- if I understood it correctly.
The software needs to be responsible for its own actions.
Does the problem you reported occur with the Olimex image(s)? If so it may be wise to report it to Olimex.
If not, which commit or change caused it?
BTW I read in the Olimex documentation that best practice is to power it differently than the way you mention.
John
Hi,
27.12.2018 23:48, 'John S' via linux-sunxi:
Well, IMHO is has to do with Olimex because they designed this specific circuitry, and the circuitry is such that 5V voltage constantly applied to the micro-usb connector might essentially become disconnected (at the board side) by programming -- if I understood it correctly.
The software needs to be responsible for its own actions.
Sure, ideally. That's why I reported.
Does the problem you reported occur with the Olimex image(s)? If so it may be wise to report it to Olimex.
I can't say about Olimex images at the moment because I switched to more current (alternative) images and subsequently to mainline openwrt and the board is already mounted for intended operation. Maybe I do some experiments later. I'd like to clarify, I don't blame Olimex, they are nice. Keeping close to mainline is not always easy.
If not, which commit or change caused it?
Again, I can't say at the moment. Not even sure if it could be fixed on the software side.
BTW I read in the Olimex documentation that best practice is to power it differently than the way you mention.
Yes, that is true. But powering through mini-usb is so wonderfully handy!
Thank you,
Regards, Nikolai
John
participants (3)
-
Diego Roversi -
John S -
Nikolai Zhubr