Rock64 Not booting

[TonyMac32]

Is it simply not pingable, or is it just plain not booting? Can you give me the age of the Rock64 and the part number of the Ethernet chip?  Power domains, time to reset/bring up the SD card might not be waiting long enough either (I think it simply waits default 10 ms, might need a little longer)

 

Can everyone involved please give me the exact SD cards being used, or even pictures of them?  

[foxriver76]

In my case it just seems to be not pingable and seldom pingable but not accessable via SSH.

Where I can find information about the ethernet chip? Is it the number on the ethernet port? And where to find information about the age? Bought it four weeks ago but that isn't a valid information.

 

It's a San Disk Ultra 32 GB, Class 10 SD card. (this one:

https://www.amazon.de/gp/product/B073JWXGNT/ref=oh_aui_detailpage_o01_s00?ie=UTF8&psc=1)

 

 

 

[xelcho]

Well I can confirm that the 155 image @igor provided works on my 4gb rock64, also.

 

I admit that I only did the most basic of testing, loging in, running nano, ps and ls etc.

 

Is there a specific command or test I should run to confirm all is as expected?

[TonyMac32]

3 hours ago, foxriver76 said:

Bought it four weeks ago but that isn't a valid information.

Well, should be valid enough.  This is a case where I don't think too many of the original are "in the wild" anyway.  The chip near the Ethernet port that has the crab on it will say RTL8211E or F (IIRC, mine is an old one).  

[TonyMac32]

13 hours ago, foxriver76 said:

The HDMI output isn't working

 

I believe the initialization got some how messed up in Rockchip kernel past June.  On Tinker Board the answer was initialize display in U-boot.  Ayafun u-boot doesn't have any of the drivers for that yet, so I have only found unplugging/replugging the HDMI, sometimes multiple times, to work.

 

Thanks all for the updates so far, this looks like an issue on multiple fronts.  I also opened a pull request to Rockchip (I'll drop Ayufan a note) about the ciu-drv clock name typo that might impact tuning phase on the eMMC. And a deprecated frequency parameter that really shouldn't do anything but gives dmesg spam.

[foxriver76]

33 minutes ago, TonyMac32 said:

Well, should be valid enough.  This is a case where I don't think too many of the original are "in the wild" anyway.  The chip near the Ethernet port that has the crab on it will say RTL8211E or F (IIRC, mine is an old one).  

 

Mine is RTL8211F. Thanks for your help.

[botfap]

1 hour ago, TonyMac32 said:

 

I believe the initialization got some how messed up in Rockchip kernel past June.  On Tinker Board the answer was initialize display in U-boot.  Ayafun u-boot doesn't have any of the drivers for that yet, so I have only found unplugging/replugging the HDMI, sometimes multiple times, to work.

 

Not a real fix but running: "xrandr && xrandr --output HDMI-1 --auto" in a startup script will init the display on our RK and AML boards

 

 

On 9/8/2018 at 6:56 PM, TonyMac32 said:

@Igor I'm not convinced 50 MHz SD communications at 3.3V are stable on any rockchip SoC using a 4mA GPIO driver limit.  I'll make a separate topic for that discussion, but it might be the problem, I see some rockchip updates to the sdmmc clock sources, it could be a similar story to the Amlogic devices where they were not clocking at the requested speed before due to an error in picking a divider or some such nonsense.

 

We drive the SD at 12mA on RK devices after having lots and lots of problems with clients swapping out the provided SD cards with cheap shitters. 8mA mostly fixed it but we were left with a few edge cases. 4mA was problem after problem for 3.3v cards even at 25Mhz signalling which should be easily within spec

 

 

 

 

 

 

 

 

 

[TonyMac32]

Right, I've set it to 8, I'm considering 12 because of the results of some oversimplification modelling I did in the tutorial section showing it not completely taking care of the issue.

For the screen, yeah, we could do that. If Ayufan rebases the u-boot it won't be an issue.

Sent from my Pixel using Tapatalk

[TonyMac32]

Also forgot:  https://github.com/rockchip-linux/kernel/pull/122

 

That's for the phase tuning on mmc's, partially separate and minor, but another twig on the heap.

[Vin]

On 9/10/2018 at 3:36 AM, TonyMac32 said:

I didn't get any feedback on you about the HDMI hotplug, however on some favorable testing on a Renegade I've pushed a patch into the build system to set the drive levels on the SD to 8mA, up from 4mA.  I don't have a failuyre to boot issue with my SD cards, supply and Rock64, however a sample size of 1 is not representative, and my conclusion was that bumping all rk3328 devices to 8mA was advisable.  At @Igor's discretion this can be built for nightly or test.

 

sorry was cought up at work this week.

 

Regarding the SD card, I m using two different Samsung EVO Plus Micro SDXC 64GB, as well as a Samsung Speicherkarte MicroSDHC 32GB EVO Plus UHS-I and various different SanDisk and Samsung Cards

My Ethernet Chip is as well a RTL8211F

 

Tried five different Cards so far with the newest 155 iso, still not booting up

[Vin]

Wanted to give you the feedback that the latest update

 

ARMBIAN 5.60 stable Ubuntu 18.04.1 LTS 4.4.156-rockchip64

 

works flawlessly, thank you so much for your hard work you put into it, very happy to be back to the armbian distro.

 

Best regards

 

Vin