Bringing up Snickerdoodle

[Nick Burkitt]

I'm going to post some of my experiences in bringing up a snickerdoodle board (assuming I eventually accomplish that feat), and I encourage everyone else contribute their hard-won knowledge of what works and what doesn't.

Development Platform

I'm using both Windows and Linux. Windows is my target development environment, but some of what I need to do requires Linux.

I'm using Ubuntu 14.04.3 (x64) on a physical machine. That's the version that Xilinx officially supports, and it seems to work. I tried at first to use the latest Ubuntu release (16.04 LTS), but XSDK crashes on startup. Vivado seems to run, though.

I'm using Windows 10 (x64) on a physical machine.

Booting Snickerdoodle

So far, I've followed the instructions in the Getting Started Guide to create a bootable microSD card (SanDisk 64 GB). If you're looking for the ROOTFS tarball, see this post from Bush. UPDATE: And this post (from me).

I've installed the microSD card, connected my sn'oodle via a USB cable to my Windows system, and found the appropriate (I hope) driver for the virtual COM port. So far, though, there have been no signs of life beyond a pulsing LED, which is the same whether there's an SD card in the drive or not. UPDATE: Snickerdoodle is alive! No WiFi yet, but how hard can that be? :-P

UPDATE2: Okay, pretty danged hard. Fortunately, Jamil posted these instructions which made it easy. By the way, the predefined non-root user is "snickerdoodle", with password also "snickerdoodle".

Anyone else making progress?

[Nick Burkitt]

Oh, and Ubuntu 14.04.3 can be found here.

[Rob Barris]

I don't have any hardware myself yet so I am just following along here - should it be possible to get the board to drop into u-boot and get to a prompt showing that the chip is alive and well ?  i.e. if your com port link is good, it shouldn't be dead silent.

On Friday, June 10, 2016 at 11:21:22 AM UTC-7, Nick Burkitt wrote:

I'm going to post some of my experiences in bringing up a snickerdoodle board (assuming I eventually accomplish that feat), and I encourage everyone else contribute their hard-won knowledge of what works and what doesn't.

Development Platform

I'm using both Windows and Linux. Windows is my target development environment, but some of what I need to do requires Linux.

I'm using Ubuntu 14.04.3 (x64) on a physical machine. That's the version that Xilinx officially supports, and it seems to work. I tried at first to use the latest Ubuntu release (16.04 LTS), but XSDK crashes on startup. Vivado seems to run, though.

I'm using Windows 10 (x64) on a physical machine.

Booting Snickerdoodle

So far, I've followed the instructions in the Getting Started Guide to create a bootable microSD card (SanDisk 64 GB). If you're looking for the ROOTFS tarball, see this post from Bush. 

I've installed the microSD card, connected my sn'oodle via a USB cable to my Windows system, and found the appropriate (I hope) driver for the virtual COM port. So far, though, there have been no signs of life beyond a pulsing LED, which is the same whether there's an SD card in the drive or not.

Anyone else making progress?

[Nick Burkitt]

And set your terminal emulator to 9600/N/8/1. It's alive!

Actually, set your terminal emulator to the same values as you set the virtual COM port in Device Manager (I did say I was using Windows, didn't I?)

[Nick Burkitt]

Hi Rob.

should it be possible to get the board to drop into u-boot

Yes, if you can get your terminal program connected in time, you get the u-boot "Hit any key to stop autoboot" prompt. I'm using the USB port for both power and serial communications, so the COM port is only available after applying power. That makes getting connected in time to see the prompt less than a sure thing. If you were powering the board through J2, it would be easier. The reset button doesn't seem to have any effect at this point.

Do you have hardware on the way?

[weath...@krtkl.com]

Hi Nick,

I also receiver the odd error message upon unrolling the tar file.
Bush is going to post a compressed microSD card image for a 32GB card along with instructions for creating the card in Windows.

Thank you again for being part of the Alpha group.  This is some of the usability stuff that will definitely get fixed for the production shipments.

-Jamil

On Friday, June 10, 2016 at 11:21:22 AM UTC-7, Nick Burkitt wrote:

I'm going to post some of my experiences in bringing up a snickerdoodle board (assuming I eventually accomplish that feat), and I encourage everyone else contribute their hard-won knowledge of what works and what doesn't.

Development Platform

I'm using both Windows and Linux. Windows is my target development environment, but some of what I need to do requires Linux.

I'm using Ubuntu 14.04.3 (x64) on a physical machine. That's the version that Xilinx officially supports, and it seems to work. I tried at first to use the latest Ubuntu release (16.04 LTS), but XSDK crashes on startup. Vivado seems to run, though.

I'm using Windows 10 (x64) on a physical machine.

Booting Snickerdoodle

So far, I've followed the instructions in the Getting Started Guide to create a bootable microSD card (SanDisk 64 GB). If you're looking for the ROOTFS tarball, see this post from Bush. 

I've installed the microSD card, connected my sn'oodle via a USB cable to my Windows system, and found the appropriate (I hope) driver for the virtual COM port. So far, though, there have been no signs of life beyond a pulsing LED, which is the same whether there's an SD card in the drive or not.

Anyone else making progress?

[Rob Barris]

I do expect some boards whenever the next batch of alphas goes out, I was about the 15th backer or so way down at the bottom of the list :)

[Nick Burkitt]

Hi Jamil.

The problem is just that the instructions don't agree with the tarball as posted - the paths aren't three extra levels deep. If you leave off the "--strip-components=3" option, then all is well, you get no errors and you end up with a working ROOTFS.

Now on to WiFi! Any hints? :-)

-Nick

[weath...@krtkl.com]

The correct terminal emulator settings are 115200, n, 8, 1

-Jamil

[weath...@krtkl.com]

These are the notes I took on that.   I'm actually the hardware/FPGA/baremetal guy not the Linux expert so hopefully my naivety is helpful:

1) Clear out the contents of /etc/udev/rules.d/70-persistent-net.rules with vi and execute a reboot so that the wireless interface comes up at wlan0
2) ifconfig wlan0 up to bring up the WLAN radio
2) Do a iw wlan0 scan | grep SSID  to find the SSID of your base station
3) use vi to edit the template in /etc/wpa_supplicant.conf for your basestation SSID (ssid=) and WPA password (psk=)
4) execute wpa_supplicant -d -D nl80211 -c /etc/wpa_supplicant.conf -i wlan0 -B to bring up the WPA supplicant --- apparently that's what they call the software to support WPA encryption on the WLAN station
5) execute dhclient wlan0 to get your dhcp address etc from the router
6) use iw wlan0 link to show radio link statistics
7) use ifconfig wlan0  to show network interfaces and IP/ethernet addresses

That should help.

-Jamil

[Bush]

I've updated the getting started guide to remove the "--strip-components=3" flag.

On Friday, June 10, 2016 at 12:12:01 PM UTC-7, Nick Burkitt wrote:

If you use the ROOTFS tarball from here, the command shown in the Getting Started Guide for untarring it is wrong. Omit the "--strip-components=3" option, and use just "tar -C /media/user/ROOTFS -xvzf snickerdoodle-ubuntu-14.04.tar.gz".

[Nick Burkitt]

Your naivety perfectly matches my own on this point. And your instructions were spot on, as well - I can rsh into the board, and bring up the default web page, too.! Not bad for a hardware guy. :-P

Thanks!

-Nick

[weath...@krtkl.com]

Hi Rob,

As of this moment our CM is assembling those starting on 6/14.  Accounting for delivery to us for final test and shipping up to CrowdSupply for fulfillment I would say that your board should ship on 6/27 from Crowd Supply.
Thanks for hanging in there.

-Jamil

[Rob Barris]

Sounds great; wasn't complaining!  Looking forward to it.

[Tom Olenik]

Nick or anyone else,

How did you get from the "no signs of life other than the pulsing LED" and "Snickerdoodle is Alive!" That is where I am stuck at the moment. The details of where I'm at are:

• I'm using Ubuntu 14.04 LTS (64 bit) on a dedicated machine, not a virtual machine.
• I'm using the preloaded microSD card that arrived with the snickerdoodle. The case has a sticker than says "/01 /02" on it. I'm not sure what that means.
• When I use ls /dev/tty* before and after plugging in the snickerdoodle I see it show up as /dev/ttyACM0
• When I plug the snickerdoodle into my computer via USB the white light start pulsing in a heartbeat like pattern.
• I've tried using screen typing "sudo screen /dev/ttyACM0 115200"
• I then just get a blank terminal with a blinking cursor. No prompt, nothing. Is just stays there until I kill it. 
• I tried minicom as well and same result. I also tried dropping 115200 from the above and same thing.
• I have both a snickerdoodle standard and a snickerdoodle black. Same results with either one.

My next thing to try is to make my own microSD card and try that, but thought I would check to see if there could be something else anyone else suggests. 

Thanks,

Tom Olenik

[Bush]

Hey Tom,

Can you connect the microSD card to your Ubuntu host and confirm it's contents?

On Monday, July 11, 2016 at 4:54:18 PM UTC-7, Tom Olenik wrote:

Nick or anyone else,

How did you get from the "no signs of life other than the pulsing LED" and "Snickerdoodle is Alive!" That is where I am stuck at the moment. The details of where I'm at are:

• I'm using Ubuntu 14.04 LTS (64 bit) on a dedicated machine, not a virtual machine.
• I'm using the preloaded microSD card that arrived with the snickerdoodle. The case has a sticker than says "/01 /02" on it. I'm not sure what that means.
• When I use ls /dev/tty* before and after plugging in the snickerdoodle I see it show up as /dev/ttyACM0
• When I plug the snickerdoodle into my computer via USB the white light start pulsing in a heartbeat like pattern.
• I've tried using screen typing "sudo screen /dev/ttyACM0 115200"

• I then just get a black screen with a blank terminal with a blinking cursor. No prompt, nothing. Is just stays there until I kill it. 

• I tried minicom as well and same result. I also tried dropping 115200 from the above and same thing.
• I have both a snickerdoodle standard and a snickerdoodle black. Same results with either one.

[Tom Olenik]

Yes, I forgot to mention I did that. I see a ROOTFS and BOOT. Is there anything in particular I should look for? At first glance it looks like other Linux images I've seen, but I will not claim to be a Linux expert. 

Sent from my iPhone

--
You received this message because you are subscribed to the Google Groups "snickerdoodle forum" group.
To unsubscribe from this group and stop receiving emails from it, send an email to krtkl-snickerdo...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/krtkl-snickerdoodle/3d4a78ff-d4de-4b9f-9e68-7eaa0646a71c%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Bush]

There should be a BOOT.bin, devicetree.dtb, uImage, uEnv.txt (at a minimum) along with some additional update scripts and pre-built bitstreams and configurations. If you type into the terminal do you get momentary blinking on the green LED nearest to the microUSB jack?

[Rob Barris]

Just throwing out some things

a - I was kind of a doofus when I first tried to emplace the microsd card.  I thought I was supposed to lift the little metal lid, lay down the card, close the lid on top of it... nope.  (Never got so far as trying to boot in this sad shape, I knew it wasn't right...)  Lid lifts up, card goes in lthe id, lid clamps back down nicely.  If this was in the docs, I missed it.

b - I use minicom @ /dev/ttyACM0 and that works for me on Ubuntu 14.04.  I think I usually run it 'sudo' however. It can notice the port going missing if I physically unplug the USB tether, and then recover and notice its return when replugged.  115200 baud rate set in the comm settings there.

c - do you have a JTAG probe for the JTAG port ?  I do a fair amount of code & test just using baremetal mode in Vivado + XSDK, in which case the sdcard is somewhat superfluous.  That could be a way to test some board behavior via the hardware manager in vivado.

A krtkl person can probably explain where u-boot comes from, is it there even if no sdcard present, is it in the QSPI flash ?  You should see u-boot saying *something* as power is applied, maybe spam some keys to interrupt the autoboot.

[Bush]

Thanks, Rob. The SD card cage can be a bit tricky. The card should indeed be held by the tabs IN the lid rather than just being held down by it. Hopefully you can see that in my (somewhat blurry) image. To confirm it's an issue with the SD card, and not a problem with the device, you can try to boot from the QSPI by holding down the RESET button when connecting power. After holding the button down for 3 seconds, you should see the normal "heartbeat" light pulse and you can release the button and connect using your terminal emulator.

[Tom Olenik]

Confirmed I see:

BOOT.bin

config.txt

devicetree.dtb

uEnv.txt

uImage

update_qspi.img

uramdisk.image.gz

I also see a folder called "config".

When I first plug in I see activity on the green LED. I also do see activity on the green LED when I type into the terminal using screen. I do not see green LED activity when using minicom.

[Bush]

Can you try to boot from the QSPI flash using the instructions above and let me know your results? You should contact me directly to do more troubleshooting to avoid a long-winded back-and-forth on this thread.

[Rob Barris]

With the snickerdoodle attached, try running minicom as root (sudo minicom) and check that its comm device is /dev/ttyACM0?

(Ctrl-A, then letter O )

Attaching snap of what it looks like on my machine.

 

[weath...@krtkl.com]

I'm trying to figure out exactly what "hardware flow control = YES" does on USB CDC link.
I would generally set that to "NO" unless there exists actual RTS/CTS pins

[Tom Olenik]

Rob,

That was exactly it although I had a compounding issue described below. Two screen shots before and after showing the issue I was having with minicom specifically. 

Bush,

The other issue was I was mostly working with the black version to troubleshoot this and my preloaded card was for the standard version. Your instructions in the snickerdoodle book for creating a microSD card in Windows worked perfectly. 

I now have coms with both the standard and the black via both minicom and screen. In both I still don't see anything unless I try entering a command followed by enter. Then I will get an error and the snickerdoodle> prompt. 

On to the next steps which seem to be to figure out why there are no files listed when I use ls. See third screenshot where help returns a description of what ls should do like normal, but ls doesn't return anything. It doesn't seem like there is a file system. This might just be my lack of Linux experience, but I'm guessing I won't be the only one that runs into it. 

Thanks for the help. 

Tom Olenik

[flashp...@gmail.com]

Although I'm unable to establish a connection via the BT interface. Neither my phone or PC running Ubuntu 14.10 see any activity, I've established connectivity via a USB serial interface and putty. I've even compiled and run the two examples in the workspace.

So, pretty good progress.

Not sure at the moment on the best path for debugging the BT interface.

Regards,

Flash, aka Chuck Glasser

[Rob Barris]

The prompt you are typing 'ls' at is the u-boot prompt I believe.  That's not Linux, it's the preboot environment.  Try typing "boot" as a command to that prompt.

I can totally see why the SD cards are not directly interchangeable, it's because there is a bitstream file on the BOOT partition that is Zynq chip model specific.  So I will have to go through the card-brewing cycle for my non-black as I don't have a prebuilt SD card to clone here for it..

[Bush]

Yes, that I U-Boot, not Linux. There must be something wrong with your boot configuration or some of the boot images are not loading. Can you post the result of entering printenv into the U-Boot prompt and the contents of uEnv.txt in your boot partition on the SD card.

[Tom Olenik]

I tried the reset command before seeing the help above and that got me into Linux. The boot command also work. Attached are the results from the printenv command before getting to Linux. I've also attached my uEnv.txt. It seems like I can get things going, but there does seem to be a hickup somewhere. 

Thanks again for all the help. I plan to consolidate the lessons learned from this initial set-up and make a video out of it to help others.

[Bush]

What I think happened is you attempted to boot into the QSPI using the reset button and were presented with a U-Boot prompt which was booted from the QSPI flash. You should be able to confirm that by simply removing the SD card and powering the board with the reset button pressed. It sounds like you were able to get the SD card booted, though. It's good to get these sticking points documented to avoid people similarly pulling their hair out when trying to get up and running. Thanks for your patience.

[Tom Olenik]

This is very repeatable. I am greeted with UBOOT and the snickerdoodle> prompt whether I press the reset button or not. I get the same if I remove the sd card and hold reset when plugging in. I get the same on the standard snickerdoodle with the preloaded image or on the black with the image I loaded myself. Both will re-boot into Linux if I enter the reset or boot command.

I did capture a screenshot that might have a clue as to what is happening. By running screen immediately after plugging in the snickerdoodle into my USB, I was able to capture the attached which reports a bad CRC and then there is the message to "Hit any key to stop autoboot: 0".  Followed by some indication that the terminal emulator was sending commands. 

I agree about documenting these. Some little glitches are to be expected at this point. You can never find all of them until you get some units into the hands of a few people not involved with the development. I'm glad to be part of it.

[Bush]

That is strange behavior, indeed. It looks as though your host machine is sending characters over the connection before you are able to establish a connection with your emulator. The normal autoboot countdown timer lasts for 3 seconds and will boot using the commands specified in the uEnv.txt file if no characters are received before the 3 second timer expires. The "*** Warning - bad CRC, using default environment" message is letting you know that there is not a valid stored environment on the QSPI flash because no environment has been saved. This can be ignored. This message should go away if you save the environment using saveenv from the U-Boot prompt.

Do you get the same output when using minicom vs screen? I would like to understand the origin of the interrupting characters.

Also, can you output (at least the last half or so) the boot messages you get when you boot into Linux? I believe you may be booting into the uramdisk and not the Ubuntu filesystem on the SD card. You can try run uenvboot (which is the normal boot command) to boot using the commands and variables specified in uEnv.txt, which should use the Ubuntu filesystem. Let me know your results. Thanks, again.

[Rob Barris]

Is there any pin where the UART data that normally flows out the USB path (ttyACM0) is physically mirrored on the board, so one could attach an always-ready serial port on separate cable, to watch what comes out at first device power on ?

[Tom Olenik]

I get the same result in minicom as in screen. See attached screen capture. The linux boot messages are in the attached pdf. It appears to me it is accessing the microSD card. 

[Tom Olenik]

 Another update. If I power up the snickerdoodle first with an external power supply connected to breaky board, wait a couple minutes, then connect the usb, then I get linux as soon as I start minicom. See attached screen shot. This does seem to be something that my host machine is causing. If I can get it to stop sending characters when screen or minicom starts, it seems the issue will be solved. Again, it is a dedicate linux machine. Not a virtual or even a dual boot. I'm sure it is some setting somewhere now.

I have a work around. So I'm going to move on for now.

[Rob Barris]

Minicom does have some "modem setup strings" that I went into prefs and cleared out.  It might be that minicom is trying to init your "modem" and if you do that during the autoboot countdown, plop, you wind up in u-boot.  But I would think if that happened, you could just type "boot" there and proceed.

[Tom Olenik]

That makes sense. I figure it is something like that. The characters showing up are AT. There is a tickle of a memory in the back of my mind telling me that is what a modem handshake might start off like but it's been a long time. 

Sent from my iPhone

On Jul 12, 2016, at 6:48 PM, Rob Barris <rba...@gmail.com> wrote:

Minicom does have some "modem setup strings" that I went into prefs and cleared out.  It might be that minicom is trying to init your "modem" and if you do that during the autoboot countdown, plop, you wind up in u-boot.  But I would think if that happened, you could just type "boot" there and proceed.

--

You received this message because you are subscribed to the Google Groups "snickerdoodle forum" group.
To unsubscribe from this group and stop receiving emails from it, send an email to krtkl-snickerdo...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/krtkl-snickerdoodle/065ce041-ce56-476a-94f5-eb073fdc70c4%40googlegroups.com.

[Bush]

The "ATAT\rAT\r" part seems suspiciously like some handshaking protocol left over from the old Unix days. I'd be interested to see what your Ubuntu host output of stty -F /dev/ttyACM0 is. Can you confirm that the characters are sent only after you've started your terminal session (i.e. connect the microUSB but don't start a session for ~5 seconds to allow the boot process to initiate).

[weath...@krtkl.com]

'AT' codes were used on serial attached dialup modems to control the modem modes inband with the data traffic.
For instance to dial out and set baud rates.  Fun days.

[Charles Glasser]

I've noticed that several times the system will come up an stop in U-Boot. If you type reset the system will continue on to the snickerdoodle prompt.

cg

--
You received this message because you are subscribed to a topic in the Google Groups "snickerdoodle forum" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/krtkl-snickerdoodle/E9Oe-cnFGJE/unsubscribe.
To unsubscribe from this group and all its topics, send an email to krtkl-snickerdo...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/krtkl-snickerdoodle/4b236c30-a4f7-4eb0-abd5-05aea72546b0%40googlegroups.com.

[mathew...@gmail.com]

If you're using Ubuntu for your host, remove the Modem Manager with "sudo apt-get remove modemmanager".  Modem Manager considered /dev/ttyACM0 to be a modem and tries to configure it by sending a series of AT commands to it.  The first command interrupts the uboot process. 

[Tom Olenik]

Hello all!

I just wanted to say Mathew's instructions below fixed the issue I was having yesterday. Spread the word. 

Tom Olenik

[Bush]

Can you confirm the detection of the USB connection on your Linux host with dmesg?

On Tuesday, February 28, 2017 at 4:59:02 AM UTC-8, MichaelS wrote:

I seem to have a problem getting snickerdoodle to come alive. Specifically, it seems that my host machine is unable to recognize the device. When connected the green LED near the USB connection will flash very briefly, and the main LED will start pulsing in a heartbeat pattern. The issue is that the device will not show up in device manager (on windows). On linux it will not show up as ttyACM0 (or anything else it seems).

I have done the following (on a windows 8.1 host machine)

1.      1.  Format a high speed microSD card

2.      2. Downloaded the snickerdoodle Ubuntu SD card image from http://krtkl.com/downloads/

3.      3.  Used Win32DiskImager to write the image to the SD Card.

4.      4.  Ejected the SD card and placed it in the receptacle of snickerdoodle.

Furthermore I have tried:

1.      1.  Downloading STMicroelectronics driver

2.      2. Run the installer (VCP_1.4.0_Setup)

3.      3.  Run dpinst_amd64

But the device does not show up in device manager.

I’ve also tried connecting to an Ubuntu host, but there seems to be no change to the /dev directory when I connect the snickerdoodle.

Could anyone point me in the right direction? Any help or hint much appreciated.

[Nick Burkitt]

If you have one, it's never a bad idea to try a second cable, just in case.

-Nick

On Wednesday, March 1, 2017 at 7:22:41 AM UTC-8, MichaelS wrote:

Running;

sudo dmesg | grep STM

as suggested in the book doesn't show anything related to STMicroelectronics, and shows no change whether or not it is connected.

It should be noted that i am very new to linux.