We have a Serial-to-USB cable connected to the Serial interface of the board and a USB port of the testing machine. Then you can catch the output via 'picocom' for instance. The implementation of the UART driver is [1]. It uses 1 stop bit, a baudrate of 115200, and no parity. So, the following works for me:
sudo picocom -p n -b 115200 /dev/ttyUSB0
Even if your Genode doesn't come up you should at least see some output from the Uboot like this:
U-Boot 2015.10-00237-g92e40f5-dirty (Dec 04 2015 - 15:37:04 +0100) CPU: Freescale i.MX6Q rev1.2 at 792 MHz Reset cause: POR Board: Wandboard rev B1 I2C: ready DRAM: 2 GiB MMC: FSL_SDHC: 0, FSL_SDHC: 1 No panel detected: default to HDMI Display: HDMI (1024x768) In: serial Out: serial Err: serial Net: FEC [PRIME] Hit any key to stop autoboot: 1 U-Boot SPL 2015.10-00237-g92e40f5-dirty (Dec 04 2015 - 15:37:04)
Did this help?
[1] base/include/drivers/uart/imx.h
Am 18.08.2017 um 18:03 schrieb Steven Harp:
Martin:
Yes, that worked well. Happily I had the older toolchain handy, and it compiles uboot without error.
Unfortunately, the output is not yet working for me (no serial out). I'm still trying to puzzle out the recipe to get Genode images running on the Wand Quad. Genode Labs perhaps has a preferred solution for testing on this platform--is there a web page or README that discusses? Maybe tftpboot from one of the Wand linux/uboot images could work?
Any advice or examples would be much appreciated.
Thanks, Steve
On 08/18/2017 07:10 AM, Martin Stein wrote:
Hi Steven Harp,
Thank you for this hint! It's obviously a bug in the 'create_uboot' tool. The tool uses the Genode tool chain and since the last update of this tool cain I didn't test whether they're still fine with each other.
One quick work-around for you would be to download the 16.05 toolchain [1], unpack it to an individual directory (no 'P' flag) [2], and adapt your create_uboot to use the other tool chain [3].
I will have a more detailed look at the problem these days and provide a long-term fix as soon as possible.
I hope this helped?
Cheers, Martin
Martin:
That is much like the output I see with similar serial wiring and using the uboot that is included with the Wand Debian image. My Wand seems to be rev. D1 instead of B1, but hopefully that's not a problem. However the create_uboot image results in no serial output whatsoever.
1. Maybe some additional steps are required to prepare the microSD or make the image bootable, e.g. something analogous to repos/os/run/tz_vmm.run?
2. Is it feasible to just reconfigure the working uboot (from Debian) to bootstrap genode? If so, what files from wand_quad/var/run/... need to be loaded onto the microSD, and what uboot environment and commands are required?
// Steve
On 08/21/2017 04:49 AM, Martin Stein wrote:
We have a Serial-to-USB cable connected to the Serial interface of the board and a USB port of the testing machine. Then you can catch the output via 'picocom' for instance. The implementation of the UART driver is [1]. It uses 1 stop bit, a baudrate of 115200, and no parity. So, the following works for me:
sudo picocom -p n -b 115200 /dev/ttyUSB0
Even if your Genode doesn't come up you should at least see some output from the Uboot like this:
U-Boot 2015.10-00237-g92e40f5-dirty (Dec 04 2015 - 15:37:04 +0100) CPU: Freescale i.MX6Q rev1.2 at 792 MHz Reset cause: POR Board: Wandboard rev B1 I2C: ready DRAM: 2 GiB MMC: FSL_SDHC: 0, FSL_SDHC: 1 No panel detected: default to HDMI Display: HDMI (1024x768) In: serial Out: serial Err: serial Net: FEC [PRIME] Hit any key to stop autoboot: 1 U-Boot SPL 2015.10-00237-g92e40f5-dirty (Dec 04 2015 - 15:37:04)
Did this help?
[1] base/include/drivers/uart/imx.h
Am 18.08.2017 um 18:03 schrieb Steven Harp:
Martin:
Yes, that worked well. Happily I had the older toolchain handy, and it compiles uboot without error.
Unfortunately, the output is not yet working for me (no serial out). I'm still trying to puzzle out the recipe to get Genode images running on the Wand Quad. Genode Labs perhaps has a preferred solution for testing on this platform--is there a web page or README that discusses? Maybe tftpboot from one of the Wand linux/uboot images could work?
Any advice or examples would be much appreciated.
Thanks, Steve
On 08/18/2017 07:10 AM, Martin Stein wrote:
Hi Steven Harp,
Thank you for this hint! It's obviously a bug in the 'create_uboot' tool. The tool uses the Genode tool chain and since the last update of this tool cain I didn't test whether they're still fine with each other.
One quick work-around for you would be to download the 16.05 toolchain [1], unpack it to an individual directory (no 'P' flag) [2], and adapt your create_uboot to use the other tool chain [3].
I will have a more detailed look at the problem these days and provide a long-term fix as soon as possible.
I hope this helped?
Cheers, Martin
Check out the vibrant tech community on one of the world's most engaging tech sites, Slashdot.org! http://sdm.link/slashdot _______________________________________________ genode-main mailing list genode-main@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/genode-main
I am increasingly suspicious that the uboot that is built by create_uboot is not happy with rev D1 of the wandboard quad. The SPL first stage, installed in sector 1 per instructions at contrib/uboot/board/wandboard, seems to be a non-starter: no serial output at all. The source code suggests it knows about rev B1 and C1, and anything that is not rev C1 is assumed to be rev B1.
Using the bootstrap from the wandboard Debian, the u-boot.img from create_uboot will load, and provides the messages shown below. It is clearly having difficulties doing some i2c transaction. MMC commands beyond "info" all hang the cpu, so no Genode loading possible.
* * * * * * * * * * * * * * * * * * * * * * * * *
U-Boot SPL 2015.04-gc42b934 (May 02 2017 - 11:51:42) reading u-boot.img reading u-boot.img
U-Boot 2015.10-rc4-gd797dd5 (Aug 18 2017 - 14:10:37 -0500)
CPU: Freescale i.MX6QP rev1.0 at 792 MHz Reset cause: POR Board: Wandboard rev B1 I2C: ready DRAM: 2 GiB MMC: FSL_SDHC: 0, FSL_SDHC: 1 *** Warning - bad CRC, using default environment
wait_for_sr_state: failed sr=81 cr=a0 state=2020 i2c_init_transfer: failed for chip 0x10 retry=0 wait_for_sr_state: failed sr=81 cr=a0 state=2020 i2c_init_transfer: failed for chip 0x10 retry=1 wait_for_sr_state: failed sr=81 cr=a0 state=2020 i2c_init_transfer: failed for chip 0x10 retry=2 i2c_init_transfer: give up i2c_regs=0x21a4000 No panel detected: default to HDMI Display: HDMI (1024x768) In: serial Out: serial Err: serial Net: FEC [PRIME] Hit any key to stop autoboot: 0 => * * * * * * * * * * * * * * * * * * * * * * * * *
Maybe there is something else missing....
Happily, the uboot that came with Debian is able to load and start my application uImage (seL4)--this is a good enough work-around for me for now.
I did notice two, apparently non-fatal, Genode errors: ... [init] child "timer" announces service "Timer" Error: mapping failed 0x4a9000 -> 0xa6000 Error: mapping failed 0x476000 -> 0x73000 ...
This seems like a different topic. I'll start another thread if anyone has cycles to look at this.
// Steve
On 08/21/2017 04:49 AM, Martin Stein wrote:
We have a Serial-to-USB cable connected to the Serial interface of the board and a USB port of the testing machine. Then you can catch the output via 'picocom' for instance. The implementation of the UART driver is [1]. It uses 1 stop bit, a baudrate of 115200, and no parity. So, the following works for me:
sudo picocom -p n -b 115200 /dev/ttyUSB0
Even if your Genode doesn't come up you should at least see some output from the Uboot like this:
U-Boot 2015.10-00237-g92e40f5-dirty (Dec 04 2015 - 15:37:04 +0100) CPU: Freescale i.MX6Q rev1.2 at 792 MHz Reset cause: POR Board: Wandboard rev B1 I2C: ready DRAM: 2 GiB MMC: FSL_SDHC: 0, FSL_SDHC: 1 No panel detected: default to HDMI Display: HDMI (1024x768) In: serial Out: serial Err: serial Net: FEC [PRIME] Hit any key to stop autoboot: 1 U-Boot SPL 2015.10-00237-g92e40f5-dirty (Dec 04 2015 - 15:37:04)
-
Martin Stein -
Steven Harp