On Mon, 21 Dec 2020 00:14:47 +0200, in gmane.comp.hardware.beagleboard.user
Robert Berger <
robert.karl.berger-R...@public.gmane.org>
wrote:
>
>Pocketbeagle
>
>I posted it under "pocketbeagle" and thought that would be more obvious,
>but it does not seem to be.
>
I read via gmane's NNTP server where things are just
Newsgroups: gmane.comp.hardware.beagleboard.user
I don't see any sub-categories.
(I used to reply via gmane also, but a year or so ago something
glitched and my replies began to bounce. I now have to remember to use
"reply via email" and change the To: address to Beagleboard
<
beagl...@googlegroups.com>)
>>
>>> This seems extremely unstable and random.
>>
>> What results have you obtained that cause you to make that statement.
>
>screen /dev/ttyUSB4 115200
>
> From "nothing" to some gibberisch (non ASCII characters, sometimes some
>ASCII characters)
>
>after the board is booted screen /dev/ttyACM0 115200 works fine, but
>that's NOT what I want.
>
Both of those are host side references -- I'm on Windows so can't
really help, though if there is an option to ensure 8N1 that might be
useful.
>Board Adapter
>
>Txd Rxd
>Rxd Txd
>GND GND
>
Which board PINS? {Just for verification}
>
>Not really everything the board sends. Out of the box only stuff from
>the time the "virtual" serial port starts.
>
There is no "virtual" serial port on the beagle side. Virtual serial
ports exist on the host side of a USB adapter, and since the adapter is
powered from the host computer, it should be available as soon as the host
enumerates the port. This is not dependent on having connections on the
3.3V TTY lines (I just hung a FTDI RS232 [one with a DB-9 connector, so has
much higher voltages:
https://www.amazon.com/Gearmo-RS-232-Adapter-Indicators-Windows/dp/B00AHYJWWG/ref=redir_mobile_desktop?ie=UTF8&aaxitk=qFyAraevFSYJRhyMvpzZ6g&hsa_cr_id=6549072620501&ref_=sbx_be_s_sparkle_mcd_asin_0
] off a USB port on my monitor, and am able to configure it without having
anything on the DB-9 side -- I do have the Adafruit [I think] FTDI board,
but have never used it; besides BeagleBone Black routes the boot console to
a set of dedicated debug pins, not the regular jumpers).
>I see the "virtual" serial port, but I would like to see the "console"
>output starting from the first line of MLO.
Until u-boot configures the console UART parameters, there should be
nothing available. The UART is a hardware serial port, nothing "virtual" on
that side.
>
>debian@beaglebone:~$ cat /proc/cmdline
>console=ttyO0,115200n8 root=/dev/mmcblk0p1 ro rootfstype=ext4 rootwait
>coherent_pool=1M net.ifnames=0 lpj=1990656 rng_core.default_quality=100
>quiet
>
Interesting -- my BBB cmdline includes a clause for the capemgr:
bone_capemgr.uboot_capemgr_enabled=1
>I am also a bit confused about the ttyO0, although there is also a ttyS0;)
ttyOx was the older TI "OMAP" serial ports, newer kernels replace them
with ttySx. From dmesg (so Linux stage) on a BeagleBone Black:
[ 0.002772] WARNING: Your 'console=ttyO0' has been replaced by 'ttyS0'
[ 0.002776] This ensures that you still see kernel messages. Please
[ 0.002780] update your kernel commandline.
{later}
[ 0.991655] Serial: 8250/16550 driver, 6 ports, IRQ sharing disabled
[ 0.994459] 44e09000.serial: ttyS0 at MMIO 0x44e09000 (irq = 30,
base_baud = 3000000) is a 8250
[ 1.004605] console [ttyS0] enabled
[ 1.005573] 48022000.serial: ttyS1 at MMIO 0x48022000 (irq = 31,
base_baud = 3000000) is a 8250
[ 1.006394] 48024000.serial: ttyS2 at MMIO 0x48024000 (irq = 32,
base_baud = 3000000) is a 8250
[ 1.007214] 481a6000.serial: ttyS3 at MMIO 0x481a6000 (irq = 33,
base_baud = 3000000) is a 8250
[ 1.008252] 481a8000.serial: ttyS4 at MMIO 0x481a8000 (irq = 34,
base_baud = 3000000) is a 8250
[ 1.009041] 481aa000.serial: ttyS5 at MMIO 0x481aa000 (irq = 35,
base_baud = 3000000) is a 8250
Once the kernel is booted, the ttyOx devices are links to the ttySx
devices.
lrwxrwxrwx 1 root root 5 Dec 11 13:48 /dev/ttyO0 -> ttyS0
lrwxrwxrwx 1 root root 5 Dec 11 13:48 /dev/ttyO1 -> ttyS1
lrwxrwxrwx 1 root root 5 Dec 11 13:48 /dev/ttyO2 -> ttyS2
lrwxrwxrwx 1 root root 5 Dec 11 13:48 /dev/ttyO3 -> ttyS3
lrwxrwxrwx 1 root root 5 Dec 11 13:48 /dev/ttyO4 -> ttyS4
lrwxrwxrwx 1 root root 5 Dec 11 13:48 /dev/ttyO5 -> ttyS5
crw--w---- 1 root tty 4, 64 Dec 11 13:48 /dev/ttyS0
crw-rw---- 1 root dialout 4, 65 Dec 11 13:48 /dev/ttyS1
crw-rw---- 1 root dialout 4, 66 Dec 11 13:48 /dev/ttyS2
crw-rw---- 1 root dialout 4, 67 Dec 11 13:48 /dev/ttyS3
crw-rw---- 1 root dialout 4, 68 Dec 11 13:48 /dev/ttyS4
crw-rw---- 1 root dialout 4, 69 Dec 11 13:48 /dev/ttyS5
debian@beaglebone:~$
{Not sure why they show December 11 -- unless that is the last time I'd
booted the board, and the device entries are created before NTP reset the
time}
>Not really.
>
>1) Quite a few boards have TTL or real RS-232 outputs. So why go to USB
>and not directly to good old RS-232?
My point was that many modern host computers have dropped RS-232 ports.
(heck, neither my decade old laptop, and my nearly as old desktop have
RS-232 ports native) This means that, to use an RS-232 connection, requires
one to have a USB<>RS-232 adapter in the system. So having such a cable on
the host computer, and then using an RS-232<>TTL adapter on the Beagle, is
just adding complications -- especially as the latter adapter likely needs
to be powered by the Beagle. Just use a USB<>TTL adapter
https://www.adafruit.com/product/954 and avoid the voltage shifting (3.3V
TTL -> +/- 12V RS-232 -> 5V USB)
>Unfortunately some newer boards only have USB connectors. Those I
>connect via USB cables to USB hubs. But in any case I would like to see
>the complete boot log (from MLO) and not just when the kernel is available.
>
Unless MLO output is considered part of u-boot, it may not have console
output. I confess I'm not cognizant of how the boot files are made and what
the difference is between MLO and u-boot.
--
Dennis L Bieber