Add IPM console

[finikorg]

This PR adds IPM console which might be tested with up_squared_adsp.
Following options need to be enabled: CONSOLE & IPM and disable LOG_PRINTK

After that you can "flash" the board with:

$ west flash zephyr/boards/xtensa/up_squared_adsp/tools/up_squared_adsp_flash.sh <KEY>

And in another terminal

$ sudo boards/xtensa/up_squared_adsp/tools/mbterm.py 
*** Booting Zephyr OS build zephyr-v2.3.0-2086-g83a5af82007c  ***
Hello World! up_squared_adsp

Can be used with sanitycheck also, for example from inside zephyrproject:

$ zephyr/scripts/sanitycheck -M -c -p up_squared_adsp --device-test \
--west-flash="zephyr/boards/xtensa/up_squared_adsp/tools/up_squared_adsp_flash.sh,otc_private_key.pem" \
--device-serial-pty="$HOME/set_ambient zephyr/boards/xtensa/up_squared_adsp/tools/mbterm.py" \
-T zephyr/samples/hello_world \
-x="CONFIG_IPM=y" -x="CONFIG_CONSOLE=y" -x="CONFIG_LOG_PRINTK=n" -v

where $HOME/set_ambient is described here: #22724 (comment)

[andrewboie]

I'm super confused.

Something wrong with the existing IPM console driver?
See drivers/console/ipm_console_sender.c, drivers/console/ipm_console_receiver.c

[finikorg]

I'm super confused.

Something wrong with the existing IPM console driver?
See drivers/console/ipm_console_sender.c, drivers/console/ipm_console_receiver.c

@andrewboie I can modify this ipm_console_sender.c to make it usable as a console. At the moment it is a kind of library used in tests/drivers/ipm/ fot tests only as names sender and receiver implies (It can be probably moved to tests)

So it was easier to write proper ipm_console then modify those console senders and receivers and also tests.

@andyross also does not like the method used to send character: See https://github.com/zephyrproject-rtos/zephyr/pull/22724/commits/2419dac1060bb11757519f17e43da7a694a27967 (github generates wrong link)

[andrewboie]

At the moment it is a kind of library used in tests/drivers/ipm/ fot tests only as names sender and receiver implies (It can be probably moved to tests)

No, it's not a test library. We used it to forward console output on Arduino 101 from the ARC CPU to the Lakemont, it was enabled by the default for that board.

The name "sender" is the side sending the console messages (on Arduino 101, the ARC CPU)
The name "receiver" is the code running on the other CPU (on Arudiuno 101, the Lakemont) which receives the console data and prints it.

I don't care if you rewrite it but if you want to rewrite how IPM console works to make it line-buffered you need to remove all of this code or amend it. Not have two completely different implementations for the same thing in-tree, that's a hard NACK.

[finikorg]

At the moment it is a kind of library used in tests/drivers/ipm/ fot tests only as names sender and receiver implies (It can be probably moved to tests)

No, it's not a test library. We used it to forward console output on Arduino 101 from the ARC CPU to the Lakemont, it was enabled by the default for that board.

But now it is enabled only by test.

The name "sender" is the side sending the console messages (on Arduino 101, the ARC CPU)
The name "receiver" is the code running on the other CPU (on Arudiuno 101, the Lakemont) which receives the console data and prints it.

I don't care if you rewrite it but if you want to rewrite how IPM console works to make it line-buffered you need to remove all of this code or amend it. Not have two completely different implementations for the same thing in-tree, that's a hard NACK.

As I said @andyross gave strong NACK with one character per blocking send method, even wrote small essay ;).

So at the moment nobody uses this code, if everybody OK I will remove it. I am also OK reusing console sender, I still have support in my python terminal tool with getting console characters in ID field. I just need for that enabling console sender and removing or modifying test.

[nashif]

why is this being removed? I know it was used in A101, but there is nothing here that is related to 101 and we have a test. So I would leave it, we might get users for this in the future.

[finikorg]

@nashif Do you mean drop the commit 58e0c009ab5fb0df07d33fa365f655d1d0956b6f ?

[nashif]

yes