Armaan Chowfin commented on a discussion: 
https://gitlab.rtems.org/rtems/rtos/rtems/-/merge_requests/1094#note_144216


Hi Gedare

I discovered a few more lm3s69xx uart problems- this time I was trying to get 
lm4f120 builds (testsuites/samples) to run on my tm4c123 target board, with 
picocom on host to display text.

1. _None of the samples (all use the simple console driver) could display 
text_. The simple console driver path does not initialize the uart port 
(enable, set baud). For VMs, qemu most probably does uart port initialization 
during machine init - `NEEDS_SIMPLE_CONSOLE` samples all work on the 
lm3s6965evb.
2. _ticker.exe exceeds tm4c123 memory limits in every configuration (1 task/4 
task/simple console/legacy console)._ This was creating a double fault:  
`_Internal_error() -> printk() -> bus error at uart->fr during write_polled()` 
. (verified by reading BFSR and MMFAR regs, see 
https://interrupt.memfault.com/blog/cortex-m-hardfault-debug) 

I moved uart initialization code (uart_clock_enable, baudrate set) to 
`bsp_start_hook_1()`). This fixes the uart prints for all samples and double 
fault too.

3. The TM4C123 datasheet (https://www.ti.com/lit/ds/symlink/tm4c123gh6pm.pdf, 
page \~650, 660) mentions a GPIOPCTL register `0x52C`) that needs to be 
configured to enable uart. The port register struct in rtems lm3s69xx does not 
track this, but the uart prints were fixed without adding pctl code. I'm not 
sure why this should happen...

Should I create an issue in rtos/rtems for all the lm3s69xx fixes? I could have 
a larger MR in a day or so.

Regarding the rtems-tools, it might be better to create a new issue in tools/.

-- 
View it on GitLab: 
https://gitlab.rtems.org/rtems/rtos/rtems/-/merge_requests/1094#note_144216
You're receiving this email because of your account on gitlab.rtems.org.


_______________________________________________
bugs mailing list
[email protected]
http://lists.rtems.org/mailman/listinfo/bugs

Reply via email to