split up serial writes

[DavePutz]

Issue #1639 Application exception output to serial (over USB) appears truncated due to buffering issue

Added code in serial_write_substring() to split large writes into chunks to help avoid buffer overflows.
A limit of CFG_TUD_CDC_EPSIZE was used, since that is the limit used in tud_cdc_n_write_flush() from lib/tinyusb/src/class/cdc/cdc_device.c.
Prior to the fix there was a reproduction of the issue about 80% of the time using
>>> raise AttributeError('abcdefghij' * 12)
After the fix 50 cycles were run with no truncated outputs.

[tannewt]

I thought TinyUSB did this already. @hathach can you take a look at this?

[hathach]

@DavePutz could you please tell me how to reproduce the issue, also on which board as well. I will try to see if there is a bug within tinyusb stack.

[DavePutz]

Hi, I was able to reproduce it about 80% of the time using a CircuitPlayground Express connected to a Windows 10 system using Teraterm. I ran from the REPL:

>> raise AttributeError('abcdefghij' * 12)

This should return Traceback (most recent call last): File "<stdin>", line 1, in <module> AttributeError: abcdefghijabcdefghijabcdefghijabcdefghijabcdefghijabcdefghijabcdefghijabcdefghijabcdefghijabcdefghijabcdefghijabcdefghij When it fails it only returns Traceback (most recent call last): File "<stdin>", line 1, in <module> AttributeError The rest of the output only prints after you press any key. Please note that running on Linux on the same system I was not able to reproduce the error at all. Let me know if any more info would be of help. Thanks! -Dave

…

On Wed, Apr 15, 2020 at 1:42 AM Ha Thach ***@***.***> wrote: @DavePutz <https://github.com/DavePutz> could you please tell me how to reproduce the issue, I will try to see if there is a bug within tinyusb stack. — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#2771 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AFNJKEREEZO3IRCYNZBOXRDRMVJPFANCNFSM4MIDORUA> .

[hathach]

Ok, so it is windows only issue. It sounds like this tinyusb issue. There is a pending PR in the stack repo hathach/tinyusb#260

I will try to reproduce first then get back to you. One last question, does this happen to the latest CPY version 5.2 right ?

[DavePutz]

Hi, Yes, this reproduces on CPY 5.2. Also, I rewrote to use tud_cdc_write_available() but that still fails in the same way. -Dave

…

On Wed, Apr 15, 2020 at 11:40 AM Ha Thach ***@***.***> wrote: Ok, so it is windows only issue. It sounds like this tinyusb issue. There is a pending PR in the stack repo hathach/tinyusb#260 <hathach/tinyusb#260> I will try to reproduce first then get back to you. One last question, does this happen to the latest CPY version 5.2 right ? — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#2771 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AFNJKERMEAJFE53JT4BW3ADRMXPPJANCNFSM4MIDORUA> .

[hathach]

Hi, Yes, this reproduces on CPY 5.2. Also, I rewrote to use tud_cdc_write_available() but that still fails in the same way. -Dave

Thanks, I will try to reproduce it tomorrow to see if it is usb stack bug :)

[hathach]

@DavePutz I could reproduce it on my local machine. Will try to troubleshoot it. We will move back to the #1639 for discussion since this PR is already closed.