Sorry i haven't followed up- this time of year is obscenely busy for me both at work and outside.
The boot process simply output text + ANSI colorization escape sequences. they are output at multiple times during the 1st and 2nd stages of the linux boot process for each step in the rc.d/init.d sequence.
As soon as I can come up for air - hopefully next week - i'll sit down with this and do some more digging.
Some further testing -- when the BPv4 locks up, the USB device actually times out, such that I get the following spewed into my dmesg: tty_port_close_start: tty->count = 1 port count = 0.
Got to console, and created a 12kb text file. 'cat' to the console should lock up the device, right?
but once didn't... and 3MB of text ( base64 </boot/uImage) didn't lock it up either....
So now I get to go through my boot process and see if something strange is being sent to the serial port that is causing the lockup -- or if it's a weird timing issue that only shows up in RECV before XMIT has been used much...
*edit* I've re-written one of the startup scripts that was the last to run before UART lockup, and I'm no longer getting lockup on boot. It did output some ANSI escape sequences, so maybe a weird bit combination and timing combination?
I'll look at some more debugging probably next week, but for now, I've got a workaround in place, and I'm able to debug my Shiva system which is what this is all about, in the end.
115200,N,8,1 normal output mode. Oddly, as soon as I configure the serial port, but before entering into bridge mode, it tells me there's a "possible bufferoverflow" -- even tho the serial port has no data (the client is powered off)
after turning on the 'plug and letting it try to boot, i get to the lock-up after ~11,945 bytes. Holding down the Normal button for 30s a few times does nothing.
Also just noticed this: While received characters are formatted correctly, transmitted data that is echoed back is not correct.
.. and the auto-built hexfile (which is 1/3 as small as the official 6.1 !!) actually works on my BPv4.
Caveat being that UART mode is still buggy - I don't have a exact byte count for when it locks up, but it's about 5/8's the way through my (custom) Shivaplug boot process.
I can't send chars to interrupt the boot process (they echo tho.. that's odd)
Once the UART lock-up happens, I can no longer 'Normal' to escape bridge mode, but have to full-on 'Reset'
Just received my BPv4 today, and I've already run into the UART issue...
I don't really have the background to help out, but I'm definitely willing to test pre-built firmwares against various uart-y things.
I did download MPLABX and tried to compile from SVN... but I just get lots of build errors. this whole IDE thing is new to me. Give me a bare makefile and I'm happy.
