Phase 2: UART the one

Finally resolved.

Modified test_determine_fractional_divider_values() to not read back the values and instead just write 0, 1, and 2 followed by the current MULVAL and DIVADDVAL values. I saw several times where it appeared ("012") under minicom so once again I dumped the data to a file

stty raw
cat >/tmp/serial.dat </dev/ttyS0
hexdump /tmp/serial.dat

Got back the following:

0000200 c0e0 4416 0327 fe0f 24c0 e6ce 40b6 fe18
0000210 6c80 8c1c 806c f823 4c80 8c3c 804c f04c
0000220 c800 c678 361c 0660 00ff f0d8 380c 0032
0000230 c0e0 e098 7818 e666 9880 18e0 98c3 3006
0000240 e07e c038 8738 0638 0060 aab4 4c82 0c15
0000250 2b88 e108 a458 144b 90ff 962c f880 3130
0000260 0632 7005 7271 0606 8d60 6626 e0d0 1ae0
0000270 2364 0460 34c0 7347 b073 c006 8e24 c4c4
0000280 16c0 80fc 1c6c 8c8c 2780 80f8 384c 8c8c
0000290 7080 00f0 78c8 1e42 6038 fc0e c800 84f0
00002a0 c338 f800 00c0 e098 7818 0086 80fe eab4
00002b0 0482 0c55 a128 e08d 5648 c11c 2498 48d2
00002c0 b0fc 9999 f890 3130 0732 7006 7271 0707
00002d0 8560 7626 e118 0ae0 f424 82b0 16c0 8c23
00002e0 220f c0f0 c624 f2e2 8360 24c0 c48e c0e4
00002f0 fc38 6c80 8c1c 80ec f87b 4c80 8c38 80cc
0000300 f07c cc00 6378 3e0e c060 2aa4 04c2 1955
0000310 d52c f848 42c8 c110 a458 248b 98ff 1624
0000320 fc50 b130 88b2 3086 3231 0708 7170 0872
0000330 6008 2685 5886 e0e1 240b 1004 e0c1 6412
0000340 0423 fe17 34c0 7347 1081 c007 8e24 04c6
0000350 0720 2cc0 cc8e 810c f83c 6c80 8c1c 030c
0000360 f07e 6ab4 1482 10d5 b56c fe01 2b88 e114
0000370 9658 1211 a418 5433 90fe 562c fc68 3130
0000380 0932 3007 3231 0809 7170 0972 6019 2685
0000390 8896 60f0 260b d096 e0c1 641a 06a3 fe18
00003a0 16c0 3127 9633 c0f0 c624 12e2 e020 24c0
00003b0 c48e 8124 fc3f 2aa4 14c2 0055 356c ff28
00003c0 aba8 f891 52c8 c218 a458 24cb 98ff 9624
00003d0 fc60 99b0 4599 30f0 3231 080a 3130 0a32


Randomly chose one occurance of 3130 0732 xx06. This amounts to a MULVAL of 7 and a DIVADDVAL of 6.

Then i modified IE_setup.c's setup_uart() with the above fractional divide values then main.c to write out 0's non stop. This produced the following output under minicom (using the db8 external ttl-rs232 adapter).

000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000


Occasionally, it would deteriorate and when I checked the output of RX (null modem) on a scope I saw the signal attenuated to between 0 and 3v but squeezing the max232 chip fixed it ... which means its probably a solder issue (now that the fractional divider value has been found).

Ill fix up the mainboard db9 connector and use that. I might even wire it up in a null modem (swap RX and TX) so I can stick with the straight through serial cable I am using ...