There is a small ~10uS positive-going glitch on the nTRST signal, due to the way that the MPSSE mode of ftdi chips is configured in the file ./openocd/share/openocd/scripts/interface/ftdi/olimex-...
The line
ftdi_layout_init 0x0808 0x0a1b
should rather be
ftdi_layout_init 0x0b08 0x0b1b
Bits 8 and 9 of the first 16-bit word of ftdi_layout_init
should be '1', so that the internal pull-ups for nTRST and nSRST are enabled.
Bit 8 of the second 16-bit word of ftdi_layout_init
should be '1', so that nTRST is initially configured in an output direction.
This is somewhat related to the discussions in Olimex ARM-USB-OCD SRST pull-down resistance (https://www.olimex.com/forum/index.php?topic=8010.0) and Can the ARM-USB-TINY-H put the SRST pin into a high impedance state? (https://www.olimex.com/forum/index.php?topic=8590.0)
The following table might be helpful:
Sig MPSSE PIN P/U DIR
TCK TCK/SK ADBUS0 0 1
TDI TDI/DO ADBUS1 0 1
TDO TDO/DI ADBUS2 0 0
TMS TMS/CS ADBUS3 1 1
??? GPIOL0 ADBUS4 0 1
. GPIOL1 ADBUS5 0 0
. GPIOL2 ADBUS6 0 0
. GPIOL3 ADBUS7 0 0
TRST GPIOH0 ACBUS0 1 1
SRST GPIOH1 ACBUS1 1 1
. GPIOH2 ACBUS2 0 0
LED GPIOH3 ACBUS3 1 1
. GPIOH4 ACBUS4 0 0
. GPIOH5 ACBUS5 0 0
. GPIOH6 ACBUS6 0 0
. GPIOH7 ACBUS7 0 0
Thank you for this contribution. Please also notify the OpenOCD community about your findings if you haven't already. Also please specify which unit exactly it applies too.
In our OpenOCD debuggers the SRST has big pull-down resistor. Value depends on exact unit: in ARM-USB-OCD-H it is 470k, in ARM-USB-TINY-H it is 1M, in ARM-USB-OCD it is 270k, in ARM-UBS-TINY it is again 270k. It might be good to have some pull-up on the SRST at your target (10k would be good value for the pull-up at the target, but no more than 50k). And yes, as you noted it can be toggled via software means it is connected via transistor to FT2232's ACBUS1, but maybe the software configuration is not proper.