Hi Jörg and Andreas, On 03/06/13 08:33, Andreas Fritiofson wrote: > On Sun, Jun 2, 2013 at 3:07 PM, Jörg Fischer <tur...@gmx.de > <mailto:tur...@gmx.de>> wrote: > Hi, > Am 02.06.2013 06:57, schrieb Stuart Longland (VK4MSL): > > I'm just trying to debug a problem that has cropped up out of the > blue. > > I have a couple of STM32F103VE-based microcontrollers which I'm > > programming using a Olimex ARM-USB-TINY-H JTAG programmer. > > > > Host environment is Open On-Chip Debugger > 0.7.0-dev-00080-g59c051c-dirty > > (2013-01-25-19:17), from the Gentoo ebuild. > > > > I'm invoking openocd with the following: > >> openocd -f interface/olimex-arm-usb-tiny-h.cfg -f target/stm32f1x.cfg > > > It probably won't help with your problem, but my standard reply to > people who are still using the old, flawed "ft2232" adapter driver is to > suggest to switch to the new "ftdi" driver instead. Assuming that > OpenOCD was built with --enable-ftdi, it should be a simple matter of > using the similarly named interface config file from the ftdi subfolder > instead, i.e openocd -f interface/ftdi/olimex-arm-usb-tiny-h.cfg ... in > your case.
Is that using the proprietary library from FTDI themselves or is there another one I missed? I'll use proprietary stuff if I have to, but I'd really like to avoid it where possible as the other driver has worked fine up until now. > > > > which up until now has worked fine, I've been able to telnet in to > port > > 4444, and flash my devices, even hook up gdb. I was doing some > > debugging today, no problems flashing the devices until suddenly, it > > stopped halting the device. No configuration files had been changed. > > Do you use Sleep or low power modes? Did you mess with the clock? > > > Usually, it won't even find the TAPs if the core is in low power mode. No sleep modes at this point. I might consider setting something up later but to my knowledge, there is no low power mode being invoked at this time. > >> in procedure 'reset' > > > > Turning up debugging I get: > > [..skipped..] > >> Debug: 398 74782 cortex_m.c:974 cortex_m3_assert_reset(): > target->state: reset > >> Debug: 399 74788 cortex_m.c:1065 cortex_m3_assert_reset(): Using > Cortex-M3 SYSRESETREQ > >> Debug: 400 74790 cortex_m.c:597 cortex_m3_halt(): target->state: > reset > >> [...] > > > > Is there any way, using openocd, to determine if it's something wrong > > with the targets (all three seem to have failed, and all were working > > earlier), or if the JTAG device has failed? > > Have you connected the SRST line? Then you could try: > "reset_config srst_only separate" > > > I agree that it sounds like an SRST problem. I know that some versions > of the Olimex dongles suffers from a design fault causing the reset line > to be pulled down slightly (or even clamped) even when not asserted. On > an STM32 this typically causes the level on the pin to be below the > guaranteed high level. Depending on a lot of factors (processor batch, > temperature, Vdd level, ...) this could be interpreted as an asserted > reset or not. If the reset pin is not properly released, you'll get > symptoms like you describe. Measure the SRST voltage and make sure it's > at Vdd level. Also check any other circuitry you may have connected to > the SRST pin. If you find a problem with the dongle, you may simply > disconnect the reset line since in many cases you can manage without it > (and the default is to not use it anyway). Ahh okay, so possibly environment related. I wasn't aware of the design fault on the Olimex dongles. I'll make a note to check this next weekend when I'm tinkering with the device. For the record, the device I'm programming is one of these: http://www.cutedigi.com/development-tools/arm-cortex-development-board/arm-cortex-m3-mini-stm32-dev-board-2-4-tft-lcd-touchscreen.html I'll have a look at the schematic later and see if SRST is connected. > Using a low JTAG clock might also be helpful: > "adapter_khz 10" > > > Unless there are electrical problems, the STM32 will reliably start up > with 1 MHz since it defaults to the 8 MHz RC at reset. I've been leaving at the defaults so far. As I say, it has been working rather well. After a little tweaking when I first got the dongle, things seemed to just work. Just that it stopped working out of the blue. OpenOCD has been a great asset, and I do appreciate your efforts in its maintenance and support. I'll investigate this weekend and see where I get. -- Stuart Longland (aka Redhatter, VK4MSL) I haven't lost my mind... ...it's backed up on a tape somewhere. ------------------------------------------------------------------------------ Get 100% visibility into Java/.NET code with AppDynamics Lite It's a free troubleshooting tool designed for production Get down to code-level detail for bottlenecks, with <2% overhead. Download for free and get started troubleshooting in minutes. http://p.sf.net/sfu/appdyn_d2d_ap2 _______________________________________________ OpenOCD-devel mailing list OpenOCD-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/openocd-devel
