Closed ivankravets closed 6 years ago
I'll investigate.
Edit: sorry - ignore my post as I just saw the pull request dealing with the issue above!
_This may be due to the GDB/OpenOCD startup/init "race condition" whereby GDB gets started and tries to make a GDB Remote Serial Protocol connection before OpenOCD is quite ready to handle it? This can happen with the socket/port connection but maybe the same issue can arise with the pipe/stdio connection too? See here for more on this issue when using the socket/port connection:
https://github.com/gnu-mcu-eclipse/eclipse-plugins/issues/186
If you can launch OpenOCD standalone and then get GDB to connect to it OK then that may be evidence in support of this race condition being the problem.
Hope this helps._
I know there were some changes in OpenCOD that I did not study thoroughly, so it might be so.
as far as I know, there are no changes in GNU MCU Eclipse OpenOCD compared to master related to this issue, and if ther are, they probably come from merging the RISC-V code.
I get this error message also after exporting a project from mbed. Target is a nrf51822, debug probe a LPC-LINK2 with cmsis-dap firmware. Calling oocd 0.10.0-3-20170826-1813-dev manually for flashing works, but starting to debug gives the e0e error code. Also when I try to start gdb manually and to connect with 'target remote localhost:3333'. When I use oocd 0.10.0 compiled from F. Chopin the debugger starts without errors.
looks the same as https://github.com/riscv/riscv-openocd/issues/115
can you confirm if this was fixed in the latest binary release?
Re-tested. It works with the latest GNU MCU Eclipse OpenOCD v0.10.0-6-20180112
Thanks and Happy New Year!
you're welcome, happy new year!
Hi @ilg-ul,
I don't know where is better to post this issue, would be thankful for a help. I can't reproduce this issue with
20170607-2132-dev
but withv0.10.0-2-20170622-1535
I can.Just pipe OpenOCD to GDB: