Re: [Openocd-development] Segfault in rev 983

[Duane E. <du...@du...>]

Nick Foster wrote:

[help me - openocd is crashing, and gave a very good state of the problem]

The problem does not occur w/ my arm7tdmi target.

I can confirm I have the same type (not exactly same) of problem with my 
stm32 board.

You are running Linux, I am running under cygwin.

That is good news - You are not alone.

Now the bad news ;-(

Part 1 of the problem -
    Was caused by Oyvinds recent changes dealing with "keep_alive".
   
    This has introduced a problem, keep_alive() needs a recursion block.  
    [Perhaps there is more to it, see below]

    Your back trace shows the proof.
    @ backtrace stack positions #4, #9, and #14 - show "keep_alive()" 
has gone recursive.
       
        #4 0x080995a3 in keep_alive () at log.c:37
        #9 0x080995ad in keep_alive () at log.c:383
        #14 0x080995ad in keep_alive () at log.c:383

    That is *not* good.
    That needs a fix.

Part 2 - Sadly there is a part 2.

    I put a recursion block in - keep_alive (located in 'log.c' of all 
places)
    I still get an error :-( - it dies in another place.
    I'm not as familiar with that section.

    Perhaps - something needs to be done "higher up".
    I'm not familiar with how Oyvind made that change.

    FYI - it crashes does *WITH* and *WITHOUT* my new patch
    (I just submitted)

    In my case (part 2) it is crashing deep inside the jtag code.

    I can't do much else tonight.

-Duane.