Re: Segfault after upgrade.

[Kai Harrekilde-Petersen]

David,

The OS never generates a segmentation violation, only programs does that 
and the OS reports back that the program did something illegal.
If the OS makes an error, you usually get a kernel panic/oops message 
(and the system goes belly-up).

Most of the strace just shows the dynamic libraries being set up (much 
akin to .DLL on Windows).

The bit you quoted below shows that the program tries to access an 
unmapped memory location, at address 0x5343ce. You would have to run 
cgps under a debugger (e.g. ddd or gdb) to determine why.

Kai

On 24 May 2020 12.49, David J Taylor wrote:
> The next step would be to run the command under strace -
>
>     strace -f <path to gpsd>/gpsd <and whatever options you are 
> feeding it>
>
> At least, that's the form I always use, the -f means it 'follows' child
> processes that are created by the main process. Beware that strace can
> generate scads of output, and much of it can appear near 
> incomprehensible.
> You can direct the output to a file using the -o option.
>
> I can never remember whether strace is part of the default installation,
> so you may have to 'apt install strace'.
>
> Despite the inscrutable nature of the output, when something is
> segfaulting, the clues tend to stick out, either by breaking a pattern in
> the output that you can perceive, or in a rather blatant chunk of output
> at the very end before it exits.
>
> Paul Theodoropoulos
> www.anastrophe.com
> ====================================================
>
> Thanks, Paul.  Yes, strace is installed, so I ran it on cgps which 
> also segfaults.
>
> Here are the last few lines of output:
>
> ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> mprotect(0x507000, 4096, PROT_READ)     = 0
> mprotect(0xb6fa5000, 4096, PROT_READ)   = 0
> munmap(0xb6f4e000, 80160)               = 0
> set_tid_address(0xb6f9e078)             = 19536
> set_robust_list(0xb6f9e080, 12)         = 0
> rt_sigaction(SIGRTMIN, {sa_handler=0xb6ddf8e8, sa_mask=[], 
> sa_flags=SA_RESTORER|SA_SIGINFO, sa_restorer=0xb6cba130}, NULL, 8) = 0
> rt_sigaction(SIGRT_1, {sa_handler=0xb6ddf9a4, sa_mask=[], 
> sa_flags=SA_RESTORER|SA_RESTART|SA_SIGINFO, sa_restorer=0xb6cba130}, 
> NULL, 8) = 0
> rt_sigprocmask(SIG_UNBLOCK, [RTMIN RT_1], NULL, 8) = 0
> ugetrlimit(RLIMIT_STACK, {rlim_cur=8192*1024, rlim_max=RLIM_INFINITY}) 
> = 0
> brk(NULL)                               = 0x17e6000
> brk(0x1807000)                          = 0x1807000
> --- SIGSEGV {si_signo=SIGSEGV, si_code=SEGV_MAPERR, si_addr=0x5343ce} ---
> +++ killed by SIGSEGV +++
> Segmentation fault
> pi@RasPi-3:~ $
> ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
>
> I've kept the whole 430+ lines if needed and uploaded it to:
>
>  https://www.satsignal.eu/temp/seg-fault.txt
>
> I can't see whether this is in the OS or the cgps program, though.  
> Again, it's /both/ cgps and gpsd which segfault, so perhaps something 
> in a common library?
>
> Cheers,
> David