Re: [hatari-devel] Improved internal timers performances in cycInt.c

[Christian Zietz <czietz@xxxxxxx>]

gdb intercepts the signal, afaik, and breaks into the debugger. Resend SIGINT to Hatari by entering at the gdb prompt:

 

signal 2

 

Then Hatari should get the SIGINT (Ctrl+C). See if you get a segmentation violation then.

 

Regards

Christian

 

Gesendet: Montag, 06. Dezember 2021 um 13:59 Uhr
Von: "Uwe Seimet" <Uwe.Seimet@xxxxxxxxx>
An: hatari-devel@xxxxxxxxxxxxxxxxxxx
Betreff: Re: [hatari-devel] Improved internal timers performances in cycInt.c

Looks as if I only send my latest reply to you, not to the list. All in all,
in gdb I don't get a segmentation fault. I also get it with kill SIGINT,
but when I do this in gdb, gdb just stops hatari and sys that SIGINT was
received.

> Le 06/12/2021 à 13:26, Uwe Seimet a écrit :
> > I just double-checked: The problem is caused by commit 12495caf. With
> > commit 7c701c53 (the one before) I don't get a segmentation fault.
> >
>
> ok ; but as Christian wrote, do you have a gdb backtrace when SEGV
> occurs instead of the one with SIGINT ?
>
> >> Uwe Seimet schrieb:
> >>
> >>> Thread 1 "hatari" received signal SIGINT, Interrupt.
> >>
> >> Wait, isn't SIGINT the *expected* reaction to Ctrl+C? If so, then you
> >> did not capture the actual crash / segmentation fault (SIGSEGV) there.
> >>
> >> Regards
> >> Christian
> >> --
> >> Christian Zietz - CHZ-Soft - czietz@xxxxxxx
> >> WWW: https://www.chzsoft.de/
> >> PGP/GnuPG-Key-ID: 0x52CB97F66DA025CA / 0x6DA025CA
> >>
> >>
> >
> >
>