[minicoredumper] [External] Infinite loop while executing read_remote

[Jos Hulzink (Ellips B.V.)]

> -----Original Message-----
> From: John Ogness <john.ogness at linutronix.de>
> Sent: dinsdag 3 augustus 2021 22:53
> To: Jos Hulzink (Ellips B.V.) <jos.hulzink at ellips.com>;
> A few more questions:
> 
> Are you using a special (maybe hardened) dynamic loader that is not
> setting DT_DEBUG?

Not that we are aware of. We use the default dynamic loader, and we only tell the loader to load immediately (-Wl,-z,now) , but that't about it. Since we are able to debug the application with the generated core dump if minicoredumper succeeds, I assume we have DT_DEBUG.

> Have you tested minicoredumper on your system? (For example, with a
> simple crashing program or by sending the SEGV signal to a program?)

It works reliable with small test programs, and also most of the time with the application that we are discussing now. The issue we are facing only occurs sporadically, but is annoying for us since it locks up systems without keyboard or display attached at the customer site.

> If you send the SEGV signal to the reported program, does it always
> result in this error?

Unfortunately not. It only occurs sporadically. We have multiple deliberate exception triggers in our software, but none of them is a reliable tool for reproduction:
- write to a null pointer
- write to random addresses
- stack overflow
- division by 0
- kill SIGABRT

Kind regards,

Jos Hulzink