Welcome, Guest. Please login or register.
Amiga Kit Amiga Store Iridium Banner AMIStore App Store A1200/A600 4xIDE Interface

AuthorTopic: Pagestream4 Crash and MuForce Hit in 3.1.4  (Read 1847 times)

0 Members and 1 Guest are viewing this topic.

guest11527

  • Guest
Re: Pagestream4 Crash and MuForce Hit in 3.1.4
« Reply #15 on: December 05, 2019, 07:10:59 PM »
A few other hits I got in Pagestream and also FinalWriter97 (these don't look like the previous issue).
I afraid there is not sufficient information in this output to say anything. Please run segtracker in the startup-sequence so we know where the PC is located, and please also add "stacklines=32" to MuForce which would allow to trace back where the call came from.

Thanks,
Thomas
 

Offline my_pc_is_amiga

Re: Pagestream4 Crash and MuForce Hit in 3.1.4
« Reply #16 on: December 06, 2019, 05:31:26 AM »
I'm just focusing on the FW one for the time being.   I actually had SegTracker in the startup-sequence and I have things in this order.

SetPatch
SegTracker
MuForce  (and I actually had stacklines=32)
MuGa

What is strange is if I remove MuGa from my startup, it doesn't seem to crash. I tried several times and FW is not crashing.  I am using the MMU libs 68060 library., etc.   Once I have MuGa there in the startup then it crashes.   

Also, what is strange is the "FW97_Process.1" does now show up in the SegTracker List.  There is a "FinalWriter' and a bunch of FWLibs in the list but not the "FW97_Process.1"

I have another dump but with stacklines=128.   

 

guest11527

  • Guest
Re: Pagestream4 Crash and MuForce Hit in 3.1.4
« Reply #17 on: December 06, 2019, 07:22:56 AM »
What is strange is if I remove MuGa from my startup, it doesn't seem to crash. I tried several times and FW is not crashing.  I am using the MMU libs 68060 library., etc.   Once I have MuGa there in the startup then it crashes.   
So MuGA works. It is supposed to trigger a hit for bad memory access, and apparently, FW accesses memory it already has released.

Also, what is strange is the "FW97_Process.1" does now show up in the SegTracker List.  There is a "FinalWriter' and a bunch of FWLibs in the list but not the "FW97_Process.1"
That's not strange. The process name does not have to be identical to the name of the binary where the process was loaded from.

But if the issue goes away without MuGA, we already know that FW accesses memory it released before, i.e. performs an invalid memory access.