I've never seen it, but I can tell you what is going on which would
hopefully help you get to the bottom of the problem. At the beginning
of the simulation any device that responds to a range sends a
sendStatusChange(Status:RangeChange) message to its peer. If the peer
is interested in then asks for the addresses and they are propagated
through the memory system. If you run with the trace flag
BusAddrRanges set you'll see the range changes happen.
I would guess you now have a new device in the system that isn't
connected to anyone and so it's sends status change and it has no
peer. Although I don't know why a device can do that would have a
port that was defaulting to NULL.
The trick is to find out which device it is and then see why that
port isn't connected. If you're in gdb you should be able to call name
() when sendStatusChange() is called and hopefully that will tell you
what device/object that port belongs to. If not doing up one frame in
the call stack and calling name() should provide some info.
Ali
On Oct 30, 2007, at 6:33 PM, Rick Strong wrote:
Hi,
I have recently tried check-pointing two threads (ammp, applu)
running on two separate AtomicSimpleCPUs when the first thread
reaches 100,000 instructions. An attempt to restore the checkpoint
meets the error given below as "ERROR". Anyone see something like
this before.
-Richard
ERROR:
"""
Copyright (c) 2001-2006
The Regents of The University of Michigan
All Rights Reserved
M5 compiled Oct 30 2007 10:02:18
M5 started Tue Oct 30 15:16:22 2007
M5 executing on 0-16-cb-b8-f-b2.dynamic.ucsd.edu
command line: /Library/WebServer/Documents/research/m5/m5-2.0b3/
build/ALPHA_SE/m5.debug -d ../results configs/example/myse.py -r 1
--bench=applu;ammp --checkpoint_dir=checkpoints -d --caches --
l2cache -n 2
SMT workload :['ammp', 'applu']
['ammp', 'applu']
Running benchmark: applu;ammp
Accessing cpu_class
Global frequency set at 1000000000000 ticks per second
Program received signal EXC_BAD_ACCESS, Could not access memory.
Reason: KERN_PROTECTION_FAILURE at address: 0x00000008
0x00184ed3 in Port::sendStatusChange (this=0x0,
status=Port::RangeChange) at port.hh:210
210 void sendStatusChange(Status status) {peer-
>recvStatusChange(status); }
(gdb)
"""
_______________________________________________
m5-users mailing list
[email protected]
http://m5sim.org/cgi-bin/mailman/listinfo/m5-users
_______________________________________________
m5-users mailing list
[email protected]
http://m5sim.org/cgi-bin/mailman/listinfo/m5-users
