Svadev mailing list

[John Criswell <criswell AT illinois.edu>]

On 8/29/11 10:52 AM, Will Dietz wrote:
> It's my understanding that SAFECode makes no guarantees about the
> state of the program after an error.  After all, how do you continue
> execution of a program that contains illegal operations?  Not just
> aborting the program is something we do to be nice :).
>
> That said in your particular use-case here's what's happening:
>
> SAFECode detects the load/store error before it happens, this is the
> first error.  Then SAFECode returns to the code to execute the next
> instruction...which is the load/store that we already know is invalid.
>   From there we just spin: segfault fires the signal handler which
> reports the error (not sure why it's categorized the way it is,
> perhaps John can comment), and in good signal handler fashion returns
> to the instruction we were executing when the signal was fired...
> which is the same instruction that segfaulted... and the signal
> handler is called.... and so on.

It's also because I disabled the cut-off of 20 error reports in 
runtime/DebugRuntime/Report.cpp.  You can modify that to re-enable the 
limit.

I disabled it earlier this month because I thought it would make sense 
for the debugging tool to keep reporting errors.  However, perhaps we 
should just make the limit higher (maybe 50-100 error reports) or make 
the signal handler terminate the program since it knows that recovery is 
not possible.

-- John T.

> Hope this helps!
>
> ~Will
>
> On Sun, Aug 28, 2011 at 10:16 PM, Matthew 
> Wala<wala1 AT illinois.edu>
>   wrote:
> > Hi everyone:
> > The most recent version of SAFECode is giving me trouble with very
> > simple test cases.
> >
> > For instance, the program
> >
> > int main()
> > {
> >   int i[100];
> >   i[101] = 10;
> >   return 0;
> > }
> >
> >
> > when instrumented and run results in the following output:
> >
> > SAFECode:Violation Type 0x6 when accessing  0x617934 at IP=0x401fc0
> >
> > =======+++++++    SAFECODE RUNTIME ALERT +++++++=======
> > = Error type                            :       Load/Store Error
> > = Faulting pointer                      :       0x617934
> > = Program counter                       :       0x401fc0
> > = Fault PC Source                       :       main:6
> > =
> > = Object allocated at PC                :       0x401ef4
> > = Allocated in Source File              :       main:1
> > = Object allocation sequence number     :       1
> > = Object start                          :       0x6177a0
> > = Object length                         :       0x64
> > SAFECode: Fault!
> > SAFECode:Violation Type 0x9 when accessing  0xc0000001 at IP=0x411b64
> >
> > =======+++++++    SAFECODE RUNTIME ALERT +++++++=======
> > = Error type                            :       Uninitialized/NULL Pointer 
> > Error
> > = Faulting pointer                      :       0xc0000001
> > = Program counter                       :       0x411b64
> > = Fault PC Source                       :<unknown>:0
> > SAFECode: Fault!
> > SAFECode:Violation Type 0x9 when accessing  0xc0000001 at IP=0x411b64
> >
> > =======+++++++    SAFECODE RUNTIME ALERT +++++++=======
> > = Error type                            :       Uninitialized/NULL Pointer 
> > Error
> > = Faulting pointer                      :       0xc0000001
> > = Program counter                       :       0x411b64
> > = Fault PC Source                       :<unknown>:0
> > SAFECode: Fault!
> > SAFECode:Violation Type 0x9 when accessing  0xc0000001 at IP=0x411b64
> >
> > The last Unitialized / NULL Pointer Error statement is repeated without end.
> > Is there an explanation for this behavior and how can I get it to work 
> > properly?
> >
> > Thanks,
> > Matt
> > _______________________________________________
> > svadev mailing list
> > svadev AT cs.uiuc.edu
> > http://lists.cs.uiuc.edu/mailman/listinfo/svadev
> _______________________________________________
> svadev mailing list
> svadev AT cs.uiuc.edu
> http://lists.cs.uiuc.edu/mailman/listinfo/svadev