simulavr-devel
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[Simulavr-devel] Re: raising signals from gdb


From: Keith Gudger
Subject: [Simulavr-devel] Re: raising signals from gdb
Date: Thu, 9 Oct 2003 14:28:51 -0700 (PDT)

Ted:

Your latest patches work great.  I have some information for you on the
SIGXX values.  The '14' offset is, I think, a fixed value.  If you look at
the output from "info signal" you will find that SIG96 is entry #110.
That explains it!  Here's my suggestion.  Why don't you suggest that
interrupts start at "SIG100" and we'll use the offset of 86 in the code.
This way, it's easy to think "If I want interrupt 11, I type 'signal
SIG111'".  What do you think?

Keith

On Thu, 9 Oct 2003, Theodore A. Roth wrote:

> 
> 
> On Thu, 9 Oct 2003, Theodore A. Roth wrote:
> 
> >
> >
> > On Thu, 9 Oct 2003, Keith Gudger wrote:
> >
> > > Ted:
> > >
> > > I used your patch, and recompiled, so I'm trying out the signal command.
> > > When I type:
> > >
> > > "signal 96+12" I get:
> > > Only signals 1-15 are valid as numeric signals.
> > > Use "info signals" for a list of symbolic signals.
> > > (gdb)
> > >
> > > So I tried:
> > > (gdb) signal SIG108
> > > Continuing with signal SIG108.
> > >
> > > But nothing happens.  I tried SIG96, SIG97, SIG98, and set breakpoints at
> > > __vector_1 and __vector_2 to try and catch anything that happened.  No
> > > dice.  Any ideas are appreciated.  Thanks.
> >
> >
> > Do you have signal handlers for __vector_1 and __vector_2?
> >
> > I was able to get it to work, but discovered that 'signal SIG96'
> > doesn't send the actual value of 96. It sends 110 instead. So, there
> > seems to be an signal number offset of 14. This may be different on
> > different systems.
> >
> > To get gdb to send a 96 (interrupt vector 0), you need to use the
> > 'signal SIG82' command. Ick.
> >
> > Well, I said the patch was a hack...
> >
> > Attached is an updated patch which prints out some diagnostics to see
> > what vector has been raised and if it is been vectored to.
> 
> I forget to attach the test program I used. I was able to set break
> points at __vector_1 and __vector_default and then could trigger those
> by giving the appropriate signal command from gdb.
> 
> Ted Roth






reply via email to

[Prev in Thread] Current Thread [Next in Thread]