The threads saga: watchpoints

The threads saga: watchpoints

[Daniel Jacobowitz]

It was pointed out to me today that watchpoints and threads really don't get
along.  It's even in the manual:

     _Warning:_ In multi-thread programs, watchpoints have only limited
     usefulness.  With the current watchpoint implementation, GDB can
     only watch the value of an expression _in a single thread_.  If
     you are confident that the expression can only change due to the
     current thread's activity (and if you are also confident that no
     other thread can become current), then you can use watchpoints as
     usual.  However, GDB may not notice when a non-current thread's
     activity changes the expression.


I think some of our hardware breakpoint implementations have the same issue.

It seems to me that, in general, this should be pretty easy to fix.  But it
requires some definite changes in the current infrastructure.  Shouldn't we
be able to insert the watchpoint in all threads?

-- 
Daniel Jacobowitz
MontaVista Software                         Debian GNU/Linux Developer

Re: The threads saga: watchpoints

[Elena Zannoni]

Daniel Jacobowitz writes:
 > It was pointed out to me today that watchpoints and threads really don't get
 > along.  It's even in the manual:
 > 
 >      _Warning:_ In multi-thread programs, watchpoints have only limited
 >      usefulness.  With the current watchpoint implementation, GDB can
 >      only watch the value of an expression _in a single thread_.  If
 >      you are confident that the expression can only change due to the
 >      current thread's activity (and if you are also confident that no
 >      other thread can become current), then you can use watchpoints as
 >      usual.  However, GDB may not notice when a non-current thread's
 >      activity changes the expression.
 > 
 > 
 > I think some of our hardware breakpoint implementations have the same issue.
 > 
 > It seems to me that, in general, this should be pretty easy to fix.  But it
 > requires some definite changes in the current infrastructure.  Shouldn't we
 > be able to insert the watchpoint in all threads?
 > 
 > -- 
 > Daniel Jacobowitz
 > MontaVista Software                         Debian GNU/Linux Developer


There was a discussion on this some time ago:

http://sources.redhat.com/ml/gdb-patches/2002-01/msg00383.html
and follow-ups.


elena

Re: The threads saga: watchpoints

[Daniel Jacobowitz]

On Fri, Mar 07, 2003 at 09:41:33AM -0500, Elena Zannoni wrote:
> Daniel Jacobowitz writes:
>  > It was pointed out to me today that watchpoints and threads really don't get
>  > along.  It's even in the manual:
>  > 
>  >      _Warning:_ In multi-thread programs, watchpoints have only limited
>  >      usefulness.  With the current watchpoint implementation, GDB can
>  >      only watch the value of an expression _in a single thread_.  If
>  >      you are confident that the expression can only change due to the
>  >      current thread's activity (and if you are also confident that no
>  >      other thread can become current), then you can use watchpoints as
>  >      usual.  However, GDB may not notice when a non-current thread's
>  >      activity changes the expression.
>  > 
>  > 
>  > I think some of our hardware breakpoint implementations have the same issue.
>  > 
>  > It seems to me that, in general, this should be pretty easy to fix.  But it
>  > requires some definite changes in the current infrastructure.  Shouldn't we
>  > be able to insert the watchpoint in all threads?
>  > 
>  > -- 
>  > Daniel Jacobowitz
>  > MontaVista Software                         Debian GNU/Linux Developer
> 
> 
> There was a discussion on this some time ago:
> 
> http://sources.redhat.com/ml/gdb-patches/2002-01/msg00383.html
> and follow-ups.

Thanks!  I missed that one.  I think that perhaps I should audit the
available hardware watchpoint mechanisms for similar problems...

-- 
Daniel Jacobowitz
MontaVista Software                         Debian GNU/Linux Developer