Re: [ppc-linux-nat]: set access flag for h/w watchpoint even if it is only read or write

[Daniel Jacobowitz <drow at false dot org>]

On Fri, Jul 07, 2006 at 01:04:36PM +0300, Eli Zaretskii wrote:
> I think the reason it works for me is that on x86, there's no real
> support for read watchpoints, so we actually set a read-write
> watchpoint, and then the logic of watchpoint_check does TRT.

Precisely.

> The question is, how to do that without breaking other platforms, like
> x86, which we cannot trust.  If you can come up with a design that
> accommodates both types of situations, I will be happy to review it.
> 
> Daniel, could you please point me to Ulrich's change, either in
> ChangeLogs or in the sources?  I cannot find it forf some reason.

I must be mistaken; our S/390 support doesn't have any read watchpoints
(I don't know if the architecture does or not).  In fact I can't find
any architecture that does this.  But I immediately recognized
the description of the problem... so it must have happened somewhere.

I can't find the discussion of it, but the gdbserver crisv32 port does
the same thing:

  /* Read watchpoints are set as access watchpoints, because of GDB's
     inability to deal with pure read watchpoints.  */
  if (type == '3')
    type = '4';

Here's some more about it:
  http://sources.redhat.com/ml/gdb/2005-11/msg00231.html

In which I also claimed S/390 did it, which doesn't appear to be true,
but at least I've had the same misconception for a while now.


-- 
Daniel Jacobowitz
CodeSourcery