Re: [PATCH 2/2] [PowerPC Book3E] Introduce new ptrace debug feature flag

[Thiago Jung Bauermann <bauerman@br.ibm.com>]

On Wed, 2011-08-24 at 14:00 +1000, David Gibson wrote:
> On Tue, Aug 23, 2011 at 02:57:56PM +0530, K.Prasad wrote:
> > On Tue, Aug 23, 2011 at 03:09:31PM +1000, David Gibson wrote:
> > > On Fri, Aug 19, 2011 at 01:23:38PM +0530, K.Prasad wrote:
> > > > 
> > > > While PPC_PTRACE_SETHWDEBUG ptrace flag in PowerPC accepts
> > > > PPC_BREAKPOINT_MODE_EXACT mode of breakpoint, the same is not intimated to the
> > > > user-space debuggers (like GDB) who may want to use it. Hence we introduce a
> > > > new PPC_DEBUG_FEATURE_DATA_BP_EXACT flag which will be populated on the
> > > > "features" member of "struct ppc_debug_info" to advertise support for the
> > > > same on Book3E PowerPC processors.
> > > 
> > > I thought the idea was that the BP_EXACT mode was the default - if the
> > > new interface was supported at all, then BP_EXACT was always
> > > supported.  So, why do you need a new flag?
> > > 
> > 
> > Yes, BP_EXACT was always supported but not advertised through
> > PPC_PTRACE_GETHWDBGINFO. We're now doing that.
> 
> I can see that.  But you haven't answered why.

BookS doesn't support BP_EXACT, that's why I suggested this flag.

A BP_EXACT watchpoint triggers only when there's a memory access exactly
at the given address. It doesn't trigger when there's (for example) a
4-byte write at an address immediately before which also changes the
memory contents of the byte watched by the BP_EXACT watchpoint. a ranged
watchpoint would trigger, so the semantics are different.

As a general rule, GDB only sets ranged watchpoints and only uses
BP_EXACT ones when the user sets a flag. I want GDB to fail when the
user sets the flag on BookS since it can't provide the feature.
-- 
[]'s
Thiago Jung Bauermann
IBM Linux Technology Center