[m-dev.] For review: calling MR_collect_filter() in MR_trace_real()

[Fergus Henderson]

On 06-Mar-2001, Erwan Jahier <Erwan.Jahier at irisa.fr> wrote:
> Fergus wrote :
> > [Erwan wrote:]
> | > @@ -1504,7 +1489,7 @@
> | >  			arguments = MR_list_empty()
> | >  		);
> | >  	}
> | > -	MR_TRACE_CALL_MERCURY((*filter_ptr)(
> | > +	(*filter_ptr)(
> | >  		MR_trace_event_number,
> | >  		seqno,
> | >  		depth,
> | 
> | What happened to the MR_TRACE_CALL_MERCURY() macro here?
> | Why did that go away?
> 
> Because I think they are unnecessary, indeed: 
> MR_copy_regs_to_saved_regs() calls 
> 	restore_transient_registers();
> 	save_registers();
> 
> MR_copy_saved_regs_to_regs() calls
> 	restore_registers();
> 	save_transient_registers();
> 
> Since MR_COLLECT_filter() is called just between MR_copy_regs_to_saved_regs()
> and MR_copy_saved_regs_to_regs(),

No, it's not; there are calls to MR_list_empty() and MR_trace_make_var_list(),
both of which can allocate heap.  They increment the real MR_hp,
whereas the Mercury code that you're calling will expect the heap
pointer to be in the saved copy of MR_hp, so it will overwrite
the heap space that they allocate.

(This is assuming --gc none and non-SPARC architecture,
i.e. no transient registers.)

> I thought the macro MR_TRACE_CALL_MERCURY()
> was not necessary in that case. 

I think it is, for the reason explained above.

Please send another diff when you've fixed that one
and the stuff mentioned in my other mail.

-- 
Fergus Henderson <fjh at cs.mu.oz.au>  |  "I have always known that the pursuit
                                    |  of excellence is a lethal habit"
WWW: <http://www.cs.mu.oz.au/~fjh>  |     -- the last words of T. S. Garp.
--------------------------------------------------------------------------
mercury-developers mailing list
Post messages to:       mercury-developers at cs.mu.oz.au
Administrative Queries: owner-mercury-developers at cs.mu.oz.au
Subscriptions:          mercury-developers-request at cs.mu.oz.au
--------------------------------------------------------------------------