On 30 Jan 2001 eeh@netbsd.org wrote:

> 
> 	> In any case, this should not be a fatal condition,
> 	> and you can probably work around it by disablinng
> 	> the streaming cache entirely or using only COHERENT
> 	> mappings.
> 	> 
> 	> Altought, looking at the code:
> 	> 
> 
> 	There are also:
> 
> 		if (!is->is_sb)
> 			return (0);
> 
> Obviously if it prints out that message then the streaming
> cache is enabled.
> 

What seems strange to me that this code should called all the time from
the beginning because of call chain

lsi64854_setup->sbus_dmamap_load->iommu_dvmamap_load->iommu_enter->iommu_strbuf_flush

and I don't see how it would be possible to avoid it except that
is_sb is not set, which also can't be true because of sbus_attach sets
it unconditionally. I must miss something...
I'll add some printouts to verify. Not sure about today though. 


> 		is->is_flush = 0;
> 
> This is where it resets the bit that the IOMMU sets when
> the flush completes.
> 
> 	> #ifdef DIAGNOSTIC
> 	> 	if (!is->is_flush) {
> 	> 		printf("iommu_strbuf_flush: flush timeout %p at %p\n",
> 	> 		    (void *)(u_long)is->is_flush, 
> 	> 		    (void *)(u_long)is->is_flushpa); /* panic? */
> 	> 
> 
> 	and another one:
> 			Debugger();
> 
> 	So I'd expect this code to fire up debugger right after it gets there
> 	after streaming buffer gets enabled. And this happen after a while.
> 	I havn't find yet a place where that buffer is enabled.
> 
> This happens in the bus controller attach routine, specifically 
> `sbus_attach()' in your case.
> 
> 	I'll try to continue from ddb and see what happens.
> 
> 	> if is->is_flush really is 0x100000000, either the code generation
> 	> is broken, or the value was set between the test and the printf ()
> 	> call.  Maybe you should just increase the timeout a bit.
> 
> 	I can compile without optimization, it helped me before with 
> 	code generation. Don't understand yet that 'timeout' magic. 
> 
> Hm.  It should work.  is_flush is volatile.  Could be D$
> coherency issues.  Try adding a call to `blast_vcache()' just
> before the test of `is->is_flush' to see if it fixes the problem
> (this is DIAGNOSTIC code, after all.)

I checked bus_dmamap_load and it uses COHERENT. I'll try this as well.
uh, will need a list.

	Andrey