Index | Thread | Search

From:
hshoexer <hshoexer@yerbouti.franken.de>
Subject:
Re: acpidmar(4): Move debug printf to show the actual mapping
To:
tech@openbsd.org
Date:
Mon, 10 Aug 2026 16:09:58 +0200

Download raw body.

Thread
Hi,

On Wed, Jul 29, 2026 at 11:15:14PM +0200, Mark Kettenis wrote:
> > Date: Tue, 28 Jul 2026 16:04:39 +0200
> > From: hshoexer <hshoexer@yerbouti.franken.de>
> > 
> > Hi,
> > 
> > trivial tweak to debug print the single mappings.  Note, this
> > requires to adjust debugme() to actually print something.
> > 
> > ok?
> 
> I think I had something similar in my tree at some point.  So ok
> kettenis@.  Although I question the usefulness of most of the debug
> code in acpidmar(4).  I'd like to see most of it removed eventually.

sure.  It's on my todo list.

> > 
> > ------------------------------------------------------------------------
> > diff --git a/sys/dev/acpi/acpidmar.c b/sys/dev/acpi/acpidmar.c
> > index cf18265fd09..b680b858411 100644
> > --- a/sys/dev/acpi/acpidmar.c
> > +++ b/sys/dev/acpi/acpidmar.c
> > @@ -857,16 +857,16 @@ domain_load_map(struct domain *dom, bus_dmamap_t map, int flags, int pteflag,
> >  			if (error)
> >  				goto fail;
> >  
> > +			if (debugme(dom)) {
> > +				printf("  LOADMAP: %.16llx %x => %.16llx\n",
> > +				    (uint64_t)seg->ds_addr,
> > +				    (uint32_t)seg->ds_len, (uint64_t)res);
> > +			}
> > +
> >  			/* Reassign DMA address */
> >  			seg->ds_addr = res | (seg->ds_addr & VTD_PAGE_MASK);
> >  		}
> >  
> > -		if (debugme(dom)) {
> > -			printf("  LOADMAP: %.16llx %x => %.16llx\n",
> > -			    (uint64_t)seg->ds_addr, (uint32_t)seg->ds_len,
> > -			    (uint64_t)res);
> > -		}
> > -
> >  		mapped_nsegs = i + 1;
> >  
> >  		for (idx = 0; idx < alen; idx += VTD_PAGE_SIZE) {
> > 
> >