Re: cdparanoia patch for ahci(4)/siis(4) (next version)

On Fri, Aug 07, 2009 at 10:19:26AM +0300, Alexander Motin wrote:
> Scott Long wrote:
> > Juergen Lock wrote:
> >> On Thu, Aug 06, 2009 at 10:07:15PM +0200, Juergen Lock wrote:
> >>> On Thu, Aug 06, 2009 at 10:46:48PM +0300, Alexander Motin wrote:
> >>>> Juergen Lock wrote:
> >>>>> 2. cdda/dae seems to be broken entirely with ahci(4) as well as
> >>>>> siis(4) (I remember a report about it being broken for usb optical
> >>>>> drives too so maybe this is related?) - I tested with the
> >>>>> audio/cdparanoia port as well as with
> >>>>>     mplayer -cdrom-device /dev/cd{0,1} cdda://...
> >>>>> (mplayer needs to be built with the libparanoia knob on for this) -
> >>>>> this
> >>>>> does work with atapicam(4) without ahci/siis so it can't be cd(4)'s
> >>>>> fault alone.  On siis(4) it seems to just fail while on ahci(4) (I
> >>>>> still
> >>>>> have another optical drive on there, it's on the board's amd sb700)
> >>>>> it causes the sata channel to be reset endlessly until I ^C mplayer:
> >>>>>
> >>>>>  Soo, anyone have ideas/patches/things they want me to check for this?
> >>>> But this appeared to to be really trivial. cdparanoia uses extremely
> >>>> simple method for detecting ATAPI devices - it checks that SIM is
> >>>> named "ata". Trivial single line hack made it successfully play some
> >>>> old AudioCD in SATA drive on SiI3132 controller for me, while I am
> >>>> typing this. Probably we should invent better way to do this.
> >>> Oooh! :)  I need to test this...
> >>
> >> Yup, works here too on siis and ahci with the following patch:
> >> (maintainer Cc'd)
> >>
> >> Index: interface/scsi_interface.c
> >> _at__at_ -1480,9 +1480,12 _at__at_
> >>      /*
> >>       * if the bus device name is `ata', we're (obviously)
> >>       * running ATAPICAM.
> >> +     * XXX same for the new ahci(4) and siis(4) drivers...
> >>       */
> >>  
> >> -    if (strncmp(d->ccb->cpi.dev_name, "ata", 3) == 0) {
> >> +    if (strncmp(d->ccb->cpi.dev_name, "ata", 3) == 0 ||
> >> +        strncmp(d->ccb->cpi.dev_name, "ahcich", 6) == 0 ||
> >> +        strncmp(d->ccb->cpi.dev_name, "siisch", 6) == 0) {
> >>          cdmessage(d, "\tDrive is ATAPI (using ATAPICAM)\n");
> >>          d->is_atapi = 1;
> >>      } else {
> >>
> >>  Thanx, :)
> >>     Juergen
> > 
> > This is fine for the moment, but unmaintainable in the long run as more
> > and more drives are written.  cdparanoia needs to look at protocol and
> > transport attributes, not device names.
> 
> CAM reports SCSI protocol for ATAPI devices at this moment. It is not
> good probably. but changing it now may be painful. Checks like
> 	d->ccb->cpi.transport == XPORT_ATA ||
> 	d->ccb->cpi.transport == XPORT_SATA
> should be for now. "ata" hack should also stay there for now, as
> ATAPICAM emulates SCSI transport now, but not a new ATA one.

Ok checking for XPORT_(S)ATA works for me as well:

Index: interface/scsi_interface.c
_at__at_ -1480,10 +1480,16 _at__at_
 	/*
 	 * if the bus device name is `ata', we're (obviously)
 	 * running ATAPICAM.
+	 * same for the new ahci(4) and siis(4) drivers and future others
+	 * which use SATA transport too...
 	 */
 
-	if (strncmp(d->ccb->cpi.dev_name, "ata", 3) == 0) {
-		cdmessage(d, "\tDrive is ATAPI (using ATAPICAM)\n");
+	if (strncmp(d->ccb->cpi.dev_name, "ata", 3) == 0 ||
+#if __FreeBSD_version >= 800102
+	    d->ccb->cpi.transport == XPORT_SATA ||
+#endif
+	    d->ccb->cpi.transport == XPORT_ATA) {
+		cdmessage(d, "\tDrive is ATAPI (using ATAPICAM or direct CAM (S)ATA transport)\n");
 		d->is_atapi = 1;
 	} else {
 		cdmessage(d, "\tDrive is SCSI\n");

 One question remains tho:  What if someone connects a sata drive
to a sas controller, will that still be XPORT_SAS then?  In that case
I'd say we'd need to check for that as well, or maybe just assume there
are no `real' sas optical drives and treat everything sas as sata here...

 Thanx,
	Juergen