On Wed, 16 Jun 1999 19:38:07 -0700 
 Jonathan Stone <jonathan@DSG.Stanford.EDU> wrote:

 > Below are the acutal `patches from Hayakawa-san's CardBus patches.  I
 > dont know enough about COM IRQ usage to judge teh pros and cons of
 > avoiding the standard COM1 and COM2 values on PCMCIA, but the rest look OK.
 > 
 > Any objection to committing these, to make life simpler for the (many)
 > people using those patches?

Yes.  The cardslot stuff isn't appropriate.  We don't have "PC-Card" (like
the older patches from Japan for PCMCIA had).

I have a brand-spankin'-new CardBus laptop, and am willing to do a LITTLE
of the work to get this integrated.

Please be patient while I get some UVM work finished, tho :-)

 > 
 > 
 > 
 > 
 > Index: sys/conf/files
 > diff -c sys/conf/files:1.1.1.1 sys/conf/files:1.1.1.1.4.1
 > *** sys/conf/files:1.1.1.1	Wed Jan 20 22:27:45 1999
 > --- sys/conf/files	Wed Feb 17 01:32:49 1999
 > ***************
 > *** 348,353 ****
 > --- 348,354 ----
 >   define	usbus { }			# USB attachment
 >   define	vmebus { }			# VME bus attachment
 >   define	pcmciabus { [controller = -1], [socket = -1]}	# PCMCIA bus attachment
 > + define	cardslot {[slot = -1]}		# PC-Card attachment
 >   
 >   # We need the USB bus controllers here so different busses can
 >   # use them in an 'attach-with'.
 > Index: sys/dev/pci/files.pci
 > diff -c sys/dev/pci/files.pci:1.1.1.1.2.1 sys/dev/pci/files.pci:1.1.1.1.2.1.2.1
 > *** sys/dev/pci/files.pci:1.1.1.1.2.1	Sun Jan 24 20:01:38 1999
 > --- sys/dev/pci/files.pci	Wed Feb 17 01:41:08 1999
 > ***************
 > *** 180,182 ****
 > --- 180,187 ----
 >   # OHCI USB controller
 >   attach	ohci at pci with ohci_pci
 >   file	dev/pci/ohci_pci.c		ohci
 > + 
 > + # YENTA PCI-Cardbus bridge 
 > + device  cbb: cardslot, pcmciabus
 > + attach	cbb at pci with cbb_pci
 > + file	dev/pci/pccbb.c			cbb
 > Index: sys/dev/pcmcia/com_pcmcia.c
 > diff -c sys/dev/pcmcia/com_pcmcia.c:1.1.1.1 sys/dev/pcmcia/com_pcmcia.c:1.1.1.1.4.1
 > *** sys/dev/pcmcia/com_pcmcia.c:1.1.1.1	Wed Jan 20 22:28:24 1999
 > --- sys/dev/pcmcia/com_pcmcia.c	Mon Feb 22 03:08:41 1999
 > ***************
 > *** 234,239 ****
 > --- 234,247 ----
 >   			continue;
 >   
 >   		if (autoalloc == 0) {
 > + 		/* XXX: I dont want to use com0 and com1.  Usually,
 > + 		   The addresses of com0 and com1 have been already
 > + 		   occupied by ISA com devices. */
 > + 		
 > + 		  if (cfe->iospace[0].start == IO_COM1 ||
 > + 		      cfe->iospace[0].start == IO_COM2) {
 > + 		    continue;
 > + 		  }
 >   			/* 
 >   			 * cfe->iomask == 3 is our test for the "generic" config table 
 >   			 * entry, which we want to avoid on the first pass and use
 > Index: sys/dev/pcmcia/pcmcia.c
 > diff -c sys/dev/pcmcia/pcmcia.c:1.1.1.1 sys/dev/pcmcia/pcmcia.c:1.1.1.1.4.1
 > *** sys/dev/pcmcia/pcmcia.c:1.1.1.1	Wed Jan 20 22:28:25 1999
 > --- sys/dev/pcmcia/pcmcia.c	Mon Mar 22 23:15:54 1999
 > ***************
 > *** 111,116 ****
 > --- 111,119 ----
 >   	struct cfdata *match;
 >   	void *aux;
 >   {
 > +   if (strcmp(match->cf_driver->cd_name, "pcmcia")) {
 > +     return 0;
 > +   }
 >   	/* if the autoconfiguration got this far, there's a socket here */
 >   	return (1);
 >   }

        -- Jason R. Thorpe <thorpej@nas.nasa.gov>