d/freebsd-nq
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Add preliminary support for the OLICOM OC2231 and OC2232. This is the

Browse Source 
first driver that does the configuration dance with CFE's.  There's
likely some additional configuration that's needed to get things
working completely...
This commit is contained in:
Warner Losh 2008-08-27 04:11:03 +00:00
parent d6742bfbd3
commit 505c6d6ae9
1 changed files with 55 additions and 12 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

67
sys/dev/ex/if_ex_pccard.c
View File

@ -49,7 +49,6 @@ __FBSDID("$FreeBSD$");
#include <dev/pccard/pccardvar.h>
#include <dev/pccard/pccard_cis.h>
#include "pccarddevs.h"
static const struct pccard_product ex_pccard_products[] = {
@ -104,23 +103,66 @@ static int
ex_pccard_probe(device_t dev)
{
const struct pccard_product *pp;
int error;
int error, i, j;
uint32_t fcn = PCCARD_FUNCTION_UNSPEC;
/* Make sure we're a network function */
if ((pp = pccard_product_lookup(dev, ex_pccard_products,
sizeof(ex_pccard_products[0]), NULL)) == NULL)
return (EIO);
if (pp->pp_name != NULL)
device_set_desc(dev, pp->pp_name);
/*
* Olicom 22.8k and 33.6k modems need to activate the right
* CFE. The odd formula below replicates the sequence of cfes
* that have multiple resources:
* 9, 11, 13, 15, 0 + 9
* 25, 27, 29, 31, 16 + 9
* 41, 43, 45, 47, 32 + 9
* 57, 59, 61, 63 48 + 9
* (entries 8, 24, 40 and 56 are single resoruce cfes)
* Fortunately the code that enables and disables the multiple
* fuctions of the card won't mess with the lower bit for cards
* that aren't stanards conforming MFC cards (which these olicom
* cards aren't).
*
* Note: These cards still don't get interrupts for reasons
* unknown, even when the right cfe is selected. There's likely
* something in the CCR that needs to be manually tweaked, but
* the COR bits seem to all be used. Bit 0 and 3 are always set
* and the other bits select the config to use. Maybe one of those
* two bits needs to be cleared, or there's something else in the
* CCR that needs tweaking. The pattern of resources suggests
* bit 0 turns on the ethernet, however...
*/
if (pp->pp_vendor == PCMCIA_VENDOR_OLICOM &&
(pp->pp_product == PCMCIA_PRODUCT_OLICOM_OC2231 ||
pp->pp_product == PCMCIA_PRODUCT_OLICOM_OC2232)) {
if (pccard_select_cfe(dev, 1) == 0)
goto good;
for (i = 0; i < 4; i++) {
for (j = 0; j < 4; j++) {
printf("Trying %d %d\n", i, j);
if (pccard_select_cfe(dev,
(i << 4) + (j << 1) + 9) == 0)
goto good;
}
}
/* Can't activate the net entries, punt */
return (EIO);
}
/*
* All other cards supported by this driver don't need specail
* treatment, so just filter based on the type of card. The
* special treatment ones are setup to 'fail safe' to a modem so
* this check would effectively filter them out as well.
*/
error = pccard_get_function(dev, &fcn);
if (error != 0)
return (error);
if (fcn != PCCARD_FUNCTION_NETWORK)
return (ENXIO);
if ((pp = pccard_product_lookup(dev, ex_pccard_products,
sizeof(ex_pccard_products[0]), NULL)) != NULL) {
if (pp->pp_name != NULL)
device_set_desc(dev, pp->pp_name);
return 0;
}
return EIO;
return (EIO);
good:;
return (0);
}
static int
@ -185,3 +227,4 @@ static driver_t ex_pccard_driver = {
};
DRIVER_MODULE(ex, pccard, ex_pccard_driver, ex_devclass, 0, 0);
MODULE_DEPEND(ex, pccard, 1, 1, 1);