In message: <[EMAIL PROTECTED]> Bob Bishop <[EMAIL PROTECTED]> writes: : At 17:44 23/7/02, M. Warner Losh wrote: : >In message: <[EMAIL PROTECTED]> : > Bob Bishop <[EMAIL PROTECTED]> writes: : >: [...] : >: none2@pci0:11:0: class=0x028000 card=0x777010b7 chip=0x777010b7 : >: rev=0x02 hdr=0x00 : >: vendor = '3COM Corp, Networking Division' : >: class = network : > : >Maybe this is the droid that you are looking for :-) : : OK, OK. Here's a less nutty patch which gets the card probed every time. : : diff -u -r1.8.2.3 if_wi_pci.c : --- if_wi_pci.c 2002/06/16 18:07:18 1.8.2.3 : +++ if_wi_pci.c 2002/07/23 22:01:33 : @@ -113,6 +113,7 @@ : int i; : : sc = device_get_softc(dev); : + DELAY(100000); : for(i=0; pci_ids[i].vendor != 0; i++) { : if ((pci_get_vendor(dev) == pci_ids[i].vendor) && : (pci_get_device(dev) == pci_ids[i].device)) {
So if you have a saner delay value, does it probe? 100ms is a very long time to delay before getting the vendor info (which should already be present before the call to pci_get_vendor due to caching, iirc). Warner To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-hackers" in the body of the message
