cvs commit: src/sys/dev/ep if_ep.c if_ep_isa.c if_ep_pccard.c
if_epreg.h if_epvar.h
Andrea Campi
andrea at webcom.it
Tue Oct 21 00:42:41 PDT 2003
On Sat, Oct 18, 2003 at 08:22:43AM -0700, Warner Losh wrote:
> imp 2003/10/18 08:22:43 PDT
>
> FreeBSD src repository
>
> Modified files:
> sys/dev/ep if_ep.c if_ep_isa.c if_ep_pccard.c
> if_epreg.h if_epvar.h
> Log:
> Transition to using bus_space macros rather than the inb/outb/etc.
[...]
>
> I've tested this with only PCMCIA cards: 3CCE589EC and 3CCSH572BT.
> I've not tried with my more extensive ISA, EISA and cbus collection.
This change (and subsequent fixes) breaks my 3C589D. As soon as dhclient
is started at boot (but the same happens if I ifconfig by hand), I get a
bunch of:
arp: unknown hardware address format (0xac40)
arp: unknown hardware address format (0xac40)
arp: unknown hardware address format (0xa840)
arp: unknown hardware address format (0xac40)
arp: unknown hardware address format (0xac40)
Actually, the bogus number vary a bit more:
brian# dmesg | grep arp | sort | uniq -c
37 arp: unknown hardware address format (0x0000)
1 arp: unknown hardware address format (0xa840)
149 arp: unknown hardware address format (0xac40)
27 arp: unknown hardware address format (0xb040)
20 arp: unknown hardware address format (0xb440)
10 arp: unknown hardware address format (0xb840)
3 arp: unknown hardware address format (0xbc40)
4 arp: unknown hardware address format (0xc040)
221 arp: unknown hardware address format (0xc440)
6 arp: unknown hardware address format (0xcc40)
I am quite sure it's this group of commits, as I cvsup'ed immediately
before and immediately after in order to verify this.
Incidentally, even when ep is sane, I still need to remove MPSAFE from
pccbb.c, or it stops after a few dozens of packets are sent/received.
Rerunning ifconfig on the interface brings it back for a few more packets.
Bye,
Andrea
--
Secret hacker rule #11: hackers read manuals.
More information about the cvs-src
mailing list