Hi,
We've got an onboard bge that has got this annoying printf that
also shows up on every up and down (comes from bge_reset):
bge0 at pci4 dev 0 function 0 "Broadcom BCM5718" rev 0x10, BCM5717 B0 \
(0x5717100), APE firmware NCSI 1.1.19.0: msi, address fc:aa:14:16:14:26
brgphy0 at bge0 phy 1: BCM5717C 10/100/1000baseT PHY, rev. 0
bge1 at pci4 dev 0 function 1 "Broadcom BCM5718" rev 0x10, BCM5717 B0 \
(0x5717100), APE firmware NCSI 1.1.19.0bge1: firmware handshake timed out
: msi, address fc:aa:14:16:14:27
brgphy1 at bge1 phy 2: BCM5717C 10/100/1000baseT PHY, rev. 0
At first I thought that we should just silence it, but after
double checking against the BCM5718 Programmer's Guide it became
apparent that we're not waiting long enough for chips with SEEPROM:
we're delay()ing for 1000ms which is good enough for "Flash" devices,
but need to delay() for 10000ms when used with SEEPROM ones. This
knowledge comes from Chapter 7 "Device Control", section "Device
Reset Procedure", point 13.
While the diff increases maximum timeout for "Flash" devices as
well as for SEEPROM ones, I don't think it would harm anything
as we're most likely to break out of the loop early. Needless
to say that it works fine here. OK?
diff --git sys/dev/pci/if_bge.c sys/dev/pci/if_bge.c
index b761ea9..7d8c970 100644
--- sys/dev/pci/if_bge.c
+++ sys/dev/pci/if_bge.c
@@ -3363,11 +3363,11 @@ bge_reset(struct bge_softc *sc)
*/
for (i = 0; i < BGE_TIMEOUT; i++) {
val = bge_readmem_ind(sc, BGE_SOFTWARE_GENCOMM);
if (val == ~BGE_MAGIC_NUMBER)
break;
- DELAY(10);
+ DELAY(100);
}
if (i >= BGE_TIMEOUT && (!(sc->bge_flags & BGE_NO_EEPROM)))
printf("%s: firmware handshake timed out\n",
sc->bge_dev.dv_xname);
