kern/79324: Broadcom bge chip initialization failure
sdebnath at mac.com
Mon Mar 28 16:40:03 PST 2005
>Synopsis: Broadcom bge chip initialization failure
>Arrival-Date: Tue Mar 29 00:40:02 GMT 2005
>Originator: Shawn Debnath
Sun Microsystems, Inc
Don't have output but running on dual opterons freebsd 5.3 release (amd64) 1gb memory
This is similar to what I am seeing during bootup of FreeBSD:
bge0: <Broadcom BCM5705M Dual Gigabit Ethernet, ASIC rev. 0x3001> mem
0x90000000-0x9000ffff irq 11 at device 14.0 on pci2
bge0: firmware handshake timed out
bge0: RX CPU self-diagnostics failed!
bge0: chip initialization failed
device_probe_and_attach: bge0 attach returned 6
However, I do have Solaris successfully running on the machine, and it thinks its not a BCM750x series card:
Mar 28 14:04:36 tornado bge: [ID 801725 kern.warning] WARNING: bge0: 5705-based
subsystem 'pci14e4,1653' not supported
Mar 28 14:04:36 tornado unix: [ID 954099 kern.info] NOTICE: IRQ16 is being share
d by drivers with different interrupt levels.
Mar 28 14:04:36 tornado This may result in reduced system performance.
Mar 28 14:04:37 tornado genunix: [ID 454863 kern.info] dump on /dev/dsk/c0d0s1 s
ize 2047 MB
Mar 28 14:04:37 tornado genunix: [ID 678236 kern.info] Ethernet address = 0:11:9
Mar 28 14:04:37 tornado gld: [ID 944156 kern.info] bge0: BCM579x driver v0.44: t
ype "ether" mac address 00:11:09:a0:03:eb
Mar 28 14:04:37 tornado pci: [ID 370704 kern.info] PCI-device: pci14e4,1653 at b, b
Mar 28 14:04:37 tornado genunix: [ID 936769 kern.info] bge0 is /pci at 0,0/pci14e4,
1653 at b
Just for curiosity sake, I tried increasing the TIMEOUT (src/sys/dev/bge/if_bgereg.h) from:
#define BGE_TIMEOUT 100000 /* 5 zeros */
#define BGE_TIMEOUT 1000000 /* 6 zeros */
and still no luck. Any help will be appreciated. If you need me to test code, send the src files to dump in src/dev/bge and I can test it out.
I am experiencing this on the MSI MASTER2-FAR motherboard (on board broadcom chipset). The MSI website says that its Broadcom BCM5705 Gigabit LAN controllers but they often make mistakes. I can't decipher the model # on the chipset itself, but if I do I will add to this report.
More information about the freebsd-bugs