From Bugzilla Helper: User-Agent: Mozilla/4.78 [en] (X11; U; Linux 2.4.3-12smp i686) Description of problem: This /etc/modules.conf entry used to work to put my dell 2550 eth2 broadcom gigabit interface into 100 Mbit/s, full duplex mode in rh7.1 kernels 2.4.2-* and 2.4.3-* : options bcm5700 line_speed=100 full_duplex=1 however under kernel 2.4.9-6 and -12 it now forces the interface down: Nov 6 09:58:30 meggan network: Bringing up interface eth2: succeeded Nov 6 09:58:43 meggan kernel: eth2: Broadcom BCM5700 found at mem feb00000, IRQ 17, node addr 00b0d0df0f71 Nov 6 09:58:43 meggan kernel: eth2: Broadcom BCM5401 Copper transceiver found Nov 6 09:58:43 meggan kernel: eth2: Scatter-gather ON, 64-bit DMA ON, Tx Checksum OFF, Rx Checksum OFF Nov 6 09:58:43 meggan kernel: bcm5700: eth2 NIC Link is DOWN I have to comment out that /etc/modules.conf entry and reboot for the interface to start working again. However it starts up as 100Mbit/s and HALF-duplex. Then I have to use mii-tool to force to FULL-duplex and that works. Version-Release number of selected component (if applicable): How reproducible: Always Steps to Reproduce: 1. have dell 2550 with on-board broadcom gigabit rh45 recepticable 2. connect to a cisco switch forcing cisco port to full/100 3. install rh7.1 kernels 2.4.9-6 or -12 with all latest rpm patches 4. add this /etc/modules.conf entry: alias eth2 bcm5700 5. add this too: options bcm5700 line_speed=100 full_duplex=1 6. reboot and you'll see link light on gigabit port go out mid-boot Actual Results: [root@meggan /root]# cat /etc/modules.conf alias eth0 eepro100 alias eth1 e1000 alias eth2 bcm5700 alias scsi_hostadapter aic7xxx alias parport_lowlevel parport_pc alias usb-controller usb-ohci options eepro100 options=48 options e100 e100_speed_duplex=4 # options bcm5700 line_speed=100 full_duplex=1 [root@meggan /root]# uname -a Linux meggan.989studios.com 2.4.9-12smp #1 SMP Tue Oct 30 18:16:48 EST 2001 i686 unknown with bcm5700 options reinstated: Nov 6 09:58:30 meggan network: Bringing up interface eth2: succeeded Nov 6 09:58:43 meggan kernel: eth2: Broadcom BCM5700 found at mem feb00000, IRQ 17, node addr 00b0d0df0f71 Nov 6 09:58:43 meggan kernel: eth2: Broadcom BCM5401 Copper transceiver found Nov 6 09:58:43 meggan kernel: eth2: Scatter-gather ON, 64-bit DMA ON, Tx Checksum OFF, Rx Checksum OFF Nov 6 09:58:43 meggan kernel: bcm5700: eth2 NIC Link is DOWN althought ifconfig eth2 shows the interface is 'up': eth2 Link encap:Ethernet HWaddr 00:B0:D0:DF:0F:71 ^M inet addr:208.236.12.134 Bcast:208.236.12.255 Mask:255.255.255.0^M UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1^M RX packets:0 errors:0 dropped:0 overruns:0 frame:0^M TX packets:0 errors:0 dropped:0 overruns:0 carrier:0^M collisions:0 txqueuelen:100 ^M Interrupt:17 ^M ^M comment out the bcm options and reboot: Nov 6 10:02:12 meggan network: Bringing up interface eth2: succeeded Nov 6 10:02:24 meggan kernel: eth2: Broadcom BCM5700 found at mem feb00000, IRQ 17, node addr 00b0d0df0f71 Nov 6 10:02:24 meggan kernel: eth2: Broadcom BCM5401 Copper transceiver found Nov 6 10:02:24 meggan kernel: eth2: Scatter-gather ON, 64-bit DMA ON, Tx Checksum OFF, Rx Checksum OFF Nov 6 10:02:24 meggan kernel: bcm5700: eth2 NIC Link is UP, 100 Mbps half duplex Expected Results: expect the gigabit interface to come up as 100 Mbit/s full-duplex and stay up. Additional info:
Working with Chip, and doing some testing in the lab here. This is related to the switch settings. When the switch is set to: set port duplex <portnum> full set port speed <portnum> 100 The bcm5700 module will only work without the option line and will only come up half-duplex. With the port set to auto negotiate, the bcm5700 module will honor the option line, and will also auto negotiate to full duplex. Not having admin access to the switch, I can't test this last bit in house here. Everything else is currently set up on perf54 internally.
Thanks for the bug report. However, Red Hat no longer maintains this version of the product. Please upgrade to the latest version and open a new bug if the problem persists. The Fedora Legacy project (http://fedoralegacy.org/) maintains some older releases, and if you believe this bug is interesting to them, please report the problem in the bug tracker at: http://bugzilla.fedora.us/