Skip to contentSkip to conversation

various hotplugging problems

2.417rc1 Andreas Hartmann 1 participant 1 message Dec 16, 2001

lore
Share & tools
lore
#0 lore raw

Hello!

I'm running kernel 2.4.17rc1 and I have some strange problems with hotplugging and module xircom_tulip_cb (Xircom Cardbus ethernet 10/100+Modem 56 (CBEM56G-100) with Thinkpad T21 laptop).

If I put in my card, the networkdriver always starts with 100 MB/halfduplex (autonegotiation).

The linkpartner says:
Setting half-duplex based on auto-negotiated partner ability 0000.

Next, I forced both sides to do 100Mb full-duplex. As a result, I got on the laptop the following errormessage:
kernel: eth0: Link forced to 100Mbit full-duplex
kernel: xircom_tulip_cb: outl_CSR6 too many attempts,csr5=0xfffe0000 but the connect is working with a lot of errors on TX carrier side.

eth0 Link encap:Ethernet HWaddr 00:10:A4:A5:7B:F8

          inet addr:192.168.2.4  Bcast:192.168.2.255  Mask:255.255.255.0
          UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
          RX packets:71404 errors:0 dropped:0 overruns:0 frame:0
          TX packets:0 errors:36955 dropped:0 overruns:0 carrier:36955
          collisions:0 txqueuelen:100
          RX bytes:103113288 (98.3 Mb)  TX bytes:0 (0.0 b)
          Interrupt:11 Base address:0x4000

Otherwise, I didn't get any error on the device of the linkpartner.

If I'm running kernel 2.4.16 with pcmcia_cs package (no hotplugging), all is working fine - with using autonegotiation on both sides.

eth0 Link encap:Ethernet HWaddr 00:10:A4:A5:7B:F8

          inet addr:192.168.2.4  Bcast:192.168.2.255  Mask:255.255.255.0
          UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
          RX packets:83080 errors:1 dropped:0 overruns:0 frame:0
          TX packets:47330 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:100
          RX bytes:103882334 (99.0 Mb)  TX bytes:3124000 (2.9 Mb)
          Interrupt:11 Base address:0x200

(the one error on RX-side is from startup of the device).

There is one more problem with the hotplugging-function of kernel 2.4.17rc1:
in /lib/modules/2.4.17-rc1/modules.pcimap, the module serial_cs is missing - so the serial port of the Xircom-card can't be loaded (pcimodules can't show it).
You can insert a line, beginning with serial_cs and the following data are the same as in the xircom_tulip_cb - line. These datas are surely not correct - but the serial device of the Xircom-card is now configured automatically.

Unfortunately, this wasn't the last problem:
if you do a apm -s with kernel 2.4.17rc1, it is working fine unless the hotplugging hasn't been started.
If hotplugging has been started once, the laptop "wakes up" and hangs after apm -s (apm -s doesn't come back again). You can't even ping the laptop. I have to do a hw-reset. I tested those kernel options:
-> Allow interrupts during APM BIOS calls
-> Ignore USER SUSPEND

The result was always the same.

The HW-suspend-mode with Fn-F4 isn't working, too.

If I do the same with kernel 2.4.16 and pcmcia_cs-package, the hw-suspend is working fine as long as the cardmgr wasn't stopped or restarted. apm -s doesn't behave other as with kernel 2.4.17 and hotplugging.

Could you please fix these bugs or did I do some configuration errors? If you want, I can do some tests for you.

Thanks for any hint,
regards,
Andreas Hartmann