|
Insmod "unresolved symbol" errors
Okay, I realize that this is a problem with a driver that probably no one has ever heard of, but I'm not getting much help on their forum. I'm trying to install the open-source driver for the ACX100-chipset network card (my D-Link DWL-520+). It is one of those "$ make" supposedly simple installations that does access kernel includes. Anyway, I do that and then run the install script, which is mainly just an '$ insmod acx100_pci.o firmware_dir=firmware debug=0x9b' command. The problem is that I get the following load of unresolved symbols:
+ insmod ../src/acx100_pci.o debug=0x9b firmware_dir=../firmware
../src/acx100_pci.o: unresolved symbol pci_set_power_state
../src/acx100_pci.o: unresolved symbol eth_type_trans
../src/acx100_pci.o: unresolved symbol schedule_timeout
../src/acx100_pci.o: unresolved symbol filp_open
../src/acx100_pci.o: unresolved symbol __kfree_skb
../src/acx100_pci.o: unresolved symbol alloc_skb
../src/acx100_pci.o: unresolved symbol pci_register_driver
../src/acx100_pci.o: unresolved symbol __generic_copy_from_user
../src/acx100_pci.o: unresolved symbol __release_region
../src/acx100_pci.o: unresolved symbol ether_setup
../src/acx100_pci.o: unresolved symbol kmalloc
../src/acx100_pci.o: unresolved symbol pci_free_consistent
../src/acx100_pci.o: unresolved symbol pci_enable_device
../src/acx100_pci.o: unresolved symbol alloc_etherdev
../src/acx100_pci.o: unresolved symbol __get_free_pages
../src/acx100_pci.o: unresolved symbol vfree
../src/acx100_pci.o: unresolved symbol cpu_raise_softirq
../src/acx100_pci.o: unresolved symbol free_irq
../src/acx100_pci.o: unresolved symbol unregister_netdev
../src/acx100_pci.o: unresolved symbol __out_of_line_bug
../src/acx100_pci.o: unresolved symbol iounmap
../src/acx100_pci.o: unresolved symbol free_pages
../src/acx100_pci.o: unresolved symbol pci_alloc_consistent
../src/acx100_pci.o: unresolved symbol filp_close
../src/acx100_pci.o: unresolved symbol __ioremap
../src/acx100_pci.o: unresolved symbol register_netdev
../src/acx100_pci.o: unresolved symbol dev_alloc_name
../src/acx100_pci.o: unresolved symbol iomem_resource
../src/acx100_pci.o: unresolved symbol kfree
../src/acx100_pci.o: unresolved symbol request_irq
../src/acx100_pci.o: unresolved symbol netif_rx
../src/acx100_pci.o: unresolved symbol pci_unregister_driver
../src/acx100_pci.o: unresolved symbol skb_over_panic
../src/acx100_pci.o: unresolved symbol pci_set_master
../src/acx100_pci.o: unresolved symbol sprintf
../src/acx100_pci.o: unresolved symbol jiffies
../src/acx100_pci.o: unresolved symbol __vmalloc
../src/acx100_pci.o: unresolved symbol softnet_data
../src/acx100_pci.o: unresolved symbol __request_region
../src/acx100_pci.o: unresolved symbol printk
../src/acx100_pci.o: unresolved symbol __generic_copy_to_user
Can anybody recognize these symbols as being part of a certain kernel option, or software package? Any guesses or recommendations of another possible problem are very welcome.
As far as I know, I have all the necessary kernel options (CONFIG_NET_RADIO, CONFIG_NET_WIRELESS) installed, as well as the required Wireless Tools package installed. I cleaned up all the leftovers from other kernel installations and recompiled the kernel that I did have. /lib/modules/`uname -r`/build points to the right spot (that being /usr/src/linux). I've tried installing with gcc 2.9 as well as gcc 3.3 installed and with four different versions of the open-source driver. The driver is technically in alpha stages, but there have been so many success stories that I'm convinced it can work, and no one else has posted any similar problem to this on their forum (acx100.sf.net).
Finally, those of you who know about the TI ACX100 chipset, I was able to install and insmod Ivor's acx100sta module just fine, and get associated and connected to the internet. However, it consistently locked up the card after transmitting a few kilobytes (15KB or so) and the only way to fix it was to reinstall the driver ... again and again. I've checked other forum threads and the people that have had this problem haven't seemed to have found a way around it. However, if anyone reading this does know a way to fix that, I'll be happy. I would prefer to get the open-source driver going, since it's got a lot more options, etc., but any internet connection will do.
|
