Re: [Madwifi-users] Problem on Latitude D830
Status: Beta
Brought to you by:
otaku
From: Pavel R. <pr...@gn...> - 2008-03-22 05:52:34
|
On Fri, 2008-03-21 at 15:02 +0100, Diego Dujovne wrote: > > I have installed fc9_x64 on the D830, and I receive the following > error when I connect the Proxim card after booting: What is fc9_x64? Google finds no hits for that. I assume you are booting a 64-bit kernel from some prerelease of Fedora 9. > pccard: CardBus card inserted into slot 0 > PCI: Enabling device 0000:04:00.0 (0000 -> 0002) > ACPI: PCI Interrupt 0000:04:00.0[A] -> GSI 19 (level, low) -> IRQ 19 > ath5k_pci 0000:04:00.0: registered as 'phy1' > ath5k phy1: failed to wakeup the MAC Chip > ACPI: PCI interrupt for device 0000:04:00.0 disabled > ath5k_pci: probe of 0000:04:00.0 failed with error -5 Actually, I was asking you to test ath5k on a 32-bit kernel, but I'm quite sure now it would fail because of the resource issues. "failed to wakeup" is something more low-level and possibly specific to ath5k. I've seen it on newer cards. Either your card is still unsupported by ath5k, or maybe CardBus cannot work with addresses over the 4G boundary at all. Perhaps you should ask in the PCI list about it, not here. > [root@localhost ~]# modinfo ath5k > filename: > /lib/modules/2.6.24-2.fc9/kernel/drivers/net/wireless/ath5k/ath5k.ko Yes, that looks like a Fedora kernel. I think if you install "kernel-devel" in additional to the "kernel" package, you might be able to compile MadWifi for that kernel even with 32-bit tools. > If I reboot with the card inserted, the dmesg ath5k lines are: > (I have added pci=assign-busses for the kernel parameters) > > ath5k_pci 0000:07:00.0: registered as 'phy1' > ath5k phy1: failed to wakeup the MAC Chip > ath5k_pci: probe of 0000:07:00.0 failed with error -5 That's expected, considering that the I/O memory is still in the same place. > I have also tried: > > pci=assign-busses reservetop=64K cbmemsize=128K > > with no success. Have you read kernel-parameters.txt? reservetop is for 32-bit kernels only, and the syntax for cbmemsize is different. Actually, I don't think cbmemsize is what you need. > Shall I discard a problem with the Cardbus bridge configuration? I don't understand. What I would do would be to establish some facts, such as what would happen with less memory, whether ath5k works with the device in some configuration, whether MadWifi works in some configuration. And then I would try to find out what changes make the card stop working. Now you have data points that are too far from each other. MadWifi fails on a 32-bit kernel and ath5k fails on a 64-bit kernel. -- Regards, Pavel Roskin |