From: Harry O. <ho...@ho...> - 2005-12-12 03:34:52
|
>From: Denis Vlasenko <vd...@il...> >To: acx...@li... >CC: "Harry Orenstein" <ho...@ho...> >Subject: Re: [Acx100-users] HWP54G works (and a question) >Date: Sun, 11 Dec 2005 18:35:15 +0200 > >On Saturday 10 December 2005 17:38, Harry Orenstein wrote: > > Just wanted to report that I got the Hawking HWP54G working. Details >below: > > > > Driver: 20051202 > > Firmware: tiacx111c16 1.2.1.34 > > Kernel: 2.6.10-1.770_FC3 > > > > I got this working with auto rate (I've seen anywhere from 1M to 36M) >and > > WAP using a restricted mode key and it works automatically on boot up. > > > > I did have a couple of questions: I order to get the driver to compile I >had > > to change both "#ifdef CONFIG_PM" in pci.c to "#ifndef CONFIG_PM" since >my > > kernel has CONFIG_PM and the code within the #ifdef does not want to >compile > > on my system. There is mention in the comments that "PM code needs to >be > >What was the error? > > > fixed / debugged / tested". Is it a problem not enabling power >management? > > Is there any info I can supply that will assist the developers to get >this > > working? > >By describing in detail how it does not work. Compile errors, logs of >suspend/resume cycle, etc... Actually, I did some more research. The problem is with acxpci_e_suspend. The parameter "state" is defined as pm_message_t in kernel 2.6.12 and u32 in 2.6.10 (I don't have 2.6.11, so I don't know exactly when it changed). The following patch fixes it for me: --- pci.c.orig 2005-12-11 22:23:56.046909787 -0500 +++ pci.c 2005-12-11 00:06:06.000000000 -0500 @@ -1853,7 +1853,11 @@ #ifdef CONFIG_PM static int if_was_up = 0; /* FIXME: HACK, do it correctly sometime instead */ static int +#if LINUX_VERSION_CODE >= KERNEL_VERSION(2, 6, 11) acxpci_e_suspend(struct pci_dev *pdev, pm_message_t state) +#else +acxpci_e_suspend(struct pci_dev *pdev, u32 state) +#endif { struct net_device *dev = pci_get_drvdata(pdev); wlandevice_t *priv = netdev_priv(dev); I got the info about what to change from (IIRC) one of the forums. > > > Also: when the driver loads it seems to configure USB for the card >(usbcore: > > registered new driver acx_usb in the log) even though it is PCI. Is >this > >It registers the driver, but finds no usb devices to operate on. > > > the result of the PCI and USB modules having been combined (the first > > version of the driver I tried to run was acx100-0.2.0pre8_plus_fixes_57, > > which had separate PCI and USB modules)? Is this normal? > >Yes, it's ok. Most of driver code is shared between pci and usb. >-- >vda Thanks for the reply. Sorry for the incomplete error report. I was really just trying to find out if I could safely skip over the code within the #ifdefs. Guess I should have thought about how my question might help others. I hope the patch above contributes. -- Harry O. |