Renee Danson wrote: > On Wed, May 28, 2008 at 03:28:18PM +0800, Fu, Zhonghui wrote: >> Kaiwai Gardiner, >> >> The following is my test and it's result: >> >> Parameters: >> >> test case: surfing the internet >> time: 4 hours >> OS: B81 and B86 >> driver: iwk1.1 >> security: open >> configure mode: nwamd and manual >> >> Result: >> >> B81 + nwamd: work normally and work normally after OS reboot >> B81 + manual: work normally and work normally after OS reboot >> B86 + nwamd: "link down and up" occur and work normally after OS >> reboot B86 + manual: work normally and work normally after OS reboot >> >> Taking account of above results, I think iwk driver is good and the >> problem is from nwamd. So, could the developer of nwamd please give >> more explanations about this problem? > > Could you explain what you mean by "link down and up"? How did you > observe this? Did you see syslog messages from the mac module, like > those that Matthew first reported? Zhonghui: Yes, I saw the same syslog messages from the mac module as what Matthew first reported.
> > Also, what exactly does link down/up mean for the iwk driver? This > concept is (at least for me) much less clear with wireless than with > wired links. Does "link up" map to "connected to an AP"? Or is there > some other meaning? Zhonghui: "link up/down" means connected/disconnected to an AP for iwk driver. > > There are only two cases in which nwam would disconnect (from an AP) > a wifi link: if it decided a preferred link was available (i.e. you > plugged in your ethernet device); or if it determined that the signal > strength was "very weak" (as reported by libdladm). Is it possible > that signal strength was the problem during your tests? Zhonghui: I am sure that above two cases had not occured during my test. > > Regarding the problem Matthew reported (link down following the error > report from ieee80211_crypto_enmic), it sounds like Quaker had some > possibilities to investigate. In particular, there's the CR that he > mentioned, 6698754, which is still open. Zhonghui: I used the Open mode in my test, and had not occured ieee80211_crypto_enmic error. > > There were also several bugs fixed in the iwk driver in nevada build > 87; Matthew was running build 90, but your tests were against builds > 81 and 86, and would not have included those fixes. Zhonghui: Although I done my test using B81 and B86, but I used the same version of iwk driver as that in B90. BTW, I am one of developers of iwk driver in Intel. Thanks, Zhonghui > > I also have a question for Matthew: > >>> Kaiwai Gardiner wrote: >>>> On Tue, 2008-05-27 at 10:07 +0800, Fu, Zhonghui wrote: >>>>> Kaiwai Gardiner, >>>>> >>>>> Have you used other wireless cards in addition to 4965? Is there >>>>> the same phenomena as 4965 on them? >>>>> Could you please give out your operation steps in detail? >>>> >>>> No, unfortunately not. I only have this one laptop. I do have a >>>> athereos one in my bedroom in a desktop and it doesn't display the >>>> same rate of dropping out - and dladm states that the reception is >>>> poor. Soomething is going wrong when a poor signal is maintaining a >>>> conneciton whilst this laptop classes the signal as excellent - the >>>> reception keeps falling out of the sky. > > I'm a little confused about the signal levels. Which device reports > the poor signal strength: the atheros or the iwk? > > Thanks, > renee
