bcm4312 on again, off again
It took a BS programmer over five hours, but he got my bcm4318 chip talking to my computer with the NDISWrapper and a kernel build. Before that the wireless was showing on the laptop's wifi switch light but not making up to the computer to actually use the signal. Now, after six days, it suddenly stops. Give it a few days, and it's on again. Any clues? ___ Bcm43xx-dev mailing list Bcm43xx-dev@lists.berlios.de https://lists.berlios.de/mailman/listinfo/bcm43xx-dev
Re: bcm4312 on again, off again
You know, people read this list TO HELP PEOPLE. That's right. I'm up at 2307 my time hoping that I can provide some information that will help you get your bcm43xx device functional, or that failing that I can help you provide information so the folks that work between the firmware, the reverse-engineered speficiations, and the driver can make it work. NOTHING would please me more than to see your device work. I feel I would not be happy if it were not to work. And yet, unfortunately, in this imperfect world I must work with some imperfections... like Zero diagnostic information. Ndiswrapper typically works in about 5 seconds, so not only no diagnostic information, but your "BS programmer" is 4 hours 59 minutes and 55 seconds full of "BS". You don't mention your kernel version. You don't include a uname -a or a dmesg | grep bcm43 Now let's talk English "chip talking" -- meaningless "a kernel build" -- meaningless "wireless was showing" -- meaningless "wifi switch light" -- meaningless "not making up to" -- meaningless "actually use the signal" -- meaningless "it suddenly stops" -- almost meaningless. SOmething that might have been previously defined (but wasn't) no longer words. "it's on again" -- same comment as above "Any clues" -- no, you don't have any SO... I'd say "good luck" but it looks like you don't have any of that either. Ehud Robert Easter wrote: It took a BS programmer over five hours, but he got my bcm4318 chip talking to my computer with the NDISWrapper and a kernel build. Before that the wireless was showing on the laptop's wifi switch light but not making up to the computer to actually use the signal. Now, after six days, it suddenly stops. Give it a few days, and it's on again. Any clues? ___ Bcm43xx-dev mailing list Bcm43xx-dev@lists.berlios.de https://lists.berlios.de/mailman/listinfo/bcm43xx-dev smime.p7s Description: S/MIME Cryptographic Signature ___ Bcm43xx-dev mailing list Bcm43xx-dev@lists.berlios.de https://lists.berlios.de/mailman/listinfo/bcm43xx-dev
Re: bcm4312 on again, off again
On Thu, Jul 05, 2007 at 12:44:23AM -0500, Robert Easter wrote: > It took a BS programmer over five hours, but he got my bcm4318 chip > talking to my computer with the NDISWrapper and a kernel build. It should have taken him only one kernel build to get the 4318 working with the open source driver bcm43xx driver shipped with the latest -rc kernel. > Before > that the wireless was showing on the laptop's wifi switch light but not > making up to the computer to actually use the signal. Now, after six > days, it suddenly stops. Give it a few days, and it's on again. Any clues? Eh yeah. No driver loaded means no working network device. Hardly a surprise, I'd say. If you're using ndiswrapper right now, you're at the vendor's whim. We don't have the source of the Windows driver you loaded, so we can't help you. You *might* try talking to your vendor, but I don't think they consider Linux + ndiswrapper as a supported configuration for their driver. They joys of proprietary software... If you want people over here to be able to help you, use the open source driver. Oh, and please be precise: your subject says you're using a bcm4312, but in the body you talk about a bcm4318. Erik [using a 4318 with stock linux-2.6.22-rc7] -- They're all fools. Don't worry. Darwin may be slow, but he'll eventually get them. -- Matthew Lammers in alt.sysadmin.recovery signature.asc Description: Digital signature ___ Bcm43xx-dev mailing list Bcm43xx-dev@lists.berlios.de https://lists.berlios.de/mailman/listinfo/bcm43xx-dev
Re: bcm4312 on again, off again
On 7/4/07, Robert Easter <[EMAIL PROTECTED]> wrote: > It took a BS programmer over five hours, but he got my bcm4318 chip > talking to my computer with the NDISWrapper and a kernel build. Before > that the wireless was showing on the laptop's wifi switch light but not > making up to the computer to actually use the signal. Now, after six > days, it suddenly stops. Give it a few days, and it's on again. Any clues? Ok i am going to try to be nice, as i myself have been chewed out by this forum before (improper email reply edict), and as i had a similar problem i think, but as has been explained before its hard to tell. First there is a transmit power problem that has been hard to fix (correct me if i have misunderstood some of the other emails that flow though here) this is going to make connecting to some networks very flaky if not impossible when you are father away. As for the ability to connect sometimes and not others, it is something called i need to use the new kernel. Or use rmmod bcm43xx a lot. If you are using a distros kernel, it is very likely that it wont work, so i direct your attention toward www.kernel.org. At the time that my system was exhibiting that behavior, i applied this patch and it went away https://lists.berlios.de/pipermail/bcm43xx-dev/2007-February/003819.html . With the newer kernels, it should not be a problem. I still have problems when i am to far away from my AP, but in a reasonable range, it is very functional. a uname -a would be helpful lspci | grep Broadcom dmesg | grep bcm43xx Also that tidbit about the light was completely useless, it is hardware not software (the filament in your ceiling light does not effect whether or not the wall switch opens and closes) If wont use the bcm43xx driver you are not going to find what people say here very useful. Well that is my best guess at what you were trying to say. -- Brennan Ashton Bellingham, Washington "The box said, 'Requires Windows 98 or better'. So I installed Linux" ___ Bcm43xx-dev mailing list Bcm43xx-dev@lists.berlios.de https://lists.berlios.de/mailman/listinfo/bcm43xx-dev
Re: bcm4312 on again, off again
On 7/5/07, Robert Easter <[EMAIL PROTECTED]> wrote: > Thanks, man- Just after I decided to make a display of my native > asininity the thing came back on, but I am keeping this letter as a > first try for next time it craps out. > > Brennan Ashton wrote: > > > > First there is a transmit power problem that has been hard to fix > > (correct me if i have misunderstood some of the other emails that flow > > though here) this is going to make connecting to some networks very > > flaky if not impossible when you are father away. > I guess that would explain the light-related problem, especially if it's > a power indicator and not a tx/rx detector... I would guess that the like is a link tx/rx indicator. There is usually a light that tells you if the wireless is turned on. > > > > As for the ability to connect sometimes and not others, it is > > something called i need to use the new kernel. > He pulled in the latest kernel last week off kernel.org, so I would at I wonder if he got 2.6.21.4 or 2.6.22.rc7 there is a big difference. run uname -a that will tell us > least hope it will be top kernel for a while before the next F7 build > comes out, & even then there's at least the hope that the new one will > adapt... (No?) > > Or use rmmod bcm43xx a > > lot. > I'll have to check that out & see what it is. it is a command. > > If you are using a distros kernel, it is very likely that it > > won't work, so i direct your attention toward www.kernel.org. At the > > time that my system was exhibiting that behavior, i applied this patch > > and it went away > > https://lists.berlios.de/pipermail/bcm43xx-dev/2007-February/003819.html > I will hang on to that link- thanx! wont work on the newer kernels > > > . With the newer kernels, it should not be a problem. I still have > > problems when i am to far away from my AP, but in a reasonable range, > > it is very functional. > > > > a uname -a would be helpful > > lspci | grep Broadcom > > dmesg | grep bcm43xx > What do these mean / do, by the way? it is a | not a / it is the shift \ looks like a colon, it is known as a pipe, and allows you to "pipe" the output of one command to another. Grep is used to search the result for a string (as well as other things). > > > > Also that tidbit about the light was completely useless, it is > > hardware not software (the filament in your ceiling light does not > > effect whether or not the wall switch opens and closes) > Yeah, but when the wifi was available again, the light started > flashing. When it had quit, so had the light. Maybe there's a hardware > issue I need to look into? doubt it, suggests that it is a link light. > > > > If wont use the bcm43xx driver you are not going to find what people > > say here very useful. > Well, what else is there? you said that you were using ndiswrapper you need to use that opensource driver bcm43xx with a firmware cutter. you will need to remove the ndiswrapper stuff, extract the firmware and load the modules (bcm43xx-mac80211). I believe once you have delt with ndiswrapper (cant help you there) you need to do somthing like this: wget http://downloads.openwrt.org/sources/broadcom-wl-4.80.9.2.tar.bz2 tar -xjvf broadcom-wl-4.80.9.2.tar.bz2 bcm43xx-fwcutter -w /lib/firmware/ broadcom-wl-4.80.9.2/kmod/wl_apsta.o echo 'blacklist bcm43xx' >> /etc/modprobe.d/blacklist (not positive about this) when you set up ndiswrapper you may have blacklisted bcm43xx-mac80211 if so you need to remove it. modprobe bcm43xx-mac80211 I think that is all correct. > > > > Well that is my best guess at what you were trying to say. > Sounds like you're a good guesser. I guess! > > Blessings on ya! > > r.e. > > > -- Brennan Ashton Bellingham, Washington "The box said, 'Requires Windows 98 or better'. So I installed Linux" ___ Bcm43xx-dev mailing list Bcm43xx-dev@lists.berlios.de https://lists.berlios.de/mailman/listinfo/bcm43xx-dev
Re: bcm4312 on again, off again
Brennan Ashton wrote: > On 7/4/07, Robert Easter <[EMAIL PROTECTED]> wrote: >> It took a BS programmer over five hours, but he got my bcm4318 chip >> talking to my computer with the NDISWrapper and a kernel build. Before >> that the wireless was showing on the laptop's wifi switch light but not >> making up to the computer to actually use the signal. Now, after six >> days, it suddenly stops. Give it a few days, and it's on again. Any clues? What does BS mean in this context? > Ok i am going to try to be nice, as i myself have been chewed out by > this forum before (improper email reply edict), and as i had a similar > problem i think, but as has been explained before its hard to tell. > First there is a transmit power problem that has been hard to fix > (correct me if i have misunderstood some of the other emails that flow > though here) this is going to make connecting to some networks very > flaky if not impossible when you are father away. Several of us try very hard to make this forum helpful, and not chew out the OP, but it is very hard to diagnose a problem when crucial information is missing. Suppose I asked for help in diagnosing my Ford truck and I didn't tell you that it is a diesel, and I didn't even say whether it started, or not. For the record, diagnosis of problems require the kernel version (uname -r), any patches that have been applied, the results from the command 'dmesg | grep bcm43xx', information about encryption on the access point, and any additional software being used to connect (NetworkManager, etc.). As stated below, if you want to use ndiswrapper, go elsewhere. We have done a lot of work to avoid that!! > As for the ability to connect sometimes and not others, it is > something called i need to use the new kernel. Or use rmmod bcm43xx a > lot. If you are using a distros kernel, it is very likely that it > wont work, so i direct your attention toward www.kernel.org. At the > time that my system was exhibiting that behavior, i applied this patch > and it went away > https://lists.berlios.de/pipermail/bcm43xx-dev/2007-February/003819.html > . With the newer kernels, it should not be a problem. I still have > problems when i am to far away from my AP, but in a reasonable range, > it is very functional. > > a uname -a would be helpful > lspci | grep Broadcom > dmesg | grep bcm43xx > > Also that tidbit about the light was completely useless, it is > hardware not software (the filament in your ceiling light does not > effect whether or not the wall switch opens and closes) > > If wont use the bcm43xx driver you are not going to find what people > say here very useful. > > Well that is my best guess at what you were trying to say. Larry ___ Bcm43xx-dev mailing list Bcm43xx-dev@lists.berlios.de https://lists.berlios.de/mailman/listinfo/bcm43xx-dev