“Wireless networking on Linux is entering a new era. An era of bliss and ease; where users and network administrators have abundant time for relaxing lie-abouts on sunny warm hills because their wireless systems are humming along contentedly, instead of being vexing and unreliable. OK, so maybe it’s not going to be quite that Utopian, but things are definitely looking up, thanks to a lot of hard work by a lot of talented developers. So what’s going to be different? A brand-new wireless LAN kernel stack, which is going to reduce the current WLAN herd-of-cats approach to a single unified subsystem that supports all wireless drivers.”
I’ve heard this before, and yet it still sucks on Linux, every time someone talks about it getting better, it still sucks on Linux. I somehow doubt this will fix it, people ought to work on importing OpenBSD’s wireless instead.
Oh please… At least RTFA. The whole point is they released this has been the case, and they are going to try and address whats been the root cause.
Go troll over something else thats legit w/Linux, every one knows there is enough things that you don’t have to make it up.
I am sure they realised this was the case. that everything in the Linux wireless world sucks, and I am sure they want to correct the problem, but they’ve tried this before, a couple times and each time it has failed.
Linux wireless has sucked each time someone decided they would, “fix it.” As I already said. As the article itself explains, this is something legitimately wrong with Linux, that’s why they’re doing this, again.
The great thing about Free Software is that you are allowed to fail. You can fail as many times as it needed to fix a problem.
I would have to agree. The USB stack has been rewritten several times and now it is one of the best, if not the best, USB stack out there.
Aren’t we exaggerating a bit? Mmmh?
Wireless certainly does not come on my laptop. And it’s a Broadcom chipset, too. It works wonderfully, in fact. Network Manager has *drastically* improved over the past few months. I daresay Wireless is even better than on Windows – at least on my laptop.
The fact that it works great has little to do with the inclusion of the new devicescape stack, however. It’s already good without it, it will be better with it. The new kernel will be the first one with the stack in place, incidentally, so for you to say that it sucks means that you are somehow able to see into the future. Or are a troll.
I think I’ll go with the second choice.
I didn’t say this new thing will suck, I did say that everything done so far has sucked. Rather than repeatedly screw it up, I think they should just take the one that works and use it, it’s BSD licensed afterall.
No need to throw words in my mouth, I have enough there on my own thank you.
Don’t fret about being modded down.. many osnews readers have become overly sensitive and sensationalist.
Rather than repeatedly screw it up, I think they should just take the one that works and use it, it’s BSD licensed afterall.
What’s wrong in trying different ways to reach the optimum? IMHO it’s just good that they take the time and have the will to fix and improve what is broken, and every time they learn something new. It’s not away from me anyway, in fact it’s quite the opposite: I can just enjoy the benefits. Oh, and about taking one that works…Well, you know nothing is perfect and can always be improved?
But lets remember, thats no different to Microsoft Windows. Windows XP had not wireless stack, hence the reason when you downloaded a driver off Intel it was around 39mb for the complete software stack – because each vendor provided their own stack.
Windows Vista is the first release of Windows which includes a unified share wireless stack on which all vendors can base their wireless driver upon – but from my experience, the quality has been pretty damn flaky.
Linux 2.6.22 will have the new stack, but one has to remember one thing; drivers will have to be rewritten against the new stack – Intel is doing that right now, but its going to be a hard road as not all drivers will be ready by the time of its merging.
In the mean time, you can test out Intels driver from their website: http://intellinuxwireless.org/?p=iwlwifi
Much easier to set up Wi-Fi in Vista than XP. My only real problem right now is that video playback is bloody slow, except for Flash videos, even with all the visual effects turned off (only have a 1.86Ghz Core Solo with 512MB RAM, 64 of which are used by Intel GMA 950 integrated graphics). Do I need to update the video drivers or DirectShow codecs?
Mate, the machine is so under powered for Windows Vista, it should be a criminal offence if the company you bought it off had it preloaded. 512MB is not even enough for Vista,let alone the 448mb left after graphics is taken out of the equation, it still be woefully under powered.
My suggestion, grab a copy of Ubuntu, max out the ram, and go from there, you’ll find your life will be alot easier.
As a side note, they’re still selling Core Solo machines?
The shop threw in Vista Starter Edition for free, in addition to a laptop bag and mini optical mouse. Already dual-booting Ubuntu 7.04, but sound is not working, and I have to fix the screen rez. Maybe I should try Fedora again (stopped at FC4).
Have you got the specifications of the machine? I’m very surprised that the sound isn’t working given that most machines these days are loaded with either AC97 or Intel HD sound chips.
Regarding Fedora, If I was you I’d probably wait till Fedora 7 is released or at least track down the sound issue on you laptop; also, with the screen resolution, make sure the correct driver is being used – I am not too sure whether the Intel accelerated driver has been merged yet – http://www.intellinuxgraphics.org/
Fixed the screen rez by switching to the “intel” driver instead of “i810”. There’s at least one other guy on ubuntuforums having sound problems with my laptop (Acer Aspire 5570) under 7.04.
Linux wireless has sucked each time someone decided they would, “fix it.”
As far as I know, wireless networking support in Linux has never been fixed. It sort of evolved organically into the mess it is today, hence the need to fix it. Another poster mentioned the USB stack. That was fixed at least two times. SCSI was fixed at least once. The Linux kernel devs aren’t afraid to fix things. That’s why they don’t guarantee a stable in-kernel API.
I don’t quite understand what you’re griping about. Nobody is trying to brush anything under the rug and say that all is well in Linux-land. Are you upset that something is being done about a deficiency in the Linux kernel? Is it because that would remove an item from your list of reasons why Linux sucks compared to OpenBSD?
I’m personally not familiar with the 802.11 stack in OpenBSD or how the drivers interact with it. It wouldn’t be especially off-topic for you to elaborate on its design. I’m sure the Linux developers didn’t blindly dismiss their implementation, but often porting low-level driver frameworks from one kernel to another is more trouble than it’s worth. I’m sure you realize that this would not be a drop-in solution.
It might be in OpenBSD’s best interests that Linux NOT adopt its framework. If the GPL version becomes incompatible with the BSD version somewhere down the road, OpenBSD might be stuck maintaining drivers that they used to get from upstream vendors. Instead, OpenBSD has no reason to believe that drivers will ever come from upstream, given the incompatibility with Linux and the vanishingly small installed base of OpenBSD on mobile computers. I see where you’re coming from, hoping OpenBSD can gain some hardware support on the coattails of Linux’s success, but as an OpenBSD supporter, you must realize that your OS has to sink or swim on its own.
And quit downplaying the bcm fiasco. That’s as big of faux pas as they come in free software. Committing code to an open source project, under any license, comes with its share of responsibilities. In the Linux kernel, you have to sign-off on every commit to certify that this code is freely licensable under the GPL. The OpenBSD folks should do something similar. You expect the Linux developers to apologize for going public like that guy apologized for getting in the way of Dick Cheney’s bullet.
There is one driver framework that has been supporting both BSD’s and Linux from the beginning: DRI. (and I think maybe even Solaris, but not sure)
I see a parallel to recent wireless issue:
Lately DRM/DRI is gaining needed care because it has at least one huge graphic card vendor behind: Intel. At least this is valid for Linux side of it. That is, new “TTM” memory manager, which will someday go into the upstream kernel, is not ported to BSD’s. Unless a skilled developer(s) shows up and starts caring about this, BSD might start lagging behind even in the area of free graphic drivers. Same story is valid for modesetting, which will after all end up in kernel (ported from Xorg drivers) and be tightly coupled to DRI memory manager (old GGI/KGI grail being fulfilled but in a more proper way). So again, if no BSD mantainers show up, this will be Linux-only thing, for open source drivers at least. And I bet that developer attention will shift from Xorg to those kernel-drivers, a situation that won’t be particularly good for other Unixes that use Xorg.
It is also attitude of some developers that DRI/DRM should be purely GPL to prevent another ATI case (i.e. when they took BSD (dual)licensed DRI code and made their closed-source Linux driver from it). Though at this moment this seems unlikely to happen again, as ATI is doing a unified Vista/XP/Linux OpenGL driver, while Nvidia has one already.
I never said anyone was trying to hide the poor state of Linux wireless, hell, the article talks about it. You’re not reading hard enough.
I have to use Linux, it’s a fact of life, I would love nothing more than to never have to run it for anything, the instability of interfaces alone bugs the shit out of me, but it is also extremely annoying when someone wants a wireless solution using Linux. Most of the time I can talk them down from it, but not always, and then I am stuck in the mess of the Linux wireless world, where barely anything works consistantly and there are always issues with driver X and kernel Y, not to mention userland tool Z.
OpenBSD doesn’t have this, and haven’t had this for a while now. Taking this code from OpenBSD, while a bit of work, saves the work of implenting Yet Another Wireless Solution. I don’t actually care if it’s in OpenBSD’s best interest that someone use OpenBSD’s code, that’s why it’s BSD licensed, so that it can be used. OpenBSD does not get drivers from vendors, they make them on their own or they import them from other BSDs, or reimplement them based on binary, GPL or otherwise unusable drivers. OpenBSD stands to gain nothing from someone else using their code, but the code is there and already works.
I could care less the reasoning behind Linux developers seeming insistance that BSD code is somehow evil, but it’s there and ready to be used. I am baffled by this belief of yours that I hope this should somehow help OpenBSD, I don’t care how it effects OpenBSD. The context of this article was, “Linux’s wireless sucks, this will surely save us!” And I was pointing out that there is already a working solution that does the job just fine in existance and there is no need to make the wheel again.
I am not downplaying the issue with the bcw driver, it was not theft, read the definition of the word, it was copyright infringement, as I said. I also said that the developers of the bcm driver were dicks about it, rather than even once trying to contact the developer who f–ked up instead of making a massive fiasco out of it, they made a mass mailing accusing the man instead of being civil. As if somehow Marcus were a vile and venomous viper, poised to strike at the heart of their project and destroy it with it’d deadly toxins, rather than a guy trying to figure out how the hell a driver worked, and was testing code to do so. Yes, he committed code he shouldn’t have, it was in fact illegal, and I am sure you’ve never made a mistake either. I expected the driver’s developers to act like adults, rather than children, a pity I hoped for too much.
That you would compare a guy being shot in the face with a guy who’s code was copied is rather silly, I don’t recall any harm at all coming to the bcm driver’s developers because of that commit, unless perhaps it was carpal tunnel syndrome, while Harry Whittington was hospitalized.
I actually do think the driver’s developers should apologize for being total dicks about the entire matter, but I find it highly unlikely – since they started the whole issue off on an uncivil note, it probably won’t be finished on a civil one.
Why would you say that? It seems like OpenBSD developers got a bit upset when they were caught red handed stealing Linux wireless drivers and got called out on it:
http://thread.gmane.org/gmane.linux.kernel.wireless.general/1558
You know, it’s kind of sad that you don’t know the difference between copyright infringment and stealing, the fact of the matter is that they are drastically different.
It is true that a developer of the OpenBSD driver was careless enough to commit code straight from a GPL driver into CVS, and while it was wrong, it was only a mistake that noone need to be a dick about.
Perhaps you need to read up on the definition of theft, since you’re so lacking in basic lingual comprehension.
Wireless certainly has been one of the most irritating issues I’ve encountered on notebooks with Linux… I’ve been keeping an eye on this stack just out of interest (not a Linux user personally, but have friends who are) and am glad its finally going to be replacing the current mess.
Configuration is rather tricky, too.
ipw3945 will work under Gentoo, but I still haven’t got it coming up at boot time in a stable fashion.
We’ve got firmware, we’ve got kernel driver, we’ve got regulatory daemon, we’ve got udev, we’ve got /lib/udev/firmware.sh, we’ve got the /etc/init.d stuff…
prism54, OTOH, has firmware and kernel driver, and comes up like a champ.
What PC cards currently on the market use prism54? I should like to have a few o’ them dudes in standby.
This is really great news! GO LINUX! GO GNU!
Best operating system ever.
Now if someone ported PF to Linux, I would actually be interested to play with it
If there’s anything that’s diminished my enjoyment of Linux, it’s wireless issues. Everything else is pretty good, and when I have to use a PC instead of a Mac, I will use Linux when possible. But no wi-fi = no Linux.