[opensuse-factory]problem with ath5k driver?
I have been trying to copy a file from my PowerBook to my Thinkpad over the network. The Thinkpad has a Netgear wireless card that uses the ath5k driver and the Powerbook uses a DLink card that uses the madwifi driver(the chipset isn't supported by ath5k). Well, after a lot of frustration and thinking it was the Powerbook, it turned out to be the Thinkpad(at which point my son decided to rub it in my nose because the Powerbook is his). Anyway, the problem had to do with the MAC address being lost. When I did a dmesg, I got the following: recvmsg bug: copied E638B28A seq E638B832 KERNEL: assertion (flags & MSG_PEEK) failed at net/ipv4/tcp.c (1354) Over and over again. Any idea if this is a Kernel problem or a module problem? And which way I should report it? Thanx --------------------------------------------------------------------- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-factory+help@opensuse.org
On Sun, Jun 08, 2008 at 07:27:46PM -0400, Larry Stotler wrote:
I have been trying to copy a file from my PowerBook to my Thinkpad over the network. The Thinkpad has a Netgear wireless card that uses the ath5k driver and the Powerbook uses a DLink card that uses the madwifi driver(the chipset isn't supported by ath5k). Well, after a lot of frustration and thinking it was the Powerbook, it turned out to be the Thinkpad(at which point my son decided to rub it in my nose because the Powerbook is his).
Anyway, the problem had to do with the MAC address being lost. When I did a dmesg, I got the following:
recvmsg bug: copied E638B28A seq E638B832 KERNEL: assertion (flags & MSG_PEEK) failed at net/ipv4/tcp.c (1354)
Over and over again. Any idea if this is a Kernel problem or a module problem? And which way I should report it?
Please open a bug at Novell bugzilla. With above information and if possible more dmesg output. Ciao, Marcus --------------------------------------------------------------------- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-factory+help@opensuse.org
Am Montag, 9. Juni 2008 schrieb Marcus Meissner:
On Sun, Jun 08, 2008 at 07:27:46PM -0400, Larry Stotler wrote:
I have been trying to copy a file from my PowerBook to my Thinkpad over the network. The Thinkpad has a Netgear wireless card that uses the ath5k driver and the Powerbook uses a DLink card that uses the madwifi driver(the chipset isn't supported by ath5k). Well, after a lot of frustration and thinking it was the Powerbook, it turned out to be the Thinkpad(at which point my son decided to rub it in my nose because the Powerbook is his).
Anyway, the problem had to do with the MAC address being lost. When I did a dmesg, I got the following:
recvmsg bug: copied E638B28A seq E638B832 KERNEL: assertion (flags & MSG_PEEK) failed at net/ipv4/tcp.c (1354)
Over and over again. Any idea if this is a Kernel problem or a module problem? And which way I should report it?
Please open a bug at Novell bugzilla.
With above information and if possible more dmesg output.
There is already a bugzilla about ath5k, it seems to be generally buggy: bug 398011 and we'll have to do a kernel update later. Greetings, Stephan --------------------------------------------------------------------- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-factory+help@opensuse.org
On Mon, Jun 9, 2008 at 3:08 AM, Stephan Kulow
There is already a bugzilla about ath5k, it seems to be generally buggy: bug 398011 and we'll have to do a kernel update later.
I have another one open about the compatibility problem. However, this is the first issue I've had with it. It only happens when I tried that file copy. IIRC, it didn't do that before, but I'm not certain if I had actually tried a copy since Beta3 or not. Otherwise, it's been a solid driver. Well, then with the release of 11.0 soon, maybe a warning should be added to YaST's configuration about this? Due to this and the other compatibility problem, I think we are going to see a lot of disgruntled users who have atheros based cards..... --------------------------------------------------------------------- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-factory+help@opensuse.org
On Mon, 2008-06-09 at 12:13 -0400, Larry Stotler wrote: Well, then with the release of 11.0 soon, maybe a warning should be
added to YaST's configuration about this? Due to this and the other compatibility problem, I think we are going to see a lot of disgruntled users who have atheros based cards.....
Well, considering Atheros based cards haven't been supported out of the
box in openSUSE since 10.1; I don't think most users will have a
problem. Those that do will install MadWifi like we've had to before.
--
Kevin "Yo" Dupuy
Public Mail
On Mon, Jun 9, 2008 at 3:22 PM, Kevin Dupuy
Well, considering Atheros based cards haven't been supported out of the box in openSUSE since 10.1; I don't think most users will have a problem. Those that do will install MadWifi like we've had to before.
No. The problem is that the "new" ath5k driver is now included in the kernel. You don't have to build madwifi for the 52xx series chips. However, YaST will see older chips(like the 2413 in my DLink card) and then automatically use the ath5k driver WHICH WILL NOT WORK! Therefore, someone who doesn't know will just assume that their wireless will work, and then it WON'T. Trust me, I discovered the problem on my Powerbook. It works fine after I compile the madwifi driver. But anyone who uses automatic setup for hardware will have the ath5k driver used for nonsupported cards and then they won't have wireless working. Support worked just fine for my 5212/5213 based card(although I did recently uncover that bug....). --------------------------------------------------------------------- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-factory+help@opensuse.org
2008/6/9 Larry Stotler
No. The problem is that the "new" ath5k driver is now included in the kernel. You don't have to build madwifi for the 52xx series chips. However, YaST will see older chips(like the 2413 in my DLink card) and then automatically use the ath5k driver WHICH WILL NOT WORK!
and also use the same driver for Atheros card that are not supported yet, like the AR2425 that I have... regards, Luiz --------------------------------------------------------------------- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-factory+help@opensuse.org
On Mon, 2008-06-09 at 15:37 -0400, Larry Stotler wrote:
No. The problem is that the "new" ath5k driver is now included in the kernel. You don't have to build madwifi for the 52xx series chips. However, YaST will see older chips(like the 2413 in my DLink card) and then automatically use the ath5k driver WHICH WILL NOT WORK!
Therefore, someone who doesn't know will just assume that their wireless will work, and then it WON'T. Trust me, I discovered the problem on my Powerbook. It works fine after I compile the madwifi driver. But anyone who uses automatic setup for hardware will have the ath5k driver used for nonsupported cards and then they won't have wireless working.
I know. I have the same issue. YaST detects my card, but I can't find
any wireless networks. I filed a bug report months ago, which ended up
being a duplicate of bug #373098.
--
Kevin "Yo" Dupuy
Public Mail
On Mon, Jun 9, 2008 at 5:47 PM, Kevin Dupuy
I know. I have the same issue. YaST detects my card, but I can't find any wireless networks. I filed a bug report months ago, which ended up being a duplicate of bug #373098.
I originally filed my bug because it didn't work on my Powerbook and I thought it was a PPC issue. While I'm glad they are developing a module that can be put into the kernel, I really wish they would have added full support for all the cards instead of just a subset. Of course, they ported a driver from a BSD project, so I guess YMMV..... --------------------------------------------------------------------- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-factory+help@opensuse.org
On Mon, Jun 09, 2008 at 07:14:34PM -0400, Larry Stotler wrote:
On Mon, Jun 9, 2008 at 5:47 PM, Kevin Dupuy
wrote: I know. I have the same issue. YaST detects my card, but I can't find any wireless networks. I filed a bug report months ago, which ended up being a duplicate of bug #373098.
I originally filed my bug because it didn't work on my Powerbook and I thought it was a PPC issue.
While I'm glad they are developing a module that can be put into the kernel, I really wish they would have added full support for all the cards instead of just a subset. Of course, they ported a driver from a BSD project, so I guess YMMV.....
They have added support for all cards that they currently know how to control. The blame really has to go back to the manufacturer, they are the ones not supporting Linux here, we are trying to play catchup... thanks, greg k-h --------------------------------------------------------------------- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-factory+help@opensuse.org
On Mon, Jun 9, 2008 at 7:28 PM, Greg KH
They have added support for all cards that they currently know how to control. The blame really has to go back to the manufacturer, they are the ones not supporting Linux here, we are trying to play catchup...
The madwifi driver supports the older cards. Not being a programmer, I have no idea how hard it is to do all this. --------------------------------------------------------------------- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-factory+help@opensuse.org
On Mon, 2008-06-09 at 19:34 -0400, Larry Stotler wrote:
The madwifi driver supports the older cards. Not being a programmer, I have no idea how hard it is to do all this.
The madwifi contain a binary only part that make it non-redistribuable in openSUSE. This is the missing part for which Atheros has provided NO documentation for Linux developers to write it. Without this documentation writing the code is MUCH HARDER and take MUCH LONGER. Hub --------------------------------------------------------------------- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-factory+help@opensuse.org
On Mon, Jun 09, 2008 at 07:34:08PM -0400, Larry Stotler wrote:
On Mon, Jun 9, 2008 at 7:28 PM, Greg KH
wrote: They have added support for all cards that they currently know how to control. The blame really has to go back to the manufacturer, they are the ones not supporting Linux here, we are trying to play catchup...
The madwifi driver supports the older cards. Not being a programmer, I have no idea how hard it is to do all this.
It's legally impossible to redistribute that, as it is a closed source module :( Go poke Atheros, not much we can do here... thanks, greg k-h --------------------------------------------------------------------- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-factory+help@opensuse.org
Tirsdag den 10. Juni 2008 01:45:28 skrev Greg KH:
The madwifi driver supports the older cards. Not being a programmer, I have no idea how hard it is to do all this.
It's legally impossible to redistribute that, as it is a closed source module :(
Go poke Atheros, not much we can do here...
I believe it's no longer necessary to poke Atheros... well, maybe to make them move faster... http://lwn.net/Articles/278132/ --------------------------------------------------------------------- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-factory+help@opensuse.org
Am Montag, 9. Juni 2008 21:37:09 schrieb Larry Stotler:
On Mon, Jun 9, 2008 at 3:22 PM, Kevin Dupuy
wrote: Well, considering Atheros based cards haven't been supported out of the box in openSUSE since 10.1; I don't think most users will have a problem. Those that do will install MadWifi like we've had to before.
No. The problem is that the "new" ath5k driver is now included in the kernel. You don't have to build madwifi for the 52xx series chips. However, YaST will see older chips(like the 2413 in my DLink card) and then automatically use the ath5k driver WHICH WILL NOT WORK!
Therefore, someone who doesn't know will just assume that their wireless will work, and then it WON'T. Trust me, I discovered the problem on my Powerbook. It works fine after I compile the madwifi driver. But anyone who uses automatic setup for hardware will have the ath5k driver used for nonsupported cards and then they won't have wireless working.
Support worked just fine for my 5212/5213 based card(although I did recently uncover that bug....).
not all, my 5212/13 doesn't work with the ath5k also(there exist 3 Versions of this one). Maybe we must ship an instruction to blacklist the ath5k in most cases. I think because of not working properly yet, it would be better not to include ath5k in kernel. Daniel
--------------------------------------------------------------------- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-factory+help@opensuse.org
-- Betriebssystem: Linux 2.6.22.17-0.1-default x86_64 System: openSUSE 10.3 (x86_64) KDE: 3.5.9 "release 62.3" --------------------------------------------------------------------- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-factory+help@opensuse.org
On Tue, Jun 10, 2008 at 3:27 AM, Daniel Fuhrmann
not all, my 5212/13 doesn't work with the ath5k also(there exist 3 Versions of this one). Maybe we must ship an instruction to blacklist the ath5k in most cases. I think because of not working properly yet, it would be better not to include ath5k in kernel.
Well, that's a little harsh. Considering it actually works just fine on my system and I didn't have to compile madwifi, I would agure that it just shouldn't automatically install the ath5k driver. Other than that copy issue that I started this thread with, it seems to work for me. --------------------------------------------------------------------- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-factory+help@opensuse.org
Larry Stotler escribió: I would agure that
it just shouldn't automatically install the ath5k driver.
If not installed automatically, no one ever will let us to know what the problems are, nor report bugs etc.. and will continue to have a non working card. -- "Progress is possible only if we train ourselves to think about programs without thinking of them as pieces of executable code.” - Edsger W. Dijkstra Cristian Rodríguez R. Platform/OpenSUSE - Core Services SUSE LINUX Products GmbH Research & Development http://www.opensuse.org/
On Tuesday 10 June 2008 04:36:00 pm Cristian Rodríguez wrote:
Larry Stotler escribió: I would agure that
it just shouldn't automatically install the ath5k driver.
If not installed automatically, no one ever will let us to know what the problems are, nor report bugs etc.. and will continue to have a non working card.
It was mentioned earlier that are known cards that currently don't work, so if that is right, installation in such cases is pointless. Better to save user patience to handle bugs for software that works to some extent. -- Regards, Rajko http://en.opensuse.org/Portal needs helpful hands. --------------------------------------------------------------------- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-factory+help@opensuse.org
Am Dienstag, 10. Juni 2008 22:41:51 schrieb Larry Stotler:
On Tue, Jun 10, 2008 at 3:27 AM, Daniel Fuhrmann
wrote: not all, my 5212/13 doesn't work with the ath5k also(there exist 3 Versions of this one). Maybe we must ship an instruction to blacklist the ath5k in most cases. I think because of not working properly yet, it would be better not to include ath5k in kernel.
Well, that's a little harsh. Considering it actually works just fine on my system and I didn't have to compile madwifi, I would agure that it just shouldn't automatically install the ath5k driver. Other than that copy issue that I started this thread with, it seems to work for me. Ok, I thougt for an extra package installable by Yast, and who likes to test it can do, and the others have to build themselves or use madwifi repo.
One of my atheros cards seem to work with ath5k but only without encryption (not tested).
--------------------------------------------------------------------- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-factory+help@opensuse.org
-- Betriebssystem: Linux 2.6.22.17-0.1-default x86_64 System: openSUSE 10.3 (x86_64) KDE: 3.5.9 "release 64.2 --------------------------------------------------------------------- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-factory+help@opensuse.org
On Wed, Jun 11, 2008 at 2:54 AM, Daniel Fuhrmann
Ok, I thougt for an extra package installable by Yast, and who likes to test it can do, and the others have to build themselves or use madwifi repo.
ath5k isn't a package, it's included in the kernel now.
One of my atheros cards seem to work with ath5k but only without encryption (not tested).
Can't verify encryption - don't use any. --------------------------------------------------------------------- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-factory+help@opensuse.org
Ok, I just did the kernel update for RC1 and the driver just still isn't stable enough. Went ahead and compiled the madwifi driver and now it works perfectly. Oh well. --------------------------------------------------------------------- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-factory+help@opensuse.org
participants (11)
-
Cristian Rodríguez
-
Daniel Fuhrmann
-
Greg KH
-
Hubert Figuiere
-
Kevin Dupuy
-
Larry Stotler
-
Marcus Meissner
-
Martin Schlander
-
Rajko M.
-
Stephan Kulow
-
¡ElCheVive!