[opensuse-factory] 12.2 ms 2, dropping wireless
uname -a Linux crash2 3.3.0-rc6-1-desktop #1 SMP PREEMPT Tue Mar 6 19:02:14 UTC 2012 (0d8ec4d) x86_64 x86_64 x86_64 GNU/Linux 27: None 00.0: 1070a WLAN [Created at net.124] Unique ID: AYEt.QXn1l67RSa1 Parent ID: y9sn.6YBoXuWNRV3 SysFS ID: /class/net/wlan0 SysFS Device Link: /devices/pci0000:00/0000:00:14.4/0000:02:08.0 Hardware Class: network interface Model: "WLAN network interface" Driver: "ath5k" Driver Modules: "ath5k" Device File: wlan0 HW Address: 00:0f:b5:f9:f9:c3 Link detected: yes Config Status: cfg=new, avail=yes, need=no, active=unknown Attached to: #19 (Ethernet controller) many log entries: ath5k phy0: gain calibration timeout (2462MHz) and loose wireless connection. Only can reconnect by warm/cold reboot. found on google two work-a-rounds: add to boot: processor.ignore_ppc=1 add to wireless card parameters/options: fastchanswitch the first has worked for me for about 12 hrs. did not have problems with previous kernel Have not filed bug-report ???? -- (paka)Patrick Shanahan Plainfield, Indiana, USA HOG # US1244711 http://wahoo.no-ip.org Photo Album: http://wahoo.no-ip.org/gallery2 http://en.opensuse.org openSUSE Community Member Registered Linux User #207535 @ http://linuxcounter.net -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
* Patrick Shanahan
uname -a Linux crash2 3.3.0-rc6-1-desktop #1 SMP PREEMPT Tue Mar 6 19:02:14 UTC 2012 (0d8ec4d) x86_64 x86_64 x86_64 GNU/Linux
27: None 00.0: 1070a WLAN [Created at net.124] Unique ID: AYEt.QXn1l67RSa1 Parent ID: y9sn.6YBoXuWNRV3 SysFS ID: /class/net/wlan0 SysFS Device Link: /devices/pci0000:00/0000:00:14.4/0000:02:08.0 Hardware Class: network interface Model: "WLAN network interface" Driver: "ath5k" Driver Modules: "ath5k" Device File: wlan0 HW Address: 00:0f:b5:f9:f9:c3 Link detected: yes Config Status: cfg=new, avail=yes, need=no, active=unknown Attached to: #19 (Ethernet controller)
many log entries: ath5k phy0: gain calibration timeout (2462MHz) and loose wireless connection. Only can reconnect by warm/cold reboot.
found on google two work-a-rounds: add to boot: processor.ignore_ppc=1 add to wireless card parameters/options: fastchanswitch
the first has worked for me for about 12 hrs. did not have problems with previous kernel
Have not filed bug-report ????
Neither of the above work-a-rounds do the job. Still dropping wireless connection with the "gail calibration timeout" notice. Help ???? tks, -- (paka)Patrick Shanahan Plainfield, Indiana, USA HOG # US1244711 http://wahoo.no-ip.org Photo Album: http://wahoo.no-ip.org/gallery2 http://en.opensuse.org openSUSE Community Member Registered Linux User #207535 @ http://linuxcounter.net -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
On 03/19/2012 02:35 PM, Patrick Shanahan wrote:
* Patrick Shanahan
[01-01-70 12:34]: uname -a Linux crash2 3.3.0-rc6-1-desktop #1 SMP PREEMPT Tue Mar 6 19:02:14 UTC 2012 (0d8ec4d) x86_64 x86_64 x86_64 GNU/Linux
27: None 00.0: 1070a WLAN [Created at net.124] Unique ID: AYEt.QXn1l67RSa1 Parent ID: y9sn.6YBoXuWNRV3 SysFS ID: /class/net/wlan0 SysFS Device Link: /devices/pci0000:00/0000:00:14.4/0000:02:08.0 Hardware Class: network interface Model: "WLAN network interface" Driver: "ath5k" Driver Modules: "ath5k" Device File: wlan0 HW Address: 00:0f:b5:f9:f9:c3 Link detected: yes Config Status: cfg=new, avail=yes, need=no, active=unknown Attached to: #19 (Ethernet controller)
many log entries: ath5k phy0: gain calibration timeout (2462MHz) and loose wireless connection. Only can reconnect by warm/cold reboot.
found on google two work-a-rounds: add to boot: processor.ignore_ppc=1 add to wireless card parameters/options: fastchanswitch
the first has worked for me for about 12 hrs. did not have problems with previous kernel
Have not filed bug-report ????
Neither of the above work-a-rounds do the job. Still dropping wireless connection with the "gail calibration timeout" notice.
Help ????
If the "processor.ignore_ppc=1" option fixes it, it does not seem to be a problem with ath5k, but with something deep in the kernel. I think you should post on the linux-kernel@vger.kernel.org ML. Just in case it is something wilt ath5k, also post on linux-wireless@vger.kernel.org. The wireless developers will see it there. I assume that the "previous" kernel was 3.2, but be specific. Larry -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
* Larry Finger
Neither of the above work-a-rounds do the job. Still dropping wireless connection with the "gail calibration timeout" notice.
Help ????
If the "processor.ignore_ppc=1" option fixes it, it does not seem to be a problem with ath5k, but with something deep in the kernel. I think you should post on the linux-kernel@vger.kernel.org ML.
Just in case it is something wilt ath5k, also post on linux-wireless@vger.kernel.org. The wireless developers will see it there.
"processor.ignore_ppc=1" finally *also* failed. Still have problem.
I assume that the "previous" kernel was 3.2, but be specific.
present kernel is kernel-desktop-3.3.rc6-1.3.x86_64 I fat-fingered /etc/zypp.d/zypp.conf and lost previous and now do not know what previous kernel was and there is nothing to indicate which in /usr/src/ iirc previous kernel was 3.3.rc6-..... as I recall severl versions in a short time period. tks, -- (paka)Patrick Shanahan Plainfield, Indiana, USA HOG # US1244711 http://wahoo.no-ip.org Photo Album: http://wahoo.no-ip.org/gallery2 http://en.opensuse.org openSUSE Community Member Registered Linux User #207535 @ http://linuxcounter.net -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
* Patrick Shanahan
* Patrick Shanahan
[01-01-70 12:34]: uname -a Linux crash2 3.3.0-rc6-1-desktop #1 SMP PREEMPT Tue Mar 6 19:02:14 UTC 2012 (0d8ec4d) x86_64 x86_64 x86_64 GNU/Linux
27: None 00.0: 1070a WLAN [Created at net.124] Unique ID: AYEt.QXn1l67RSa1 Parent ID: y9sn.6YBoXuWNRV3 SysFS ID: /class/net/wlan0 SysFS Device Link: /devices/pci0000:00/0000:00:14.4/0000:02:08.0 Hardware Class: network interface Model: "WLAN network interface" Driver: "ath5k" Driver Modules: "ath5k" Device File: wlan0 HW Address: 00:0f:b5:f9:f9:c3 Link detected: yes Config Status: cfg=new, avail=yes, need=no, active=unknown Attached to: #19 (Ethernet controller)
many log entries: ath5k phy0: gain calibration timeout (2462MHz) and loose wireless connection. Only can reconnect by warm/cold reboot.
found on google two work-a-rounds: add to boot: processor.ignore_ppc=1 add to wireless card parameters/options: fastchanswitch
the first has worked for me for about 12 hrs. did not have problems with previous kernel
Have not filed bug-report ????
Neither of the above work-a-rounds do the job. Still dropping wireless connection with the "gail calibration timeout" notice.
And now internet is dropping within an hour and rebooting appears only solution. Appears I will have to stop testing 12.2 :^( -- (paka)Patrick Shanahan Plainfield, Indiana, USA HOG # US1244711 http://wahoo.no-ip.org Photo Album: http://wahoo.no-ip.org/gallery2 http://en.opensuse.org openSUSE Community Member Registered Linux User #207535 @ http://linuxcounter.net -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
On 03/20/2012 09:28 AM, Patrick Shanahan wrote:
* Patrick Shanahan
[03-19-12 15:37]: * Patrick Shanahan
[01-01-70 12:34]: uname -a Linux crash2 3.3.0-rc6-1-desktop #1 SMP PREEMPT Tue Mar 6 19:02:14 UTC 2012 (0d8ec4d) x86_64 x86_64 x86_64 GNU/Linux
27: None 00.0: 1070a WLAN [Created at net.124] Unique ID: AYEt.QXn1l67RSa1 Parent ID: y9sn.6YBoXuWNRV3 SysFS ID: /class/net/wlan0 SysFS Device Link: /devices/pci0000:00/0000:00:14.4/0000:02:08.0 Hardware Class: network interface Model: "WLAN network interface" Driver: "ath5k" Driver Modules: "ath5k" Device File: wlan0 HW Address: 00:0f:b5:f9:f9:c3 Link detected: yes Config Status: cfg=new, avail=yes, need=no, active=unknown Attached to: #19 (Ethernet controller)
many log entries: ath5k phy0: gain calibration timeout (2462MHz) and loose wireless connection. Only can reconnect by warm/cold reboot.
found on google two work-a-rounds: add to boot: processor.ignore_ppc=1 add to wireless card parameters/options: fastchanswitch
the first has worked for me for about 12 hrs. did not have problems with previous kernel
Have not filed bug-report ????
Neither of the above work-a-rounds do the job. Still dropping wireless connection with the "gail calibration timeout" notice.
And now internet is dropping within an hour and rebooting appears only solution. Appears I will have to stop testing 12.2 :^(
Do you mean the unloading/reloading the ath5k module does not restore wireless? Larry -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
* Larry Finger
Do you mean the unloading/reloading the ath5k module does not restore wireless?
modprobe -r ath5k will drop wireless connection modprobe ath5k; ifup wlan0 will restore wireless connection cannot say if this works when "ath5k phy0: gain calibration timeout" is encountered, but will advise. then next would be a script to detect the timeout and initiate the unloading/loading and initialization of the wireless connection. I can handle the latter, how do I detect a *new* timeout? tks, -- (paka)Patrick Shanahan Plainfield, Indiana, USA HOG # US1244711 http://wahoo.no-ip.org Photo Album: http://wahoo.no-ip.org/gallery2 http://en.opensuse.org openSUSE Community Member Registered Linux User #207535 @ http://linuxcounter.net -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
* Patrick Shanahan
* Larry Finger
[03-20-12 10:47]: Do you mean the unloading/reloading the ath5k module does not restore wireless?
modprobe -r ath5k will drop wireless connection modprobe ath5k; ifup wlan0 will restore wireless connection
cannot say if this works when "ath5k phy0: gain calibration timeout" is encountered, but will advise.
then next would be a script to detect the timeout and initiate the unloading/loading and initialization of the wireless connection. I can handle the latter, how do I detect a *new* timeout?
OK, have written a script and added it to /etc/init.d/after.local //////////////////////////////////////////////////////////////////// #!/bin/bash TIM=$(grep ath5k.*timeout /var/log/messages|tail -n 1) sleep 600 TIM1=$(grep ath5k.*timeout /var/log/messages|tail -n 1) while [ "$TIM" = "$TIM1" ]; do sleep 600 TIM1=$(grep ath5k.*timeout /var/log/messages|tail -n 1) done modprobe -r ath5k sleep 30 modeprobe ath5k ifup wlan0 //////////////////////////////////////////////////////////////////// is this sufficient ???? tks, -- (paka)Patrick Shanahan Plainfield, Indiana, USA HOG # US1244711 http://wahoo.no-ip.org Photo Album: http://wahoo.no-ip.org/gallery2 http://en.opensuse.org openSUSE Community Member Registered Linux User #207535 @ http://linuxcounter.net -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
On 03/20/2012 04:49 PM, Patrick Shanahan wrote:
* Patrick Shanahan
[03-20-12 12:03]: * Larry Finger
[03-20-12 10:47]: Do you mean the unloading/reloading the ath5k module does not restore wireless?
modprobe -r ath5k will drop wireless connection modprobe ath5k; ifup wlan0 will restore wireless connection
cannot say if this works when "ath5k phy0: gain calibration timeout" is encountered, but will advise.
then next would be a script to detect the timeout and initiate the unloading/loading and initialization of the wireless connection. I can handle the latter, how do I detect a *new* timeout?
OK, have written a script and added it to /etc/init.d/after.local
//////////////////////////////////////////////////////////////////// #!/bin/bash
TIM=$(grep ath5k.*timeout /var/log/messages|tail -n 1)
sleep 600 TIM1=$(grep ath5k.*timeout /var/log/messages|tail -n 1)
while [ "$TIM" = "$TIM1" ]; do sleep 600 TIM1=$(grep ath5k.*timeout /var/log/messages|tail -n 1) done
modprobe -r ath5k sleep 30
modeprobe ath5k ifup wlan0 ////////////////////////////////////////////////////////////////////
is this sufficient ????
That should be if you start the whole thing with "while [ 1 ] ; do" and end with "done". In addition, it should not take 30 sec between the unload and the following reload. In addition, you have a typo in the load statement. I was merely asking "Is it necessary to reboot, or does the unload/load work too?" The reason is that some conditions in the device are not reset with the module unload/reload and require a reboot. I wanted to know which condition applied here. Larry -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
* Larry Finger
That should be if you start the whole thing with "while [ 1 ] ; do" and end with "done". In addition, it should not take 30 sec between the unload and the following reload. In addition, you have a typo in the load statement.
tks, will make adjustments and try...
I was merely asking "Is it necessary to reboot, or does the unload/load work too?" The reason is that some conditions in the device are not reset with the module unload/reload and require a reboot. I wanted to know which condition applied here.
appears reboot *is* necessary. Wireless connection dropped again while I was away and on return I was unable to restore connection with modprobe -r ath5k ath modprobe ath5k ifup wlan0 system appears connected but cannot ping out and cannot acces the box remotely :^( tks, -- (paka)Patrick Shanahan Plainfield, Indiana, USA HOG # US1244711 http://wahoo.no-ip.org Photo Album: http://wahoo.no-ip.org/gallery2 http://en.opensuse.org openSUSE Community Member Registered Linux User #207535 @ http://linuxcounter.net -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
* Patrick Shanahan
* Larry Finger
[03-20-12 18:10]: That should be if you start the whole thing with "while [ 1 ] ; do" and end with "done". In addition, it should not take 30 sec between the unload and the following reload. In addition, you have a typo in the load statement.
tks, will make adjustments and try...
I was merely asking "Is it necessary to reboot, or does the unload/load work too?" The reason is that some conditions in the device are not reset with the module unload/reload and require a reboot. I wanted to know which condition applied here.
appears reboot *is* necessary. Wireless connection dropped again while I was away and on return I was unable to restore connection with modprobe -r ath5k ath modprobe ath5k ifup wlan0
system appears connected but cannot ping out and cannot acces the box remotely :^(
Do I also need to kill dhcp processes? -- (paka)Patrick Shanahan Plainfield, Indiana, USA HOG # US1244711 http://wahoo.no-ip.org Photo Album: http://wahoo.no-ip.org/gallery2 http://en.opensuse.org openSUSE Community Member Registered Linux User #207535 @ http://linuxcounter.net -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
On 03/20/2012 09:48 PM, Patrick Shanahan wrote:
* Patrick Shanahan
[03-20-12 21:12]: tks, will make adjustments and try...
appears reboot *is* necessary. Wireless connection dropped again while I was away and on return I was unable to restore connection with modprobe -r ath5k ath modprobe ath5k ifup wlan0
system appears connected but cannot ping out and cannot acces the box remotely :^(
Do I also need to kill dhcp processes?
You should not need to; however, you might try 'ifdown' before the 'ifup'. Another thing you might consider is bisecting this problem. It seems not to affect many systems, and it might not be fixed unless you identify the commit that caused the problem. Larry -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
* Larry Finger
Do I also need to kill dhcp processes?
You should not need to; however, you might try 'ifdown' before the 'ifup'.
sadly, this does *not* work.
Another thing you might consider is bisecting this problem. It seems not to affect many systems, and it might not be fixed unless you identify the commit that caused the problem.
after a week of problems with wireless connection, I'm afraid I do not have *any* idea which commit caused the problem and I am *just* assuming that it is kernel related but the kernel did change about that time. Updates in factory are volumous and frequent and leave no documented memory trail that I know of. Can you suggest a way to find the "commit that caused the problem" or some trouble-shooting method to provide intelligent information. This is a *test* machine and I *am* willing to *test* :^) tks for your indulgence. I am open to reinstalliing via a suggested path. Present install is/was an upgrade from 12.1 standare x86_64 to factory: 16:31 crash2: ~ # zypper lr # | Alias | Name | Enabled | Refresh ---+--------------------+-------------------------------------------------------------------+---------+-------- 1 | KDE_Distro_Factory | KDE 4.8.x Packages (openSUSE_Factory) | Yes | Yes 2 | KDE_Extra | Additional packages maintained by the KDE team (openSUSE_Factory) | Yes | Yes 3 | factory-non-oss | factory-non-oss | Yes | Yes 4 | factory-oss | factory-oss | Yes | Yes 5 | home_tittiatcoke | tittiatcoke's Home Project (openSUSE_Factory) | Yes | Yes 6 | libdvdcss2 | libdvdcss2 | Yes | Yes 7 | packman | Packman repository (Factory) | Yes | Yes 8 | repo-debug | openSUSE-12.1-Debug | No | No 9 | repo-debug-update | openSUSE-12.1-Update-Debug | No | No 10 | repo-non-oss | openSUSE-12.1-Non-Oss | No | No 11 | repo-oss | openSUSE-12.1-Oss | No | No 12 | repo-source | openSUSE-12.1-Source | No | No 13 | repo-update | openSUSE-12.1-Update | No | No 14 | utilities | all the small tools for the shell (openSUSE_Factory) | Yes | Yes -- (paka)Patrick Shanahan Plainfield, Indiana, USA HOG # US1244711 http://wahoo.no-ip.org Photo Album: http://wahoo.no-ip.org/gallery2 http://en.opensuse.org openSUSE Community Member Registered Linux User #207535 @ http://linuxcounter.net -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
On 03/21/2012 04:56 PM, Patrick Shanahan wrote:
* Larry Finger
[03-20-12 23:05]: Do I also need to kill dhcp processes?
You should not need to; however, you might try 'ifdown' before the 'ifup'.
sadly, this does *not* work.
Another thing you might consider is bisecting this problem. It seems not to affect many systems, and it might not be fixed unless you identify the commit that caused the problem.
after a week of problems with wireless connection, I'm afraid I do not have *any* idea which commit caused the problem and I am *just* assuming that it is kernel related but the kernel did change about that time.
Updates in factory are volumous and frequent and leave no documented memory trail that I know of. Can you suggest a way to find the "commit that caused the problem" or some trouble-shooting method to provide intelligent information. This is a *test* machine and I *am* willing to *test* :^)
tks for your indulgence. I am open to reinstalliing via a suggested path. Present install is/was an upgrade from 12.1 standare x86_64 to factory:
I would not recommend doing this kind of debugging with the Factory kernels. Instead you should download the mainline git kernel by using sudo zypper in git git clone git://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git This will get you the full source for the kernel. Next you cd to linux, and do the following: git bisect start git bisect bad v3.3 git bisect good v3.2 This will generate a new trial kernel roughly half way between the two versions. You build, install and test that one. Once you know if it is good or bad, git bisect xxxx (either good/bad). If you know that v3.3-rc6 was bad, then use that instead of v3.3. Larry -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
* Larry Finger
I would not recommend doing this kind of debugging with the Factory kernels. Instead you should download the mainline git kernel by using
sudo zypper in git git clone git://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git
This will get you the full source for the kernel. Next you cd to linux, and do the following:
git bisect start git bisect bad v3.3 git bisect good v3.2
This will generate a new trial kernel roughly half way between the two versions. You build, install and test that one. Once you know if it is good or bad,
git bisect xxxx (either good/bad).
If you know that v3.3-rc6 was bad, then use that instead of v3.3.
I will dig into this when I have sit-down-time. You may need to answer *many* questions as I suspect you have exceeded my working-knowledge-level *:^) tks, -- (paka)Patrick Shanahan Plainfield, Indiana, USA HOG # US1244711 http://wahoo.no-ip.org Photo Album: http://wahoo.no-ip.org/gallery2 http://en.opensuse.org openSUSE Community Member Registered Linux User #207535 @ http://linuxcounter.net -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
participants (2)
-
Larry Finger
-
Patrick Shanahan