View previous topic :: View next topic |
Author |
Message |
The Doctor Moderator
Joined: 27 Jul 2010 Posts: 2678
|
Posted: Wed May 30, 2012 12:46 am Post subject: Working around 3.x kernel ath5k disconnects |
|
|
I apologize if this is in the wrong place. Please move if appropriate.
It seems to be a widespread problem with the 3.x kernels that the ath5k module periodically disconnects and leaves messages about a gain calibration timeout, for me this seems to have been associated with high internet usage. There is a workaround for those interested:
Add pcie_aspm=force to the kernel command line.
I tested this by downloading some data unnecessary data (Ubuntu iso ~700 mb) and so far this workaround has been stable.
EDIT: After a few weeks, I can say this did not entirely solve the problem, but it does make the failures much less frequent. I have only had it happen once or twice since I did this. _________________ First things first, but not necessarily in that order.
Apologies if I take a while to respond. I'm currently working on the dematerialization circuit for my blue box.
Last edited by The Doctor on Mon Jun 11, 2012 6:53 pm; edited 1 time in total |
|
Back to top |
|
|
khayyam Watchman
Joined: 07 Jun 2012 Posts: 6227 Location: Room 101
|
Posted: Mon Jun 11, 2012 2:09 pm Post subject: |
|
|
penguin ...
I've been using 3.x kernels for a while and haven't had this particular issue.
Code: | % uname -r
3.2.12-gentoo
% lspci | awk '/Atheros/'
02:00.0 Ethernet controller: Atheros Communications Inc. AR242x / AR542x Wireless Network Adapter (PCI-Express) (rev 01)
% dmesg | awk '/Atheros/'
ath5k phy0: Atheros AR5414 chip found (MAC: 0xa3, PHY: 0x61) |
This may be an issue with a specific chipset, but anyhow, I'm sure those that have this problem will recognise the symptoms so I'm being somwhat redundant.
best ... khay |
|
Back to top |
|
|
The Doctor Moderator
Joined: 27 Jul 2010 Posts: 2678
|
Posted: Mon Jun 11, 2012 6:51 pm Post subject: |
|
|
Thats interesting.
Here is what my lspci spits out:
Code: | 01:03.0 Ethernet controller: Atheros Communications Inc. AR5413/AR5414 Wireless Network Adapter [AR5006X(S) 802.11abg] (rev 01) |
So it looks like we don't have quite the same card. _________________ First things first, but not necessarily in that order.
Apologies if I take a while to respond. I'm currently working on the dematerialization circuit for my blue box. |
|
Back to top |
|
|
khayyam Watchman
Joined: 07 Jun 2012 Posts: 6227 Location: Room 101
|
Posted: Mon Jun 11, 2012 8:13 pm Post subject: |
|
|
Penguin ...
Yes, your seems to be a more recent card ... they are both 'abg' however.
Code: | % awk '{RS="\n\n"}/Atheros/{print}' <(lspci -v)
02:00.0 Ethernet controller: Atheros Communications Inc. AR242x / AR542x Wireless Network Adapter (PCI-Express) (rev 01)
Subsystem: Apple Computer Inc. AR5BXB6 802.11abg Wireless Mini PCIe Card
Flags: bus master, fast devsel, latency 0, IRQ 17
Memory at 90100000 (64-bit, non-prefetchable) [size=64K]
Capabilities: [40] Power Management version 2
Capabilities: [50] MSI: Enable- Count=1/1 Maskable- 64bit-
Capabilities: [60] Express Legacy Endpoint, MSI 00
Capabilities: [90] MSI-X: Enable- Count=1 Masked-
Capabilities: [100] Advanced Error Reporting
Capabilities: [140] Virtual Channel
Kernel driver in use: ath5k
Kernel modules: ath5k |
|
|
Back to top |
|
|
Steve_Mac_G n00b
Joined: 13 Dec 2005 Posts: 19 Location: MiddleEarth
|
Posted: Tue Jul 17, 2012 7:34 am Post subject: Re: Working around 3.x kernel ath5k disconnects |
|
|
Hi Guys
I have the same Problem with my ath5k (Kernel 3.2.12)
This is my Device:
Code: | arwen ~ # lspci -v -s 02:00.0
02:00.0 Ethernet controller: Atheros Communications Inc. AR242x / AR542x Wireless Network Adapter (PCI-Express) (rev 01)
Subsystem: Askey Computer Corp. Device 7131
Flags: bus master, fast devsel, latency 0, IRQ 16
Memory at f0100000 (64-bit, non-prefetchable) [size=64K]
Capabilities: [40] Power Management version 2
Capabilities: [50] MSI: Enable- Count=1/1 Maskable- 64bit-
Capabilities: [60] Express Legacy Endpoint, MSI 00
Capabilities: [90] MSI-X: Enable- Count=1 Masked-
Capabilities: [100] Advanced Error Reporting
Capabilities: [140] Virtual Channel
Kernel driver in use: ath5k
|
I've searched long for a Solution but didn't found any working .... Till now!
penguin swordmaster wrote: | ...
Add pcie_aspm=force to the kernel command line.
... |
works for me to , the connection to my Router is now up since over a day instead of maximum 5 min.
Thanks for the Post Penguin Swordmaster :thumbsup:
Peace
Steve |
|
Back to top |
|
|
khayyam Watchman
Joined: 07 Jun 2012 Posts: 6227 Location: Room 101
|
Posted: Tue Jul 17, 2012 3:00 pm Post subject: |
|
|
all ...
oddly I have the exact same card as Steve_Mac_G and have never experienced this issue, including with 3.2.12-gentoo which I used for some time.
Code: | 02:00.0 Ethernet controller: Atheros Communications Inc. AR242x / AR542x Wireless Network Adapter (PCI-Express) (rev 01)
Subsystem: Apple Inc. AR5BXB6 802.11abg Wireless Mini PCIe Card
Flags: bus master, fast devsel, latency 0, IRQ 17
Memory at 90100000 (64-bit, non-prefetchable) [size=64K]
Capabilities: [40] Power Management version 2
Capabilities: [50] MSI: Enable- Count=1/1 Maskable- 64bit-
Capabilities: [60] Express Legacy Endpoint, MSI 00
Capabilities: [90] MSI-X: Enable- Count=1 Masked-
Capabilities: [100] Advanced Error Reporting
Capabilities: [140] Virtual Channel
Kernel driver in use: ath5k
Kernel modules: ath5k |
best ... khay |
|
Back to top |
|
|
Logicien Veteran
Joined: 16 Sep 2005 Posts: 1555 Location: Montréal
|
Posted: Tue Jul 17, 2012 3:56 pm Post subject: |
|
|
The kernel parameter pcie_aspm=force activate the Active-State Power Management for all PCIe devices even if they do not support Aspm. It do not take in account the devices performances result. This is what the /usr/src/linux/Documentation/kernel-prameters.txt say: Code: | pcie_aspm= [PCIE] Forcibly enable or disable PCIe Active State Power
Management.
off Disable ASPM.
force Enable ASPM even on devices that claim not to support it.
WARNING: Forcing ASPM on may cause system lockups.
|
Devices who do not support PCIe ASPM can make a system lockup. It's the reason for the warning. All devices connected to PCIe have to support ASPM to advoid lokups.
Me I use on a desktop computer the pci_aspm=off by the way of
Code: | echo performance > /sys/module/pcie_aspm/parameters/policy |
The policy can be one of default (BIOS, EFI, Coreboot, etc settings for all devices), powersave and performance. The one in brakets [] is the active one.
Disable pcie_aspm make the devices connected to PCIe have the best performances. You can try pcie_aspm=off for test and compare with the other policys. _________________ Paul |
|
Back to top |
|
|
khayyam Watchman
Joined: 07 Jun 2012 Posts: 6227 Location: Room 101
|
Posted: Tue Jul 17, 2012 4:39 pm Post subject: |
|
|
paul ... thanks for the info, this may explain why Steve_Mac_G has the issue and me not, pcie_aspm/parameters/policy is currently [default] and so set via EFI ...
best ... khay |
|
Back to top |
|
|
Logicien Veteran
Joined: 16 Sep 2005 Posts: 1555 Location: Montréal
|
Posted: Tue Jul 17, 2012 6:02 pm Post subject: |
|
|
Hi everybody,
I check. I pass pcie_aspm=off at the kernel command line and /sys/module/pcie_aspm/parameters/policy is default. So It look that the pcie_aspm kernel parameter do not change the policy in sysfs. So, to change the aspm policy I continue to work with the /sys directory. _________________ Paul |
|
Back to top |
|
|
|