Slow wifi connection AR9287 with ath9k 3.x kernel
Hello,
I have submitted bug on Kernel bug tracker and provide info, but it seems that no one there is interested in this topic. I discovered that since 2.6.38 kernel version there is bug in ath9k driver (using AR9287 chipset). I've posted some additional info at this adress:
https://bugzilla.kernel.org/show_bug.cgi?id=42352
The last working version is 2.6.37.6.
I can provide more data if anyone is interested.
Cheers,
Kuba
Quoted 1 line · kuba
Hello,
hi,
Quoted 6 lines
I have submitted bug on Kernel bug tracker and provide info, but it seems that no one there is interested in this topic. I discovered that since 2.6.38 kernel version there is bug in ath9k driver (using AR9287 chipset). I've posted some additional info at this adress:
https://bugzilla.kernel.org/show_bug.cgi?id=42352
please provide as much info as possible, in which mode you are operating, how noisy is the environment, distance between STA and AP. anything specific etc.
i just tried compat-wireless-2012-02-04, distance between STA and AP at least 5ms, a bit of noisy environment, channel 6 HT40 plus, AES ping RTT is good only. i may be missing something
Quoted 2 lines
The last working version is 2.6.37.6.
if we are very clear about the working one and non-working doing the following things will make our work very easier
lets say 2.6.37.6 is working fine and 2.6.38 is broken for you
we can work with our wireless-testing tree kernel
git clone git://git.kernel.org/pub/scm/linux/kernel/git/linville/wireless-testing.git
then please do git tag
*git log --pretty=oneline v2.6.37-rc6..v2.6.38 | grep AR9287 07422063d43612762d53fac8b6df213c96f4b1f6 ath9k_hw: Fix calibration for AR9287 devicesf7ec8fb4d6f8f3ecb8b11e9e46ece95aa66139cc ath9k_hw: Fix eeprom offset for AR9287 devices (PCI/USB)
*even with 3.3-rc5 we got less commits
6321eb0977b011ac61dfca36e7c69b2c4325b104 ath9k_hw: Fix number of GPIO pins for AR9287/9300 a7be039d347743c289b7280d5de82abf7bbdf1d8 ath9k: Fix eifs/usec timeout for AR9287 v1.3+49c99520f3b15300156830904f9ffcf51cb1160e ath9k_hw: Add dump_eeprom support for AR9287087b6ff682ff62c49c9a36ddfd24fc781367b1d1 ath9k_hw: fix MAC clock rate for AR9287 v1.3+1df85ecec36ad5da3f0165760704310d6c03f65f ath9k: Fix AR9287 calibration 2d05a0c2b4ac614cb5e0eba75d39a37205d129e8 ath9k_hw: Remove unused code in AR9287 eeprom81544026e4cecb85a8b727d5f64cb3c8a8cb64a3 ath9k_hw: Fix throughput drops in HT40 mode for AR9287 chips 6d64ab7f9240e3201fde3fd16ce4227bd795d2ab ath9k_htc: Fix LED pin for AR9287 HTC device07422063d43612762d53fac8b6df213c96f4b1f6 ath9k_hw: Fix calibration for AR9287 devicesf7ec8fb4d6f8f3ecb8b11e9e46ece95aa66139cc ath9k_hw: Fix eeprom offset for AR9287 devices (PCI/USB)
*git log --pretty=oneline v2.6.37-rc6..v2.6.38 | grep ath9k_hw
*git log --pretty=oneline v2.6.37-rc6..v2.6.38 | grep ath9k
Quoted 7 lines
I can provide more data if anyone is interested.
Cheers,
Kuba
--
To unsubscribe from this list: send the line "unsubscribe linux-wireless" in the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html
Signature · 2 lines
thanks, shafi
Quoted 61 lines · Mohammed Shafi
Quoted 1 line · kuba
Hello,
hi,
Quoted 6 lines
I have submitted bug on Kernel bug tracker and provide info, but it seems that no one there is interested in this topic. I discovered that since 2.6.38 kernel version there is bug in ath9k driver (using AR9287 chipset). I've posted some additional info at this adress:
https://bugzilla.kernel.org/show_bug.cgi?id=42352please provide as much info as possible, in which mode you are operating, how noisy is the environment, distance between STA and AP. anything specific etc.
i just tried compat-wireless-2012-02-04, distance between STA and AP at least 5ms, a bit of noisy environment, channel 6 HT40 plus, AES ping RTT is good only. i may be missing somethingQuoted 2 lines
The last working version is 2.6.37.6.
if we are very clear about the working one and non-working doing the following things will make our work very easier
lets say 2.6.37.6 is working fine and 2.6.38 is broken for youwe can work with our wireless-testing tree kernel
git clonegit://git.kernel.org/pub/scm/linux/kernel/git/linville/wireless-testing.gitthen please do git tag
*git log --pretty=oneline v2.6.37-rc6..v2.6.38 | grep AR9287
07422063d43612762d53fac8b6df213c96f4b1f6ath9k_hw: Fix calibration for AR9287 devicesf7ec8fb4d6f8f3ecb8b11e9e46ece95aa66139ccath9k_hw: Fix eeprom offset for AR9287 devices (PCI/USB)*even with 3.3-rc5 we got less commits
6321eb0977b011ac61dfca36e7c69b2c4325b104ath9k_hw: Fix number of GPIO pins for AR9287/9300a7be039d347743c289b7280d5de82abf7bbdf1d8ath9k: Fix eifs/usec timeout for AR9287 v1.3+49c99520f3b15300156830904f9ffcf51cb1160eath9k_hw: Adddump_eepromsupport for AR9287087b6ff682ff62c49c9a36ddfd24fc781367b1d1ath9k_hw: fix MAC clock rate for AR9287 v1.3+1df85ecec36ad5da3f0165760704310d6c03f65fath9k: Fix AR9287 calibration2d05a0c2b4ac614cb5e0eba75d39a37205d129e8ath9k_hw: Remove unused code in AR9287 eeprom81544026e4cecb85a8b727d5f64cb3c8a8cb64a3ath9k_hw: Fix throughput drops in HT40 mode for AR9287 chips6d64ab7f9240e3201fde3fd16ce4227bd795d2abath9k_htc: Fix LED pin for AR9287 HTC device07422063d43612762d53fac8b6df213c96f4b1f6ath9k_hw: Fix calibration for AR9287 devicesf7ec8fb4d6f8f3ecb8b11e9e46ece95aa66139ccath9k_hw: Fix eeprom offset for AR9287 devices (PCI/USB)
Quoted 5 lines
*git log --pretty=oneline v2.6.37-rc6..v2.6.38 | grep
ath9k_hw*git log --pretty=oneline v2.6.37-rc6..v2.6.38 | grep ath9k
please do try with the latest wireless-testing tree if you can get the same problem, also please try to check if this issue occurs with some other AP too.
Quoted 14 lines
Quoted 8 lines
I can provide more data if anyone is interested.
Cheers,
Kuba
--
To unsubscribe from this list: send the line "unsubscribe linux-wireless" in the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html--
thanks,
shafi
Signature · 2 lines
thanks, shafi
W dniu 07.03.2012 06:37, Mohammed Shafi pisze:
Quoted 58 lines
Quoted 56 lines · Mohammed Shafi
Quoted 1 line · kuba
Hello,
hi,
Quoted 5 lines
I have submitted bug on Kernel bug tracker and provide info, but it seems that no one there is interested in this topic. I discovered that since 2.6.38 kernel version there is bug in ath9k driver (using AR9287 chipset). I've posted some additional info at this adress:
https://bugzilla.kernel.org/show_bug.cgi?id=42352please provide as much info as possible, in which mode you are operating, how noisy is the environment, distance between STA and AP. anything specific etc.
i just tried compat-wireless-2012-02-04, distance between STA and AP at least 5ms, a bit of noisy environment, channel 6 HT40 plus, AES ping RTT is good only. i may be missing somethingQuoted 1 line
The last working version is 2.6.37.6.
if we are very clear about the working one and non-working doing the following things will make our work very easier
lets say 2.6.37.6 is working fine and 2.6.38 is broken for youwe can work with our wireless-testing tree kernel
git clonegit://git.kernel.org/pub/scm/linux/kernel/git/linville/wireless-testing.gitthen please do git tag
*git log --pretty=oneline v2.6.37-rc6..v2.6.38 | grep AR9287
07422063d43612762d53fac8b6df213c96f4b1f6ath9k_hw: Fix calibration for AR9287 devicesf7ec8fb4d6f8f3ecb8b11e9e46ece95aa66139ccath9k_hw: Fix eeprom offset for AR9287 devices (PCI/USB)*even with 3.3-rc5 we got less commits
6321eb0977b011ac61dfca36e7c69b2c4325b104ath9k_hw: Fix number of GPIO pins for AR9287/9300a7be039d347743c289b7280d5de82abf7bbdf1d8ath9k: Fix eifs/usec timeout for AR9287 v1.3+49c99520f3b15300156830904f9ffcf51cb1160eath9k_hw: Adddump_eepromsupport for AR9287087b6ff682ff62c49c9a36ddfd24fc781367b1d1ath9k_hw: fix MAC clock rate for AR9287 v1.3+1df85ecec36ad5da3f0165760704310d6c03f65fath9k: Fix AR9287 calibration2d05a0c2b4ac614cb5e0eba75d39a37205d129e8ath9k_hw: Remove unused code in AR9287 eeprom81544026e4cecb85a8b727d5f64cb3c8a8cb64a3ath9k_hw: Fix throughput drops in HT40 mode for AR9287 chips6d64ab7f9240e3201fde3fd16ce4227bd795d2abath9k_htc: Fix LED pin for AR9287 HTC device07422063d43612762d53fac8b6df213c96f4b1f6ath9k_hw: Fix calibration for AR9287 devicesf7ec8fb4d6f8f3ecb8b11e9e46ece95aa66139ccath9k_hw: Fix eeprom offset for AR9287 devices (PCI/USB)
Quoted 23 lines
Quoted 5 lines
*git log --pretty=oneline v2.6.37-rc6..v2.6.38 | grep
ath9k_hw*git log --pretty=oneline v2.6.37-rc6..v2.6.38 | grep ath9k
please do try with the latest wireless-testing tree if you can get the same problem, also please try to check if this issue occurs with some other AP too.
Quoted 12 lines
Quoted 7 lines
I can provide more data if anyone is interested.
Cheers,
Kuba
--
To unsubscribe from this list: send the line "unsubscribe linux-wireless" in the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html--
thanks,
shafi
Hi,
I didn't have access to my notebook for 3 months, so I can't anwser, sorry.
I have tried kernel 3.3.4 and I can still reproduce this error. This issue occurs with all AP, even if my notebook stands close to AP (1-5 m). I've noticed that wireless card works better with 3.3.4 but there are times when connection slows down and looks like this:
--- 192.168.1.1 ping statistics ---
35 packets transmitted, 26 received, 25% packet loss, time 34017ms rtt min/avg/max/mdev = 0.986/6194.909/9659.450/2844.281 ms, pipe 10
It's a ping to gateway, ping to random website looks even worse. Wireless-testing doesn't solve this problem.
lspci -v
03:00.0 Network controller: Atheros Communications Inc. AR9287 Wireless Network Adapter (PCI-Express) (rev 01)
Subsystem: Foxconn International, Inc. Device e034
Flags: bus master, fast devsel, latency 0, IRQ 17
Memory at b5100000 (64-bit, non-prefetchable) [size=64K]
Capabilities: <access denied>
Kernel driver in use: ath9k
I can provide more data, just tell me what should I paste here.
Cheers,
Kuba