* carl9170 problem @ 2010-09-25 0:34 Dale Turner 2010-09-25 22:57 ` Christian Lamparter 0 siblings, 1 reply; 4+ messages in thread From: Dale Turner @ 2010-09-25 0:34 UTC (permalink / raw) To: linux-wireless Hello everyone. Thanks for releasing the carl9170 driver in the compat-wireless package. I have been able to compile the driver, so that is great. The connection seems to be a bit inconsistent. I get the following quite often in /var/log/messages : kernel: ieee80211 phy9: channel change: 2472 -> 2462 failed (2) and/or kernel: ieee80211 phy9: invalid plcp cck rate (0) and/or kernel: ieee80211 phy9: FW: MAC RESET When that happens (which can be as often as every few minutes) my reported connection speed in network manager drops to 6 or 1 Mb/s. The network speed is noticeably reduced (for instance download speeds drastically reduce, online videos stutter, or music streaming from a different computer pauses, etc). My previous adapter TP-Link TL-WN721N (ath9k_htc driver) did not behave like this. Any thoughts/solutions ?? Again, many thanks for providing the driver. I cannot program (I wish I could to help) but I can report my experience. Dale ^ permalink raw reply [flat|nested] 4+ messages in thread
* Re: carl9170 problem 2010-09-25 0:34 carl9170 problem Dale Turner @ 2010-09-25 22:57 ` Christian Lamparter [not found] ` <SNT113-W459FFC08EF0C4936BEDD5CF1660@phx.gbl> 0 siblings, 1 reply; 4+ messages in thread From: Christian Lamparter @ 2010-09-25 22:57 UTC (permalink / raw) To: Dale Turner; +Cc: linux-wireless On Saturday 25 September 2010 02:34:26 Dale Turner wrote: > > Hello everyone. > > Thanks for releasing the carl9170 driver in the > compat-wireless package. I recommend using wireless-testing.git over compat-wireless + old kernel. > I have been able to compile the driver, so that is great. > The connection seems to be a bit inconsistent. > I get the following quite often in /var/log/messages : > kernel: ieee80211 phy9: channel change: 2472 -> 2462 failed (2) That's a noise floor calibration error. Check if your compat-wireless source includes "carl9170: reinit phy after HT settings have changed". Also check if there is another device active on the same (primary/ secondary) channel like a microwave oven, DECT phone or another wifi network, or wifi network peer. > kernel: ieee80211 phy9: invalid plcp cck rate (0) This is a known bug, the driver received a invalid stream frame from the hardware. This only happens a lot when the AP ignores the device's ampdu spacing. (Do you have an older draft-n AP, if so what chip?) > kernel: ieee80211 phy9: FW: MAC RESET The device failed to send pending data for more than 150ms(HT40) - 300ms, either because the medium was busy, or the AP failed to response at all mostly because it has crashed first. (check if your AP's TSF is no where near your AP's uptime) > When > that happens (which can be as often as every few minutes) my reported > connection speed in network manager drops to 6 or 1 Mb/s. The network > speed is noticeably reduced (for instance download speeds drastically > reduce, online videos stutter, or music streaming from a different > computer pauses, etc). Network manager has a feature called "background scanning". The advantage is that you'll always have a up-to-date site survey. The disadvantage is that it needs to perform a periodically scan (as in every few minutes) which on some occasions result in a channel change failed hiccup. > > My previous adapter TP-Link TL-WN721N (ath9k_htc driver) did not behave like this. Why did you go back to the previous generation? ath9k_htc (for AR9271) is in fact the replacement for AR9170. Regards, Chr ^ permalink raw reply [flat|nested] 4+ messages in thread
[parent not found: <SNT113-W459FFC08EF0C4936BEDD5CF1660@phx.gbl>]
* Re: carl9170 problem [not found] ` <SNT113-W459FFC08EF0C4936BEDD5CF1660@phx.gbl> @ 2010-09-28 10:11 ` Dale Turner 2010-09-28 11:35 ` Christian Lamparter 1 sibling, 0 replies; 4+ messages in thread From: Dale Turner @ 2010-09-28 10:11 UTC (permalink / raw) To: linux-wireless >>> Hello everyone. >>> >>> Thanks for releasing the carl9170 driver in the >>> compat-wireless package. >> I recommend using wireless-testing.git over >> compat-wireless + old kernel. >> >>> I have been able to compile the driver, so that is great. >>> The connection seems to be a bit inconsistent. >>> I get the following quite often in /var/log/messages : >> >>> kernel: ieee80211 phy9: channel change: 2472 -> 2462 failed (2) >> That's a noise floor calibration error. Check if your compat-wireless >> source includes "carl9170: reinit phy after HT settings have changed". >> Also check if there is another device active on the same (primary/ >> secondary) channel like a microwave oven, DECT phone or another wifi >> network, or wifi network peer. >> >>> kernel: ieee80211 phy9: invalid plcp cck rate (0) >> This is a known bug, the driver received a invalid stream frame >> from the hardware. This only happens a lot when the AP ignores >> the device's ampdu spacing. >> (Do you have an older draft-n AP, if so what chip?) >> >>> kernel: ieee80211 phy9: FW: MAC RESET >> The device failed to send pending data for more than 150ms(HT40) - 300ms, >> either because the medium was busy, or the AP failed to response at all >> mostly because it has crashed first. >> (check if your AP's TSF is no where near your AP's uptime) >> >>> When >>> that happens (which can be as often as every few minutes) my reported >>> connection speed in network manager drops to 6 or 1 Mb/s. The network >>> speed is noticeably reduced (for instance download speeds drastically >>> reduce, online videos stutter, or music streaming from a different >>> computer pauses, etc). >> Network manager has a feature called "background scanning". >> The advantage is that you'll always have a up-to-date site survey. >> The disadvantage is that it needs to perform a periodically scan >> (as in every few minutes) which on some occasions result in a >> channel change failed hiccup. >> >>> >>> My previous adapter TP-Link TL-WN721N (ath9k_htc driver) did not behave like this. >> Why did you go back to the previous generation? >> ath9k_htc (for AR9271) is in fact the replacement for AR9170. >> >> Regards, >> Chr > Thanks for your response. I chose the TL-WN821N because it seemed to have a 'better' feature set (faster speed, MIMO, etc) plus the carl9170 driver is supposed to capable of AP mode. (I have plans in the future of constructing a wireless router). So I thought I was improving (guess I'm wrong). I don't think there are any nearby networks on my channel. I purposely change it from default ( 6 ) to 11 (actually wide 9). The access point is a Linksys wrt120n (probably not the best, I know). To be honest, I don't know what a TSF is or how to check it. Thanks again, Dale ^ permalink raw reply [flat|nested] 4+ messages in thread
* Re: carl9170 problem [not found] ` <SNT113-W459FFC08EF0C4936BEDD5CF1660@phx.gbl> 2010-09-28 10:11 ` Dale Turner @ 2010-09-28 11:35 ` Christian Lamparter 1 sibling, 0 replies; 4+ messages in thread From: Christian Lamparter @ 2010-09-28 11:35 UTC (permalink / raw) To: Dale Turner; +Cc: linux-wireless On Tuesday 28 September 2010 03:47:00 Dale Turner wrote: > >> kernel: ieee80211 phy9: FW: MAC RESET > > The device failed to send pending data for more than 150ms(HT40) - 300ms, > > either because the medium was busy, or the AP failed to response at all > > mostly because it has crashed first. > > (check if your AP's TSF is no where near your AP's uptime) > > I don't think there are any nearby networks on my channel. > I purposely change it from default ( 6 ) to 11 (actually wide 9). What's wide 9? HT40 with the primary channel on 11 and the extension on 6? > The access point is a Linksys wrt120n (probably not the best, I know). Isn't this a 1-stream align-n (AR9285) platform? If so then there is no advantage of using an AR9170 over the AR9271. > To be honest, I don't know what a TSF is or how to check it. The TSF keeps the timers for all STAs in the "network" synchronized. http://en.wikipedia.org/wiki/Timing_Synchronization_Function_%28TSF%29 It can be easily check by: iw dev wlanX scan e.g.: (sample output) BSS 00:02:32:0d:d1:32 (on wlan13) TSF: 2669831323 usec (0d, 00:44:29) <------- freq: 2437 This will tell you that the "network" has an uptime of 0 days, 0 hours, 44 minutes and 29 seconds. Regards, Chr ^ permalink raw reply [flat|nested] 4+ messages in thread
end of thread, other threads:[~2010-09-28 11:35 UTC | newest]
Thread overview: 4+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2010-09-25 0:34 carl9170 problem Dale Turner
2010-09-25 22:57 ` Christian Lamparter
[not found] ` <SNT113-W459FFC08EF0C4936BEDD5CF1660@phx.gbl>
2010-09-28 10:11 ` Dale Turner
2010-09-28 11:35 ` Christian Lamparter
This is a public inbox, see mirroring instructions for how to clone and mirror all data and code used for this inbox; as well as URLs for NNTP newsgroup(s).