Bug 40462 - RTL8187L slow rates
Summary: RTL8187L slow rates
Status: RESOLVED CODE_FIX
Alias: None
Product: Networking
Classification: Unclassified
Component: Wireless (show other bugs)
Hardware: All Linux
: P1 normal
Assignee: networking_wireless@kernel-bugs.osdl.org
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2011-08-03 11:14 UTC by ch.strobl
Modified: 2012-01-13 19:15 UTC (History)
2 users (show)

See Also:
Kernel Version:
Subsystem:
Regression: No
Bisected commit-id:


Attachments

Description ch.strobl 2011-08-03 11:14:05 UTC
Description: very slow data rate with my RTL8187L device.
just got about 1,8 MB/s in my network which isn't normal. used to be normal a few month ago.


Additional info:
2.6.39-ARCH #1
Bus 001 Device 002: ID 0bda:8187 Realtek Semiconductor Corp. RTL8187 Wireless Adapter
It's an Alfa AWUS036H

I use the kernel build-in driver.


It's my first bug report, so i hope it's enough
Comment 1 Larry Finger 2011-08-03 16:09:31 UTC
This driver has not been changed for some time.

Please post the output of iwconfig. What is the distance to the AP?
Comment 2 ch.strobl 2011-08-04 06:10:08 UTC
iwconfig:

wlan0     IEEE 802.11bg  ESSID:"myWLAN2"  
          Mode:Managed  Frequency:2.412 GHz  Access Point: 00:16:38:40:ED:C1   
          Bit Rate=54 Mb/s   Tx-Power=20 dBm   
          Retry  long limit:7   RTS thr:off   Fragment thr:off
          Power Management:off
          Link Quality=58/70  Signal level=-52 dBm  
          Rx invalid nwid:0  Rx invalid crypt:0  Rx invalid frag:0
          Tx excessive retries:33  Invalid misc:1349   Missed beacon:0


The distance to AP is about 2m trough concrete ceiling. It worked great, but since some months it slowed down.
Comment 3 Larry Finger 2011-08-04 17:37:30 UTC
There are some reports of the rate-control algorithm being too aggressive on the RTL8187L. Most of those have been for signal strengths below -70 dBm, not -52 as you see, but just in case, test the throughput after running (as root) a command of the form:

iwconfig wlan0 rate XX

where XX is 11M, 18M, 24M, 36M, or 48M.

How do you measure speed?
Comment 4 ch.strobl 2011-08-05 13:57:21 UTC
I'm measuring with via rsync --progress with a nfs folder on my Seagate Dockstar (gigabit-ethernet). It starts with about 4MB/s rises up 20MB/s to and then slows down to 1,5-1,8 MB/s within a few seconds.

Tried every rate, starts high, slows down very fast.


The most wired thing is, my internet connection is about 6 Mbit/s, on my Dockstar i got this 6Mbit/s. But with the RTL8187L just get about 1Mbit/s.
Comment 5 ch.strobl 2011-08-06 18:30:02 UTC
i also measured with GNU netcat.

"nc -ulp 5000 > /dev/null" on my Dockstar

and

"pv < /dev/zero | nc -u XX.XX.XX.XX 5000" on my 

write(net): Das Netzwerk ist nicht erreichbar              <=>                 ]
6,62GB 0:00:33 [ 202MB/s] [                                 <=>                ]
pv < /dev/zero  0,85s user 17,12s system 53% cpu 33,658 total
nc -u XX.XX.XX.XX 5000  1,87s user 25,63s system 81% cpu 33,658 total



write(net): Das Netzwerk ist nicht erreichbar
engl. Network not reachable.


It breaks up after a few seconds, and my network connection is down.

wlan0     IEEE 802.11bg  ESSID:off/any  
          Mode:Managed  Frequency:2.412 GHz  Access Point: Not-Associated   
          Tx-Power=20 dBm   
          Retry  long limit:7   RTS thr:off   Fragment thr:off
          Power Management:off



I can get the connection up again, but if i repeat the nc thing, it breaks again.
Comment 6 Larry Finger 2011-08-06 18:39:14 UTC
What happens if you do not use the "u" option on the nc call? I suspect a problem with queue overrun while using UDP mode that is not likely to happen with TCP transmissions.
Comment 7 ch.strobl 2011-08-06 20:59:46 UTC
10,9MB 0:00:53 [ 320kB/s]
Comment 8 ch.strobl 2011-08-10 09:43:48 UTC
Kernel Update to 3.0 didn't solve the problem
Comment 9 John W. Linville 2012-01-13 19:15:08 UTC
Does this problem persist with 3.2 kernels?

Note You need to log in before you can comment on or make changes to this bug.