On 08/29/2010 10:55 AM, Daniel B. Thurman wrote:
I have a gateway M-6750 laptop which sports a
Marvell TopDog wifi chip. I have obtained the
latest drivers for this chip, and used ndiswrapper
to hook itself to the XP/Vista topdog driver, and
it at first appears to be working... although there
are crash data appearing in the log files but other
than that, it continues on working.
I can use Firefox, and Pidgin and the network
appears to work fine.
But the minute I start Thunderbird, it appears the
wifi/network performance is severely degradated
as TB is trys to sync the IMAP server data with
local data storage. It knocked out pidgin. gkrellm
displayed Xorg was hitting CPU hard, Nautilus
froze for a time, gnome terminal blocked text entry,
basically, everything appeared eratic. But given
several minutes of time for TB to settle down and
to finish its tasks, the system seemed to return to
some sense of normalcy - just slightly better.
While this was going on, I thought I'd ping a local
server to get some sense of what is going on with
the network since I cannot think of a better diagnostic
test, and the summary is shown below but watching
each ping line, there appeared many times, complete
line display stoppage running several seconds before
the next display appears.
42 packets transmitted, 41 received, 2% packet loss, \
time 41873ms
rtt min/avg/max/mdev = 0.754/359.603/2005.013/581.445 \
ms, pipe 3
Line by line data, the highest delay was 4000ms at that
short time of testing.
Is there anything I can do to see why there is network
degradation and if it is related to ndiswrapper or not?
I tried wired LAN and there is no network degradation?
This is addtional data:
Wifi pings:
64 bytes from 10.1.0.100: icmp_seq=806 ttl=128 time=0.824 ms
64 bytes from 10.1.0.100: icmp_seq=807 ttl=128 time=2127 ms
64 bytes from 10.1.0.100: icmp_seq=808 ttl=128 time=1127 ms
64 bytes from 10.1.0.100: icmp_seq=809 ttl=128 time=127 ms
64 bytes from 10.1.0.100: icmp_seq=810 ttl=128 time=11999 ms
64 bytes from 10.1.0.100: icmp_seq=811 ttl=128 time=11000 ms
64 bytes from 10.1.0.100: icmp_seq=812 ttl=128 time=10000 ms
64 bytes from 10.1.0.100: icmp_seq=813 ttl=128 time=9000 ms
64 bytes from 10.1.0.100: icmp_seq=814 ttl=128 time=8000 ms
64 bytes from 10.1.0.100: icmp_seq=815 ttl=128 time=7000 ms
64 bytes from 10.1.0.100: icmp_seq=816 ttl=128 time=6000 ms
64 bytes from 10.1.0.100: icmp_seq=817 ttl=128 time=5000 ms
64 bytes from 10.1.0.100: icmp_seq=818 ttl=128 time=4000 ms
64 bytes from 10.1.0.100: icmp_seq=819 ttl=128 time=3000 ms
64 bytes from 10.1.0.100: icmp_seq=820 ttl=128 time=2000 ms
64 bytes from 10.1.0.100: icmp_seq=822 ttl=128 time=2.56 ms
64 bytes from 10.1.0.100: icmp_seq=823 ttl=128 time=2002 ms
64 bytes from 10.1.0.100: icmp_seq=824 ttl=128 time=1002 ms
64 bytes from 10.1.0.100: icmp_seq=825 ttl=128 time=1.46 ms
Wired pings:
64 bytes from 10.1.0.100: icmp_seq=1025 ttl=128 time=0.239 ms
64 bytes from 10.1.0.100: icmp_seq=1026 ttl=128 time=0.233 ms
64 bytes from 10.1.0.100: icmp_seq=1027 ttl=128 time=0.252 ms
64 bytes from 10.1.0.100: icmp_seq=1028 ttl=128 time=0.214 ms
64 bytes from 10.1.0.100: icmp_seq=1029 ttl=128 time=0.233 ms
64 bytes from 10.1.0.100: icmp_seq=1030 ttl=128 time=0.248 ms
64 bytes from 10.1.0.100: icmp_seq=1031 ttl=128 time=0.231 ms
64 bytes from 10.1.0.100: icmp_seq=1032 ttl=128 time=0.200 ms
64 bytes from 10.1.0.100: icmp_seq=1033 ttl=128 time=0.224 ms
64 bytes from 10.1.0.100: icmp_seq=1034 ttl=128 time=0.240 ms
64 bytes from 10.1.0.100: icmp_seq=1035 ttl=128 time=0.202 ms
64 bytes from 10.1.0.100: icmp_seq=1036 ttl=128 time=0.241 ms
64 bytes from 10.1.0.100: icmp_seq=1037 ttl=128 time=0.222 ms
64 bytes from 10.1.0.100: icmp_seq=1038 ttl=128 time=0.235 ms
64 bytes from 10.1.0.100: icmp_seq=1039 ttl=128 time=0.249 ms
64 bytes from 10.1.0.100: icmp_seq=1040 ttl=128 time=0.230 ms
64 bytes from 10.1.0.100: icmp_seq=1041 ttl=128 time=0.239 ms
64 bytes from 10.1.0.100: icmp_seq=1042 ttl=128 time=0.246 ms
64 bytes from 10.1.0.100: icmp_seq=1043 ttl=128 time=0.251 ms
So it appears quite clear that when there is heavy load,
the Ndiswrapper/TopDog is peglegging along sputtering
blood on the trails while the wired connections have no
problems...
Well, if anyone has any idea how I can get the topdog wifi
chip working at better performance - please let me know?
Thanks!