You are not logged in.

#1 2009-06-26 15:41:20

speng
Member
Registered: 2009-01-17
Posts: 136

Pacman update - Very strange network problems/lag spikes

So I updated my system a day ago or so, since then I have noticed a very strange behaviour.
If I'm using the network, after about 5 minutes-ish I will get a massive lag-spike and then my ping is doubled.

This is after the lag spike (when it has calmed down):

speng@spengpc ~ $ ping -c 100 www.google.com
PING www.l.google.com (66.102.9.104) 56(84) bytes of data.
64 bytes from lm-in-f104.google.com (66.102.9.104): icmp_seq=1 ttl=243 time=99.8 ms
64 bytes from lm-in-f104.google.com (66.102.9.104): icmp_seq=2 ttl=243 time=100 ms
64 bytes from lm-in-f104.google.com (66.102.9.104): icmp_seq=3 ttl=243 time=100 ms
64 bytes from lm-in-f104.google.com (66.102.9.104): icmp_seq=4 ttl=243 time=100 ms
64 bytes from lm-in-f104.google.com (66.102.9.104): icmp_seq=5 ttl=243 time=100 ms
64 bytes from lm-in-f104.google.com (66.102.9.104): icmp_seq=6 ttl=243 time=100 ms
64 bytes from lm-in-f104.google.com (66.102.9.104): icmp_seq=7 ttl=243 time=103 ms
64 bytes from lm-in-f104.google.com (66.102.9.104): icmp_seq=8 ttl=243 time=103 ms
64 bytes from lm-in-f104.google.com (66.102.9.104): icmp_seq=9 ttl=243 time=106 ms
64 bytes from lm-in-f104.google.com (66.102.9.104): icmp_seq=10 ttl=243 time=110 ms
64 bytes from lm-in-f104.google.com (66.102.9.104): icmp_seq=11 ttl=243 time=106 ms
64 bytes from lm-in-f104.google.com (66.102.9.104): icmp_seq=12 ttl=243 time=106 ms
64 bytes from lm-in-f104.google.com (66.102.9.104): icmp_seq=13 ttl=243 time=106 ms
64 bytes from lm-in-f104.google.com (66.102.9.104): icmp_seq=14 ttl=243 time=106 ms
64 bytes from lm-in-f104.google.com (66.102.9.104): icmp_seq=15 ttl=243 time=106 ms
64 bytes from lm-in-f104.google.com (66.102.9.104): icmp_seq=16 ttl=243 time=105 ms
64 bytes from lm-in-f104.google.com (66.102.9.104): icmp_seq=17 ttl=243 time=104 ms
64 bytes from lm-in-f104.google.com (66.102.9.104): icmp_seq=18 ttl=243 time=105 ms
64 bytes from lm-in-f104.google.com (66.102.9.104): icmp_seq=19 ttl=243 time=104 ms
64 bytes from lm-in-f104.google.com (66.102.9.104): icmp_seq=20 ttl=243 time=104 ms
64 bytes from lm-in-f104.google.com (66.102.9.104): icmp_seq=21 ttl=243 time=104 ms
64 bytes from lm-in-f104.google.com (66.102.9.104): icmp_seq=22 ttl=243 time=104 ms
64 bytes from lm-in-f104.google.com (66.102.9.104): icmp_seq=23 ttl=243 time=104 ms
64 bytes from lm-in-f104.google.com (66.102.9.104): icmp_seq=24 ttl=243 time=107 ms
64 bytes from lm-in-f104.google.com (66.102.9.104): icmp_seq=25 ttl=243 time=104 ms
64 bytes from lm-in-f104.google.com (66.102.9.104): icmp_seq=26 ttl=243 time=104 ms
64 bytes from lm-in-f104.google.com (66.102.9.104): icmp_seq=27 ttl=243 time=103 ms
64 bytes from lm-in-f104.google.com (66.102.9.104): icmp_seq=28 ttl=243 time=104 ms
64 bytes from lm-in-f104.google.com (66.102.9.104): icmp_seq=29 ttl=243 time=103 ms
64 bytes from lm-in-f104.google.com (66.102.9.104): icmp_seq=30 ttl=243 time=104 ms
64 bytes from lm-in-f104.google.com (66.102.9.104): icmp_seq=31 ttl=243 time=104 ms
64 bytes from lm-in-f104.google.com (66.102.9.104): icmp_seq=32 ttl=243 time=104 ms
64 bytes from lm-in-f104.google.com (66.102.9.104): icmp_seq=33 ttl=243 time=103 ms
64 bytes from lm-in-f104.google.com (66.102.9.104): icmp_seq=34 ttl=243 time=107 ms
64 bytes from lm-in-f104.google.com (66.102.9.104): icmp_seq=35 ttl=243 time=103 ms
^C
--- www.l.google.com ping statistics ---
35 packets transmitted, 35 received, 0% packet loss, time 34003ms
rtt min/avg/max/mdev = 99.871/104.251/110.063/2.397 ms

Ok, so let's restart the driver.

sudo rmmod ndiswrapper
sudo modprobe ndiswrapper
sudo /etc/rc.d/network restart
speng@spengpc ~ $ ping -c 100 www.google.com
PING www.l.google.com (66.102.9.104) 56(84) bytes of data.
64 bytes from lm-in-f104.google.com (66.102.9.104): icmp_seq=1 ttl=243 time=41.2 ms
64 bytes from lm-in-f104.google.com (66.102.9.104): icmp_seq=2 ttl=243 time=43.2 ms
64 bytes from lm-in-f104.google.com (66.102.9.104): icmp_seq=3 ttl=243 time=42.3 ms
64 bytes from lm-in-f104.google.com (66.102.9.104): icmp_seq=4 ttl=243 time=43.9 ms
64 bytes from lm-in-f104.google.com (66.102.9.104): icmp_seq=5 ttl=243 time=43.5 ms
64 bytes from lm-in-f104.google.com (66.102.9.104): icmp_seq=6 ttl=243 time=43.2 ms
64 bytes from lm-in-f104.google.com (66.102.9.104): icmp_seq=7 ttl=243 time=43.7 ms
64 bytes from lm-in-f104.google.com (66.102.9.104): icmp_seq=8 ttl=243 time=43.2 ms
64 bytes from lm-in-f104.google.com (66.102.9.104): icmp_seq=9 ttl=243 time=44.3 ms
64 bytes from lm-in-f104.google.com (66.102.9.104): icmp_seq=10 ttl=243 time=42.4 ms
64 bytes from lm-in-f104.google.com (66.102.9.104): icmp_seq=11 ttl=243 time=41.5 ms
64 bytes from lm-in-f104.google.com (66.102.9.104): icmp_seq=12 ttl=243 time=40.9 ms
64 bytes from lm-in-f104.google.com (66.102.9.104): icmp_seq=13 ttl=243 time=41.6 ms
64 bytes from lm-in-f104.google.com (66.102.9.104): icmp_seq=14 ttl=243 time=43.9 ms
64 bytes from lm-in-f104.google.com (66.102.9.104): icmp_seq=15 ttl=243 time=41.0 ms
64 bytes from lm-in-f104.google.com (66.102.9.104): icmp_seq=16 ttl=243 time=45.6 ms
64 bytes from lm-in-f104.google.com (66.102.9.104): icmp_seq=17 ttl=243 time=41.9 ms
64 bytes from lm-in-f104.google.com (66.102.9.104): icmp_seq=18 ttl=243 time=43.1 ms
64 bytes from lm-in-f104.google.com (66.102.9.104): icmp_seq=19 ttl=243 time=41.9 ms
64 bytes from lm-in-f104.google.com (66.102.9.104): icmp_seq=20 ttl=243 time=43.4 ms
64 bytes from lm-in-f104.google.com (66.102.9.104): icmp_seq=21 ttl=243 time=41.5 ms
64 bytes from lm-in-f104.google.com (66.102.9.104): icmp_seq=22 ttl=243 time=43.2 ms
64 bytes from lm-in-f104.google.com (66.102.9.104): icmp_seq=23 ttl=243 time=40.5 ms
64 bytes from lm-in-f104.google.com (66.102.9.104): icmp_seq=24 ttl=243 time=43.5 ms
64 bytes from lm-in-f104.google.com (66.102.9.104): icmp_seq=25 ttl=243 time=43.4 ms
64 bytes from lm-in-f104.google.com (66.102.9.104): icmp_seq=26 ttl=243 time=40.9 ms
^C
--- www.l.google.com ping statistics ---
26 packets transmitted, 26 received, 0% packet loss, time 25033ms
rtt min/avg/max/mdev = 40.593/42.697/45.601/1.263 ms

You get the idea...
I am using these drivers: http://www.64bitjungle.com/wp-content/u … ers.tar.gz
Which I have been using for a year with ndiswrapper, which is why I'm surprised as why this is happening.

I have adjusted my modprobe/blacklist as suggested in the news section.
I am using a wireless PCI BCM4318 card (54g).

This happens with both the b43 driver and the ndiswrapper. IPv6 is disabled as always, I don't know what more I could add. It's such a strange problem.


Edit:

speng@spengpc ~ $ lspci | grep -i airforce
02:09.0 Network controller: Broadcom Corporation BCM4318 [AirForce One 54g] 802.11g Wireless LAN Controller (rev 02)

Captured a lag spike:

speng@spengpc ~ $ ping -c 100 www.google.com
PING www.l.google.com (66.102.9.99) 56(84) bytes of data.
64 bytes from lm-in-f99.google.com (66.102.9.99): icmp_seq=1 ttl=243 time=899 ms
64 bytes from lm-in-f99.google.com (66.102.9.99): icmp_seq=2 ttl=243 time=799 ms
64 bytes from lm-in-f99.google.com (66.102.9.99): icmp_seq=3 ttl=243 time=1399 ms
64 bytes from lm-in-f99.google.com (66.102.9.99): icmp_seq=4 ttl=243 time=1200 ms
64 bytes from lm-in-f99.google.com (66.102.9.99): icmp_seq=5 ttl=243 time=1999 ms
64 bytes from lm-in-f99.google.com (66.102.9.99): icmp_seq=6 ttl=243 time=1999 ms
64 bytes from lm-in-f99.google.com (66.102.9.99): icmp_seq=7 ttl=243 time=1600 ms
64 bytes from lm-in-f99.google.com (66.102.9.99): icmp_seq=8 ttl=243 time=1099 ms
64 bytes from lm-in-f99.google.com (66.102.9.99): icmp_seq=9 ttl=243 time=800 ms
64 bytes from lm-in-f99.google.com (66.102.9.99): icmp_seq=10 ttl=243 time=1299 ms
64 bytes from lm-in-f99.google.com (66.102.9.99): icmp_seq=11 ttl=243 time=1200 ms
64 bytes from lm-in-f99.google.com (66.102.9.99): icmp_seq=12 ttl=243 time=999 ms
64 bytes from lm-in-f99.google.com (66.102.9.99): icmp_seq=13 ttl=243 time=399 ms
64 bytes from lm-in-f99.google.com (66.102.9.99): icmp_seq=14 ttl=243 time=99.5 ms
64 bytes from lm-in-f99.google.com (66.102.9.99): icmp_seq=15 ttl=243 time=98.0 ms
64 bytes from lm-in-f99.google.com (66.102.9.99): icmp_seq=16 ttl=243 time=97.0 ms
64 bytes from lm-in-f99.google.com (66.102.9.99): icmp_seq=17 ttl=243 time=96.2 ms
64 bytes from lm-in-f99.google.com (66.102.9.99): icmp_seq=18 ttl=243 time=95.0 ms
64 bytes from lm-in-f99.google.com (66.102.9.99): icmp_seq=19 ttl=243 time=93.9 ms
64 bytes from lm-in-f99.google.com (66.102.9.99): icmp_seq=20 ttl=243 time=93.4 ms
64 bytes from lm-in-f99.google.com (66.102.9.99): icmp_seq=21 ttl=243 time=92.0 ms
64 bytes from lm-in-f99.google.com (66.102.9.99): icmp_seq=22 ttl=243 time=91.0 ms
64 bytes from lm-in-f99.google.com (66.102.9.99): icmp_seq=23 ttl=243 time=90.1 ms
^C
--- www.l.google.com ping statistics ---
23 packets transmitted, 23 received, 0% packet loss, time 31509ms
rtt min/avg/max/mdev = 90.122/723.694/1999.979/645.721 ms, pipe 3

Third edit:

I think this is my problem:

irq 17: nobody cared (try booting with the "irqpoll" option)
Pid: 0, comm: swapper Tainted: P           2.6.30-ARCH #1
Call Trace:
 <IRQ>  [<ffffffff802a5330>] ? __report_bad_irq+0x30/0xb0
 [<ffffffff80214d05>] ? sched_clock+0x5/0x10
 [<ffffffff802a5558>] ? note_interrupt+0x1a8/0x1f0
 [<ffffffff802a5fa5>] ? handle_fasteoi_irq+0xf5/0x120
 [<ffffffff8020fa89>] ? handle_irq+0x29/0x50
 [<ffffffff8020f0e2>] ? do_IRQ+0x72/0x110
 [<ffffffff8020ce53>] ? ret_from_intr+0x0/0x11
 <EOI>  [<ffffffff80215e1c>] ? mwait_idle+0x8c/0x1c0
 [<ffffffff80272b0d>] ? __atomic_notifier_call_chain+0x2d/0x80
 [<ffffffff8020b52a>] ? cpu_idle+0xba/0x120
 [<ffffffff806befda>] ? start_kernel+0x3cc/0x3ec
 [<ffffffff806be44d>] ? x86_64_start_kernel+0x10b/0x12f
handlers:
[<ffffffffa0e0a300>] (io_irq_isr+0x0/0x70 [ndiswrapper])
Disabling IRQ #17

Last edited by speng (2009-06-26 16:51:16)

Offline

#2 2009-06-27 21:06:13

Hohoho
Member
Registered: 2007-06-23
Posts: 222

Re: Pacman update - Very strange network problems/lag spikes

Same exact problem, slightly different network card (Broadcom Corporation BCM4312 802.11b/g (rev 01)), using latest ndiswrapper.
What's interesting is the error doesn't appear until something network intensive is happening (like a torrent downloading).

irq 17: nobody cared (try booting with the "irqpoll" option)
Pid: 0, comm: swapper Tainted: P           2.6.30-ARCH #1
Call Trace:
 [<c0186041>] ? __report_bad_irq+0x31/0xb0
 [<c0184b85>] ? handle_IRQ_event+0x55/0x1b0
 [<c018623a>] ? note_interrupt+0x17a/0x1c0
 [<c0186ac3>] ? handle_fasteoi_irq+0xc3/0xf0
 [<c01068f5>] ? handle_irq+0x25/0x50
 [<c0106034>] ? do_IRQ+0x54/0xd0
 [<c0104350>] ? common_interrupt+0x30/0x38
 [<f8451664>] ? acpi_idle_enter_bm+0x2a5/0x2e7 [processor]
 [<c031eaee>] ? cpuidle_idle_call+0x8e/0xf0
 [<c0102ae4>] ? cpu_idle+0xa4/0xf0
 [<c04e3a25>] ? start_kernel+0x335/0x34d
 [<c04e34e1>] ? unknown_bootoption+0x0/0x1bd
handlers:
[<f8290200>] (io_irq_isr+0x0/0x60 [ndiswrapper])
[<f9d02880>] (b44_interrupt+0x0/0x130 [b44])
Disabling IRQ #17

Booting with irqpoll does seem to help, but AFAIK it incurs some kind of performance penalty, so far everything seems to work as fast.

Last edited by Hohoho (2009-06-27 21:20:39)

Offline

#3 2009-07-18 20:00:12

jigglychu
Member
Registered: 2009-07-16
Posts: 1

Re: Pacman update - Very strange network problems/lag spikes

Any progress on this front? I'm having the exact same problem...

Offline

Board footer

Powered by FluxBB