* Re: e1000e 2.6.30.1 massive packet loss
[not found] <81bfc67a0907160049q3d43093cy52276f0fe1157d89@mail.gmail.com>
@ 2009-07-16 8:04 ` Eric Dumazet
2009-07-16 8:20 ` Caleb Cushing
0 siblings, 1 reply; 3+ messages in thread
From: Eric Dumazet @ 2009-07-16 8:04 UTC (permalink / raw)
To: Caleb Cushing; +Cc: Linux Kernel Mailing List, Linux Netdev List
Caleb Cushing a écrit :
> So from 2.6.30 to .1 I noticed that via an icmp test (mtr) 20-50%
> packet loss web browsing became impractical (downgrading fixed the
> issue). I'm wondering if the change to the driver (w/ msi) is the
> cause? also wondering if this is known and will be fixed in .2 I can
> provide more information if necessary.
>
Added netdev to get more attention to this.
What do you call 20-50% packet loss ?
(Losses to icmp flood itself, or tcp losses ?
Is the machine answering to icmp flood, or is it the origin ?
Are you saying packet loss was not occurring with 2.6.30 ?
Could you describe your methodology ?
You could give us results of "cat /proc/interrupts" from both
versions, because it sounds as an interrupt affinity problem.
# cat /proc/interrupts
# tc -s -d qdisc
# ifconfig -a
^ permalink raw reply [flat|nested] 3+ messages in thread
* Re: e1000e 2.6.30.1 massive packet loss
2009-07-16 8:04 ` e1000e 2.6.30.1 massive packet loss Eric Dumazet
@ 2009-07-16 8:20 ` Caleb Cushing
2009-07-16 9:05 ` Caleb Cushing
0 siblings, 1 reply; 3+ messages in thread
From: Caleb Cushing @ 2009-07-16 8:20 UTC (permalink / raw)
To: Eric Dumazet; +Cc: Linux Kernel Mailing List, Linux Netdev List
> What do you call 20-50% packet loss ?
sorry that's just the approximate average loss mtr was reporting. I
realize it's not the most accurate test, but I have no problems before
changing to .1 or after downgrading from it.
> (Losses to icmp flood itself, or tcp losses ?
> Is the machine answering to icmp flood, or is it the origin ?
mtr uses all icmp, but considering web browsing and I'm sure even dns
was affected I'd say all of it. I haven't done further investigation.
it was the origin and the losses occured between my desktop
machine(2.6.31.x) and my linksys wrt54gl(on openwrt) router. Since I
know the cable is good, and the losses haven't happened before or
since, and I even tried rebooting the router, I'm fairly confident
that it's a result of changes between the 2 kernel versions.
> Are you saying packet loss was not occurring with 2.6.30 ?
yes.
> Could you describe your methodology ?
>
> You could give us results of "cat /proc/interrupts" from both
> versions, because it sounds as an interrupt affinity problem.
these are from 2.6.30 I'll send .1 in a bit.
> # cat /proc/interrupts
CPU0 CPU1 CPU2 CPU3
0: 1605 0 0 0 XT-PIC-XT timer
1: 2 0 0 0 XT-PIC-XT i8042
2: 0 0 0 0 XT-PIC-XT cascade
4: 3 0 0 0 XT-PIC-XT ohci1394
5: 40347741 0 0 0 XT-PIC-XT
uhci_hcd:usb3, EMU10K1
6: 2 0 0 0 XT-PIC-XT floppy
8: 4738 0 0 0 XT-PIC-XT rtc0
9: 0 0 0 0 XT-PIC-XT
acpi, ehci_hcd:usb2, uhci_hcd:usb6
10: 0 0 0 0 XT-PIC-XT
uhci_hcd:usb4
11: 923781 0 0 0 XT-PIC-XT
ehci_hcd:usb1, uhci_hcd:usb5, uhci_hcd:usb7, uhci_hcd:usb8
12: 4 0 0 0 XT-PIC-XT i8042
25: 428267 425244 423627 427244 PCI-MSI-edge i915
26: 1194972 1211065 1191149 1200914 PCI-MSI-edge ahci
27: 7287597 7270695 7291505 7280533 PCI-MSI-edge eth0
NMI: 0 0 0 0 Non-maskable interrupts
LOC: 66007880 72160316 45190647 33707930 Local timer interrupts
SPU: 0 0 0 0 Spurious interrupts
RES: 67599 68493 54742 53268 Rescheduling interrupts
CAL: 288 295 363 323 Function call interrupts
TLB: 133835 49505 43942 48309 TLB shootdowns
TRM: 0 0 0 0 Thermal event interrupts
THR: 0 0 0 0 Threshold APIC interrupts
ERR: 0
MIS: 0
> # tc -s -d qdisc
qdisc pfifo_fast 0: dev eth0 root bands 3 priomap 1 2 2 2 1 2 0 0 1 1
1 1 1 1 1 1
Sent 33951429966 bytes 12436885 pkt (dropped 0, overlimits 0 requeues 1375)
rate 0bit 0pps backlog 0b 0p requeues 1375
> # ifconfig -a
eth0 Link encap:Ethernet HWaddr 00:21:9B:06:4C:C9
inet addr:192.168.1.3 Bcast:192.168.1.255 Mask:255.255.255.0
inet6 addr: fe80::221:9bff:fe06:4cc9/64 Scope:Link
UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1
RX packets:16135626 errors:0 dropped:0 overruns:0 frame:0
TX packets:28308711 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:100
RX bytes:2181880167 (2080.8 Mb) TX bytes:34811667228 (33198.9 Mb)
Memory:fdfc0000-fdfe0000
lo Link encap:Local Loopback
inet addr:127.0.0.1 Mask:255.0.0.0
inet6 addr: ::1/128 Scope:Host
UP LOOPBACK RUNNING MTU:16436 Metric:1
RX packets:92471 errors:0 dropped:0 overruns:0 frame:0
TX packets:92471 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:0
RX bytes:11927979 (11.3 Mb) TX bytes:11927979 (11.3 Mb)
--
Caleb Cushing
http://xenoterracide.blogspot.com
^ permalink raw reply [flat|nested] 3+ messages in thread
* Re: e1000e 2.6.30.1 massive packet loss
2009-07-16 8:20 ` Caleb Cushing
@ 2009-07-16 9:05 ` Caleb Cushing
0 siblings, 0 replies; 3+ messages in thread
From: Caleb Cushing @ 2009-07-16 9:05 UTC (permalink / raw)
To: Eric Dumazet; +Cc: Linux Kernel Mailing List, Linux Netdev List
2.6.30.1 with the issue.
CPU0 CPU1 CPU2 CPU3
0: 36 0 0 0 XT-PIC-XT timer
1: 2 0 0 0 XT-PIC-XT i8042
2: 0 0 0 0 XT-PIC-XT cascade
4: 3 0 0 0 XT-PIC-XT ohci1394
5: 392 0 0 0 XT-PIC-XT
uhci_hcd:usb2, EMU10K1
6: 2 0 0 0 XT-PIC-XT floppy
8: 163 0 0 0 XT-PIC-XT rtc0
9: 0 0 0 0 XT-PIC-XT
acpi, ehci_hcd:usb3, uhci_hcd:usb6
10: 0 0 0 0 XT-PIC-XT
uhci_hcd:usb4
11: 1318 0 0 0 XT-PIC-XT
ehci_hcd:usb1, uhci_hcd:usb5, uhci_hcd:usb7, uhci_hcd:usb8
12: 4 0 0 0 XT-PIC-XT i8042
25: 801 812 813 785 PCI-MSI-edge i915
26: 3233 3217 3261 3263 PCI-MSI-edge ahci
27: 199 200 184 173 PCI-MSI-edge eth0
NMI: 0 0 0 0 Non-maskable interrupts
LOC: 14886 16536 13854 15765 Local timer interrupts
SPU: 0 0 0 0 Spurious interrupts
RES: 504 515 358 397 Rescheduling interrupts
CAL: 55 79 74 44 Function call interrupts
TLB: 649 1202 685 902 TLB shootdowns
TRM: 0 0 0 0 Thermal event interrupts
THR: 0 0 0 0 Threshold APIC interrupts
ERR: 0
MIS: 0
qdisc pfifo_fast 0: dev eth0 root bands 3 priomap 1 2 2 2 1 2 0 0 1 1
1 1 1 1 1 1
Sent 82807 bytes 499 pkt (dropped 0, overlimits 0 requeues 0)
rate 0bit 0pps backlog 0b 0p requeues 0
eth0 Link encap:Ethernet HWaddr 00:21:9B:06:4C:C9
inet addr:192.168.1.3 Bcast:192.168.1.255 Mask:255.255.255.0
inet6 addr: fe80::221:9bff:fe06:4cc9/64 Scope:Link
UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1
RX packets:366 errors:0 dropped:0 overruns:0 frame:0
TX packets:519 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:100
RX bytes:130219 (127.1 Kb) TX bytes:86900 (84.8 Kb)
Memory:fdfc0000-fdfe0000
lo Link encap:Local Loopback
inet addr:127.0.0.1 Mask:255.0.0.0
inet6 addr: ::1/128 Scope:Host
UP LOOPBACK RUNNING MTU:16436 Metric:1
RX packets:817 errors:0 dropped:0 overruns:0 frame:0
TX packets:817 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:0
RX bytes:55043 (53.7 Kb) TX bytes:55043 (53.7 Kb)
^ permalink raw reply [flat|nested] 3+ messages in thread
end of thread, other threads:[~2009-07-16 9:05 UTC | newest]
Thread overview: 3+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
[not found] <81bfc67a0907160049q3d43093cy52276f0fe1157d89@mail.gmail.com>
2009-07-16 8:04 ` e1000e 2.6.30.1 massive packet loss Eric Dumazet
2009-07-16 8:20 ` Caleb Cushing
2009-07-16 9:05 ` Caleb Cushing
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).