Prev: [PATCH]hostap:hostap_hw.c Fix typo.
Next: mount notification question
From: MadLoisae on 2 Aug 2010 17:00
Hi Guys,

I've updated from linux 2.6.34.1 to 2.6.35. With 2.6.34.1 and before I
had on my box a load average of about 0.4 to 0.5. Now with the same
processes / hardware I have got between 1.2 and 1.3.
In top I can see nothing is going on - CPU is still about 70% idle.
I am using vmware server 1.0.10 on it, the modules are the same as in
2.6.34.1. If running, vmware "eats" about 25% CPU (there is only one
machine running) or adds about 0.25 to 0.3 onto load - not more than
before 2.6.35.
If I stop vmware and remove the modules the load decreases of about 0.3
- this was also with 2.6.34, I had without vmware a load between 0.05
and 0.1.

Attached my dmesg of 2.6.34.1 and 2.6.35 as well as my actual kernel-config.
Altough I have compiled in with 2.6.35 wireless lan support I have
actually not the hardware. So the complete WIFI-stack is not used, no
modules are loaded.

Hardware / software:
Debian lenny
CPU: Via Nehemiah 1GHz, one CPU
1GB DDR-RAM, no swap configured
primary ide channel: not in use
secondary master: Transcend compactflash 16GiB, 1 partition, ext3 (/)
secondary slave: HDD Western Digital 320Gib, 4 partitions, ext4 (/home, ...)
4 Realtek RTL8139, 2 of them in use (eth0, eth1)
6 or 8 USB-Ports, none of them in use
PCMCIA sockets, not in use
Via AC97 sound, not in use
Terratec Cinergy 1200 DVB-C, in use with vdr
serial line, in use with APC-USV

There are no interrupt errors:
CPU0
0: 3946346 XT-PIC-XT timer
1: 2 XT-PIC-XT i8042
2: 0 XT-PIC-XT cascade
3: 2 XT-PIC-XT
4: 6757 XT-PIC-XT serial
7: 163510 XT-PIC-XT yenta, uhci_hcd:usb2, eth0
8: 2 XT-PIC-XT rtc
9: 0 XT-PIC-XT acpi
10: 0 XT-PIC-XT uhci_hcd:usb4, VIA8233
11: 1385932 XT-PIC-XT yenta, ehci_hcd:usb1, uhci_hcd:usb3,
saa7146 (0), eth1
12: 4 XT-PIC-XT i8042
14: 0 XT-PIC-XT pata_via
15: 25173 XT-PIC-XT pata_via
NMI: 0 Non-maskable interrupts
LOC: 0 Local timer interrupts
SPU: 0 Spurious interrupts
PMI: 0 Performance monitoring interrupts
PND: 0 Performance pending work
ERR: 0
MIS: 0

atop also does not display any other values than with 2.6.34 except the
higher load:
PRC | sys 0.38s | user 1.80s | #proc 142 | #zombie 0 |
#exit 1 |
CPU | sys 3% | user 17% | irq 0% | idle 77% |
wait 2% |
CPL | avg1 1.18 | avg5 1.22 | avg15 1.18 | csw 8090 | intr
3914 |
MEM | tot 993.3M | free 230.5M | cache 546.3M | buff 19.9M | slab
20.6M |
SWP | tot 0.0M | free 0.0M | | vmcom 581.4M | vmlim
496.6M |
DSK | sda | busy 2% | read 0 | write 8 | avio
28 ms |
NET | transport | tcpi 289 | tcpo 372 | udpi 0 |
udpo 0 |
NET | network | ipi 289 | ipo 278 | ipfrw 0 |
deliv 289 |
NET | eth1 0% | pcki 262 | pcko 345 | si 12 Kbps | so 187
Kbps |
NET | eth0 0% | pcki 133 | pcko 0 | si 6 Kbps | so 0
Kbps |
NET | lo ---- | pcki 27 | pcko 27 | si 1 Kbps | so 1
Kbps |

At this moment vmware was not running, only a few xterm-windows and
thunderbird over vnc. The high load is not I/O-load, CF as HDD are
nearly idle theo whole time (below 5blocks/second read / written). There
is nearly no network traffic (below 1kbps in/out). The DVB-C card
acutally does nothing.

Any ideas what I can search for?

Thanks!

kind regards
Alois
 | 
Pages: 1
Prev: [PATCH]hostap:hostap_hw.c Fix typo.
Next: mount notification question