Skip to contentSkip to conversation

2.6.24-rc3-git3: usb hardware interrupts using 25% of CPU time

DISCUSSION Frederik Himpe 3 participants 4 messages Nov 28, 2007

lore
Share & tools
lore
#0 lore raw

I installed Linux 2.6.24-rc3-git3 on a PPC based Apple Powerbook G4 system. Now with top I see that over 20% of CPU time is used by hardware interrupts. Looking at /proc/interrupts, the number in the row ohci_hcd:usb1 is rising very quickly:

$ cat /proc/interrupts 
           CPU0       
 21:          1   MPIC 1    Edge      i2sbus: i2s-a (tx)
 24:      11008   MPIC 1    Level     ide1
 25:      33608   MPIC 1    Level     VIA-PMU
 26:         56   MPIC 1    Level     keywest i2c
 27:          1   MPIC 1    Edge      i2sbus: i2s-a (rx)
 29:     281578   MPIC 1    Level     ohci_hcd:usb1
 30:          0   MPIC 1    Level     i2sbus: i2s-a (control)
 31:          0   MPIC 1    Level     i2sbus: i2s-b (control)
 33:          1   MPIC 1    Edge      i2sbus: i2s-b (tx)
 34:          1   MPIC 1    Edge      i2sbus: i2s-b (rx)
 39:      41611   MPIC 1    Level     ide0
 40:          2   MPIC 1    Level     ohci1394
 41:          0   MPIC 1    Level     eth0
 42:          0   MPIC 1    Level     keywest i2c
 47:       2826   MPIC 1    Level     GPIO1 ADB
 48:      58633   MPIC 1    Level     radeon@pci:0000:00:10.0
 52:      18961   MPIC 1    Level     b43
 53:          1   MPIC 1    Level     yenta
 61:          1   MPIC 1    Edge      headphone-detect
 63:          0   MPIC 1    Level     ohci_hcd:usb2, ohci_hcd:usb3, 
ehci_hcd:usb4
BAD:     187917
$ lspci -nn
0000:00:0b.0 Host bridge [0600]: Apple Computer Inc. UniNorth 2 AGP 
[106b:0034]
0000:00:10.0 VGA compatible controller [0300]: ATI Technologies Inc RV350 
[Mobility Radeon 9600 M10] [1002:4e50]
0001:10:0b.0 Host bridge [0600]: Apple Computer Inc. UniNorth 2 PCI 
[106b:0035]
0001:10:12.0 Network controller [0280]: Broadcom Corporation BCM4306 
802.11b/g Wireless LAN Controller [14e4:4320] (rev 03)
0001:10:13.0 CardBus bridge [0607]: Texas Instruments PCI1510 PC card 
Cardbus Controller [104c:ac56]
0001:10:17.0 Unknown class [ff00]: Apple Computer Inc. KeyLargo/Intrepid 
Mac I/O [106b:003e]
0001:10:19.0 USB Controller [0c03]: Apple Computer Inc. KeyLargo/Intrepid 
USB [106b:003f]
0001:10:1a.0 USB Controller [0c03]: Apple Computer Inc. KeyLargo/Intrepid 
USB [106b:003f]
0001:10:1b.0 USB Controller [0c03]: NEC Corporation USB [1033:0035] (rev 
43)
0001:10:1b.1 USB Controller [0c03]: NEC Corporation USB [1033:0035] (rev 
43)
0001:10:1b.2 USB Controller [0c03]: NEC Corporation USB 2.0 [1033:00e0] 
(rev 04)
0002:24:0b.0 Host bridge [0600]: Apple Computer Inc. UniNorth 2 Internal 
PCI [106b:0036]
0002:24:0d.0 Unknown class [ff00]: Apple Computer Inc. UniNorth/Intrepid 
ATA/100 [106b:003b]
0002:24:0e.0 FireWire (IEEE 1394) [0c00]: Apple Computer Inc. UniNorth 2 
FireWire [106b:0031] (rev 81)
0002:24:0f.0 Ethernet controller [0200]: Apple Computer Inc. UniNorth 2 
GMAC (Sun GEM) [106b:0032] (rev 80)

Full dmesg, .config and lsmod output can be found on
http://artipc10.vub.ac.be/linux-2.6.24-rc3-git3/

Signature · 1 line
Frederik Himpe
#1 lore raw
Quoted 4 lines · Frederik Himpe

I installed Linux 2.6.24-rc3-git3 on a PPC based Apple Powerbook G4 system. Now with top I see that over 20% of CPU time is used by hardware interrupts. Looking at /proc/interrupts, the number in the row ohci_hcd:usb1 is rising very quickly:

Hi Frederik,
Thanks for the report. We suspect that this is an accounting
problem introduced when powerpc enabled NO_HZ and highres timers. Fortunatley from the testing we've done it doens't actually look like the time is being spent in hardware interupts, just accounted for there.

Yours Tony

linux.conf.au http://linux.conf.au/ || http://lca2008.linux.org.au/
Jan 28 - Feb 02 2008 The Australian Linux Technical Conference!

#2 lore raw
In reply to Frederik Himpe
Quoted 4 lines · Frederik Himpe

I installed Linux 2.6.24-rc3-git3 on a PPC based Apple Powerbook G4 system. Now with top I see that over 20% of CPU time is used by hardware interrupts. Looking at /proc/interrupts, the number in the row ohci_hcd:usb1 is rising very quickly:

What's the last kernel that worked correctly?

Quoted 60 lines
$ cat /proc/interrupts 
           CPU0       
 21:          1   MPIC 1    Edge      i2sbus: i2s-a (tx)
 24:      11008   MPIC 1    Level     ide1
 25:      33608   MPIC 1    Level     VIA-PMU
 26:         56   MPIC 1    Level     keywest i2c
 27:          1   MPIC 1    Edge      i2sbus: i2s-a (rx)
 29:     281578   MPIC 1    Level     ohci_hcd:usb1
 30:          0   MPIC 1    Level     i2sbus: i2s-a (control)
 31:          0   MPIC 1    Level     i2sbus: i2s-b (control)
 33:          1   MPIC 1    Edge      i2sbus: i2s-b (tx)
 34:          1   MPIC 1    Edge      i2sbus: i2s-b (rx)
 39:      41611   MPIC 1    Level     ide0
 40:          2   MPIC 1    Level     ohci1394
 41:          0   MPIC 1    Level     eth0
 42:          0   MPIC 1    Level     keywest i2c
 47:       2826   MPIC 1    Level     GPIO1 ADB
 48:      58633   MPIC 1    Level     radeon@pci:0000:00:10.0
 52:      18961   MPIC 1    Level     b43
 53:          1   MPIC 1    Level     yenta
 61:          1   MPIC 1    Edge      headphone-detect
 63:          0   MPIC 1    Level     ohci_hcd:usb2, ohci_hcd:usb3, 
ehci_hcd:usb4
BAD:     187917
$ lspci -nn
0000:00:0b.0 Host bridge [0600]: Apple Computer Inc. UniNorth 2 AGP 
[106b:0034]
0000:00:10.0 VGA compatible controller [0300]: ATI Technologies Inc RV350 
[Mobility Radeon 9600 M10] [1002:4e50]
0001:10:0b.0 Host bridge [0600]: Apple Computer Inc. UniNorth 2 PCI 
[106b:0035]
0001:10:12.0 Network controller [0280]: Broadcom Corporation BCM4306 
802.11b/g Wireless LAN Controller [14e4:4320] (rev 03)
0001:10:13.0 CardBus bridge [0607]: Texas Instruments PCI1510 PC card 
Cardbus Controller [104c:ac56]
0001:10:17.0 Unknown class [ff00]: Apple Computer Inc. KeyLargo/Intrepid 
Mac I/O [106b:003e]
0001:10:19.0 USB Controller [0c03]: Apple Computer Inc. KeyLargo/Intrepid 
USB [106b:003f]
0001:10:1a.0 USB Controller [0c03]: Apple Computer Inc. KeyLargo/Intrepid 
USB [106b:003f]
0001:10:1b.0 USB Controller [0c03]: NEC Corporation USB [1033:0035] (rev 
43)
0001:10:1b.1 USB Controller [0c03]: NEC Corporation USB [1033:0035] (rev 
43)
0001:10:1b.2 USB Controller [0c03]: NEC Corporation USB 2.0 [1033:00e0] 
(rev 04)
0002:24:0b.0 Host bridge [0600]: Apple Computer Inc. UniNorth 2 Internal 
PCI [106b:0036]
0002:24:0d.0 Unknown class [ff00]: Apple Computer Inc. UniNorth/Intrepid 
ATA/100 [106b:003b]
0002:24:0e.0 FireWire (IEEE 1394) [0c00]: Apple Computer Inc. UniNorth 2 
FireWire [106b:0031] (rev 81)
0002:24:0f.0 Ethernet controller [0200]: Apple Computer Inc. UniNorth 2 
GMAC (Sun GEM) [106b:0032] (rev 80)

Full dmesg, .config and lsmod output can be found on
http://artipc10.vub.ac.be/linux-2.6.24-rc3-git3/

#3 lore raw
Quoted 7 lines · Rafael J. Wysocki
Quoted 4 lines · Frederik Himpe

I installed Linux 2.6.24-rc3-git3 on a PPC based Apple Powerbook G4 system. Now with top I see that over 20% of CPU time is used by hardware interrupts. Looking at /proc/interrupts, the number in the row ohci_hcd:usb1 is rising very quickly:

What's the last kernel that worked correctly?

Actually I'm not sure: the kernel I was running before was Debian's 2.6.22 kernel package

Signature · 1 line
Frederik