I decided to try out the Ubuntu 10.10 RC on my workstation today. The install worked fine, but the mouse and keyboard (both of which is usb) are unbearably laggy. This was also the case when I ran it from the live cd.
The strange part is that it works fine for the first few seconds when entering X. Then something happens in the background, and the mouse all of a sudden behaves like I'm playing some demanding 3d game on the highest setting. The keyboard is also laggy, and if I press more than a single key per second the key will be dropped, and only about every fifth character will actually appear on screen. Even stranger is that this lag also is present when a use ctrl+alt+f1 to switch to console mode.
If I immediately switch to console mode when booting, it works fine. Then when I switch back to X, it takes a couple of seconds and then it happens again.
I've checked the logs, and the most suspicious message is:
Oct 3 20:16:16 CHURCHILL kernel: [ 70.868940] irq 19: nobody cared (try booting with the "irqpoll" option)
Oct 3 20:16:16 CHURCHILL kernel: [ 70.868946] Pid: 0, comm: swapper Not tainted 2.6.35-22-generic #33-Ubuntu
Oct 3 20:16:16 CHURCHILL kernel: [ 70.868948] Call Trace:
Oct 3 20:16:16 CHURCHILL kernel: [ 70.868950] <IRQ> [<ffffffff810cba5b>] __report_bad_irq+0x2b/0xa0
Oct 3 20:16:16 CHURCHILL kernel: [ 70.868961] [<ffffffff810cbc5c>] note_interrupt+0x18c/0x1d0
Oct 3 20:16:16 CHURCHILL kernel: [ 70.868967] [<ffffffff81085d10>] ? sched_clock_tick+0x60/0x90
Oct 3 20:16:16 CHURCHILL kernel: [ 70.868970] [<ffffffff810cc45d>] handle_fasteoi_irq+0xdd/0x110
Oct 3 20:16:16 CHURCHILL kernel: [ 70.868974] [<ffffffff8100cb12>] handle_irq+0x22/0x30
Oct 3 20:16:16 CHURCHILL kernel: [ 70.868978] [<ffffffff81590b2c>] do_IRQ+0x6c/0xf0
Oct 3 20:16:16 CHURCHILL kernel: [ 70.868981] [<ffffffff81589713>] ret_from_intr+0x0/0x11
Oct 3 20:16:16 CHURCHILL kernel: [ 70.868983] <EOI> [<ffffffff810360fb>] ? native_safe_halt+0xb/0x10
Oct 3 20:16:16 CHURCHILL kernel: [ 70.868990] [<ffffffff81012ecd>] default_idle+0x3d/0x90
Oct 3 20:16:16 CHURCHILL kernel: [ 70.868995] [<ffffffff81008da3>] cpu_idle+0xb3/0x110
Oct 3 20:16:16 CHURCHILL kernel: [ 70.868999] [<ffffffff8156f6fa>] rest_init+0x8a/0x90
Oct 3 20:16:16 CHURCHILL kernel: [ 70.869004] [<ffffffff81aedc9d>] start_kernel+0x387/0x390
Oct 3 20:16:16 CHURCHILL kernel: [ 70.869007] [<ffffffff81aed341>] x86_64_start_reservations+0x12c/0x130
Oct 3 20:16:16 CHURCHILL kernel: [ 70.869011] [<ffffffff81aed43f>] x86_64_start_kernel+0xfa/0x109
Oct 3 20:16:16 CHURCHILL kernel: [ 70.869013] handlers:
Oct 3 20:16:16 CHURCHILL kernel: [ 70.869014] [<ffffffff81404bb0>] (usb_hcd_irq+0x0/0x90)
Oct 3 20:16:16 CHURCHILL kernel: [ 70.869019] [<ffffffff81404bb0>] (usb_hcd_irq+0x0/0x90)
Oct 3 20:16:16 CHURCHILL kernel: [ 70.869022] [<ffffffff81404bb0>] (usb_hcd_irq+0x0/0x90)
Oct 3 20:16:16 CHURCHILL kernel: [ 70.869026] Disabling IRQ #19
IRQ related issues, is from what I can tell, rather common. However, most of them seems to be related to shared interrupts, but on my computer IRQ 19 is used exclusively by the USB controller:
root@CHURCHILL:/home/emil# cat /proc/interrupts
CPU0
0: 26 IO-APIC-edge timer
1: 2 IO-APIC-edge i8042
7: 1 IO-APIC-edge parport0
8: 0 IO-APIC-edge rtc0
9: 1 IO-APIC-fasteoi acpi, firewire_ohci
12: 4 IO-APIC-edge i8042
14: 0 IO-APIC-edge pata_atiixp
15: 1004 IO-APIC-edge pata_atiixp
17: 154 IO-APIC-fasteoi ATI IXP
18: 6164 IO-APIC-fasteoi radeon, hda_intel
19: 2000002 IO-APIC-fasteoi ehci_hcd:usb1, ohci_hcd:usb2, ohci_hcd:usb3
20: 9097 IO-APIC-fasteoi eth0, eth1
22: 13702 IO-APIC-fasteoi sata_sil
NMI: 0 Non-maskable interrupts
LOC: 105505 Local timer interrupts
SPU: 0 Spurious interrupts
PMI: 0 Performance monitoring interrupts
PND: 0 Performance pending work
RES: 0 Rescheduling interrupts
CAL: 0 Function call interrupts
TLB: 0 TLB shootdowns
TRM: 0 Thermal event interrupts
THR: 0 Threshold APIC interrupts
MCE: 0 Machine check exceptions
MCP: 4 Machine check polls
ERR: 1
MIS: 0
The on other person I've been able to found that had similar problems resorted to using irqpoll, as suggested by the kernel message. However, he reported that the inputs didn't work at all without that option, which makes me suggest that this might not really be IRQ related after all. Especially since it works fine for a while.
lspci
lists the USB controller as ATI Technologies Inc IXP SB400 USB Host Controller
.
Any suggestions would be greatly appreciated.
Best Answer
Maybe its not kernel bug, but some user process hogs resources? For example, epic gnome-settings-daemon (https://bugs.launchpad.net/bugs/625793)? Start System Monitor, open All processes from menu and sort by CPU usage, you may find answer...