Comment 3 for bug 253089

Thank you for that hint! I tried with "pci=nomsi" and haven't noticed any slowdown so far.

Now I just get this repeatedly:

[10781.427102] ACPI: PCI Interrupt 0000:0c:00.0[A] -> GSI 17 (level, low) -> IRQ 17
[10781.427294] PM: Writing back config space on device 0000:0c:00.0 at offset 1 (was 100002, writing 100006)
[10781.427384] iwl3945: Radio disabled by HW RF Kill switch
[10781.427411] ACPI: PCI interrupt for device 0000:0c:00.0 disabled

The device in the first line is my Intel 3945 wifi card. So this blurb just seems to happen from time to time without any noticeable harm.

So it seems the third block of messages is the actual offender here, it also starts much later than the iwl3945 spewage:

[20725.262003] irq 219, desc: c0478c00, depth: 0, count: 0, unhandled: 0
[20725.262003] ->handle_irq(): c016aef0, handle_bad_irq+0x0/0x2a0
[20725.262003] ->chip(): c044d8c0, no_irq_chip+0x0/0x40
[20725.262003] ->action(): 00000000
[20725.262003] IRQ_DISABLED set
[20725.262003] IRQ_MASKED set
[20725.262003] unexpected IRQ trap at vector db

So I can live with the "pci=nomsi" workaround, but ideally we can fix this by default?