I have tested both with smp and without smp support in my kernel and I
experience the problem both ways.  I've tried driver arguments in my
xorg.conf, various kernel args at boot time, several different
versions of the nvidia driver, and the only think that worked (kind
of) was not running both the nvidia driver and and the broadcom
(ndiswrapper) driver at the same time.  I can use one of either but
not both at the same time.  (nv instead of nvidia driver is my
preferred way to go because the last time I tested it, the broadcom
driver in the kernel only allowed 802.11b for my particular 802.11g
card ( "Dell D820/Broadcom Corporation BCM4310 UART (rev 01)" ).

Chris Bozic

On 1/14/07, jinn <email address hidden> wrote:
> DMESG:
> wlan2: no IPv6 routers present
> spurious 8259A interrupt: IRQ7.
> BUG: soft lockup detected on CPU#0!                      <------------------------------ NOTICE THIS!
>  [<c014d30b>] softlockup_tick+0xab/0x110
>  [<c0131131>] update_process_times+0x31/0x80
>  [<c01178a0>] smp_apic_timer_interrupt+0x90/0xb0
>  [<c0103d43>] apic_timer_interrupt+0x1f/0x24
>  [<c014d681>] handle_IRQ_event+0x11/0x60
>  [<c014ef46>] handle_level_irq+0x96/0x120
>  [<c0105da9>] do_IRQ+0x69/0xd0
>  [<c0103cb2>] common_interrupt+0x1a/0x20
>  [<c012c1a2>] __do_softirq+0x62/0xf0
>  [<c012c26b>] do_softirq+0x3b/0x50
>  [<c012c4c5>] irq_exit+0x45/0x50
>  [<c01178a5>] smp_apic_timer_interrupt+0x95/0xb0
>  [<c0103d43>] apic_timer_interrupt+0x1f/0x24
>  [<c026a993>] acpi_processor_idle+0x1fe/0x3bb
>  [<c0405250>] unknown_bootoption+0x0/0x270
>  [<c0101c06>] cpu_idle+0x76/0xe0
>  [<c0405826>] start_kernel+0x366/0x430
>  [<c0405250>] unknown_bootoption+0x0/0x270
>  =======================
> irq 4: nobody cared (try booting with the "irqpoll" option)
>  [<c014e154>] __report_bad_irq+0x24/0x90
>  [<c014e3f8>] note_interrupt+0x238/0x270
>  [<f8b4d232>] ndis_isr+0x52/0xc0 [ndiswrapper]
>  [<c014ef9e>] handle_level_irq+0xee/0x120
>  [<c0105da9>] do_IRQ+0x69/0xd0
>  [<c0103cb2>] common_interrupt+0x1a/0x20
>  [<c014d681>] handle_IRQ_event+0x11/0x60
>  [<c014ef46>] handle_level_irq+0x96/0x120
>  [<c0105da9>] do_IRQ+0x69/0xd0
>  [<c0103cb2>] common_interrupt+0x1a/0x20
>  [<c014d681>] handle_IRQ_event+0x11/0x60
>  [<c014ef46>] handle_level_irq+0x96/0x120
>  [<c0105da9>] do_IRQ+0x69/0xd0
>  [<c0103cb2>] common_interrupt+0x1a/0x20
>  [<c012c1a2>] __do_softirq+0x62/0xf0
>  [<c012c26b>] do_softirq+0x3b/0x50
>  [<c012c4c5>] irq_exit+0x45/0x50
>  [<c01178a5>] smp_apic_timer_interrupt+0x95/0xb0
>  [<c0103d43>] apic_timer_interrupt+0x1f/0x24
>  [<c026a993>] acpi_processor_idle+0x1fe/0x3bb
>  [<c0405250>] unknown_bootoption+0x0/0x270
>  [<c0101c06>] cpu_idle+0x76/0xe0
>  [<c0405826>] start_kernel+0x366/0x430
>  [<c0405250>] unknown_bootoption+0x0/0x270
>  =======================
> handlers:
> [<f8b4d1e0>] (ndis_isr+0x0/0xc0 [ndiswrapper])
> Disabling IRQ #4
> jinn@Estergon:0 >
>
>
> Adding "nosmp" to kernel parameters seems to have done it for me. Works just fine and I am using ndiswrapper right now.
>
> I would really like to know now, to ensure that this is smp related (and
> not only in my case) that if you have smp systems as well, and if nosmp
> helps you also.
>
> thx
> Jinn
>
> --
> BCM1390M (broadcom 4311) with ndiswrapper and nvidia IRQ conflicts
> https://launchpad.net/bugs/57355
>