Re: kernel 4.4-rc1 BUG when booting as Xen dom0

[Boris Ostrovsky <boris.ostrovsky@oracle.com>]

On 11/17/2015 08:38 AM, Juergen Gross wrote:
> Trying to boot a 4.4 kernel as Xen dom0 crashes the system:
>
> [    9.949589] ACPI: Added _OSI(Module Device)
> [    9.957803] ACPI: Added _OSI(Processor Device)
> [    9.966814] ACPI: Added _OSI(3.0 _SCP Extensions)
> [    9.976346] ACPI: Added _OSI(Processor Aggregator Device)
> [    9.988715] BUG: unable to handle kernel NULL pointer dereference at
>            (null)
> [   10.004212] IP: [<ffffffff8135cf32>] xen_irq_init+0x22/0xe0
> [   10.015468] PGD 0
> [   10.019622] Oops: 0002 [#1] SMP
> [   10.026217] Modules linked in:
> [   10.032446] CPU: 0 PID: 1 Comm: swapper/0 Not tainted 4.4.0-rc1-pv+ #490
> [   10.045974] Hardware name: Dell Inc. Latitude E6440/0159N7, BIOS A07
> 06/26/2014
> [   10.060718] task: ffff8801fa570d40 ti: ffff8801fa574000 task.ti:
> ffff8801fa574000
> [   10.075798] RIP: e030:[<ffffffff8135cf32>]  [<ffffffff8135cf32>]
> xen_irq_init+0x22/0xe0
> [   10.091910] RSP: e02b:ffff8801fa577d38  EFLAGS: 00010246
> [   10.102646] RAX: 0000000000000000 RBX: 0000000000000009 RCX:
> 0000000000000000
> [   10.117089] RDX: 0000000000000001 RSI: 00000000024080c0 RDI:
> ffff8801fb000df0
> [   10.131437] RBP: 0000000000000009 R08: 0000000000000001 R09:
> ffff8801fb000d88
> [   10.145824] R10: ffff8801fb000df8 R11: 0000000000000000 R12:
> 0000000000000009
> [   10.160205] R13: ffffffff817736e5 R14: 0000000000000009 R15:
> 0000000000000000
> [   10.174592] FS:  0000000000000000(0000) GS:ffff8801ff800000(0000)
> knlGS:0000000000000000
> [   10.190895] CS:  e033 DS: 0000 ES: 0000 CR0: 0000000080050033
> [   10.202492] CR2: 0000000000000000 CR3: 000000000180c000 CR4:
> 0000000000042660
> [   10.216878] Stack:
> [   10.221029]  0000000000000009 0000000000000001 ffffffff8135dd9f
> 0000000000000000
> [   10.235931]  ffff8801fa577dd0 0000000000000009 0000000000000000
> 0000000000000001
> [   10.250836]  ffffffff81932e2e 0000000000000000 0000000000000000
> ffffffff81415c53
> [   10.265740] Call Trace:
> [   10.270774]  [<ffffffff8135dd9f>] ? xen_bind_pirq_gsi_to_irq+0x7f/0x240
> [   10.284122]  [<ffffffff81932e2e>] ? acpi_sleep_proc_init+0x22/0x22
> [   10.296601]  [<ffffffff81415c53>] ?
> xen_register_pirq.constprop.9+0x83/0xd0
> [   10.310640]  [<ffffffff81415ce8>] ? acpi_register_gsi_xen+0x48/0xc0
> [   10.323292]  [<ffffffff8104d8a6>] ? acpi_get_override_irq+0xa6/0x120
> [   10.336117]  [<ffffffff8133515c>] ? acpi_ev_sci_dispatch+0x53/0x53
> [   10.348595]  [<ffffffff81932e2e>] ? acpi_sleep_proc_init+0x22/0x22
> [   10.361071]  [<ffffffff810465f9>] ? acpi_gsi_to_irq+0x59/0x70
> [   10.372683]  [<ffffffff8131f781>] ?
> acpi_os_install_interrupt_handler+0x44/0xbf
> [   10.387427]  [<ffffffff81332da0>] ?
> acpi_ev_install_xrupt_handlers+0x12/0x5f
> [   10.401627]  [<ffffffff81932e9f>] ? acpi_init+0x71/0x289
> [   10.412368]  [<ffffffff81932e2e>] ? acpi_sleep_proc_init+0x22/0x22
> [   10.424849]  [<ffffffff81002101>] ? do_one_initcall+0xa1/0x1d0
> [   10.436633]  [<ffffffff81087700>] ? parse_args+0x270/0x460
> [   10.447724]  [<ffffffff818f81cb>] ? kernel_init_freeable+0x169/0x1f3
> [   10.460551]  [<ffffffff818f7958>] ? initcall_blacklist+0xb2/0xb2
> [   10.472681]  [<ffffffff81514750>] ? rest_init+0x80/0x80
> [   10.483251]  [<ffffffff81514755>] ? kernel_init+0x5/0xe0
> [   10.493996]  [<ffffffff8152000f>] ? ret_from_fork+0x3f/0x70
> [   10.505263]  [<ffffffff81514750>] ? rest_init+0x80/0x80
> [   10.515832] Code: fe ff ff 66 0f 1f 44 00 00 55 89 fd 53 e8 c7 f5 d5
> ff 48 85 c0 74 08 48 8b 40 10 48 83 c0 18 48 8b 15 33 d9 2a 00 be c0 80
> 40 02 <48> 89 10 48 8b 15 2c d9 2a 00 48 89 50 08 48 8b 15 29 d9 2a 00
> [   10.553606] RIP  [<ffffffff8135cf32>] xen_irq_init+0x22/0xe0
> [   10.565051]  RSP <ffff8801fa577d38>
> [   10.572157] CR2: 0000000000000000
>
> irq_get_affinity_mask(irq) in xen_irq_init() is returning NULL:
>
> static void xen_irq_init(unsigned irq)
> {
>          struct irq_info *info;
> #ifdef CONFIG_SMP
>          /* By default all event channels notify CPU#0. */
>          cpumask_copy(irq_get_affinity_mask(irq), cpumask_of(0));
> #endif
>
> Is this already known?


Yes: 
http://lists.xenproject.org/archives/html/xen-devel/2015-11/msg01800.html

I will send patch later today.

-boris