lkml.org 
[lkml]   [2010]   [Mar]   [31]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
    Patch in this message
    /
    SubjectRe: BUG: null pointer deref in dmar_ir_support
    From
    Date
    [Problem is in the x86 portion of the code, so copying x86 folks for
    their attention]

    On Wed, 2010-03-31 at 11:34 -0700, Jan.Grossmann@kielnet.net wrote:
    > >> [ 0.144012] BUG: unable to handle kernel NULL pointer dereference at
    > >> 0000000000000025
    > >> [ 0.148000] IP: [<ffffffff81199c96>] dmar_ir_support+0xb/0x14
    > >> [ 0.148000] PGD 0
    > >> [ 0.148000] Oops: 0000 [#1] SMP
    > >> [ 0.148000] last sysfs file:
    > >> [ 0.148000] CPU 0
    > >> [ 0.148000] Modules linked in:
    > >> [ 0.148000]
    > >> [ 0.148000] Pid: 1, comm: swapper Not tainted 2.6.34-rc2 #4 A8N-SLI
    > >> Premium/System Product Name
    > >> [ 0.148000] RIP: 0010:[<ffffffff81199c96>] [<ffffffff81199c96>]
    > >> dmar_ir_support+0xb/0x14
    > >> [ 0.148000] RSP: 0018:ffff88007fbe3e30 EFLAGS: 00010046
    > >> [ 0.148000] RAX: 0000000000000000 RBX: 0000000000000000 RCX:
    > >> 000000000000004c
    > >> [ 0.148000] RDX: ffffffff81762338 RSI: 0000000000000046 RDI:
    > >> 0000000000000010
    > >> [ 0.148000] RBP: ffff88007fbe3e30 R08: 0000000000000002 R09:
    > >> 000000000000000a
    > >> [ 0.148000] R10: 0000000000000005 R11: 0000000000000000 R12:
    > >> 0000000000000010
    > >> [ 0.148000] R13: 0000000000000246 R14: 0000000000000000 R15:
    > >> 000000000000e850
    > >> [ 0.148000] FS: 0000000000000000(0000) GS:ffff880001a00000(0000)
    > >> knlGS:0000000000000000
    > >> [ 0.148000] CS: 0010 DS: 0000 ES: 0000 CR0: 000000008005003b
    > >> [ 0.148000] CR2: 0000000000000025 CR3: 0000000001625000 CR4:
    > >> 00000000000006f0
    > >> [ 0.148000] DR0: 0000000000000000 DR1: 0000000000000000 DR2:
    > >> 0000000000000000
    > >> [ 0.148000] DR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7:
    > >> 0000000000000400
    > >> [ 0.148000] Process swapper (pid: 1, threadinfo ffff88007fbe2000,
    > >> task ffff88007fbde740)
    > >> [ 0.148000] Stack:
    > >> [ 0.148000] ffff88007fbe3e40 ffffffff816d1f01 ffff88007fbe3e50
    > >> ffffffff816b7fa6
    > >> [ 0.148000] <0> ffff88007fbe3ea0 ffffffff816b8095 ffff88007fbe3ec0
    > >> ffffffff8130122e
    > >> [ 0.148000] <0> ffff880000000008 ffff88007fbe3ed0 ffffffff8168ec20
    > >> 0000000000000200
    > >> [ 0.148000] Call Trace:
    > >> [ 0.148000] [<ffffffff816d1f01>] intr_remapping_supported+0x12/0x45
    > >> [ 0.148000] [<ffffffff816b7fa6>] enable_IR+0x9/0x5d
    > >> [ 0.148000] [<ffffffff816b8095>] enable_IR_x2apic+0x9b/0x1b6
    > >> [ 0.148000] [<ffffffff8130122e>] ? printk+0x3c/0x3e
    > >> [ 0.148000] [<ffffffff816b81f7>] APIC_init_uniprocessor+0x47/0x105
    > >> [ 0.148000] [<ffffffff816b62cf>] native_smp_prepare_cpus+0x190/0x365
    > >> [ 0.148000] [<ffffffff816ab5f5>] kernel_init+0x71/0x1f0
    > >> [ 0.148000] [<ffffffff810097e4>] kernel_thread_helper+0x4/0x10
    > >> [ 0.148000] [<ffffffff816ab584>] ? kernel_init+0x0/0x1f0
    > >> [ 0.148000] [<ffffffff810097e0>] ? kernel_thread_helper+0x0/0x10
    > >> [ 0.148000] Code: ff c7 06 00 00 00 00 48 8b 04 fd 80 ba 42 81 eb 0d
    > >> c7 06 02 00 00 00 48 c7 c0 e8 30 4a 81 c9 c3 48 8b 05 0e 4c 56 00 55 48
    > >> 89 e5 <0f> b6 40 25 c9 83 e0 01 c3 55 48 89 e5 e8 7c 8f 16 00 c9 c3 55
    > >> [ 0.148000] RIP [<ffffffff81199c96>] dmar_ir_support+0xb/0x14
    > >> [ 0.148000] RSP <ffff88007fbe3e30>
    > >> [ 0.148000] CR2: 0000000000000025
    > >>

    Jan, Problem is that your system doesn't have proper SMP tables and we
    were doing enable_IR_x2apic() twice finally leading to this panic. Can
    you please check if the appended patch fixes your issue? Thanks.
    ---

    From: Suresh Siddha <suresh.b.siddha@intel.com>
    Subject: x86: fix calling enable_IR_x2apic() twice with SMP kernel on !SMP boards

    Jan Grossmann reported kernel boot panic while booting SMP kernel on his
    system with a single core cpu. SMP kernels call enable_IR_x2apic() from
    native_smp_prepare_cpus() and on platforms where the kernel doesn't
    find SMP configuration we ended up again calling enable_IR_x2apic() from the
    APIC_init_uniprocessor() call in the smp_sanity_check(). Thus leading to
    kernel panic.

    Don't call enable_IR_x2apic() and default_setup_apic_routing()
    from APIC_init_uniprocessor() in CONFIG_SMP case.

    Reported-by: Jan.Grossmann@kielnet.net
    Signed-off-by: Suresh Siddha <suresh.b.siddha@intel.com>
    Cc: stable@kernel.org [2.6.32, 2.6.33]
    ---

    diff --git a/arch/x86/kernel/apic/apic.c b/arch/x86/kernel/apic/apic.c
    index 00187f1..e5a4a1e 100644
    --- a/arch/x86/kernel/apic/apic.c
    +++ b/arch/x86/kernel/apic/apic.c
    @@ -1640,8 +1640,10 @@ int __init APIC_init_uniprocessor(void)
    }
    #endif

    +#ifndef CONFIG_SMP
    enable_IR_x2apic();
    default_setup_apic_routing();
    +#endif

    verify_local_APIC();
    connect_bsp_APIC();



    \
     
     \ /
      Last update: 2010-04-01 03:09    [W:0.046 / U:28.992 seconds]
    ©2003-2016 Jasper Spaans. hosted at Digital OceanAdvertise on this site