KVM oops in svm_vcpu_run with 2.6.24-rc7

KVM oops in svm_vcpu_run with 2.6.24-rc7

[Andi Kleen]

While starting a kernel

Unable to handle kernel NULL pointer dereference at 0000000000000008 RIP: 
 [<ffffffff885698fd>] :kvm_amd:svm_vcpu_run+0x35/0x30d
PGD 471e1067 PUD 54ca9067 PMD 0 
Oops: 0000 [1] SMP 
last sysfs 
file: /devices/LNXSYSTM:00/device:00/PNP0C0A:00/power_supply/BAT1/cha
rge_full_design
CPU 0 
Modules linked in: kvm_amd kvm nls_iso8859_1 nls_cp437 vfat fat usb_storage 
af_p
acket rfkill_input sha256_generic cbc dm_crypt aes_x86_64 snd_pcm_oss 
snd_mixer_
oss snd_seq snd_seq_device cpufreq_conservative cpufreq_userspace 
cpufreq_powers
ave powernow_k8 loop dm_mod arc4 ecb blkcipher b43 rfkill mac80211 cfg80211 
led_
class input_polldev ssb pcmcia firmware_class hci_usb yenta_socket shpchp 
serio_
raw pci_hotplug tifm_7xx1 tifm_core bluetooth rtc_cmos ac snd_hda_intel 
rsrc_non
static snd_pcm snd_timer button snd_page_alloc irda crc_ccitt snd_hwdep 
pcmcia_c
ore ohci1394 rtc_core rtc_lib k8temp ieee1394 tg3 hwmon container i2c_piix4 
dock
 i2c_core snd battery power_supply joydev parport_pc parport soundcore sg 
ehci_h
cd ohci_hcd usbcore sd_mod edd ext3 mbcache jbd fan pata_atiixp sata_sil 
libata 
scsi_mod thermal processor
Pid: 19761, comm: qemu-kvm Tainted: G       N 
2.6.24-rc7-20080110105413-default 
#1
RIP: 0010:[<ffffffff885698fd>]  
[<ffffffff885698fd>] :kvm_amd:svm_vcpu_run+0x35/
0x30d
RSP: 0018:ffff81002a0b39c0  EFLAGS: 00010046
RAX: ffff810056e2e000 RBX: 0000000000000000 RCX: 0000000000000000
RDX: ffffffff80734780 RSI: ffff810056e21040 RDI: 0000000000000000
RBP: ffff810056e21040 R08: ffff810056e2e000 R09: 0000000000000000
R10: 00000000000007f9 R11: 0000000000000001 R12: ffff810056e22000
R13: ffff810056e210a0 R14: ffff810056e21700 R15: ffff810056e21500
FS:  00007fd584db16f0(0000) GS:ffffffff80592000(0000) knlGS:00000000b74e1b10
CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
CR2: 0000000000000008 CR3: 0000000034db1000 CR4: 00000000000006e0
DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
DR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7: 0000000000000400
Process qemu-kvm (pid: 19761, threadinfo ffff81002a0b2000, task 
ffff810054852080
)
Stack:  ffffffff8854edb8 00000000001280d2 ffff81002c9945c0 ffff8100000106a8
 0000000000000000 ffff810054852080 ffffffff80276e3d ffffffff80733f98
 0000000000000000 0000000000000000 00000010001280d2 ffff810000000008
Call Trace:
 [<ffffffff8854edb8>] :kvm:kvm_vcpu_ioctl+0x36d/0x11b8
 [<ffffffff802a47c9>] do_ioctl+0x21/0x6c
 [<ffffffff802a4a4a>] vfs_ioctl+0x236/0x24c
 [<ffffffff802a4ab1>] sys_ioctl+0x51/0x70
 [<ffffffff8020bfde>] system_call+0x7e/0x83
 [<00007fd583945267>]

-------------------------------------------------------------------------
This SF.net email is sponsored by: Microsoft
Defy all challenges. Microsoft(R) Visual Studio 2008.
http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/

Re: KVM oops in svm_vcpu_run with 2.6.24-rc7

[Avi Kivity]

Andi Kleen wrote:
> While starting a kernel
>
> Unable to handle kernel NULL pointer dereference at 0000000000000008 RIP: 
>  [<ffffffff885698fd>] :kvm_amd:svm_vcpu_run+0x35/0x30d
>   

Can you disassemble svm_vcpu_run() so we see where this is?

-- 
error compiling committee.c: too many arguments to function


-------------------------------------------------------------------------
This SF.net email is sponsored by: Microsoft
Defy all challenges. Microsoft(R) Visual Studio 2008.
http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/

Re: KVM oops in svm_vcpu_run with 2.6.24-rc7

[Andi Kleen]

On Wednesday 23 January 2008 12:28:06 Avi Kivity wrote:
> Andi Kleen wrote:
> > While starting a kernel
> >
> > Unable to handle kernel NULL pointer dereference at 0000000000000008 RIP:
> >  [<ffffffff885698fd>] :kvm_amd:svm_vcpu_run+0x35/0x30d
>
> Can you disassemble svm_vcpu_run() so we see where this ist

Pretty much at the beginning. It's the first reference of the first argument


00000000000008c8 <svm_vcpu_run>:
     8c8:       65 8b 0c 25 24 00 00    mov    %gs:0x24,%ecx
     8cf:       00 
     8d0:       48 63 c1                movslq %ecx,%rax
     8d3:       48 89 fe                mov    %rdi,%rsi
     8d6:       48 c7 c2 00 00 00 00    mov    $0x0,%rdx
     8dd:       48 8b 04 c5 00 00 00    mov    0x0(,%rax,8),%rax
     8e4:       00 
     8e5:       48 8b 40 08             mov    0x8(%rax),%rax
     8e9:       48 8b 3c 02             mov    (%rdx,%rax,1),%rdi
     8ed:       48 8b 86 f0 0d 00 00    mov    0xdf0(%rsi),%rax
     8f4:       c6 40 5c 00             movb   $0x0,0x5c(%rax)
     8f8:       39 4e 40                cmp    %ecx,0x40(%rsi)
     8fb:       75 0d                   jne    90a <svm_vcpu_run+0x42>
     8fd:       48 8b 47 08             mov    0x8(%rdi),%rax                

Last instruction faulted

-Andi

-------------------------------------------------------------------------
This SF.net email is sponsored by: Microsoft
Defy all challenges. Microsoft(R) Visual Studio 2008.
http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/

Re: KVM oops in svm_vcpu_run with 2.6.24-rc7

[Avi Kivity]

Andi Kleen wrote:
> On Wednesday 23 January 2008 12:28:06 Avi Kivity wrote:
>   
>> Andi Kleen wrote:
>>     
>>> While starting a kernel
>>>
>>> Unable to handle kernel NULL pointer dereference at 0000000000000008 RIP:
>>>  [<ffffffff885698fd>] :kvm_amd:svm_vcpu_run+0x35/0x30d
>>>       
>> Can you disassemble svm_vcpu_run() so we see where this ist
>>     
>
> Pretty much at the beginning. It's the first reference of the first argument
>
>
> 00000000000008c8 <svm_vcpu_run>:
>      8c8:       65 8b 0c 25 24 00 00    mov    %gs:0x24,%ecx
>      8cf:       00 
>      8d0:       48 63 c1                movslq %ecx,%rax
>      8d3:       48 89 fe                mov    %rdi,%rsi
>      8d6:       48 c7 c2 00 00 00 00    mov    $0x0,%rdx
>      8dd:       48 8b 04 c5 00 00 00    mov    0x0(,%rax,8),%rax
>      8e4:       00 
>      8e5:       48 8b 40 08             mov    0x8(%rax),%rax
>      8e9:       48 8b 3c 02             mov    (%rdx,%rax,1),%rdi
>      8ed:       48 8b 86 f0 0d 00 00    mov    0xdf0(%rsi),%rax
>      8f4:       c6 40 5c 00             movb   $0x0,0x5c(%rax)
>      8f8:       39 4e 40                cmp    %ecx,0x40(%rsi)
>      8fb:       75 0d                   jne    90a <svm_vcpu_run+0x42>
>      8fd:       48 8b 47 08             mov    0x8(%rdi),%rax                
>   

No, %rdi now points at per_cpu(svm_data, raw_smp_processor_id()); the 
instruction is referencing asid_generation.

Have you hotplugged cpus on this machine?  Maybe via suspend/resume?

-- 
error compiling committee.c: too many arguments to function


-------------------------------------------------------------------------
This SF.net email is sponsored by: Microsoft
Defy all challenges. Microsoft(R) Visual Studio 2008.
http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/