RHEL guest oops with latest kvm-userspace

RHEL guest oops with latest kvm-userspace

[Jan Kiszka]

Hi Avi,

I've a problem with the last qemu-cvs merge into kvm-userspace: Booting
the RHEL kernel 2.6.18-53.el5 in with more than one VCPU gives me

...
CPU: L1 I cache: 32K, L1 D cache: 32K
CPU: L2 cache: 2048K
QEMU Virtual CPU version 0.9.1 stepping 03
general protection fault: 0000 [1] SMP
last sysfs file:
CPU 1
Modules linked in:
Pid: 0, comm: swapper Not tainted 2.6.18-53.el5 #1
RIP: 0010:[<ffffffff8006dd6d>]  [<ffffffff8006dd6d>] write_rdtscp_cb+0xa/0xd
RSP: 0018:ffff81002fc4bf90  EFLAGS: 00000046
RAX: 0000000000000001 RBX: 0000000000000001 RCX: 00000000c0000103
RDX: 0000000000000000 RSI: 0000000000000002 RDI: 0000000000000001
RBP: ffff81002fc45e70 R08: 0000000000000002 R09: 0000000000000030
R10: 0000000000000000 R11: 0000000000000000 R12: ffffffff8006dd63
R13: 0000000000000001 R14: 0000000000000000 R15: 0000000000000000
FS:  0000000000000000(0000) GS:ffff81002fc21840(0000) knlGS:0000000000000000
CS:  0010 DS: 0018 ES: 0018 CR0: 000000008005003b
CR2: 0000000000000000 CR3: 0000000000201000 CR4: 00000000000006e0
Process swapper (pid: 0, threadinfo ffff81002fc44000, task ffff81002fc25100)
Stack:  ffffffff80022baf ffffffff80068d0e 0000000000000000 ffffffff8039e080
 ffffffff8005bc22 ffff81002fc45e70 <EOI>  0000000000000001 0000000000000000
 0000000000000000 0000000000000030 0000000000000002 0000000000000000
Call Trace:
 <IRQ>  [<ffffffff80022baf>] smp_call_function_interrupt+0x57/0x75
 [<ffffffff80068d0e>] default_idle+0x0/0x50
 [<ffffffff8005bc22>] call_function_interrupt+0x66/0x6c
 <EOI>  [<ffffffff80068d37>] default_idle+0x29/0x50
 [<ffffffff80046f8d>] cpu_idle+0x95/0xb8
 [<ffffffff80074501>] start_secondary+0x45a/0x469


Code: 0f 30 c3 41 54 55 53 89 f3 e8 42 3a 02 00 85 db 41 89 c4 0f
RIP  [<ffffffff8006dd6d>] write_rdtscp_cb+0xa/0xd
 RSP <ffff81002fc4bf90>
 <0>Kernel panic - not syncing: Fatal exception
 BUG: warning at kernel/panic.c:137/panic() (Not tainted)

Call Trace:
 <IRQ>  [<ffffffff8008ccca>] panic+0x1e3/0x1f4
 [<ffffffff80196ae8>] do_unblank_screen+0x1b/0x132
 [<ffffffff800631aa>] oops_end+0x51/0x53
 [<ffffffff80069689>] die+0x3a/0x44
 [<ffffffff80063755>] do_general_protection+0xfe/0x107
 [<ffffffff8006dd63>] write_rdtscp_cb+0x0/0xd
 [<ffffffff8005bde9>] error_exit+0x0/0x84
 [<ffffffff8006dd63>] write_rdtscp_cb+0x0/0xd
 [<ffffffff8006dd6d>] write_rdtscp_cb+0xa/0xd
 [<ffffffff80022baf>] smp_call_function_interrupt+0x57/0x75
 [<ffffffff80068d0e>] default_idle+0x0/0x50
 [<ffffffff8005bc22>] call_function_interrupt+0x66/0x6c
 <EOI>  [<ffffffff80068d37>] default_idle+0x29/0x50
 [<ffffffff80046f8d>] cpu_idle+0x95/0xb8
 [<ffffffff80074501>] start_secondary+0x45a/0x469


I suspect that is due to some lost features or regression around cpuid
after the kvm merge into qemu, but I can't bisect in that reagion. Any
ideas? Further info needed? Just let me know.

Jan

-- 
Siemens AG, Corporate Technology, CT SE 2 ES-OS
Corporate Competence Center Embedded Linux

Re: RHEL guest oops with latest kvm-userspace

[Jan Kiszka]

Jan Kiszka wrote:
> Hi Avi,
> 
> I've a problem with the last qemu-cvs merge into kvm-userspace: Booting
> the RHEL kernel 2.6.18-53.el5 in with more than one VCPU gives me
> 
> ...
> CPU: L1 I cache: 32K, L1 D cache: 32K
> CPU: L2 cache: 2048K
> QEMU Virtual CPU version 0.9.1 stepping 03
> general protection fault: 0000 [1] SMP
> last sysfs file:
> CPU 1
> Modules linked in:
> Pid: 0, comm: swapper Not tainted 2.6.18-53.el5 #1
> RIP: 0010:[<ffffffff8006dd6d>]  [<ffffffff8006dd6d>] write_rdtscp_cb+0xa/0xd
> RSP: 0018:ffff81002fc4bf90  EFLAGS: 00000046
> RAX: 0000000000000001 RBX: 0000000000000001 RCX: 00000000c0000103
> RDX: 0000000000000000 RSI: 0000000000000002 RDI: 0000000000000001
> RBP: ffff81002fc45e70 R08: 0000000000000002 R09: 0000000000000030
> R10: 0000000000000000 R11: 0000000000000000 R12: ffffffff8006dd63
> R13: 0000000000000001 R14: 0000000000000000 R15: 0000000000000000
> FS:  0000000000000000(0000) GS:ffff81002fc21840(0000) knlGS:0000000000000000
> CS:  0010 DS: 0018 ES: 0018 CR0: 000000008005003b
> CR2: 0000000000000000 CR3: 0000000000201000 CR4: 00000000000006e0
> Process swapper (pid: 0, threadinfo ffff81002fc44000, task ffff81002fc25100)
> Stack:  ffffffff80022baf ffffffff80068d0e 0000000000000000 ffffffff8039e080
>  ffffffff8005bc22 ffff81002fc45e70 <EOI>  0000000000000001 0000000000000000
>  0000000000000000 0000000000000030 0000000000000002 0000000000000000
> Call Trace:
>  <IRQ>  [<ffffffff80022baf>] smp_call_function_interrupt+0x57/0x75
>  [<ffffffff80068d0e>] default_idle+0x0/0x50
>  [<ffffffff8005bc22>] call_function_interrupt+0x66/0x6c
>  <EOI>  [<ffffffff80068d37>] default_idle+0x29/0x50
>  [<ffffffff80046f8d>] cpu_idle+0x95/0xb8
>  [<ffffffff80074501>] start_secondary+0x45a/0x469
> 
> 
> Code: 0f 30 c3 41 54 55 53 89 f3 e8 42 3a 02 00 85 db 41 89 c4 0f
> RIP  [<ffffffff8006dd6d>] write_rdtscp_cb+0xa/0xd
>  RSP <ffff81002fc4bf90>
>  <0>Kernel panic - not syncing: Fatal exception
>  BUG: warning at kernel/panic.c:137/panic() (Not tainted)
> 
> Call Trace:
>  <IRQ>  [<ffffffff8008ccca>] panic+0x1e3/0x1f4
>  [<ffffffff80196ae8>] do_unblank_screen+0x1b/0x132
>  [<ffffffff800631aa>] oops_end+0x51/0x53
>  [<ffffffff80069689>] die+0x3a/0x44
>  [<ffffffff80063755>] do_general_protection+0xfe/0x107
>  [<ffffffff8006dd63>] write_rdtscp_cb+0x0/0xd
>  [<ffffffff8005bde9>] error_exit+0x0/0x84
>  [<ffffffff8006dd63>] write_rdtscp_cb+0x0/0xd
>  [<ffffffff8006dd6d>] write_rdtscp_cb+0xa/0xd
>  [<ffffffff80022baf>] smp_call_function_interrupt+0x57/0x75
>  [<ffffffff80068d0e>] default_idle+0x0/0x50
>  [<ffffffff8005bc22>] call_function_interrupt+0x66/0x6c
>  <EOI>  [<ffffffff80068d37>] default_idle+0x29/0x50
>  [<ffffffff80046f8d>] cpu_idle+0x95/0xb8
>  [<ffffffff80074501>] start_secondary+0x45a/0x469
> 
> 
> I suspect that is due to some lost features or regression around cpuid
> after the kvm merge into qemu, but I can't bisect in that reagion. Any
> ideas? Further info needed? Just let me know.
> 

Just to clarify:

Bad:
09f10bea3072a2b17ef015410954765f6fba82bd
(Merge branch 'qemu-cvs')

Good (/wrt to this bug):
0806ff96ee7f97da60b6ef2df0e03b57c2a1cfcf
(kvm: external module: update .gitignore)

Jan

-- 
Siemens AG, Corporate Technology, CT SE 2 ES-OS
Corporate Competence Center Embedded Linux

Re: RHEL guest oops with latest kvm-userspace

[Avi Kivity]

Jan Kiszka wrote:
> Hi Avi,
>
> I've a problem with the last qemu-cvs merge into kvm-userspace: Booting
> the RHEL kernel 2.6.18-53.el5 in with more than one VCPU gives me
>
> RIP: 0010:[<ffffffff8006dd6d>]  [<ffffffff8006dd6d>] write_rdtscp_cb+0xa/0xd
>   

...

> I suspect that is due to some lost features or regression around cpuid
> after the kvm merge into qemu, but I can't bisect in that reagion. Any
> ideas? Further info needed? Just let me know.

You're right - rdtscp probably appeared out of nowhere.  I'll fix it up.

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

Re: RHEL guest oops with latest kvm-userspace

[Avi Kivity]

Jan Kiszka wrote:
> Jan Kiszka wrote:
>   
>> Hi Avi,
>>
>> I've a problem with the last qemu-cvs merge into kvm-userspace: Booting
>> the RHEL kernel 2.6.18-53.el5 in with more than one VCPU gives me
>>
>> Modules linked in:
>> Pid: 0, comm: swapper Not tainted 2.6.18-53.el5 #1
>> RIP: 0010:[<ffffffff8006dd6d>]  [<ffffffff8006dd6d>] write_rdtscp_cb+0xa/0x

I fixed a cpuid screwup which exposed rdtscp when it isn't present; this 
should be fixed in latest git.

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

Re: RHEL guest oops with latest kvm-userspace

[Jan Kiszka]

[-- Attachment #1: Type: text/plain, Size: 546 bytes --]

Avi Kivity wrote:
> Jan Kiszka wrote:
>> Jan Kiszka wrote:
>>  
>>> Hi Avi,
>>>
>>> I've a problem with the last qemu-cvs merge into kvm-userspace: Booting
>>> the RHEL kernel 2.6.18-53.el5 in with more than one VCPU gives me
>>>
>>> Modules linked in:
>>> Pid: 0, comm: swapper Not tainted 2.6.18-53.el5 #1
>>> RIP: 0010:[<ffffffff8006dd6d>]  [<ffffffff8006dd6d>]
>>> write_rdtscp_cb+0xa/0x
> 
> I fixed a cpuid screwup which exposed rdtscp when it isn't present; this
> should be fixed in latest git.

It is, thanks.

Jan


[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 257 bytes --]