Opened 18 years ago
Closed 18 years ago
#317 closed defect (fixed)
KVM & VBox conflict (Linux Host) -> fixed in SVN/1.4.2
| Reported by: | rojo | Owned by: | |
|---|---|---|---|
| Component: | other | Version: | VirtualBox 1.3.8 |
| Keywords: | GPF general protection fault freeze POST | Cc: | |
| Guest type: | other | Host type: | other |
Description
VM freezes immediately when powered on. Black VirtualBox window appears, "Starting the virtual machine" progress bar appears, and freezes at 0%. Stack trace follows. Tested on Ubuntu Edgy and Ubuntu Feisty hosts, with VirtualBox 1.3.8 and with locally-built SVN, same results, on a Gigabyte GA-M57SLI-S4 with latest production BIOS (and with a beta BIOS I made Gigabyte build me containing HPET support). KVM GPF's at the same spot and gives pretty much an identical stack trace (as far as I can tell on cursory observation). For what it's worth, kqemu didn't work until I had Gigabyte supply me a BIOS supporting HPET. Is this a hardware issue, maybe a crash when trying to access the CPU's hardware virtualization instruction set? cat /proc/cpuinfo follows the stack trace.
Message from syslogd@blister at Thu Apr 26 22:29:20 2007 ... blister kernel: [ 6514.852000] SMP
Message from syslogd@blister at Thu Apr 26 22:29:20 2007 ... blister kernel: [ 6514.852000] EIP: 0060:[<f9632990>] Tainted: P VLI
Message from syslogd@blister at Thu Apr 26 22:29:20 2007 ... blister kernel: [ 6514.852000] EFLAGS: 00010286 (2.6.20-15-generic #2)
Message from syslogd@blister at Thu Apr 26 22:29:20 2007 ... blister kernel: [ 6514.852000] EIP is at svm_vcpu_run+0xc0/0x560 [kvm_amd]
Message from syslogd@blister at Thu Apr 26 22:29:20 2007 ... blister kernel: [ 6514.852000] esi: 00000000 edi: 00000000 ebp: da200464 esp: db6afd14
Message from syslogd@blister at Thu Apr 26 22:29:20 2007 ... blister kernel: [ 6514.852000] ds: 007b es: 007b ss: 0068
Message from syslogd@blister at Thu Apr 26 22:29:20 2007 ... blister kernel: [ 6514.852000] CPU: 0
Message from syslogd@blister at Thu Apr 26 22:29:20 2007 ... blister kernel: [ 6514.852000] Process kvm (pid: 6511, ti=db6ae000 task=c1a61a90 task.ti=db6ae000)
Message from syslogd@blister at Thu Apr 26 22:29:20 2007 ... blister kernel: [ 6514.852000] Stack: dfdff5c0 00000001 00000006 df9fd020 f7ed3800 c0250bad c025a2b5 00000001
Message from syslogd@blister at Thu Apr 26 22:29:20 2007 ... blister kernel: [ 6514.852000] f7c1d840 db6afd58 db6afec8 00016935 da200464 00000000 db6afec8 00000000
Message from syslogd@blister at Thu Apr 26 22:29:20 2007 ... blister kernel: [ 6514.852000] da200464 f9646a61 00000000 dfadec80 bfdb1680 00016935 00000000 f7c1d840
Message from syslogd@blister at Thu Apr 26 22:29:20 2007 ... blister kernel: [ 6514.852000] Call Trace:
Message from syslogd@blister at Thu Apr 26 22:29:20 2007 ... blister kernel: [ 6514.852000] [input_event+77/1216] input_event+0x4d/0x4c0
Message from syslogd@blister at Thu Apr 26 22:29:20 2007 ... blister kernel: [ 6514.852000] [dma_pool_free+197/320] dma_pool_free+0xc5/0x140
Message from syslogd@blister at Thu Apr 26 22:29:20 2007 ... blister kernel: [ 6514.852000] [<f9646a61>] kvm_vcpu_ioctl+0x491/0x9d0 [kvm]
Message from syslogd@blister at Thu Apr 26 22:29:20 2007 ... blister kernel: [ 6514.856000] [<f8eab5b1>] finish_urb+0x81/0x100 [ohci_hcd]
Message from syslogd@blister at Thu Apr 26 22:29:20 2007 ... blister kernel: [ 6514.856000] [<f8eab766>] dl_done_list+0x136/0x1f0 [ohci_hcd]
Message from syslogd@blister at Thu Apr 26 22:29:20 2007 ... blister kernel: [ 6514.856000] [<f905680c>] reiserfs_dirty_inode+0x8c/0xa0 [reiserfs]
Message from syslogd@blister at Thu Apr 26 22:29:20 2007 ... blister kernel: [ 6514.856000] [blockable_page_cache_readahead+114/208] blockable_page_cache_readahead+0x72/0xd0
Message from syslogd@blister at Thu Apr 26 22:29:20 2007 ... blister kernel: [ 6514.856000] [file_read_actor+228/240] file_read_actor+0xe4/0xf0
Message from syslogd@blister at Thu Apr 26 22:29:20 2007 ... blister kernel: [ 6514.856000] [do_generic_mapping_read+1132/1392] do_generic_mapping_read+0x46c/0x570
Message from syslogd@blister at Thu Apr 26 22:29:20 2007 ... blister kernel: [ 6514.856000] [generic_file_aio_read+257/592] generic_file_aio_read+0x101/0x250
Message from syslogd@blister at Thu Apr 26 22:29:20 2007 ... blister kernel: [ 6514.856000] [file_read_actor+0/240] file_read_actor+0x0/0xf0
Message from syslogd@blister at Thu Apr 26 22:29:20 2007 ... blister kernel: [ 6514.856000] [activate_page+32/176] activate_page+0x20/0xb0
Message from syslogd@blister at Thu Apr 26 22:29:20 2007 ... blister kernel: [ 6514.856000] [mark_page_accessed+45/64] mark_page_accessed+0x2d/0x40
Message from syslogd@blister at Thu Apr 26 22:29:20 2007 ... blister kernel: [ 6514.856000] [autoremove_wake_function+0/80] autoremove_wake_function+0x0/0x50
Message from syslogd@blister at Thu Apr 26 22:29:20 2007 ... blister kernel: [ 6514.856000] [<f96465d0>] kvm_vcpu_ioctl+0x0/0x9d0 [kvm]
Message from syslogd@blister at Thu Apr 26 22:29:20 2007 ... blister kernel: [ 6514.856000] [do_ioctl+43/144] do_ioctl+0x2b/0x90
Message from syslogd@blister at Thu Apr 26 22:29:20 2007 ... blister kernel: [ 6514.856000] [vfs_ioctl+92/672] vfs_ioctl+0x5c/0x2a0
Message from syslogd@blister at Thu Apr 26 22:29:20 2007 ... blister kernel: [ 6514.856000] [sys_ioctl+114/144] sys_ioctl+0x72/0x90
Message from syslogd@blister at Thu Apr 26 22:29:20 2007 ... blister kernel: [ 6514.856000] [sysenter_past_esp+105/169] sysenter_past_esp+0x69/0xa9
Message from syslogd@blister at Thu Apr 26 22:29:20 2007 ...
Message from syslogd@blister at Thu Apr 26 22:29:20 2007 ... blister kernel: [ 6514.856000] Code: 0f 85 41 03 00 00 8b 75 24 85 f6 75 0e 8b 44 24 28 80 78 08 00 0f 84 2c 03 00 00 8b 43 0c 8b 53 10 83 c8 10 89 43 0c 89 53 10 90 <0f> 01 dd 65 8b 15 04 00 00 00 b8 f4 87 42 c0 03 04 95 00 fc 3c blister kernel: [ 6514.856000] =======================
Message from syslogd@blister at Thu Apr 26 22:29:20 2007 ... blister kernel: [ 6514.856000] EIP: [<f9632990>] svm_vcpu_run+0xc0/0x560 [kvm_amd] SS:ESP 0068:db6afd14
Message from syslogd@blister at Thu Apr 26 22:29:20 2007 ... blister kernel: [ 6514.852000] eax: bd040017 ebx: c61e1000 ecx: f9634bc0 edx: 0000007f
Message from syslogd@blister at Thu Apr 26 23:38:54 2007 ... blister kernel: [10688.920000] general protection fault: 0000 #3
Message from syslogd@blister at Thu Apr 26 23:38:55 2007 ... blister kernel: [10688.920000] SMP
Message from syslogd@blister at Thu Apr 26 23:38:55 2007 ... blister kernel: [10688.920000] CPU: 0
Message from syslogd@blister at Thu Apr 26 23:38:55 2007 ... blister kernel: [10688.920000] EIP: 0060:[<f97164ca>] Tainted: P VLI
Message from syslogd@blister at Thu Apr 26 23:38:55 2007 ... blister kernel: [10688.920000] EFLAGS: 00210002 (2.6.20-15-generic #2)
Message from syslogd@blister at Thu Apr 26 23:38:55 2007 ... blister kernel: [10688.920000] EIP is at 0xf97164ca
Message from syslogd@blister at Thu Apr 26 23:38:55 2007 ... blister kernel: [10688.920000] eax: 00001000 ebx: 00001000 ecx: c0000080 edx: 00000000
Message from syslogd@blister at Thu Apr 26 23:38:55 2007 ... blister kernel: [10688.920000] esi: 00000000 edi: b38b0000 ebp: c364bf94 esp: c364bec4
Message from syslogd@blister at Thu Apr 26 23:38:55 2007 ... blister kernel: [10688.920000] ds: 007b es: 007b ss: 0068
Message from syslogd@blister at Thu Apr 26 23:38:55 2007 ... blister kernel: [10688.920000] Process VirtualBox (pid: 8189, ti=c364a000 task=f6c3da90 task.ti=c364a000)
Message from syslogd@blister at Thu Apr 26 23:38:55 2007 ... blister kernel: [10688.920000] Stack: 01000020 69746e65 c364bedc c011e341 00000000 00000000 00200286 00000000
Message from syslogd@blister at Thu Apr 26 23:38:55 2007 ... blister kernel: [10688.920000] 00000000 f3dd6c98 fffffffd d9903de1 00000000 00000000 00000000 00000000
Message from syslogd@blister at Thu Apr 26 23:38:55 2007 ... blister kernel: [10688.920000] 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
Message from syslogd@blister at Thu Apr 26 23:38:55 2007 ... blister kernel: [10688.920000] Call Trace:
Message from syslogd@blister at Thu Apr 26 23:38:55 2007 ... blister kernel: [10688.920000] [activate_task+33/64] activate_task+0x21/0x40
Message from syslogd@blister at Thu Apr 26 23:38:55 2007 ... blister kernel: [10688.920000] [work_resched+5/35] work_resched+0x5/0x23
Message from syslogd@blister at Thu Apr 26 23:38:55 2007 ... blister kernel: [10688.920000] =======================
Message from syslogd@blister at Thu Apr 26 23:38:55 2007 ... blister kernel: [10688.920000] Code: 00 00 89 55 cc 31 d2 09 c2 89 85 40 ff ff ff c7 85 44 ff ff ff 00 00 00 00 75 12 8b 5d c8 8b b5 4c ff ff ff 80 cf 10 89 d8 89 f2 <0f> 30 8b 85 50 ff ff ff 8b 95 4c ff ff ff c7 85 3c ff ff ff 00
Message from syslogd@blister at Thu Apr 26 23:38:55 2007 ... blister kernel: [10688.920000] EIP: [<f97164ca>] 0xf97164ca SS:ESP 0068:c364bec4
rojo@blister:~$ cat /proc/cpuinfo
processor : 0
vendor_id : AuthenticAMD
cpu family : 15
model : 75
model name : AMD Athlon(tm) 64 X2 Dual Core Processor 4200+
stepping : 2
cpu MHz : 1000.000
cache size : 512 KB
physical id : 0
siblings : 2
core id : 0
cpu cores : 2
fdiv_bug : no
hlt_bug : no
f00f_bug : no
coma_bug : no
fpu : yes
fpu_exception : yes
cpuid level : 1
wp : yes
flags : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt lm 3dnowext 3dnow pni cx16 lahf_lm cmp_legacy svm cr8legacy ts fid vid ttp tm stc
bogomips : 2011.95
clflush size : 64
processor : 1
vendor_id : AuthenticAMD
cpu family : 15
model : 75
model name : AMD Athlon(tm) 64 X2 Dual Core Processor 4200+
stepping : 2
cpu MHz : 1000.000
cache size : 512 KB
physical id : 0
siblings : 2
core id : 1
cpu cores : 2
fdiv_bug : no
hlt_bug : no
f00f_bug : no
coma_bug : no
fpu : yes
fpu_exception : yes
cpuid level : 1
wp : yes
flags : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt lm 3dnowext 3dnow pni cx16 lahf_lm cmp_legacy svm cr8legacy ts fid vid ttp tm stc
bogomips : 2011.95
clflush size : 64
Change History (19)
comment:1 by , 18 years ago
comment:2 by , 18 years ago
Checked out from svn tonight and rebuilt. Same symptom -- freezes when VM is powered on at 0%. Stack regurgitated into the console as follows:
rojo@blister:/usr/src/vbox/out/linux.x86/release/bin$ LD_LIBRARY_PATH=. ./Virtua lBox Message from syslogd@blister at Thu May 10 23:53:47 2007 ... blister kernel: [113177.968000] SMP Message from syslogd@blister at Thu May 10 23:53:47 2007 ... blister kernel: [113177.968000] CPU: 1 Message from syslogd@blister at Thu May 10 23:53:47 2007 ... blister kernel: [113177.968000] EIP: 0060:[<f90c68ca>] Tainted: P VLI Message from syslogd@blister at Thu May 10 23:53:47 2007 ... blister kernel: [113177.968000] EFLAGS: 00210002 (2.6.20-15-generic #2) Message from syslogd@blister at Thu May 10 23:53:47 2007 ... blister kernel: [113177.968000] EIP is at 0xf90c68ca Message from syslogd@blister at Thu May 10 23:53:47 2007 ... blister kernel: [113177.968000] eax: 00001000 ebx: 00001000 ecx: c0000080 edx: 00000000 Message from syslogd@blister at Thu May 10 23:53:47 2007 ... blister kernel: [113177.968000] esi: 00000000 edi: b414d000 ebp: f62e7f74 esp: f62e7ea4 Message from syslogd@blister at Thu May 10 23:53:47 2007 ... blister kernel: [113177.968000] ds: 007b es: 007b ss: 0068 Message from syslogd@blister at Thu May 10 23:53:47 2007 ... blister kernel: [113177.968000] Process VirtualBox (pid: 14681, ti=f62e6000 task =f52cba90 task.ti=f62e6000) Message from syslogd@blister at Thu May 10 23:53:47 2007 ... blister kernel: [113177.968000] Stack: 01000020 69746e65 00000000 00000001 00000 000 00000000 00000001 00000000 Message from syslogd@blister at Thu May 10 23:53:47 2007 ... blister kernel: [113177.968000] 00000000 c0112eb8 00000000 c15de3e0 00000 000 00000000 00000000 00000000 Message from syslogd@blister at Thu May 10 23:53:47 2007 ... blister kernel: [113177.968000] 00000000 00000000 00000000 00000000 00000 000 00000000 00000000 00000000 Message from syslogd@blister at Thu May 10 23:53:47 2007 ... blister kernel: [113177.968000] Call Trace: Message from syslogd@blister at Thu May 10 23:53:47 2007 ... blister kernel: [113177.968000] [do_flush_tlb_all+24/80] do_flush_tlb_all+0x18/ 0x50 Message from syslogd@blister at Thu May 10 23:53:47 2007 ... blister kernel: [113177.968000] [__switch_to+156/496] __switch_to+0x9c/0x1f0 Message from syslogd@blister at Thu May 10 23:53:47 2007 ... blister kernel: [113177.968000] [vfs_ioctl+92/672] vfs_ioctl+0x5c/0x2a0 Message from syslogd@blister at Thu May 10 23:53:47 2007 ... blister kernel: [113177.968000] [sys_ioctl+125/144] sys_ioctl+0x7d/0x90 Message from syslogd@blister at Thu May 10 23:53:47 2007 ... blister kernel: [113177.968000] ======================= Message from syslogd@blister at Thu May 10 23:53:47 2007 ... blister kernel: [113177.968000] Code: 00 00 89 55 cc 31 d2 09 c2 89 85 40 ff ff ff c7 85 44 ff ff ff 00 00 00 00 75 12 8b 5d c8 8b b5 4c ff ff ff 80 cf 10 89 d8 89 f2 <0f> 30 8b 85 50 ff ff ff 8b 95 4c ff ff ff c7 85 3c ff ff ff 00 Message from syslogd@blister at Thu May 10 23:53:47 2007 ... blister kernel: [113177.968000] EIP: [<f90c68ca>] 0xf90c68ca SS:ESP 0068:f62e7ea 4
comment:3 by , 18 years ago
Another similar report: http://forums.virtualbox.org/viewtopic.php?p=645#645
comment:4 by , 18 years ago
I have the same problem. I'm using Ubuntu Feisty (7.04) and when I try to start my Windows XP in order to install it, the VirtualBox windows freezes black and the progress bar of starting the Virtual Machine freezes at 0%.
Then I must reboot because after waiting hours all stills freeze.
Is there a solution for that?
comment:6 by , 18 years ago
Replying to sandervl73:
Try again with 1.4.0.
1.4.0 has the same symptom. VM freezes at "Starting the virtual machine..." at 0%.
comment:7 by , 18 years ago
I tried it with 1.4.0 too - VirtualBox still freezes!!!
I've replaced the AMD Athlon 64 X2 3800+ (the 65W variant) with a Sempron 3600+ and VirtualBox works (same Host OS and Mainboard!). So this definitly seems to be an issue with the AMD Athlon 64 X2 3800+ CPU. (I'm using the more energy efficient 65W variant of this CPU)
follow-up: 10 comment:9 by , 18 years ago
rojo, do you have the kvm module loaded or is kvm compiled into your host Linux kernel? Try to remove the module.
comment:10 by , 18 years ago
Replying to frank:
rojo, do you have the kvm module loaded or is kvm compiled into your host Linux kernel? Try to remove the module.
Same effect whether kvm is loaded or not. 0_o
comment:11 by , 18 years ago
Hi all. Had the same problem as described. Removed the kvm_intel and kvm modules and it appears to be working fine...
comment:12 by , 18 years ago
| Summary: | VM freezes immediately on power on → KVM & VBox conflict (Linux Host) |
|---|
follow-up: 14 comment:13 by , 18 years ago
According to rumors:
The real problem is that VT-instructions have no way to know, whenever other app uses VT.
If two programs access VT CPU at the same time, the host OS will crash.
-Technologov
comment:14 by , 18 years ago
Replying to Technologov:
According to rumors:
The real problem is that VT-instructions have no way to know, whenever other app uses VT.
If two programs access VT CPU at the same time, the host OS will crash.
Untrue. First of all this defect is not about two programs using VT at the same time. Second if people read the Intel manuals properly, we wouldn't have this problem in the first place.
I've changed VBox to detect whether somebody (in this case KVM) has left the CPU in VMX root mode. If so, VBox will simply refuse to run. Obviously the KVM guys should fix their code, but this will at least prevent VBox from crashing the host.
comment:15 by , 18 years ago
| Summary: | KVM & VBox conflict (Linux Host) → KVM & VBox conflict (Linux Host) -> fixed in SVN/1.4.2 |
|---|
comment:16 by , 18 years ago
sandervl73 wrote:
Obviously the KVM guys should fix their code, but this will at least prevent VBox from crashing the host.
Well, you can email the main KVM developer: "avi at qumranet dot com". I don't understand the VMX deeply enough to explain him the problem.
VBox will simply refuse to run.
I think a better approach is to fallback to software virtualization. What do you think ?
-Technologov
comment:17 by , 18 years ago
If by software virtualization you mean the standard way VBox virtualizes the guest, then you are wrong, because that's exactly when the problem occurs.
Falling back to QEmu like speed would only make our product look bad. Better to force the user to get rid of KVM.
comment:18 by , 18 years ago
I meant the standard way for VBox, kernel acceleration on, but without VMX.
So ?!
Is it possible to run VBox and KVM side by side at the same time ?
-Technologov
comment:19 by , 18 years ago
| Resolution: | → fixed |
|---|---|
| Status: | new → closed |
Without the noise:
The above process isn't VBox from what I can tell and it looks nothing like the VirtualBox crash below.
Executing "00 00 89 55 cc..." looks kind of like we're executing the wrong code.
You mention you needed bios fix for HEP to make it work with some other product, AMD X2 and VirtualBox 1.3.8 isn't 100% stable because of TSC drift between the two cores (results in bad timing). That's a known problem. However, I cannot really tell this is what's going wrong here or not. Since we've fixed all known AMD X2 problems, building and trying the current code (from subversion) would help finding out if this is a new problem or not.