guest hanging on rc1
From: Borislav Petkov
Date: Mon Dec 22 2014 - 09:34:29 EST
Hi Paolo,
so I installed an old SUSE guest (SLES10, kernel is 2.6.16 + enterprise
stuff) and it was booting and all was fine but this week not anymore.
Host kernel is 3.19-rc1 + tip/master. I did miss some kvm config options
initially so I did
$ make kvmconfig
and it added those (see diff at the end).
Now, when I go to the monitor and do
(qemu) p $eip
0xffffffff80120e7e
This address is in kvm_register_clock() on the guest kernel:
ffffffff80120dd2 <kvm_register_clock>:
ffffffff80120dd2: 55 push %rbp
ffffffff80120dd3: 53 push %rbx
ffffffff80120dd4: 48 83 ec 08 sub $0x8,%rsp
ffffffff80120dd8: 65 8b 34 25 24 00 00 mov %gs:0x24,%esi
ffffffff80120ddf: 00
ffffffff80120de0: 31 c0 xor %eax,%eax
ffffffff80120de2: 83 3d 7f a8 33 00 00 cmpl $0x0,0x33a87f(%rip) # ffffffff8045b668 <use_kvm_time>
ffffffff80120de9: 0f 84 91 00 00 00 je ffffffff80120e80 <kvm_register_clock+0xae>
ffffffff80120def: 48 63 d6 movslq %esi,%rdx
ffffffff80120df2: 48 c7 c0 80 04 43 80 mov $0xffffffff80430480,%rax
ffffffff80120df9: 48 8b 14 d5 20 42 3f mov -0x7fc0bde0(,%rdx,8),%rdx
ffffffff80120e00: 80
ffffffff80120e01: 4c 8b 42 08 mov 0x8(%rdx),%r8
ffffffff80120e05: 4a 8d 14 00 lea (%rax,%r8,1),%rdx
ffffffff80120e09: 48 b8 ff ff ff 7f ff movabs $0xffffffff7fffffff,%rax
ffffffff80120e10: ff ff ff
ffffffff80120e13: 48 39 c2 cmp %rax,%rdx
ffffffff80120e16: 8d 8a 00 00 00 80 lea -0x80000000(%rdx),%ecx
ffffffff80120e1c: 77 02 ja ffffffff80120e20 <kvm_register_clock+0x4e>
ffffffff80120e1e: 89 d1 mov %edx,%ecx
ffffffff80120e20: 48 c7 c0 80 04 43 80 mov $0xffffffff80430480,%rax
ffffffff80120e27: 83 c9 01 or $0x1,%ecx
ffffffff80120e2a: 4a 8d 14 00 lea (%rax,%r8,1),%rdx
ffffffff80120e2e: 48 b8 ff ff ff 7f ff movabs $0xffffffff7fffffff,%rax
ffffffff80120e35: ff ff ff
ffffffff80120e38: 89 cd mov %ecx,%ebp
ffffffff80120e3a: 48 39 c2 cmp %rax,%rdx
ffffffff80120e3d: 76 07 jbe ffffffff80120e46 <kvm_register_clock+0x74>
ffffffff80120e3f: b8 00 00 00 80 mov $0x80000000,%eax
ffffffff80120e44: eb 0a jmp ffffffff80120e50 <kvm_register_clock+0x7e>
ffffffff80120e46: 48 b8 00 00 00 00 00 movabs $0x7f0000000000,%rax
ffffffff80120e4d: 7f 00 00
ffffffff80120e50: 48 8d 04 02 lea (%rdx,%rax,1),%rax
ffffffff80120e54: 48 89 c2 mov %rax,%rdx
ffffffff80120e57: 48 c1 ea 20 shr $0x20,%rdx
ffffffff80120e5b: 48 85 ff test %rdi,%rdi
ffffffff80120e5e: 89 d3 mov %edx,%ebx
ffffffff80120e60: 74 11 je ffffffff80120e73 <kvm_register_clock+0xa1>
ffffffff80120e62: 49 89 f8 mov %rdi,%r8
ffffffff80120e65: 31 c0 xor %eax,%eax
ffffffff80120e67: 48 c7 c7 f4 7a 31 80 mov $0xffffffff80317af4,%rdi
ffffffff80120e6e: e8 b7 4a 01 00 callq ffffffff8013592a <printk>
ffffffff80120e73: b9 12 00 00 00 mov $0x12,%ecx
ffffffff80120e78: 89 e8 mov %ebp,%eax
ffffffff80120e7a: 89 da mov %ebx,%edx
ffffffff80120e7c: 0f 30 wrmsr
ffffffff80120e7e: 31 c0 xor %eax,%eax <--- rIP
ffffffff80120e80: 5a pop %rdx
ffffffff80120e81: 5b pop %rbx
ffffffff80120e82: 5d pop %rbp
ffffffff80120e83: c3 retq
right after we've written the 0x12 MSR, i.e. MSR_KVM_SYSTEM_TIME.
So something broke somewhere and the guest won't boot anymore. Even if I
try to boot the iso image which I used to install it, its kernel hangs
there too. Not the same VA though:
(qemu) p $eip
0xffffffff8011b172
(qemu)
but the exact same place:
(qemu) x/20i 0xffffffff8011b150
0xffffffff8011b150: je 0x7fff8011b167
0xffffffff8011b152: mov %rdi,%r8
0xffffffff8011b155: mov %ebp,%ecx
0xffffffff8011b157: xor %esi,%esi
0xffffffff8011b159: mov $0xffffffff802ee1ac,%rdi
0xffffffff8011b160: xor %eax,%eax
0xffffffff8011b162: callq 0x7fff8012a562
0xffffffff8011b167: mov $0x12,%ecx
0xffffffff8011b16c: mov %ebp,%eax
0xffffffff8011b16e: mov %ebx,%edx
0xffffffff8011b170: wrmsr
0xffffffff8011b172: xor %eax,%eax <--- rIP
0xffffffff8011b174: pop %rdx
0xffffffff8011b175: pop %rbx
0xffffffff8011b176: pop %rbp
0xffffffff8011b177: retq
0xffffffff8011b178: push %rbp
0xffffffff8011b179: mov %edi,%ebp
0xffffffff8011b17b: push %rbx
0xffffffff8011b17c: sub $0x8,%rsp
Right after the MSR write.
Any ideas how to debug this further? :)
Thanks.
--- /boot/config-3.19.0-rc1+ 2014-12-22 11:20:13.015922953 +0100
+++ .config 2014-12-22 13:31:24.030189420 +0100
@@ -336,7 +336,15 @@ CONFIG_X86_FEATURE_NAMES=y
CONFIG_X86_SUPPORTS_MEMORY_FAILURE=y
CONFIG_SCHED_OMIT_FRAME_POINTER=y
# CONFIG_KVMTOOL_TEST_ENABLE is not set
-# CONFIG_HYPERVISOR_GUEST is not set
+CONFIG_HYPERVISOR_GUEST=y
+CONFIG_PARAVIRT=y
+# CONFIG_PARAVIRT_DEBUG is not set
+# CONFIG_PARAVIRT_SPINLOCKS is not set
+# CONFIG_XEN is not set
+CONFIG_KVM_GUEST=y
+# CONFIG_KVM_DEBUG_FS is not set
+# CONFIG_PARAVIRT_TIME_ACCOUNTING is not set
+CONFIG_PARAVIRT_CLOCK=y
CONFIG_NO_BOOTMEM=y
# CONFIG_MEMTEST is not set
CONFIG_MK8=y
@@ -1126,6 +1134,7 @@ CONFIG_VIRTIO_BLK=y
# CONFIG_SENSORS_APDS990X is not set
# CONFIG_HMC6352 is not set
# CONFIG_DS1682 is not set
+# CONFIG_VMWARE_BALLOON is not set
# CONFIG_BMP085_I2C is not set
# CONFIG_USB_SWITCH_FSA9480 is not set
# CONFIG_SRAM is not set
@@ -2708,6 +2717,7 @@ CONFIG_VIRTIO_PCI=y
#
# Microsoft Hyper-V guest support
#
+# CONFIG_HYPERV is not set
# CONFIG_STAGING is not set
# CONFIG_X86_PLATFORM_DEVICES is not set
# CONFIG_CHROME_PLATFORMS is not set
--
Regards/Gruss,
Boris.
Sent from a fat crate under my desk. Formatting is fine.
--
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/