Hi there.
Under some circumstances Windows HVM with vcpu>=2 stuck with vcpu (one or
all) runstate offline.
Next parts of trace file shows that vcpu start yielding the pcpu after
PAUSE_INSTRUCTION vmexit reason and finally runstate_change to offline
without waking back.
]178.854695610 |||-|||||||-|----|--|||x d808v1 mmio_assist r gpa fee00300
data 2f
[vla] d808v1 icr status idle
]178.854705889 ||--|||||||-|----||-|||x d808v1 vmentry cycles 24664 !
]178.854707083 ||--|||||||-|----||-|||x d808v1 vmexit exit_reason
APIC_ACCESS eip fffff80001810f82
]178.854707083 ||--|||||||-|----||-|||x d808v1 mmio_assist w gpa fee00310
data 1000000
]178.854713398 ||--|||||||-|- --||-|||x d808v1 vmentry cycles 15152 !
]178.854714458 ||--|||||||-|-|--||-|||x d808v1 vmexit exit_reason
APIC_ACCESS eip fffff80001810f89
]178.854714458 ||--|||||||-|-|--||-|||x d808v1 mmio_assist w gpa fee00300
data 8e1
[vla] d808v1 icr vec 225 dest_field
]178.854720294 |||-|||||||---|--||||||x d808v1 vmentry cycles 14004 !
]178.854723805 |||-|||||||-------|||||x d808v1 vmexit exit_reason
PAUSE_INSTRUCTION eip fffff8000190157f
]178.854724469 |||-|||||||--------||||x d808v1 28005(2:8:5) 2 [ 328 1 ]
178.854726237 |||-|||||||--------||||x d808v1 runstate_continue d808v1
running->running
]178.854727041 |||-|||||||--------||||x d808v1 vmentry cycles 7764
]178.854729877 |||-|||||||--------||||x d808v1 vmexit exit_reason
PAUSE_INSTRUCTION eip fffff8000190157f
]178.854730157 |||-|||||||--------||||x d808v1 28005(2:8:5) 2 [ 328 1 ]
178.854731035 |||-|||||||--------||||x d808v1 runstate_continue d808v1
running->running
...
<huge amount of PAUSE_INSTRUCTION/28005>
...
]178.857438583 |||--------------|---|-x d808v1 vmentry cycles 1972
]178.857440840 |||--------------|---|-x d808v1 vmexit exit_reason
PAUSE_INSTRUCTION eip fffff8000190157f
]178.857441010 |||--------------|---|-x d808v1 28005(2:8:5) 2 [ 328 1 ]
178.857441405 |||--------------|---|-x d808v1 runstate_continue d808v1
running->running
]178.857441634 |||--------------|---|-x d808v1 vmentry cycles 1904
]178.857443891 |||--------------|---|-x d808v1 vmexit exit_reason
PAUSE_INSTRUCTION eip fffff8000190157f
]178.857444064 |||--------------|---|-x d808v1 28005(2:8:5) 2 [ 328 1 ]
178.857444458 |||--------------|---|-x d808v1 runstate_continue d808v1
running->running
]178.857444686 |||--------------|---|-x d808v1 vmentry cycles 1908
]178.857446940 |||--------------|---|-x d808v1 vmexit exit_reason
PAUSE_INSTRUCTION eip fffff8000190157f
]178.857447253 |||--------------|---|-x d808v1 28005(2:8:5) 2 [ 328 1 ]
]178.857449979 |||--------------|---|-x d808v1 2800e(2:8:e) 2 [ 328
1e1278 ]
]178.857450122 |||--------------|---|-x d808v1 2800f(2:8:f) 3 [ 7fff
82fb0c ffffffff ]
]178.857450264 |||--------------|---|-x d808v1 2800a(2:8:a) 4 [ 328 1
7fff 17 ]
178.857450394 |||--------------|---|-x d808v1 runstate_change d808v1
running->offline
process_lost_records: setting d808v1 to RUNSTATE_LOST
[w2h] Clearing w2h state for d808v1
process_lost_records: setting d808v1 to RUNSTATE_LOST
[w2h] Clearing w2h state for d808v1
process_lost_records: setting d808v1 to RUNSTATE_LOST
[w2h] Clearing w2h state for d808v1
process_lost_records: setting d808v1 to RUNSTATE_LOST
[w2h] Clearing w2h state for d808v1
process_lost_records: setting d808v1 to RUNSTATE_LOST
....
d808v0 stuck a little bit later:
]179.737341628 -|------------|x|-|---|- d808v0 vmexit exit_reason
PAUSE_INSTRUCTION eip fffff800018d4dde
]179.737341876 -|------------|x|-|---|- d808v0 28005(2:8:5) 2 [ 328 0 ]
179.737343055 -|------------|x|-|---|- d808v0 runstate_continue d808v0
running->running
]179.737343785 -|------------|x|-|---|- d808v0 vmentry cycles 5176
]179.737347024 -|------------|x|-|---|- d808v0 vmexit exit_reason
PAUSE_INSTRUCTION eip fffff800018d4dde
]179.737347306 -|------------|x|-|---|- d808v0 28005(2:8:5) 2 [ 328 0 ]
179.737348500 -|------------|x|-|---|- d808v0 runstate_continue d808v0
running->running
]179.737349432 -|------------|x|-|---|- d808v0 vmentry cycles 5776
]179.737352022 -|------------|x|-|---|- d808v0 vmexit exit_reason
PAUSE_INSTRUCTION eip fffff800018d4dde
]179.737354408 -|------------|x|-|---|- d808v0 28005(2:8:5) 2 [ 328 0 ]
]179.737356033 -|------------|x|-|---|- d808v0 2800e(2:8:e) 2 [ 328
1c6e79 ]
]179.737356177 -|------------|x|-|---|- d808v0 2800f(2:8:f) 3 [ 7fff
1caa98 ffffffff ]
]179.737356367 -|------------|x|-|---|- d808v0 2800a(2:8:a) 4 [ 328 0
7fff f ]
179.737356528 -|------------|x|-|---|- d808v0 runstate_change d808v0
running->offline
process_lost_records: setting d808v0 to RUNSTATE_LOST
[w2h] Clearing w2h state for d808v0
process_lost_records: setting d808v0 to RUNSTATE_LOST
[w2h] Clearing w2h state for d808v0
...
and that''s it - no more information about domid 808.
Is this patch
http://lists.xen.org/archives/html/xen-devel/2011-12/msg01884.html can be
suitable in my situation or maybe some more info needed to solve this issue.
Thanks!
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xen.org
http://lists.xen.org/xen-devel