Krysan, Susan
2006-Aug-14 19:33 UTC
[Xen-devel] fatal page fault crash during xm-test local live migration when host has >32GB RAM
Running xm-test with xen-unstable changeset 10949 on Unisys ES7000/one, x86_64, with 32 physical processors and 128 GB RAM results in a fatal page fault crash during the migration test. Bug 729 has been opened. http://bugzilla.xensource.com/bugzilla/show_bug.cgi?id=729 Tail of serial console output shows: Xen call trace: (XEN) [<ffff83000013be35>] free_shadow_page+0x165/0xba0 (XEN) [<ffff83000013d34f>] free_shadow_pages+0x23f/0x2d0 (XEN) [<ffff83000013de3b>] shadow_mode_control+0x18b/0x500 (XEN) [<ffff83000011b791>] arch_do_dom0_op+0x101/0xc50 (XEN) [<ffff830000120a1a>] apic_timer_interrupt+0x2a/0x30 (XEN) [<ffff830000120a1a>] apic_timer_interrupt+0x2a/0x30 (XEN) [<ffff830000106b3c>] do_dom0_op+0xe1c/0xf10 (XEN) [<ffff830000111b63>] csched_vcpu_wake+0x213/0x220 (XEN) [<ffff8300001089f0>] evtchn_set_pending+0x80/0x110 (XEN) [<ffff830000108cb4>] evtchn_send+0xf4/0x110 (XEN) [<ffff830000109531>] do_event_channel_op+0x861/0xc90 (XEN) [<ffff8300001696cb>] do_iret+0x7b/0x140 (XEN) [<ffff830000168702>] syscall_enter+0x62/0x67 (XEN) (XEN) Pagetable walk from 0000000000000000: (XEN) L4 = 0000000056e7b067 0000000001f8b67b (XEN) L3 = 00000001e88b5067 0000000001f8ec0a (XEN) L2 = 0000000000000000 ffffffffffffffff (XEN) (XEN) **************************************** (XEN) Panic on CPU 4: (XEN) CPU4 FATAL PAGE FAULT (XEN) [error_code=0000] (XEN) Faulting linear address: 0000000000000000 (XEN) **************************************** (XEN) (XEN) Reboot in five seconds... When the host has 64 GB RAM, we still get a crash. Only when we booted the host with 32GB RAM does the localhost live migration work. Does anyone have an idea about what might be causing the crash? Thanks, Sue Krysan Linux Systems Group Unisys Corporation _______________________________________________ Xen-devel mailing list Xen-devel@lists.xensource.com http://lists.xensource.com/xen-devel