I report a benchmark result of last week on IPF using
xen-unstable and linux-2.6.18-xen. # not ia64

- Building xen failed on ia64.
  This issue has already fixed by 19957:1ebe34ccd271.

- The hypervisor hung up while booting.
  I reverted 19913:ef38784f9f85 and 19914:d6c1d7992f43.

- Xend did not run on Python 2.3.
  I sent patches (19958:ad5a697d701a and 19959:69cd50c86a79) 
  and removed codes that Python 2.3 do not support.

- DomVTi cannot use tap:aio. When a guest started, dom0 hit BUG_ON()
  at mm/rmap.c:522 and the guest hung up. Thus we used vbd.
  I attached the dom0 call trace.
  # Notice that DomU can use tap:aio.

- On domU, syslog09 failed, but passed when tested again.

- The hypervisor crashed while Unixbench was running on DomVTi.
  (XEN) ****************************************
  (XEN) Panic on CPU 0:
  (XEN) PREEMPT happen in multicall
  (XEN) ****************************************

    Machine          : Tiger4
    Kernel           :
    Changeset        : 19946:91407452cdb6 (xen-unstable)
                       917:b420e936c022   (linux-2.6.18-xen)
                       137:3ad73b4314e3   (efi-vfirmware)
                       5cc34ea            (qemu-xen-unstable)
    Dom0 OS          : RHEL4 U2 (2P)
    DomU OS          : RHEL4 U2 (8P, using tap:aio)
    DomVTi OS        : RHEL4 U2 (8P, with PV-on-HVM drivers)

    unixbench4.1.0    : Pass
    bonnie++-1.03     : Pass
    ltp-full-20070930 : 1/835 FAIL (syslog09)
    iozone3_191       : Pass
    lmbench-3.0-a5    : Pass
  DomVTi: # file:
    unixbench4.1.0    : CRASH (hypervisor panic)
    bonnie++-1.03     : Pass
    ltp-full-20070930 : N/A
    iozone3_191       : N/A
    lmbench-3.0-a5    : N/A

Best regards,
KUWAMURA Shin'ya and FUJITSU members

Xen-ia64-devel mailing list

Reply via email to