Hi Stephen,

I got this warning when testing kernels in kvm:

[    0.000000] --------------------------------------------------------
[    0.000000] 133 out of 218 testcases failed, as expected. |
[    0.000000] ----------------------------------------------------
[    0.000000] ODEBUG: 8 of 8 active objects replaced
[    0.000000] ------------[ cut here ]------------
[    0.000000] WARNING: at /c/kernel-tests/src/x86/lib/debugobjects.c:908 check_results.constprop.9+0x129/0x153()
[    0.000000] Hardware name: Bochs
[    0.000000] ODEBUG: selftest warnings failed 4 != 5
[    0.000000] Pid: 0, comm: swapper/0 Not tainted 3.2.0-06292-g59a5680 #1
[    0.000000] Call Trace:
[    0.000000]  [<ffffffff810587ea>] warn_slowpath_common+0x7a/0xb0
[    0.000000]  [<ffffffff818ff01f>] ? check_results.constprop.9+0x35/0x153
[    0.000000]  [<ffffffff81058896>] warn_slowpath_fmt+0x46/0x50
[    0.000000]  [<ffffffff81435c81>] ? _raw_spin_lock_irqsave+0x61/0x70
[    0.000000]  [<ffffffff818ff01f>] ? check_results.constprop.9+0x35/0x153
[    0.000000]  [<ffffffff818ff113>] check_results.constprop.9+0x129/0x153
[    0.000000]  [<ffffffff818ff6d6>] debug_objects_mem_init+0x412/0x4a6
[    0.000000]  [<ffffffff818e0a15>] start_kernel+0x3c0/0x443
[    0.000000]  [<ffffffff818e0000>] ? __init_begin+0x14000/0x14000
[    0.000000]  [<ffffffff818e0156>] x86_64_start_reservations+0x131/0x135
[    0.000000]  [<ffffffff818e0259>] x86_64_start_kernel+0xff/0x10e
[    0.000000] ---[ end trace 4eaa2a86a8e2da22 ]---
[    0.000000] hpet clockevent registered
[    0.000000] Fast TSC calibration using PIT
[    0.000000] Detected 3092.385 MHz processor.
[    0.008004] Calibrating delay loop (skipped), value calculated using timer frequency.. 6184.77 BogoMIPS (lpj=12369540)

feac18dda25134005909e7770c77464e65608bd8 is the first bad commit

commit feac18dda25134005909e7770c77464e65608bd8
Author: Stephen Boyd <sboyd@codeaurora.org>
Date:   Mon Nov 7 19:48:26 2011 -0800

    debugobjects: Be smarter about static objects

    Make debugobjects use the return code from the fixup function. That
    allows us better diagnostics in the activate check than relying on a
    WARN_ON() in the object specific code.

    [ tglx@linutronix.de: Split out the debugobjects vs. the timer change ]

    Signed-off-by: Stephen Boyd <sboyd@codeaurora.org>
    Cc: Christine Chan <cschan@codeaurora.org>
    Cc: John Stultz <john.stultz@linaro.org>
    Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
    Link: http://lkml.kernel.org/r/1320724108-20788-2-git-send-email-sboyd@codeaurora.org
    Signed-off-by: Thomas Gleixner <tglx@linutronix.de>

Thanks,
Fengguang