Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1755601Ab0DVO5w (ORCPT ); Thu, 22 Apr 2010 10:57:52 -0400 Received: from e6.ny.us.ibm.com ([32.97.182.146]:40436 "EHLO e6.ny.us.ibm.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1755568Ab0DVO5r (ORCPT ); Thu, 22 Apr 2010 10:57:47 -0400 Message-ID: <4BD063E4.5030400@us.ibm.com> Date: Thu, 22 Apr 2010 07:57:40 -0700 From: Darren Hart User-Agent: Thunderbird 2.0.0.24 (X11/20100411) MIME-Version: 1.0 To: Yong Zhang CC: "lkml, " , Peter Zijlstra , Ingo Molnar Subject: Re: Lockdep usage References: <4BCFEBFF.4090505@us.ibm.com> <20100422070625.GA17515@windriver.com> In-Reply-To: <20100422070625.GA17515@windriver.com> Content-Type: text/plain; charset=UTF-8; format=flowed Content-Transfer-Encoding: 7bit Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 4059 Lines: 105 Yong Zhang wrote: > On Wed, Apr 21, 2010 at 11:26:07PM -0700, Darren Hart wrote: >> I'm having some trouble with lockstat. I've enabled >> CONFIG_LOCK_STATS and then run the following commands to collect the >> lock statistics for the futex_lock test case. I expect this test >> case to generate a fair amount of contention on the futex hash >> bucket locks, but they don't show up in the stats (in fact, nothing >> shows up in the stats). I have seen much longer lock_stat files when >> I forgot to clear them before the test, but no hb locks showed up >> there either. >> >> Is my usage below incorrect? Do I have to somehow annotate the >> hb->lock in order for lockstat to track it? > > IMHO, you don't need to do that. > >> lock.lockstat; echo 0 > /proc/sys/kernel/lock_stat >> futex_lock: Measure FUTEX_LOCK operations per second >> Arguments: iterations=10000000 threads=256 adaptive=no >> period=1000 duty-cycle=20% >> Result: 731 Kiter/s >> >> # cat lock.lockstat >> lock_stat version 0.3 >> *WARNING* lock debugging disabled!! - possibly due to a lockdep warning > > This means lockdep has detect there is something wrong. > Can you show your 'dmesg'? dmesg includes the following during bootup, but nothing new while running the tests: BUG: key ffff88024e657858 not in .data! ------------[ cut here ]------------ WARNING: at kernel/lockdep.c:2706 lockdep_init_map+0x226/0x5c0() Hardware name: BladeCenter LS21 -[7971AC1]- Pid: 1, comm: swapper Not tainted 2.6.34-rc1dvhv4ls #32 Call Trace: [] ? lockdep_init_map+0x226/0x5c0 [] warn_slowpath_common+0x87/0xb0 [] warn_slowpath_null+0x14/0x20 [] lockdep_init_map+0x226/0x5c0 [] ? lock_release_holdtime+0x83/0xf0 [] sysfs_add_file_mode+0x71/0xd0 [] sysfs_add_file+0x11/0x20 [] sysfs_create_file+0x30/0x40 [] device_create_file+0x19/0x20 [] ipmi_register_smi+0xba5/0xfc0 [] ? put_device+0x17/0x20 [] ? autoremove_wake_function+0x0/0x40 [] try_smi_init+0x5d2/0xa20 [] init_ipmi_si+0x432/0xb80 [] ? init_ipmi_si+0x0/0xb80 [] do_one_initcall+0x3d/0x180 [] kernel_init+0x16c/0x200 [] kernel_thread_helper+0x4/0x10 [] ? kernel_init+0x0/0x200 [] ? kernel_thread_helper+0x0/0x10 ---[ end trace 30838b0dc5ec48e4 ]--- Several more "BUG: key fffXXXXXXXXXXXX not in .data!" lines show up throughout the boot process. I vaguely recall this being normal for lockstat at initial boot. -- Darren > > Thanks, > Yong > >> ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- >> class name con-bounces >> contentions waittime-min waittime-max waittime-total >> acq-bounces acquisitions holdtime-min holdtime-max >> holdtime-total >> ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- >> >> # >> >> Thanks, >> >> -- >> Darren Hart >> IBM Linux Technology Center >> Real-Time Linux Team >> -- >> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in >> the body of a message to majordomo@vger.kernel.org >> More majordomo info at http://vger.kernel.org/majordomo-info.html >> Please read the FAQ at http://www.tux.org/lkml/ -- Darren Hart IBM Linux Technology Center Real-Time Linux Team -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/