Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1752348AbZK1Prg (ORCPT ); Sat, 28 Nov 2009 10:47:36 -0500 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1752120AbZK1Prg (ORCPT ); Sat, 28 Nov 2009 10:47:36 -0500 Received: from mta3.srv.hcvlny.cv.net ([167.206.4.198]:51400 "EHLO mta3.srv.hcvlny.cv.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751539AbZK1Prf (ORCPT ); Sat, 28 Nov 2009 10:47:35 -0500 Date: Sat, 28 Nov 2009 10:47:38 -0500 From: Michael Breuer Subject: Re: Problem? intel_iommu=off; perf top shows acpi_os_read_port as extremely busy In-reply-to: <20091128071808.GA32183@elte.hu> To: Ingo Molnar Cc: Peter Zijlstra , Thomas Gleixner , Len Brown , Arjan van de Ven , Arnaldo Carvalho de Melo , linux-kernel@vger.kernel.org Message-id: <4B11461A.3020408@majjas.com> MIME-version: 1.0 Content-type: text/plain; charset=ISO-8859-1; format=flowed Content-transfer-encoding: 7BIT References: <4B106CC0.4050406@majjas.com> <20091128071808.GA32183@elte.hu> User-Agent: Thunderbird 2.0.0.23 (Windows/20090812) Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 5346 Lines: 138 Ok - did the following in runlevel 3 to avoid the dmar errors I'm getting with nouveau & vt-d. In theory, the system was similarly loaded (i.e., doing pretty much nothing) for both runs. The sample is consistent with what I've seen previously. Perhaps there's no issue, or perhaps the issue is with my broken bios and intel_iommu=on. Perf top with intel_iommu=off: (snapshop) - acpi_os_read_port is often #1 and I've seen it over 30%. ------------------------------------------------------------------------------ PerfTop: 3957 irqs/sec kernel:84.0% [100000 cycles], (all, 8 CPUs) ------------------------------------------------------------------------------ samples pcnt kernel function _______ _____ _______________ 3183.00 - 16.7% : _spin_lock 3167.00 - 16.7% : acpi_os_read_port 1053.00 - 5.5% : io_apic_modify_irq 810.00 - 4.3% : hpet_next_event 529.00 - 2.8% : _spin_lock_irqsave 522.00 - 2.7% : io_apic_sync 283.00 - 1.5% : tg_shares_up 270.00 - 1.4% : acpi_idle_enter_bm 259.00 - 1.4% : irq_to_desc 222.00 - 1.2% : i8042_interrupt 213.00 - 1.1% : acpi_hw_validate_io_request 204.00 - 1.1% : ktime_get 180.00 - 0.9% : find_busiest_group 169.00 - 0.9% : _spin_unlock_irqrestore 168.00 - 0.9% : sub_preempt_count Performance counter stats for 'sleep 1' (10 runs): 8021.581362 task-clock-msecs # 8.009 CPUs ( +- 0.033% ) 607 context-switches # 0.000 M/sec ( +- 4.251% ) 27 CPU-migrations # 0.000 M/sec ( +- 11.455% ) 408 page-faults # 0.000 M/sec ( +- 34.557% ) 311405638 cycles # 38.821 M/sec ( +- 6.887% ) 85807775 instructions # 0.276 IPC ( +- 13.824% ) 2300079 cache-references # 0.287 M/sec ( +- 6.859% ) 77314 cache-misses # 0.010 M/sec ( +- 11.184% ) 1.001616593 seconds time elapsed ( +- 0.009% ) intel_iommu on: ------------------------------------------------------------------------------ PerfTop: 9941 irqs/sec kernel:81.9% [100000 cycles], (all, 8 CPUs) ------------------------------------------------------------------------------ samples pcnt kernel function _______ _____ _______________ 11465.00 - 20.8% : _spin_lock 3679.00 - 6.7% : io_apic_modify_irq 3295.00 - 6.0% : hpet_next_event 2172.00 - 3.9% : _spin_lock_irqsave 2111.00 - 3.8% : acpi_os_read_port 1094.00 - 2.0% : io_apic_sync 904.00 - 1.6% : find_busiest_group 695.00 - 1.3% : _spin_unlock_irqrestore 686.00 - 1.2% : tg_shares_up 620.00 - 1.1% : acpi_idle_enter_bm 577.00 - 1.0% : add_preempt_count 568.00 - 1.0% : sub_preempt_count 475.00 - 0.9% : audit_filter_syscall 470.00 - 0.9% : schedule 450.00 - 0.8% : tick_nohz_stop_sched_tick Performance counter stats for 'sleep 1' (10 runs): 8015.967731 task-clock-msecs # 8.003 CPUs ( +- 0.024% ) 2628 context-switches # 0.000 M/sec ( +- 20.053% ) 124 CPU-migrations # 0.000 M/sec ( +- 20.561% ) 3014 page-faults # 0.000 M/sec ( +- 35.573% ) 850702031 cycles # 106.126 M/sec ( +- 10.601% ) 311032631 instructions # 0.366 IPC ( +- 17.859% ) 8578386 cache-references # 1.070 M/sec ( +- 13.894% ) 333768 cache-misses # 0.042 M/sec ( +- 21.894% ) 1.001656333 seconds time elapsed ( +- 0.008% ) Ingo Molnar wrote: > * Michael Breuer wrote: > > >> Having given up for now on VT-D, I rebooted 2.6.38 rc8 with >> intel_iommu=off. Whilst my myriad of broken bios issues cleared, I now >> see in perf top acpi_os_read_port as continually the busiest function. >> With intel_iommu enabled, _spin_lock was always on top, and nothing >> else was notable. >> >> This seems odd to me, perhaps this will make sense to someone else. >> >> FWIW, I'm running on an Asus p6t deluxe v2; ht enabled; no errors or >> oddities in dmesg or /var/log/messages. >> > > Could you post the perf top output please? > > Also, could you also post the output of: > > perf stat -a --repeat 10 sleep 1 > > this will show us how idle the system is. (My guess is that your system > is idle and perf top shows acpi_os_read_port because the system goes to > idle via ACPI methods and PIO is slow. In that case all is nominal and > your system is fine. But it's hard to tell without more details.) > > Thanks, > > Ingo > -- 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/