[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <alpine.DEB.2.02.1410031542440.15421@chino.kir.corp.google.com>
Date: Fri, 3 Oct 2014 15:44:22 -0700 (PDT)
From: David Rientjes <rientjes@...gle.com>
To: Fengguang Wu <fengguang.wu@...el.com>
cc: Dave Hansen <dave.hansen@...el.com>,
Ingo Molnar <mingo@...nel.org>,
LKML <linux-kernel@...r.kernel.org>, lkp@...org
Subject: Re: [perf/x86/intel] WARNING: CPU: 48 PID: 7540 at
arch/x86/kernel/cpu/perf_event_intel_ds.c:316
reserve_ds_buffers+0x208/0x430()
On Fri, 3 Oct 2014, Fengguang Wu wrote:
> Hi David,
>
> FYI, we noticed the below changes on commit
>
> 4485154138f6ffa5b252cb490aba3e8eb30124e4 ("perf/x86/intel: Avoid spamming kernel log for BTS buffer failure")
>
> +---------------------------------------------------------------------------+------------+------------+
> | | 338b522ca4 | 4485154138 |
> +---------------------------------------------------------------------------+------------+------------+
> | boot_successes | 3 | 5 |
> | boot_failures | 21 | 15 |
> | page_allocation_failure:order:,mode | 11 | |
> | backtrace:reserve_ds_buffers | 11 | 5 |
> | backtrace:x86_pmu_event_init | 11 | 5 |
> | backtrace:perf_init_event | 11 | 5 |
> | backtrace:SYSC_perf_event_open | 11 | 5 |
> | backtrace:SyS_perf_event_open | 11 | 5 |
> | BUG:kernel_test_crashed | 10 | 10 |
> | WARNING:at_arch/x86/kernel/cpu/perf_event_intel_ds.c:reserve_ds_buffers() | 0 | 5 |
> +---------------------------------------------------------------------------+------------+------------+
>
>
> 2014-10-03 01:27:35 ./usemem --runtime 300 -t 64 -f /tmp/vm-scalability/sparse-mremap-xread-rand-mt-60 -E --prealloc --readonly --random 17179869184
> 2014-10-03 01:27:36 ./usemem --runtime 300 -t 64 -f /tmp/vm-scalability/sparse-mremap-xread-rand-mt-62 -E --prealloc --readonly --random 17179869184
> [ 192.496567] ------------[ cut here ]------------
> [ 192.505045] WARNING: CPU: 48 PID: 7540 at arch/x86/kernel/cpu/perf_event_intel_ds.c:316 reserve_ds_buffers+0x208/0x430()
> [ 192.524168] alloc_bts_buffer: BTS buffer allocation failure
> [ 192.533339] Modules linked in: loop ipmi_watchdog nfsd auth_rpcgss dm_mod fuse sg sd_mod crct10dif_generic sr_mod crc_t10dif cdrom crct10dif_common ata_generic pata_acpi mgag200 syscopyarea snd_pcm sysfillrect sysimgblt snd_timer ttm snd drm_kms_helper soundcore ata_piix pcspkr i7core_edac drm libata megaraid_sas i2c_i801 edac_core ipmi_si ipmi_msghandler acpi_cpufreq
> [ 192.581450] CPU: 48 PID: 7540 Comm: perf Not tainted 3.16.0-rc2-00011-g4485154 #1
> [ 192.592393] Hardware name: Quanta MP Server/QSSC-S4R, BIOS QSSC-S4R.QCI.01.00.0028.090220101116 09/02/2010
> [ 192.606455] 0000000000000009 ffff88005de53ce8 ffffffff81844e55 ffff88005de53d30
> [ 192.618370] ffff88005de53d20 ffffffff8106de6d 0000000000000000 0000000000010528
> [ 192.630786] 0000000000000000 0000000000000000 0000000000000000 ffff88005de53d80
> [ 192.642465] Call Trace:
> [ 192.647913] [<ffffffff81844e55>] dump_stack+0x4d/0x66
> [ 192.656643] [<ffffffff8106de6d>] warn_slowpath_common+0x7d/0xa0
> [ 192.666961] [<ffffffff8106dedc>] warn_slowpath_fmt+0x4c/0x50
> [ 192.677047] [<ffffffff81030863>] ? reserve_ds_buffers+0x2b3/0x430
> [ 192.687420] [<ffffffff810307b8>] reserve_ds_buffers+0x208/0x430
> [ 192.697838] [<ffffffff8118abec>] ? handle_mm_fault+0x8ac/0xf70
> [ 192.707522] [<ffffffff8102a354>] x86_pmu_event_init+0x434/0x450
> [ 192.717351] [<ffffffff8115241b>] perf_init_event+0xdb/0x140
> [ 192.727292] [<ffffffff81152808>] perf_event_alloc+0x388/0x440
> [ 192.738117] [<ffffffff81152c60>] SYSC_perf_event_open+0x3a0/0xd20
> [ 192.750511] [<ffffffff81153979>] SyS_perf_event_open+0x9/0x10
> [ 192.761657] [<ffffffff8184da29>] system_call_fastpath+0x16/0x1b
> [ 192.771904] ---[ end trace 6d601571820ebee0 ]---
Thanks for the report.
This is the intended behavior. Before commit 4485154138f6
("perf/x86/intel: Avoid spamming kernel log for BTS buffer failure") this
would result in page allocation failure warnings everytime the kzalloc()
failed. The patch turns this into a WARN_ON_ONCE() to only print it once
to avoid spamming the kernel log, it just has a different format than page
allocation failure warnings.
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
Powered by blists - more mailing lists