| lists.openwall.net | lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC | |
|
Open Source and information security mailing list archives
| ||
|
Date: Fri, 30 Jun 2017 13:21:47 +0300
From: Alexey Budankov <alexey.budankov@...ux.intel.com>
To: Mark Rutland <mark.rutland@....com>
Cc: Peter Zijlstra <peterz@...radead.org>,
Ingo Molnar <mingo@...hat.com>,
Arnaldo Carvalho de Melo <acme@...nel.org>,
Alexander Shishkin <alexander.shishkin@...ux.intel.com>,
Andi Kleen <ak@...ux.intel.com>,
Kan Liang <kan.liang@...el.com>,
Dmitri Prokhorov <Dmitry.Prohorov@...el.com>,
Valery Cherepennikov <valery.cherepennikov@...el.com>,
David Carrillo-Cisneros <davidcc@...gle.com>,
Stephane Eranian <eranian@...gle.com>,
linux-kernel@...r.kernel.org
Subject: Re: [PATCH v3 1/n] perf/core: addressing 4x slowdown during
per-process profiling of STREAM benchmark on Intel Xeon Phi
Hi,
On 19.06.2017 16:37, Alexey Budankov wrote:
> On 19.06.2017 16:26, Mark Rutland wrote:
>> On Mon, Jun 19, 2017 at 04:08:32PM +0300, Alexey Budankov wrote:
>>> On 16.06.2017 1:10, Alexey Budankov wrote:
>>>> On 15.06.2017 22:56, Mark Rutland wrote:
>>>>> On Thu, Jun 15, 2017 at 08:41:42PM +0300, Alexey Budankov wrote:
>>>>>> This series of patches continues v2 and addresses captured comments.
>>
>>>>>> Specifically this patch replaces pinned_groups and flexible_groups
>>>>>> lists of perf_event_context by red-black cpu indexed trees avoiding
>>>>>> data structures duplication and introducing possibility to iterate
>>>>>> event groups for a specific CPU only.
>>
>>>>> Have you thrown Vince's perf fuzzer at this?
>>>>>
>>>>> If you haven't, please do. It can be found in the fuzzer directory of:
>>>>>
>>>>> https://github.com/deater/perf_event_tests
>>>>
>>>> Accepted.
>>>
>>> I run the test suite and it revealed no additional regressions in
>>> comparison to what I have on the clean kernel.
>>>
>>> However the fuzzer constantly reports some strange stacks that are
>>> not seen on the clean kernel and I have no idea how that might be
>>> caused by the patches.
>>
>> Ok; that was the kind of thing I was concerned about.
>>
>> What you say "strange stacks", what do you mean exactly?
>>
>> I take it the kernel spewing backtraces in dmesg?
>>
>> Can you dump those?
>
> Here it is:
>
> list_del corruption. prev->next should be ffff88c2c4654010, but was ffff88c31eb0c020
> [ 607.632813] ------------[ cut here ]------------
> [ 607.632816] kernel BUG at lib/list_debug.c:53!
> [ 607.632825] invalid opcode: 0000 [#1] SMP
> [ 607.632898] Modules linked in: fuse xt_CHECKSUM iptable_mangle ipt_MASQUERADE nf_nat_masquerade_ipv4 iptable_nat nf_nat_ipv4 nf_nat nf_conntrack_ipv4 nf_defrag_ipv4 xt_conntrack nf_conntrack libcrc32c tun bridge stp llc ebtable_filter ebtables ip6table_filter ip6_tables nfsv3 rpcsec_gss_krb5 nfsv4 arc4 md4 nls_utf8 cifs nfs ccm dns_resolver fscache rpcrdma ib_isert iscsi_target_mod ib_iser libiscsi scsi_transport_iscsi ib_srpt target_core_mod ib_srp scsi_transport_srp ib_ipoib rdma_ucm ib_ucm ib_uverbs ib_umad rdma_cm ib_cm iw_cm intel_rapl sb_edac x86_pkg_temp_thermal intel_powerclamp hfi1 coretemp crct10dif_pclmul crc32_pclmul ghash_clmulni_intel intel_cstate iTCO_wdt intel_uncore iTCO_vendor_support joydev rdmavt intel_rapl_perf i2c_i801 ib_core ipmi_ssif mei_me mei ipmi_si ipmi_devintf tpm_tis
> [ 607.633954] lpc_ich pcspkr ipmi_msghandler acpi_pad tpm_tis_core shpchp tpm wmi acpi_power_meter nfsd auth_rpcgss nfs_acl lockd grace sunrpc mgag200 drm_kms_helper ttm drm igb crc32c_intel ptp pps_core dca i2c_algo_bit
> [ 607.634262] CPU: 271 PID: 28944 Comm: perf_fuzzer Not tainted 4.12.0-rc4+ #22
> [ 607.634363] Hardware name: Intel Corporation S7200AP/S7200AP, BIOS S72C610.86B.01.01.0190.080520162104 08/05/2016
> [ 607.634505] task: ffff88c2d5714000 task.stack: ffffa6f9352c8000
> [ 607.634597] RIP: 0010:__list_del_entry_valid+0x7b/0x90
> [ 607.634670] RSP: 0000:ffffa6f9352cbad0 EFLAGS: 00010046
> [ 607.634746] RAX: 0000000000000054 RBX: ffff88c2c4654000 RCX: 0000000000000000
> [ 607.634845] RDX: 0000000000000000 RSI: ffff88c33fdce168 RDI: ffff88c33fdce168
> [ 607.634944] RBP: ffffa6f9352cbad0 R08: 00000000fffffffe R09: 0000000000000600
> [ 607.635042] R10: 0000000000000005 R11: 0000000000000000 R12: ffff88c2e71ab200
> [ 607.635140] R13: ffff88c2c4654010 R14: 0000000000000001 R15: 0000000000000001
> [ 607.635240] FS: 0000000000000000(0000) GS:ffff88c33fdc0000(0000) knlGS:0000000000000000
> [ 607.635351] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
> [ 607.635431] CR2: 00000000026be1e4 CR3: 0000002488e09000 CR4: 00000000001407e0
> [ 607.635531] Call Trace:
> [ 607.635583] list_del_event+0x1d7/0x210
> [ 607.635646] ? perf_cgroup_attach+0x70/0x70
> [ 607.635711] __perf_remove_from_context+0x3e/0x90
> [ 607.635783] ? event_sched_out.isra.90+0x300/0x300
> [ 607.635854] event_function_call+0xbf/0xf0
> [ 607.635918] ? event_sched_out.isra.90+0x300/0x300
> [ 607.635991] perf_remove_from_context+0x25/0x70
> [ 607.636060] perf_event_release_kernel+0xda/0x250
> [ 607.636132] ? __dentry_kill+0x10e/0x160
> [ 607.636192] perf_release+0x10/0x20
> [ 607.636249] __fput+0xdf/0x1e0
> [ 607.636299] ____fput+0xe/0x10
> [ 607.636350] task_work_run+0x83/0xb0
> [ 607.636408] do_exit+0x2bc/0xbc0
> [ 607.636460] ? page_add_file_rmap+0xaf/0x200
> [ 607.636526] ? alloc_set_pte+0x115/0x4f0
> [ 607.636587] do_group_exit+0x3f/0xb0
> [ 607.636644] get_signal+0x1cc/0x5c0
> [ 607.636703] do_signal+0x37/0x6a0
> [ 607.636758] ? __perf_sw_event+0x4f/0x80
> [ 607.636821] ? __do_page_fault+0x2e1/0x4d0
> [ 607.636885] exit_to_usermode_loop+0x4c/0x92
> [ 607.636953] prepare_exit_to_usermode+0x40/0x50
> [ 607.637023] retint_user+0x8/0x13
> [ 607.640312] RIP: 0033:0x40f4a9
> [ 607.643500] RSP: 002b:00007ffc62d00668 EFLAGS: 00000206 ORIG_RAX: ffffffffffffff02
> [ 607.646678] RAX: 0000000000000000 RBX: 0000000000000000 RCX: 000000000001100a
> [ 607.649777] RDX: 00007fd5dd25bae0 RSI: 00007fd5dd259760 RDI: 00007fd5dd25a640
> [ 607.652791] RBP: 00007ffc62d00680 R08: 00007fd5dd45e740 R09: 0000000000000000
> [ 607.655709] R10: 00007fd5dd45ea10 R11: 0000000000000246 R12: 0000000000401980
> [ 607.658530] R13: 00007ffc62d02a80 R14: 0000000000000000 R15: 0000000000000000
> [ 607.661249] Code: e8 3a f7 d8 ff 0f 0b 48 89 fe 31 c0 48 c7 c7 20 66 cb ac e8 27 f7 d8 ff 0f 0b 48 89 fe 31 c0 48 c7 c7 e0 65 cb ac e8 14 f7 d8 ff <0f> 0b 48 89 fe 31 c0 48 c7 c7 a8 65 cb ac e8 01 f7 d8 ff 0f 0b
> [ 607.666819] RIP: __list_del_entry_valid+0x7b/0x90 RSP: ffffa6f9352cbad0
> [ 607.683316] ---[ end trace 34244c35550e0713 ]---
> [ 607.691830] Fixing recursive fault but reboot is needed!
addressed in:
[PATCH v5 4/4] perf/core: addressing 4x slowdown during per-process
profiling of STREAM benchmark on Intel Xeon Phi
>
> 2:
>
> [ 467.942059] unchecked MSR access error: WRMSR to 0x711 (tried to write 0x00000000e8cc0055) at rIP: 0xffffffffac05fbd4 (native_write_msr+0x4/0x30)
> [ 467.942068] Call Trace:
> [ 467.942073] <IRQ>
> [ 467.942094] ? snbep_uncore_msr_enable_event+0x54/0x60 [intel_uncore]
> [ 467.942107] uncore_pmu_event_start+0x9b/0x100 [intel_uncore]
> [ 467.942119] uncore_pmu_event_add+0x235/0x3a0 [intel_uncore]
> [ 467.942126] ? sched_clock+0xb/0x10
> [ 467.942132] ? sched_clock_cpu+0x11/0xb0
> [ 467.942140] event_sched_in.isra.100+0xdf/0x250
> [ 467.942145] sched_in_group+0x210/0x390
> [ 467.942150] ? sched_in_group+0x390/0x390
> [ 467.942155] group_sched_in_flexible_callback+0x17/0x20
> [ 467.942160] perf_cpu_tree_iterate+0x45/0x75
> [ 467.942165] ctx_sched_in+0x97/0x110
> [ 467.942169] perf_event_sched_in+0x77/0x80
> [ 467.942174] ctx_resched+0x69/0xb0
> [ 467.942179] __perf_event_enable+0x208/0x250
> [ 467.942184] event_function+0x93/0xe0
> [ 467.942188] remote_function+0x3b/0x50
> [ 467.942194] flush_smp_call_function_queue+0x71/0x120
> [ 467.942200] generic_smp_call_function_single_interrupt+0x13/0x30
> [ 467.942206] smp_call_function_single_interrupt+0x27/0x40
> [ 467.942212] call_function_single_interrupt+0x93/0xa0
> [ 467.942217] RIP: 0010:native_restore_fl+0x6/0x10
> [ 467.942220] RSP: 0000:ffff88c33ba03e00 EFLAGS: 00000206 ORIG_RAX: ffffffffffffff04
> [ 467.942224] RAX: ffff88c33c8dca08 RBX: ffff88c33c8dc928 RCX: 0000000000000017
> [ 467.942227] RDX: 0000000000000000 RSI: 0000000000000206 RDI: 0000000000000206
> [ 467.942229] RBP: ffff88c33ba03e00 R08: 0000000000000001 R09: 0000000000007151
> [ 467.942232] R10: 0000000000000000 R11: 000000000000005d R12: ffff88c33c8dca08
> [ 467.942234] R13: ffff88c33c8dc140 R14: 0000000000000001 R15: 00000000000001d8
> [ 467.942241] _raw_spin_unlock_irqrestore+0x16/0x20
> [ 467.942245] update_blocked_averages+0x2cf/0x4a0
> [ 467.942251] rebalance_domains+0x4b/0x2b0
> [ 467.942256] run_rebalance_domains+0x1d7/0x210
> [ 467.942260] __do_softirq+0xd1/0x27f
> [ 467.942267] irq_exit+0xe9/0x100
> [ 467.942271] scheduler_ipi+0x8f/0x140
> [ 467.942275] smp_reschedule_interrupt+0x29/0x30
> [ 467.942280] reschedule_interrupt+0x93/0xa0
> [ 467.942284] RIP: 0010:native_safe_halt+0x6/0x10
> [ 467.942286] RSP: 0000:fffffffface03de0 EFLAGS: 00000246 ORIG_RAX: ffffffffffffff02
> [ 467.942291] RAX: 0000000000000000 RBX: fffffffface104c0 RCX: 0000000000000000
> [ 467.942293] RDX: 0000000000000000 RSI: 0000000000000000 RDI: 0000000000000000
> [ 467.942295] RBP: fffffffface03de0 R08: 0000006d2c5efae3 R09: 0000000000000000
> [ 467.942298] R10: 0000000000000201 R11: 0000000000000930 R12: 0000000000000000
> [ 467.942300] R13: fffffffface104c0 R14: 0000000000000000 R15: 0000000000000000
> [ 467.942302] </IRQ>
> [ 467.942308] default_idle+0x20/0x100
> [ 467.942313] arch_cpu_idle+0xf/0x20
> [ 467.942317] default_idle_call+0x2c/0x40
> [ 467.942321] do_idle+0x158/0x1e0
> [ 467.942325] cpu_startup_entry+0x71/0x80
> [ 467.942330] rest_init+0x77/0x80
> [ 467.942338] start_kernel+0x4a7/0x4c8
> [ 467.942342] ? set_init_arg+0x5a/0x5a
> [ 467.942348] ? early_idt_handler_array+0x120/0x120
> [ 467.942352] x86_64_start_reservations+0x29/0x2b
> [ 467.942357] x86_64_start_kernel+0x151/0x174
> [ 467.942363] secondary_startup_64+0x9f/0x9f>
>
>>
>> Thanks,
>> Mark.
>>
>
>
Powered by blists - more mailing lists