From: "Luck, Tony" <tony.luck@intel.com>
To: "Luck, Tony" <tony.luck@intel.com>,
Peter Zijlstra <peterz@infradead.org>
Cc: Mel Gorman <mgorman@techsingularity.net>,
Vincent Guittot <vincent.guittot@linaro.org>,
Patrick Bellasi <patrick.bellasi@arm.com>,
"Ingo Molnar" <mingo@kernel.org>,
Norbert Manthey <nmanthey@amazon.de>,
"Frederic Weisbecker" <frederic@kernel.org>,
"linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>
Subject: RE: v4.16+ seeing many unaligned access in dequeue_task_fair() on IA64
Date: Mon, 2 Apr 2018 23:39:02 +0000 [thread overview]
Message-ID: <3908561D78D1C84285E8C5FCA982C28F7B3C0F02@ORSMSX110.amr.corp.intel.com> (raw)
In-Reply-To: <20180402232448.fbop7k5xicblski5@agluck-desk>
> kernel unaligned access to 0xe00000031660fd74, ip=0xa0000001000f23e0
> kernel unaligned access to 0xe00000033bdffbcc, ip=0xa0000001000f2370
Here's the disassembly of dequeu_task_fair() in case it would help to see
which two instructions are getting all the faults:
a0000001000f21c0 <dequeue_task_fair>:
a0000001000f21c0: 08 28 29 0e 80 05 [MMI] alloc r37=ar.pfs,10,7,0
a0000001000f21c6: c0 00 33 7e 46 00 adds r12=-32,r12
a0000001000f21cc: c2 08 85 84 adds r16=4236,r33
a0000001000f21d0: 09 58 40 ab 16 27 [MMI] addl r11=-685232,r1
a0000001000f21d6: 80 02 84 02 42 20 adds r40=128,r33
a0000001000f21dc: 05 10 01 84 mov r41=r34;;
a0000001000f21e0: 08 00 00 00 01 00 [MMI] nop.m 0x0
a0000001000f21e6: f0 00 2c 00 42 80 mov r15=r11
a0000001000f21ec: 04 00 c4 00 mov r36=b0
a0000001000f21f0: 19 30 00 50 07 39 [MMB] cmp.eq p6,p7=0,r40
a0000001000f21f6: 30 00 c0 a3 4e 03 mov r3=-219008
a0000001000f21fc: 60 00 00 41 (p06) br.cond.spnt.few a0000001000f2250 <dequeue_task_fair+0x90>;;
a0000001000f2200: 0b 50 00 20 10 10 [MMI] ld4 r10=[r16];;
a0000001000f2206: 90 50 3c 24 40 00 shladd r9=r10,3,r15
a0000001000f220c: 00 00 04 00 nop.i 0x0;;
a0000001000f2210: 0b 40 00 12 18 10 [MMI] ld8 r8=[r9];;
a0000001000f2216: 60 1a 20 00 40 00 add r38=r3,r8
a0000001000f221c: 00 00 04 00 nop.i 0x0;;
a0000001000f2220: 11 18 71 4c 01 21 [MIB] adds r35=156,r38
a0000001000f2226: 70 02 98 02 42 00 adds r39=128,r38
a0000001000f222c: 68 f3 ff 58 br.call.sptk.many b0=a0000001000f1580 <dequeue_entity>;;
a0000001000f2230: 0a 10 00 46 10 10 [MMI] ld4 r2=[r35];;
a0000001000f2236: e0 f8 0b 7e 46 00 adds r14=-1,r2
a0000001000f223c: 00 00 04 00 nop.i 0x0
a0000001000f2240: 0a 00 00 00 01 00 [MMI] nop.m 0x0;;
a0000001000f2246: 00 70 8c 20 23 00 st4 [r35]=r14
a0000001000f224c: 00 00 04 00 nop.i 0x0
a0000001000f2250: 09 98 c0 02 d7 26 [MMI] addl r19=-2069584,r1
a0000001000f2256: 50 21 80 00 42 00 adds r21=4,r32
a0000001000f225c: 83 01 05 84 adds r24=152,r32;;
a0000001000f2260: 09 90 00 26 18 10 [MMI] ld8 r18=[r19]
a0000001000f2266: 60 01 54 20 20 00 ld4 r22=[r21]
a0000001000f226c: 00 00 04 00 nop.i 0x0;;
a0000001000f2270: 09 a0 fc 2d 3f 23 [MMI] adds r20=-1,r22
a0000001000f2276: 10 01 48 a0 20 00 ld4.a r17=[r18]
a0000001000f227c: 00 00 04 00 nop.i 0x0;;
a0000001000f2280: 08 00 00 00 01 00 [MMI] nop.m 0x0
a0000001000f2286: 00 a0 54 20 23 00 st4 [r21]=r20
a0000001000f228c: a1 8a 24 50 tbit.z p8,p9=r17,21
a0000001000f2290: 18 00 00 00 01 00 [MMB] nop.m 0x0
a0000001000f2296: 10 d9 00 80 02 00 chk.a.clr r17,a0000001000f2440 <dequeue_task_fair+0x280>
a0000001000f229c: 00 00 00 20 nop.b 0x0
a0000001000f22a0: 10 00 00 00 01 00 [MIB] nop.m 0x0
a0000001000f22a6: 00 00 00 02 00 04 nop.i 0x0
a0000001000f22ac: e0 00 00 43 (p08) br.cond.dpnt.few a0000001000f2380 <dequeue_task_fair+0x1c0>
a0000001000f22b0: 0b b8 00 30 10 10 [MMI] ld4 r23=[r24];;
a0000001000f22b6: b0 00 5c 14 73 00 cmp4.eq p11,p10=0,r23
a0000001000f22bc: 00 00 04 00 nop.i 0x0;;
a0000001000f22c0: 71 01 81 40 02 e1 [MIB] (p11) adds r32=288,r32
a0000001000f22c6: 12 01 00 00 42 05 (p11) mov r17=r0
a0000001000f22cc: e0 00 00 42 (p10) br.cond.dptk.few a0000001000f23a0 <dequeue_task_fair+0x1e0>;;
a0000001000f22d0: 09 f8 40 18 00 21 [MMI] adds r31=16,r12
a0000001000f22d6: 00 88 80 60 23 c0 st4.rel [r32]=r17
a0000001000f22dc: 00 10 1d 50 tbit.z p6,p7=r34,0;;
a0000001000f22e0: 10 00 44 3e 90 11 [MIB] st4 [r31]=r17
a0000001000f22e6: 00 00 00 02 00 03 nop.i 0x0
a0000001000f22ec: a0 00 00 43 (p06) br.cond.dpnt.few a0000001000f2380 <dequeue_task_fair+0x1c0>
a0000001000f22f0: 09 88 e0 42 02 21 [MMI] adds r17=312,r33
a0000001000f22f6: e0 80 85 04 42 40 adds r14=304,r33
a0000001000f22fc: c4 0b 09 84 adds r34=316,r33;;
a0000001000f2300: 09 00 01 22 10 10 [MMI] ld4 r32=[r17]
a0000001000f2306: 00 01 88 20 20 00 ld4 r16=[r34]
a0000001000f230c: 00 00 04 00 nop.i 0x0;;
a0000001000f2310: 10 00 00 00 01 00 [MIB] nop.m 0x0
a0000001000f2316: 90 00 80 10 28 04 tbit.z p9,p8=r32,0
a0000001000f231c: 70 00 00 43 (p08) br.cond.dpnt.few a0000001000f2380 <dequeue_task_fair+0x1c0>
a0000001000f2320: 0b 48 01 1c b8 10 [MMI] ld8.acq r41=[r14];;
a0000001000f2326: 10 0a a4 5c 40 00 or r33=1,r41
a0000001000f232c: 00 00 04 00 nop.i 0x0;;
a0000001000f2330: 0b 40 85 20 05 20 [MMI] sub r40=r33,r16;;
a0000001000f2336: 70 4a a0 00 42 00 adds r39=9,r40
a0000001000f233c: 00 00 04 00 nop.i 0x0;;
a0000001000f2340: 0a 58 48 4e 8a f5 [MMI] cmp4.ltu p11,p10=18,r39;;
a0000001000f2346: 82 82 a0 22 c0 e5 (p11) shladd r40=r16,2,r40
a0000001000f234c: 04 62 00 84 (p11) adds r39=32,r12
a0000001000f2350: 62 81 70 18 00 21 [MII] (p11) adds r16=28,r12
a0000001000f2356: 00 00 00 02 80 05 nop.i 0x0;;
a0000001000f235c: 45 40 75 52 (p11) extr.u r40=r40,2,30
a0000001000f2360: 6a 01 84 20 90 d1 [MMI] (p11) st4 [r16]=r33;;
a0000001000f2366: 02 40 9d 20 23 00 (p11) st4 [r39]=r40
a0000001000f236c: 00 00 04 00 nop.i 0x0
*a0000001000f2370: 6a 39 01 20 18 d0 [MMI] (p11) ld8 r39=[r16];;
a0000001000f2376: 02 38 45 70 23 00 (p11) st8.rel [r17]=r39
a0000001000f237c: 00 00 04 00 nop.i 0x0
a0000001000f2380: 00 00 00 00 01 00 [MII] nop.m 0x0
a0000001000f2386: 00 28 01 55 00 00 mov.i ar.pfs=r37
a0000001000f238c: 40 0a 00 07 mov b0=r36
a0000001000f2390: 19 00 00 00 01 00 [MMB] nop.m 0x0
a0000001000f2396: c0 00 31 00 42 80 adds r12=32,r12
a0000001000f239c: 08 00 84 00 br.ret.sptk.many b0;;
a0000001000f23a0: 18 f0 e0 42 02 21 [MMB] adds r30=312,r33
a0000001000f23a6: 00 02 81 04 42 00 adds r32=288,r32
a0000001000f23ac: 00 00 00 20 nop.b 0x0
a0000001000f23b0: 09 e0 50 18 00 21 [MMI] adds r28=20,r12
a0000001000f23b6: f0 81 30 00 42 c0 adds r31=16,r12
a0000001000f23bc: 00 10 1d 50 tbit.z p6,p7=r34,0;;
a0000001000f23c0: 09 e8 00 3c b8 10 [MMI] ld8.acq r29=[r30]
a0000001000f23c6: 10 01 80 20 20 00 ld4 r17=[r32]
a0000001000f23cc: 00 00 04 00 nop.i 0x0;;
a0000001000f23d0: 00 00 00 00 01 00 [MII] nop.m 0x0
a0000001000f23d6: b0 01 76 3e 29 40 shr.u r27=r29,32
a0000001000f23dc: 03 e8 00 84 mov r26=r29
*a0000001000f23e0: 0b 00 74 38 98 11 [MMI] st8 [r28]=r29;;
a0000001000f23e6: f0 e8 6c 1c 69 00 cmp4.ltu p15,p14=r29,r27
a0000001000f23ec: 00 00 04 00 nop.i 0x0;;
a0000001000f23f0: eb d1 00 36 00 21 [MMI] (p15) mov r26=r27;;
a0000001000f23f6: 90 09 68 5c 40 00 or r25=1,r26
a0000001000f23fc: 00 00 04 00 nop.i 0x0;;
a0000001000f2400: 02 68 44 32 8c 34 [MII] cmp4.ltu p13,p12=r17,r25
a0000001000f2406: 00 00 00 02 00 26 nop.i 0x0;;
a0000001000f240c: 12 c9 14 80 (p12) sub r17=r17,r25
a0000001000f2410: a2 89 44 22 05 20 [MII] (p13) sub r17=r17,r17
a0000001000f2416: 00 00 00 02 00 00 nop.i 0x0;;
a0000001000f241c: 00 00 04 00 nop.i 0x0
a0000001000f2420: 18 00 44 40 b0 11 [MMB] st4.rel [r32]=r17
a0000001000f2426: 00 88 7c 20 a3 03 st4 [r31]=r17
a0000001000f242c: d0 fe ff 4a (p07) br.cond.dptk.few a0000001000f22f0 <dequeue_task_fair+0x130>
a0000001000f2430: 10 00 00 00 01 00 [MIB] nop.m 0x0
a0000001000f2436: 00 00 00 02 00 00 nop.i 0x0
a0000001000f243c: 50 ff ff 48 br.few a0000001000f2380 <dequeue_task_fair+0x1c0>
a0000001000f2440: 09 00 00 00 01 00 [MMI] nop.m 0x0
a0000001000f2446: 10 01 48 20 20 00 ld4 r17=[r18]
a0000001000f244c: 00 00 04 00 nop.i 0x0;;
a0000001000f2450: 11 00 00 00 01 00 [MIB] nop.m 0x0
a0000001000f2456: 80 50 45 12 28 00 tbit.z p8,p9=r17,21
a0000001000f245c: 50 fe ff 48 br.few a0000001000f22a0 <dequeue_task_fair+0xe0>;;
next prev parent reply other threads:[~2018-04-02 23:39 UTC|newest]
Thread overview: 11+ messages / expand[flat|nested] mbox.gz Atom feed top
2018-04-02 23:24 v4.16+ seeing many unaligned access in dequeue_task_fair() on IA64 Luck, Tony
2018-04-02 23:39 ` Luck, Tony [this message]
2018-04-03 7:37 ` Peter Zijlstra
2018-04-03 18:58 ` Luck, Tony
2018-04-04 0:04 ` Luck, Tony
2018-04-04 7:25 ` Peter Zijlstra
2018-04-04 16:38 ` Luck, Tony
2018-04-04 16:53 ` Peter Zijlstra
2018-04-05 8:05 ` Peter Zijlstra
2018-04-05 8:56 ` Ingo Molnar
2018-04-05 9:18 ` [tip:sched/urgent] sched/core: Force proper alignment of 'struct util_est' tip-bot for Peter Zijlstra
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=3908561D78D1C84285E8C5FCA982C28F7B3C0F02@ORSMSX110.amr.corp.intel.com \
--to=tony.luck@intel.com \
--cc=frederic@kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=mgorman@techsingularity.net \
--cc=mingo@kernel.org \
--cc=nmanthey@amazon.de \
--cc=patrick.bellasi@arm.com \
--cc=peterz@infradead.org \
--cc=vincent.guittot@linaro.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.