From owner-freebsd-arm@freebsd.org Thu Jun 28 16:33:22 2018 Return-Path: Delivered-To: freebsd-arm@mailman.ysv.freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2610:1c1:1:606c::19:1]) by mailman.ysv.freebsd.org (Postfix) with ESMTP id 62426101F0C1 for ; Thu, 28 Jun 2018 16:33:22 +0000 (UTC) (envelope-from fbsd@www.zefox.net) Received: from www.zefox.net (www.zefox.net [50.1.20.27]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (Client CN "www.zefox.org", Issuer "www.zefox.org" (not verified)) by mx1.freebsd.org (Postfix) with ESMTPS id BB16483660 for ; Thu, 28 Jun 2018 16:33:21 +0000 (UTC) (envelope-from fbsd@www.zefox.net) Received: from www.zefox.net (localhost [127.0.0.1]) by www.zefox.net (8.15.2/8.15.2) with ESMTPS id w5SGXTVA033483 (version=TLSv1.2 cipher=DHE-RSA-AES256-GCM-SHA384 bits=256 verify=NO); Thu, 28 Jun 2018 09:33:30 -0700 (PDT) (envelope-from fbsd@www.zefox.net) Received: (from fbsd@localhost) by www.zefox.net (8.15.2/8.15.2/Submit) id w5SGXTbR033482; Thu, 28 Jun 2018 09:33:29 -0700 (PDT) (envelope-from fbsd) Date: Thu, 28 Jun 2018 09:33:29 -0700 From: bob prohaska To: Mark Millard Cc: freebsd-arm@freebsd.org, bob prohaska Subject: Re: RPI3 swap experiments Message-ID: <20180628163328.GA33408@www.zefox.net> References: <20180626151843.GD17293@www.zefox.net> <3525D7C7-F848-45A1-BD85-2DAC895DF48C@yahoo.com> <20180626222834.GA20270@www.zefox.net> <28012DFB-37A0-461A-BB62-CD3EE61E82F0@yahoo.com> <20180627054027.GA22144@www.zefox.net> <20180627194217.GA27793@www.zefox.net> <20180628022457.GA30110@www.zefox.net> <7B9D272D-3EDE-46FA-8A1C-AEE65047167C@yahoo.com> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <7B9D272D-3EDE-46FA-8A1C-AEE65047167C@yahoo.com> User-Agent: Mutt/1.5.24 (2015-08-30) X-BeenThere: freebsd-arm@freebsd.org X-Mailman-Version: 2.1.26 Precedence: list List-Id: "Porting FreeBSD to ARM processors." List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Thu, 28 Jun 2018 16:33:22 -0000 It turns out that Peter Holm's stress2 suite will trigger a crash whose console messages look superficially like those produced by -j4 buildworld. The storage configuration is the same: /var and /usr on USB flash, 1 GB swap on microSD. The gstat/swapinfo log file is at http://www.zefox.net/~fbsd/rpi3/swaptests/r335655/1gbsdflash/stress2/swapuse.log An early sign of trouble is dT: 10.042s w: 10.000s L(q) ops/s r/s kBps ms/r w/s kBps ms/w d/s kBps ms/d %busy Name 16 57 0 0 0.0 57 6351 237.6 0 0 0.0 100.2 mmcsd0 15 57 0 0 0.0 57 6351 237.6 0 0 0.0 100.2 mmcsd0s3 14 57 0 0 0.0 57 6351 237.7 0 0 0.0 100.2 mmcsd0s3a Thu Jun 28 00:52:49 PDT 2018 Device 1K-blocks Used Avail Capacity /dev/mmcsd0s3b 1048576 32320 1016256 3% Jun 28 00:52:05 www bob[2122]: Starting test df.cfg Jun 28 00:52:45 www kernel: swap_pager: indefinite wait buffer: bufobj: 0, blkno: 3467, size: 4096 dT: 10.004s w: 10.000s L(q) ops/s r/s kBps ms/r w/s kBps ms/w d/s kBps ms/d %busy Name 19 51 0 0 0.0 51 5796 239.5 0 0 0.0 100.2 mmcsd0 17 51 0 0 0.0 51 5796 239.5 0 0 0.0 100.2 mmcsd0s3 17 51 0 0 0.0 51 5796 239.6 0 0 0.0 100.2 mmcsd0s3a Thu Jun 28 00:52:59 PDT 2018 Device 1K-blocks Used Avail Capacity /dev/mmcsd0s3b 1048576 32320 1016256 3% Jun 28 00:52:05 www bob[2122]: Starting test df.cfg Jun 28 00:52:45 www kernel: swap_pager: indefinite wait buffer: bufobj: 0, blkno: 3467, size: 4096 Here's the spot where the /dev/da0 errors begin in the gstat/swapinfo log: dT: 10.002s w: 10.000s L(q) ops/s r/s kBps ms/r w/s kBps ms/w d/s kBps ms/d %busy Name 4 9 1 3 131.4 9 186 46.9 0 0 0.0 28.5 mmcsd0 1 1 0 0 0.0 1 34 51.6 0 0 0.0 7.2 da0 4 9 1 3 133.3 9 186 53.3 0 0 0.0 32.9 mmcsd0s3 3 9 0 0 0.0 9 186 53.3 0 0 0.0 32.9 mmcsd0s3a 1 1 1 3 133.3 0 0 0.0 0 0 0.0 8.0 mmcsd0s3b 1 1 0 0 0.0 1 34 51.7 0 0 0.0 7.2 da0a Thu Jun 28 02:34:11 PDT 2018 Device 1K-blocks Used Avail Capacity /dev/mmcsd0s3b 1048576 37736 1010840 4% Jun 28 02:35:15 www kernel: (da0:umass-sim0:0:0:0): Error 5, Retries exhausted Jun 28 02:35:15 www kernel: g_vfs_done():da0a[WRITE(offset=827129856, length=16384)]error = 5 dT: 10.002s w: 10.000s L(q) ops/s r/s kBps ms/r w/s kBps ms/w d/s kBps ms/d %busy Name 0 38 38 452 285.2 0 0 0.0 0 0 0.0 35.3 mmcsd0 0 38 38 456 286.1 0 0 0.0 0 0 0.0 35.5 mmcsd0s3 0 38 38 456 286.3 0 0 0.0 0 0 0.0 35.5 mmcsd0s3b Thu Jun 28 02:40:26 PDT 2018 Device 1K-blocks Used Avail Capacity /dev/mmcsd0s3b 1048576 58408 990168 6% Jun 28 02:40:15 www kernel: (da0:umass-sim0:0:0:0): CAM status: CCB request completed with an error Jun 28 02:40:15 www kernel: (da0:umass-sim0:0:0:0): Retrying command, 3 more tries remain dT: 10.002s w: 10.000s L(q) ops/s r/s kBps ms/r w/s kBps ms/w d/s kBps ms/d %busy Name Thu Jun 28 02:40:37 PDT 2018 Device 1K-blocks Used Avail Capacity /dev/mmcsd0s3b 1048576 58408 990168 6% Jun 28 02:40:15 www kernel: (da0:umass-sim0:0:0:0): CAM status: CCB request completed with an error Jun 28 02:40:15 www kernel: (da0:umass-sim0:0:0:0): Retrying command, 3 more tries remain dT: 10.076s w: 10.000s L(q) ops/s r/s kBps ms/r w/s kBps ms/w d/s kBps ms/d %busy Name Thu Jun 28 02:40:50 PDT 2018 Device 1K-blocks Used Avail Capacity /dev/mmcsd0s3b 1048576 58408 990168 6% Jun 28 02:41:38 www kernel: (da0:umass-sim0:0:0:0): WRITE(10). CDB: 2a 00 00 18 a6 80 00 00 40 00 Jun 28 02:41:38 www kernel: (da0:umass-sim0:0:0:0): CAM status: CCB request completed with an error Jun 28 02:41:38 www kernel: (da0:umass-sim0:0:0:0): Retrying command, 1 more tries remain Jun 28 02:41:38 www kernel: smsc0: warning: Failed to read register 0x114 Jun 28 02:41:38 www kernel: smsc0: warning: MII read timeout [da0 error flood continues] Curiously, the machine kept running (top output updated) until the plug was pulled next morning. This test is certainly quicker than using -j4 buildworld, but it isn't obvious it's doing the same thing since /dev/da0 is exercised far less. Thanks for reading, bob prohaska