[00:06:21] *** Joins: gila (~gila@5ED4D979.cm-7-5d.dynamic.ziggo.nl) [01:00:50] *** Quits: ziyeyang_ (ziyeyang@nat/intel/x-tblzdqkzlldgbrst) (Remote host closed the connection) [01:23:52] *** Joins: mszwed (~mszwed@134.134.139.72) [02:46:50] *** Joins: travis-ci (~travis-ci@ec2-54-198-100-54.compute-1.amazonaws.com) [02:46:51] (spdk/master) OCF: tests: add flush and unmap test (Vitaliy Mysak) [02:46:51] Diff URL: https://github.com/spdk/spdk/compare/ec415110c1b0...c61be1af59ad [02:46:51] *** Parts: travis-ci (~travis-ci@ec2-54-198-100-54.compute-1.amazonaws.com) () [08:50:01] peluse: Test: test_crypto_op_complete .../var/jenkins/workspace/Other/blockdev_autotest/spdk/lib/bdev/crypto/vbdev_crypto.c: 426:_crypto_operation_complete: *ERROR*: Unknown bdev type 9 on crypto operation completion [08:50:10] type 9 is WRITE_ZEROES [08:50:34] have you seen that anywhere? [09:00:09] uh, that's not a real failure, just a negative test [09:00:23] the crypto_ut still failed for me: [09:00:32] FAILED 1. crypto_ut.c:825 - result_ [09:06:40] darsto, one sec... should be a negative tests if my memory severs... do you know if ASAN was enabled on that run? [09:08:34] SPDK_RUN_ASAN=0 [09:09:11] here's the full log https://ci.spdk.io/spdk-jenkins/results/autotest-per-patch/builds/21254/archive/blockdev_autotest/build.log [09:10:38] it happened twice in a row on my patch, although I can't see how exactly I broke it [09:10:52] intermittent failure is my guess [09:12:42] its not intermittent, it fails when ASAN is not enabled and I don't know why [09:14:18] sethhowe, or klateck may have changed something but we weren't running ctypto UT on Jenkins as early as last week because of a config mistake. Sounds like it was fixed but ASAN not enabled [09:15:02] if you run the UT locally you'll see it will fail w/o asan and pass with it. It hasn't always been that way, sethhowe fixed some scan-build issues a while back nad it started happening at that time.... [09:19:36] BTW, the not running UT for crypto is on the unittets sub job in Jenkins though, not on the blockdev one. I'm not sure why blockdev isn' failing every time right now, looks like ASAN has been off for a while [09:22:13] darsto, Oh, I see it's your patch that's exposing this https://review.gerrithub.io/c/spdk/spdk/+/442327/1/scripts/vagrant/autorun-spdk.conf [09:22:26] this meaning the known issue about crypto and asan and UT [09:24:46] the patch is just a PoC, but that crypto failure got me intrigued [09:32:41] yeah, seth and I were just talking about it. I suppose its time to root cause it :) [09:32:55] I'll step through it and see what jumps out at me if anything... [09:42:57] darsto, OK I think I see the problem. Have a 1:1 right now but will submit a one liner afterwards if I'm correct, just need to spend a few more min with it [09:44:18] sweet [10:02:47] *** Joins: travis-ci (~travis-ci@ec2-54-83-78-36.compute-1.amazonaws.com) [10:02:48] (spdk/master) nvmf: add TCP sample items to nvmf.conf.in (Changpeng Liu) [10:02:48] Diff URL: https://github.com/spdk/spdk/compare/c61be1af59ad...6ee1f99b3f93 [10:02:48] *** Parts: travis-ci (~travis-ci@ec2-54-83-78-36.compute-1.amazonaws.com) () [10:03:59] *** Joins: travis-ci (~travis-ci@ec2-54-198-79-67.compute-1.amazonaws.com) [10:04:00] (spdk/master) rdma: split PENDING_DATA_TRANSFER into two states. (Seth Howell) [10:04:00] Diff URL: https://github.com/spdk/spdk/compare/6ee1f99b3f93...1d0a8e1cec5a [10:04:00] *** Parts: travis-ci (~travis-ci@ec2-54-198-79-67.compute-1.amazonaws.com) () [10:06:26] *** Joins: travis-ci (~travis-ci@ec2-54-80-46-148.compute-1.amazonaws.com) [10:06:27] (spdk/master) test/vhost: test multiple connections to a single vhost device (Pawel Kaminski) [10:06:27] Diff URL: https://github.com/spdk/spdk/compare/1d0a8e1cec5a...766b1f62d935 [10:06:27] *** Parts: travis-ci (~travis-ci@ec2-54-80-46-148.compute-1.amazonaws.com) () [10:41:09] bwalker: can we remove the deprecated construct_nvmf_subsystem RPC? [10:41:13] before 19.01? [10:41:24] sure [10:41:26] will submit patch [10:41:33] i can do it [10:41:54] just wanted to confirm before i spent time on it :) i was just reviewing a patch where someone was updating the deprecated RPC [10:42:18] go for it [10:42:24] we should audit deprecated RPCs probably [10:42:30] as we do release prep [10:44:05] yeah there's a few other ones I'm going to kill [10:44:31] need to confirm when they were deprecated actually [10:44:56] we marked construct_nvmf_subsystems deprecated before v18.10 was released [10:45:05] yeah that one is good [10:45:11] lots of others marked deprecated that I'm auditing [10:56:57] darsto, so yeah the problem was a prev scan-build fix conflicting with asan being off. I don't have an old enough scan-build to repro so sethhowe is submitting the patch, super simple change.... [11:07:04] jimharris: that whole nvmf_rpc_deprecated file can be deleted actually [11:19:24] jimharris: What do you think about deleting deprecated options in the config file? There are at least two of them marked deprecated with a warning for at least one release [11:19:44] since the whole config file format is deprecated, I'm not sure we should bother [11:30:55] *** Joins: travis-ci (~travis-ci@ec2-34-230-73-6.compute-1.amazonaws.com) [11:30:56] (spdk/master) nvme/perf: enable keep alive feature with perf (Changpeng Liu) [11:30:56] Diff URL: https://github.com/spdk/spdk/compare/766b1f62d935...a4c344c1a4b4 [11:30:56] *** Parts: travis-ci (~travis-ci@ec2-34-230-73-6.compute-1.amazonaws.com) () [11:47:08] 46.6k lines of code added, 9.3k removed this release [11:47:23] that's a lot of activity [11:48:13] 706 commits [11:48:24] in 3 months, which included the holidays [11:55:44] what holidays? [11:58:41] *** Joins: travis-ci (~travis-ci@ec2-54-87-88-57.compute-1.amazonaws.com) [11:58:42] (spdk/master) OCF: rpc: add get_ocf_bdevs method (Vitaliy Mysak) [11:58:42] Diff URL: https://github.com/spdk/spdk/compare/a4c344c1a4b4...199080cfa2b6 [11:58:42] *** Parts: travis-ci (~travis-ci@ec2-54-87-88-57.compute-1.amazonaws.com) () [11:59:41] *** Joins: travis-ci (~travis-ci@ec2-34-207-242-93.compute-1.amazonaws.com) [11:59:42] (spdk/master) vhost: fail I/Os with payload length = 0 (Darek Stojaczyk) [11:59:42] Diff URL: https://github.com/spdk/spdk/compare/199080cfa2b6...4d74fad8b5ad [11:59:42] *** Parts: travis-ci (~travis-ci@ec2-34-207-242-93.compute-1.amazonaws.com) () [12:01:22] *** Joins: travis-ci (~travis-ci@ec2-54-196-226-99.compute-1.amazonaws.com) [12:01:23] (spdk/master) virtio: don't reinitialize entire virtqueues on mem hotplug (Darek Stojaczyk) [12:01:23] Diff URL: https://github.com/spdk/spdk/compare/4d74fad8b5ad...148eda65c1b4 [12:01:23] *** Parts: travis-ci (~travis-ci@ec2-54-196-226-99.compute-1.amazonaws.com) () [12:29:04] *** Joins: travis-ci (~travis-ci@ec2-54-226-232-213.compute-1.amazonaws.com) [12:29:05] (spdk/master) OCF: use rte_pause in wait loop (Vitaliy Mysak) [12:29:05] Diff URL: https://github.com/spdk/spdk/compare/148eda65c1b4...b2d5bfd0d192 [12:29:05] *** Parts: travis-ci (~travis-ci@ec2-54-226-232-213.compute-1.amazonaws.com) () [12:45:19] I had to leave out 2 OCF patches and rebase the remaining ones on top [12:46:40] they will slowly make their way through the TP - It seems like I can't reschedule them on neither CTP nor Jenkins [12:47:20] could someone verify if it's broken or is it just me? [12:57:35] darsto: you wouldn't know about the holidays - you worked right through them [13:32:02] darsto: It seemed to work for me. I just rescheduled them on the CTP [15:25:38] *** Joins: travis-ci (~travis-ci@ec2-54-80-199-181.compute-1.amazonaws.com) [15:25:39] (spdk/master) bdev/virtio: Remove deprecated remove_virtio_scsi_bdev rpc (Ben Walker) [15:25:39] Diff URL: https://github.com/spdk/spdk/compare/b2d5bfd0d192...b5c7ae072dc4 [15:25:39] *** Parts: travis-ci (~travis-ci@ec2-54-80-199-181.compute-1.amazonaws.com) () [15:35:56] *** Quits: gila (~gila@5ED4D979.cm-7-5d.dynamic.ziggo.nl) (Quit: My Mac Pro has gone to sleep. ZZZzzz…) [15:51:07] jimharris, see the note I just added to the ISAL patch... let me know what you think, I can fix it up now if you like [15:55:33] isal builds the manuals by default? [15:55:42] ideally we just turn that off [16:01:07] yeah, greg mentioned we might be able to do that. after merge though or now? [16:01:23] i don't even see how the docs are getting built [16:02:28] let me repro and I'll pinpoint it [16:03:15] does it happen during configure or make? [16:04:47] make [16:04:48] "[peluse@pels-64 spdk]$ make -j72 [16:04:49] make[2]: warning: -jN forced in submake: disabling jobserver mode. [16:04:49] Configuration done using x86_64-native-linuxapp-gcc [16:04:49] make[3]: Entering directory '/home/peluse/spdk/dpdk' [16:04:49] make[3]: help2man: Command not found [16:04:49] make[3]: [Makefile:3723: programs/igzip.1] Error 127 (ignored) [16:04:51] == Build lib" [16:05:08] oh wait, one sec [16:10:13] find anything? [16:10:26] rerunning a build with one proc, wil be a min [16:10:36] * peluse twiddling thumbs [16:23:00] even after removing help2man which I just installed it still works. Rebooting [16:27:37] OK, I think this is fine in the ISAL patch. It's happening only on my vbdev rebased rebased on ISAL patch so I must have munged a merge or something. Will track it down and go leave a note on the ISAL patch [16:28:02] ok [16:48:46] BTW sethhowe `and I just talked about the other warning in my paste about -jN and its coming from ipsecbuild/Makefile when we specify a different number of procs than what is used in there. seems harmless though [17:15:57] *** Joins: travis-ci (~travis-ci@ec2-54-80-199-181.compute-1.amazonaws.com) [17:15:58] (spdk/master) examples/ioat: Improve error check of input parsing by spdk_strtol (Shuhei Matsumoto) [17:15:58] Diff URL: https://github.com/spdk/spdk/compare/b5c7ae072dc4...47fadb7f9b17 [17:15:58] *** Parts: travis-ci (~travis-ci@ec2-54-80-199-181.compute-1.amazonaws.com) () [17:16:27] *** Joins: travis-ci (~travis-ci@ec2-54-80-46-148.compute-1.amazonaws.com) [17:16:28] (spdk/master) check_format: Verify #include syntax (Ben Walker) [17:16:28] Diff URL: https://github.com/spdk/spdk/compare/47fadb7f9b17...9fd924cb40e7 [17:16:28] *** Parts: travis-ci (~travis-ci@ec2-54-80-46-148.compute-1.amazonaws.com) () [18:34:44] *** Joins: travis-ci (~travis-ci@ec2-34-207-242-93.compute-1.amazonaws.com) [18:34:45] (spdk/master) nvmf: Remove deprecated set_nvmf_target_options rpc (Ben Walker) [18:34:45] Diff URL: https://github.com/spdk/spdk/compare/9fd924cb40e7...d1f943239ee8 [18:34:45] *** Parts: travis-ci (~travis-ci@ec2-34-207-242-93.compute-1.amazonaws.com) () [18:49:26] *** Joins: ziyeyang__ (~ziyeyang@192.55.46.36) [18:49:26] *** ChanServ sets mode: +o ziyeyang__ [20:30:34] Got this in CH TP (passed on Jenkins) and this is unrelated to the patch it came from. I've not seen this before for sure though: "Starting SPDK v19.01-pre / DPDK 18.11.0 initialization... [20:30:34] [ DPDK EAL parameters: nvme_cli -c 0x1 -m 512 --base-virtaddr=0x200000000000 --file-prefix=spdk0 --proc-type=auto ] [20:30:35] EAL: Detected 32 lcore(s) [20:30:35] EAL: Detected 2 NUMA nodes [20:30:35] EAL: Auto-detected process type: SECONDARY [20:30:37] EAL: Multi-process socket /var/run/dpdk/spdk0/mp_socket_369004_6cf839e1303f [20:30:39] EAL: Probing VFIO support... [20:30:41] EAL: VFIO support initialized [20:30:43] nvme.c: 337:nvme_driver_init: *ERROR*: primary process is not started yet [20:30:45] spdk_nvme_connect() failed" [20:30:49] full log: https://ci.spdk.io/spdk/builds/review/e4708ae9b74fb25b07520ba5a56c4cd80f91cd26.1548731229/fedora-03/build.log [20:43:17] *** Quits: ziyeyang__ (~ziyeyang@192.55.46.36) (Ping timeout: 268 seconds) [23:15:32] *** Joins: travis-ci (~travis-ci@ec2-54-224-208-44.compute-1.amazonaws.com) [23:15:33] (spdk/master) scripts: vm_setup.sh add OCF dependencies (Vitaliy Mysak) [23:15:33] Diff URL: https://github.com/spdk/spdk/compare/d1f943239ee8...ec2e2594023b [23:15:33] *** Parts: travis-ci (~travis-ci@ec2-54-224-208-44.compute-1.amazonaws.com) () [23:36:17] *** Joins: ziyeyang__ (ziyeyang@nat/intel/x-inpleevtimuetpuf) [23:36:17] *** ChanServ sets mode: +o ziyeyang__ [23:52:48] *** Joins: travis-ci (~travis-ci@ec2-54-87-88-57.compute-1.amazonaws.com) [23:52:49] (spdk/master) vhost/nvme: fix vhost-nvme target with session feature (Changpeng Liu) [23:52:49] Diff URL: https://github.com/spdk/spdk/compare/ec2e2594023b...f2b480cc24cc [23:52:49] *** Parts: travis-ci (~travis-ci@ec2-54-87-88-57.compute-1.amazonaws.com) () [23:58:32] *** Joins: travis-ci (~travis-ci@ec2-54-91-149-97.compute-1.amazonaws.com) [23:58:33] (spdk/master) ftl: verify LBA format is supported (Konrad Sztyber) [23:58:34] Diff URL: https://github.com/spdk/spdk/compare/f2b480cc24cc...a3777e3f1143 [23:58:34] *** Parts: travis-ci (~travis-ci@ec2-54-91-149-97.compute-1.amazonaws.com) ()