[00:54:47] *** Joins: tkulasek (~tkulasek@192.55.54.42) [01:17:24] *** Joins: gila_ (~gila@5ED74129.cm-7-8b.dynamic.ziggo.nl) [01:18:08] *** Quits: gila (~gila@5ED74129.cm-7-8b.dynamic.ziggo.nl) (Read error: Connection reset by peer) [03:50:13] *** Quits: dlw (~Thunderbi@114.255.44.143) (Ping timeout: 252 seconds) [06:17:50] *** Joins: dlw (~Thunderbi@123.116.59.238) [06:53:14] *** Quits: dlw (~Thunderbi@123.116.59.238) (Ping timeout: 260 seconds) [09:55:16] *** Joins: travis-ci (~travis-ci@ec2-54-197-120-202.compute-1.amazonaws.com) [09:55:17] (spdk/master) test/rpc: add multiple NVMe bdev support in the rpc_config.py (Liang Yan) [09:55:18] Diff URL: https://github.com/spdk/spdk/compare/3148c4807973...04ad0ad9f118 [09:55:18] *** Parts: travis-ci (~travis-ci@ec2-54-197-120-202.compute-1.amazonaws.com) () [09:59:00] *** Quits: tkulasek (~tkulasek@192.55.54.42) (Ping timeout: 245 seconds) [10:02:43] *** Joins: travis-ci (~travis-ci@ec2-54-162-4-12.compute-1.amazonaws.com) [10:02:44] (spdk/master) dpdk: update submodule to latest spdk-18.02 branch (Daniel Verkamp) [10:02:44] Diff URL: https://github.com/spdk/spdk/compare/04ad0ad9f118...b0cfbb4fc096 [10:02:44] *** Parts: travis-ci (~travis-ci@ec2-54-162-4-12.compute-1.amazonaws.com) () [10:03:14] *** Joins: travis-ci (~travis-ci@ec2-54-162-4-12.compute-1.amazonaws.com) [10:03:15] (spdk/master) bdev/virtio: fix memleak in construct_virtio_user_blk_bdev (Dariusz Stojaczyk) [10:03:15] Diff URL: https://github.com/spdk/spdk/compare/b0cfbb4fc096...267318d67104 [10:03:15] *** Parts: travis-ci (~travis-ci@ec2-54-162-4-12.compute-1.amazonaws.com) () [10:19:01] bwalker: can you review the multi-process AER patch? https://review.gerrithub.io/#/c/spdk/spdk/+/410953/ [10:22:00] I only haven't hit +2 because I'm not sure if we need a lock in nvme_ctrlr_async_event_cb [10:23:01] I guess it didn't take one before, so if it is a problem it's not a new one [10:23:27] it should already hold the lock because it's inside polling of the admin queue [10:23:32] ah ok [10:23:52] it would be nice to have some way to assert that we are holding a particular lock, but I couldn't find a way to do that with pthread recursive mutexes [10:24:07] you can sort of hack around it with pthread_mutex_trylock, but that will actually acquire the mutex if it's recursive [10:39:28] is open channel called open channel or lightnvm? [10:40:19] or denali [11:10:26] all three [11:10:26] lol [11:10:39] lightnvm is the name for the library that Matias/Cnex developed [11:10:47] open channel is the name of the current spec [11:10:59] denali is the name for the industry effort to formalize open channel [11:11:08] i would stick with names based around open channel [11:12:13] I just read the whole spec [11:12:19] seems reasonable to me [11:12:26] intuitive, if you know how NAND works [11:12:45] except for the part of how you recognize an open channel ssd :) [11:12:46] yeah, so we should probably change all the "lvnm" to "oc" if that's what we want to use [11:12:48] my only complaint is that they don't define a new value for the PPA command set [11:13:07] drv and I just worked out how an open channel ssd should be identified [11:13:12] we solved world hunger here [11:13:55] it's a quick fix to the current spec and it fits in nicely with SPDK's model [11:21:44] *** Joins: travis-ci (~travis-ci@ec2-54-166-169-32.compute-1.amazonaws.com) [11:21:45] (spdk/master) nvme: make AER callback per-process (Daniel Verkamp) [11:21:45] Diff URL: https://github.com/spdk/spdk/compare/267318d67104...955b295a5838 [11:21:45] *** Parts: travis-ci (~travis-ci@ec2-54-166-169-32.compute-1.amazonaws.com) () [13:55:29] this is how OCSSDs should be detected: https://review.gerrithub.io/#/c/spdk/spdk/+/412102/ [13:56:00] (no actual open channel changes in there - it's just NVMe 1.3b stuff that we should have done anyway) [14:07:09] *** Joins: travis-ci (~travis-ci@ec2-54-162-4-12.compute-1.amazonaws.com) [14:07:10] (spdk/master) test/vm_setup.sh: update nvml-tools -> pmempool (Daniel Verkamp) [14:07:10] Diff URL: https://github.com/spdk/spdk/compare/955b295a5838...700477897a5b [14:07:10] *** Parts: travis-ci (~travis-ci@ec2-54-162-4-12.compute-1.amazonaws.com) () [14:08:11] *** Joins: travis-ci (~travis-ci@ec2-54-162-4-12.compute-1.amazonaws.com) [14:08:12] (spdk/master) jsonrpc: make "jsonrpc" request field optional (Daniel Verkamp) [14:08:12] Diff URL: https://github.com/spdk/spdk/compare/700477897a5b...aa67900a2e42 [14:08:12] *** Parts: travis-ci (~travis-ci@ec2-54-162-4-12.compute-1.amazonaws.com) () [14:09:27] bwalker - can you look at https://review.gerrithub.io/#/c/spdk/spdk/+/411740/ [14:09:52] *** Joins: travis-ci (~travis-ci@ec2-54-166-169-32.compute-1.amazonaws.com) [14:09:53] (spdk/master) scripts/pkgdep.sh: fix astyle package name on RH (Daniel Verkamp) [14:09:53] Diff URL: https://github.com/spdk/spdk/compare/aa67900a2e42...3d22c5088eba [14:09:53] *** Parts: travis-ci (~travis-ci@ec2-54-166-169-32.compute-1.amazonaws.com) () [14:11:16] bwalker: are you just saying the open channel commands should have their own command set and not reuse NVM? [14:14:04] I think that's what bwalker thinks, but it's also not what the open channel spec currently says :) [14:15:56] *** Joins: travis-ci (~travis-ci@ec2-54-197-120-202.compute-1.amazonaws.com) [14:15:57] (spdk/master) include/lvol.h: add comments for public APIs (Yanbo Zhou) [14:15:57] Diff URL: https://github.com/spdk/spdk/compare/3d22c5088eba...b31330acaacd [14:15:57] *** Parts: travis-ci (~travis-ci@ec2-54-197-120-202.compute-1.amazonaws.com) () [14:16:09] exactly [14:16:24] that patch doesn't help us today [14:20:48] yeah, I think we should get the infrastructure in place anyway, but it doesn't have any real-world impact as of now [14:20:59] agreed [14:24:53] jimharris: is this issue resolved after your RTE_BAD_IOVA fix? https://github.com/spdk/spdk/issues/260 [14:28:57] the open channel specification has a split personality from my reading [14:29:20] it refers to the "physical page address command set" [14:29:23] and even has a table of its commands [14:29:43] but it never defines it as a command set in terms of CAP.CSS and CC.CSS [14:30:08] I think defining it as an official command set is a trivial addition to the current OC spec [14:31:06] and doesn't materially change how it works, except now there is a real way to detect support [14:32:28] it's already the case that the drive accepts "physical page address" commands or "NVM" commands, but not both [14:32:45] and while the read/write/trim are similar to the NVM versions, they aren't the same [14:33:28] it may be a trivial addition but none of the OC SSDs out there today work that way [14:34:14] the OC SSDs out there today are just FPGAs [14:34:15] i would be shocked if this doesn't get resolved if/when OC becomes part of the official NVMe spec though [14:34:35] I'm not worried about making it work on the ones out there today - we can use a quirk [14:51:10] *** Joins: travis-ci (~travis-ci@ec2-54-162-4-12.compute-1.amazonaws.com) [14:51:11] (spdk/master) nvmf/rdma: Update handling of RDMA_CM_EVENTS (Ben Walker) [14:51:12] Diff URL: https://github.com/spdk/spdk/compare/b31330acaacd...f0191857a5e8 [14:51:12] *** Parts: travis-ci (~travis-ci@ec2-54-162-4-12.compute-1.amazonaws.com) () [14:57:39] *** Joins: travis-ci (~travis-ci@ec2-54-166-169-32.compute-1.amazonaws.com) [14:57:40] (spdk/master) bdev: Make the QoS channel use a regular channel (Ben Walker) [14:57:40] Diff URL: https://github.com/spdk/spdk/compare/f0191857a5e8...aac1f5f9349f [14:57:40] *** Parts: travis-ci (~travis-ci@ec2-54-166-169-32.compute-1.amazonaws.com) () [16:05:53] *** Joins: Shuhei (caf6fc61@gateway/web/freenode/ip.202.246.252.97) [16:06:15] *** Joins: travis-ci (~travis-ci@ec2-54-166-169-32.compute-1.amazonaws.com) [16:06:16] (spdk/master) nbd/rpc: make get_nbd_disks nbd_device param optional (Daniel Verkamp) [16:06:17] Diff URL: https://github.com/spdk/spdk/compare/aac1f5f9349f...e95f0bb75654 [16:06:17] *** Parts: travis-ci (~travis-ci@ec2-54-166-169-32.compute-1.amazonaws.com) () [16:11:11] *** Joins: travis-ci (~travis-ci@ec2-54-166-169-32.compute-1.amazonaws.com) [16:11:12] (spdk/master) bdev/lvol: inherit need_aligned_buffer of base bdev (Daniel Verkamp) [16:11:12] Diff URL: https://github.com/spdk/spdk/compare/e95f0bb75654...475165dffe97 [16:11:12] *** Parts: travis-ci (~travis-ci@ec2-54-166-169-32.compute-1.amazonaws.com) () [16:11:30] *** Joins: travis-ci (~travis-ci@ec2-54-162-4-12.compute-1.amazonaws.com) [16:11:31] (spdk/master) bdev/virtio_blk: fix double free on init failure (Dariusz Stojaczyk) [16:11:32] Diff URL: https://github.com/spdk/spdk/compare/475165dffe97...3949140b2eea [16:11:32] *** Parts: travis-ci (~travis-ci@ec2-54-162-4-12.compute-1.amazonaws.com) () [16:14:09] ziyeyang: I was off yesterday. now I'm reading your scsi patch and I'll update to you later. [18:18:28] *** Joins: dlw (~Thunderbi@114.255.44.143) [18:24:53] *** Quits: dlw (~Thunderbi@114.255.44.143) (Remote host closed the connection) [18:27:18] *** Joins: dlw (~Thunderbi@114.255.44.143) [18:33:53] *** Quits: dlw (~Thunderbi@114.255.44.143) (Remote host closed the connection) [18:34:17] *** Joins: dlw (~Thunderbi@114.255.44.143) [20:14:00] *** Quits: Shuhei (caf6fc61@gateway/web/freenode/ip.202.246.252.97) (Ping timeout: 260 seconds) [20:19:28] *** Joins: Shuhei (caf6fc61@gateway/web/freenode/ip.202.246.252.97) [20:37:55] *** Quits: Shuhei (caf6fc61@gateway/web/freenode/ip.202.246.252.97) (Ping timeout: 260 seconds) [20:50:48] *** Joins: Shuhei (caf6fc61@gateway/web/freenode/ip.202.246.252.97) [21:14:04] *** Quits: Shuhei (caf6fc61@gateway/web/freenode/ip.202.246.252.97) (Ping timeout: 260 seconds) [23:22:49] *** Joins: Shuhei (caf6fc61@gateway/web/freenode/ip.202.246.252.97)