[00:03:48] *** Joins: tkulasek (tkulasek@nat/intel/x-cvaoasgtxsvntgcc) [00:31:13] *** Quits: tkulasek (tkulasek@nat/intel/x-cvaoasgtxsvntgcc) (Ping timeout: 248 seconds) [00:35:24] *** Joins: tkulasek (~tkulasek@134.134.139.73) [00:36:20] *** Quits: tkulasek (~tkulasek@134.134.139.73) (Remote host closed the connection) [00:38:06] *** Joins: tkulasek (tkulasek@nat/intel/x-bzincyxldttxepon) [00:40:01] *** Joins: tkulasek_ (tkulasek@nat/intel/x-iciknmxcsyrsxymm) [00:40:01] *** Quits: tkulasek (tkulasek@nat/intel/x-bzincyxldttxepon) (Remote host closed the connection) [01:04:49] *** Quits: tkulasek_ (tkulasek@nat/intel/x-iciknmxcsyrsxymm) (Quit: Leaving) [01:05:26] *** Joins: tkulasek (~tkulasek@134.134.139.73) [02:05:07] *** Joins: tkulasek_ (~tkulasek@134.134.139.73) [02:05:22] *** Quits: tkulasek_ (~tkulasek@134.134.139.73) (Remote host closed the connection) [02:05:35] *** Joins: tkulasek_ (tkulasek@nat/intel/x-gzbxdunrzpmhmznx) [02:08:10] *** Quits: tkulasek (~tkulasek@134.134.139.73) (Ping timeout: 264 seconds) [02:39:35] *** Quits: tkulasek_ (tkulasek@nat/intel/x-gzbxdunrzpmhmznx) (Remote host closed the connection) [02:39:46] *** Joins: tkulasek (tkulasek@nat/intel/x-jmniztmqtxwwbgzu) [03:26:09] on each merged patch on gerrithub there's now a "INCLUDED IN" button showing a single branch - master [03:26:37] does it mean that a single commit can now be merged to multiple branches? [03:53:28] *** Joins: guerby_ (~guerby@ip165.tetaneutral.net) [03:56:33] *** Quits: guerby (~guerby@april/board/guerby) (Remote host closed the connection) [04:00:24] *** Quits: guerby_ (~guerby@ip165.tetaneutral.net) (Excess Flood) [04:00:32] *** Joins: guerby_ (~guerby@ip165.tetaneutral.net) [05:03:04] *** guerby_ is now known as guerby [05:03:16] *** Quits: guerby (~guerby@ip165.tetaneutral.net) (Changing host) [05:03:16] *** Joins: guerby (~guerby@april/board/guerby) [06:32:17] *** Quits: tomzawadzki (tomzawadzk@nat/intel/x-ivjndhpnoykxyuou) (Ping timeout: 256 seconds) [07:07:31] *** Quits: nKumar (sid239884@gateway/web/irccloud.com/x-qgtvcjojffypjtzl) () [07:07:44] *** Joins: nKumar (sid239884@gateway/web/irccloud.com/x-bczbwisazlxnwxlp) [08:02:36] *** Quits: tkulasek (tkulasek@nat/intel/x-jmniztmqtxwwbgzu) (Ping timeout: 255 seconds) [09:06:25] *** Joins: travis-ci (~travis-ci@ec2-54-198-157-128.compute-1.amazonaws.com) [09:06:26] (spdk/master) vhost/nvme: remove unnecessary vva casts (Daniel Verkamp) [09:06:26] Diff URL: https://github.com/spdk/spdk/compare/9c7d842cc429...5d5e4028100d [09:06:26] *** Parts: travis-ci (~travis-ci@ec2-54-198-157-128.compute-1.amazonaws.com) () [09:08:32] *** Joins: travis-ci (~travis-ci@ec2-54-198-157-128.compute-1.amazonaws.com) [09:08:33] (spdk/master) env_dpdk: fix build with DPDK 16.07 (Daniel Verkamp) [09:08:33] Diff URL: https://github.com/spdk/spdk/compare/5d5e4028100d...5ac612990803 [09:08:33] *** Parts: travis-ci (~travis-ci@ec2-54-198-157-128.compute-1.amazonaws.com) () [09:12:40] drv: could we use "Virtio Block Disk" as a bdev->product_name? [09:12:57] sure, we aren't super consistent with those, but that sounds fine to me [09:12:58] i guess if we're ever going to change that, it should be now [09:13:08] as long as it fits in the SCSI field that it ends up in, which I think is 20 chars? [09:14:21] 16 [09:15:38] oh, this means that even "Virtio SCSI Disk" doesn't fit [09:16:50] it's probably not the end of the world if it is truncated [09:16:56] but we should at least keep it in mind [09:17:55] I think scsi_bdev is the only place we really use the product name right now [09:21:23] jimharris: please review https://review.gerrithub.io/#/c/spdk/spdk/+/408771/ - we should definitely get this in before release [09:22:11] *** Joins: travis-ci (~travis-ci@ec2-54-198-157-128.compute-1.amazonaws.com) [09:22:12] (spdk/master) bdev/gpt: allocate buffer with correct alignment (Daniel Verkamp) [09:22:12] Diff URL: https://github.com/spdk/spdk/compare/5ac612990803...b7060cb21ac0 [09:22:12] *** Parts: travis-ci (~travis-ci@ec2-54-198-157-128.compute-1.amazonaws.com) () [09:23:11] well, i'm not touching "VirtioBlk Disk" then [09:26:17] long ago, they used to be "LUN" instead of "Disk" [09:26:24] saving one character :) [09:33:16] interesting not-super-important thing... was testing crypto vbdev with various core masks and numbers of underlying bdevs. With 8 malloc disks and a core mask of either 1 of ff things seem to work. Other values cause some exceptions in malloc/bdev, heap use after free in touching a channel that was previously freed by crypto vbddev, HOWEVER [09:33:50] *** Joins: tomzawadzki (tomzawadzk@nat/intel/x-srqsidkdafuxosll) [09:34:11] if I use split to create 8 bdevs on top of NVMe device all core masks and combinations of bdevs work without issue... haven't really looked into it much. Any existing known issues with malloc bdev module and multiple core/bdev usage? [09:36:04] peluse: do you have i/oat hardware on that machine? (malloc will use the copy engine module, which will use ioat if available) [09:36:38] yeah [09:36:45] you can try adding an [Ioat] Disable Yes to your config file [09:36:56] if that "fixes" the problem, it's probably something in the ioat/copy engine layer [09:36:59] and I think I saw some ioat shit in the dump. I'll try that real quick [09:40:48] it does indeed "fix" it [09:43:09] I'll look further into it, thanks [09:43:59] I managed to review the nvmf shutdown patch again [09:44:07] taking a look at the remaining 18.04 patches [09:44:22] *** Joins: travis-ci (~travis-ci@ec2-54-198-157-128.compute-1.amazonaws.com) [09:44:23] (spdk/master) vhost-nvme: reserve a bigger enough opcode to avoid conflict with QEMU (Changpeng Liu) [09:44:23] Diff URL: https://github.com/spdk/spdk/compare/b7060cb21ac0...c2b66a83bc9a [09:44:23] *** Parts: travis-ci (~travis-ci@ec2-54-198-157-128.compute-1.amazonaws.com) () [09:49:32] I think some of the ones that just got tagged as 18.04 are not actually release critical [09:49:52] the getpid wrapping thing is definitely not [09:54:34] I'm reviewing the stat rpc right now [10:01:07] I put some comments on it: https://review.gerrithub.io/#/c/spdk/spdk/+/393130/ [10:07:59] hmm, yeah, I think it makes sense for the user to pass the stat object to the initial spdk_bdev_get_device_stat() call [10:08:06] we should get the API right at least before we merge thi [10:08:36] yeah - the API thing needs to be right. But also it added a dependency on the event framework [10:08:39] that wasn't necessary [10:09:28] bwalker, jimharris: any opinions on https://review.gerrithub.io/#/c/spdk/spdk/+/408084/ (lvol delete RPC) [10:09:42] I'd like to get the RPC method in place, at least - the implementation still isn't right, but it's not worse than what we have now [10:22:05] I don't see any issue with the RPC interface you added [10:49:29] *** Quits: tomzawadzki (tomzawadzk@nat/intel/x-srqsidkdafuxosll) (Ping timeout: 260 seconds) [10:50:41] *** Joins: tomzawadzki (~tomzawadz@192.55.54.42) [11:22:56] drv: yeah - reverting that freebsd commit fixed the problem in my vm [11:23:10] ok, I'm good with putting that revert into our DPDK submodule for now [11:23:16] and updating the submodule before the spdk 18.04 release [11:23:25] my VM was actually super old - 10.2-RC2 - from 2015 when I did the original port of the driver to FreeBSD :) [11:23:34] but it reproduced with latest SPDK [11:23:51] i'll respond to rajesh on the mailing list too [12:40:35] *** Joins: tzawadzki (~tomzawadz@192.55.54.38) [12:40:36] *** Quits: tomzawadzki (~tomzawadz@192.55.54.42) (Remote host closed the connection) [12:53:10] *** Quits: tzawadzki (~tomzawadz@192.55.54.38) (Remote host closed the connection) [14:36:59] *** Joins: travis-ci (~travis-ci@ec2-54-198-157-128.compute-1.amazonaws.com) [14:37:00] (spdk/master) lvol: add destroy_lvol_bdev RPC method (Daniel Verkamp) [14:37:00] Diff URL: https://github.com/spdk/spdk/compare/c2b66a83bc9a...f9d52c9ebdaa [14:37:00] *** Parts: travis-ci (~travis-ci@ec2-54-198-157-128.compute-1.amazonaws.com) () [14:41:25] *** Joins: travis-ci (~travis-ci@ec2-54-156-92-113.compute-1.amazonaws.com) [14:41:26] (spdk/master) env: add malloc variants with DMA/shareable flags (zkhatami88) [14:41:26] Diff URL: https://github.com/spdk/spdk/compare/f9d52c9ebdaa...a77cd3f75c0f [14:41:26] *** Parts: travis-ci (~travis-ci@ec2-54-156-92-113.compute-1.amazonaws.com) () [14:49:34] *** Joins: travis-ci (~travis-ci@ec2-54-156-92-113.compute-1.amazonaws.com) [14:49:35] (spdk/master) nbd: add JSON config dump (Pawel Wodkowski) [14:49:35] Diff URL: https://github.com/spdk/spdk/compare/a77cd3f75c0f...5b424282c1d2 [14:49:35] *** Parts: travis-ci (~travis-ci@ec2-54-156-92-113.compute-1.amazonaws.com) () [14:57:31] *** Joins: travis-ci (~travis-ci@ec2-54-156-92-113.compute-1.amazonaws.com) [14:57:32] (spdk/master) Make iteration over I/O types safe against further type additions. (Andrey Kuzmin) [14:57:32] Diff URL: https://github.com/spdk/spdk/compare/c769e03fd760...d5a9fca816cc [14:57:32] *** Parts: travis-ci (~travis-ci@ec2-54-156-92-113.compute-1.amazonaws.com) () [16:00:15] *** Joins: Shuhei (caf6fc61@gateway/web/freenode/ip.202.246.252.97) [19:41:37] *** Joins: travis-ci (~travis-ci@ec2-54-198-19-7.compute-1.amazonaws.com) [19:41:38] (spdk/master) vhost/nvme: fix doorbell buffer size calculation (Daniel Verkamp) [19:41:38] Diff URL: https://github.com/spdk/spdk/compare/9acb995ededc...130307862b12 [19:41:38] *** Parts: travis-ci (~travis-ci@ec2-54-198-19-7.compute-1.amazonaws.com) () [20:38:29] *** Quits: Shuhei (caf6fc61@gateway/web/freenode/ip.202.246.252.97) (Ping timeout: 260 seconds) [23:28:33] *** Joins: Shuhei (caf6fc61@gateway/web/freenode/ip.202.246.252.97) [23:48:35] *** Joins: tkulasek (~tkulasek@192.55.54.44)