[08:15:58] *** ChanServ sets mode: +o jimharris [08:33:21] *** Joins: stefanha (~stefanha@yuzuki.vmsplice.net) [08:44:05] *** Joins: tomzawadzki (~tomzawadz@134.134.139.76) [09:35:21] hi stefanha [09:58:50] hi jimharris :) [10:04:30] jimharris: Given the vhost-user-blk and vhost-user-nvme activity, I figured it would be a good idea to subscribe to SPDK mailing lists and follow activity. [10:06:40] drv: Do we know each other from Etherboot/gPXE/SYSLINUX? :) [10:06:40] that's great stefanha - changpe1 (Changpeng Liu) has some additional performance/efficiency data for vhost-user-nvme, I asked him offline if it's ready to share on the qemu mailing list [10:06:55] ha - yeah, drv was just telling me about that in the lab [10:07:05] Nice :) [10:16:26] *** Quits: tomzawadzki (~tomzawadz@134.134.139.76) (Ping timeout: 256 seconds) [10:21:02] hey stefanha! [10:21:21] long time no see :) [10:33:58] bwalker: i pushed first three patches from my series - fourth patch failed due to scan-build issue - if you're not already look at it I can do it [10:35:01] I'll do it - have it pulled up [10:43:31] drv: can you look at https://review.gerrithub.io/#/c/396790/ [10:47:51] hmm, this is getting a little icky [10:48:10] as long as it still works when running 'make' before 'configure', I'm OK with it [10:49:09] seems like distros that ship only Python 3 should still have a 'python' binary that points to python3, but I understand that has compatibility problems with scripts expecting 'python' to be Python 2 [10:50:19] or we could just stop supporting systems without a 'python3' binary, but I'd like to avoid that if possible [10:59:00] I added some comments to the review [11:05:40] thanks [11:06:08] drv: we should remove 18.01 tag from most of gangcao's qos patches right? [11:06:14] agreed [11:06:25] we probably shouldn't merge any of them if we're not going to put the whole series in [11:06:34] ok - going to do that now [11:13:55] *** Joins: AlanA (94571712@gateway/web/freenode/ip.148.87.23.18) [12:42:57] *** Joins: sethhowe (sethhowe@nat/intel/x-cpdfswcvvqmxwtjo) [12:48:15] *** Joins: nvme_noob (d05b0202@gateway/web/freenode/ip.208.91.2.2) [12:53:53] @dev Thanks for working on the ipv6 support(https://review.gerrithub.io/#/c/396494/). When can I expect the build to be available with that change? I am sorry I don't understand the spdk dev process and might be asking soemthing obvious.. [12:54:10] @drv * [12:54:57] the patch has been merged to the master branch now (as of today), and it will be in the next release (SPDK v18.01) [12:55:23] Is there an expected date for that release? [12:55:42] well, ideally in January, but we are still working on getting the last few features merged [12:55:53] definitely in the next week or so [12:56:09] you can download the code from the master branch today if you prefer not to wait [12:56:45] yeah I am considering that.. Let me give that a try [12:57:34] Thanks @drv for the timely response and for working on that patch [13:11:05] sure, let us know if you get a chance to test it out :) [15:20:08] hmm, test/vhost/initiator/blockdev.sh seems to use a mix of $vm_no and $vm_num, seemingly unintentionally [15:20:12] not sure how this works right now [15:22:11] or maybe it's intentional, and $vm_num is supposed to be set by vm_setup() in test/vhost/common/common.sh, but the live migration patch makes $vm_num a local, so that fails [15:22:35] either way, seems like we should clean up the variable names so the intent is clear [15:46:18] *** Joins: Shuhei (caf6fc61@gateway/web/freenode/ip.202.246.252.97) [16:03:25] jimharris, not a big priority item but see my comment on https://review.gerrithub.io/#/c/394116/ when you get a chance [16:08:01] yeah, I'm in the same boat - I agree that "benchmarking" is not quite the right name for the content it has right now, but we can probably rename it later [16:08:10] or we can come up with a good name now - shouldn't be too hard, hopefully [16:13:06] peluse: I made some comments on a previous rev of the match patch (just noticed you uploaded a new rev in between, don't want the comments to get missed) [16:14:55] thanks, yeah a few quick things... the trailing white space was in both a file that I checked in and also in the output file that the CLI generates so I fixed the checked in file and added a line to strip the trailers from the CLI output instead of making a CLI change for no good reason [16:16:00] and for the DPDK and EAL output stuff, you can't just delete it from the match file or it won't match anything when compared to the real output. You'd also have to strip it from the real output so that's why I added the .ignore file feature - much less work :) [16:16:52] drv, and no hurry on that one of course. Just been on my todo list for to long so wanted to get it going finally... [18:40:24] *** Quits: nvme_noob (d05b0202@gateway/web/freenode/ip.208.91.2.2) (Ping timeout: 260 seconds) [22:16:21] *** Joins: ziyeyang_ (ziyeyang@nat/intel/x-uxhhuwvpsvvamrtg) [23:51:29] *** Joins: tomzawadzki (~tomzawadz@192.55.54.44)