18:08:27 danmcd hey! So I noticed this yesterday. if I do `dladm remove-bridge -l vnic0 bridge0` I get `link vnic0 is not on bridge bridge0`, but I think it should say `bridge bridge0 does not exist`, since, indeed bridge0 does not exist. Should this be considered a bug? 18:11:10 Probably? Lemme look really quickly... 18:13:50 The function this is in is dladm.c:do_remove_bridge(). I'm not 100% sure, however, if libdladm is expressing all the failure modes. 18:14:01 I think the catch-all is that for either: 18:14:13 - link exists, but not on specific bridge 18:14:39 - link doesn't exist, bridge does 18:14:53 - link doesn't exist, and neither does bridge 18:15:20 the same message might get printed. If it's a bug, it's a bug in expressiveness of failures. And do you care if the bridge exists or not if the link doesn't exist? 18:15:25 in my case it was - link exists, but not on that sspecific bridge; however the bridge didn't exist anyway 18:15:57 danmcd no idea :)) coming from ioctl(ERRNOBRIDGE) to this, it's already better :P 18:16:38 If it's a bug it's a low-priority one that'll require some modicum of testing (to make sure older cases still work as expected). 18:17:30 wooh, I like that kind of things! last question: should I fork and work on illumos-gate or should I do the work on the illumos-gate specific to my variant? (in my case, OmniOS) 19:45:37 antranivg: depends on what sort of setup you have for testing and whether it depends on the omnios added kernel features. 19:47:03 you should be able to move patches between variants fairly easily with git cherry-pick or the like, so depending on the nature of the change it might be simpler to develop in the tree you're running. 19:47:20 and then cherry-pick to illumos-gate for final test & integration. 23:40:27 [illumos-gate] 16129 mlxcx_explore_pcam() in mlxcx(4D) doesn't play nice with ConnectX-4 parts anymore -- Jason King