Switch to DuckDuckGo Search
   June 9, 2020  
< | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 | 12 | 13 | 14 | 15 | 16 | 17 | 18 | 19 | 20 | 21 | 22 | 23 | 24 | 25 | 26 | 27 | 28 | 29 | 30 | >


NOTICE: This channel is no longer actively logged.

Toggle Join/Part | bottom
[00:15:11] *** kohju <kohju!~kohju@gw.justplayer.com> has quit IRC (Ping timeout: 265 seconds)
[00:15:33] *** kohju <kohju!~kohju@gw.justplayer.com> has joined #illumos
[00:16:23] *** lblume <lblume!~lblume@greenviolet/laoyijiehe/lblume> has quit IRC (Ping timeout: 272 seconds)
[00:19:11] *** amrfrsh <amrfrsh!~Thunderbi@213.232.87.96> has quit IRC (Quit: amrfrsh)
[00:29:06] *** bahamas10 <bahamas10!~dave@cpe-72-231-182-75.nycap.res.rr.com> has quit IRC (Remote host closed the connection)
[00:29:31] *** Obscurax_ <Obscurax_!~Obscurax@unaffiliated/obscurax> has joined #illumos
[00:29:58] *** Cthulhux <Cthulhux!cthulhu@piratenpartei/ni/tux> has quit IRC (Ping timeout: 256 seconds)
[00:30:14] *** Obscurax <Obscurax!~Obscurax@unaffiliated/obscurax> has quit IRC (Ping timeout: 240 seconds)
[00:31:18] *** Cthulhux <Cthulhux!cthulhu@rosaelefanten.org> has joined #illumos
[00:43:02] *** amrfrsh <amrfrsh!~Thunderbi@213.232.87.96> has joined #illumos
[00:47:17] *** freakazoid0223 <freakazoid0223!~matt@pool-96-227-117-103.phlapa.fios.verizon.net> has left #illumos ("Leaving")
[00:49:19] *** amrfrsh <amrfrsh!~Thunderbi@213.232.87.96> has quit IRC (Ping timeout: 272 seconds)
[01:00:06] *** amrfrsh <amrfrsh!~Thunderbi@213.232.87.96> has joined #illumos
[01:07:48] *** psarria__ <psarria__!~psarria@74.red-79-146-99.dynamicip.rima-tde.net> has joined #illumos
[01:10:12] *** psarria_ <psarria_!~psarria@74.red-79-146-99.dynamicip.rima-tde.net> has quit IRC (Ping timeout: 256 seconds)
[01:12:40] *** dodobrain <dodobrain!~dodobrain@unaffiliated/freakabcd> has joined #illumos
[01:34:37] *** Cthulhux <Cthulhux!cthulhu@rosaelefanten.org> has quit IRC (Ping timeout: 246 seconds)
[01:34:55] *** amrfrsh <amrfrsh!~Thunderbi@213.232.87.96> has quit IRC (Ping timeout: 260 seconds)
[01:36:51] *** Cthulhux <Cthulhux!cthulhu@rosaelefanten.org> has joined #illumos
[01:47:11] *** amrfrsh <amrfrsh!~Thunderbi@213.232.87.96> has joined #illumos
[02:06:08] *** gh34 <gh34!~textual@cpe-184-58-181-106.wi.res.rr.com> has quit IRC (Quit: Textual IRC Client: www.textualapp.com)
[02:17:27] <danmcd> If anyone has a stock OI or illumos-gate running, could you please utter:
[02:17:30] <danmcd> echo "::print -ath struct proc" | mdb -k | grep HOLE | wc -l
[02:17:34] <danmcd> and tell me what the count is?
[02:21:16] * danmcd noticed one could eliminate some of the holes and make sizeof (proc_t) exactly 4k.
[02:21:51] <pmooney> danmcd: 19
[02:22:16] <danmcd> That's with stock, huh? Same as smartos and omnios which both have the new thing JBK's upstreaming.
[02:22:29] <danmcd> It bugs me, but it's probably not worth fixing.
[02:23:33] <pmooney> danmcd: sizeof 'struct proc' is < 4k on stock
[02:24:13] <pmooney> at least as of the last build I did for this thing, which is circa a week ago or so
[02:27:50] <danmcd> It's 4136 on SmartOS and OmniOS
[02:28:15] <pmooney> probably LX related, if I had to guess
[02:28:26] <danmcd> (You're likely right.) But even on those two, killing some holes might reduce it.
[02:29:11] <pmooney> I think if someone was going to go off on that quest, they'd want to think pretty hard about optimizing for cachelines in addition to size
[02:29:42] <danmcd> Like I said, no big, and yes, cachelines are the bigger fish to fry. (Hell, there's still false-sharing in TCP/IP hash buckets, e.g. ipsaf_t.)
[02:29:48] <pmooney> and definitely about preserving the readable ordering with respect to what parts are covered by which locks
[02:29:57] <danmcd> Jerry fixed it for conn_t, IIRC.
[02:30:00] <pmooney> (that's the part I'm post worried about)
[02:33:02] <danmcd> Thanks for cashing the reality check. I'm gonna bail.
[02:43:17] <pmooney> np
[02:56:47] *** jocthbr <jocthbr!~salci@138-122-44-143.host.cicloti.com.br> has quit IRC (Ping timeout: 246 seconds)
[03:02:17] *** liv3010m <liv3010m!~liv3010m@77-72-245-190.fibertel.com.ar> has quit IRC (Ping timeout: 260 seconds)
[03:03:44] *** liv3010m <liv3010m!~liv3010m@77-72-245-190.fibertel.com.ar> has joined #illumos
[03:06:01] *** jocthbr <jocthbr!~salci@138-122-44-143.host.cicloti.com.br> has joined #illumos
[03:10:25] *** benfelin <benfelin!~benfelin@cpc91226-cmbg18-2-0-cust35.5-4.cable.virginm.net> has quit IRC (Ping timeout: 264 seconds)
[03:10:58] *** jcea <jcea!~Thunderbi@2001:bc8:2ecd:caed:7670:6e00:7670:6e00> has quit IRC (Quit: jcea)
[03:26:07] *** spicywolf <spicywolf!~spicywolf@2601:281:c000:4ad2:4494:ee03:6e08:a786> has joined #illumos
[03:27:35] *** spicywolf <spicywolf!~spicywolf@2601:281:c000:4ad2:4494:ee03:6e08:a786> has quit IRC (Client Quit)
[03:27:52] *** spicywolf <spicywolf!~spicywolf@2601:281:c000:4ad2:4494:ee03:6e08:a786> has joined #illumos
[05:41:07] *** BOKALDO <BOKALDO!~BOKALDO@91.105.119.15> has joined #illumos
[06:03:04] *** amrfrsh <amrfrsh!~Thunderbi@213.232.87.96> has quit IRC (Ping timeout: 246 seconds)
[06:31:13] *** wl_ <wl_!~wl_@2605:6000:1b0c:4825::87c> has joined #illumos
[06:56:57] *** cypa <cypa!~cypam]_@5.79.181.101> has joined #illumos
[07:05:06] *** cypa <cypa!~cypam]_@5.79.181.101> has quit IRC (Remote host closed the connection)
[07:36:27] *** leah2 <leah2!~leah@vuxu.org> has quit IRC (Ping timeout: 260 seconds)
[07:36:49] *** tomww <tomww!~tom@unaffiliated/tomww> has quit IRC (Ping timeout: 256 seconds)
[07:55:14] *** leah2 <leah2!~leah@vuxu.org> has joined #illumos
[08:03:14] *** leah2 <leah2!~leah@vuxu.org> has quit IRC (Ping timeout: 256 seconds)
[08:05:15] *** tomww <tomww!~tom@unaffiliated/tomww> has joined #illumos
[08:08:27] *** leah2 <leah2!~leah@vuxu.org> has joined #illumos
[08:38:01] *** BOKALDO <BOKALDO!~BOKALDO@91.105.119.15> has quit IRC (Ping timeout: 264 seconds)
[08:42:10] *** alanc <alanc!~alanc@inet-hqmc02-o.oracle.com> has quit IRC (Remote host closed the connection)
[08:42:37] *** alanc <alanc!~alanc@inet-hqmc02-o.oracle.com> has joined #illumos
[08:59:30] *** storkone <storkone!~storkone@a80-101-69-245.adsl.xs4all.nl> has quit IRC (Read error: Connection reset by peer)
[09:00:10] *** storkone <storkone!~storkone@a80-101-69-245.adsl.xs4all.nl> has joined #illumos
[09:05:22] *** lblume <lblume!~lblume@greenviolet/laoyijiehe/lblume> has joined #illumos
[09:16:37] *** amrfrsh <amrfrsh!~Thunderbi@213.232.87.96> has joined #illumos
[09:28:03] *** andy_js <andy_js!~andy@94.11.185.123> has joined #illumos
[09:32:55] *** jimklimov <jimklimov!~jimklimov@78.80.224.132> has joined #illumos
[09:43:18] *** eki <eki!~eki@dsl-hkibng41-54f858-46.dhcp.inet.fi> has quit IRC (Quit: leaving)
[09:43:39] *** ptribble <ptribble!~ptribble@cpc92716-cmbg20-2-0-cust138.5-4.cable.virginm.net> has joined #illumos
[09:45:19] *** eki <eki!~eki@dsl-hkibng41-54f858-46.dhcp.inet.fi> has joined #illumos
[09:53:48] *** clapont <clapont!~clapont@unaffiliated/clapont> has quit IRC (Ping timeout: 256 seconds)
[09:55:35] *** clapont <clapont!~clapont@unaffiliated/clapont> has joined #illumos
[10:00:38] *** neuroserve <neuroserve!~toens@ip-88-152-243-61.hsi03.unitymediagroup.de> has quit IRC (Ping timeout: 260 seconds)
[10:53:38] *** fgudin_ <fgudin_!fgudin@sdf-eu.org> has joined #illumos
[10:59:31] *** BOKALDO <BOKALDO!~BOKALDO@91.105.119.15> has joined #illumos
[11:39:13] *** cantstanya <cantstanya!~chatting@gateway/tor-sasl/cantstanya> has quit IRC (Remote host closed the connection)
[11:40:57] *** cantstanya <cantstanya!~chatting@gateway/tor-sasl/cantstanya> has joined #illumos
[11:41:07] *** spicywolf <spicywolf!~spicywolf@2601:281:c000:4ad2:4494:ee03:6e08:a786> has quit IRC (Ping timeout: 260 seconds)
[12:04:15] *** neuroserve <neuroserve!~toens@ip-84-118-130-44.unity-media.net> has joined #illumos
[12:23:03] *** benfelin <benfelin!~benfelin@cpc91226-cmbg18-2-0-cust35.5-4.cable.virginm.net> has joined #illumos
[12:24:06] *** amrfrsh <amrfrsh!~Thunderbi@213.232.87.96> has quit IRC (Remote host closed the connection)
[12:24:21] *** amrfrsh <amrfrsh!~Thunderbi@213.232.87.96> has joined #illumos
[12:31:35] <Agnar> erm, does our nfs server support versions 4.1 and 4.2?
[12:33:34] <wilbury> freebsd's does :-P
[12:47:26] <Agnar> well, good, but illumos? :)
[12:51:18] <v_a_b> agnar I don't think so. NFS v4.1 is pNFS which I guess illumos doesn't have.
[12:51:38] <Agnar> *sigh*
[12:52:35] <v_a_b> NFS v4.2 is quite new. I actually had to look up the new features just now to find the differences. I would like to have sparse files over NFS ;-)
[12:54:05] <Agnar> server side copy and labeled nfs
[12:54:37] <v_a_b> yep, nice stuff
[12:58:09] *** cantstanya <cantstanya!~chatting@gateway/tor-sasl/cantstanya> has quit IRC (Remote host closed the connection)
[13:00:32] *** cantstanya <cantstanya!~chatting@gateway/tor-sasl/cantstanya> has joined #illumos
[13:27:10] *** BOKALDO_ <BOKALDO_!~BOKALDO@91.105.127.77> has joined #illumos
[13:30:47] *** BOKALDO <BOKALDO!~BOKALDO@91.105.119.15> has quit IRC (Ping timeout: 260 seconds)
[13:32:11] *** benfelin <benfelin!~benfelin@cpc91226-cmbg18-2-0-cust35.5-4.cable.virginm.net> has quit IRC (Ping timeout: 260 seconds)
[13:40:45] *** chandlore____ <chandlore____!sid259138@gateway/web/irccloud.com/x-awlaqtkpyjnxhqrn> has quit IRC (Quit: Connection closed for inactivity)
[13:43:53] *** benfelin <benfelin!~benfelin@cpc91226-cmbg18-2-0-cust35.5-4.cable.virginm.net> has joined #illumos
[14:22:23] *** jollyd1 <jollyd1!~alarcher@lfbn-nic-1-142-60.w2-15.abo.wanadoo.fr> has joined #illumos
[14:24:08] *** jollyd <jollyd!~alarcher@lfbn-nic-1-142-60.w2-15.abo.wanadoo.fr> has quit IRC (Ping timeout: 272 seconds)
[14:28:08] *** gh34 <gh34!~textual@cpe-184-58-181-106.wi.res.rr.com> has joined #illumos
[14:40:18] *** tsoome <tsoome!~tsoome@0c47-94d8-80bf-5902-2f80-4a40-07d0-2001.sta.estpak.ee> has quit IRC (Quit: My MacBook has gone to sleep. ZZZzzz…)
[14:45:54] *** jollyd1 <jollyd1!~alarcher@lfbn-nic-1-142-60.w2-15.abo.wanadoo.fr> has quit IRC (Quit: Leaving.)
[15:00:53] *** phyre <phyre!~phyre___@78.30.22.107> has joined #illumos
[15:41:29] *** vgusev_ <vgusev_!~vgusev@188.187.60.230> has quit IRC (Read error: Connection reset by peer)
[15:42:06] *** vgusev <vgusev!~vgusev@188.187.60.230> has joined #illumos
[15:42:28] * sjorge googles pNFS
[15:57:11] *** igork <igork!~igork@178.162.50.254> has quit IRC (Ping timeout: 256 seconds)
[15:57:59] *** igork <igork!~igork@91.204.56.74> has joined #illumos
[15:59:21] *** tsoome <tsoome!~tsoome@fc35-68b8-42e2-2989-2f80-4a40-07d0-2001.sta.estpak.ee> has joined #illumos
[15:59:25] *** jellydonut <jellydonut!~quassel@s91904424.blix.com> has quit IRC (Quit: jellydonut)
[15:59:51] *** jellydonut <jellydonut!~quassel@s91904424.blix.com> has joined #illumos
[16:02:37] *** amrfrsh <amrfrsh!~Thunderbi@213.232.87.96> has quit IRC (Ping timeout: 246 seconds)
[16:04:40] *** amrfrsh <amrfrsh!~Thunderbi@213.232.87.96> has joined #illumos
[16:08:17] *** tsoome <tsoome!~tsoome@fc35-68b8-42e2-2989-2f80-4a40-07d0-2001.sta.estpak.ee> has quit IRC (Ping timeout: 260 seconds)
[16:08:59] *** tsoome <tsoome!~tsoome@fc35-68b8-42e2-2989-2f80-4a40-07d0-2001.sta.estpak.ee> has joined #illumos
[16:11:47] *** phyre <phyre!~phyre___@78.30.22.107> has quit IRC (Remote host closed the connection)
[16:20:55] <sjorge> Well... that was new!
[16:21:26] <sjorge> I rewrote the read/write function a bit, then added 2 more wrappers as both openbsd and freebsd use them and well it makes sense
[16:21:46] <sjorge> so I adapted the first wrapper to what we need to do... like 8 lines ~ 12 including the definition
[16:21:50] <sjorge> and it compiled... first try
[16:22:03] <sjorge> not sure if it works or not :p but it compiles
[16:22:17] <sjorge> So sorry if I someone used all luck for today in the world
[16:24:17] *** andyf <andyf!~andyf@omniosce.ee.ethz.ch> has quit IRC (Quit: nn)
[16:27:32] *** hadfl <hadfl!~hadfl@omniosce.ee.ethz.ch> has quit IRC (Ping timeout: 260 seconds)
[16:29:03] *** elenium <elenium!~elenium@46.180.115.3> has quit IRC (Ping timeout: 260 seconds)
[16:33:49] *** andyf <andyf!~andyf@omniosce.ee.ethz.ch> has joined #illumos
[16:34:18] *** hadfl <hadfl!~hadfl@omniosce.ee.ethz.ch> has joined #illumos
[16:36:30] *** Kurlon <Kurlon!~Kurlon@cpe-67-253-136-97.rochester.res.rr.com> has quit IRC (Remote host closed the connection)
[16:37:08] *** Kurlon <Kurlon!~Kurlon@167-48.gwi.net> has joined #illumos
[16:41:42] <LeftWing> Agnar: I seem to recall seeing some work in that direction somewhere
[16:42:50] <gitomat> [illumos-gate] 12815 ucode/update.intel needs pkgfmt fix -- Dan McDonald <danmcd at joyent dot com>
[16:42:51] <gitomat> [illumos-gate] 12817 ucode/update.intel should be able to use git repos -- Dan McDonald <danmcd at joyent dot com>
[16:42:52] <gitomat> [illumos-gate] 12390 AES: aes_mech_info_tab error -- Nick Ewins <nick at krypton dot computer>
[16:43:09] <LeftWing> Agnar: https://www.illumos.org/issues/12606
[16:44:46] <rmustacc> danmcd: Um, mind backing out the ucode changes until they've been RTI'd?
[16:44:57] <danmcd> Oh fuck.
[16:45:20] <rmustacc> It's fine. Just git revert the two + squash it as one backout.
[16:46:50] <Agnar> LeftWing: ah, thanks!
[16:47:53] <gitomat> [illumos-gate] Revert "12817 ucode/update.intel should be able to use git repos" -- Dan McDonald <danmcd at joyent dot com>
[16:49:10] <Agnar> LeftWing: amazing, thanks for the pointer.
[17:02:28] <danmcd> There's WIP on NFSv4.[12]. pNFS (which is almost it's own thing) isn't part of that WIP.
[17:03:16] <danmcd> I've not done any work on it, but I did create an `nfsv41-2` branch in illumos-joyent so the people who ARE doing the work can get SmartOS testing.
[17:04:04] <danmcd> Ahh, vgusev is here. He might be able to speak more to it.
[17:05:12] <LeftWing> I am somewhat sceptical of pNFS but it'd be good to get the other new 4.1/4.2 features
[17:10:37] <LeftWing> danmcd: It'd be good to lay the backout commit message like we generally do next time so that the automation catches it properly
[17:12:33] <danmcd> Was that due to the squashing of two? Or because it wasn't "backout: XXXXX <comment>" ?
[17:12:40] <danmcd> (I'm guessing the latter.)
[17:12:49] <LeftWing> the "backout:" one yea
[17:13:04] <LeftWing> Squashing is OK
[17:16:51] <vgusev> danmcd: Are you asking about nfsv4.[12]?
[17:17:09] <danmcd> Not me... @Agnar.
[17:18:26] <LeftWing> I'm curious too though, vgusev -- how's the NFSv4.X stuff coming along?
[17:19:34] <vgusev> This week probably I will update repo https://github.com/racktopsystems/illumos-gate/tree/nfs4xserver and then will prepare review. Are you interested in that ?
[17:20:13] <dodobrain> how come the openindiana.org site is showing lots of old info?
[17:20:58] <dodobrain> the gallery page shows stuff from 2016. the Roadmap page shows the latest "on the menu for hipster 2017.10" details
[17:21:22] <dodobrain> am i whining in the wrong channel or is this tolerated here? :)
[17:21:56] <LeftWing> dodobrain: I suspect it'd be better to ask in #openindiana -- but I also suspect the answer is just that they have limited resources and have preferred to focus on the software rather than the site.
[17:22:00] <LeftWing> So, be gentle :)
[17:23:24] <dodobrain> sure, thats the point. i'm not going to be "heavy" in terms of providing contributions to important things in the os/distro
[17:23:30] <dodobrain> so i'll focus on the easy bits for now :D
[17:25:46] <danmcd> @vgusev --> Please ping me when you've updated that branch.
[17:26:02] <vgusev> danmcd: yes, sure
[17:26:09] *** Kurlon_ <Kurlon_!~Kurlon@cpe-67-253-136-97.rochester.res.rr.com> has joined #illumos
[17:26:16] <danmcd> (mail is best)
[17:26:35] <vgusev> LeftWing: are you interested in nfsv4.[12] ?
[17:27:10] <LeftWing> I am interested in the status! I don't think I am going to have a lot of cycles to review a large change in an unfamiliar area just at the moment
[17:27:14] *** Kurlon_ <Kurlon_!~Kurlon@cpe-67-253-136-97.rochester.res.rr.com> has quit IRC (Remote host closed the connection)
[17:27:51] *** Kurlon_ <Kurlon_!~Kurlon@167-48.gwi.net> has joined #illumos
[17:28:19] <dodobrain> anyway, its getting pretty late (0127 now) to be playing with what once was osol. i'll be back later tomorrow
[17:28:24] *** Kurlon <Kurlon!~Kurlon@167-48.gwi.net> has quit IRC (Ping timeout: 258 seconds)
[17:29:34] *** Kurlon <Kurlon!~Kurlon@cpe-67-253-136-97.rochester.res.rr.com> has joined #illumos
[17:30:42] *** clapont <clapont!~clapont@unaffiliated/clapont> has quit IRC (Ping timeout: 258 seconds)
[17:30:42] *** Kurlon <Kurlon!~Kurlon@cpe-67-253-136-97.rochester.res.rr.com> has quit IRC (Remote host closed the connection)
[17:32:32] *** clapont <clapont!~clapont@unaffiliated/clapont> has joined #illumos
[17:33:18] *** Kurlon_ <Kurlon_!~Kurlon@167-48.gwi.net> has quit IRC (Ping timeout: 256 seconds)
[17:33:31] <vgusev> LeftWing: Maybe are you going to use that ?
[17:34:42] <LeftWing> vgusev: I use NFS a lot myself, but I'd need to see a summary of the new features to know how much of an impact it'll have. I'm mostly just keen to encourage folks to upstream work like this; e.g., like the SMB updates from Nexenta, etc.
[17:36:04] <rmustacc> w/in 6
[17:36:06] <rmustacc> Oops.
[17:36:08] <LeftWing> :D
[17:37:12] <bahamat> LeftWing: I set up a TimeMachine back up using the new SMB stuff. I've been able to do a full back up and restore of individual files, but I haven't tested a full restore yet.
[17:37:20] <LeftWing> Neat!
[17:37:39] <LeftWing> Did you have to do anything unusual to get it to work?
[17:37:50] <LeftWing> I know people have experienced some challenges with Time Machine in the past haha
[17:38:50] <bahamat> You need some dns-sd services registered
[17:39:43] <bahamat> I added it to the end of this: https://docs.smartos.org/configuring-smb-in-smartos/
[17:40:25] <bahamat> So if you've got a new enough PI, standard SMB set up plus dns-sd will make time machine work (at least to some degree).
[17:41:18] <bahamat> I don't have a system I can blow away and test the restore on.
[17:41:34] <LeftWing> Would that they had an enlightened policy on VMs
[17:41:43] <bahamat> Even with a second back up on a real TM disk, I'd be out of a machine for a day to try it.
[17:41:49] <danmcd> I thought the plan was to upstream that work. It's why I branched for it in illumos-joyent.
[17:42:54] <LeftWing> danmcd: vgusev said above he was going to prepare a review, so I believe that's the case
[17:43:53] <danmcd> Yep. I've been able to use the nfs-zone test rig I had to help him.
[17:44:07] <bahamat> I actually got a VM to work in Parallels, but it's been quite some time since I've used it. And this laptop with only 16G of RAM strains to just run as it is. Adding an 8G or 12G VM on top of that is so difficult.
[17:44:28] <bahamat> Even with CoaL, I have to kill a bunch of things to use it, and I can't use it for very long.
[17:45:13] <bahamat> I'd probably still be unable to use my laptop for regular work even with doing a full TM restore in a VM.
[17:49:10] <bahamat> When I get a new laptop (whenever that might be...), I'll try to restore it from the TM first, before doing a migration from my current laptop. If that works, I think we're golden. At least until Apple changes the protocol.
[17:54:16] *** dodobrain <dodobrain!~dodobrain@unaffiliated/freakabcd> has quit IRC (Ping timeout: 256 seconds)
[17:56:37] *** yuripv <yuripv!~yuripv@91.240.124.138> has quit IRC (Read error: Connection reset by peer)
[17:57:17] *** yuripv <yuripv!~yuripv@91.240.124.138> has joined #illumos
[18:17:31] <sjorge> danmcd sounds like you needed some extra coffee :p
[18:17:41] <sjorge> And interesting about the newer NFS work!
[18:19:22] <sjorge> LeftWing rmustacc : for uchcom I just going to adapt some of the 'vendor' functions form the freebsd driver and use those in our params function
[18:19:29] <sjorge> Seems cleaner and easier to understand haha
[18:19:34] <sjorge> which... I do not fully yet
[18:20:09] <sjorge> But i got version reading from teh chip working and I can erm write to the ctrl registers and fuck them up so the stick is dead for a few minutes after unplug
[18:20:27] <sjorge> So it's doing something, how common is it to mess up hardware while writing a driver?
[18:20:40] <sjorge> (assume that is why a lot of dev hardware has seperate debug ports)
[18:20:43] <sjorge> *assumes
[18:28:31] *** neuroserve <neuroserve!~toens@ip-84-118-130-44.unity-media.net> has quit IRC (Ping timeout: 260 seconds)
[18:34:30] <am11> jlevon: it seems like the uprlim fix proposed in #4963 does not elide the EPERM when process is running with basic privileges. it still requires proc_prioup to raise thread priority.
[18:34:43] <am11> my attempt was https://github.com/am11/runtime/commit/c4cffb92be17228a5c84872d2bc83015ab5b76e5#diff-d1222fd142b2a107514deade7201f2ecR1257.
[18:36:18] <jlevon> pcparms_t pcparms;
[18:36:34] <jlevon> tsparms_t *tsp = (tsparms_t *)pcparms.pc_clparms;
[18:36:34] <jlevon> if (tsp->ts_uprilim < schedParam.sched_priority)
[18:36:39] <jlevon> isn't that stack junk?
[18:38:30] <am11> it's cpp so RAII is in effect?
[18:40:22] <jlevon> I'm not following how RAII plays in here? pcparms_t is an uninited POD struct isn't it?
[18:40:39] <jlevon> also if it is zeroed for you, why even bother checking?
[18:42:16] <jlevon> i.e. don't you want a PC_GETPARMS first
[18:54:41] <danmcd> @sjorge yes, was woken up earlier than I'd planned.
[19:05:01] *** clapont <clapont!~clapont@unaffiliated/clapont> has quit IRC (Ping timeout: 264 seconds)
[19:10:46] *** clapont <clapont!~clapont@unaffiliated/clapont> has joined #illumos
[19:11:58] *** Kurlon <Kurlon!~Kurlon@cpe-67-253-136-97.rochester.res.rr.com> has joined #illumos
[19:15:31] *** jollyd <jollyd!~alarcher@2a01:cb1d:611:bb00:8820:7cbd:8965:af56> has joined #illumos
[19:15:36] *** jollyd <jollyd!~alarcher@2a01:cb1d:611:bb00:8820:7cbd:8965:af56> has quit IRC (Client Quit)
[19:16:11] *** jollyd <jollyd!~alarcher@2a01:cb1d:611:bb00:8820:7cbd:8965:af56> has joined #illumos
[19:27:13] <nahamu> bahamat: does dns_sd need to be rerun every time the SMB server zone reboots, or does that command set a property in an existing service??
[19:27:34] <nahamu> (when making a Time Machine target via SMB)
[19:28:12] <nahamu> err... dns-sd
[19:29:02] * nahamu goes looking at the man pages.
[19:30:51] <nahamu> I guess it interacts with mdnsd.
[19:36:16] <am11> jlevon: i followed the path of pthread_setschedparam down to priocntl in illumos-gate; we do not call PC_GETPARMS there (with policy SCHED_OTHER).
[19:40:47] <bahamat> nahamu: It needs to be running/backgrounded.
[19:41:07] <bahamat> I wrote a simple smf service for it that basically does `dns-sd ... &`
[19:41:20] <nahamu> got it. makes sense. Thanks!
[19:41:57] <bahamat> I can show you what I have if you want to see it, but it's not really worth putting in the wiki.
[19:43:28] *** liv3010m <liv3010m!~liv3010m@77-72-245-190.fibertel.com.ar> has quit IRC (Ping timeout: 246 seconds)
[19:44:47] *** liv3010m <liv3010m!~liv3010m@77-72-245-190.fibertel.com.ar> has joined #illumos
[19:53:08] *** psarria__ <psarria__!~psarria@74.red-79-146-99.dynamicip.rima-tde.net> has quit IRC (Quit: Leaving)
[19:53:22] *** psarria__ <psarria__!~psarria@74.red-79-146-99.dynamicip.rima-tde.net> has joined #illumos
[19:53:33] *** psarria__ <psarria__!~psarria@74.red-79-146-99.dynamicip.rima-tde.net> has quit IRC (Client Quit)
[19:53:52] *** psarria <psarria!~psarria@74.red-79-146-99.dynamicip.rima-tde.net> has joined #illumos
[20:00:23] <wiedi> hmm newer mdnsresponder supports reading service definitions from a file with -f but I don't think we have that yet
[20:01:26] <bahamat> wiedi: That would probably be handy. To do time machine you need at least two dns-sd processes running.
[20:02:17] <jlevon> am11: where do you think pcparms is getting set in your code?
[20:02:30] <bahamat> It'd also be nice to have a "respond to, but don't advertise" flag to mimic the behavior of macos for _device-info
[20:06:35] *** ryao <ryao!~ryao@gentoo/developer/ryao> has joined #illumos
[20:06:46] <ryao> rmustacc: Just in case you guys have not seen it yet: https://www.vusec.net/projects/crosstalk/
[20:08:00] <rmustacc> I only learned of it today, but I presume some others were roped in.
[20:15:22] *** jimklimov <jimklimov!~jimklimov@78.80.224.132> has quit IRC (Ping timeout: 256 seconds)
[20:20:37] *** BOKALDO_ <BOKALDO_!~BOKALDO@91.105.127.77> has quit IRC (Quit: Leaving)
[20:30:57] *** wacki <wacki!~wacki@i577B890F.versanet.de> has joined #illumos
[20:32:12] *** neuroserve <neuroserve!~toens@ip-178-202-213-124.hsi09.unitymediagroup.de> has joined #illumos
[20:34:14] *** jimklimov <jimklimov!~jimklimov@ip-86-49-254-26.net.upcbroadband.cz> has joined #illumos
[21:00:49] <sjorge> wiedi last time I needed that i... think I just installed avahi from pkgsrc but it's fugly
[21:00:51] <am11> jlevon: i understood what you mentioned above, then followed how it is done in libc: https://github.com/illumos/illumos-gate/blob/4e0c5eff9af325c80994e9527b7cb8b3a1ffd1d4/usr/src/lib/libc/port/rt/sched.c#L418. this function is called in path followed by pthread_setschedparam (for which we are making this workaround). as you can see my approach is aligned with that.
[21:02:04] <am11> in set_priority, we never call PC_GETPARMS for SCHED_OTHER policy.
[21:02:48] <jlevon> am11: but you read the lim before it's ever set!
[21:03:50] <am11> in libc code? where exactly/
[21:04:01] <jlevon> in your code you posted
[21:04:05] <jlevon> sorry, lost the link
[21:04:31] <am11> that is aligned with what libc is doing, no?
[21:04:41] <jlevon> no
[21:04:44] <jlevon> since libc never reads it
[21:04:46] <jlevon> just sets it
[21:05:03] <jlevon> as I pointed out in the bug, that's probably why it's doing what it's doing
[21:08:58] <am11> jlevon: if i try to get parms like this https://github.com/am11/runtime/commit/eaec4b2d46e292382669890ac1766c004eae9e37#diff-d1222fd142b2a107514deade7201f2ecR1265, it gives me ESRCH.
[21:13:04] *** freakazoid0223 <freakazoid0223!~matt@pool-96-227-117-103.phlapa.fios.verizon.net> has joined #illumos
[21:15:50] <jlevon> am11: hmm that's annoying. indicates the thread isn't in the class which doesn't make much sense. not sure offhand.
[21:21:41] <jlevon> am11: you might be able to pass in -1 as the cid for the getparms
[22:12:22] *** clapont <clapont!~clapont@unaffiliated/clapont> has quit IRC (Ping timeout: 260 seconds)
[22:22:54] <am11> jlevon: tried with cid -1; GETPARMS goes through, but SETPARMS still errors out with EPERM. one thing i noticed while reading libc code is that PC_GETPARMS with P_LWPID expects lpwid, but PC_SETPARMS expects pthread_t with P_LWPID. nonetheless, flipping those has no effect on the fact that priocntl(type, id, PC_SETPARMS, pcparms) expects priv_proc_prioup.
[22:25:05] <jlevon> am11: you're confident this time your parms values are correct
[22:25:13] <jlevon> ?
[22:28:57] <am11> jlevon: they are the same except for cid -1, which you mentioned.
[22:31:24] <am11> jlevon: can PC_SETPARMS ever succeed without priv_proc_prioup?
[22:32:07] <am11> that permission checking flow is in libc's prriocntl_common(), i don't see how the propsed change will elide the privilge requirement.
[22:32:34] <jlevon> ??
[22:32:42] <jlevon> obviously libc can't be checking perms for anything
[22:32:58] <jlevon> secpolicy_raisepriority() is what requires the UP variant
[22:33:26] *** wacki <wacki!~wacki@i577B890F.versanet.de> has left #illumos
[22:33:27] <jlevon> oh you mean the kernel's
[22:33:40] <am11> yup, sorry. usr/src/uts/common/disp/priocntl.c to be exact.
[22:33:42] <jlevon> can you be specific about which lines of code you mean
[22:34:42] <am11> jlevon: the case which starts here: https://github.com/illumos/illumos-gate/blob/4e0c5eff9af325c80994e9527b7cb8b3a1ffd1d4/usr/src/uts/common/disp/priocntl.c#L270
[22:34:49] <jlevon> presuming you're in the TS class, ts_parmsset() is where the up priv is checked
[22:35:17] <jlevon> if you're failing that check, something is wrong in what you're passing in
[22:35:24] <am11> i followed various paths, it looks like the EPERM i am seeing (which goes away when priv_proc_prioup is set), is comming from there.
[22:35:26] <jlevon> that's just a basic "am I operating on myself" test
[22:35:44] <jlevon> 619 prochasprocperm(proc_t *tp, proc_t *sp, const cred_t *scrp)
[22:35:51] <jlevon> 626 if (tp == sp)
[22:35:51] <jlevon> 627 return (1);
[22:36:22] <jlevon> am11: I don't know if you've used dtrace much to confirm?
[22:37:41] *** freakazoid0223 <freakazoid0223!~matt@pool-96-227-117-103.phlapa.fios.verizon.net> has left #illumos ("Leaving")
[22:37:56] <am11> i can check with dtrace. but do you expect priocntl(..PC_SETPARMS...) to succeed without priv_proc_prioup?
[22:38:02] <am11> my main prusit is that.
[22:38:09] <jlevon> yes if otherwise allowed.
[22:39:21] <jlevon> have to run now.
[22:47:19] *** jimklimov <jimklimov!~jimklimov@ip-86-49-254-26.net.upcbroadband.cz> has quit IRC (Quit: Leaving.)
[22:54:54] *** ptribble <ptribble!~ptribble@cpc92716-cmbg20-2-0-cust138.5-4.cable.virginm.net> has quit IRC (Quit: Leaving)
[23:03:49] *** liv3010m <liv3010m!~liv3010m@77-72-245-190.fibertel.com.ar> has quit IRC (Ping timeout: 264 seconds)
[23:04:34] *** liv3010m <liv3010m!~liv3010m@77-72-245-190.fibertel.com.ar> has joined #illumos
top

   June 9, 2020  
< | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 | 12 | 13 | 14 | 15 | 16 | 17 | 18 | 19 | 20 | 21 | 22 | 23 | 24 | 25 | 26 | 27 | 28 | 29 | 30 | >