IRC logs for #cip for Thursday, 2018-08-30

*** iwamatsu has joined #cip04:09
*** iwamatsu has quit IRC04:10
*** iwamatsu has joined #cip04:12
*** iwamatsu has quit IRC04:13
*** iwamatsu has joined #cip04:16
*** iwamatsu has joined #cip04:16
*** iwamatsu has quit IRC04:17
*** iwamatsu has joined #cip04:18
*** iwamatsu has quit IRC04:29
*** iwamatsu has joined #cip04:29
*** iwamatsu has quit IRC04:53
*** iwamatsu has joined #cip04:53
*** iwamatsu has quit IRC04:55
*** iwamatsu has joined #cip04:57
*** iwamatsu_ has joined #cip04:59
*** robertmarshall has joined #cip07:09
*** patersonc has joined #cip07:16
*** robertmarshall has quit IRC07:18
*** robertmarshall has joined #cip07:18
*** robertmarshall is now known as rajm07:20
*** iwamatsu__ has joined #cip07:36
*** iwamatsu__ has quit IRC07:37
*** iwamatsu__ has joined #cip07:39
*** iwamatsu__ has quit IRC07:40
* rajm returns and has run a successful healthcheck on BBB with an email to the list07:51
*** iwamatsu__ has joined #cip07:52
*** hungtran has joined #cip07:57
*** HoloIRCUser has joined #cip07:57
patersoncWelcome back rajm07:58
rajmthanks patersonc!07:58
*** HoloIRCUser is now known as iwamatsu____07:58
*** Tzongyen_Lin has joined #cip07:59
szlin#roll call08:00
szlinplease say hi if you're here.08:00
patersoncNot many here today then?08:05
*** mungaip has joined #cip08:06
patersoncHi szlin, is the meeting going ahead?08:09
*** iwamatsu__ has joined #cip08:10
*** iwamatsu__ has joined #cip08:10
szlin#topic action item review08:11
szlinAbout having a git group account in - still waiting for response from Ben08:11
szlin#topic kernel maintenance updates08:11
szlinI've sent some overlayfs backport patches to stable-4.4.y08:11
szlinThese patches fix read-only issue when non-empty workdir occurred08:11
szlinin overlayfs, the non-empty workdir could be easily reproduced in08:11
szlinpower-failure test during write operations.08:11
szlinI will continue to send bugfix patches in overlayfs in these day.08:11
szlinpatersonc: yes08:11
szlinpatersonc: the network status seems weird.08:11
patersoncszlin: Yes, I just got 10 lines at once!08:11
szlinpatersonc: can you see my updates?08:11
*** toscalix has joined #cip08:11
szlinAccording to, the next stable kernel might be 4.1908:12
patersoncHas anyone seen anything concrete about LTS v4.19?08:12
*** toscalix has quit IRC08:13
patersoncszlin: Do we know when Debian will make a decision on their next Kernel?08:14
szlinNot from me08:14
szlinpatersonc: we need to ask Ben08:14
*** toscalix has joined #cip08:15
szlinBut the date of 4.19 might be happened in Oct or Nov.08:15
patersoncI would guess v4.19 will be out mid Oct08:16
szlinpatersonc: about the kernel of buster, you may refer to here08:17
*** moxavictor has joined #cip08:17
*** sangorrin has joined #cip08:17
szlinthat's why I mentioned that we need to ask Ben08:18
szlinany other point?08:18
sangorrinthe bug that jan mentioned the other day08:19
sangorrinL1FT or something08:19
patersoncThat's the new Intel bug08:19
sangorrinit was fixed in 4.4 from what i see08:19
sangorrinin the patch queue08:19
toscalixsangorrin: bwh is working on the new CIP 4.4 kernel and includes that patch08:20
sangorrinok cool08:20
toscalixI believe I mentioned a few days ago somewhere (I cannot remember when or where, but CIP channel for sure)08:21
szlintoscalix: did you get the feedback from Ben? About having a git group account in kernel.org08:21
sangorrinit will be good to run LTP cve tests often08:21
patersoncAre there any specific test cases for it?08:21
sangorrini think the spectre bugs appear in /proc somewhere08:21
sangorrinso we could add a test checking /proc/cpuinfo08:22
toscalixszlin: no about that point no08:22
sangorrinalso in /sys/devices/system/cpu/vulnerabilities/l1tf08:22
toscalixszlin: I thought it was solved. I mean, we were going to ask08:23
sangorrintoscalix: wagner mentioned an email address08:23
* toscalix have a release next week. I have been disconnected the last few days08:24
sangorrinthe problem is that we need someone with reputation to back us, so Ben could help us with that08:24
szlintoscalix: sorry, I don't get it. who will ask group account in for CIP08:24
toscalixszlin: should it be Ben?08:26
toscalixor somebody with certain level of representation will?08:26
toscalixwho will manage it?08:26
szlinok, I suggest to discuss this topic in TSC meeting next Monday08:27
sangorrintoscalix: good point, maybe we can wait until the next kernel maintainer is completely decided08:27
patersoncTSC should probably set it up, but having Ben's/Greg's ack would help08:28
toscalixpatersonc: agree. I believe there is a mail I sent last week about it and Ben haven't answered it. I will ping him about it08:29
szlin#action Agustin will ping Ben for the feedback of git group account in kernel.org08:30
szlinpatersonc: please08:30
szlinany other point?08:31
patersoncIf CIP chooses v4.19 for the next Kernel, when do you think we'd aim to make the first v4.19.y-cip release?08:31
sangorrinit would be nice to have support for RX-G1M as soon as it is released08:32
sangorrinRZ i mean08:32
patersoncRZ/G1M should be supported by default as it is upstreamed08:33
sangorrinthat's great08:33
szlinI think it depends on the maintainer of 4.19 in CIP. AFAIK, we don't have release policy about that.08:33
patersoncRZ/G2M will not be, as we're still upstreaming it. First patches will hit v4.2008:33
szlin(please correct me if I'm wrong08:33
patersoncszlin: I don't think there is a policy, no.08:34
szlinpatersonc: so you may contact maintainer of 4.19.y in CIP for that08:35
szlinany other point?08:36
szlin#topic kernel testing08:36
szlinAny update?08:37
toscalixrajm ?08:37
* toscalix pnging rajm08:37
szlinthanks. let's move forward08:38
szlin#topic AOB08:38
toscalixmeanwhile, I have a question08:38
toscalixabout kernel testing08:38
toscalixis there a prospect date to have the infra in place for testing it?08:38
rajmsorry - missed the ping08:41
toscalixa line or two of your recent progress?08:41
rajmto give a late update - moving forward with containering the kernelci and LAVA - some issues with networking and inter docker package communication08:42
rajmnow that I'm back off a week away I'll look to addressing there - thanks08:43
toscalixwhere can we follow progress?08:43
toscalixtickets to follow?08:44
rajmhopefully here, and in the ticket / journal08:44
rajmjust a mo08:44
sangorrinrajm: are you using compose? one reason it may not communicate is if you don't put the compose version08:44
rajmI am using compose08:44
*** moxavictor has quit IRC08:45
rajmjournal at
rajmI was using lucj's version which uses compose in the script but I'm still finding issues08:45
szlintoscalix: we want to build the infra. at our own site in Q1/ Q2 next year, and we're working on it.08:46
toscalixand check related issues08:46
szlinAlign with CIP testing is our next step.08:46
rajmthanks toscalix (and #197)08:46
szlinpatersonc: sangorrin do you have any prospect date to have the infra in place for testing it?08:47
toscalixin order for each one of us to align, we need to know details about the CIP infra08:47
sangorrinrajm: I added stuff for proxy's in lucj's version. I hope i didnt break that :)08:47
rajmsangorrin: I'll check the commits - the kernelci looks fine it's 'just' an inability to download any builds08:48
sangorrinszlin: i don't have time to maintain a lab, I use fuego for testing CIP08:48
patersoncszlin: We're currently trying to get B@D up and running for iwg20m behind our firewalls etc. Next step is to get a lava-slave up and running for CIP.08:48
patersonc(thanks to rajm for his help)08:48
rajmta patersonc!08:49
toscalixsangorrin: we do not either. We hope to use B@D and send results to the central infra08:49
sangorrinrajm: yes i have seen that problem08:49
sangorrini think you still can access the builds if you input the url by hand08:49
patersoncThere is already a central kernelci server, so I guess B@D could be configured to report there?08:49
sangorrinbut the links are wrong08:49
toscalixpatersonc: yes08:49
rajmah - they're trying to pull off the local docker and they need to use one of the other instances08:50
sangorrintoscalix: yeah, the central infra is what i am waiting for as well08:51
patersoncWe need to assign a KernelCI maintainer/admin I guess.08:51
sangorrinpatersonc: yes, someone that assigns the tokens etc08:52
toscalixpatersonc: yes08:52
patersoncDoes anyone have and experience with KernelCI?08:52
toscalixpatersonc: since this person needs to be assigned internally, because it will require some work, this is the perfect topic to raise at TSC08:53
sangorrinpatersonc: I have a little of experience, but not as much as rajm or robert08:53
patersonctoscalix: I was about to say that :)08:53
toscalixit cannot be on best effort basis. It requires internal support from the company to do this08:54
szlin#action discuss the KernelCI maintainer/admin at TSC meeting08:54
rajmNeeds to do some looking on those issues - I was thinking it might need knowlege of the guts of the system but maybe not08:54
sangorrinin agl the maintenance is done by kernelci creators, isn't it?08:55
toscalixthere is a contract in place and jan-simon also helps08:55
toscalixbut the contract is not just for this, but for other tasks also08:55
toscalixI would assume a CIP Member would contribute with a person for this08:56
sangorrindo you mean at full time?08:57
patersoncszlin: The TSC topic should cover all testing infrastructure (e.g. lava master), not just KernelCI08:57
toscalixsince every company tests kernels, so this can be a good place to learn or an easy infra to maintain if you are already maintaining one08:57
patersoncsangorrin: I wouldn't have thought so08:57
toscalixsangorrin: no full time no, but more time at the beginning, less overtime08:57
szlin #action discuss the testing infra. including maintainer/admin at TSC meeting08:58
* toscalix need to go08:58
szlinany other point?08:58
szlinthanks all08:58
sangorrinoh too late, but i was going to ask if we are going to review the lts paches at some point08:59
rajmta szlin!09:00
szlinsangorrin: we can discuss it all the time, not just in IRC weekly meeting :D09:01
*** Tzongyen_Lin has quit IRC09:01
iwamatsu____thanks all09:01
sangorrini was reviewing the changes to modify_ldt, but i have no idea how to test that...09:01
sangorrini guess i will have to enable kasan and run the test in the commit message09:02
patersonciwanatsu___: Are you officially the CIP maintainer now?09:03
sangorrinpatersonc: not officially09:06
sangorrinbtw bootlin rejected our offer09:06
patersoncOh really?09:06
patersoncDo you know why?09:06
sangorrinno, I guess kobayashi-san will mention it in the next meetign09:07
patersonciwanatsu___: Do you know anything more about the potential LTS v4.19?09:08
patersonciwamatsu____ ^^09:09
*** iwamatsu____ has quit IRC09:11
szlinsangorrin: do you have the commit id?09:15
sangorrinccd5b32351 upstream09:16
sangorrini found a cve test in ltp that addresses that patch!09:17
szlinsangorrin: I found that patch is marked as Cc: <> [v4.6+]09:23
szlinsangorrin: This patch depends on 39a0526fb3f7d93433d146304278477eb463f8af, which is not in 4.410:07
sangorrinI can't check now, but i think it is in 4.4 because the patch that was merged assumes that init_new_context_ldt exists10:38
sangorrinI will check tomorrow10:38
*** sangorrin has quit IRC10:39
*** tpollard has quit IRC11:02
* rajm sees that the lucj/kernelci-proxy:latest docker contains today's build but kernelci web i/f is not showing them (also that the docker still contains the builds from July - rather contradicting what I said last week!11:06
*** hungtran has quit IRC11:17
rajmand now the web i/f does show today's build?!12:03
rajmsangorrin: I replaced a approp line in flask_settings with FILE_SERVER_URL = "" and kernelci now retrieves the builds - doesn't look like a change from you :)12:39
*** rajm has quit IRC14:16
*** rajm has joined #cip14:16
*** patersonc has quit IRC14:28
*** rajm has quit IRC14:30
*** rajm has joined #cip14:43
*** rajm has quit IRC15:45
*** toscalix has quit IRC18:01

Generated by 2.15.3 by Marius Gedminas - find it at!