**** BEGIN LOGGING AT Tue Jun 02 02:59:57 2020 Jun 02 05:03:00 I'm using mint 19.3, upgraded to kernel 5.0, my pc is crashing a lot. Is 5.0 bleeding edge? Jun 02 05:03:00 Earlier i had 4.15 Jun 02 05:25:27 vedant16[m]: Which PC do you have? Is it relatively old or new? Jun 02 05:26:03 The latest kernel as of today is 5.7, so 5.0 has been around for a while Jun 02 05:41:43 New Jun 02 05:41:48 1yr old Jun 02 05:42:16 Noveau driver causes problems Jun 02 05:48:31 vedant16[m]: Use nvidia proprietary driver, you can install from within the linux mint itself Jun 02 06:12:54 > vedant16: Use nvidia proprietary driver, you can install from within the linux mint itself Jun 02 06:12:54 Pc doesn't boot if use this with 5.0 kernel. Jun 02 06:13:11 I downgraded and it now works okay. Jun 02 06:13:15 Okay Jun 02 06:13:25 Does it crash on boot? Jun 02 06:13:56 It doesn't boot after splash screen Jun 02 06:14:06 Screen turns off Jun 02 06:14:10 Okay Jun 02 06:30:21 Fedora is a better option than Debian based distros if you want up to date kernel and other software packages all the time. i am using 5.6.13 with my fedora kde no problem at all. Jun 02 07:57:01 Abhishek_ jkridner ds2 pdp7: meeting in 4 minutes Jun 02 07:58:49 *yawn* Jun 02 07:59:54 Hi Jun 02 08:00:11 Hello Jun 02 08:00:23 let see participation is better then last week :D Jun 02 08:00:33 s/see/see if/ Jun 02 08:02:28 is video all done and ready to go? :D Jun 02 08:03:02 I was working on LCD4-01 cape overlay but i don't know why i2c1 is the target for touch controller can anybody help, here is the code written by Robert https://github.com/beagleboard/bb.org-overlays/blob/master/src/arm/BBAI_BB-BONE-LCD7-01-00A2.dts Jun 02 08:03:33 what do you mean by you don't know? Jun 02 08:03:47 selection of I2C buses are tied in with the way it is wired Jun 02 08:04:28 I am not able to find any connection between the touch controller and the i2c1 bus Jun 02 08:05:13 I am working on the video not done yet. Jun 02 08:05:51 looking up the lcd4... Jun 02 08:07:16 lorforlinux[m]: have you see this: https://www.openhacks.com/uploadsproductos/beaglebone-lcd4-reva1-schematic.pdf Jun 02 08:07:31 Search for R111.. that is by default connected to I2C1 Jun 02 08:09:44 does `T Point A` means `Touch point A` Jun 02 08:09:53 where does it say that? Jun 02 08:10:43 I see it... not sure...that's just notes Jun 02 08:10:57 the wiring is pretty clear - TS_I2C is the I2C bus on the TSC (Controller) Jun 02 08:11:10 it is connected to I2C1 via the 0's Jun 02 08:12:53 Ohh ohkay, got it now. Jun 02 08:15:25 any more things to sort out? Jun 02 08:15:40 If i write an overlay for servo cape, The input which are used for remote control should have pinmode set to GPIO right? Jun 02 08:17:23 got a link for the schematic? I donno that cape off hand and servos can mean a few things Jun 02 08:17:27 (to me that is) Jun 02 08:18:29 https://github.com/beagleboard/capes/blob/master/beaglebone/Servo/Servo_Cape_sch.pdf Jun 02 08:20:45 looks like GPIO is right Jun 02 08:21:38 Okay GPIO it is. Jun 02 08:24:01 Also, I wanted to know I am working on bb.org-overlay repo and BeagleBoard-DeviceTree repo, what is the process of syncing the DTs on the linux repo for builing dtbs? Jun 02 08:24:51 Someone else should answer that... don't think my opinions is appropriate in this context Jun 02 08:25:36 Robert has some script in `/opt/scrips` i think for testing the up to date DTs on your board directly from BeagleBoard-DeviceTree repo. Jun 02 08:26:31 I normally work in the kernel so synchronization doesn't make sense for me hence the deference Jun 02 08:27:16 Okay, For BBAI the best way to use a cape is to use UBoot overlays right? due to dynamic pinmuxing issue? Jun 02 08:27:48 unfortunately, that is the case right now Jun 02 08:28:16 remuxing it at a later time w/o understanding the underlying processor errata could cause a hang :( Jun 02 08:28:58 For creating cape compatibility layer i have to work on overlays and beagle-tester, what else we can do to have a seemless experience switching boards while using some cape? Jun 02 08:30:26 a graceful way of flagging capes that can't work Jun 02 08:30:57 since a lot of users don't have (or don't care to) use the proper serial console, there needs to be a different UI...prehaps LED patterns when that is detected Jun 02 08:33:38 okay, when we plug a cape in BBB it detects it through the information provided by EEPROM and loads the correct overlay for it right? I have not tested any cape yet :( Jun 02 08:35:02 that is the way it is suppose to work Jun 02 08:36:41 The best thing for creating a cape compatibility layer would be to detect the cape on BBAI, edit the uEnv.txt and ask the user to reboot, what are your thoughts on this? Jun 02 08:37:23 editing uEnv.txt to set UBoot overlays for that cape. Jun 02 08:39:12 Is it possible to detect a cape during boot time and load the appropriate UBoot overlay without any glitches on BBAI? Jun 02 08:39:35 I don't think that Jun 02 08:39:59 it ideally should be the same process for the user as in BB and editing a config dynamically seems like a bad idea Jun 02 08:42:37 Okay, things are more clear to me now. Thanks ds2 Jun 02 08:43:05 glad to help Jun 02 08:44:32 pdp7 Abhishek_ jkridner please provide some insights into this :) Jun 02 08:55:25 if there is nothing else, I'll be sleeping (0155) Jun 02 09:15:53 lorforlinux[m]: contact rcn-ee for integration into kernel tree Jun 02 09:33:00 okay i will discuss this with Robert. Jun 02 09:35:12 please do, I don't have anythings else to ask for now. Jun 02 10:08:31 jkridner: The FexEx delivery guy is asking to share the Customs Payment proof on 711911@fedex.com, He told me to share NEFC details (I don't know what that is). The package will not be delivered until they receive the payment details. Jun 02 10:11:25 I think It's National Electronic Funds Transfer (NEFT) not NEFC Jun 02 10:11:47 Probably yes Jun 02 10:12:08 The payment would be through a shipping account in this case Jun 02 10:12:18 No NEFT Jun 02 10:12:38 what should i do? Jun 02 10:13:44 delivery will be delayed for one more day If it's not resolved now :( Jun 02 10:19:53 jkridner: I also got a call again for customs payment of around 8200 INR i.e around 115 dollers Jun 02 10:31:18 jkridner: Just talked to FedEx they are saying that in the details it's mentioned that recipient will pay the Customs Fee :( Jun 02 10:31:43 Can you please look into the issue? Jun 02 11:19:42 so sorry. fail on real try #1 to be up at 4AM. Jun 02 11:20:04 lorforlinux: was ds2 still up? Jun 02 11:20:18 looks like it. lots of backlog. Jun 02 11:21:26 yes, ds2 cleared a lot of my doubts. I can now work things forward. Jun 02 11:21:55 cool. I didn't see an answer on the video. know what you are going to do for it? Jun 02 11:22:18 NEFC.... hmmm. Jun 02 11:22:39 * NEFT he meant Jun 02 11:24:24 does it work if you send me invoices and I send the money by paypal? Jun 02 11:24:53 I will share the presentation with you once it's complete, is that okay? It would be nice if you could just approve the content. Jun 02 11:25:44 I am not sure, are you able to contact FedEx? I can try contacting them and ask if that's possible to do. Jun 02 11:25:45 sure. Jun 02 11:27:18 it would mean me paying you and you paying them. DHL told me the money has to come from the recipient, but then told me that someone would call me to pay by account, since I had an account, and they never called me. FedEx told me they would reverse the charges to my account and then, well, here we are. :-( Jun 02 11:27:24 I'll call FedEx again. Jun 02 11:27:42 they are super nice and don't get anything done. Jun 02 11:28:21 🤣 Jun 02 11:28:59 I'll look into the DHL issue. Jun 02 11:33:20 jkridner: Can you pay directly here -> https://del.dhl.com/IN/6yAlCZ6Bv6 Jun 02 11:33:51 This was shared by DHL in a message ^ Jun 02 11:34:13 I cannot. I've tried. Only takes India banks or credit carts. Jun 02 11:34:16 * I cannot. I've tried. Only takes India banks or credit cards. Jun 02 11:35:33 on phone with Fedex again now.... I'll call DHL next. Jun 02 11:36:44 okay if it doesn't work out, I will pay the DHL fee. Jun 02 11:40:36 deepankarmaithani: can you remind me the DHL airway bill Jun 02 11:42:08 jkridner: i dont have the DHL airway bill as the shipment has not been shiped yet. i only have Web Order Number : 19014051 Jun 02 11:42:20 deepankarmaithani: doh. ok. Jun 02 11:42:41 I'll call Mouser again Jun 02 11:44:14 jkridner: saketh saketh told me to send mouser a mail today and tell them the intended purpose of the product order . and i sent them a mail. I dont know why it was needed but i did as he said. :P Jun 02 11:49:01 for those with shipments stuck at Mouser, when I spoke to them last week, they said that they'd stopped processing shipments to India because the carrier wasn't accepting them. They said that the carrier started accepting them again and that they should go out quickly.... then they didn't, so I'll call them when they open today. Jun 02 11:52:08 jkridner: did you talk to FedEx? Jun 02 11:52:23 deepankarmaithan: that's what mouser folks asked me to do when I asked them Jun 02 11:53:32 okay. :) Jun 02 11:53:49 I thought my messages were going into thin air Jun 02 11:53:52 xD Jun 02 11:53:53 yeah all the necessery stuff is there Jun 02 11:55:16 > jkridner: I wasnt able to get 5.4-rt running on BBAI, as I was not able to fix Oops message from yesterday. rma31 and I agreed, that it would be best to switch to a target kernel which runs smoothly on BBAI as a starting point so I do not loose so much time right now. Which target RT kernel would you prefer? Jun 02 11:55:16 @jkridner:matrix.org : I think you overlooked this yesterday because the chat was quite busy so I ask again. Which kernel version would you prefer? Jun 02 12:15:41 If DHL issue doesn't resolve, this is my paypal https://www.paypal.me/deepaklorkhatri Jun 02 16:34:35 nwan[m]: whatever works for your development. chatting with rcn-ee soon on this. he said yesterday he'd just released an untested 5.4 yesterday. Jun 02 16:37:00 jkridner: did you talk to FedEx? Jun 02 16:37:36 yes. they said it should be cleared with Beagle paying for the duties. Jun 02 16:38:00 jkridner: I compiled and tried the latest 5.4 on my BeagleBone (AM335x), seems to be OK with respect to USB. Jun 02 16:38:25 Haven't put it on the AI yet. Jun 02 16:39:18 Okay that's nice, what about DHL? should i proceed to pay? Jun 02 16:39:52 Please let us know when they deduct the customs from the beagle account so that we can ask them to deliver. Also better if you have a receipt that that we can show them sometimes the guys on the last leg are to fussy Jun 02 16:40:28 Do you have a beagle image with everything working with beaglelogic Jun 02 16:40:36 Yes they were asking for proof of any kind. Jun 02 16:41:08 deepankarmaithan: Not yet, making a little progress though Jun 02 16:41:08 the guy on the phone had no proof to offer. :-( Jun 02 16:41:21 okay :) Jun 02 16:41:47 did your account already got deducted with the amount Jun 02 16:42:07 deepankarmaithan: Use this as a starting point - https://beaglelogic.readthedocs.io/en/latest/beaglelogic_system_image.html Jun 02 16:42:50 downloading right now. Jun 02 16:43:23 Once you download, you need to chdir to /opt/BeagleLogic , do a git pull and reinstall. Jun 02 16:43:53 Tracking page suggests so, it was written there that payment is required now it's not showing that. jkridner can you please confirm? Jun 02 16:44:58 you mean on the DHL page or FEDEX? Jun 02 16:45:12 FedEx Jun 02 16:45:37 Okay i need to get a new sd card and will try tommrrow and then will get back if i will have any querries Jun 02 16:45:52 it never showed me any such thing Jun 02 16:45:54 Sure Jun 02 16:46:32 lorforlinux: you mean on this page https://www.fedex.com/apps/fedextrack/index.html?tracknumbers=182485642510&cntry_code=us Jun 02 16:46:51 Don't know what's the deal with FedEx. Jun 02 16:47:33 yes Jun 02 17:21:44 nwan: can you try disabling the WiFi driver? Jun 02 17:23:16 nwan[m]: are you still having problem with the AI? Jun 02 17:26:27 @jkridner:matrix.org: i will try when i get back home. Jun 02 17:29:10 @pdp7: Yes I tried running kernel 5.4-rt on AI but henrix said that it would be better to work with a stable kernel for GSoC. That would be 4.14-rt at the moment right? Jun 02 18:12:18 I think 5.4 should be ok Jun 02 18:12:30 you just need to not use the broadcom driver Jun 02 18:12:38 as it will break PREEMPT_RT Jun 02 18:13:28 /etc/modprobe.d/brcmfmac-blacklist.conf Jun 02 18:13:28 blacklist brcmfmac Jun 02 18:13:40 should fix it Jun 02 18:14:20 that brcmfmac is a nasty vendor driver that we have to use because they have not gotten SoftAP into their mainline driver and we use SoftAP for the new user experience Jun 02 18:14:27 but it does not matter for your application Jun 02 18:15:08 please try 5.4-rt, blacklist brcmfmac and let me know if it boots ok Jun 02 18:15:25 it would be good to know if there are any other issues Jun 02 18:16:42 the #linux-rt folks have told they will help investigate if we run into issues (of course outside the nasty vendor wifi driver) Jun 02 18:18:37 and for RT, I think there are some important improvements from 4.19 to 5.4 Jun 02 18:18:58 PREEMPT_RT has come a long way :) Jun 02 18:19:05 and 4.19 was 2 years ago Jun 02 18:25:00 nwan[m]: this is a useful page that show the RT versions and who maintains them https://wiki.linuxfoundation.org/realtime/preempt_rt_versions Jun 02 18:40:05 pdp7: Very helpfull. Thanks. I just compiled the kernel completely without wifi modules as jason told. I am currently flashing. Will tell you if this fixes things Jun 02 18:53:47 jkridner: disabling wifi drivers didnt work out. Got the same errors. I am trying now pdp7 's idea Jun 02 18:58:42 pdp7: I made a file brcmfmac-blacklist.conf with the line blacklist brcmfmac. But sadly Kernel still crashes during boot. This is the error message: https://pastebin.com/riG4x0sA Jun 02 19:16:40 nwan[m] interesting results Jun 02 19:16:50 how did you build the kernel? Jun 02 19:16:54 i will try to reproduce Jun 02 19:18:23 it looks like it is failing in the threaded interrupt handler for i2c Jun 02 19:19:05 are you using this https://github.com/RobertCNelson/ti-linux-kernel-dev/tree/ti-linux-rt-5.4.y ? Jun 02 19:19:28 I'll ask over in #linux-rt to see if they have any suggestions Jun 02 19:20:03 but i want to make sure I am looking at the same kernel source as what generated that backtrace Jun 02 19:24:19 pdp7: No I built from this tree https://github.com/beagleboard/linux/tree/5.4-rt Jun 02 19:29:21 * nwan[m] sent a long message: < https://matrix.org/_matrix/media/r0/download/matrix.org/lmpcbwsdYCJNKUpVDmsmPVTh > Jun 02 19:53:05 pdp7: ah I built it from my dev branch, which is quite old: https://github.com/NiklasWan/linux/tree/dev_ctag_5.4-rt Jun 02 19:53:19 Trying to buiild from current 5.4-rt now Jun 02 20:33:49 Still the same error Jun 02 20:48:32 how do you build https://github.com/NiklasWan/linux/tree/dev_ctag_5.4-rt ? Jun 02 20:48:53 meaning, which config do you use Jun 02 20:49:07 bb.org_defconfig Jun 02 20:49:11 ok Jun 02 20:49:19 i'll give it a try Jun 02 20:49:25 cool Jun 02 20:49:47 I really appreciate it. Thanks for your help Jun 02 20:49:54 meanwhile, have you built 4.19? it would be good to compare the results Jun 02 20:50:41 yes I did it today. It worked, but I got some error messages while running and while running sudo poweroff BBAI crashed Jun 02 20:51:22 do you have a pastebin with those errors? Jun 02 20:51:36 Wait i made pictures of it. I will send it over. Sadly no pastebin Jun 02 20:53:24 * nwan[m] uploaded an image: 4.19errors.jpeg (257KB) < https://matrix.org/_matrix/media/r0/download/matrix.org/tgNjDCKXrzPjRDPTdBqFnwUn > Jun 02 20:53:51 No errors but blocking infos Jun 02 21:18:58 interesting to take a look at Jun 02 21:19:24 but yeah probably not urgent unless you have experiencing issues during run-time Jun 02 21:19:37 I'd say it does make sense for you to continue with 4.19 Jun 02 21:20:08 I'll try out 5.4 and see what the linux-rt people think, but not worth you blocking on 5.4 working Jun 02 21:20:56 Okay then I will start on 4.19, maybe during the course of the project 5.4 starts working and if porting existing stuff is not too much work we could do that Jun 02 21:23:23 yeah, that sounds like a good approach Jun 02 21:24:18 a good outcome from GSoC would be that 5.4 RT is validated to work on the AI... or least we can recommend people stay on 4.19 RT Jun 02 21:24:24 Cool, thank you very much for your help. I'll go to sleep now. See you Jun 02 21:24:42 right now, I don't think many people have tried PREEMPT_RT on the BeagleBone AI at all Jun 02 21:24:59 yes I think so too Jun 02 21:25:13 so this a good exercise to see how well it works and what the issues are Jun 02 21:25:42 hopefully it will benefit others, like Bela, who I would really like to see working on AI Jun 02 21:26:20 Yes they could make interesting stuff with this guy Jun 02 21:26:33 they are a little different in terms of requirements, given Xenomai and the PRU usage, but i would like to be able to tell them, you this RT kernel. We know it works Jun 02 21:27:10 Do you know on which kernel version they currently are? Jun 02 21:27:22 it need to check their image Jun 02 21:27:31 in fact, I should touch base with them Jun 02 21:27:41 it's been months since i've checked in with them Jun 02 21:28:43 https://github.com/BelaPlatform/bela-image-builder/releases/tag/v0.3.8 Jun 02 21:28:52 they are still on 4.14 Jun 02 21:29:17 so they updating might be interesting for them Jun 02 21:41:18 yeah, they also have the constraint of Xenomai support, and they need to also get their PRU code working on the AM5729... which hopefully isn't too hard Jun 02 21:53:49 nwan[m]: is this just meant to be your own branch in case you change anything https://github.com/NiklasWan/linux/tree/dev_ctag_5.4-rt or did you already make changes? Jun 02 21:53:55 I'm cloning (slowly) Jun 02 21:54:14 github can't handle a compare across kernel branches :) Jun 02 21:55:02 i should have probably just added a remote, oh well Jun 02 21:55:19 No changes so far just branch from 5.4-rt. I think from the mid of march Jun 02 21:55:38 ah ok Jun 02 21:55:52 nwan[m]: when is your weekly meeting time? Jun 02 21:56:17 dev_ctag_4.19-rt has allready a working driver for ctag face on BBB Jun 02 21:56:19 i think i have it some calendar but can't seem to find it Jun 02 21:56:25 ok, nice Jun 02 21:56:41 Thursday at 17.30 Jun 02 21:56:53 You are also CEST right? Jun 02 21:57:05 yeah, Berlin Jun 02 21:57:11 what city are you? Jun 02 21:57:25 Kiel, Germany Jun 02 21:57:37 All the way up in the north Jun 02 21:58:16 ah, yes, henrix[m] was telling me about it. he just so happened to be visiting my next neighbor a couple weeks ago. small world :) Jun 02 21:59:14 Haha yes. henrix actually lives 3 streets away from my place :D Jun 02 23:48:55 nwan[m]: compile finished, but might need to wait until morning to try it since late. Jun 02 23:48:55 btw, this might be of interest from #linux-rt over on OFTC Jun 02 23:48:55 rt has remained in the 300-500 commit range for 10+ years. But each kernel release is about 13,000+ commits. Jun 02 23:48:55 12:39 AM DrewFustini, so the choice becomes one of "what are they trying to do/achieve?" -- and hence what is more important to them - the rt changes, or the 50,000+ commits between v4.19 and v5.4? Jun 02 23:48:55 12:40 AM if they are trying to align with an existing product/distro-kernel then it makes sense to pick what is closest to that. Jun 02 23:48:55 12:41 AM if they are trying to work on making "new" drivers rt-safe, then using what is newest makes sense. Jun 02 23:49:20 i thought it was a interesting perspective **** ENDING LOGGING AT Wed Jun 03 02:59:57 2020