| *** tpb has joined #timvideos | 00:00 | |
| *** tweakoz has quit IRC | 00:35 | |
| xobs | CarlFK: thanks for the report. fixed! | 00:36 |
|---|---|---|
| CarlFK | \o/ | 00:36 |
| CarlFK | xobs: how much do you know about the low end power supply market? like do either of these look familiar: | 00:47 |
| CarlFK | https://www.ebay.com/i/361282386766 | 00:47 |
| tpb | Title: Brand Adapter 12V + 5V For HARD DISK Drive Power Supply CD/DVD-Rom with AC Cord | eBay (at www.ebay.com) | 00:47 |
| CarlFK | https://www.aliexpress.com/item/32842702383.html | 00:48 |
| tpb | Title: vovotrade Power Supply 12V+5V AC Adapter for Hard Disk Drive HDD CD DVD ROM Drop Shipping-in Chargers from Consumer Electronics on Aliexpress.com | Alibaba Group (at www.aliexpress.com) | 00:48 |
| mithro | CarlFK: should be a new firmware build up for the opsis now | 00:49 |
| mithro | CarlFK: All green now -> https://travis-ci.org/timvideos/HDMI2USB-litex-firmware | 00:49 |
| tpb | Title: Travis CI - Test and Deploy Your Code with Confidence (at travis-ci.org) | 00:49 |
| *** tweakoz has joined #timvideos | 00:50 | |
| CarlFK | mithro: yay! thank you! and thank you for sorting out my udev rules problem I created for myself | 00:50 |
| xobs | CarlFK: hard to say. Can't really tell much just by a photo, which might not even be what you get. The second one is pretty light, and is cheaper, so I'd go for the ebay one instead. | 00:50 |
| CarlFK | xobs: thanks | 00:50 |
| CarlFK | aliexpress Weight: Approx 120g , did you see a weight for the ebay one? | 00:54 |
| xobs | No, unfortunately. | 00:56 |
| CarlFK | mithro: https://travis-ci.org/timvideos/HDMI2USB-litex-firmware/jobs/553467849#L3974 const char* git_describe = "v0.0.4-502-g2bc1d19"; | 01:10 |
| tpb | Title: Travis CI - Test and Deploy Your Code with Confidence (at travis-ci.org) | 01:10 |
| CarlFK | mithro: python3 ../HDMI2USB-mode-switch/bin/download-prebuilt-firmware.py --platform opsis | 01:10 |
| CarlFK | Channel unstable is at rev v0.0.4-502-g2bc1d19 | 01:10 |
| CarlFK | rev: v0.0.4-502-g2bc1d19 | 01:10 |
| CarlFK | Found platforms: arty, basys3, ice40_hx8k_b_evn, ice40_up5k_b_evn, icebreaker, mimas_a7, neso, netv2, nexys_video, tinyfpga_bx, upduino_v1 | 01:11 |
| CarlFK | mithro: no opsis | 01:11 |
| mithro | Hrm, looks like it uploaded successfully -> https://travis-ci.org/timvideos/HDMI2USB-litex-firmware/jobs/553467849#L5722 | 01:12 |
| tpb | Title: Travis CI - Test and Deploy Your Code with Confidence (at travis-ci.org) | 01:12 |
| mithro | Looks like it's there -> https://github.com/timvideos/HDMI2USB-firmware-prebuilt/tree/master/archive/master/v0.0.4-502-g2bc1d19/opsis/hdmi2usb/lm32 | 01:13 |
| tpb | Title: HDMI2USB-firmware-prebuilt/archive/master/v0.0.4-502-g2bc1d19/opsis/hdmi2usb/lm32 at master · timvideos/HDMI2USB-firmware-prebuilt · GitHub (at github.com) | 01:13 |
| CarlFK | mithro: got it. rm github.pickle ... caching is hard. | 01:15 |
| CarlFK | mithro: no hdmi2usb for Atlys: https://github.com/timvideos/HDMI2USB-firmware-prebuilt/tree/master/archive/master/v0.0.4-502-g2bc1d19/atlys | 01:21 |
| tpb | Title: HDMI2USB-firmware-prebuilt/archive/master/v0.0.4-502-g2bc1d19/atlys at master · timvideos/HDMI2USB-firmware-prebuilt · GitHub (at github.com) | 01:21 |
| mithro | https://travis-ci.org/timvideos/HDMI2USB-litex-firmware/jobs/553467848 | 01:30 |
| tpb | Title: Travis CI - Test and Deploy Your Code with Confidence (at travis-ci.org) | 01:30 |
| mithro | - Did not find atlys/hdmi2usb/lm32 at v0.0.4-502-g2bc1d19 | 01:30 |
| mithro | Transmitting file data .......................................................................svn: E160024: Commit failed (details follow): | 01:31 |
| mithro | svn: E160024: MERGE failed | 01:31 |
| CarlFK | oh swell. | 01:31 |
| mithro | Will try kicking it off again | 01:31 |
| CarlFK | opsis tests pass!! | 01:33 |
| CarlFK | er no. tests did not pass :( | 01:36 |
| * CarlFK[m] uploaded a video: VectorVideo_2019-07-04_084315.mp4 (8828KB) < https://matrix.org/_matrix/media/v1/download/matrix.org/ixNBUFXNZDEZGJYLfbsoHLuK > | 01:52 | |
| CarlFK | mithro: that is H2U 00:20:58>x c 0 0 ; Connecting input0 to output0 | 01:53 |
| CarlFK | output0 to an lcd monitor | 01:54 |
| mithro | CarlFK: Committed revision 7403. | 02:01 |
| mithro | https://travis-ci.org/timvideos/HDMI2USB-litex-firmware/jobs/553467848 | 02:01 |
| tpb | Title: Travis CI - Test and Deploy Your Code with Confidence (at travis-ci.org) | 02:01 |
| CarlFK | mithro: opsis: output1 shows pattern and expected version: v0.0.4-502-g2bc1d19 | 02:08 |
| CarlFK | mithro: gst-launch-1.0 v4l2src num-buffers=3 device=/dev/hdmi2usb/by-num/opsis0/video ! jpegdec ! fakesink | 02:09 |
| CarlFK | no valid frames found | 02:09 |
| *** tweakoz has quit IRC | 02:51 | |
| mithro | CarlFK: does it include your changes to the output? | 03:03 |
| CarlFK | mithro: I don't think so - but I'm also not sure what that "black code" made black? | 03:04 |
| mithro | Well it hasn't been merged I don't think? https://github.com/timvideos/HDMI2USB-litex-firmware/pull/475 | 03:09 |
| tpb | Title: Pattern closes #474 by CarlFK · Pull Request #475 · timvideos/HDMI2USB-litex-firmware · GitHub (at github.com) | 03:09 |
| CarlFK | mithro: I thought I saw that it was merged. let me see if "pattern" comnand works | 03:10 |
| Kamilion | CarlFK: I remember getting one of those with a cheap USB to SATA dongle that just popped on the back of a drive. Havn't seen any for close to a decade. | 03:11 |
| CarlFK | Kamilion: yeah, that's how i discovered them too. was hoping you used them at work for .. um. drives :p | 03:12 |
| Kamilion | Nothing's used molex since the IDE days. | 03:12 |
| Kamilion | just about everything's moved to SFF-8482 | 03:13 |
| Kamilion | CarlFK: https://www.amazon.com/AGPtek-Drive-Adapter-Converter-External/dp/B00BIE996S/ | 03:15 |
| CarlFK | mithro: you are right, that PR not merged) H2U 00:45:52>pattern = nothing | 03:15 |
| Kamilion | $14 good enough deal? | 03:15 |
| CarlFK | Kamilion: https://www.ebay.com/i/361282386766 $10 free shipping arrives in 5ish days. | 03:18 |
| tpb | Title: Brand Adapter 12V + 5V For HARD DISK Drive Power Supply CD/DVD-Rom with AC Cord | eBay (at www.ebay.com) | 03:18 |
| CarlFK | if I could see that both grounds were hooked up, I might pay the extra $4 | 03:18 |
| CarlFK | actually the sticker shows they are not hooked up | 03:19 |
| Kamilion | ah, i assumed arrival time was more important, so I looked for anything Prime | 03:20 |
| CarlFK | I need them glued in place by ~ july 22 | 03:20 |
| CarlFK | I do see 2.5A, which might be a little more real than the Approx 120g 2A (which I don't really believe, but I'm not sure how much I need anyway? | 03:22 |
| CarlFK | I am using them to power the 12v fpga board and a 5v sbc, either c2 or turbot | 03:22 |
| CarlFK | turbot has an atom with a heat sink and does draw enough current to not work right on what I setup for the C2 | 03:23 |
| CarlFK | so something is an issue, I just have no idea what the real numbers are | 03:23 |
| Kamilion | ah. I dunno either. I've pretty much been using nothing but those FSP084-DIBAN 12V/7A bricks | 03:24 |
| Kamilion | I don't have any spare 12V->5V on hand right now either | 03:24 |
| Kamilion | but I normally use https://www.amazon.com/SMAKN®-Converter-Supply-Module-OUTPUT/dp/B00R5JL8WI/ (but from aliexpress instead, at 1/3rd the unit cost) | 03:26 |
| Kamilion | either the SMAKN or DROK brand is what I go with. | 03:27 |
| Kamilion | used my last spare one installing a bluetooth head unit in a friend's tacoma truck | 03:28 |
| CarlFK | I like the 12/5 in one ps - less stuff to assemble and wonder about | 03:28 |
| *** rohitksingh has joined #timvideos | 03:28 | |
| Kamilion | yeah, but then you've got a pile of weird one-off parts like that molex brick that are hard to reproduce/restock | 03:30 |
| Kamilion | plus you have no idea the quality of those bricks, and *they were meant for external harddrives* with noisy DC motors. There's no way they bothered spending enough money for a really prim and clean output | 03:31 |
| Kamilion | and more or less I bet those bricks just have a 12V-5V linear reg after the switchmode AC->12V | 03:32 |
| Kamilion | Oh, sorry, it sounds like I'm arguing with you; that's not my intent. | 03:32 |
| Kamilion | I'm just really picky about clean power, it resolves so many quirky behaviors | 03:33 |
| Kamilion | oh, reminds me, if you're around the bay area and you see a raspberry pi 4 for sale; lemme know. I can't seem to find any yet. | 03:34 |
| CarlFK | na, thats the kind of feedback I'm interested in | 03:34 |
| CarlFK | https://www.ebay.com/itm/AC-to-12V-5-Volt-DC-Power-Supply-Adapter-w-4-Pin-Molex-Connector-Plug-SW12V5V2A-/132218098446 | 03:34 |
| CarlFK | I want to know what the AC plug is.... | 03:35 |
| Kamilion | "brand: ITE" | 03:35 |
| CarlFK | Connector A: 4 Pin 12 V 5V; Connector B: 4 Pin 12 V 5V | 03:35 |
| CarlFK | um.. no. | 03:35 |
| Kamilion | yeah, that thing isn't even branded at all | 03:36 |
| Kamilion | but the AC side's probably the normal IEC C13 | 03:37 |
| Kamilion | very unlikely it's IEC C5 (Clover) | 03:38 |
| *** rohitksingh has quit IRC | 03:42 | |
| CarlFK | why unlikely? | 03:43 |
| CarlFK | other than there seems to be some pattern: 2 grounds on the molex side has clover AC side. | 03:43 |
| Kamilion | Clover doesn't tend to show up a lot outside of laptop power supplies, for some reason. | 03:44 |
| Kamilion | I *think* it is because of older airplane AC provisions | 03:45 |
| Kamilion | and "travel" adapters tend to have C5 Clover, while stationary adapters tend to have C13 | 03:46 |
| Kamilion | at this point, C13's just about the easiest to find worldwide | 03:46 |
| Kamilion | to the point of a friend of mine that travels a lot carries a C14-to-C5 dongle in his bag and just buys C13 leads locally wherever he travels. | 03:48 |
| Kamilion | https://www.amazon.com/CLOVER-FEMALE-POWER-ADAPTER-SOCKET/dp/B0793J6LCP/ | 03:48 |
| CarlFK | https://www.amazon.com/Coolerguys-100-240v-Molex-Power-Adapter/dp/B000MGG6SC/ | 03:48 |
| Kamilion | wow, more expensive, and not even grounded. | 03:49 |
| CarlFK | both are easy to come by, I like the c13 because I have lots of 12" cords | 03:49 |
| *** tweakoz has joined #timvideos | 03:50 | |
| CarlFK | the molex has both grounds, | 03:50 |
| CarlFK | on one of mine, um.. the 2 grounds aren't directly hooked together | 03:51 |
| Kamilion | yep, and from the customer photos, that looks like a branded supply, has proper warnings, and is likely double insulated (to have an earth-less inlet) | 03:51 |
| CarlFK | continuity tester... only one direction | 03:51 |
| CarlFK | does that imply anything about what's inside? | 03:51 |
| CarlFK | like 2 power supplies, or anything I might care about | 03:52 |
| Kamilion | Generally not; the cost of having two inductors would be prohibitive | 03:52 |
| CarlFK | good - I guess :p | 03:53 |
| Kamilion | so in almost all cases, a switching brick is gonna drop to the highest output voltage; and then have a regulator hanging off that output to provide any lower voltages | 03:53 |
| Kamilion | if the two grounds aren't connected; they may be using a nicely isolated 12V->5V device | 03:53 |
| Kamilion | that's definitely a plus, if so | 03:54 |
| CarlFK | swell. so I should track that one down and get more :p | 03:55 |
| CarlFK | its at PS1 - about 10 min drive. Ill go in tomorrow | 03:56 |
| Kamilion | but noise/load on the 12V side can still affect the 5V side... but with these modern 8V-36V buck regulators and a good quality low esr bulk capacitor, even a noisy 12V in can have a clean 5.08V output with a ripple under 5mv | 03:56 |
| Kamilion | those dual voltage bricks have been kicking around since the 90s, when I got my first 5.25" external CD-ROM chassis | 03:57 |
| Kamilion | almost a given that a few corners will be cut over 25 years | 03:58 |
| Kamilion | Least it's better than my atari's big ass transformer -> linear reg supply | 03:58 |
| CarlFK | the 12v is the fpga, which doesn't even have a heat sink, and nothing gets hot to the touch, so I don't think it's going to put much load on the 12v side | 03:59 |
| Kamilion | yeah, i doubt you'll have an issue unless you were doing anything like 3D printing / stepper control / motor control, with a lot of EMF noise | 03:59 |
| Kamilion | however, you can't really tell the age of these dual voltage bricks, and it's certainly possible to get old stock with craptastic dried out capacitors that'll only get worse | 04:00 |
| Kamilion | the Coolerguys brick does seem to be recent though, at least from the customer pics | 04:00 |
| Kamilion | https://images-na.ssl-images-amazon.com/images/I/615M3boQzpL._SL1600_.jpg | 04:01 |
| CarlFK | mithro: can you flip the order of the builds around so that hdmi2usb/opsis is done first? | 04:28 |
| CarlFK | mithro: I'm kinda assuming the rest of them are to confirm it builds, not for anyone to use | 04:29 |
| *** samsagaz has quit IRC | 05:18 | |
| *** rohitksingh_work has joined #timvideos | 05:26 | |
| *** rails has quit IRC | 05:41 | |
| *** CuriousLearner has quit IRC | 05:42 | |
| *** sb0 has joined #timvideos | 05:52 | |
| *** m4ssi has joined #timvideos | 06:02 | |
| CarlFK | mithro: find Opsis? | 07:05 |
| *** jea has quit IRC | 10:23 | |
| *** sb0 has quit IRC | 12:26 | |
| *** jea has joined #timvideos | 12:30 | |
| *** rohitksingh_work has quit IRC | 13:07 | |
| *** tweakoz has quit IRC | 13:22 | |
| *** rohitksingh has joined #timvideos | 14:25 | |
| *** rohitksingh has quit IRC | 14:30 | |
| *** rohitksingh has joined #timvideos | 14:32 | |
| *** rohitksingh has quit IRC | 14:57 | |
| *** rohitksingh_ has joined #timvideos | 14:57 | |
| *** rohitksingh_ has quit IRC | 15:04 | |
| *** rohitksingh has joined #timvideos | 15:04 | |
| *** rohitksingh has quit IRC | 15:42 | |
| *** rohitksingh_ has joined #timvideos | 15:42 | |
| *** aiena has joined #timvideos | 15:52 | |
| aiena | Hi CarlFK | 15:52 |
| CarlFK[m] | Hi aiena | 15:55 |
| aiena | the kernel has so many make commands (: | 16:02 |
| aiena | my kernel didnt build without running make defconfig and then make later | 16:02 |
| aiena | now someone is saying there is a make modules-config too | 16:02 |
| CarlFK | did you read all 120 of Rustys slides ? | 16:04 |
| *** m4ssi has quit IRC | 16:06 | |
| aiena | CarlFK: not fully its a lot of info but I am trying to understand as I read. | 16:23 |
| aiena | I mean each slide makes you want to read related stuff to understand it sometimes yesterday I watched the talk you linked to | 16:24 |
| aiena | It was very interesting and helped me understand USB better | 16:24 |
| aiena | today morning was spent on understanding USB rather than the kernle (: | 16:24 |
| aiena | kernel | 16:24 |
| aiena | CarlFK: in the slides they say "System calls are the legal method of trapping into the kernel from user-space" | 16:26 |
| aiena | what are system calls and traps here. | 16:26 |
| aiena | hmm dir structure changed a bit cool | 16:32 |
| CarlFK | aiena: system calls are function calls | 16:42 |
| aiena | and traps are like interrupts | 16:42 |
| CarlFK | I think so. that one I have never used | 16:43 |
| aiena | but syncronous so a trap can pause a system ca;; | 16:43 |
| aiena | call | 16:43 |
| aiena | so void main() would be a sys cal | 16:43 |
| aiena | call | 16:43 |
| CarlFK | void main() is a function | 16:43 |
| aiena | err actually no | 16:43 |
| aiena | hmm ok so when the function is run | 16:44 |
| aiena | the memory and the processing done is the sys call | 16:44 |
| aiena | i mean like mem allocation computation, freeing | 16:44 |
| aiena | hmm wait sorry | 16:45 |
| aiena | I think its like when you pass params to a function and run it | 16:45 |
| aiena | so like main(a); is the system call and void main() is the function | 16:47 |
| CarlFK | yes. | 16:47 |
| aiena | *void main (int foo) | 16:47 |
| aiena | ok like normal programming | 16:48 |
| CarlFK | except main is your function, not a function in the kernel | 16:48 |
| * aiena nods | 16:48 | |
| aiena | the entry point in the rusty docs is an assembly file for syscalls | 16:49 |
| aiena | since I don't know assembly I cannot make a lot of sense of it | 16:49 |
| CarlFK | https://medium.com/anubhav-shrimal/adding-a-hello-world-system-call-to-linux-kernel-dad32875872 | 16:50 |
| tpb | Title: Adding a Hello World System Call to Linux Kernel - Anubhav Shrimal - Medium (at medium.com) | 16:50 |
| CarlFK | more kenel bulding dos ;) | 16:50 |
| *** rohitksingh_ has quit IRC | 16:50 | |
| aiena | hehe | 16:52 |
| aiena | CarlFK: is obj-y : for assembly in cmake? | 16:53 |
| aiena | ah it means builtin lol\ | 16:54 |
| CarlFK | obj-y ? | 16:55 |
| aiena | oh cool its linked to the kconfig system | 16:55 |
| aiena | so obj-y would make the file be built if its part of the core kernel config only the tick boxes in make-xconfig I think | 16:56 |
| aiena | CarlFK: yes in the makefiles | 16:56 |
| *** rohitksingh has joined #timvideos | 16:59 | |
| CarlFK | ah, i see it now. um.. it's a make label. that is referenced somewhere else. (im really just guessing, only 1/2 paying attention to it) | 17:01 |
| aiena | CarlFK: hmm basically the kernel has a kconfig system | 17:04 |
| aiena | so you can choose what support you need built itno the kernel etc. | 17:05 |
| aiena | from what I saw for example which crypto cypers etc | 17:05 |
| aiena | CarlFK: like http://paste.opensuse.org/view/raw/62b361bc | 17:08 |
| aiena | it added that entry to the kconfig | 17:08 |
| aiena | then if it is ticked then the corresponding .c or .S file is compiled into object code for linking | 17:08 |
| aiena | so thats how you enable disable kernel features like protocols etc. that a linux kernel thingy which is pretty awesome in a way. | 17:09 |
| CarlFK | yep | 17:10 |
| aiena | the obj-$() stuff is the makefile level then the kconfig is at this level | 17:10 |
| aiena | so kconfig sets the variable for the makefile to use which in turn tells make to build or not build the whatever | 17:10 |
| CarlFK | sounds about right | 17:11 |
| aiena | ok | 17:11 |
| CarlFK | there are a few goals of the kernel config and make process: keep the size of the kernel small (the less ram it needs the more there is for your apps) and don't spend time compiling things you don't want | 17:13 |
| aiena | Ok that's the impression I got. | 17:16 |
| aiena | so amm these menuconfig /xconfig etc. just write to the .config file in the source root tree | 17:16 |
| aiena | and this file in turn feeds make etc. | 17:17 |
| aiena | I'll just copy over the system config file just to see what ubuntu has enabled by default | 17:18 |
| CarlFK | yep | 17:18 |
| aiena | if I use "defconfig" type-c support is disabled I feel ubuntu would keep it enabled by default | 17:18 |
| CarlFK | I have no idea | 17:19 |
| aiena | but it doesnt matter for me at the moment I think defconfig is sufficient | 17:19 |
| aiena | hmm actually wait it may be better to use the ubuntu config | 17:19 |
| aiena | something may break if the new kernel is loaded and missing something used by all the other packages on the host | 17:20 |
| CarlFK | it's very unlikely that anything will break | 17:20 |
| aiena | if I use defconfig | 17:21 |
| aiena | all my fs's are ext4 but I use one ntfs partition | 17:21 |
| CarlFK | odd hardware might not be supported because ubuntu enabled it and vanilla didn't | 17:21 |
| aiena | wonder if defconfig drops ntfs support | 17:22 |
| CarlFK | I doubt it - many thumb drives are ntfs | 17:22 |
| aiena | CarlFK: strange | 17:24 |
| aiena | how come I cannot see NTFS anywhere under filesystems | 17:24 |
| aiena | I checked under the misc FS subcategory too | 17:24 |
| aiena | and the network FS ones | 17:24 |
| aiena | do you know | 17:26 |
| CarlFK | yes.. I think it is called hpfs | 17:26 |
| aiena | CarlFK: found it | 17:28 |
| aiena | its in DOS/FAT/NT Filesystems category and | 17:29 |
| aiena | "OFF" by default | 17:29 |
| CarlFK | here is the last thing I did http://paste.ubuntu.com/p/S6xYWb27tX/ | 17:29 |
| tpb | Title: Ubuntu Pastebin (at paste.ubuntu.com) | 17:29 |
| aiena | scary like type-c | 17:29 |
| aiena | CarlFK: what is the difference between make oldconfig and make defconfig? | 17:30 |
| CarlFK | to test if vizzini.patch still worked. it did. now if only someone would get it submitted into upstream linux grumble... | 17:30 |
| CarlFK | oldconfig is what your current kernel is running | 17:30 |
| CarlFK | defconfig is what Linus thinks is a safe config for eveyone | 17:31 |
| aiena | oh so I sohuld use oldconfig | 17:32 |
| aiena | and that will use the .config file of the distro as a recipe automagically? | 17:32 |
| * aiena likes to understand some of the magic wherever possible | 17:32 | |
| CarlFK[m] | I think so,. | 17:33 |
| * aiena wonders how oldconfig werks | 17:33 | |
| CarlFK[m] | Look in /boot | 17:33 |
| CarlFK[m] | You should see a bunch of config files with the same name as the colonel files | 17:34 |
| aiena | i am looking at the linux makefile to see oldconfig logic | 17:35 |
| aiena | I cannot find it in the root makefile where is this config stuff defined | 17:35 |
| aiena | yes ./config-5.0.0-20-generic looks exactly like the .config that defconfig would generate but with different options.pages | 17:36 |
| aiena | s/pages/ | 17:37 |
| aiena | oh found it | 17:38 |
| aiena | it runs the conf executable and oldconfig is in C code | 17:39 |
| aiena | CarlFK: How does oldconfig automatically pickup the current kernle conf though can't make out in the C code | 17:40 |
| aiena | Hmm why is oldconfig not taking the distro options | 17:42 |
| aiena | after running make oldconfig I just ran xconfig and looked at NTFS | 17:43 |
| aiena | In ubuntu's config file `CONFIG_NTFS_FS=m` is set but running config oldconfig alone seems to keep it unset | 17:43 |
| aiena | on my source code kernel tree's ,config | 17:44 |
| aiena | yes but copying the .config file preserves the correct settings | 17:45 |
| aiena | CarlFK: so there is something I am missing with make oldconfig | 17:45 |
| aiena | that alone is almost identical in behaviour to make defconfig | 17:45 |
| aiena | wonder what am I missing | 17:46 |
| CarlFK | im not sure - hard to understand exactly what you have done - personally I wouldn't spend any more time on it | 17:47 |
| CarlFK | it is good to understand, but not that good, so spend your time on getting your kernel module working | 17:48 |
| * aiena nods I just copied over ubuntu's config | 17:48 | |
| aiena | from /boot/.config-5.x.x. /source/tree/.config | 17:49 |
| aiena | CarlFK: how do I set the extraversion | 17:49 |
| aiena | I am guessing I need to make install my built kernel too | 17:50 |
| CarlFK | I did extraversion thing like this: make -j `getconf _NPROCESSORS_ONLN` deb-pkg LOCALVERSION=-viz2-torv | 17:51 |
| CarlFK | but I'm sure it is in the .conf somswehre | 17:51 |
| CarlFK | juser@gator:~/vizzini/shenki/linux$ grep VERSION .config | 17:52 |
| CarlFK | CONFIG_LOCALVERSION="" | 17:52 |
| aiena | yep seems like I need to edit it in the file the xconfig area doenst have a way to enter strings | 17:56 |
| aiena | ok now the rusty thing only mentions building the linux kernel and then loads it in qemu | 17:57 |
| aiena | the commands are a bit confusing as in in the sildes he seems to create a ramdisk with mount then unmount it | 17:57 |
| aiena | wont the ramdisk get just destroyed | 17:57 |
| CarlFK | whats the command to make the ramdisk? | 17:58 |
| CarlFK | it has been a year or two sense I read it | 17:58 |
| aiena | hmm wait sorry no ramdisk | 17:59 |
| aiena | command is like this mount -o loop -o offset=32256 ../debian /mnt/qemu | 17:59 |
| aiena | I need to install qemu first but do you have instructions sans virtualisation | 17:59 |
| CarlFK | apt install qemu | 18:00 |
| aiena | ok I unstalled qemu | 18:02 |
| aiena | now ehat is this ../debian /mnt/qemu with the offset | 18:03 |
| CarlFK | I always remember "mount something somewhere" | 18:04 |
| CarlFK | in this case, add some -options: mount -options something somewhere | 18:04 |
| CarlFK | options are: -o loop -o offset=32256 | 18:05 |
| CarlFK | do you know about loopback ? | 18:05 |
| aiena | a loopback is A read only filesystem afaik | 18:06 |
| CarlFK | no | 18:06 |
| aiena | hmm let me see then | 18:06 |
| CarlFK | a fs (filesystem) is a bunch of bytes: filenames, pointers and data and unused space | 18:06 |
| CarlFK | normally those bytes are on a device, either the whole disk, or a partition | 18:07 |
| aiena | nods | 18:07 |
| aiena | nods | 18:07 |
| CarlFK | they can also all be contained in a file | 18:07 |
| aiena | ok like iso or dd output | 18:07 |
| CarlFK | so instead of mounting a device found under /dev you can mount a file | 18:07 |
| CarlFK | right | 18:07 |
| aiena | or squashfs | 18:08 |
| CarlFK | maybe. I think squashfs can be on a device | 18:09 |
| aiena | so ../debian is an iso | 18:09 |
| aiena | or some kind of image file | 18:09 |
| CarlFK | right | 18:09 |
| aiena | and it is mounted to /mnt/usb skipping 32256 bytes or something | 18:09 |
| aiena | err /mnt/qemu | 18:09 |
| CarlFK | ah - offset... | 18:09 |
| aiena | since the ../debian is a relative path idk where this ../debian is or if qemu provide it by default I did apt-install qemu | 18:10 |
| CarlFK | I'll deduce that "debian" is a ... um.. disk with partitions | 18:10 |
| CarlFK | I bet you need to download debian | 18:11 |
| aiena | hmm that looks like it is using the debian images default built kernel with modules from the source treee of the host kernel | 18:12 |
| CarlFK | about offset: if you dd a disk with partititons, like dd if=/dev/sda of=disk.img - and then you want to mount -o loop the first partition, you need to skip over the partition table | 18:12 |
| aiena | oh | 18:12 |
| aiena | so it is a clone of some live installed system | 18:13 |
| aiena | mighty difficult that one tiny line | 18:13 |
| aiena | err bare metal system | 18:13 |
| CarlFK | right | 18:13 |
| aiena | I think the non qemu route not documented in the sildes may be easier | 18:14 |
| aiena | then | 18:14 |
| CarlFK | could be | 18:14 |
| aiena | i can't dd my system | 18:14 |
| CarlFK | why do you think you need to dd your system? | 18:15 |
| aiena | how did that ../debian image used get made? | 18:15 |
| aiena | if the offset is there then | 18:15 |
| aiena | that ../debian must be like you suggested dd'd from some minimal host | 18:16 |
| aiena | it has to be an image file since it is a loopback device | 18:16 |
| aiena | ah wait the image file probably is made simply by installing the distro in the vm like a VDI | 18:17 |
| aiena | I am guessing I can use the ubuntu live cd but will have to update it to use the latest ubuntu kernel | 18:18 |
| aiena | so CarlFK how was this debian image created this is all qemu stuff now | 18:20 |
| aiena | i think its an install debian system on qemu but the slides dont mention anything | 18:20 |
| CarlFK | let me look... | 18:21 |
| aiena | slide #19 | 18:22 |
| aiena | I am sorry I have only used virtualbox for virtualisation. I dod try QEMU but it was difficult because you need root access to make each vm etc. and I was a bit scared. | 18:23 |
| CarlFK | http://mirror.slingshot.co.nz/pub/linux.conf.au/2005/cdrom-beta-2/kerneltute/ | 18:26 |
| tpb | Title: Index of /pub/linux.conf.au/2005/cdrom-beta-2/kerneltute (at mirror.slingshot.co.nz) | 18:26 |
| aiena | CarlFK: https://wiki.debian.org/QEMU | 18:26 |
| tpb | Title: QEMU - Debian Wiki (at wiki.debian.org) | 18:26 |
| CarlFK | debian.gz 19 mb | 18:26 |
| aiena | hmm thats from 2005 | 18:26 |
| aiena | I am building against the 5.x kernel | 18:27 |
| CarlFK | it should work | 18:27 |
| CarlFK | you need it to boot, some simple stdio and bash | 18:28 |
| aiena | yes that code is pretty basic and it builds | 18:28 |
| aiena | let me try I am unfamiliar with qemu | 18:28 |
| aiena | do I ned to extract the gz first? | 18:28 |
| CarlFK | yes | 18:29 |
| aiena | hmm the mirror you pointed to is all the stuff in the state that the kernel was in during that tut | 18:30 |
| aiena | i am getting the debian.gz | 18:30 |
| CarlFK | I like qemu - maybe because it is what I have always used. but it seems like it is easier to use from the command line, so it is easier to document and use for lessons | 18:30 |
| * aiena nods | 18:31 | |
| aiena | I heard qemu with kvm is also faster and more bug free than say virtualbox | 18:31 |
| aiena | people compain about virtualbox creating a mess but as just a user I don't know anything | 18:31 |
| aiena | just you need the root shell lots unless you do some workarounds for qemu so that was ... for me | 18:32 |
| aiena | ok i donwloaded the gz | 18:32 |
| aiena | CarlFK: I wonder if when he says "as root" for those 3 commands if he means "sudo" or "sudo -i" | 18:34 |
| aiena | because sudo -i will remove you from your pwd too | 18:34 |
| CarlFK | it doesnt' really matter - sudo -i is less typing but more dangerous | 18:35 |
| aiena | i am guessing the debian is outside the source tree so it must be sudo | 18:35 |
| aiena | ok | 18:35 |
| CarlFK | um, what does source tree have to do with sudo? | 18:35 |
| aiena | sudo -i will remove the pwd you are cd'd into | 18:36 |
| aiena | so the ../debian makes no more sense | 18:36 |
| aiena | but the ..debian makes sense if you followed http://mirror.slingshot.co.nz/pub/linux.conf.au/2005/cdrom-beta-2/kerneltute/ dir structure | 18:36 |
| tpb | Title: Index of /pub/linux.conf.au/2005/cdrom-beta-2/kerneltute (at mirror.slingshot.co.nz) | 18:36 |
| aiena | *../debian | 18:36 |
| CarlFK | you need root for the mount command | 18:37 |
| aiena | nods | 18:38 |
| aiena | i think all 3 | 18:38 |
| aiena | you cannot make any folder in /mnt without it anyway | 18:38 |
| CarlFK | this 1/2 works: http://paste.ubuntu.com/p/n6xGFSYn4m/ | 18:43 |
| tpb | Title: Ubuntu Pastebin (at paste.ubuntu.com) | 18:43 |
| CarlFK | I'm actually not sure if the debian image is doing anything - it may all be just the ubuntu kernel booting and then crashing | 18:44 |
| aiena | ok before all this qemu sauce my l3rk driver is not compiling | 18:48 |
| aiena | I get a make[2]: *** No rule to make target 'drivers/crypto/lr3k.c', needed by 'drivers/crypto/lr3k.o'. Stop. | 18:48 |
| aiena | error but I followed all the instructions in tohse 3 slides | 18:48 |
| aiena | I dont understnad what I missed this error comes after running "make modules" | 18:48 |
| CarlFK | No rule - that sounds like the Makefile doesn't say how too build lr3k.c | 18:49 |
| CarlFK | or it can't find lr3k.c | 18:49 |
| CarlFK | (I'm not good at make) | 18:49 |
| aiena | oh wait | 18:49 |
| aiena | i got it I think the c file is named wrong lol | 18:50 |
| aiena | yes that was it lol | 18:50 |
| aiena | it built | 18:50 |
| aiena | i think the make-j8 modules step is the long time compiling part of linux | 18:51 |
| aiena | CarlFK: in the slide #19 there is the 'make INSTALL_MOD_PATH=/mnt/qemu modules_install' line I think modules_install will compile if not yet compiled the modules | 18:52 |
| aiena | because of the sudo is everything built now owned by root | 18:52 |
| aiena | do the object files etc. move to root somehow I have had that mess before with building userland apps | 18:52 |
| aiena | hmm no ownership of the object file is still with the user | 18:54 |
| aiena | hmm no ignore that I am running make modules as regular user so that is correct | 18:55 |
| aiena | what is sudo ' 'make INSTALL_MOD_PATH=/mnt/qemu modules_install' expected to do? | 18:55 |
| aiena | i feel it will copy the modules over to /mnt/qemu | 18:55 |
| CarlFK | I'm kinda guessing, but I think it just sets up where they will be installed when you do: "sudo make install modules" later | 18:56 |
| CarlFK | no wait... | 18:57 |
| aiena | CarlFK: the facinating thing is that debian.gz has no kernel but just the basic linux dir structire and some userspace binaries | 18:57 |
| CarlFK | as root: ... make INSTALL_MOD_PATH=/mnt/qemu modules_install\ | 18:57 |
| aiena | ls /mnt/qemu/boot turns out empty | 18:58 |
| aiena | would modules_install also install the kernel? | 18:58 |
| CarlFK | no ... | 18:58 |
| aiena | the /boot is totally empty if you mout the extracted debian image | 18:59 |
| aiena | mount | 18:59 |
| CarlFK | this is setup to make building/booting new kerenels in fewer steps... | 18:59 |
| CarlFK | slide 20 Run qemu: qemu -hda ../debian -kernel arch/i386/boot/bzImage -append “ro root=/dev/hda1” | 18:59 |
| aiena | yes | 19:00 |
| CarlFK | that tells the VM to use a bzImag from the host fs, not debian file | 19:00 |
| aiena | that add the kernel so whatever kernel you built | 19:00 |
| aiena | nods got it | 19:00 |
| CarlFK | right - so debian doesn't need any kernel. | 19:00 |
| CarlFK | (I kinda don't like using "debian" ad the file name here. makes this conversation confusing. | 19:01 |
| aiena | yes | 19:01 |
| aiena | its better to call it skelinux | 19:01 |
| aiena | its barebones devoid of a kernel but it still has chmod etc in it | 19:01 |
| aiena | i guess those interface with the higher level kerne API so they will work as is | 19:02 |
| aiena | and hence are kept in there | 19:02 |
| aiena | oh boy out of disk space | 19:03 |
| CarlFK | lol | 19:03 |
| aiena | kernel hog waaaaaaaah | 19:03 |
| CarlFK | I rand out of memory too. thats what we get using linux with 14 years of additions | 19:03 |
| aiena | 14 gb and the modules aint fully built | 19:04 |
| aiena | how big is your built kernel tree | 19:04 |
| aiena | CarlFK: and this is without the git load too | 19:05 |
| aiena | the nast .git munchkin | 19:05 |
| aiena | nasty | 19:05 |
| aiena | oh it needed 19.7GB | 19:08 |
| aiena | i did some aggressive freeing thank heavens for this beauty called Filelight | 19:08 |
| aiena | so no ican go to the qemu step :P | 19:09 |
| CarlFK | 556M linux-image-5.2.0-rc1-viz2-torv-dbg_5.2.0-rc1-viz2-torv-1_amd64.deb | 19:09 |
| aiena | yeah | 19:09 |
| CarlFK | im not really sure where to get stats on how big my compiled binaries are | 19:10 |
| aiena | that is the packed file | 19:10 |
| aiena | compiling exploads all that exponentially lol | 19:10 |
| aiena | thats why building from source is annoying | 19:10 |
| aiena | 1 app after building maybe say 1GB but will need at least 10GB to build | 19:10 |
| aiena | that 556M needs 19.7 GB :P | 19:11 |
| aiena | when we donwload built apps its like oh its only 5556MB :P | 19:11 |
| aiena | anyway now for the qemu stuff | 19:12 |
| aiena | lol now I cannot install my built modules hahahahahaha | 19:13 |
| aiena | CarlFK: is qemu bound to /mnt/ | 19:14 |
| aiena | I have another drive with more free space | 19:14 |
| aiena | my root partition is on an SSD so its tight (: | 19:15 |
| CarlFK | qemu knows nothing about your /mnt/ | 19:15 |
| CarlFK | mnt is used to alter the debian file, then qemu reads the debian file | 19:15 |
| aiena | oh lol i have no space there too haha | 19:16 |
| aiena | oh now it makes sense | 19:17 |
| aiena | so it updates the skelinux with the kernel and mudles you made | 19:17 |
| aiena | then the notsoskelinux is run by qemu | 19:17 |
| aiena | *kernel and modules | 19:17 |
| CarlFK | modules yes, kernel maybe (I dont' think kernel) | 19:18 |
| aiena | oh right | 19:18 |
| CarlFK | ah right. if it put the kernel there, it would have to tell the boot loader about it (grub maybe) | 19:18 |
| aiena | kernel is just passed as a param | 19:18 |
| CarlFK | yeah - qemu -kernel foo is much easier | 19:18 |
| aiena | nods | 19:19 |
| CarlFK | you are dealing with the birth of the universe - things are not like you are used to :p | 19:19 |
| aiena | so the deb file is updated ony with your modules the rest is static | 19:19 |
| aiena | CarlFK: I know that is what is scary and fun about the kernel | 19:19 |
| CarlFK | what deb file? | 19:19 |
| aiena | everything run on top of it and it interfaces with the hardware directly | 19:20 |
| aiena | CarlFK: lol the debian | 19:20 |
| aiena | the skelinux | 19:20 |
| aiena | why that name is idiotic | 19:20 |
| CarlFK | lol | 19:20 |
| aiena | the deb file is used for deb packages but that file is also called debian | 19:20 |
| CarlFK | I think we can call it debian. we have no need for Debian the distro right now | 19:21 |
| aiena | lets call it skelinux | 19:21 |
| aiena | so the skelinux gets updated and rebundled with your drivers | 19:21 |
| aiena | and then qemu loads notsoskelinux with the built kernel :P | 19:21 |
| aiena | i am stuck at the make the notsoskelinux part | 19:22 |
| aiena | lets see what else I can trash (: | 19:22 |
| aiena | CarlFK: can you do a mock test and see how big your notsoskelinux becomes after installing the modules | 19:23 |
| aiena | oh boy | 19:27 |
| aiena | i know whats happening now | 19:27 |
| aiena | im so silly | 19:27 |
| aiena | the /mnt is on the root partition and I am freeing space on the home partition grrrr | 19:27 |
| aiena | i think I will just unmount /mnt | 19:28 |
| aiena | why is the sudo needed for the makeinstall | 19:28 |
| aiena | make modules_install | 19:28 |
| aiena | CarlFK: do the modules need some root ownership or something in the loop device fs? | 19:29 |
| aiena | i mean can I change /mnt/qemu to /home/user/qemu | 19:29 |
| CarlFK | yes, you can change the mount point | 19:29 |
| aiena | what reverse the make modules_install thing | 19:30 |
| CarlFK | permissions: I suspect you will need to give write perms to your user | 19:30 |
| aiena | is there a make modules_uninstall? | 19:30 |
| aiena | CarlFK: hmm the home directlry is writable non root | 19:30 |
| CarlFK | I doubt it - but you can just delete whtever yo uneed | 19:30 |
| aiena | its not an issue I can always re extract the debian.gz and start afresh :P | 19:33 |
| aiena | ah now I understand what you mean | 19:35 |
| aiena | root will take ownership of the debian file and the mount point | 19:35 |
| aiena | CarlFK: ok now I am sooooo cofunsed | 19:37 |
| aiena | I have 9.3 Gigs free on the partition that I made the mount point on | 19:37 |
| aiena | I get an error like this cp: error writing '/home/aiena/BUILD-AREA/LINUX/qemu//lib/modules/5.0.0aiena/kernel/arch/x86/crypto/cast6-avx-x86_64.ko': No space left on device | 19:38 |
| aiena | i dont know what I am goofing up | 19:40 |
| aiena | CarlFK: is this device the loopback device? | 19:42 |
| aiena | that its complaining has no space? | 19:42 |
| aiena | CarlFK: seems like it lsblk says `loop0 7:0 0 50M 0 loop /home/aiena/BUILD-AREA/LINUX/qemu` this is where the debian file is mounted | 19:44 |
| aiena | so is the loop device size ony 50Mb how do I add more space? | 19:44 |
| CarlFK | hmm.. i may know this... | 19:45 |
| aiena | yup | 19:45 |
| aiena | `df -h` says `/dev/loop0 49M 49M 0 100% /home/aiena/BUILD-AREA/LINUX/qemu` | 19:46 |
| aiena | so the loop device is full or nothing can be added | 19:46 |
| aiena | nothing to do with the block devices storage | 19:46 |
| aiena | any idea how to fix this .... so much fun :) | 19:48 |
| aiena | CarlFK: I was looking at this https://www.thegeekdiary.com/how-to-create-virtual-block-device-loop-device-filesystem-in-linux/ | 19:52 |
| tpb | Title: How to create virtual block device (loop device/filesystem) in Linux – The Geek Diary (at www.thegeekdiary.com) | 19:52 |
| aiena | what is losetup ? | 19:52 |
| aiena | seems like he makes the file system etc without mounting it | 19:52 |
| aiena | i am giessing I need to resize the debian partition | 19:52 |
| aiena | then remoount it and install into it i.e. make the debian image larger with free space | 19:52 |
| CarlFK | well.. I fond what I was looking for, but it wont help you: https://salsa.debian.org/debconf-video-team/ansible/blob/master/usbinst/mk_usb_installer.sh#L113 | 20:00 |
| tpb | Title: usbinst/mk_usb_installer.sh · master · DebConf video team / ansible · GitLab (at salsa.debian.org) | 20:00 |
| CarlFK | fatresize -p -v -s 1G ${dev} | 20:00 |
| CarlFK | we need some other fs resize | 20:00 |
| aiena | CarlFK: what do you think of this https://askubuntu.com/questions/260620/resize-dev-loop0-and-increase-space psuki's solution? | 20:04 |
| tpb | Title: partitioning - resize /dev/loop0 and increase space - Ask Ubuntu (at askubuntu.com) | 20:04 |
| aiena | *psusi | 20:05 |
| aiena | anyway goodnight CarlFK I need to sleep I can try to pursue this on the morrow lots of hurdles lol | 20:08 |
| aiena | thank you so much for your time | 20:08 |
| *** aiena has quit IRC | 20:09 | |
| *** rohitksingh has quit IRC | 22:00 | |
| *** CarlFK has quit IRC | 22:53 | |
| *** thaytan has quit IRC | 23:17 | |
| *** thaytan has joined #timvideos | 23:17 | |
| *** ChanServ sets mode: +v thaytan | 23:17 | |
| *** sc00bz has quit IRC | 23:47 | |
| *** sc00bz1 has joined #timvideos | 23:49 | |
| *** CarlFK has joined #timvideos | 23:56 | |
| *** ChanServ sets mode: +v CarlFK | 23:56 | |
Generated by irclog2html.py 2.13.1 by Marius Gedminas - find it at mg.pov.lt!
