**** BEGIN LOGGING AT Sat Jul 23 02:59:56 2022 Jul 23 10:12:13 Morning! Jul 23 10:12:20 Herrie: did you also see this patch set ? https://lore.kernel.org/linux-arm-kernel/CACRpkday7mRWPfirjY+VcS-tB8CwqWQCyunkTUh0om9DXF4F3g@mail.gmail.com/ Jul 23 10:12:55 it's quite recent, probably not merged yet, but goes together with the graphics branch you pointed to earlier Jul 23 10:22:39 morning Jul 23 10:26:59 Herrie: also I found his github repo: https://github.com/lopsided98/linux/commits/hp-tenderloin Jul 23 10:50:24 Tofe: Yeah I saw that patchset, seems 1/3 patches need rework Jul 23 10:50:51 So I'm trying to see if Ben (lopsided98) can resend it: https://github.com/lopsided98/linux/commits/hp-tenderloin Jul 23 10:51:14 There's also this: https://github.com/mbd98/hp-touchpad-linux/commits/tenderloin-5.15 Jul 23 10:52:19 There's also this patch set still pending, so wrote to Linus as well ;) https://patches.linaro.org/project/linux-arm-msm/patch/20170419091326.11226-3-linus.walleij@linaro.org/ Jul 23 10:52:48 And there's this bit specifically for graphics as well.... https://git.kernel.org/pub/scm/linux/kernel/git/linusw/linux-integrator.git/commit/?h=apq8060-dragonboard-graphics&id=a0cdde9dc378768de25538e6fc9d7f80840e4fc2 Jul 23 10:53:38 Then we might need some of the pending patches that are listed here as well: https://dflund.se/~triad/krad/dragonboard/ Jul 23 10:54:44 I rebased some on top of 5.19 (mostly cleanly applied): https://github.com/Herrie82/linux/commits/herrie/tenderloin-5.19 Jul 23 10:54:56 Just struggling to get a defconfig generated with the supplied .mak file ;) Jul 23 10:55:05 https://github.com/Herrie82/linux/commit/6d76455e1dba2ef3be566326ce821abec0c709b2 Jul 23 10:57:01 And then there's probably some old patches from Rob Clark that might still be needed for Freedreno that were on the 3.0 branch: https://github.com/freedreno/kernel-msm/commits/hp-tenderloin-3.0 Jul 23 11:01:50 It would help if we would have the schematics of the Tenderloin, but I guess those aren't available anywhere.... Jul 23 11:03:13 The scripts/config takes config option to enable and seems to enable the rest of the config options that are dependencies as well (similar to menuconfig I guess, but automated). Just not really sure how to get this running correctly Jul 23 11:03:38 It keeps complaining about my CROSS_COMPILE even when trying inside of Yoto Jul 23 11:03:42 Yocto Jul 23 11:07:18 I could of course send a Tweet from @webosports requesting schematics, who knows what turns up ;) Jul 23 11:10:36 I mean something like this would be of great help: https://files.pine64.org/doc/PinePhonePro/PinephonePro-Schematic-V1.0-20211127.pdf Jul 23 11:12:40 There is this one for the TP Go (Opal): https://forums.webosnation.com/attachments/other-webos-tablets/69363-go-schematics-opal_schematic_inventec.png Jul 23 11:14:33 yes, and of course it's not the same screen... Jul 23 11:14:58 though the naming must be very close Jul 23 11:16:23 Tofe: Got 2 PDFs with schematics for Opal Jul 23 11:16:32 It's not 100% identical of course, but it will be of help Jul 23 11:16:37 Let me put them on the usual place Jul 23 11:17:39 Opal_6050A2430401-mb_dvt_0704.pdf Jul 23 11:18:09 And Opal 6050A2430401 rX01.pdf Jul 23 11:18:18 Those are quite some detailed schematics Jul 23 11:18:29 Now if we would have the same for tenderloin that would be of great help Jul 23 11:18:34 In creating the DTB Jul 23 11:18:54 got them thx Jul 23 11:21:01 They're very detailed, but from what I understood it helps with which registers etc to use in a DTB Jul 23 11:21:08 And it lists all the components Jul 23 12:15:38 Tofe: Sent them to lopsided98 as well, might spark his interest a bit ;) Jul 23 12:52:22 Tofe: You see anything in this that looks suspicious? This is latest PPP image I built: https://bpa.st/DQFA Jul 23 12:52:25 And doesn't boot to UI Jul 23 13:05:16 Tofe; OK was able to extract the defconfig from the 4.20 zImage: https://pastebin.ubuntu.com/p/T9FynQyFmN/ Jul 23 13:19:12 Herrie: you don't have the journal ? at least concerning surface-manager Jul 23 13:19:31 Tofe: No this is all I can get using the serial Jul 23 13:19:35 THere's nothing on display Jul 23 13:19:42 I'll revert to older kernel to see if that works Jul 23 13:21:11 you can't log in either ? Jul 23 13:21:17 Nope Jul 23 13:21:20 Well I need WiFi for that ;) Jul 23 13:21:36 ah ok, I could do that on PP using serial Jul 23 13:22:17 Well here not now Jul 23 13:22:19 It worked before Jul 23 13:22:32 For Tenderloin mainline it doesn't compile because it doesn't seem to like the gold linker Jul 23 13:22:36 Not really sure how to fix thta? Jul 23 13:23:21 Herrie: we probably have a fix in our commit history for tenderloin Jul 23 13:23:22 https://bpa.st/PZYA Jul 23 13:23:38 Tofe: Well not sure... Jul 23 13:23:51 But PP(P) should have the same issue with mainline kernel, no? Jul 23 13:25:36 probably; I remember talking about it a while ago, at least Jul 23 13:27:06 Dec 19 10:25:49 JaMa: for linux 5.4, I get "scripts/Kconfig.include:39: gold linker 'aarch64-webos-linux-ld' not supported"... you know something about that one? Jul 23 13:27:20 Dec 19 10:38:01 Tofe: we have gold as default ld (ld-is-gold DISTRO_FEATURE) but you should be able to easily disable that with -fuse-ld=bfd, but the check for ld itself might be an issue Jul 23 13:29:43 Let me try that Jul 23 13:32:28 would be worth comparing scripts/Kconfig.include between PP and tenderloin Jul 23 13:35:28 Yes will do that Jul 23 13:37:05 https://megous.com/git/linux/tree/scripts/ld-version.sh?h=orange-pi-5.13#n42 surprisingly this doesn't fail for PP, I wonder why Jul 23 13:37:54 I guess it's the include I use Jul 23 13:38:31 Anyway will get some sunshine with kids now and check later Jul 23 13:47:03 I guess I need to rework the recipe a bit Jul 23 13:49:24 Will do that in the evening Jul 23 13:50:17 Because I was using a slightly modified original one for android Jul 23 13:50:23 So that needs reworking anyway Jul 23 13:50:45 So will take PP(P) as basis and add the other bits in later for the various manipulations Jul 23 13:51:40 looks like a plan Jul 23 13:52:03 I'll go for a walk too, see you later ! Jul 23 13:52:24 Because kernel building should be identical, rest is done afterwards like the uImage etc Jul 23 15:47:57 Tofe: what year was that? :) Jul 23 16:05:42 normal kernel recipes set LD to use .bfd, compare the kernel recipes (or bitbake -e from them) to see why it's missing for this tenderloin kernel recipe Jul 23 16:06:47 https://git.openembedded.org/openembedded-core/commit/?id=66c1f4b0abd3483759f9e4141a80f982eaf8ccfd Jul 23 16:09:01 there is also https://git.openembedded.org/openembedded-core/commit/?id=36fe2efc91fec5a52374a2dc4814a69076a2d28e but I don't expect you to use linux-yocto machinery for tenderloin Jul 23 16:11:16 Tofe: https://github.com/flto/linux/wiki Jul 23 16:11:21 Also interesting Jul 23 16:11:31 JaMa: Well we do for PP(P) Jul 23 16:11:42 We can do the same for tenderloin really Jul 23 16:11:49 I'll toy a bit more tonight Jul 23 16:25:26 JaMa: webos-ports.20191220.txt :) Jul 23 16:26:43 Tofe: ok that predates kernel.bbclass fix a bit Jul 23 16:27:44 JaMa: I'll rewrite the kernel recipe for tenderloin mainline a bit so it's more aligned with the PP(P) ones and then will see. The manipulations afterwards anyway are a separate step Jul 23 16:28:17 I guess it might be worth to check hammerhead's status again too for mainline at some point Jul 23 16:28:23 It should be pretty OK I guess? Jul 23 16:31:03 This does seem to suggest it's pretty complete and that's with a 4.x kernel: https://github.com/flto/linux/wiki/hammerhead-upstream Jul 23 16:40:24 Ah for Hammerhead there are 5.16 patches: Jul 23 16:40:29 https://github.com/masneyb/nexus-5-upstream/tree/master/out-of-tree-patches/upstream-patches/v5.16 Jul 23 17:26:50 OK this kernel compile looks a bit better: https://pastebin.ubuntu.com/p/FMVM9R2HDM/ Jul 23 17:27:14 Now it's up tot the packaging of stuff it seems Jul 23 18:00:52 Oh, good find for hammerhead, I was wondering the same questions Jul 23 18:02:58 And I think the main remarkable issue was audio not really working Jul 23 18:03:09 the rest could be worked around I think Jul 23 18:03:43 I could revive my hammerhead-mainline branch and rebase it on top of kirkstone Jul 23 18:04:57 Tofe: Good idea I would say Jul 23 18:05:25 Maybe keep 2 machines for now so we can more easily test for now in paralel? Jul 23 18:06:12 Ie hammerhead-halium and hammerhead-mainline and both point to their own kernel? Jul 23 18:06:40 ah, well, good idea but not sure how much rework that implies Jul 23 18:07:09 let me investigate a bit Jul 23 18:08:34 Tofe: Well it should be pretty straightforward in most places I guess Jul 23 18:11:06 Should be mainly renaming files and folders in most places Jul 23 18:26:35 ok, basic rebase is done I think Jul 23 18:28:13 Tofe: OK did you push it already somewhere? Jul 23 18:29:02 not yet, one moment Jul 23 18:30:51 https://github.com/Tofee/meta-smartphone-1/tree/kirkstone/mainline-hammerhead part 1 Jul 23 18:31:36 https://github.com/Tofee/meta-webos-ports/tree/tofe/kirkstone/mainline-hammerhead part 2 Jul 23 18:32:02 so far it's just a basic rebasing, I didn't check if it builds or anything Jul 23 18:34:53 Tofe: Understandable Jul 23 18:34:59 Let me put kids to bed and check some stuff here Jul 23 18:35:21 I'm checking if bitbake is happy at least Jul 23 18:35:46 ah, there's a github URL to switch to htpps Jul 23 18:35:48 https Jul 23 18:36:16 that's the only parsing warning Jul 23 18:37:59 Tofe: We should be able to use 5.16 upstream kernel + these patches I guess? https://github.com/masneyb/nexus-5-upstream/tree/master/out-of-tree-patches/upstream-patches/v5.16 Jul 23 18:38:57 I guess so too yes Jul 23 18:40:18 oh, something familiar :) https://github.com/Tofee/meta-smartphone-1/blob/kirkstone/mainline-hammerhead/meta-lg/recipes-kernel/linux/linux-lg-hammerhead/reversed-disable-gold-linker.patch Jul 23 18:45:27 I'm building masneyb's 5.6 kernel just to be sure the current rebase works, then we can go further Jul 23 18:50:54 Tofe: Yeah good approach Jul 23 19:00:30 I'll try to build a new qemu image with waydroid Jul 23 19:00:46 Because that was one of the reasons we started this whole LSM migration to start with if I recall correctly LOL Jul 23 19:14:03 ahah, well yes :) Jul 23 19:18:19 Some cryptic python error messages... Jul 23 19:18:23 Let me try to debug that Jul 23 19:33:54 on my side I'm fixing some build issues, looks like the yocto-meta sidekick was missing Jul 23 19:40:07 Tofe; Yeah that one is new it seems... Jul 23 19:40:09 https://github.com/waydroid/gbinder-python/pull/2 Jul 23 19:40:17 Easy enough in general, but the error was cryptic Jul 23 19:44:41 ok, my kernel builds fine it seems Jul 23 19:44:47 I'll push this little fix Jul 23 19:44:48 :D Jul 23 19:46:55 Tofe; You want me to propose something for the hammerhead-halium and hammerhead-mainline machines? Jul 23 19:47:54 Well, if you're willing to :) Jul 23 19:48:49 Tofe: Well I can try Jul 23 19:48:57 It's probably not very pretty but will do the trick ;0 Jul 23 19:49:53 should we keep one as "hammerhead" ? Jul 23 19:50:07 maybe mainline ? Jul 23 19:51:22 to me it looks like the halium variant has a fragile future... Jul 23 19:52:09 (even though the mainline one is still functionnally incomplete) Jul 23 19:52:17 Tofe: Well yeah in general I agree Jul 23 19:52:30 I will just do hammerhead-halium and hammerhead for now then? Jul 23 19:52:40 So we can drop the halium one later when we are ready Jul 23 19:53:25 yes, looks good to me Jul 23 19:54:21 We shouldn't need luna-next-conf anymore right? Jul 23 19:55:08 correct Jul 23 19:55:23 it was in the old commits, I didn't remove it yet Jul 23 19:55:38 I added you in my repos, in case it makes things easier for you Jul 23 19:56:51 Well I just am doing a new branch in our meta-webos-ports Jul 23 19:56:54 Can PR it from there Jul 23 19:56:59 ok alright Jul 23 20:05:16 https://github.com/webOS-ports/meta-webos-ports/commits/herrie/hammerhead-split Jul 23 20:05:40 I noticed you added this patch, but it's not in the SRC_URI: https://github.com/Tofee/meta-webos-ports/commit/766af4b3dc15d7b7ce908489873cccfcf5eb5d01#diff-edd5522962a825f2ea09ed4eb254253eda2ab6ae8a54c47a513d8b200dfada8e Jul 23 20:05:44 Do we still need that one? Jul 23 20:07:44 well, it could be I was still trying to make it work Jul 23 20:07:59 this stuff is still WIP after all Jul 23 20:37:28 OK let me add it back then just Jul 23 20:39:08 no hurry :) Jul 23 20:45:37 Was on the phone, doing meta-smartphone now as well Jul 23 20:45:42 Then can do a test build Jul 23 21:04:46 https://github.com/Herrie82/meta-smartphone/commits/herrie/hammerhead-split Jul 23 21:05:57 Let me do a build or at least try Jul 23 21:09:41 Building Jul 23 21:09:43 Let's see Jul 23 21:11:11 It will be a while Jul 23 21:11:23 2400/8300 packages Jul 23 21:14:11 Tofe: You have the hammerhead firmware package laying around somewhere? Jul 23 21:14:32 hammerhead-firmware.tar.gz Jul 23 21:16:35 It's trying to fetch it from build.webos-ports.org which is no good at the moment Jul 23 21:35:20 Let me try to build the rest of the image without it at least Jul 23 22:24:32 Seems pretty OK so far, busy compiling nodejs, webruntime & qtbase Jul 23 22:24:38 So I guess it will be a little while still :P Jul 23 22:41:21 Herrie: let me look for the firmware; not sure though Jul 23 22:41:45 Tofe: Well I guess we can find it somewhere anyway Jul 23 22:42:02 For now I removed it from the machine.conf to let build continue Jul 23 22:46:31 we can probably rebuild it from https://gitlab.com/postmarketOS/pmaports/-/blob/master/device/testing/firmware-lg-hammerhead/APKBUILD Jul 23 22:47:14 Yes Jul 23 22:51:53 Seems webruntime is at 65%, rest is almost done Jul 23 22:52:24 8262 of 8329 Jul 23 22:52:34 ok, looks good then Jul 23 22:55:00 Herrie: I retrieved it from milla Jul 23 22:55:29 Tofe: Ah nice Jul 23 22:58:26 Herrie: https://github.com/Tofee/mainline-kernel-firmware Jul 23 22:58:52 it'll do for now Jul 23 23:01:38 Yes thnx Jul 23 23:05:29 time for sleep :) gn8 ! Jul 23 23:07:26 Same here Jul 23 23:07:31 Expect an image in the morning :P Jul 23 23:07:43 I can try on my hammerhead too Jul 24 00:01:17 Tofe: luneos-dev-package-hammerhead-mainline-20220724.zip is there for you **** ENDING LOGGING AT Sun Jul 24 02:59:57 2022