**** BEGIN LOGGING AT Sun Mar 11 03:00:03 2018 Mar 11 03:47:26 Has anyone else run into problems with sensors not working after reinstalling Android? My accelerometer, proximity sensor, and finger print sensor hasn't been working on my axon since reinstalling. Mar 11 03:50:33 Check your persist partition, it may have been wiped out Mar 11 03:55:48 persist isn't empty. Mar 11 03:56:56 @wickman72, [Edit] You're probably getting a bit ahead of yourself. The stuff compiling doesn't mean it's going to work. It's time to start debugging your halium boot. Mar 11 04:02:14 @bhushanshah I'm not sure what to be checking for in there. Mar 11 04:06:48 There's generally sensor calibration data in that partition Mar 11 04:06:53 Is it present? Mar 11 04:07:49 I don't think it's sensor calibration honestly, I downloaded an app to show the sensor data. It doesn't report _any_ values. Mar 11 04:07:51 (Photo, 720x1280) https://irc.ubports.com/AKBFk29o.png Mar 11 04:09:03 in sensors there's only sns.reg. No other files. Mar 11 04:09:05 Um maybe logcat might show some interesting information Mar 11 04:13:39 THere is some interesting stuff in there. Look at the proximity sensors stuff. "Sensor or listener is null" https://gist.github.com/rockybulwinkle/11e7deb8a09766805ae9b2dc582fbb7f Mar 11 13:36:50 @UniversalSuperBox, i can ,,.. what i have to do there Mar 11 13:40:23 Delete the connection that appears when you plug in your phone Mar 11 13:40:38 Also, replying to 2-month old messages is really confusing Mar 11 13:45:13 @redshaman, Be sure to stitch it all together for us. Mar 11 13:48:33 Or just contribute it to the upstream docs Mar 11 13:49:29 Especially since the complaint he had about the docs is actually not a problem at all. Breakfast is mentioned on that page: https://docs.halium.org/en/latest/porting/build-sources.html#for-halium-7-1 Mar 11 13:49:32 @UniversalSuperBox, deleted Mar 11 13:49:40 @abhishek_01, Now add your IP configuration Mar 11 13:52:40 @UniversalSuperBox, connection again got reset.. after deleting and then running the script again Mar 11 13:53:00 "Connection reset" is way better than "No route to host" Mar 11 13:53:09 Congrat, you have contacted your device. Mar 11 13:53:15 Are you running PM right now? Mar 11 13:53:30 yes Mar 11 13:53:53 Hmm, I'm not entirely sure what could be going on there. Sounds like you might be missing your SSH keys. Mar 11 13:54:00 Did you add the `.writable_image` file? Mar 11 13:54:29 no .. i just used the jbb's script Mar 11 13:54:48 https://github.com/Halium/projectmanagement/issues/40 Mar 11 13:55:44 @abhishek_01, It doesn't createe this file itself yet. Of course you can make a pull request :) Mar 11 13:55:48 @abhishek_01, [Edit] It doesn't create this file itself yet. Of course you can make a pull request :) Mar 11 14:28:15 so i followed all steps on #40 .. now i am gtting no route to host on PM Mar 11 15:18:39 halium-boot refused to work for me with plasma Mar 11 15:19:08 What happened? Mar 11 15:19:11 if i understood what you're talking about correctly Mar 11 15:19:18 Because @abhishek_01 is having the same problem Mar 11 15:40:01 heya, im trying to put the halium rootfs on using halium-install but i keep getting asked to install android-tools-fsutils and android-tools-adb and android-tools-fsutils, these packages arent in arch. Mar 11 15:45:12 @wickman72, try to get them from aur Mar 11 15:45:18 Maybe you can find them on aur. This are the ubuntu package names, so the package couls also just be called "adb" and "simg2img" Mar 11 15:45:28 [Edit] Maybe you can find them on aur. This are the ubuntu package names, so the package could also just be called "adb" and "simg2img" Mar 11 15:47:09 Most of ubuntu packages are converted for Arch and available from aur. If you are a newbie and prefer graphical things to terminal then install pamac and search from there Mar 11 15:47:40 @wickman72, they are Mar 11 15:47:46 You can always search in here https://aur.archlinux.org/packages/ download packages from this website and install locally Mar 11 15:47:50 so i already have simg2img installed apparently Mar 11 15:48:01 adb is installed Mar 11 15:48:40 fsutils ? Mar 11 15:49:01 what you mean? Mar 11 15:49:32 @wickman72, are you sure you installed halium-devel from aur ? Mar 11 15:49:41 yep Mar 11 15:50:03 if i try install android-tools it conflicts with bootimg-tools-git Mar 11 15:50:13 and that is part of halium-devel Mar 11 15:50:27 so i cant install android-tools Mar 11 15:50:40 unless i remove halium-devel Mar 11 15:51:00 error: file owned by 'android-tools' and 'bootimg-tools-git': 'usr/bin/mkbootimg' Mar 11 15:52:22 remove mkbootimg Mar 11 15:52:32 it's for decompiling kernels Mar 11 15:52:45 you can always install it later Mar 11 15:54:38 i installed android-tools Mar 11 15:54:43 still same error in halium script Mar 11 15:56:27 @wickman72, paste bin the error Mar 11 15:56:32 full Mar 11 15:58:15 https://pastebin.com/VXRQxP15 Mar 11 15:58:32 same error ragardless of venv btw Mar 11 15:59:26 it appears im missing make_ext4fs Mar 11 16:00:05 @wickman72, https://github.com/JBBgameich/halium-install use this script to install rootfs .. how to use is in readme Mar 11 16:54:54 am getting this error Mar 11 16:54:56 https://pastebin.com/1fUFrwQu Mar 11 17:01:41 do i need a toolchain or something? since I'm guessing it's because im on x86_64 and phone is arm64? Mar 11 17:03:48 oh ... you need qemu userspace emulation, but the script should have warned you if it's missing Mar 11 17:04:27 ```update-binfmts: warning: qemu-arm not in database of installed binary formats. … update-binfmts: exiting due to previous errors … update-binfmts: warning: qemu-arm not in database of installed binary formats.``` Mar 11 17:05:03 sooo Mar 11 17:05:41 Any arch linux user around who has already done this? Mar 11 17:06:01 I found this package on aur, no clue if it's the correct way to solve that https://aur.archlinux.org/packages/qemu-arm-static/ Mar 11 17:06:21 yep trying it Mar 11 17:08:56 already have qemu-user-static installed and qemu-arm-static Mar 11 17:09:40 nothing changes :/ Mar 11 17:11:33 JBBgameich: https://github.com/Halium/halium-devices/pull/63 and https://github.com/Halium/halium-devices/pull/64 :) Mar 11 17:14:31 Why two PRs? btw thank you for the Pr, it really makes initial building easier Mar 11 17:14:48 one for 5.1 and one for 7.1 Mar 11 17:15:04 Oh right, missed that :) Mar 11 17:17:43 JBBgameich: Yeah to get the initial builds to work this is the minimal usually ;) Mar 11 17:17:52 And for newcombers it can be quite tedious to figure out Mar 11 17:20:44 JBBgameich: I guess another general one might be something that would sort stuff like: https://github.com/Halium/android_device_htc_pme/commit/0df96e78aca29e287cb57436a0c8ed320b1de52c Mar 11 17:45:54 It seems like we have an issue with our system.img for mido: the produced system.img is a sparse img, not the usual ext4 fs image Mar 11 17:46:18 Therefore halium-boot doesn't know how to mount it to /system Mar 11 17:46:40 Is that a known problem on other devices ? Mar 11 17:46:44 You can just mount a sparse image? Mar 11 17:47:12 no, seems we can't, and that's the issue Mar 11 17:47:20 Our halium-install scripts convert the sparse image to a normal ext4 image Mar 11 17:47:34 oooh :) that's why. Mar 11 17:48:01 All the images we had before (tenderloin, mako, hammerhead) were already ext4 images Mar 11 17:48:24 Ok, we'll take that into account, no pb Mar 11 17:49:44 ok i might cry in a second, i cant find a way to get qemu-system-arm on my arch box Mar 11 17:50:28 you could download the debian static version and unpack it into the system manually Mar 11 17:51:18 you mean https://packages.debian.org/stretch/qemu-system-arm? Mar 11 17:51:54 qemu-arm-static Mar 11 17:53:05 error: failed to commit transaction (conflicting files) … qemu-arm-static: /usr/bin/qemu-aarch64-static exists in filesystem … qemu-arm-static: /usr/bin/qemu-arm-static exists in filesystem … qemu-arm-static: /usr/bin/qemu-armeb-static exists in filesystem Mar 11 17:55:28 @wickman72, i created that halium-devel package .. do let me know if you install any other package which was not in the haliu-devel .. so i can add it Mar 11 17:56:13 qemu-user-static conflictswith qemu-arm-static Mar 11 17:56:43 you need to add binfmt-support like you mentioned yesterday Mar 11 17:57:14 @wickman72, obvously it will .. if you have qemu and qemu-user-static installed .. you dont need to explicitly install arm-static Mar 11 17:57:23 whattttt Mar 11 17:57:51 wait so why isnt the halium installer working and giving chroot errors? Mar 11 17:58:11 what errors ? pastebin Mar 11 17:58:12 ? Mar 11 17:58:32 https://pastebin.com/1fUFrwQu Mar 11 17:59:39 can you do … ```adb shell``` Mar 11 17:59:43 ? Mar 11 17:59:45 yep Mar 11 17:59:51 no issues with adb for sure Mar 11 18:00:03 i sideloaded my j5 this moring with it Mar 11 18:01:05 do you have binfmt-support installed ? Mar 11 18:01:10 yep Mar 11 18:02:42 what other packages you installed after halium ? Mar 11 18:03:10 qemu-user-static Mar 11 18:03:22 no thats included so ignore that Mar 11 18:04:01 binfmt-support Mar 11 18:04:11 and Mar 11 18:04:50 debtap Mar 11 18:04:53 e2fsprogs Mar 11 18:07:03 thats all Mar 11 18:07:15 debtap is unrelated to this tho Mar 11 18:07:22 was used to help with binfmt-support Mar 11 18:11:17 these errors should not cause any issues Mar 11 18:11:30 [Edit] these packages should not cause any issues Mar 11 18:22:14 eh imma make a ubuntu kvm and see how that goes Mar 11 18:22:30 in fact no ill just use an old laptop passthrough is longgg Mar 11 19:14:14 works fine in ubuntu Mar 11 19:14:27 just had to install android-sdk as well for adb support Mar 11 19:21:49 right, ive gotten aosp master working, how do i go about porting hallium Mar 11 19:22:21 Start with reading and following the porting guide: http://docs.halium.org/en/latest/ Mar 11 19:28:04 righty ho, booted up the phone Mar 11 19:28:04 https://pastebin.com/m7j3xxGH Mar 11 19:29:08 ubptgbot: uhh Mar 11 19:29:15 [ 4.302990] WARNING: at /home/user/Projects/halium/kernel/xiaomi/msm8956/drivers/gpio/gpiolib.c:126 gpio_to_desc+0x2c/0x50() … [ 4.314168] invalid GPIO -517 Mar 11 19:29:15 my trees dont work with 7.1 Mar 11 19:29:20 i doubt they ever will Mar 11 19:29:58 ubptgbot: thats defered probe Mar 11 19:30:05 thats fine Mar 11 19:30:47 how do i go about finding whats up in the last_kmsg? Mar 11 19:32:03 https://pastebin.com/m7j3xxGH wait did i just simply end the boot too soon? Mar 11 19:32:11 i see no kernel panic Mar 11 19:36:28 @JBBgameich actually Mar 11 19:43:19 is here any oreo hallium at all Mar 11 19:43:36 it dont look like it Mar 11 19:43:55 i had to build from 7.1 sources (cm14.1) Mar 11 19:44:31 yeah, guess ill have to wait till hallium has oreo then Mar 11 19:44:54 why do you need oreo? Mar 11 19:45:11 you using a really new phone? Mar 11 19:54:52 nope Mar 11 19:54:54 z2 tablet 2014 Mar 11 19:55:06 but im using the mainline kernel with freedreno Mar 11 19:55:33 which needs oreo to work properly Mar 11 19:56:39 ah okay so from my pov you tryna do a 100% open src? Mar 11 19:56:40 it also means that hallium is pointless as i have proper linux drivers but its needed for ubports soo Mar 11 19:56:50 yep Mar 11 19:57:04 nice, would love to do that at some point Mar 11 19:57:15 am gonna grab the librem 5 when it comes out Mar 11 19:57:21 http://sendvid.com/63l53wvo heres a video Mar 11 19:57:35 wish librem5 was qcom Mar 11 19:58:13 why? Mar 11 19:58:56 qcom is like Mar 11 19:59:03 ... Mar 11 19:59:06 i dont know Mar 11 19:59:11 i just really like qcom Mar 11 19:59:19 eh Mar 11 19:59:27 and you have rob for freedreno Mar 11 19:59:33 whos an amazing dev Mar 11 19:59:37 cool Mar 11 19:59:44 i kinda see qcom as a massive monopoly tbh from one pov Mar 11 19:59:53 and 820 has really good support upstream Mar 11 20:00:01 but like its popularity kinda makes it well known among devs Mar 11 20:00:32 yeah Mar 11 20:00:46 i mean, i dislike qcom as a company Mar 11 20:01:12 i mean is there really an alternative Mar 11 20:01:20 but i really like what they do with mainline Mar 11 20:01:28 like mediatek is just insecure at chip level Mar 11 20:01:59 xiaomi huawei and samsung can make chips in house but are 99% of the time hush hush on their sources Mar 11 20:02:01 that can be a good thing Mar 11 20:02:21 like you can bypass any pin on a mediatek phone Mar 11 20:02:28 Yeah, if the chip is insecure by default we may be able to wipe its bad parts and replace them with our own Mar 11 20:02:29 i know someone whos porting an open source baseband to mtk Mar 11 20:02:41 and another guy whos doing the same with bootloader Mar 11 20:03:06 but if its at hardware level isnt that gonna be hard to implement (straying out of my knowledge here) Mar 11 20:03:32 yes but hes amazing at re Mar 11 20:03:46 My guess is a bug like a PIN bypass would be in their Trusted Execution Environment, not the chip itself Mar 11 20:04:26 McBitter (matrix) + unreasonable (irc) Mar 11 20:04:58 The TEE being another OS that runs on the same chip that's "Trusted". It can see the memory of the untrusted 'world', but the untrusted world can't see it. Mar 11 20:05:50 thats ring -3 rite in terms of access, so like Intel ME and AMD PSP level rite? correct me if im wrong Mar 11 20:06:14 Somewhere around there. Mar 11 20:06:43 Honestly I'm not too hip with the rings Mar 11 20:06:48 -3 is lowest really Mar 11 20:06:52 invisible to os Mar 11 20:06:53 though, is there any work being done on oreo Mar 11 20:07:17 At the moment, we're waiting on libhybris support for Oreo Mar 11 20:07:46 For context, the TEE commonly verifies the boot image for Fastboot. So... it's lower level than that. Mar 11 20:07:53 not getting adb output from my phone anymore? Mar 11 20:08:29 @UniversalSuperBox, hmm, im pretty sure thats ring -3 Mar 11 20:08:45 yea it is Mar 11 20:09:39 Okay, looks like ARM processors don't have rings. They do have Privilege modes though Mar 11 20:10:07 ah Mar 11 20:10:49 im personally waiting for the day an x86 CPU comes out with no TEE Mar 11 20:10:55 thats modern that is Mar 11 20:11:04 and x86_64 to be precise Mar 11 20:11:17 like VIA make some Mar 11 20:11:27 but they all like intel atom levels of performance Mar 11 20:11:44 Wait 3 years Mar 11 20:11:49 Well, even the i.MX8 that Purism is planning to use has TrustZone embedded. Mar 11 20:11:58 Honestly I kinda hope they do something with it. Software TPM? Mar 11 20:12:04 They will get support from chinese companies and get to actual modern production process)) Mar 11 20:12:14 isnt that open source though? Mar 11 20:12:14 Fwd from UniversalSuperBox: Well, even the i.MX8 that Purism is planning to use has TrustZone embedded. Mar 11 20:12:21 TrustZone is just another processor mode. Mar 11 20:12:26 It's what you run in it that matters Mar 11 20:12:37 the modem is isolated in librem 5 Mar 11 20:12:51 That's the plan, yes. Mar 11 20:13:05 the trustzone is open source though no? Mar 11 20:13:16 libreboot if i remember Mar 11 20:13:19 It's part of the ARM specs for most designs: https://www.arm.com/products/security-on-arm/trustzone Mar 11 20:13:27 Most modern designs that is Mar 11 20:15:02 ah ok Mar 11 20:15:16 so its open source by arm themselves Mar 11 20:15:32 Right, but that doesn't mean anything if you don't have the code that runs in the trusted world Mar 11 20:15:33 and arms equivilant of x86 ring -3 is Exception Level 3 Mar 11 20:15:46 what do you mean? Mar 11 20:16:31 https://security.stackexchange.com/a/56221 Mar 11 20:16:36 This is a pretty good explanation Mar 11 20:16:53 reading now Mar 11 20:18:37 ah okay, kinda get that Mar 11 20:18:56 You need to actually have code to put in the trusted world for it to mean anything Mar 11 20:19:34 That's kinda what I was talking about, I'm hoping that Purism would put something there like a firmware TPM... or just put a TPM in the phone. Unfortunately I don't have high hopes for them having enough time for that. Mar 11 20:19:57 tpms are insecure fundamentally on a hardware level Mar 11 20:20:25 on a virtual level, perfect Mar 11 20:21:05 Well, if someone has taken apart your phone it's kinda violated your trust anyway Mar 11 20:21:19 true true Mar 11 20:21:31 if they put in a TPM, embedded in the TEE, you got a winner, theres no way you could get anything off that pretty much Mar 11 20:22:43 So... you work with this kind of stuff often? Mar 11 20:23:15 nope, im just really in the know with this kinda stuff Mar 11 20:23:51 i only got into linux around february as my main os Mar 11 20:24:27 and then my evenings from there got taken over Mar 11 20:24:33 linux on all the things! Mar 11 20:25:32 Interesting. I ask because I've been recently thinking a lot about how we can protect user trust and data under Halium Mar 11 20:25:45 (Photo, 604x520) https://irc.ubports.com/S00aSUWX.png Mar 11 20:26:00 luks? Mar 11 20:26:13 Well, that protects data Mar 11 20:26:25 But what happens when someone flashes a kernel with a keylogger Mar 11 20:26:43 over an existing one? Mar 11 20:26:45 or first time? Mar 11 20:26:54 Over the existing one. Mar 11 20:26:59 We can lock the bootloader for that Mar 11 20:27:01 but Mar 11 20:27:15 The rootfs is booted blindly at that point Mar 11 20:27:25 lokcing bootloader is a nogo on a scarily large chunk of devices Mar 11 20:27:35 like if you dont run miui on a xiaomi phone Mar 11 20:27:38 and lock it Mar 11 20:27:44 it refuses to boot Mar 11 20:28:16 Did you sign the boot image? Mar 11 20:28:30 no, am i have no idea how Mar 11 20:29:49 Well, if the bootloader can't be locked it's all a lost cause. Mar 11 20:29:56 So I suppose there's no use worrying on those devices. Mar 11 20:30:14 But some devices do allow you to lock the bootloader, as specified in the fastboot documentation Mar 11 20:30:20 well yes Mar 11 20:30:22 why not generate an md5 of kernel Mar 11 20:30:28 and do a boot check Mar 11 20:30:46 in a similar manner to how grub handles luks Mar 11 20:31:09 That requires something running before the kernel Mar 11 20:31:23 what if you have a second bootloader Mar 11 20:31:24 @wickman72, Linux on my brain so i can rm -rf bad memories! Mar 11 20:31:35 I've been doing mental gymnastics on this for a week or so now. :P Mar 11 20:31:43 like how they did on the … otter2 (kindle fire 2nd gen) Mar 11 20:32:02 @VeryOriginalUsername, ahh i wish.... Mar 11 20:33:13 @wickman72, Not terribly interested in reverse engineering anything Mar 11 20:33:21 I think we can safely say "no" to a device with no bootloader lock. Mar 11 20:33:32 relock, that is Mar 11 20:33:45 okay Mar 11 20:34:03 It'll still run fine, but it just can't be a "Trusted Platform" Mar 11 20:34:16 yep Mar 11 20:34:39 what if you hid the OS (probably massive job and would be in a fork) Mar 11 20:34:46 a bit like veracrypt does Mar 11 20:35:00 and exposed a dummy os Mar 11 20:35:15 would make debugging a bitch Mar 11 20:35:15 That doesn't really solve the problem Mar 11 20:36:04 Veracrypt does that to give plausible deniability that you were running another OS Mar 11 20:36:15 It doesn't provide trust for your software Mar 11 20:36:37 im saying as a means of deterrent, not prevention Mar 11 20:37:09 No, you're still dependent on your boot image executing the rest of the system Mar 11 20:39:09 android phones dont have secure boot rite? Mar 11 20:39:45 if they did or if you could get something of that nature, you could verify your kernel Mar 11 20:39:52 and refuse to boot Mar 11 20:40:21 That's what a locked bootloader is Mar 11 20:41:25 hmmph Mar 11 20:41:30 im stumped then Mar 11 20:41:31 :P Mar 11 20:42:52 So yeah, locked bootloader is precisely what we need in that case. An attempt to unlock the bootloader destroys all data on the `/data` partition Mar 11 20:43:05 By way of whatever secure erase is supported on the flash, assuming it was implemented correctly Mar 11 20:45:46 That much I'm reasonably sure of. ;) Mar 11 20:46:23 The question is... how do you make sure the rest of the system is protected? :) Mar 11 20:47:14 One way is to encrypt the `/data` partition and implement a framebuffer GUI in the boot image. Mar 11 20:47:25 That's what PostMarketOS did. Mar 11 20:47:43 agree with ya Mar 11 20:49:51 also Mar 11 20:50:16 the os enables debugging at boot Mar 11 20:50:16 whilst useful Mar 11 20:50:19 can be a risk Mar 11 20:50:30 Well yeah, of course that's fixable. Mar 11 20:50:45 yup Mar 11 20:51:30 I'd also ideally like to provide integrity validation of the rootfs. Encryption alone doesn't do that Mar 11 20:51:52 Unless you're using a VERY new kernel and enable the appropriate luks target Mar 11 20:51:59 how new? Mar 11 20:52:15 3.18 is what halium uses rite? Mar 11 20:52:25 Whatever the device came with Mar 11 20:52:32 ahk Mar 11 20:52:33 Let's see... first available in 4.12 Mar 11 20:52:49 yea thats really new then Mar 11 20:52:52 That's probably four years away on Android devices. Mar 11 20:52:58 :/ Mar 11 20:53:13 Given that 4.9 was just named the new "Blessed 5-year release" Mar 11 20:53:19 oh Mar 11 20:53:27 latest is 4.16 on mainline... Mar 11 20:53:59 And it takes about two years (Qualcomm numbers) to bring up an Android platform on a given kernel Mar 11 20:54:18 So my calculation is... this year's into next year's devices are 4.4. Mar 11 20:54:27 4.9 comes out after that Mar 11 20:54:34 And whatever kernel is blessed next after that Mar 11 20:55:02 Er, anyway... `dm-integrity` is not an option Mar 11 20:56:15 We might get kernels starting with 5 this year Mar 11 20:56:22 damn... Mar 11 20:56:29 time flies Mar 11 20:56:38 2.6 times xd Mar 11 20:56:43 lol my router still runs 2.6 kernel Mar 11 20:56:43 Well once we broke out of 2.6 things started flying Mar 11 20:56:44 The newest Mediatek chipsets are still 3.18 Mar 11 20:56:56 lol everyone comments about 2.6 Mar 11 20:57:17 in a nice way Mar 11 20:57:20 it needs to die Mar 11 20:57:20 Well... 😂 Mar 11 20:57:23 Uh Mar 11 20:57:29 UniversalSuperBox Mar 11 20:57:42 It took me 2 weeks Mar 11 20:57:42 Hi, am I wrong about something? I'd love to be Mar 11 20:57:44 (Photo, 500x355) https://irc.ubports.com/v74yOOAM.png Mar 11 20:59:45 how much effort to make a script to auto compile latest kernel for a device, given the right project files? Mar 11 21:00:09 That's ignoring the proprietary drivers that are included and compiled for a specific kernel version Mar 11 21:00:14 Bringing up android doesn't take too long Mar 11 21:00:27 Yeah Mar 11 21:00:39 That's what takes time Mar 11 21:00:54 I'm using freedreno Mar 11 21:01:09 Which is the culmination of tons of work Mar 11 21:01:44 Which always tracks latest kernel Mar 11 21:01:44 Yeah Mar 11 21:02:04 say this was done for qcom Mar 11 21:02:09 at its most basic Mar 11 21:02:33 Here's the devices supported by freedreno: https://github.com/freedreno/freedreno/wiki#devices Mar 11 21:02:59 Nope Mar 11 21:03:33 Freedreno supports all a3xx-5xx chips Mar 11 21:03:46 universal open source driver set? now that would be cool Mar 11 21:03:53 That's really out of date Mar 11 21:04:04 (almost oldest-newest) Mar 11 21:05:53 With 3D acceleration? Mar 11 21:06:08 Is there any documentation on how it comes together then? Mar 11 21:07:37 *heavy sigh as androidfilehost gives 503 errors* Mar 11 21:07:39 Yep Mar 11 21:07:45 What do you mean? Mar 11 21:08:11 Documentation on the project. What it supports, how to do things with it Mar 11 21:11:53 Should work out of the box if you are using DRM/msm in kernel Mar 11 21:11:55 Status page on wiki also Mar 11 21:20:23 @UniversalSuperBox, No mt6750 some devices has 4.4.x Mar 11 21:21:38 but %90 uses 3.18-3.10-3.4 Mar 11 21:22:44 Ah yep, my mistake. Confused new devices with new chipsets. Mar 11 22:59:18 * lambdaperl[m] sent a long message: lambdaperl[m]_2018-03-11_22:59:18.txt Mar 11 22:59:39 Sorry for long post; TL;DR N00b, Want to port Halium, how dangerous should I expect it to be Mar 11 23:00:16 in my experience it's very hard to brick a device. Mar 11 23:01:54 there's usually some DFU, fastboot, or something you can fall back on if the software gets fubar'd Mar 11 23:02:42 not to say I haven't had scary moments when I *thought* Mar 11 23:02:49 I'd bricked it. Mar 11 23:05:08 that's encouraging Mar 11 23:05:09 I don't have to wipe out fastboot right? Mar 11 23:05:10 and interesting to hear of DFU, I heard about that doing microcontroller development recently, that's some USB protocol from what I can tell, right? Mar 11 23:07:42 have you had moments when fastboot hasn't been accessible to you then? Mar 11 23:07:58 I'm refering broadly to any hardware based recovery. For example my ZTE Axon 7, I used MiFlash to flash the right version of the firmware when I was unlocking it. I'm near 100% certain that that software would have worked even if fastboot wasn't. Mar 11 23:08:20 Yes, fastboot wasn't accessible before I unlocked the firmware on my phone. Or at least not a *useful* fastboot. Mar 11 23:08:58 Ah, I already have unlocked bootloader here (using LineageOS), you reckon I should probably be fine as long as I'm not too stupid then? Mar 11 23:09:37 I don't believe there should ever be a reason to wipe out fastboot. I'm not totally sure if that's even possible? Mar 11 23:10:45 that should be fine then, as long as I can fastboot I can just boot TWRP Mar 11 23:10:55 yup Mar 11 23:11:03 cheers for your advice/expertise :-) Mar 11 23:11:27 https://forum.xda-developers.com/zenfone2/help/hardbrick-overwrite-fastboot-t3101290 Mar 11 23:11:38 Apparently it is possible. Don't flash to fastboot lol. Mar 11 23:12:27 lol, I shall endeavour not to Mar 11 23:22:13 I've found someone who hard-bricked his Moto G5 and there seems to be a Qualcomm utility to reflash the bootloader, so I shall back up my bootloader and it seems like there would still be a chance if everything goes awry Mar 11 23:23:06 rockybulwinkle: Thanks a lot for your help! Hopefully I will get somewhere with this port and I'll see you soon with progress Mar 11 23:23:40 cool. Mar 11 23:57:03 Is there a link to the LuneOS rootfs that I could try on my device? Mar 11 23:57:48 https://paste.ubuntu.com/p/YS6Rx4TDP5/ … Just taken, any ideas? Mar 11 23:58:51 Well, you aren't supposed to use f2fs as your filesystem Mar 11 23:59:07 You should be using ext3 or ext4 Mar 11 23:59:09 I know that, the images are ext4 Mar 11 23:59:32 Why is it trying to mount f2fs partitions then? Mar 12 00:00:01 Cause the ext4 seems to be corrupted or something and its a brand spanking new one Mar 12 00:00:40 But I tried jbbs script as well as the rootstock one, dunno what else I can really do Mar 12 00:08:05 What do you have to do to get the camera working in PM? Mar 12 02:52:04 @Qiangong2, Dunno, haven't played with pm yet, thinking a waving dead chicken and chanting will help lol **** ENDING LOGGING AT Mon Mar 12 03:00:01 2018