Magisk General Support / Discussion

[ctcx]

PixelFlasher 5 released, hooray.

Now, given this, my previous attempt, and your notes:

Some cursory comments.

MAGISK_APK=$(pm path com.topjohnwu.magisk)   # If hidden, replace with package name

Keep in mind that, if Magisk is hidden, even if you have the correct package name, you only have the stub, which is not enough to create a patch.

../busybox unzip -o ../pf.zip

PF copies this to /data/local/tmp, you're expecting it to be there, but I don't see where is the copying part.
Not all systems have busybox, so you have to account for it.


If you're on x86_64, then you need to copy x86 version, that's one of the checks that PF does.
The other check is for this special Magisk by @Namelesswonder which provides 32 bit functionality to 64 bit only phones, you still want to copy the 32 bit Magisk if the Magisk in use is Magisk Zygote64_32, if you don't care about those, then your check is enough.

I find that this is enough to check for SYSTEM_ROOT

SYSTEM_ROOT=false
grep ' / ' /proc/mounts | grep -qv 'rootfs' && SYSTEM_ROOT=true

For Pixel devices, I never needed to bother for Recovery mode, did you do any testing on a device with recovery mode? I'm curious about your results.

SHA1=$(./magiskboot sha1 "$1" 2> /dev/null)
cp -f new-boot.img /sdcard/Download/magisk_patched_$SHA1.img

You want to add 40 character SHA1 to the filename?
just add | cut -c-8

Do these notes still apply for everything I should fix in that first attempt of wrapper I did, or should I now fix more beyond that?

Thanks again.

 

[badabing2003]

Do these notes still apply for everything I should fix in that first attempt of wrapper I did, or should I now fix more beyond that?

Mostly yes, let me try to answer them individually.

MAGISK_APK=$(pm path com.topjohnwu.magisk) # If hidden, replace with package name

Keep in mind that, if Magisk is hidden, even if you have the correct package name, you only have the stub, which is not enough to create a patch.

Still the case, even though the stub might have what we need in data portion, it's not accessible without root, and even if we're rooted, properly extracting it would require some sleuthing / reverse engineering.
Besides unless the app is newer (or Magisk delta) root option is always a preferred option.
(I just discovered yesterday that Magisk delta does not have the stub.apk in /data/adb/magisk)

../busybox unzip -o ../pf.zip

PF copies this to /data/local/tmp, you're expecting it to be there, but I don't see where is the copying part.
Not all systems have busybox, so you have to account for it.

Yep, you have to make sure it exists somehow.
Some do, and some don't, specially a phone that has never been rooted before.

SYSTEM_ROOT=false
grep ' / ' /proc/mounts | grep -qv 'rootfs' && SYSTEM_ROOT=true

For me this is still working and good enough

For Pixel devices, I never needed to bother for Recovery mode, did you do any testing on a device with recovery mode? I'm curious about your results.

I added an checkbox option in PF where the user can select if they are patching for recovery or not.
I think for PF that should be more than enough, I doubt that it would be used.

You want to add 40 character SHA1 to the filename?
just add | cut -c-8

PF does that, the generated patch is named something like this. magisk_patched_25210_bac0e555_e3c084bc.img[/icode Magisk version, stock sha1, patch sha1

 

[ipdev]

No issues here on the OnePlus 8 considering that I am still using the Alpha fork.

For what it's worth, I didn't have this problem using Canary 26102 on my Pixel 2 XL (Android 11) and on my Pixel 6 Pro (Android 14 Beta 2). 26102 shows as installed on both devices.

Haven't tried on my Hammerhead, but if I can find it maybe I will

Thank you for checking.
Seems the issue was fixed earlier today and will be in the next canary release.

Assume you've checked for previous hidden Magisk App in device settings, Apps... (It is actually possible to have multiple stub apps installed!)... Also, nothing in a second space, workspace etc?... And checked magisk -v? PW

App was never hidden (renamed) and no second version installed.
No other workspace or such setup on device.

Both the `-v` and `-V` options returned.
No daemon is currently running!

The daemon was started according to the magisk log.

---

Looks like it was related to:
Root is not defined
Issue #6976

All good now with current commits.

hammerhead:/ $ magisk -c
dd93556ad-ip:MAGISK:D (26102)
hammerhead:/ $ magisk -v
dd93556ad-ip:MAGISK:D
hammerhead:/ $ magisk -V
26102
hammerhead:/ $

Cheers all.

 

[pndwal]

...
Still the case, even though the stub might have what we need in data portion, it's not accessible without root, and even if we're rooted, properly extracting it would require some sleuthing / reverse engineering.
Besides unless the app is newer (or Magisk delta) root option is always a preferred option.
(I just discovered yesterday that Magisk delta does not have the stub.apk in /data/adb/magisk)

Not Delta specifically...

This commit:
Remove unexpected files
vvb2060 authored and topjohnwu committed on Feb 26 removed stub.apk from /data/adb/magisk prior to 25207...

this broke boot_patch.sh as it's required per the header, and consequently broke addon.d script...

@osm0sis discovered the issue and stub.apk was restored to /data/adb/magisk in 26100 with:
Fix scripts …
manager.sh + boot_patch.sh:
- all listed files from boot_patch.sh header are required for boot patching, but stub.apk was being removed so install_magisk via addon.d.sh would fail without it; leave it in place
addon.d.sh:
- remove old redundant recovery_actions call (it's also performed by setup_flashable in initialize)
- print ABI to match flash_script.sh output
boot_patch.sh:
- catch and abort on any errors from ramdisk.cpio patching in the future
util_functions.sh:
- fix hiding of mount_partitions /system_root umount stderr
- quote mount_apex .pb DEST name parsing charset for safety even though both work
Fixes #6828
osm0sis authored and topjohnwu committed on Apr 9

... Delta is just getting old (not updated recently); Latest is still based on 25210. (Latest 'Stable' indicates base is 25206. : https://huskydg.github.io/magisk-files/intro.html ). So if your tester is using latest Delta lacking Stub is expected, and that was the same for official Magisk from 25207 till 26100 arrived...

PW

 

[badabing2003]

... Delta is just getting old (not updated recently); Latest is still based on 25210. (Latest 'Stable' indicates base is 25206. : https://huskydg.github.io/magisk-files/intro.html ). So if your tester is using latest Delta lacking Stub is expected, and that was the same for official Magisk from 25207 till 26100 arrived...

Thanks for the heads up,
Right I was aware about broken/missing stub.apk on version 25206-2521x, I don't know what I was thinking to expect anything different on Delta for version 25210

 

[pndwal]

...
Looks like it was related to:
Root is not defined
Issue #6976

All good now with current commits.

View attachment 5919061

hammerhead:/ $ magisk -c
dd93556ad-ip:MAGISK:D (26102)
hammerhead:/ $ magisk -v
dd93556ad-ip:MAGISK:D
hammerhead:/ $ magisk -V
26102
hammerhead:/ $

Cheers all.

Nice work by @pasha.mcr! ...

Model way to get Devs to re-open and respond. ... Thanks to @vvb2060...

... Just noticed you seem to have tried to install debug build on Nexus 5...

Because fix turned out to be speed related (Magisk too slow; Faster get magisk tmpfs path needed) and debug Magisk is known to cause issues with slower/older devices, perhaps that was a factor?... Anyway, faster Magisk seems a good fix.

PW

 

[dobi3]

Canary 25206 works fine... Think newer 26.1 and Canary 26102 have issues on my OnePlus 7 Pro running OOS.v11.0.9.1...get Qualcomm crash dump on newer versions... Canary 25206 works flawlessly...
This is what is causing it I believe:

- Pre-init storage partition device ID: metadata

I don't get that on Canary 25206

Thanks for any help or fix..

 

[spida_singh]

FYI husky has advised Delta has merged with 26.1 should be released soon. From TG channel.

 

[martyfender]

Has anyone else on Mediatek ARMv8-A devices had trouble getting the new Magisk update install the apk or update Magisk? See my previous posts.

 

[martyfender]

Has anyone else on Mediatek ARMv8-A devices had trouble getting the new Magisk update install the apk or update Magisk? See my previous posts.

What would be needed to troubleshoot this, logcats?

Thanks

 

[pndwal]

Has anyone else on Mediatek ARMv8-A devices had trouble getting the new Magisk update install the apk or update Magisk? See my previous posts.

Bit confused by previous posts... Do you have any Magisk App that opens? Please screenshot home screen... If not, what happens? PW

 

[martyfender]

Bit confused by previous posts... Do you have any Magisk App that opens? Please screenshot home screen... If not, what happens?

I have to leave for radiation therapy and work. However I had reverted to the action build I had installed dated about April 10th.

 

[DaOldMan]

In my ViperOS rom I now have a new problem with Magisk 6.x . Any app included in zygisk denylist crashes when run :-( . If I use the older Magisk 5.x eveything is totaly fine, I can even pass safetynet. Alpha version has the same problem.
Rom is Nugat version 7.1. Tried to remove Magisk completely then reinstall, still same problem.
Thanks guys for any help.

 

[osm0sis]

Whoa, magiskboot = payload_dumper now:

Support extracting any partition from payload.bin · topjohnwu/[email protected]

The Magic Mask for Android. Contribute to topjohnwu/Magisk development by creating an account on GitHub.

github.com

 

[tim1aust]

Question: I am new here, please redirect me to appropriate forum if this isn't where I should post my Q.

I have pixel phones rooted with Magisk, thus no OTA updates for me. So I download them from the Google site when the monthly update arrives, and flash using adb sideload. But it's a PITA wrt Magisk, I usually have to fuss around to get Magisk back running, although it's still worth it.

My Q is: if I update the monthly update zip with the appropriately patched magisk_patched file, will that OTA update install successfully and also preserve Magisk?

Thank you for any advise.
Tim

 

[Lughnasadh]

Question: I am new here, please redirect me to appropriate forum if this isn't where I should post my Q.

I have pixel phones rooted with Magisk, thus no OTA updates for me. So I download them from the Google site when the monthly update arrives, and flash using adb sideload. But it's a PITA wrt Magisk, I usually have to fuss around to get Magisk back running, although it's still worth it.

My Q is: if I update the monthly update zip with the appropriately patched magisk_patched file, will that OTA update install successfully and also preserve Magisk?

Thank you for any advise.
Tim

If you want an easy and painless way to update each month while keeping root, I would suggest Pixel Flasher.

 

[tim1aust]

If you want an easy and painless way to update each month while keeping root, I would suggest Pixel Flasher.

Thank you! I am currently building it from the .tar.gz on my Manjaro box. I will try it when the next update comes round.

 

[andy242]

Thank you! I am currently building it from the .tar.gz on my Manjaro box. I will try it when the next update comes round.

You don't need to build it the Ubuntu version runs on manjaro, I'm using it. Version 5 had an issue on manjaro but it's been fixed with the latest test build

 

[badabing2003]

You don't need to build it the Ubuntu version runs on manjaro, I'm using it. Version 5 had an issue on manjaro but it's been fixed with the latest test build

It's not actually fixed, if I build it on my machine manually, it works for you guys, but if I let Github build it, it doesn't.
I can't put my finger on it as to why?
I don't want to build manually anymore for each release, but I also don't want users having to build it manually either (unless they want to).
That's why this is bugging me.

 

[tim1aust]

You don't need to build it the Ubuntu version runs on manjaro, I'm using it. Version 5 had an issue on manjaro but it's been fixed with the latest test build

Thanks again for the info. But... too late! I already got it built and running, after learning more about building wxPython (or any python module, actually), and also installing the nautilus browser to select boot images. It's been interesting! This tool seems very well done, I am hopeful to use it successfully soon. Thank you Thank you Thank you, for bringing it to my attention.
Tim