Device compatibility list (add yours)

[ExTV]

This is the running list of devices Podroid has been seen on. Add yours in a comment and I will fold it into the table.

What to post

Device:         e.g. Samsung Galaxy A24 (SM-A245F)
Android / skin: e.g. Android 16, One UI 8.0
Podroid:        e.g. 1.2.6
Backend:        the Backend row on the Status screen, e.g. Auto (qemu)
AVF reported:   Settings, Diagnostics, "AVF (pKVM) diagnostic". The second
                block prints "Feature: virtualization_framework / supported = true|false"
Anything odd:   only if something failed or behaved differently

The AVF line is worth including even when the app lands on QEMU. It is a property of the device, it needs no permission and no settings change to read, and it is the single fact that decides whether a device can ever be fast.

Why the list is shaped this way

Podroid has two backends, and which one a device can use matters more than anything else about it:

• AVF (pKVM) is hardware accelerated. Boots in seconds and a desktop is genuinely usable. The device has to report android.software.virtualization_framework, which today is mostly Pixel-class hardware.
• QEMU (TCG) is software emulation, and the fallback everywhere else. It runs on any arm64 Android 8 or newer device, which is why most entries below are QEMU, but it is much slower and heavy GUI work is where you feel it.

There are deliberately no performance numbers here. A boot time on QEMU depends on the device's single-core speed, the RAM and cores given to the VM, its thermal state, and whether load balancing is on, so two people with the same phone can land far apart. A table of numbers would read as authoritative and push people toward or away from hardware on bad evidence. What is stable and worth recording is: does it run, which backend it lands on, and any device-specific problem worth warning about.

The list

Device	Android / skin	Podroid	Backend in Auto	AVF reported	Notes
Nothing Phone (1) (A063)	Android 15, Nothing OS 3.2	1.2.8	QEMU (TCG)	no	maintainer test device
Google Pixel 10 Pro XL	Android 17	1.2.8	AVF (pKVM)	yes	maintainer test device
Google Pixel 6 Pro (Verizon)	Android 17	1.2.7	AVF (pKVM)	yes	locked bootloader; QEMU also confirmed working when forced; reported by @crazybilly
Google Pixel 11	Android 17	1.2.8	AVF (pKVM)	yes	reported by @WriteCodeEveryday
Google Pixel Tablet	Android 17	1.2.8	AVF (pKVM)	yes	locked bootloader; all vCPUs come up; QEMU also confirmed working when forced; Waydroid does not run on the stock guest kernel; reported by @Peterlovesvscode
Motorola Moto G06	Android 15	not given	AVF (pKVM)	yes	boots, but the guest gets no network interface at all (see below); reported by @kassadorrmax
Samsung Galaxy A24	Android 16, One UI 8.0	1.2.8	QEMU (TCG)	no	reported by @red7nikotin
Samsung Galaxy Tab A11+ (SM-X230)	Android 16, One UI 8.5	1.2.8	QEMU (TCG)	no	reported by @ProSteve123GH
Samsung Galaxy Tab S7+ (SM-T975)	Android 13, One UI 5.1.1	1.2.8	QEMU (TCG)	no	also runs inside Samsung Secure Folder; reported by @Entersjkhdfkjdhfksjf
Motorola Moto G7	Android 10	1.2.8	QEMU (TCG)	no	reported by @WriteCodeEveryday
Nothing Phone (1)	Android 13	1.2.8	QEMU (TCG)	no	reported by @WriteCodeEveryday
Honor Play 60 Plus	Android 14, Magic OS 8.0	1.2.8	QEMU (TCG)	no	reported by @red7nikotin
Huawei nova Y72S	EMUI 14.2.0	1.2.8	QEMU (TCG)	no	reported by @red7nikotin
Huawei Pura 70 Pro	EMUI 14.2.0	1.2.8	QEMU (TCG)	no	reported by @red7nikotin
ZTE Blade A75 5G (Z2357N)	Android 13, MyOS 13	1.2.8	QEMU (TCG)	no	reported by @ProSteve123GH
FriendlyElec NanoPi R5C	Android 14	1.2.8	QEMU (TCG)	no	single-board computer, not a phone; reported by @red7nikotin

All reported devices started with default settings and no configuration. Entries that said "not checked" for AVF before 1.2.8 were re-read by their reporters once 1.2.8 started showing the real backend on the Status screen.

Known device-specific problems

• Some Tensor G3 and G4 Pixels cannot bring up every vCPU under AVF and end up on fewer cores ([Bug] VM exits on boot with AVF backend on Pixel 8a #29).
• Some devices get no network adapter at all under AVF, so the guest boots without eth0: a Lenovo Tab M11 (AVF Backend: Any plans for USB passthrough support and network issue fix? #43, still open) and a Moto G06 so far. If you hit this, switching the backend to QEMU in Settings still gives you a working network, just slower.

If your device is in either group, say so in your entry. Those are the entries that save someone else an evening.

[crazybilly]

I tried a Google Pixel 6 Pro, a Verizon version with locked bootloader--went pretty great!

Device: Google Pixel 6 Pro, a Verizon version with locked bootloader.
Android / skin: Android 17
Podroid: 1.2.6 and 1.2.7
Backend: successful with both QEMU and AVF
AVF reported: Feature: virtualization_framework / supported = true

[kassadorrmax]

My Moto G06 boots the vm with AVF, but cannot access internet. There is no network interfaces.

[red7nikotin]

Now, in version 1.2.8, the issue with the correct display of the emulation mode has been fixed. I’ve checked all my devices again and, unfortunately, all of them don’t support the AVF mode.

So, I ask you to correct my data in this table and also add this smartphone —

“Samsung Galaxy A24”,

which, unfortunately, also doesn’t support AVF.

[red7nikotin]

Device: NanoPi R5C
Android: 14
Podroid: 1.2.8
Backend: Auto (qemu)
AVF reported: Feature: virtualization_framework / supported = false

[ProSteve123GH]

Device: Samsung Galaxy Tab A11+ (SM-X230)
Android: 16/OneUI 8.5
Podroid: 1.2.8
Backend: QEMU (TCG)
AVF Reported: virtualization_framework : supported = false

[ProSteve123GH]

Device: ZTE Blade A75 5G (Z2357N)
Android: 13/MyOS 13
Podroid: 1.2.8
Backend: QEMU (TCG)
AVF Reported: virtualization_framework : supported = false

[ExTV]

Table updated with everything posted since the last fold:

• Pixel 6 Pro (Verizon, locked bootloader): AVF, both backends confirmed. Thanks @crazybilly.
• Moto G06: boots under AVF but no network interface. That is the same symptom as the Tab M11 in AVF Backend: Any plans for USB passthrough support and network issue fix? #43, so it is now listed under known problems. @kassadorrmax, if you can post the Android version and the output of the AVF diagnostic from Settings, Diagnostics, it would help pin down what these two devices have in common. Forcing the QEMU backend in Settings should give you a working network in the meantime.
• Galaxy A24, Honor Play 60 Plus, nova Y72S, Pura 70 Pro: AVF column corrected to no, versions bumped to 1.2.8. NanoPi R5C added. Thanks @red7nikotin.
• Galaxy Tab A11+ (SM-X230) and ZTE Blade A75 5G: added, both QEMU. Thanks @ProSteve123GH.

[kassadorrmax]

Android 15, Moto G06:

[WriteCodeEveryday]

Here's some of my current test devices on 1.2.8

Motorola G7, Android 10, No AVF but works under QEMU
Nothing Phone 1, Android 13, No AVF but works under QEMU
AVF Reported: virtualization_framework : supported = false

Google Pixel 11, Android 17, AVF and working perfectly (high score on sysbench)

[Peterlovesvscode]

Device: Pixel Tablet
Android/Skin: Android 17, Pixel UI
Podroid: 1.2.8
Backend: AVV
AVF reported: Fully working, no multi core bugs
Anything odd: waydroid doesn't work on the stock kernel, this fixes it: https://github.com/Peterlovesvscode/Podroid-forked/releases, locked bootloader, qemu also works but is much slower

[Entersjkhdfkjdhfksjf]

Device: Samsung Galaxy Tab S7+ (T975)
Android / skin: e.g. Android 13, One UI 5.1.1
Podroid: 1.2.8
Backend: Auto (qemu)
AVF reported: Feature: virtualization_framework / supported = false
Anything odd: none
It also runs under Samsung's secure folder

[karuboniru]

Device: Redmi K Pad 2
Android / skin: Android 16, HyperOS 3.0.309.0.WAKCNXM.C11
Podroid: 1.2.8
Backend: auto (avf)
AVF reported: N/A, see below
Anything odd: any operation related to avf hang the device, followed by a automatic device reboot (something related to watchdog?)

A simple glance into the dmesg suggested that there is something todo with RCU stall due to the VM holding a CPU and blocking intrs. The android's builtin AVF application works so I suspect this is related to how we start the VM?

[ExTV]

Thanks to both of you, two very different data points here.

@Peterlovesvscode: Pixel Tablet on Android 17 with AVF fully working and no multi-core trouble is a useful contrast to the reports below, and your Waydroid finding was the real catch. The stock kernel had CONFIG_PSI off, so /proc/pressure did not exist and Android's lmkd could not start. PSI is now enabled in the kernel for the next release, so Waydroid should no longer need your fork. Discussion #50 has the details.

@karuboniru: a full device hang followed by a watchdog reboot is the most serious AVF report so far, since everything else we have seen stays inside the VM. Your RCU stall reading fits the pattern tracked in #29, where the AVF backend requests a host-matched CPU topology and the guest kernel is missing pieces that topology assumes. On Tensor devices that ends in a VM reboot; taking the host down with it is worse, and the K Pad 2's hardware is untested territory for Podroid. Google's Terminal app working is worth knowing but does not isolate the cause on its own, since it boots a different guest kernel and declares a GPU.

Three things would narrow it down a lot:

1. Set CPU cores to 1 in Settings, then start the VM. Does it still hang the device?
2. Under Settings, Show advanced, force the QEMU backend. That path never touches AVF, so it should be safe; please confirm whether it is.
3. If you can capture them, the dmesg lines around the stall, plus Settings, Diagnostics, Export diagnostic log.

Please be careful with repeated attempts given that the device reboots itself. If a single core avoids the hang, that points at the same root cause as #29 and I can gate the topology for this hardware.

[karuboniru]

@ExTV

1. Already tried single CPU, no luck.
2. QEMU backend works as expected (otherwise we are talking about a CVE rather than a bug).
3. Please see attached, this contains dmesg before the crash and the application's diagnostics (but not from the invocation with AVF, since it hangs).

01_pre_reboot_kernel.sanitized.log
app.log

For the dmesg, the interesting spots are around lines with rcu_preempt detected stalls and kick=0xdf,check=0xff, the first is the message where RCU detects a stall on CPU 5 and the second is the watchdog fails to get reply from CPU5, and reports crosvm_vcpu0 is running on that CPU).