My Snapdragon Dev Kit was healthy and working fine until a Windows update failed

Considering that you were seeing unpredictable behavior in the boot selector, with it randomly freezing, I would assume a hardware component (RAM?) kicked the bucket. If it were firmware corruption, it would consistently fail to present the menu, or wouldn't boot at all.

Microsoft's code quality might not be at its peak right now, but blaming them for what's most likely a hardware fault isn't very productive IMO.

After decades of experience, my normal practice is to make regular multi-generational backups of the entire disk (usually compressed). When something like this goes wrong, I can revert to the last known-good image and go from there. It saves a lot of time and trouble.

These devices are nightmares. I'm sure things will pay off at some point but this feels like all those years where everyone was cursing Nvidia on Linux and praising AMD's dedication to open source but my computer would constantly lock up regardless until I switched to Nvidia. There was this massive disconnect between my experience and what everyone told me was best supported.

Similarly, I'm constantly hearing about Qualcomm's renewed interest in Linux and this and that and how the X2 Elite will be fully supported but I have never known them to be like this. A decade or so ago we were trying to work for a school project on one of their dev kits and the documentation was so sparse.

Then I see that the Snapdragon X Elite comes in this Ideacentre stuff but looking online no one has gotten Linux anywhere close to as good as Linux is on a Mac M2. That, for me, is the marker. If a Mac can run Linux better than whatever chipset you've released, it's just not hardware worth buying. If you're not Apple, you have to support Linux. Otherwise, to borrow Internet lingo, you're "deeply unserious".

Yeah they (probably) did not.

Almost certainly a soft hardware failure, likely the SSD.

I've run into a similar situation - except the culprit was Linux not Windows. Tossed the machine in a closet for a few months, when it miraculously started working again. Until it broke again a day and a half later. It's disk or RAM corruption.

Give it up dude, it's the hardware, but let not an opportunity to smash Microsoft go unfulfilled.

I was pleased to discover recently that Ubuntu is supporting my NVidia Jetson going forward after NVidias official support period ends:

https://canonical.com/blog/ubuntu-now-officially-supports-nv...

So there is at least one ARM devkit with long term Linux support.

Sounds like this could potentially be some defective RAM. Memtest86 can boot from UEFI directly, so it should hopefully show up in BDS. A run should tell you what regions of RAM are bad, if any.

Not that you are at fault here, but I'd be very hesitant to install any system updates so shortly after they brick my computer, especially when Microsoft is involved.

Only adding to this because it's likely a hardware failure, and I had a challenging time debugging a similar issue in an aftermarket engine ECU years ago. Thought it might make a fun anecdote.

The car would run fine once started, but the car just wouldn't start sometimes (quite modified so I knew the systems well). The started would turn as that was a simple relay, but all ECU controlled devices wouldn't trigger. Plugging into the ECU, no error codes and all looked normal.

Eventually we tracked the issue to some corruption in the ROM that was only getting read in certain circumstances, since the ECU stores maps for engine parameters based on things like pressure and temperature you might only hit the corrupted bits of a table in very specific circumstances.

Reflashed the ROM and all was good afterwards. The suspected cause of corruption was intermittent power supply that had been fixed a while earlier.

I doubt this is a Windows issue.

I would replace your ram sticks. I had a similar mysterious issue on an old Intel nuc. Got some new sticks off Amazon and never had the problem again

Hopefully this serves as a reminder to decision makers with Web backgrounds to NOT push random non-critical _firmware_ updates without clear merit, or random updates in general.

Security is not fluids. It doesn't naturally evaporate. So don't try to add like they're washer fluids.

Those low-level software and associated hardware don't take software overwrites very well, even today. They might have total cumulative max overwrites, or manufacturer supplied update codes can still be dubious. It's (not)okay if you are meaning it to be a tool for your planned obsolescence strategy, otherwise, just don't touch it for the sake of doing it.

If it updated the UEFI it could be that somehow the UEFI now starts the hardware wrong (RAM setup maybe?) that causes it to be unstable.

The holy trinity of modern hardware: update servers, firmware locks, and abandoned devkits

change the SSD and retry (the same ssd in another machine may not trigger the same error btw, this is not unilateral process of elimination) - those windows updates do a lot of disk writes and a small miss there can screw up an entire install since it shuffles things around in preboot environment (moving them on disk) and that can corrupt things and prevent a new install in the same way.

You can also try to live boot into Ubuntu 25.04 arm64 since that iso has experimental snapdragon elite support and has some built-in drivers for storage and network - you can extract firmware from the windows drivers with qcom-firmware-extract - they recommend doing this from a windows partition which you should have (albeit possibly corrupted).

If that still fails - you have a ram issue as others have pointed out. I've had the exact same symptoms (hardware instability after windows update) and it was nvme ssd (an early samsung one) and ram, in both instances.

Not saying the windows update didn't also come with some junk firmware that got loaded into some of your devices, but that would be a distant diagnosis from ssd/ram (and many others would have seen the exact same thing during their update if it was that).

Nobody has the time or energy to chase companies up for this stuff, and you know somewhere in the T&C they inserted a legal clause which is expensive to contest or un-contestable to liability.

But, that said, it saddens me we've normalised "oh well" when it comes to kit. even dev kit. If MS can't manage release engineering to keep dev/test things alive, then it's not helpful to the belief they can do it for production things either.

I inherited an IBM PC/RT back in the 90s. It was well outside what most people would consider its support lifetime. IBM could not have been more helpful working out how to keep it alive. I suspect this influences why when I later had some financial authority I was happier to buy thinkpad, than any other hardware we had available: I knew from experience they stood behind their maintenance guarantees. The device was configured to run BSD, not the IBM supported OS of the day, made no difference. It was end of life product line, made no difference.

This was before Lenovo of course. But the point stands: people with positive support stories, keep that vendor in their top-set

If it's a proper devkit it should have accessible and documented test points for all voltage rails and should have come with a complete schematic. It should be possible to go through them with a voltmeter or oscilloscope and see if everything looks ok.

Given the symptoms (random crashes not right away at boot), and given that qcom is anal about secure boot, my guess is that it's unlikely that it's a firmware (in SPI-NOR or wherever) corruption that initially caused this. Firmware is checked each boot.

Might be as simple as degraded capacitor, or something similar.

And I can imagine that it's not hard to destroy this kind of HW physically with a SW update. PMICs can often produce voltages way higher than Vmax of connected components. But it's unlikely that if bug like that happened, that it would only affect one devkit out there, and not a whole range of devices.

Not sure how much it could help, but is there a possibility you connect the SSD to another machine with the same architecture, run Windows install in it, then once Windows is installed and running, shut down, move the SSD back to the Snapdragon kit and attempt to boot? Just an idea...

We just bought a Snapdragon Windows device.

I trust Microsoft 0% to keep developing Windows for it.

Man, it looks like a memory issue. Is the memory user replaceable on this machine?

As a data point, my Windows Dev Kit 2023 (also a Snapdragon) has been working great so far.

are you able to stop at uefi stage and system is stable in bootloader stage? if yes than it may not be software issue. Others have covered checking ram and ssd. I suspect it could also be thermal or voltage issue.

Soo.. Qualcomm can use a Windows drive to receive calibration data and other configurations. If you have a virus or something, you might brick a board, if its connected. We used 3-4 days in the factory to figure out why our boards were bricked. The PCs on the production line were all infected.

Cool. Running development for fork Ruby on Rails /src/ crates to match kit for enterprise-level software firmware recovery paths.

Seems like the typical Microsoft experience nowadays.

My ROG Ally ran fine on Windows 11 at the beginning, but a year later always randomly crashed, even when idle, on a fresh OS install. After switching to SteamOS it runs stable again.

THere's a chance that there was a hardware failure and _that's_ why the update was failing. And not a failed update that caused the hardware failure.

Either way, may the memroy of your Snapdragon Dev Kit be a blessing.

Completely off topic: the font on that site is quite readable and easy on the eyes and brain, the way the dyslexia font on the Kindle is.

[flagged]

Snapdragon is already canceled so I guess, they just don't care about this device. It's Microsoft on ARM. Sad to say, but don't expect full support or quality update on this.

Ref:

- https://www.youtube.com/watch?v=XrA2Xe9f7e8 - https://www.jeffgeerling.com/blog/2024/qualcomm-snapdragon-d...