Kernel panic with Tumbleweed but not Leap

I recently upgraded my hardware to a Ryzen cpu and AM4 motherboard. The first time I tried to boot Tumbleweed from the same ssd I used before the upgrades, I got a kernel panic and the system wouldn’t boot. I thought maybe it was all the new hardware that was the issue but I didn’t want to reinstall TW until I got a new NVMe drive so I installed Leap on another ssd. No kernel panic this time after installing Leap and all the new hardware is working fine. I got the NVMe drive today and installed Tumbleweed on it with no issues until I booted up the first time and got another kernel panic. I looked on Bugzilla but didn’t see any pattern of problems with the 4.10.* kernel experiencing kernel panics.

Anyone know what might be going on here? There’s also another odd thing. The ssd isn’t recognized by the bios or the TW installer when the NVMe drive is installed. I had to remove the NVMe drive in order to boot Leap again.

Since there have been no replies to this thread, I’ll assume everyone else is as puzzled as I am. I have made some progress though. Seems not all of the sata connectors on this AM4 motherboard are created equally. Two of them are called ASATA and the others are just SATA. The bios says the two ASATA are APU connectors while the SATA one’s are called chipset connectors. I had my ssd plugged into one of the ASATA connectors and that was the one that wasn’t recognized when my NVMe was installed. Plugging the ssd into one of the SATA connectors solved that problem.

I also discovered that when I get a kernel panic, if I go into the bios and select the drive to boot manually, I don’t get a kernel panic. Strange! This has worked with both an ssd and the NVMe drive, but only with Tumbleweed. I don’t get any kernel panics with Leap.

So I’m considering this issue mostly resolved. At least if I get a kernel panic, I know a workaround. I guess this what I get for upgrading to hardware as soon as it’s released.

Never heard of asata.
Googling returns only some questionable posts so I doubt “asata” disk connections are real.
If you believe I’m mistaken, pls post a technical reference.

Are you sure you don’t mean “eSata” which is literally the connections for external sata cable connectors?
eSata connections should be electrically the same as internal sata connections, but supports a shielded connector connected to cable shielding to protect externally running lines from EMF.

I don’t know how your BIOS looks, but it’s entirely conceivable that external connectors are managed differently than internal connectors and may need to be enabled.

TSU

A different search turned up this <solitary> post
https://www.reddit.com/r/Amd/comments/61dx4f/asata_slots_on_the_gigabyte_b350_gaming_3_board/

If there is such a thing, I’d want to know what this SoC (System on Chip) you have on your board.
If you believe this is something real on your board, pls post a link to your board’s technical specs.

TSU

Did you see the last post on the reddit link you provided? It says “you lose those if you plug in a nvme”. That’s exactly what happened to me. I also didn’t find anything when I did a search for asata. I think all Ryzen cpu’s are really apu’s. Even the user’s manual doesn’t explain the difference between asata and sata. It just says two of the ports are asata and the rest are sata. It’s only in the bios that it says two of the ports are apu and the rest powered by the chipset.

APU is AMD speak for GPU+CPU on a chip. No clue what asata is, Not that most CPUs these days come with built in GPU. Some models of the AMD FX series don’t have GPU but they are about the only consumer chips that don’t have GPU.

The Ryzen 7 1700 that I have doesn’t support on board video, even though the Gigabyte motherboard I have has HDMI and DVI connectors. The more expensive Ryzen cpu’s do support on board video.

OK,
Did a read on this, heavily relying on the Wikipedia entry…
https://en.wikipedia.org/wiki/Zen_(microarchitecture)

First thing to note is that AMD’s “Advanced Processing Unit” (APU) systems are just now being rolled out… First the GPU-less systems 4Q 2016 and just now in March and April systems which include an integrated GPU.

The design approach is innovative and revolutionary, attempting to inject a new level of parallelism and management within each core as opposed to conventional approaches to manage “per core” instead of “within the core.” Besides that, automatic frequency management based on multiple sensors within each core (surely because heat is “the” major factor that causes internal electrical resistance resulting in degraded performance) sounds like a really cool idea.

Because SATA connections can and are managed within an APU core, it seems reasonable to guess that an “asata” connection is one of these… Not a conventional SATA connection managed by an independent disk controller but rather managed from within the APU.

So…
Bottom line…
Like every other time revolutionary hardware is rushed to market (And I was one of those that bought an i7 laptop 3 months after Intel’s launch), it’s problematic that drivers will be integrated into the Linux kernel so soon. Based on my experience, it’ll take at least 3 weeks after launch for drivers to just be introduced upstream, then there will be a weeks-long (possibly a couple months at least) wait as that kernel works its way through down through the distros’ maintainers before the first kernels will be offered to Users, and then there will be a few more weeks before the kernel would be released as “stable.”

Therefor, it wouldn’t be surprising for instance if there is no Linux kernel driver support for “asata” yet…

Although today would still possibly be too soon, if you want to jump the gun a bit I’d recommend anyone using this technology to install a kernel from the special kernel repositories to hopefully get hardware support… But, of course YMMV regarding stability and support just about anything else.

IMO,
TSU

I think I’ll submit a bug report on Bugzilla about kernel panics with the 4.10 kernel but before I do I wanted ask about these messages I’m getting when I do get Tumbleweed to boot on the NVMe drive. This output comes after the boot splash screen:

“A start job is running for flush journal to persistent storage.”

And then this line keeps repeating:

“Unexpected IRQ trap at vector 07”

This will keep repeating for a couple of minutes until X starts. If I boot into runlevel 3, the message repeats itself indefinitely.

Anyone know what this means?

Nope but it is something to report