Couldn't get any useful info from hwdb but dmidecode came up trumps:

I'm opening up a new line of attack but I could do with some help on it!

I grepped for that throttling message in the kernel source and found it came from the mce driver section. Now I have in Slackware a program called mcelog and a script for running it in daemon mode. The trouble is, it won't start because the script looks for /dev/mcelog and exits when it fails to find it.

It looks from here as if this device was deprecated in 2017. The Slackware kernel config shows:
I assume the legacy version is this /dev/mcelog. But what has replaced it and why does this script still look for it?

OOps! I only remembered late last night that I built a new kernel to incorporate the spectre corrections. I used Patrick's config file, but of course new options like "legacy mce" got their default values and that was one of them.

I shall need to rebuild it with that enabled. Then we'll see what mcelog says.

Yes, it works with the rebuilt kernel. Stupid me! I have added it to the multiuser level, logging into its own file. We'll see what, if anything, it says.

It happened again just now. Yet there have been no reports of core overheating in syslog since Oct 1st. Weird!

Just wanted to say I am now eagerly following this as I have seen one or two other reports of something that sounds similar amongst the Vivaldi users base but it is not something anyone in the team has encountered, so we don't really know what to do make of it. Like you I suspect that there is a hardware issue but why we trigger it and other software doesn't I don't know. Out of interest, have you seen it with any other chromium based browsers?

Oh and one other thing, one of few other people reporting it mentioned that he know longer had the issue after the second to last snapshot (our test versions) update. I don't know if that is a coincidence or it could be related to the fact that there was a minor Chromium update in that build and this changed something in some way.

[EDIT] I did read your comment again more carefully and saw this:

but I still wonder if we could find out what triggers it with Vivaldi perhaps that would clarify things.

If you can trigger with Vivaldi easily perhaps try running it with `--enable-logging --v=1` which should create a debug log inside your settings folder that you could inspect later (after you force a reboot). This might give you some clue as to what Vivaldi was doing at the time

interesting thread

I've had pretty much the same issue intermittently running Firefox on a D-10 system. The O/S is on SSD -- this is the first box I've built running the O/S off SSD. Also, it has a Ryzen5 (2600 ) CPU and DDR4 memory -- ASUS STRIX/B450 MoBo

Symptoms start with poor mouse response which then ends in a dead system: no mouse; no keyboard.

I've been suspecting a hardware error: when I restart the system Firefox goes right back to the same page it failed on and zips right thru it without further trouble.

I get those, too, sometimes but the initial poor mouse response usually means that I am running low on RAM. I happens if I open too many tabs and am running another memory intensive task in parallel. If I manage to close Firefox fast enough the system goes back to normal response. The issue in this thread happens suddenly without prior warning.

It stopped happening. I haven't had one of these freezes for a couple of weeks now, touch wood!!
Mind you, I'm no longer using vivaldi. But when they were occurring regularly, they happened with FF too and sometimes in other programs. And I am still none the wiser.

The last cpu throttle was on October 2nd. And there's nothing in mcelog at all.

1 members found this post helpful.

how far did you get with "living" in another LFS install for a while?

Not very well. The current LFS is LFS 9, but it's almost impossible to build because of all the extra dependencies. I have a desktop but I can't build gtk because that now needs docbook, and docbook comes with a sheaf of other things that I don't want. And without gtk, most of my apps won't work. I suppose I could just build everything in the book and then it would all work but it seems pointless. If I want a distro with everything in it, I can install Debian or something like that. Or just stick with Slackware. Why go to all the trouble of building from source? The only reason for LFS, as far as I was concerned, was to make a distro that just contained what I needed and nothing else. I really appreciated that but I can't do it any more.

The old LFS (8) still works fine but it's out of date now and will get more and more so with time.

I got zapped this morning at about 0845. Same symptoms: very slow, jerky mouse. I waited to see if it would clear; no luck.

I hit the reset button. when the system came back online Firefox reported that Adblock Plus had been installed. I had been running Adblock plus -- but had not re-installed it; it may have been a routine update.

Anyway I removed Adblock Plus and substituted the more popular just plain adblock. However: as noted earlier in this thread: this problem tends to be dormant for -- weeks -- and then pops up again.

as we all know: web pages are fighting the ad-blockers with everything they can come up with,...... hmmmmmm

But I wasn't using an ad blocker.

hmmmm. I've not used Vivaldi. I was using Opera for a while. Opera has a built-in ad-block; I don't know about Vivaldi.

So far though: the stoppage has only occurred when attempting to load a web page.

I had not had this happen on my earlier box; only since I started using this new build. This build has a number of changes:
Ryzen 5 ( 2600 ) cpu
DDR 4 memory
Radeon RX 460 graphics card
O/S loaded on SSD
STRIX B450-F MoBo
D-10 O/S 4.19 kernel
Currently: Firefox-ESR browser

we may find folks having this particular halt have a common use of 1 or more of these components -- dunno