Just opened my ThinkPad and for the second time in three days the act of closing the laptop to sleep it, then opening it again had caused Ubuntu to hardlock. This time I got a very brief small printout about "amd ring 0 error", then it went back to a black screen and I had to hold down the power button again.
I thought getting a ThinkPad, getting AMD cpu/gpu and picking the Linux distro Lenovo lists as supported would mean I got a minimally functioning computer but I guess not.
I do complain about computers a lot but I feel like ultimately I don't ask a lot out of a computer. I just want, like… all the computer parts to work. The wifi seems to work but also sometimes it seems slower than it should be. The builtin microphone is like, I'd say a 60% chance on any given bootup it will turn out to be working. The computer itself, like. Does not know how to power down and power back up. This has been a problem from day one. It got better then it got worse
@mcc I just fucking hate laptops. They ride the edge of HW all the time, and all their power saving gizmos work the instant they ship, and then not at all, because oh my gods the bullshit they pull to make it work...
@TomF@mcc the trick is to use Linux so they don't work when they ship either because the kernel still needs a couple of iterations to get support for them.
@mcc I’ve got a framework 13 that I upgraded to the amd 7840 board, latest bios. I’ve been running fedora 39 and now fedora 40. After the amd and bios upgrade this is the best Linux laptop I’ve ever owned and I’ve had a lot of Lenovos and thinkpads. It might be the fedora dist but either way I’m satisfied. Also, it had a lot of those issues with the old intel board.
We spent all these years making jokes about "Linux on the desktop" but Linux on desktops is actually pretty good now. The issue is when you try to use a laptop and suddenly it turns out Linux has phenomenally broken support for the things laptops have but desktops don't (trackpads, high-dpi screens, sleep/wake)
So yeah— I've heard this from a few sources, that part of the problem is Microsoft recently introduced something called "Modern standby" and it fucks up Linux hard on any laptops that have it: https://social.treehouse.systems/@megmac/112357580797659333
I've seen several sets of instructions on how to switch back to traditional sleep:
@Di4na My three biggest issues with linux on a laptop are failures with sleep/wake, failures to properly handle the hidpi screen, and failures with handling multi-touch gestures. When I first started using Linux, none of these three features (to the extent SioX can be construed as a feature) existed
@mcc even on windows there are laptops out there that don't handle it right. it's so frustrating to get a new laptop today just to find it'll get stuck in sleep/hibernation or even crash and basically defeat using it at all.
@mcc That also did manage to fuck up my Linux desktop quite hard, and no matter what I did, refucked it each time I booted Windows. The only workaround I could find was to never boot Windows, which as far as workarounds goes is um... severe.
@mcc the real killer here is that "modern standby" also fucks up Windows, which has similar fail-to-sleep-and-get-dangerously-hot-in-a-closed-bag failure mode. I think LTT did a video on this a while back and I've heard others complain about it too.
sometimes using a mac feels like giving up on general purpose computing but at this point I question whether a portable computer is "general purpose" if, when I take it out of my bag, it doesn't turn on because its battery is dead
@mcc@flyingsaceur oh, I thought that matthew garrett was working on this like 3 years ago? I had assumed it must have been landed but yeah I don't see any progress since he was posting about it in 2021
@rotopenguin I don't understand what it is siox offers that I would want on a laptop computer. I would like wake on ethernet. But I don't need this on a laptop computer.
@mcc I think it was supposed to offer "waking up as fast as your phone does. Fast enough to even have Cortana catch what you're asking without missing a word!"
What it really offers is "this is the only interface that MS is using, so we OEMs will make sure it works exactly according to how MS uses it and not one step farther".
@rotopenguin Yeah, this doesn't match how I use computers. My computer takes several seconds to wake up because I need to enter a password and such and anyway the screen seems to take a few seconds to realize it's supposed to turn on even with the siox thingy
@mcc I believe the option was removed in Gen 3. My X13 Gen 2 AMD has (had?) it, but it was implemented very poorly and caused more problems than
I've been using the modern standby stuff ("S0ix" is the tech name vs. the branding name) and it works better. I think I had to add iommu=pt to the kernel args for it to work right, but this was ~2 years ago.
I very, very, very seldomly have WiFi dead when I wake it; like once every 2-3 months. Turning WiFi off and on again fixes it. It's not great, but it somehow still feels better than the work MacBook Pro, which has WiFi break when I'm actively using it, having done nothing special and changed not a thing.
@ieure You have to understand I've spent the last ten years using ancient macbooks manufactured in eras when apple made quality products so "wifi is hardly ever dead on wake" is not very impressive to me
So the correct thing to do is not upgrade to Ubuntu 24.04 now, instead I am told I want to wait until 24.04.1 and then upgrade. Except the thing is, about the exact same time 24.04 got released my 23.10 install started having perpetual, obnoxious problems with crashing when I close the laptop lid. Like this wasn't a problem a month ago, now out of nowhere it's really obnoxious. And 24.04.1 isn't coming out until like, August. That's nearly the 24.10 release.
@mcc cw: unsolicited and utterly unhelpful advice: can u configure to "do nothing" when u close the lid? (yes, i kill all my laptop batteries by treating like desktops and plugging them into the wall 24x7 :( )
Ok ok maybe on a more serious note: is the crash from handling what i assume is an interrupt from closing the lid regardless of configured action or does it stop happening if closing the lid is a no-op? Might narrow it down (i appreciate ur probably not looking to debug this)
@snosrapkungfu I don't know. I don't feel like I'm interested in debugging this unless I'm talking to an actual Ubuntu engineer who I know will do something with the information. My attempts to file bugs with Ubuntu haven't gotten good responses so far
@mcc ugh, sorry to hear that. I'm a little out of the loop myself, from the aforementioned hw abuse and from using xubuntu for any box running canonical sw bc i find both gnome and kde horribly bloated (yeah, i'm old)
@mcc I have now completely given up on the intermediate releases, and only use LTS, and this is what I recommend to everyone. The support period for the intermediate releases was shortened drastically at the same time as LTS support was extended, so now you barely have the chance to use them before they're EOL and you have to upgrade.
(The LTS releases are also more likely to be supported officially by vendors outside the traditional Linux ecosystem.)
@confluency I had a sense I probably should have done this, but 23.04 ran well on my desktop so I figured 23.10 would run well too. But 23.10 has been no end of problems!
I knew I should have done the LTS when I installed this one but I wanted to use the newest version so I could use Wayland. But Wayland doesn't really work.
@mcc Yeah, I keep trying Wayland (usually by accident, because it was turned on by default), finding something that doesn't work, and then going back to X. One day I'm sure it will officially arrive, but I think X will be with us for a lonnnnng time.
@confluency The thing that killed it for me is if you have a fractional dpi monitor (like, 1.5x) you can't render X apps normal in Wayland mode. X apps render normal in 1.5x in X, Wayland apps render normal in 1.5X in Wayland, X apps render wrong (blurry) in 1.5X in Wayland. The Qt Wayland server doesn't have this problem because it's more flexible, but Mutter does have this problem. And the X apps in question turn out to be things like "every Electron app".
@mcc if you happen to be on btrfs (or zfs, I guess), why choose? You could take a Minotaur snapshot, leap right in to Numbat, and hop back if it's worse.
(Ideally, you would have /home on a subvolume (or, sigh, its own partition, i guess), and have that mounted in no matter which OS is running)
My use pattern has been different from normal this week and I'm now up to the ThinkPad failing to wake up after closed-lid sleep 3-4 times a day. I am dying
I have committed no sin here except the hubris of believing I could run Linux on a laptop
@mcc Bummer. I hate "modern standby" (it's a terrible idea; IIRC it was introduced because sleep is hard and manufacturers don't want to support it), and was able to restore normal sleep. But I have a much older Thinkpad.
@mcc I'll have to research this the next time I need to pick a new work laptop. It's going on the "hardware enshittification to watch out for" list under "soldered RAM".
I have a bad feeling it's going to be unavoidable. I guess at least an SSD makes it more feasible to power your laptop off whenever you don't want it to run down the battery and set your bag on fire. :blobfoxdisapprove:
@confluency apparently intel and amd just stopped making chipsets that support s3. I think someone said there's a Framework blog post where they complain that they want to support s3 but it's not actually possible to make a laptop anymore that does
@mcc It looks like even recent-ish Lenovo laptops still support S3 in the BIOS, but only for Intel chips (not AMD). But if Intel has also dropped support, that can't last forever.
@mcc I've had a Framework Laptop 12th Gen (Intel) for more than a year now, with Arch Linux and the latest kernel (6.8.7) and the latest Gnome (46) on Wayland. No sleep/wake issues in general. I also set the power button to suspend, and it puts the laptop to sleep consistently as well. I haven't tried hibernate.
In the past I'd seen maybe two instances when closing the lid didn't put the laptop to sleep, but not anytime recently. Maybe the latest kernel and/or drivers have fixed something?
@mcc not sure if this is the issue but a bunch of intel chipsets have a bug where if you close the laptop while plugged into power, it wouldn't sleep correctly, and would lock up and kill the battery.
This would happen on piles of intel laptops, even macs.
@mcc New one or old one? My experience is that you need to wait a year for Linux to catch up. I reluctantly bought first gen "fresh" Lenovo P16 for the lab I worked at. Took 6 months for the standard sleep/wake issues to be resolved. It has been like that forever IMO. Also, it being listed as "supported" doesn't mean much. Unless linux comes pre-installed you can bet it is broken at some level.
@mcc I just accepted that closing the lid doesn't equal locking the computer. If you open my laptop, you see what I was doing it before I closed it and have a few seconds to interact with it. If there's a terminal window, you have time to type "rm -rf /"
Add comment