New Runtime Standby ABI Proposed for Linux Like Microsoft Windows' 'Modern Standby' (phoronix.com) 59
Phoronix reports on "an exciting post-Christmas patch series out on the Linux kernel mailing list" proposing "a new runtime standby ABI that is similar in nature to the 'Modern Standby' functionality found with Microsoft Windows..."
Modern Standby is a low-power mode on Windows 11 for letting systems remain connected to the network and appear "sleeping" but will allow for instant wake-up for notifications, music playback, and other functionality. The display is off, the network remains online, and background tasks can wake-up the system if needed with Microsoft Modern Standby...
"This series introduces a new runtime standby ABI to allow firing Modern Standby firmware notifications that modify hardware appearance from userspace without suspending the kernel," [according to the email about the proposed patch series]. "This allows userspace to set the inactivity state of the device so that it looks like it is asleep (e.g., flashing the power button) while still being able to perform basic computations..."
Those interested can see the RFC patch series for the work in its current form, in particular the documentation patch outlines the proposed /sys/power/standby interface.
"This series introduces a new runtime standby ABI to allow firing Modern Standby firmware notifications that modify hardware appearance from userspace without suspending the kernel," [according to the email about the proposed patch series]. "This allows userspace to set the inactivity state of the device so that it looks like it is asleep (e.g., flashing the power button) while still being able to perform basic computations..."
Those interested can see the RFC patch series for the work in its current form, in particular the documentation patch outlines the proposed /sys/power/standby interface.
Standby on Linux (Score:3)
Re: Standby on Linux (Score:2)
That's what a kernel developer told me when I reported this problem. In particular, they said I had a machine with an nVidia device that was different from all the similar devices they'd seen in not having the second copy engine actually function. They gave me one line of config for my PCI ID, and then it worked perfectly.
They found and fixed the issue immediately upon seeing logs from my machine, but they'd never seen exactly that issue before to know in advance that it was a possibility.
Re: (Score:2)
Literally all my linux systems can suspend and resume.
Older ones suspend to S3, newer ones suspend to S2idle.
Re:Standby on Linux (Score:5, Informative)
Suspend issues in Linux are persistent problem with Nvidia's drivers. If you use other hardware or drivers and your distro's permissions backend packages are properly installed you shouldn't have a problem. For hibernate though, you need to make sure your swap space is equal to or greater than the size of your RAM, and if you have an old BIOS with the option "Memory hole at ..." turn it off.
Re: (Score:2)
ah, I will check out memory hole on some of my systems.
At least on 6.1 you have to be below 50% RAM usage too.
I found this in a RHEL doc that pointed to a kernel README that looked old af but said the same thing.
I have a few systems that run an app on solar during the day at 80% RAM and I had to stop the service before suspend to get it to work.
Yet it worked for a couple months in disk hibernate but then stopped and only memory sleep would work. On a Debian Bookworm stable kernel, so who the heck knows wha
Re: (Score:2)
Last system I had that couldn't sleep/wake was a Dell laptop running Windows 10. 5 years of OS/BIOS updates and reinstalls never fixed it. What finally fixed it was installing Debian a couple months ago.
The only problem I've had suspending Linux in the last decade was that it liked to forget my monitor configuration about once a week on resuming. This resolved when I changed to a distro using Wayland.
Re: (Score:2)
The last system I had that couldn't sleep is my current Linux build with a Gigabyte motherboard that is notorious (I discovered after-the-fact) for not sleeping under Linux. There is no known cure, and it seems to have been the case for 20+ years. My prior system, which had the same NVidia graphics card, slept and resumed every day just fine for months on end. I will never buy another Gigabyte motherboard.
Re: (Score:2)
My Gigabyte board was apportioned with the Incredible Disappearing NIC. It hasn't used its superpower in a few weeks, but I bought a USB networking dongle to keep handy just in case. They say it doesn't always come back.
Re:Standby on Linux (Score:5, Informative)
Modern Standby isn't what you want anyway. Mostly what it does is ensure that when you open your bag on Monday morning it's burning hot and your laptop has 1% battery left. While it was on "standby" it decided to install some updates and restart a few times, until the battery reached critical level.
Re:Standby on Linux (Score:4, Interesting)
That's more of a Windows bug, because Modern Standby with Networking is only supposed to be entered when you're plugged in. The moment you unplug you should transition to Modern Standby only which disables everything and puts you into the low power mode.
But the reason you end up with hot laptop problem is because you're not exiting Modern Standby with Networking (or Powered Modern Standby) into Battery Modern Standby.
Powered Modern Standby means your laptop is still plugged in but in low power state (monitor off, interactive sessions off, it just does essential network activity like updates and such in the background), and able to respond to wakeup commands to perform say, a backup.
Battery Modern Standby is basically traditional standby but gives a moment to let things prepare for it - close network connections, reset the firewall etc., so things don't assume that existing connections are still usable and accidentally leak information in the clear where they may accidentally retry a connection.
Re: (Score:2)
Re: (Score:2)
That's more of a Windows bug, because Modern Standby with Networking is only supposed to be entered when you're plugged in.
It's not a bug, it's an optional configuration. The whole point of modern standby in windows is for tablet like devices, not for devices which are plugged in all the time. It was literally introduced for the Surface line and you absolutely want networking.
Also with networking does not mean a hot laptop. That still very much is a problem with a device that is sitting on and in some light sleep state, not in standby with networking. That's not a windows issue, it's a device specific issue.
I get this too, my D
Re: (Score:2)
"Monday morning it's burning hot and your laptop has 1% battery left"
that's been happening to me with the past 2 corporate Lenovo laptops i've been given and changing the power options hasn't worked
Re: Standby on Linux (Score:1)
Re: (Score:1)
Completely agree, and that's why the best way to single out Windows users at conferences is by looking at those who are desperately looking for a mains outlet
Hibernate? (Score:5, Informative)
I know it's more of a distro thing than a kernel thing, but seriously... How about being able to enable hibernate without having to have a PhD in hibernate?
Re:Hibernate? (Score:4, Funny)
They could have fixed this stuff years ago, if it wasn't for the political opposition and FUD coming from Big Sleep.
Re: (Score:2)
They could have fixed this stuff years ago, if it wasn't for the political opposition and FUD coming from Big Sleep.
What FUD? There's very little use case for Hibernate these days, unless you intend to leave your computer for a week or so standby uses so little power these days that hibernate is borderline irrelevant.
Also add to the fact that hibernate requires reloading your RAM which these days is slower than simply booting your computer.
Re: (Score:3)
Oh boy! (Score:2)
Re: (Score:2)
Re:Oh boy! (Score:4, Informative)
Re: (Score:2)
So, it's just another bullshit marketing term that doesn't accomplish what you think it does, like "Responsive Design", or "Secure Boot".
Re: (Score:2)
Re: (Score:2)
It's a copy of the "power nap" function that Apple have had for many years...
The Apple implementation is actually pretty good, fans don't spin up and the laptop will keep syncing your messages etc while it's plugged in.
The fact that MS have a half assed copy of the Apple feature is nothing new.
Re: (Score:2)
It *seems* to be fixed right now, at this current moment, but Apple being Apple, it may break again next update.
It's frankly fucking annoying there's no way to turn it off.
Re: (Score:2)
Beyond me why it would be a great idea to foist this on Linux users. In my experience the juice is not worth the squeeze at all. Maybe it is something that has to be implemented due to the horrible state of many mobile PC firmware (UEFI) these days.
Re: (Score:2)
Re: (Score:3)
Right, Microsoft started doing this with the use case to take Skype (RIP) calls while the computer is in "standby", but that was before everyone had a smartphone and any messaging app connected there.
The problem is that Microsoft disabled the "real" standby in a series of cat and mouse game with what's the workaround of the year for this, and let's disable it. With th
Only if it doesn't ... (Score:3)
What will this gain? (Score:3)
What would this gain, I'm curious. If the CPU isn't on standby, why have the OS go into a special mode which will introduce new bugs and security threat vectors.
Worst case, why not just have a "run level" that freezes and saves state of things, stops all apps, and effectively drops the machine into single user mode with some program watching ports and other input, to reverse things? However, this duplicates standby and suspend functionality, so why bother?
Re: (Score:2)
The point is to enable you to put the computer into a low-power state without stopping all the apps. I'm guessing this is something certain application developers have been asking for.
It could be useful for embedded systems, and maybe something like a smartphone. But for a general-purpose, traditional computer, it would be pretty useless and just muddy things up when you're trying to determine the state of your computer and operating system. Starting with the most basic: Is it on?
Re: (Score:2)
It could be useful for embedded systems, and maybe something like a smartphone. But for a general-purpose, traditional computer, it would be pretty useless
The average person probably wouldn't see too much difference between the two, in terms of what they're used for. I'm not saying it's ultimately worth it to me, but I can see the general appeal (potential issues aside) of having everything updated in the background. It's not a "linux purist" feature, but a feature that is probably beneficial for user adoption.
Re: (Score:3)
When your aunt click on the "sleep" button after an update and now the fans keeps spinning, and the device wakes up unexpectedly to do some "helpful" thing no user asked for, the appeal will quickly wane. On embedded/mobile/whatever device, power consumption is a must, and most of them are *designed* to be woke regularly in regular use. On a desktop/laptop, the "surprise, I wasn't sleeping" would require some adaptations. A laptop "waking up" in a backpack because you thought it was asleep would be bad, fo
Re: (Score:2)
Basically, on a "full-fledged" computer, if I hit sleep, it should sleep. If I want it to be super available, I just leave it there, CPU will get to his lowest power state to do basically nothing, screens will turn off, fans, will ramp down (assuming it's not burning summer), and there will be no chance for confusion. This would consume more power, but probably not that much. When idle, a modern computer goes very low in that regard.
That is your preference, and nothing wrong with it. However, a lot of people would have a different opinion. From my experience providing support to the public, most people just hit the power button and let the computer figure it out. I'm not saying there wouldn't be exceptions or problems, but for the average person it works well enough and they can't be bothered to care.
As for power usage, I don't have first hand numbers to go off of. However, consensus is that modern standby is much lower than idle. Th
Re: (Score:3)
Supporting end-users I have definitely seen the "laptop wakes up in bag and suffocates itself" scenario a bunch.
The "Shut Down" option no longer shutting down creates other confusions as well. If you want to completely shut down the OS and reload everything from disk, the only way to do that now in Windows is the "Restart" button. "Shut Down" and powering back on is no longer the same. If you've been doing that, try running "systeminfo" and looking at the date Windows booted. You will be surprised.
Re: (Score:2)
Supporting end-users I have definitely seen the "laptop wakes up in bag and suffocates itself" scenario a bunch.
I've seen that too, but the vast majority of the laptop users I deal with only have one so they can move about the house.
The "Shut Down" option no longer shutting down creates other confusions as well. If you want to completely shut down the OS and reload everything from disk, the only way to do that now in Windows is the "Restart" button. "Shut Down" and powering back on is no longer the same. If you've been doing that, try running "systeminfo" and looking at the date Windows booted. You will be surprised.
Sure, but that ultimately isn't overly important to how most people use their computer. That is, it's generally better for the computer to keep itself up to date in the background than to expect them to ever take the time to do it themselves. Having them restart the computer usually has less friction than having them run updates.
Ultimately, adding the option seems like an overall benefit
Useful for privacy invasion (Score:2)
This warm standby will be very useful for anyone wishing to remotely activate your PC and, for example, turn on the camera or microphone.
Modern Standby AKA "Power Nap" (Score:4, Insightful)
I think Macs did this first, introducing Power Nap back in 2012 - although to be honest I've never found it particularly useful. I can't say I've ever wished it were on any of my Linux systems.
Re: (Score:2)
Apple being Apple, you can't put it into a deep sleep, because they know better than you.
I'm with you- I can't say I've ever wished this particular form of computing were on my linux laptop. It's in fact the only one I have currently immune to it.
Oh great (Score:1)
Now linux machines will be able to lie to me about when they are awake. Great innovation guys.
Mobile devices (Score:2)
Standby on Linux? What for? (Score:2)
Serious question. I never found any good use-case for that.
Re: (Score:3)
It means the room is a more pleasant temperature when I come back, and also it means I don't have to enter my disk encryption password again.
Re: (Score:2)
and also it means I don't have to enter my disk encryption password again.
Ah, so you want to be less secure. I admit that is something standby does for you.
Re: (Score:2)
It seems you don't understand the purpose of disk encryption. It is designed to protect data "at rest" from someone with physical access to the machine. No one has physical access to my desktop when I put it in suspend. They may have access if I leave the house entirely, which is why I shut the PC down entirely when I leave.
The most secure thing is to never boot the machine at all. I want to be somewhat less secure than that.
Kernel Panics! (Score:4, Interesting)
I can't wait for this new level of kernel panics!!
I only state this because I have a Dell laptop for work, and Windows will consistently kernel panic (blue/black screen) when entering this mode.
Something devs forget to check is to see how well PCIe devices, in my case, Thunderbolt, goes in/out of sleep. These devices have direct memory access, and if you don't properly put them to sleep, shit breaks hard! It honestly took me well over a year of troubleshooting to figure out exactly it was these sleep states + thunderbolt at the same time causing the issue. Take either out of the equation, and shit works just fine.
Re: (Score:2)
This is frequently the case... The manufacturer will have tested the laptop itself and all its functionality, but there's no way for them to test it with every possible external peripheral and combination of drivers for those external peripherals.
Dancing Forwards and Backwards (Score:2)
Here I am just wishing my Ubuntu Mate laptop would lock the screen before going to sleep. The OS some versions ago would do that. Now I've got to write a custom systemd script to look for sleep events and lock the system prior to that. Though that doesn't solve the issue of the desktop being fully visible on wakeup prior to the lock screen blocking it and visible while locked when plugging in or unplugging an external display. I'm glad linux is progressing. I just wish its "user experience" was also pro
this needs to be fixed before it spreads (Score:1)
Part of the Reason Why I Left Windows (Score:2)
modern (Score:2)
Modern standby is not a low power mode. It's a normal power mode.