Check this out …
Eventually that is related?
2 Likes
My distro just got 6.10 and I haven’t been affected by the issue since, but it’s been less than a day. Will update if the suspend-crash happens again.
The wake-crash happened again a few times, so 6.10 definitely didn’t solve the issue.
Having this issue too for a few month… Happens every 20-25 wake from suspend instances, running Linux 6.10 but this has been going on long before 6.10 release. Machine just crashes and reboots, last log are always from suspending the computer, nothing from wake and afterwards.
For me it’s 100% of the time. Though it’s probably worth noting that sometimes instead of black screen -> reboot it’s frozen screen/graphical glitches -> reboot
Can also confirm that no changes to files in the brief wake period persist after the crash.
Not unless there already was one, no.
@Max_Pearce_Basman Alright…Nevermind, then. I had one set (i.e. storing the disk encryption passphrase in the TPM), removed the BIOS Boot PW and so far, so good (been only a few days). I read in the Release Notes from the 0.0.3.5 FW update [1] that there has been an issue where the system hangs and the EC resets the system if it can’t access the nvme after resuming from suspend.
But anyhow, that also wouldn’t explain the artifacts you describe you’re seeing shortly before system reset.
I’m getting this issue as well, happens maybe 1 in 10 lid opens - you open it up and after a few seconds see the Framework logo and a cold boot. I’m using the AMD 13 7840U model with the latest firmware (03.05). There’s lots of swap space (which shouldn’t matter given that it’s doing a suspend via s2idle). I’m using Debian and tried 3 different kernels now with the same issue - 6.10.6, 6.9.7, 6.1.0. There doesn’t seem to be many power options in the UEFI config except allowing the PCI devices to drop to gen3 for power savings which seems to have zero effect.
I was going to switch the system to the deep/suspend-to-RAM method but that doesn’t appear to be available when I check /sys/power/mem_sleep.
I’ve disabled quick boot, quiet boot and secure boot to make troubleshooting easier but probably unsurprisingly nothing has helped as this seems like something in the firmware is deciding the system can’t resume. Very frustrating.
Edit: Also side note despite using ext4 which should be fairly crash resistant one of these accidental reboots destroyed my Signal messages database. I was under the impression that the system syncs filesystems on suspend (and there’s a “1” in /sys/power/sync_on_suspend) so that’s even more frustrating.
2 Likes
I am affected by this issue too.
I am on Ubuntu 22.04 with the oem 22.04d kernel. Not sure if it has the patch mentioned above included. What I see in the logs, though, is a bunch of messages like this:
kernel: xhci_hcd 0000:c1:00.3: Refused to change power state from D0 to D3hot
Heads up: Frustrated, disillusioned user rant.
I honestly thought that having removed the BIOS password would have fixed at least this hard reset issues, but unfortunately, i did not. The EC still hard resets the machine after suspend every 5-10 wakeup, and i’m still losing all my unsaved work.
Having said that, i feel disappointed that nearly a year after launch there are still this kind of issues with the embedded controller and/or the firmware of the device.
These half-working USB ports, these constant hard resets, the USB-C PD issues, the flaky firmware and amount of issues and quirks, other people are having to workaroung with their machines…
In my personal opinion, i now regret spending the 1700€ i paid for the issue-ridden experience i have had so far, and i cannot recommend (at least this AMD version) this laptop in good faith to anyone any longer, seeing all the kinds of issues it has and the amount of compromises one has to accept, even if you are willing to make sacrifices for the “greater good”.
In this state, the machine is indeed barely usable for any productive means.
With all the love and respect i have for Framework’s Vision and my willingness to accept some drawbacks, as a whole, truly unfortunately but nonetheless, i feel like this vision has failed me. The truth (= my personal experience) simply is, that i had never as many issues/drawbacks with any laptop i ever owned before, than i have had with this FW 13" AMD model.
It is as simple as it is sad.
1 Like
Did anyone try to contact official support for this issue?
Yeah I have a support request open that I’m going to work with them on.
Great. Please, keep us all updated on this thread if you can. That would be appreciated very much.
1 Like
I highly doubt that this is an kernel issue, considering the circumstances. After having calmed down (a bit), i think this might very well be a AMD Phoenix Platform issue, perhaps even out of the purview/reach of framework as a company at all. After having a look over the wall, it appears there are even M* powered Macs that are unable to cycle suspend/resume on a reliable basis.
Nonetheless, as @Garro pointed out, it might be very interesting having a follow-up from framework themselfs, if it doesn’t just boil down to “have you tried changing your FS” or “does it occur also with OS/2 warp”…
Windows Modern Standby truly is the IE6 of Platform Engineering these days…
So I opened a line to support a few days ago and their first recommendation was to reset the BIOS to optimal settings; I did that and the only settings I’ve changed afterwards (it was already on 3.05) was disabling quick boot, disabling quiet boot and disabling secure boot. So far it’s been about 4 days with no crashes whereas before I was getting a couple a day so fingers crossed that I’m in the clear!
@_luke Thank you for the follow-up… Are you intending to turn back on “secure boot” or are you going to leave it that way?
I’m not sure. I have everything working great on the latest kernel version which is unsigned as it’s not in Debian and honestly I’m not sure figuring out the machine key infrastructure is worth having it on.
Hi, I seem to have a similar / related issue.
Observed behaviour : once in a while (one every 10 ?) the FW 13 AMD Ryzen 7 7840U will not resume, and needs a hard reset.
system/kernel : Ubuntu 24.04.1 LTS / kernel: 6.8.0-45-generic
from journalctl -b | grep -i suspend
Oct 10 19:18:56 jb-fw kernel: nvme 0000:02:00.0: platform quirk: setting simple suspend
nvme ssd hardware
description: NVMe device
product: WD_BLACK SN850X 1000GB
vendor: Sandisk Corp
I have seen a number of posts related, but cannot find a solution, it looks like the new kernel version (6.11) may still have that “platform quirk: setting simple suspend” issue
Any help most welcome !
1 Like