Hey y’all! A few things on:
- btrfs/zram hibernation (which I’ve been running successfully with 64GB ram), and
- a hibernation resumption issue that popped up a few months ago and my solution.
For those of you that haven’t seen it, there’s a great article on how to setup hibernation with btrfs/zram:
It started from this wonderful GitHub Gist:
I have hibernation/suspend-then-hibernate setup on my system where it will first S0ix sleep and then 2 hours later hibernate. Also hibernates at I believe 7% battery. It’s been working flawlessly since Fedora 35-36 until a few months ago, where sometimes on hibernation resumption, the system would freeze on a black screen. I troubleshooted today, and here’s my comment/solution:
If you have hibernation resumption issues, I recommend reading that, but to summarize so these words are populated in this forums search:
I added this to my kernel parameters to see debug logs upon system/kernel boot and hibernation resumption:
ignore_loglevel=1 initcall_debug=1
And saw that with resumption, my system froze here:
intel_ish_ipc 0000:00:12.0: [ishtp-ish]: Timed out waiting for FW-initiated reset
intel_ish_ipc 0000:00:12.0: ISH: hw start failed.
initcall ish_driver_init+0x0/0x1000 [intel_ish_ipc] returned 0 after 219091 usecs
To fix the issue, I simply blacklisted the intel_ish_ipc module by adding this to the kernel parameter:
modprobe.blacklist=intel_ish_ipc
Though it should also work by adding this to e.g. /etc/modprobe.d/blacklist.conf (or best-practice-naming as /etc/modprobe.d/intel_ish_ipc.conf) + rebooting:
blacklist intel_ish_ipc
After blacklisting, finally my system no longer freezes on that module and resumes normally!
Though it seems that the intel_ish_ipc module still loads afterwards since it’s a dependent module for intel_ishtp:
lsmod | grep ish
intel_ishtp 61440 2 intel_ishtp_hid,intel_ish_ipc
@ the Framework team
I’m still not sure if my issue was caused by a kernel regression, a Framework BIOS update, a Fedora update, or something else. These all kind of happened around the time this issue popped up. But I do remember this started happening right around upgrading to BIOS 3.09/3.10.
It could also be from a recent Linux kernel regression (a Kernel bug report I found), as I’ve seen reports on different systems (Ubuntu 22.04 Dell XPS 9500 and Arch Linux).
Is there any downside to blacklisting the intel_ish_ipc module? Though it seems to get loaded later as a dependent module for intel_ishtp anyways.
Everything still seems to work normally on my system + hibernation resumption works everytime again.