Framework Laptop Review (Intel 12th G. Laptop) with Linux: The Definitive Review
boilingsteam.com
boilingsteam.com
EDIT: Oh and the speakers are bad and the mic sometimes doesn't work after waking up from suspend but I read that there's a fix for it.
Great machine overall—if they can fix a few little bugs it will be pretty much perfect.
Panel Self Refresh (PSR) seems to be at least the most common cause:
> @kelnos, have you tried without psr? E.g. add i915.enable_psr=0 into your kernel boot parameters or disable it via debugfs (as a root):
echo 1 > /sys/kernel/debug/dri/0/i915_edp_psr_debug # yes it's '1'
> You can check that psr got really disabled by: sudo cat /sys/kernel/debug/dri/0/i915_edp_psr_status
Sink support: yes [0x03]
PSR mode: disabled
PSR sink not reliable: no
I'm on a Framework 12th Gen running Fedora 37, kernel 6.1.9, with no freezes for several weeks after having loads of them under both GNOME and KDE on Ubuntu and pop_OS. [root@fwk ~]# ls -alF /sys/kernel/debug/dri/1/i915_edp_psr_debug
-rw-r--r--. 1 root root 0 Feb 15 09:09 /sys/kernel/debug/dri/1/i915_edp_psr_debug
[root@fwk ~]# echo 1 > /sys/kernel/debug/dri/1/i915_edp_psr_debug
-bash: /sys/kernel/debug/dri/1/i915_edp_psr_debug: Operation not permitted
I'm not sure why I'm unable to write. Do you have to mount debugfs differently to do writes or something like that?EDIT: debugfs is mounted rw and I tried to disable SELinux temporarily with setenforce 0 but I still can't write.
[root@fwk ~]# mount | grep debugfs
debugfs on /sys/kernel/debug type debugfs (rw,nosuid,nodev,noexec,relatime,seclabel)
[root@fwk ~]# setenforce 0
[root@fwk ~]# sestatus | grep Current
Current mode: permissive
[root@fwk ~]# echo 1 > /sys/kernel/debug/dri/1/i915_edp_psr_debug
-bash: /sys/kernel/debug/dri/1/i915_edp_psr_debug: Operation not permitted
I also got a kernel update today (6.1.11). According to [the changelog](https://cdn.kernel.org/pub/linux/kernel/v6.x/ChangeLog-6.1.1...) there are a lot of changes regarding the i915 driver but it still freezes on my system. Feb 15 09:12:53 fwk kernel: Asynchronous wait on fence 0000:00:02.0:gnome-shell[2332]:1252 timed out (hint:intel_atomic_commit_ready [i915])
Feb 15 09:12:57 fwk kernel: i915 0000:00:02.0: [drm] GPU HANG: ecode 12:0:00000000
Feb 15 09:12:57 fwk kernel: i915 0000:00:02.0: [drm] Resetting chip for stopped heartbeat on rcs0
Feb 15 09:12:57 fwk kernel: i915 0000:00:02.0: [drm] GuC firmware i915/adlp_guc_70.bin version 70.5.1
Feb 15 09:12:57 fwk kernel: i915 0000:00:02.0: [drm] HuC firmware i915/tgl_huc.bin version 7.9.3
Feb 15 09:12:57 fwk kernel: i915 0000:00:02.0: [drm] HuC authenticated
Feb 15 09:12:57 fwk kernel: i915 0000:00:02.0: [drm] GuC submission enabledThe Fedora docs cover setting module parameters.[1] This is how setting the kernel module parameter manifests on my system:
$ sudo grep -iR i915.enable_psr /etc/modprobe.d
/etc/modprobe.d/i915.conf:options i915 enable_psr=0
$ sudo cat /etc/modprobe.d/i915.conf
options i915 enable_psr=0
Assuming you use the default GRUB 2 boot loader, the Fedora docs also cover setting the kernel boot parameters.[2]1: https://docs.fedoraproject.org/en-US/fedora/latest/system-ad...
2: https://docs.fedoraproject.org/en-US/fedora/latest/system-ad...
[root@fwk ~]# cat /etc/modprobe.d/i915.conf
options i915 enable_psr=0
[root@fwk ~]# cat /sys/kernel/debug/dri/1/i915_edp_psr_status
Sink support: yes [0x03]
PSR mode: PSR2 enabled
Source PSR ctl: enabled [0x80000226]
Source PSR status: SLEEP [0x30000114]
Busy frontbuffer bits: 0x00000000
Frame: PSR2 SU blocks:
0 0
1 0
2 0
3 0
4 0
5 0
6 0
7 0
It also has no effect when I append i915.enable_psr=0 to GRUB_CMDLINE_LINUX in /etc/default/grub and reboot. $ sudo cat /sys/kernel/debug/dri/1/i915_edp_psr_status
Sink support: yes [0x03]
PSR mode: disabled
PSR sink not reliable: no
I'd raise a proper support request[1] with Framework then, who are generally more responsive and thorough with these kinds of problems than most laptop tech support (they put me down this path to start with).1: https://framework.kustomer.help/contact/support-request-ryon...
[root@fwk ~]# fwupdmgr security
Host Security ID: HSI:3 (v1.8.10)
HSI-1
CSME manufacturing mode: Locked
CSME override: Locked
CSME v0:16.0.15.1810: Valid
MEI key manifest: Valid
Platform debugging: Disabled
SPI BIOS region: Locked
SPI lock: Enabled
SPI write: Disabled
Supported CPU: Valid
TPM empty PCRs: Valid
TPM v2.0: Found
UEFI platform key: Valid
UEFI secure boot: Enabled
HSI-2
IOMMU: Enabled
Intel BootGuard: Enabled
Intel BootGuard ACM protected: Valid
Intel BootGuard OTP fuse: Valid
Intel BootGuard verified boot: Valid
Platform debugging: Locked
TPM PCR0 reconstruction: Valid
HSI-3
Intel BootGuard error policy: Valid
Intel CET Enabled: Enabled
Pre-boot DMA protection: Enabled
Suspend-to-idle: Enabled
Suspend-to-ram: Disabled
HSI-4
Intel SMAP: Enabled
Encrypted RAM: Disabled
Runtime Suffix -!
Intel CET Active: Supported
Linux kernel: Untainted
Linux kernel lockdown: Enabled
Linux swap: Encrypted
fwupd plugins: Untainted
I was wondering whether maybe one of those (maybe the kernel lockdown) prevents me from switching the PSR mode?I'm also looking forward to upgrading when new boards come out.
Finally, there's a huge opportunity to DIY expansion ports, it's just USBc on the inside.
So bust for now :( I'll stick with Lenovo Thinkpad for now.
i don't know what the root cause is. it's most likely that linux isn't keeping up with the latest chips. maybe they need to cobble up parts from 5 years ago that do have linux support and release that. at least it would work.
1. This is a Linux-friendly laptop for the most part, but it's still built with Windows as the primary OS (and the only one you can get preinstalled). When you're building things with Linux kernels in mind, it's easy to pick current hardware that supports Linux optimally (see also: every Chromebook).
2. The places where this laptop isn't perfect are _nothing_ like the problems Linux laptops had twenty, or even ten, years ago. I used a Thinkpad with Ubuntu in 2012, and it worked pretty well... but also ran super-hot and burned through battery ridiculously quickly. The Framework doesn't do that. Sure, it doesn't have amazing battery life (around 5-6 hours), but it also doesn't have amazing battery life in Windows; it's just the nature of the hardware. The automatic-brightness thing is real, but relatively minor compared to all the rest of the hardware (Wifi, fingerprint reader, etc) working smoothly.
Apple will keep milking us until the community can come together and address the compatibility issue. We need an open linux laptop consortium that can address these issues across the board and then manufacturers can deliver a working product and compete on other dimensions, like how rounded the corners are and thin the laptop is. As it stands, manufacturs appear to be cheering they've sold 10 laptops with broken wifi! But it's wifi6e+8g revision 2023 and has a fingerprint reader. Disconnected.
Meanwhile, Apple is laser dialed on every relevant metric and they're rewarded with a closed eco-system and high margins.
https://www.tuxedocomputers.com/en/TUXEDO-InfinityBook-Pro-1...
My understanding is that some of this is because the Intel 12th gen 15-watt U-series chips only had 2 performance cores so Framework went the 28W P-series instead to get 4 performance cores.
Using socketed ram is also going to use more power than lpddr.
> sleep efficiency under Linux
What are you comparing it to? Macbooks can sleep a week and lose 5% of battery but I don't know any non-apple machines that can do that.
The Framework sleep perf seems comparable to my XPS 13.
IME all non-apple laptops suck at suspend.
Sleep works out of the box with Fedora 36/37.
> By the time they fixed this to be on par with the best of the best in linux notebooks it's already old hardware.
It seems to work as well as it will already? It's easy to swap out the mainboard if more efficient chips come out.
Do your old laptops use 5% of battery after a week of sleep? That's the performance I've observed with Macbooks.
Older laptops don't have this issue so bttery use is no problem.
https://community.frame.work/t/guide-fedora-36-hibernation-w...
They didn't do any work yet. It doesn't look like there is anyone at Framework working on Linux issues.