HCL - Lenovo Thinkpad P14s Gen 1 (AMD)

kellyling · February 9, 2022, 4:38pm

Qubes 4.1 (R4.1)

Fresh install with all default options (Fedora 34, Debian 11, Whonix 16)

Installation and first boot:

The installation had a "PCI 002… " error upon completion of install but booted to login screen
Network Manager service had to be manually added to NetVM to enable Wireless
Manually remove PCI002 (Realtek Ethernet Controller) to remove error
Reboot required after above to enable wireless configuration/detection

Confirm working:

- Volume control and mute (F1/F2/F3) with OSD
- FN-lock/FN-lock indicator (FNLOCK/ESC)
- Mic mute/unmute (F4)
- Screen brightness (F5/F6)
- Wifi on/off (F8)
- Capslock/Capslock indicator
- Touchpad (tap/click/scroll)
- Trackpoint and buttons
- Power control (Power button)
- Temperature/fanspeed sensors
- Ethernet (RJ45 , plugged in/out indicated)
- MicroSD reader (detected/readable/writable)
- Integrated speaker (Sound output)
- Keyboard (Tested US keyboard layout)
- Logitech USB Unifying receiver (for Logitech Mouse + Keyboard)

Not tested yet:

- Bluetooth

DisplayPort out

Not Working:

- Hibernation (Blank screen)
- Suspend-to-RAM (Close lid,opens to blank screen. Need to hard reboot)

Performance observations:

Boot time is extremely slow: 
    Total startup time is 2mins 50s !!! 
    for comparison on same laptop (total startup, fresh install), 
              Arch/KDE ~ 25s
              Ubuntu/GNOME ~ 30s
              Fedora 34/GNOME ~ 32s

HCL Output

layout:
‘hcl’
type:
‘notebook’
hvm:
‘yes’
iommu:
‘yes’
slat:
‘yes’
tpm:
‘unknown’
remap:
‘yes’
brand: |
LENOVO
model: |
20Y1CTO1WW
bios: |
R1BET67W(1.36 )
cpu: |
AMD Ryzen 7 PRO 4750U with Radeon Graphics
cpu-short: |
FIXME
chipset: |
Advanced Micro Devices, Inc. [AMD] Renoir Root Complex [1022:1630]
chipset-short: |
FIXME
gpu: |
Advanced Micro Devices, Inc. [AMD/ATI] Renoir [1002:1636] (rev d1) (prog-if 00 [VGA controller])
gpu-short: |
FIXME
network: |
Realtek Semiconductor Co., Ltd. RTL8111/8168/8411 PCI Express Gigabit Ethernet Controller (rev 0e)
Intel Corporation Wi-Fi 6 AX200 (rev 1a)
Realtek Semiconductor Co., Ltd. RTL8111/8168/8411 PCI Express Gigabit Ethernet Controller (rev 15)
memory: |
22247
scsi: |

usb: |
4
versions:

works:
‘FIXME:yes|no|partial’
qubes: |
R4.1
xen: |
4.14.3
kernel: |
5.10.90-1
remark: |
FIXME
credit: |
FIXAUTHOR
link: |
FIXLINK

fsflover · February 10, 2022, 8:26pm

It’s not expected to work (yet):

github.com/QubesOS/qubes-issues

Simulate Hibernation / Suspend-To-Disk

opened 03:20PM - 03 Nov 16 UTC

xloem

T: enhancement help wanted C: core P: default

I think this is a known idea but I didn't find an issue for it. For laptops, …hibernation is great to just close the lid and come back to your work a week later. For security, this is important because so much data is in RAM. Xen does not support suspend-to-disk, but Qubes could emulate suspend-to-disk by hibernating all the VMs, and then simply shutting down. It would be great if this were formally supported some day.

Sven · February 12, 2022, 3:22am

Thank you @kellyling for your R4.1 HCL report, which is online now.

kellyling · February 12, 2022, 4:56am

Notes with thanks @fsflover and @Sven

Further update to the above,
Tested the system on kernel 5.15.14-1.fc32.qubes…86_64 (kernel-latest).
Hardware compatibility stands, no new issues.

Performance update to the above,
I should clarify that ‘some’ of the slow boot performance might be attributed to the disk being fully encrypted. For those who want to gain a few more seconds during boot , you can try disabling plymouth from the grub config (GRUB_CMDLINE_LINUX=“rd.plymouth=0 plymouth.enable=0”) on dom0 . Real-world hand timing indicated a gain of ~10secs (Graphical target improved from ~49s to now ~37s). Might gain more by going the minimal template route - haven’t the time to test it out yet.

Overall, R41 (with defaults and no workarounds) has been working good on this laptop as a daily driver for the past 3 days of 8-hour shift work/zoom meetings/heavy worksheet crunching + a few hours of Steam gaming sessions.

deeplow · April 15, 2022, 2:50pm

A post was split to a new topic: Motion of Windows Have Time Lags in Between Moving Them (Thinkpad L15 Gen1 Ryzen Pro 7)

alzer89 · April 15, 2022, 3:56pm

Just out of curiosity, @kellyling, what does your journalctl say when resuming from S3 sleep. I know you get a black screen, but does it say anything about ring gfx timeout?

I have encountered bugs in amdgpu when running through Xen, causing Xorg to coredump. That’s why I’m asking…

Also, how on earth did you manage to get a Steam session in with integrated graphics? That’s incredible! I can barely get 10fps with a 4800U with RX Vega 8…

kellyling · May 4, 2022, 10:02am

@alzer89 , I don’t see any ring gfx timeout entries in my journalctl. In saying that, I’ve not seen much amdgpu errors lately.

As for Steam, I guess it depends on the games you’re planning to run. I tested a few ‘older’ titles and they ran pretty decent on low/mid settings. I did set vram settings to 2048M in BIOS, and also used a custom Debian minimal VM .

alzer89 · May 4, 2022, 10:10am

So, if you S3 sleep (suspend to RAM), does everything work as expected?

Like, it doesn’t implode on resume? No black screen? No crashes?

The context for why I’m asking this:

github.com/QubesOS/qubes-issues

Cannot Resume from S3 Sleep - amdgpu (I think...)

opened 02:22AM - 30 Mar 22 UTC

alzer89

T: bug C: other P: default hardware support needs diagnosis

### Qubes OS release 4.1 ### Brief summary Basically, the machine can't suc…cessfully resume from S3 sleep under Qubes OS. ```amdgpu``` (I think...) cannot successfully resume VGA controller from S3 sleep, causing Xorg to coredump, and machine needs to be hard reset. ### Steps to reproduce 1. Boot machine 2. Initiate S3 sleep 3. Resume from S3 sleep. ### Expected behavior The machine wakes up, and doesn't implode. ### Actual behavior 1. The display successfully resumes to ```xscreensaver``` for 2-3 seconds. a. If audio was playing before S3 suspend, it now continues playing fine for 2-3 seconds. 2. The screen flashes like it is changing resolution. a. If audio was playing before S3 suspend, it now becomes incredibly choppy and stutters. 3. The screen shows whatever was on the screen at Step 1, but the cursor cannot be moved. 4. Steps 2 & 3 repeat 4-6 times. 5. The screen flashes again. 6. The screen backlight stays on, but every pixel is black. a. If audio is playing before S3 suspend, it now stops. 7. Cannot switch to a TTY (Ctrl + Alt + F2). 8. The machine appears to have imploded, and a hard reset is needed. ### Additional information - Hardware: GPD Win Max 2021 - AMD Ryzen 7 4800U with AMD Radeon Vega 8 (4000) Integrated Graphics - - S3 sleep/resume works perfectly in the Mainline Kernel, as well as on every single other Linux distro I have tried (even the obscure ones) that are not based on Xen. * I'm not saying Xen is the issue, but it's worth noting that the machine is actually capable of resuming successfully from S3 sleep, so it is definitely not a hardware issue... - Videos of machine sleeping and resuming: * kernel-latest-5.16.15: https://user-images.githubusercontent.com/66879804/160737835-d93608a4-ecdf-4b99-aa1c-eaa7e79c0db5.mp4 * Audio does not resume because audio hardware is in ```sys-usb```, and it doesn't seem to get passed through before ```Xorg``` implodes. * Mainline Kernel 5.17.1 built with Qubes OS config and patches added in: https://user-images.githubusercontent.com/66879804/160737859-b2c086ab-f641-4b82-a55d-62fd0cc9f4eb.mp4 * Audio does not resume because audio hardware is in ```sys-usb```, and it doesn't seem to get passed through before ```Xorg``` implodes. - The machine is currently not used for anything other than Qubes OS testing. Happy to provide: - Logs - Output of any commands ### Final Word A massive thank you to all the Qubes OS team. Your work is endlessly appreciated!