thanks for the comments in my last two posts – I didn’t acknowledge all of them (some looked like spam, wasn’t sure). Anyway. I’m currently running Debian Buster and noticed a few things…
Intel GPU
In my previous Posts I’ve shown how to enable GuC and such stuff. It seems those aren’t available anymore:
[ 5.408288] i915: unknown parameter 'enable_guc_loading' ignored [ 5.408290] i915: unknown parameter 'enable_guc_submission' ignored [ 5.408292] i915: unknown parameter 'enable_rc6' ignored
Which means that one can modify the file /etc/modprobe.d/i915.conf because it won’t do anything:
options i915 enable_guc_loading=1 enable_guc_submission=1 enable_rc6=7
enable_rc6 seems to be gone completely. enable_guc_loading and enable_guc_submission can be replaced by just enable_guc:
options i915 enable_guc=1
ACPI_OSI
You might remember, I had trouble getting this notebook running without issues until I did set:
GRUB_CMDLINE_LINUX="acpi_osi=! acpi_osi=\"Windows 2009\""
in /etc/default/grub. I just checked the HP page and there is an BIOS update (yay!) which I just applied:
root@asuna:/home/jean# dmidecode -t bios | grep "Version" Version: F.17
The Release date is 06/19/2018. Not sure if it has to do with the BIOS update or updates in the Linux Kernel, though the acpi_osi stuff is not required anymore. The system boots up fine and no graphic issues without the above lines. Just to make sure:
root@asuna:/home/jean# uname -a Linux asuna 4.18.0-2-amd64 #1 SMP Debian 4.18.10-2 (2018-11-02) x86_64 GNU/Linux
However, I do still have 181 ACPI Errors shown in dmesg. The Intel GPU, Bumblebee / the nvidia-card still works fine without the acpi_osi stuff.
Bumblebee / Discrete graphics
It seems the above modifications cause a tiny probably ignorable message in dmesg. If you start nvidia-settings with optirun -c :8 and close the window, dmesg shows:
[ 82.618653] ------------[ cut here ]------------ [ 82.618656] pci 0000:01:00.0: disabling already-disabled device [..]
However, since the card is still disabled, you might ignore that.
On the BUS thingy
I’m still not able to select the new bus for my touchpad:
root@asuna:/home/jean# grep RMI4_SMB /boot/config-4.18.0-2-amd64 # CONFIG_RMI4_SMB is not set
While it still tells me:
[ 2.445681] psmouse serio1: synaptics: The touchpad can support a better bus than the too old PS/2 protocol. Make sure MOUSE_PS2_SYNAPTICS_SMBUS and RMI4_SMB are enabled to get a better touchpad experience.
TPM
I’ll have to check more, probably a separate blog post. Seems I can do something with the TPM chip in this notebook. Not absolutely sure, though. Still getting:
[ 0.000000] ACPI: TPM2 0x000000006A672DA8 000034 (v03 HPQOEM 838F 00000001 HP 00000000) [ 5.305272] tpm_crb MSFT0101:00: [Firmware Bug]: ACPI region does not cover the entire command/response buffer. [mem 0xfed40000-0xfed4087f flags 0x201] vs fed40080 f80 [ 5.306104] tpm_crb MSFT0101:00: [Firmware Bug]: ACPI region does not cover the entire command/response buffer. [mem 0xfed40000-0xfed4087f flags 0x201] vs fed40080 f80
While this looks promising:
root@asuna:/home/jean# time tpm2_hash foo hash(sha1):22596363b3de40b06f981fb85d82312e8c0ed511 real 0m0.024s user 0m0.001s sys 0m0.007s root@asuna:/home/jean# time sha1sum foo 22596363b3de40b06f981fb85d82312e8c0ed511 foo real 0m0.009s user 0m0.000s sys 0m0.003s
With the new kernel in-place, the acpi_osi stuff removed and GuC enabled, pcie_aspm forced my notebook shows me 11:59 hours to go on battery instead of 4-6. This is quite interesting. However, Let’s see where it sets down and how long it does run in reality. Started with battery usage at 15:00 and will update this post later.
Update: It’s 21:32 and my Notebook shows 03:15h to go. I’m working with it all the time by the way. Some browsing, chatting, sometimes calculating photo-hashes. My IDE is running as well as chrome, thunderbird, pidgin, and ~9 terminals. So, whatever (updates? kernel-update? bios-update? the acpi-osi-switch?) caused this, massively increased battery time. And ~9-10 hours seems do-able.
No Comments