[Kernel 5.2] Looking for brave testers

I'm looking for a (or even more than one) tester for my Intel-tuned kernel, consider this kernel as being an experiment :wink:
It is based on Manjaro 5.2, and includes among others the following patches:

  • BFQ updates assembled by Piotr Górski
  • Some patches from Clear Linux
  • Graysky's CPU optimisations
  • Futex-Steam patch
  • CFS scheduler tuned for lower latency
  • a few fixes for bugs

So far it received only minimal testing on my laptop (I noticed a small improvement in boot time) and it would be cool if someone could give some more feedback.

You should ideally have a pure Intel platform, Haswell or later, with Intel graphics and Intel networking, like for example many high-end ThinkPads (X/T).
EDIT: You don't have to have such a platform. It should run on any Intel and even AMD, but I think most gains are only achieved on pure platforms.

It uses the "performance" governor and workqueue.power_efficient=0 by default, so if it draws too much power, choose "powersave" and use boot parameter workqueue.power_efficient=1.
It also uses Haswell CPU optimisations by default, you can change it in the kernel config to fit your processor.
Included are PKGBUILDs for nvidia, virtualbox and acpi_call extramodules.

Thank you!

N.B: make sure you checkout intel branch, not master!
Recommended(?) boot options on stable platforms (for testing):
nohz_full=0-63 cpuidle.governor=teo


4 Likes

Wanted to try this on my t460s. For some reason it fails, one or more files did not pass validity check.


bfq-sirlucjan-20190904.patch ... FAILED

Summary
filip@laptop Linux 5.3.0-1-MANJARO x86_64 18.0.4 Illyria
❯❯❯❯ git clone https://github.com/torvic9/linux52-vd.git                                                   ~/Projecten
Cloning into 'linux52-vd'...
remote: Enumerating objects: 117, done.
remote: Counting objects: 100% (117/117), done.
remote: Compressing objects: 100% (92/92), done.
remote: Total 117 (delta 46), reused 94 (delta 23), pack-reused 0
Receiving objects: 100% (117/117), 239.78 KiB | 919.00 KiB/s, done.
Resolving deltas: 100% (46/46), done.
❯❯❯❯ cd linux52-vd                                                                                         ~/Projecten
❯❯❯❯ git checkout intel                                                                          ~/P/linux52-vd master
Branch 'intel' set up to track remote branch 'intel' from 'origin'.
Switched to a new branch 'intel'
❯❯❯❯ LANG=C makepkg -sri                                                                          ~/P/linux52-vd intel
==> Making package: linux52-vd 5.2.13-3 (Mon Sep  9 20:39:42 2019)
==> Checking runtime dependencies...
==> Checking buildtime dependencies...
==> Installing missing dependencies...
[sudo] password for filip: 
resolving dependencies...
looking for conflicting packages...

Packages (2) elfutils-0.176-1  xmlto-0.0.28-3

Total Installed Size:  1.73 MiB

:: Proceed with installation? [Y/n] 
(2/2) checking keys in keyring                                          [----------------------------------------] 100%
(2/2) checking package integrity                                        [----------------------------------------] 100%
(2/2) loading package files                                             [----------------------------------------] 100%
(2/2) checking for file conflicts                                       [----------------------------------------] 100%
(2/2) checking available disk space                                     [----------------------------------------] 100%
:: Running pre-transaction hooks...
(1/1) Making BTRFS snapshot...
Deleted old snapshot @-drgy6o
Created snapshot @-1l8q0t!
:: Processing package changes...
(1/2) installing xmlto                                                  [----------------------------------------] 100%
(2/2) installing elfutils                                               [----------------------------------------] 100%
:: Running post-transaction hooks...
(1/1) Arming ConditionNeedsUpdate...
==> Retrieving sources...
  -> Downloading linux-5.2.13.tar.xz...
  % Total    % Received % Xferd  Average Speed   Time    Time     Time  Current
                                 Dload  Upload   Total   Spent    Left  Speed
100  102M  100  102M    0     0  4392k      0  0:00:23  0:00:23 --:--:-- 4710k
  -> Downloading linux-5.2.13.tar.sign...
  % Total    % Received % Xferd  Average Speed   Time    Time     Time  Current
                                 Dload  Upload   Total   Spent    Left  Speed
100   989  100   989    0     0   1818      0 --:--:-- --:--:-- --:--:--  1818
  -> Found config.x86_64
  -> Found config
  -> Found config.vd
  -> Found config.x270
  -> Found linux52-vd.preset
  -> Found 60-linux.hook
  -> Found 90-linux.hook
  -> Found x509.genkey
  -> Found 0001-add-sysctl-to-disallow-unprivileged-CLONE_NEWUSER-by.patch
  -> Found 0002-ZEN-Add-CONFIG-for-unprivileged_userns_clone.patch
  -> Found clearlinux-20190902.patch
  -> Found cpu-optimisations-gcc9-graysky.patch
  -> Found kconfig-20190507-add-a-CONFIG-option-that-sets-O3.patch
  -> Found sched-fixes-hho.patch
  -> Found cfs-lowlatency.patch
  -> Downloading bfq-sirlucjan-20190904.patch...
  % Total    % Received % Xferd  Average Speed   Time    Time     Time  Current
                                 Dload  Upload   Total   Spent    Left  Speed
100  176k  100  176k    0     0   315k      0 --:--:-- --:--:-- --:--:--  314k
  -> Downloading bfq-sirlucjan-20190904-revert.patch...
  % Total    % Received % Xferd  Average Speed   Time    Time     Time  Current
                                 Dload  Upload   Total   Spent    Left  Speed
100  1926  100  1926    0     0   4205      0 --:--:-- --:--:-- --:--:--  4205
  -> Downloading futex-steam-sirlucjan.patch...
  % Total    % Received % Xferd  Average Speed   Time    Time     Time  Current
                                 Dload  Upload   Total   Spent    Left  Speed
100 18876  100 18876    0     0  23893      0 --:--:-- --:--:-- --:--:-- 23863
  -> Downloading fix-awk-regexp.patch...
  % Total    % Received % Xferd  Average Speed   Time    Time     Time  Current
                                 Dload  Upload   Total   Spent    Left  Speed
100  4495  100  4495    0     0  10805      0 --:--:-- --:--:-- --:--:-- 10779
  -> Downloading btrfs-001-keep-pages-dirty-when-using-btrfs_writepage_fixup_worker.patch...
  % Total    % Received % Xferd  Average Speed   Time    Time     Time  Current
                                 Dload  Upload   Total   Spent    Left  Speed
100  4887  100  4887    0     0  10811      0 --:--:-- --:--:-- --:--:-- 10835
  -> Downloading btrfs-002-fix-use-after-free.patch...
  % Total    % Received % Xferd  Average Speed   Time    Time     Time  Current
                                 Dload  Upload   Total   Spent    Left  Speed
100  4591  100  4591    0     0  11252      0 --:--:-- --:--:-- --:--:-- 11224
==> Validating source files with sha256sums...
    linux-5.2.13.tar.xz ... Passed
    linux-5.2.13.tar.sign ... Skipped
    config.x86_64 ... Passed
    config ... Passed
    config.vd ... Passed
    config.x270 ... Passed
    linux52-vd.preset ... Passed
    60-linux.hook ... Passed
    90-linux.hook ... Passed
    x509.genkey ... Passed
    0001-add-sysctl-to-disallow-unprivileged-CLONE_NEWUSER-by.patch ... Passed
    0002-ZEN-Add-CONFIG-for-unprivileged_userns_clone.patch ... Passed
    clearlinux-20190902.patch ... Passed
    cpu-optimisations-gcc9-graysky.patch ... Passed
    kconfig-20190507-add-a-CONFIG-option-that-sets-O3.patch ... Passed
    sched-fixes-hho.patch ... Passed
    cfs-lowlatency.patch ... Passed
    bfq-sirlucjan-20190904.patch ... FAILED
    bfq-sirlucjan-20190904-revert.patch ... Passed
    futex-steam-sirlucjan.patch ... Passed
    fix-awk-regexp.patch ... Passed
    btrfs-001-keep-pages-dirty-when-using-btrfs_writepage_fixup_worker.patch ... Passed
    btrfs-002-fix-use-after-free.patch ... Passed
==> ERROR: One or more files did not pass the validity check!
==> Removing installed dependencies...
checking dependencies...

Packages (2) elfutils-0.176-1  xmlto-0.0.28-3

Total Removed Size:  1.73 MiB

:: Do you want to remove these packages? [Y/n] y
:: Running pre-transaction hooks...
(1/1) Making BTRFS snapshot...
Deleted old snapshot @-g8vkeh
Created snapshot @-gi48q8!
:: Processing package changes...
(1/2) removing xmlto                                                    [----------------------------------------] 100%
(2/2) removing elfutils                                                 [----------------------------------------] 100%
:: Running post-transaction hooks...
(1/1) Arming ConditionNeedsUpdate...
❯❯❯❯                        
1 Like

Oops.
Run updpkgsums , I think I forgot to update the checksums.
Thanks!

Ok did that, but I think I have to add your public key?

==> Verifying source file signatures with gpg...
    linux-5.2.13.tar ... FAILED (unknown public key 38DBBDC86092693E)

Either use --skippgpcheck or add the key.
It's not mine, but Greg Kroah-Hartmann's.

See here: https://www.kernel.org/category/signatures.html

this worked, compiling

1 Like

My partner might just kill me, but ...

[I'll get back to you]

1 Like

I might run into the same issue😏, but could contribute a Lenovo Ideapad 330 tomorrow...

1 Like

Which CPU does it have? Celeron or iX?

Can't remember, will check and get back to you tomorrow...

Reason I ask is because it has RAID6 algo fixed to AVX2(*) which (I think) the Celeron doesn't support.

(*) Normally the algo is tested at boot and it chooses the best one, but it costs a tiny amount of time. This has been patched so that you can choose the algo in kernel config without testing it at boot.

[  +0,009562] raid6: Forced to use recovery algorithm avx2x2
[  +0,000001] raid6: Forced gen() algo avx2x4

Here's the inxi of said machine:

CPU:       Topology: Quad Core model: Intel Core i5-8250U bits: 64 type: MT MCP L2 cache: 6144 KiB 
           Speed: 700 MHz min/max: 400/3400 MHz Core speeds (MHz): 1: 700 2: 700 3: 700 4: 700 5: 700 6: 700 7: 700 8: 700 

While getting the data I noticed that there are quite a few of these...


ACPI Error: Aborting method \_TZ.FNCL due to previous error (AE_NOT_FOUND) (20190509/psparse-529)
…
[Di Sep 10 08:33:03 2019] ath10k_pci 0000:01:00.0: pci irq msi oper_irq_mode 2 irq_mode 0 reset_mode 0
[Di Sep 10 08:33:03 2019] pcieport 0000:00:1c.0: AER: Multiple Corrected error received: 0000:00:1c.0
[Di Sep 10 08:33:03 2019] pcieport 0000:00:1c.0: AER: PCIe Bus Error: severity=Corrected, type=Physical Layer, (Receiver ID)

...in dmesg, they may mess up the messages we want to see from the new kernel?
Think I need to do a BIOS upgrade first, which requires Windows :frowning: , and I can't just boot from my dirty Windows USB drive, tried that before...need to swap the internal drive :frowning: :frowning:

One of these cases where something small leads to a cascade of other actions...

Thanks for reporting, unfortunately I have no clue where the error comes from.
AER is advanced error recovery for PCIe, which can be disabled with boot parameter pci=noaer. Doesn't solve the problem but should get rid of the messages.

Wait... is this on my kernel or Manjaro's, or both?

1 Like

Don't worry, not your kernel yet :slight_smile:
This is the default Manjaro Stable one (5.2.11-1-MANJARO), but I was wondering if it doesn't make more sense to sort out these errors first before we switch the kernel, otherwise we won't know where potential new errors come from.

And getting rid of the existing errors will include the BIOS upgrade which in turn requires.....and on and on it goess :slight_smile:

Well it would make sense, but if your computer runs well even with these errors (it says "corrected" after all) then I don't think we should worry.
However it would be interesting to know where the errors come from.

What device is 0000:00:1c.0?
See lspci.

FOLLOW >>

Ignore.

2 Likes

Thanks!
The machine seems to be fine re. the Atheros driver.

sudo dmesg|grep firmware|grep ath10k_pci

Gives me a similar, error free result.

The PCI are gone thanks to this:

And I understand I can ignore ACPI errors. And while I was on the case I also fixed the quiet boot so my wife doesn't get to see the error messages during boot :slight_smile:
Everythings fixed now...

00:1c.0 PCI bridge: Intel Corporation Sunrise Point-LP PCI Express Root Port #6 (rev f1)

Handle 0x0019, DMI type 9, 17 bytes
System Slot Information
	Designation: J6C1
	Type: x1 PCI Express x1
	Current Usage: In Use
	Length: Other
	ID: 1
	Characteristics:
		PME signal is supported
		Hot-plug devices are supported
	Bus Address: 0000:00:1c.0

And now over to the actual topic of this thread :slight_smile:

I can give it a try if it works and report back, what info do I need to provide about it? I am little new to Manjaro so I am not up to date on all, but I got experience running Linux.
Do you want the test done on a newer PC or older.
I can do it on a VirtualMachine Now, for a PC I can do it end of the week when I get my system76.

Ari.

Think I got the kernel running.
It's the 5.2.13-3-vd one is it?

So far no new error messages in dmesg, journalctl is clean but my touchpad isn't working any longer
Anything specific I should test?

Forum kindly sponsored by Bytemark