Thanks to visit codestin.com
Credit goes to github.com

Skip to content

Conversation

mmstick
Copy link
Member

@mmstick mmstick commented Jul 6, 2023

Kernel Builder (sita) and others added 9 commits July 6, 2023 17:42
…flags

BugLink: [Replace -fcf-protection=none patch with new version]

The gcc -fcf-protection=branch option is not compatible with
-mindirect-branch=thunk-extern. The latter is used when
CONFIG_RETPOLINE is selected, and this will fail to build with
a gcc which has -fcf-protection=branch enabled by default. Adding
-fcf-protection=none when building with retpoline support to
prevents such build failures.

Signed-off-by: Seth Forshee <[email protected]>
BugLink: http://bugs.launchpad.net/bugs/1585311

Signed-off-by: Andy Whitcroft <[email protected]>
Acked-by: Tim Gardner <[email protected]>
Acked-by: Brad Figg <[email protected]>
Signed-off-by: Kamal Mostafa <[email protected]>
Ubuntu's mainline kernels have been missing config,
this commit adds the missing configs set to the
default recommended config from the kernel.

Once Ubuntu's mainline kernels startbuilding
correctly it is likely best to adopt their changes
and drop this commit as well as `base packaging`
commit and pull `base backaging` from their branch.

Below is a lost of changes:

`ANDROID_BINDER_IPC` that has changed to be
mandatory, as such is is a `y`, but others are same
as 6.0.6

fix decnet
fix adp5588
fix cpia2
fix ZR364XX
fix TM6000
fix vaio picturebook
fix HEIUM
fix siemens nixdorf
fix SAA7146 DVB cards
fix kselftests
fix AV7110 cards
fix syntek camera
fix firewire serial
fix supported serial port count
add android binder ipc
fix rave sp eeprom

---

Add CONFIG_INET_TABLE_PERTURB_ORDER

> The default is almost always what you want. Only change this if you know what you are doing.

Kernel docs say to leave the default. We will do the same here,
https://cateee.net/lkddb/web-lkddb/INET_TABLE_PERTURB_ORDER.html

---

Add kernel config for Intel MAX10 BMC Secure Update driver
Add CONFIG_TEST_MAPLE_TREE=n
@leviport
Copy link
Member

leviport commented Jul 6, 2023

My oryp9 also froze in Integrated graphics mode with my USB-C docking display plugged in. dmesg had pretty much the same output as #263 (comment)

@mmstick
Copy link
Member Author

mmstick commented Jul 6, 2023

I wonder if any of our patches are causing this. I could open a separate branch without our patches.

@mmstick
Copy link
Member Author

mmstick commented Jul 6, 2023

Also, I wonder if disabling system76-power would prevent this if it's something related to runtime power management.

@leviport
Copy link
Member

leviport commented Jul 6, 2023

Sounds good, I can give it a go. I suppose it could also be my install though. I could throw in a second drive and see if it happens on a fresh install

mmstick and others added 16 commits July 7, 2023 16:57
Updates packaging information to reflect that WireGuard kernel support is present.

Co-authored-by: thomas-zimmerman <[email protected]>
The pin fixup is required to detect headset microphones on the oryp5.

Fixes: 80690a2 ("ALSA: hda/realtek - Add quirk for Tuxedo XC 1509")
Signed-off-by: Tim Crawford <[email protected]>
Asus released motherboard(s) with an alternate ALC4080 that lacks
a SPDIF jack, and requires applying this map.
The Aorus Xtreme uses the same ID for audio controller, but the
maps are very different. This successfully fixes all of the
audio jacks on the back.
When plugging of unplugging an audio jack on this motherbaord,
sometimes the audio jacks would stop appearing to
pipewire/pulseaudio. Interestingly `cat`-ing out the file
`/proc/asound/<card number>/codec#0`, and or restarting pipewire
fixes the issue temporarily.

This PR improves the current functionality by making hotplug with
one 3.5mm jack work, it still breaks if hotplug is between multiple
jacks though.
@leviport
Copy link
Member

leviport commented Jul 7, 2023

This bug is tricky. To summarize what I've found so far:

  • This affects more than just oryp9, and more than just ADL-H laptops with switchable graphics. I've Also seen it on darp8, darp9, oryp11, and serw13.
  • I tested the linux-6.4.2-testing branch which has our patches removed, and the bug still happens.
  • I tested with intel_idle.max_cstate=1, but that seemed to have no effect.
  • A USB-C to DP cable also had the freeze on oryp10. I will also test DP to MiniDP to see if it happens there as well.

For testing, I'm playing a video in Firefox, as well as adding a bit of load with stress-ng -c 1 -m 2.

@jglathe
Copy link

jglathe commented Jul 7, 2023

mm interesting. My observations with completely different hw: Have a newer NV card (RTX3060 or 3090), it is enough to scroll wrong in firefox wth a video playing, and the whole sys frezes. This is annoying. The same with a GTX1660, no issue. All on AMD m/b, B550 or X570 with R9-3900X..R9-5950X. And this since at least 6.0. I am happy on the GTX1660 box, the others not so much.
Oh, and no issues on a really old Lenovo T500 with Core2 Duo and IGP "graphics". Profits greatly from pop-os and newer kernels, though.

@mmstick
Copy link
Member Author

mmstick commented Jul 7, 2023

Any difference if using our 535 NVIDIA driver?
pop-os/nvidia-graphics-drivers#184

@leviport
Copy link
Member

leviport commented Jul 7, 2023

Ah, forgot to mention that. 535 vs 525 doesn't seem to make any difference. The darp8 & darp9 didn't even have the Nvidia driver installed.

@nerdboy4200
Copy link

One thing you might want to look at before releasing this is the possible bug related to PER_VMA_LOCK functionality - it appears to cause issues with user-land programs (crashing in bad ways) - see:

https://bugzilla.kernel.org/show_bug.cgi?id=217624

They point to a mailing list with possible patches (possibly a combo of patches plus partial combo fixes

@jackpot51
Copy link
Member

I see a recommendation to revert 0bff0aa, we can try it

@jackpot51
Copy link
Member

@nerdboy4200
Copy link

Hunting through the email chain on lore - I see this finally showed up:

https://lore.kernel.org/all/[email protected]/

and I think the second patch in that set will do the proper disabling for that new feature.

@mmstick
Copy link
Member Author

mmstick commented Jul 10, 2023

I wonder if it would be worth it to test the 6.3 rc kernels from https://kernel.ubuntu.com/~kernel-ppa/mainline/ to narrow the scope of when our i915 drm crash started happening with USB-C/docked displays.

13r0ck and others added 4 commits July 10, 2023 13:48
A memory corruption was reported in [1] with bisection pointing to the
patch [2] enabling per-VMA locks for x86.
Disable per-VMA locks config to prevent this issue until the fix is
confirmed. This is expected to be a temporary measure.

[1] https://bugzilla.kernel.org/show_bug.cgi?id=217624
[2] https://lore.kernel.org/all/[email protected]

Reported-by: Jiri Slaby <[email protected]>
Closes: https://lore.kernel.org/all/[email protected]/
Reported-by: Jacob Young <[email protected]>
Closes: https://bugzilla.kernel.org/show_bug.cgi?id=217624
Fixes: 0bff0aa ("x86/mm: try VMA lock-based page fault handling first")
Cc: [email protected]
Signed-off-by: Suren Baghdasaryan <[email protected]>
@mmstick mmstick force-pushed the linux-6.4.2_jammy branch from 63347f3 to dc66530 Compare July 10, 2023 11:49
jackpot51
jackpot51 previously approved these changes Jul 10, 2023
@XV-02
Copy link

XV-02 commented Jul 10, 2023

When I was trying to test this a bit on Ubuntu, I was seeing issues with nvidia 525 dkms. Nvidia 535 seemed to be behaving though, and I haven't seen any issues with 525 ans 6.4.2 on Pop.

jackpot51
jackpot51 previously approved these changes Jul 10, 2023
@XV-02
Copy link

XV-02 commented Jul 14, 2023

I think we may not release our virtualbox package as part of our PPA for Ubuntu, which is causing DKMS failures during install/kernel upgrade. When it was initially tested, that wasn't an issue, as the build for virtualbox was explicitly added to the list of sources to pull from.

On the other side of that, I haven't seen any issues with the 525 dkms package during kernel upgrade outside of the specific install where it cropped up. Not sure what caused the conflict, but it seems not to be inherent to our packaging. Scratch that. Still seeing an error when processing 525-dkms - something about it not being configured. Still not having any issues with 535. Not sure what's going on there. I wonder if it's an order-of-operations issue. Looking into that now - but if not, we'll need to push the nvidia 535 driver with the kernel.

@truebit
Copy link

truebit commented Jul 17, 2023

I got two SSDs with the same "MAXIO MAP1602" nvme controller chipset, I found the controller would not init or only init one SSD, After I update my kernel to latest 6.4.3 using mainline due to commit, the 1 or all SSDs would still not initialize with success due to dmesg
Device not ready; aborting initialisation, CSTS=0x0:
dmesg screenshot

After applying patch https://lore.kernel.org/lkml/[email protected]/, it works. Could we add this patch in the upcoming release?
Thanks

@ragardner
Copy link

Thanks for all your hard work on pop os,

Just to say that a lot of users are reporting problems with 7000 series amd gpus on kernel 6.4.x for example:
https://www.reddit.com/r/linux_gaming/comments/14whwac/psa_possible_issue_with_kernel_642_amd_graphics/

cheers

@mmstick
Copy link
Member Author

mmstick commented Jul 17, 2023

@ragardner That's probably pop-os/linux-firmware#36 rather than the kernel.

@mmstick mmstick mentioned this pull request Jul 17, 2023
@mmstick mmstick dismissed jackpot51’s stale review August 3, 2023 16:00

The merge-base changed after approval.

@mmstick mmstick closed this Aug 3, 2023
@mmstick mmstick deleted the linux-6.4.2_jammy branch August 3, 2023 16:00
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.