Moving Linux Kernel to 5.4

[airon11]

Looks similar but even more garbled...
Thank you for pointing me there...I had the xcompmgr disabled to prevent screen tearing, but it looks like reenabking it make it work in full screen...even if some tearing comes back specially in webvideos...but at least mpv works fullscreen also..

Could you post a picture so I can judge whether it is the same issue?

of course..here it is

IMG_20200721_172851.jpg (119.44 KiB) Viewed 2856 times

by the way using kms module with mpv give me black screen when moving the mouse over the windows playing...any hint to solve this also?
thanks

[AkulaMD]

Hi, may i know how to disable it? Thank you in advance.

To disable compositor:

Code: Select all

sudo rm /etc/xdg/autostart/xcompmgr.desktop

Too bad, it's not working for me. Thanks anyway.

[dom]

by the way using kms module with mpv give me black screen when moving the mouse over the windows playing...any hint to solve this also?
thanks

We are aware of that bug. We're currently backporting a newer version of the vc4/kms driver to 5.4. Hopefully that will fix some issues.

[hjimbens]

of course..here it is

I am convinced that this is the same issue. I have not had any response on https://github.com/raspberrypi/linux/issues/3665. I am not sure it cannot be reproduced by the people working on the kernel or that it is considered low priority.

by the way using kms module with mpv give me black screen when moving the mouse over the windows playing...any hint to solve this also?

The last time I checked the kms driver with 5.4 with a fullscreen app I got only a black screen, nothing else. At that time I was told "KMS is still under test on the Pi4, there are likely to be issues with it. You should be using FKMS" https://github.com/raspberrypi/firmware ... -628697225.

[airon11]

by the way using kms module with mpv give me black screen when moving the mouse over the windows playing...any hint to solve this also?
thanks

We are aware of that bug. We're currently backporting a newer version of the vc4/kms driver to 5.4. Hopefully that will fix some issues.

Thank you very much !! So I will wait for this backport and recheck !

[airon11]

of course..here it is

I am convinced that this is the same issue. I have not had any response on https://github.com/raspberrypi/linux/issues/3665. I am not sure it cannot be reproduced by the people working on the kernel or that it is considered low priority.

by the way using kms module with mpv give me black screen when moving the mouse over the windows playing...any hint to solve this also?

The last time I checked the kms driver with 5.4 with a fullscreen app I got only a black screen, nothing else. At that time I was told "KMS is still under test on the Pi4, there are likely to be issues with it. You should be using FKMS" https://github.com/raspberrypi/firmware ... -628697225.

Yes I think it is the same bug...I read both the reports...and I manage to get full screen working without garbage with Compton o xcompmgr enabled...( Which I would not like to use)
Hope it gets fixed, or that the backport of KMS they are doing solves the bigger bugs...

[AkulaMD]

by the way using kms module with mpv give me black screen when moving the mouse over the windows playing...any hint to solve this also?
thanks

We are aware of that bug. We're currently backporting a newer version of the vc4/kms driver to 5.4. Hopefully that will fix some issues.

Thanks. Glad to hear that a backport is coming. Will wait for the solution then.

[ghanta]

I am also getting garbled screen.
Disabling compositor gives a lot of screen tearing (unacceptable).
For now, I am downgrading to 4.19.118-v7l+
For anyone interested, the terminal command with corresponding hash is:

sudo rpi-update e1050e94821a70b2e4c72b318d6c6c968552e9a2

[airon11]

I am also getting garbled screen.
Disabling compositor gives a lot of screen tearing (unacceptable).
For now, I am downgrading to 4.19.118-v7l+
For anyone interested, the terminal command with corresponding hash is:

sudo rpi-update e1050e94821a70b2e4c72b318d6c6c968552e9a2

Actually should be the opposite....no compositor no screen tearing and fullscreen garbling...
Compositor enable no garbling in full screen but some more screen tearing...

[knute]

The new kernel has killed my little 3.5 inch displays. Apparently their drivers aren't compatible with 5.4. I reverted the kernel with rpi-update and that solves the display issues for now. I emailed the maker and if I hear back on a solution I'll post it here.

[LTolledo]

I updated and upgrade my RPi3B+, now the kernel is "5.4"
after upgrade my mouse and keyboard weren't working anymore
the keyboard and mouse are connected to the RPi via KVM switch
when the keyboard and mouse's wireless dongle was connected directly to the RPi3B+'s USB port, mouse and keyboard were functional

therefore:
the 5.4 kernel killed the usage of my KVM switch

am foregoing the update and upgrade of my other KVM connected RPi (this RPi4B secondary desktop and the RPi4B-8G main/media desktop)
as a precaution....

as I need the RPi3B+ to gain functionality with a KVM connected keyboard and mouse, please advice on procedure to revert back to the previous stable kernel (4.19.x)

[dom]

therefore:
the 5.4 kernel killed the usage of my KVM switch

Can you create an issue here?
This sounds like something that will need extra information to diagnose.

[LTolledo]

thanks for the reply @dom

please forgive my ignorance, but how do I create an "issue" at github.
I have not signed up for that...(yet)
and what are the required info so I may be able to prepare those...(I do know dmesg output to be one of those)

[dom]

please forgive my ignorance, but how do I create an "issue" at github.
I have not signed up for that...(yet)
and what are the required info so I may be able to prepare those...(I do know dmesg output to be one of those)

Yes, create a github account, and use the "new issue" button. You'll get a template that suggests info to include.
Someone will ask further questions. output of "lsusb -v" is likely to be one.

[ghanta]

Actually should be the opposite....no compositor no screen tearing and fullscreen garbling...
Compositor enable no garbling in full screen but some more screen tearing...

Yes, true.. thanks for pointing out the mistake. It's my bad.

[Sedeptro]

I'm using the PL011 as primary UART on a RPi 3B. Bluetooth is disabled in config.txt.
The serial device used to be /dev/ttyAMA0. After upgrading to kernel 5.4 it's /dev/ttyS0.

Is this expected behaviour or a bug?

[Sedeptro]

I have to correct myself. According to dmesg the PL011 still is /dev/ttyAMA0:

Code: Select all

[    2.378486] 3f201000.serial: ttyAMA0 at MMIO 0x3f201000 (irq = 81, base_baud = 0) is a PL011 rev2
[    2.380981] 3f215040.serial: ttyS0 at MMIO 0x0 (irq = 53, base_baud = 50000000) is a 16550

But the serial console on GPIO 14+15 reports ttyS0 instead of ttyAMA0 with kernel 4.19.

[PhilE]

The PL011 is not the natural console UART on a Pi 3B. What did you do to make it so in the past?

[Sedeptro]

I've used this in config.txt:

Code: Select all

dtoverlay=pi3-disable-bt
enable_uart=1

The overlay file /boot/overlays/pi3-disable-bt.dtbo has been removed. After changing to

Code: Select all

dtoverlay=disable-bt
enable_uart=1

everything is working as before.

[trejan]

I've used this in config.txt:

Code: Select all

dtoverlay=pi3-disable-bt
enable_uart=1

The overlay file /boot/overlays/pi3-disable-bt.dtbo has been removed. After changing to

Code: Select all

dtoverlay=disable-bt
enable_uart=1

everything is working as before.

pi3-disable-bt was renamed to disable-bt but an alias was added for backwards compatibility so pi3-disable-bt should still work. Did it not work for you?

Code: Select all

$ dtoverlay -h pi3-disable-bt
Name:   pi3-disable-bt

Info:   This overlay has been renamed disable-bt, keeping pi3-disable-bt as an
        alias for backwards compatibility.

Usage:  <Deprecated>

[PhilE]

For the aliasing to work you need a fairly recent firmware and a file called overlay_map.dtb in the overlays directory.

[hippy]

The serial device used to be /dev/ttyAMA0. After upgrading to kernel 5.4 it's /dev/ttyS0.

My primary serial device, GPIO 14/15, with both available is showing up as /dev/ttyS0 (/dev/serial0), under 4.9 and 5.4 on my Pi 3B (non-plus) and Zero W, so I'm not seeing any change there.

[PhilE]

We've already established that the change seen was due to the replacement of pi3-disable-bt with disable-bt.

[mpratt14]

Hi all, got Kernel 5.4 recently through sudo apt-get full-upgrade.

I am having OOM issues (oom-killer kills processes) when running a lot of things (KDE, chromium, building openwrt), but still having 500 MB free ram at the time...

Also, there are new kernel warning lines I don't remember being there before...but it might have been...
I am currently network booting over NFS, which might be related
I don't use wifi or bluetooth so I don't notice anything.

Code: Select all

pi@SDDesk:~ $ dmesg | grep brcmfmac
[   10.440625] brcmfmac: F1 signature read @0x18000000=0x15264345
[   10.448544] brcmfmac: brcmf_fw_alloc_request: using brcm/brcmfmac43455-sdio for chip BCM4345/6
[   10.448995] usbcore: registered new interface driver brcmfmac
[   10.476918] brcmfmac mmc1:0001:1: Direct firmware load for brcm/brcmfmac43455-sdio.raspberrypi,4-model-b.txt failed with error -2
[   10.695404] brcmfmac: brcmf_fw_alloc_request: using brcm/brcmfmac43455-sdio for chip BCM4345/6
[   10.713297] brcmfmac: brcmf_c_preinit_dcmds: Firmware: BCM4345/6 wl0: Mar  2 2020 23:30:41 version 7.45.202 (r724630 CY) FWID 01-72f6ece2
[   12.832600] brcmfmac: brcmf_cfg80211_set_power_mgmt: power save enabled
[   14.321350] brcmfmac: brcmf_cfg80211_set_power_mgmt: power save enabled

[mtinman]

Getting the same thing here using Linux 5.4.51-v7+ #1327 SMP,
on a Raspberry 3 B+, only problem with my unit is that the WiFi is not working properly. Not sure about bluetooth, I'm not using it.

$ dmesg | grep brcmfmac
[ 8.089458] brcmfmac: F1 signature read @0x18000000=0x1541a9a6
[ 8.103112] brcmfmac: brcmf_fw_alloc_request: using brcm/brcmfmac43430-sdio for chip BCM43430/1
[ 8.103406] usbcore: registered new interface driver brcmfmac
[ 8.140278] brcmfmac mmc11: Direct firmware load for brcm/brcmfmac43430-sdio.raspberrypi,3-model-b.txt failed with error -2
[ 8.339535] brcmfmac: brcmf_fw_alloc_request: using brcm/brcmfmac43430-sdio for chip BCM43430/1
[ 8.339669] brcmfmac: brcmf_c_process_clm_blob: no clm_blob available (err=-2), device may have limited channels available
[ 8.340569] brcmfmac: brcmf_c_preinit_dcmds: Firmware: BCM43430/1 wl0: Oct 22 2019 01:59:28 version 7.45.98.94 (r723000 CY) FWID 01-3b33decd
[ 12.112387] brcmfmac: brcmf_cfg80211_set_power_mgmt: power save enabled
[ 17.178368] brcmfmac: brcmf_cfg80211_set_power_mgmt: power save enabled