The sound is back with 5.11.0-2
Another problem the cmdline.txt hasnāt the bluetooth modification ttyAMA0/serial0.
Tested with the last kernel after removing the old file (and .pacsave).
What ever you had for cmdline.txt the 2 new kernels will not be replaced as I tried to explain above when I first posted the new kernels.
The .txt files has been removed so they will not be installed. This change was done so the config files do not get replaced on every kernel upgrade to help with people with custom configs.
There is a process after the package install that checks if .txt or .pacsave files are present. Depending on which is present it will either keep the config.txt file or move the .pacsave to .txt respectively or if none is present it will create new default .txt files.
So while using the new kernels you can make your custom configs within the .txt files and they will stay intact after each new kernel upgrade in the future.
If you go back to an older kernel before the change and switch back to the newer kernel with the change the older kernel when it gets uninstalled will remove config.txt and cmdline.txt as these files are part of itās packages and the new kernel in itās post install process will not see the files and will install the default .txt files.
Disregard that last post. I am leaving it there though as it is a good explanation. I found where I forgot to change to serial0 in post_upgrade(). I had only made the changes in post_install().
Thanks to @tartanpion running into an issue I made an correction involving the post_upgrade () section in the .INSTALL had not been changed to serial0 for the new packages code change for bluetooth to work. New packages pushed to testing and unstable:
@Darksky I ran across this link which I believe provides context for the serial0 change we made. I noticed on my SD image, that I used to help debug my A/V jack issue, had a /proc/cmdline that included ttyS0 rather than ttyAMA0.
Note: `/dev/serial0` and `/dev/serial1` are symbolic links which point to either `/dev/ttyS0` or `/dev/ttyAMA0`.
That is pretty much what the RPi people said when I asked when ttyAMA0 quit working in cmdline.txt. Something about some aliases and said ttyAMA0 was out dated in cmdline.txt. But in the end ttyAMA0 is used so it looks like from my end they are circling around the block when itās easier to walk across the street. But what do I knowā¦
I mis-read they are going to initramfs images with zst; not packages.
Looks like arch-arm has patched their mkinitcpio 30-1 so it still uses gzip. Thinking about it more they supposrt a lot of arm boards and some use kernels versions less than 5.10.