@0n0w1c Some how I gave you the wrong commit. Guess I was on the wrong tree. You are going to have to start over.
f4fe899e803de922c8f36e92a731ca023afaf3d6
You always want the last commit made. Click on the last commit at the top of the page and it will give you the full commit hash.
That build was successful, no config questions, thank you. 
And confirmed that my distcc config is working properly.
Great!. It built here also. Do you have your x86 hooked with the distcc also. I have a vim3, rock64, pi4 and my x86 desktop all crunching with distccd.
No, this is an all pi4 setup, four of them. I am trying to live an x86_64-free life. 
==> Finished making: linux-rpi4-rc 5.14.rc6-1 (Sun 22 Aug 2021 04:43:01 PM CDT)
real 40m39.065s
user 104m57.260s
sys 34m6.996s
Unfortunately, the 5.14-rc6 kernel does not boot for me with either u-boot or uefi. I have only tried with kms, but lack the motivation to mess with it further. This was more about getting distcc working than it was booting the kernel.
I gave the 5.13.12 kernel a try, it boots on SD with firmware but then fails with UEFI + grub on USB. Grub runs and at least starts to load the kernel, but it appears to hang at that point, no messages.
@Darksky did you try 5.13.12 ? Is it still broken ? Or do you wait for 5.14.x one ?
Compiled and tested 5.13.12 this mourning. Trying to use fkms with V3D rendering is still broken. kms with V3D rendering works or if you comment out the config lines as to disable all VC4 overlays and install xf86-video-fbturbo-git it works.
I am waiting for things to get fixed with VC4 that upstream broke starting with 5.13. According to the post above @0n0w1c posted it looks like it will be at least kernel 5.15.
If you want to play with 5.13.12 here is a link to the packages I built this mourning.
md5sum:
7ea8a113bfdf2b38d40ee169eb7cef35 linux-rpi4-mainline-5.13.12-1.tar
I tried your 5.13.12 package, same results for me. Should I be concerned that now both 5.13 and 5.14 fail to boot with UEFI? 
Looking to me like a regression somewhere.
Someone has opened an issue with the uefi project, but I bet this is a pi4 kernel code issue. And since they do not “endorse” the uefi project, they may not be too inclined to the resolve this issue, given the current kms/fkms issue.
Probably not but it appears they had it working in 5.12 though. Just for laughs have you tried the bcm2711-rpi-4-b.dtb from your 5.12 kernel on the 5.13 kernel.
No, I have not tried changing the dbt, worth a shot. In the ticket, he stated the uefi 1.28 works, while 1.29 fails. So there are a few things to try. I’ll try this evening and see what I get working. I’ll need to find a way off of 5.12 before too long.
Edit: The OP also states the upstream kernel boots, so it does indeed appear to be a rpi kernel issue.
I found this that looks interesting. It appears to be specially for uefi boot for the pi4. It also appears to be a turn key job when the packages are built where you install them and reboot like on a fresh pi image.
I am building it right now. Seems like I noticed some stuff getting built where it might be a hardened kernel. It starts with the RPi 5.14 tree and bcm2711_defconfig and patches the hell out of it. It also downloads a ton of other trees for use; some I noticed are in our repo. I noticed it uses dracut.
If all compiles well I will give you the packages to install. I will not have time to mess with it as I am going to town today and tomorrow.
That looks very interesting.
Be away for a couple of hours. It is still compiling.
@0n0w1c Here is the kernel packages. I looked around in the packages some and it will take me some time to figure out all they are doing. There does not seem to be any mention of prep work on the image before installing that I can see.
II would disable the fkms overlay and gpu_mem= but really not sure if config.txt is supposed to be there. I think it will be uninstalled (along with cmdline.txt) if the kernel package was arch-arm. I believe it might be part of the arch-arm kernel package. If this is the case then fbturbo needs to be installed. You may get files exist and have to force a overwrite.
This new kernel package seem to copy over cmdline.txt from my existing system to the package but there is no config.txt. They put a config-5.14.0-rc6-1-v8+ the put in /boot but it is the kernel config file; strange…
Here is what gets put in /boot:
The kernel is hardened:
Good luck with this. It may take some time to figure things out.
md5sum:
76c65d5fad9b4ca5f5739e507d0f3da8 raspberrypi4-uefi.tar
https://drive.google.com/file/d/1-BOdGnPe5Mnod8nlCHtNatNPxmwp3JCE/view?usp=sharing
Unfortunately, it does not boot, no messages. I do see the grub messages about loading the kernel, and then loading the initramfs but then, nothing. I also tried booting with firmware and it did not boot.
He makes quite a few changes to the kernel config. Some I recognized and agreed with, but there were several I had no idea what they did.
To my eyes, he does good work. The integration of parts is very similar to what I have setup. I hope it catches on.
Very close, it seems it is a firmware issue but I have not yet confirmed it. Rolling back the start4.elf and fixup4.dat to the version that comes with the 1.28 uefi release is reported to work.
