Help for a newbie

[telnetdoogie]

I received the v1.11.10-23 GPL archive from Ubiquiti, but I think what I was expecting to receive was something that made a bit more sense to me...

The archive has:

buildroot.config
host-manifest.csv
manifest.csv

and four folders:

host-licenses/
host-sources/
licenses/
sources/

inside each of the sources folder are module / dependencies with patches, and the src.tar for each of those modules. For example:

sources/alsa-lib-1.1.5/
  - alsa-lib-1.1.5.tar.bz2
  - 0001-no-mmu.patch
  - 0002-dlmisc.patch
  - 0003-conditional-enabling-of-libdl-in-m4.patch
  - series (appears to be a list of the patches for this module)

How does one untangle what differences are pertinent between the previous version committed to this repo, versus what I received?

If it seems like I'm way too much of a newbie, I won't be offended if you let me know that ;) I would love to learn though.

[telnetdoogie]

...perhaps there's a lot more building and kernel compilation experience than I was expecting before I'm able to diff the two sets of artifacts

[fabianishere]

What I do is I extract the contents of the Linux tarball (located in sources/linux-xxxx) into the udm-kernel repository. This allows Git to collect the changes between the last commit and the new firmware release.

[telnetdoogie]

SUPER helpful. Now a git status shows me the difference in the files... looks like this:

Changes not staged for commit:
  (use "git add <file>..." to update what will be committed)
  (use "git restore <file>..." to discard changes in working directory)
        modified:   drivers/gpio/gpio-pca953x.c
        modified:   drivers/net/ethernet/al/al_eth.c
        modified:   drivers/net/ethernet/al/al_mod_hal_eth_epe.c
        modified:   drivers/net/ethernet/al/al_mod_hal_eth_main.c
        modified:   drivers/net/ethernet/al/al_mod_init_eth_lm.c
        modified:   drivers/net/phy/Kconfig
        modified:   drivers/net/phy/ar8216.h
        modified:   drivers/net/phy/ar8327.c
        modified:   drivers/net/phy/rtl8370.h
        modified:   drivers/net/phy/swconfig_ubnt.c
        modified:   include/linux/switch_ubnt.h
        modified:   include/net/netfilter/nf_conntrack_dpi.h
        modified:   include/net/udp_tunnel.h
        modified:   include/uapi/linux/netfilter/nfnetlink_conntrack.h
        modified:   net/ipv4/udp_tunnel.c
        modified:   net/ipv6/ip6_udp_tunnel.c
        modified:   net/netfilter/nf_conntrack_netlink.c

...looks like a manageable list

[telnetdoogie]

So now I have a branch in my fork with the changes applied, do you want me to do a pull request just for those changes? I assume I want to be able to actually build the kernel with those changes to make sure things work, but that looks like it's a different process (I don't see the .github folder for example in my branch which is a branch off of flavour/stock-v4.19

https://github.com/telnetdoogie/udm-kernel/commit/c1c6f387a454ffffd431dc23c85acc4afc524268

[fabianishere]

You'll have to merge these changes on top of the flavour/edge-v4.19 branch to get access to the .github folder.

[fabianishere]

If you send a pull request, I'll integrate those changes into the repository.

[telnetdoogie]

I super appreciate your help. I merged master into my branch (maybe should've done flavour-edge-v4.19 instead and have now done a successful build of the kernel which I'll test out here shortly. I'll submit a PR for my updates for 1.11.10-23 here shortly.

...looks like master is up to date with flavour-edge-v4.19 so I guess no harm no foul

[telnetdoogie]

When attempting to load the built kernel it appears to cause a kernel panic and then the original stock kernel is loaded.

# cat /sys/fs/pstore/*
[    0.575326] pci-pf-stub 0000:00:04.0: writing to VF config space
[    0.575391] pci-pf-stub 0000:00:05.0: writing to VF config space
[    0.594244] ahci 0001:00:00.0: writing to VF config space
[    0.601907] al_eth 0000:00:00.0: writing to VF config space
[    0.613631] al_eth 0000:00:01.0: writing to VF config space
[    2.295691] al_eth 0000:00:02.0: writing to VF config space
[    2.307391] al_eth 0000:00:03.0: writing to VF config space
[    7.751487] Unable to handle kernel NULL pointer dereference at virtual address 0000000000000020
[    7.760341] Mem abort info:
[    7.763167]   ESR = 0x96000005
[    7.766248]   Exception class = DABT (current EL), IL = 32 bits
[    7.772222]   SET = 0, FnV = 0
[    7.775305]   EA = 0, S1PTW = 0
[    7.778469] Data abort info:
[    7.781405]   ISV = 0, ISS = 0x00000005
[    7.785268]   CM = 0, WnR = 0
[    7.788264] user pgtable: 4k pages, 39-bit VAs, pgdp = 00000000511f7f89
[    7.794930] [0000000000000020] pgd=0000000000000000, pud=0000000000000000
[    7.801744] Internal error: Oops: 96000005 [#1] SMP
[    7.806672] Modules linked in: ubnt_common(PFO)
[    7.811228] Process modprobe (pid: 348, stack limit = 0x00000000f7e3803f)
[    7.818066] CPU: 1 PID: 348 Comm: modprobe Tainted: PF          O      4.19.152-v1.10.0.3686-a2edd0c #1
[    7.827503] Hardware name: Annapurna Labs Alpine V2 UBNT (DT)
[    7.833270] pstate: a0000005 (NzCv daif -PAN -UAO)
[    7.838116] pc : ref_module+0x48/0x110
[    7.841890] lr : resolve_symbol.isra.0+0xa0/0x108
[    7.846616] sp : ffffff8009f13bb0
[    7.849982] x29: ffffff8009f13bb0 x28: ffffff8000b28000
[    7.855318] x27: ffffff8009f13df0 x26: ffffff80089d8670
[    7.860679] x25: ffffff8000b226c8 x24: ffffff80009e4080
[    7.866013] x23: ffffff80009df2b8 x22: ffffff8009f13d70
[    7.871346] x21: ffffff80009e4080 x20: ffffff80009e4378
[    7.876706] x19: ffffff8000b223c0 x18: 0000000000000000
[    7.882039] x17: 0000000000000000 x16: 0000000000000000
[    7.887400] x15: 5400160b13131717 x14: ff00000000000000
[    7.892732] x13: 0000000000000000 x12: 0000000000000007
[    7.898093] x11: 0000000000000030 x10: 0101010101010101
[    7.903427] x9 : 0000000000000003 x8 : 7f7f7f7f7f7f7f7f
[    7.908789] x7 : 6bff636064715e6d x6 : 0000000000008018
[    7.914122] x5 : 1800000000000000 x4 : 0080000000000000
[    7.919482] x3 : ffffff80009df2b8 x2 : 0000000000000000
[    7.924816] x1 : ffffff80009e4080 x0 : ffffff8008131a00
[    7.930177] Call trace:
[    7.932651]  ref_module+0x48/0x110
[    7.936077]  resolve_symbol.isra.0+0xa0/0x108
[    7.940484]  load_module+0x1108/0x22e8
[    7.944257]  __se_sys_finit_module+0xc8/0xd8
[    7.948550]  __arm64_sys_finit_module+0x28/0x38
[    7.953133]  el0_svc_handler+0xc8/0x1a8
[    7.956993]  el0_svc+0x8/0xc4
[    7.959985] Code: 1400000c f9400042 eb02029f 54000120 (f9401043)
[    7.966128] ---[ end trace 7197ffc4a5b46163 ]---
[    7.974407] Kernel panic - not syncing: Fatal exception
[    7.979684] SMP: stopping secondary CPUs
[    7.983650] Kernel Offset: disabled
[    7.987163] CPU features: 0x0,20006008
[    7.990963] Memory Limit: none
[    7.997672] Rebooting in 3 seconds..

[telnetdoogie]

...would I need to rebuild udm-kernel-tools with the commit # from my latest changeset for these changes before I'm able to apply that kernel? I think when I updated udm-kernel-tools I used the previous kernel commit versus my own?

[telnetdoogie]

I think I also didn't use the config.udm file from that build so the kernel I'm attempting to apply is wrong... I'll tinker for a bit

[telnetdoogie]

never mind, I got this build working. I needed to rebuild udm-kernel-tools, merge this with the master branch, AND update the config.stock.udm with the image from the stock kernel. ALL good