Switching SUNXI-DEV to 5.5.y

[dolphs]

While 5.4.y is not in ( sunxi64_common.inc ) current yet, at this time of writing I like to explore  " KERNELBRANCH="branch:orange-pi-5.5" ".

Except for just updating to this branch, I noticed it is just too early as megous github branch is not prepared yet for 5.5-RC1.

I am curious to see if any of these have been applied already and what new fun stuff needs to be considered in the kernel .config

[Igor]

If you like to explore, just do it. If something doesn't work or if patches are not applying well - which is expected - dig in and fix them.

[Werner]

https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/diff/?id=v5.5-rc1&id2=v5.4&dt=2

Check this for changes compared to 5.4

 

Examples for things you can find:

https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=2fabf6dd77014f19d45fc71c01f1b073c03df255

https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=46b257b1852fccdc4edaf4bbc6fdb59be0b94f0d

[dolphs]

First attempt (5.5-rc2), thanks to @megi who added rc2 to the 5.5-branch

 

This build resulted in compile errors at once, which was to be expected.

Just reporting here in case DEV will be switched to 5.5 this month.

The log attached holds "duplicate_label" as "well duplicate_node_names"-errors.

compilation.logFetching info...

[Igor]

  On 12/16/2019 at 8:55 AM, dolphs said:

Just reporting here in case DEV will be switched to 5.5 this month.

Expand  

 

It will 100% not unless you will do the job of adjusting and clearing patches.

[Igor]

I was too quick. Even you do the job ... 

[Werner]

  On 12/16/2019 at 8:55 AM, dolphs said:

compilation.log 2.97 kB · 7 downloads

Expand  

Probably some patches are new obsolete since their content has been merged into mainline. Though they getting applied anyways which results in duplicate note errors. Same procedure as every time when a new kernel is released.

Try to find the unneeded patches and disable them.

[martinayotte]

I've committed the sunxi-dev switch to 5.5.y !

The only patches that I didn't took time to fix, so I've disabled them, are :

 

0128-drm-bridge-extract-some-Analogix-I2C-DP-common-code.patch-disabled
0130-drm-bridge-Add-Analogix-anx6345-support.patch-disabled
v3-1-2-si2168-Set-TS-clock-mode-and-frequency.patch-disabled
v3-2-2-dvbsky-Add-support-for-MyGica-T230C-v2.patch-disabled

But I will try again later ...

[Werner]

aufs5-standalone.patch fails to apply to the sources and therefore the build fails obviously. I will retry without AuFS.

 

Edit: Yep, works.

[guidol]

  On 12/27/2019 at 9:40 PM, martinayotte said:

I've committed the sunxi-dev switch to 5.5.y !

Expand  

"only" sunxi-dev or also sunxi64-dev?  

[martinayotte]

  On 12/28/2019 at 3:27 PM, guidol said:

only" sunxi-dev or also sunxi64-dev?

Expand  

Both !

I still have to test some boards but 75% of my garden is now on 5.5.0-rc2 ...

[guidol]

  On 12/28/2019 at 3:37 PM, martinayotte said:

Both !

I still have to test some boards but 75% of my garden is now on 5.5.0-rc2 ...

Expand  

My sunxi64-dev SBCs (3x NPI Neo2 + NPi A64)  does startup fine:
 

Neo2:
Welcome to Armbian Buster with Linux 5.5.0-rc2-sunxi64
package bsp-kernel[19.11.5] u-boot[19.11.5] dtb[19.11.5] firmware[19.11.5] config[19.11.5]
branch[dev]

A64:
Welcome to Debian Buster with Armbian Linux 5.5.0-rc2-sunxi64
package bsp-kernel[19.11.5] u-boot[19.11.5] dtb[19.11.3] firmware[19.11.5] config[19.11.5]
branch[dev]

Thanks
Now I could start to compile for sunxi-dev  

[Igor]

  On 12/28/2019 at 11:01 AM, Werner said:

aufs5-standalone.patch fails to apply to the sources and therefore the build fails obviously.

Expand  

Its expected to fail.

 

It would be probably best to add AUFS="no" here:
https://github.com/armbian/build/blob/master/config/sources/families/include/sunxi64_common.inc#L30

or even to 
https://github.com/armbian/build/blob/master/config/sources/arm64.conf
under if $BRANCH = dev

[Werner]

  On 12/28/2019 at 7:26 PM, Igor said:

Its expected to fail.

 

 

Expand  

Of course I did not expect everything works out of the box. Though I wanted to mention it anyways ;).

[guidol]

  On 12/29/2019 at 9:40 AM, Werner said:

Of course I did not expect everything works out of the box. Though I wanted to mention it anyways ;).

Expand  

I did compile with 

./compile.sh EXPERT=yes WIREGUARD=no AUFS=no BUILD_KSRC=no

and it did also work for for my non-64Bit-sunxi SBCs OPi Zero, OPi One and my BPi M2 Berry

At this time I have no need for WireGuard.

[Igor]

  On 12/29/2019 at 9:40 AM, Werner said:

Though I wanted to mention it anyways ;).

Expand  

Yes, I know. Thanks!

I only proposed to by default disable advanced features such as AUFS, WG, BUILD_KSRC perhaps even WIFI on DEV branches since they - in many cases - fail to build, build longer and its unnecessary trouble in development / testing stage.

[martinayotte]

  On 12/28/2019 at 11:01 AM, Werner said:

aufs5-standalone.patch fails to apply to the sources

Expand  

If I remember, it was already broken in 5.4 ...

  On 12/29/2019 at 10:16 AM, guidol said:

WIREGUARD=no AUFS=no

Expand  

That is what I'm doing since months ...

[Igor]

  On 12/29/2019 at 1:45 PM, martinayotte said:

it was already broken in 5.4 ...

Expand  

 

Yes, it was only fixed few days ago by the AUFS maintainer.

[martinayotte]

  On 12/29/2019 at 1:47 PM, Igor said:

Yes, it was only fixed few days ago by the AUFS maintainer.

Expand  

So, no more needs to default it to "no" ?

[Igor]

  On 12/29/2019 at 1:49 PM, martinayotte said:

So, no more needs to default it to "no" ?

Expand  

 

5.4.y is supported, while for 5.5.y I didn't tried it yet.

[Werner]

  On 12/29/2019 at 10:16 AM, guidol said:

I did compile with 

./compile.sh EXPERT=yes WIREGUARD=no AUFS=no BUILD_KSRC=no

and it did also work for for my non-64Bit-sunxi SBCs OPi Zero, OPi One and my BPi M2 Berry

At this time I have no need for WireGuard.

Expand  

Wireguard seems to work in 5.5.0-rc2. It compiled at least.

  LD [M]  drivers/net/wireguard/wireguard.ko
  LD [M]  net/wireguard/wireguard.ko
  INSTALL drivers/net/wireguard/wireguard.ko
  INSTALL net/wireguard/wireguard.ko

 

[Werner]

Looking decent.

  ___  ____  _    ___
 / _ \|  _ \(_)  / _ \ _ __   ___
| | | | |_) | | | | | | '_ \ / _ \
| |_| |  __/| | | |_| | | | |  __/
 \___/|_|   |_|  \___/|_| |_|\___|

Welcome to Armbian Buster with Linux 5.5.0-rc2-sunxi
[...]
root@honeypot:~# lsmod
Module                  Size  Used by
wireguard              57344  0

 

 

Edit: OPi0+ looking good too.

  ___  ____  _   _____                ____  _
 / _ \|  _ \(_) |__  /___ _ __ ___   |  _ \| |_   _ ___
| | | | |_) | |   / // _ \ '__/ _ \  | |_) | | | | / __|
| |_| |  __/| |  / /|  __/ | | (_) | |  __/| | |_| \__ \
 \___/|_|   |_| /____\___|_|  \___/  |_|   |_|\__,_|___/

Welcome to Armbian buster with Linux 5.5.0-rc2-sunxi64

Though CPU is capped to 1.01GHz while the H5 should work at 1.2GHz.

Found this:

[    6.037059] core: _opp_supported_by_regulators: OPP minuV: 1320000 maxuV: 1320000, not supported by regulator
[    6.037083] cpu cpu0: _opp_add: OPP not supported by regulators (1104000000)
[    6.037167] core: _opp_supported_by_regulators: OPP minuV: 1320000 maxuV: 1320000, not supported by regulator
[    6.037174] cpu cpu0: _opp_add: OPP not supported by regulators (1200000000)
[    6.037301] core: _opp_supported_by_regulators: OPP minuV: 1340000 maxuV: 1340000, not supported by regulator
[    6.037309] cpu cpu0: _opp_add: OPP not supported by regulators (1296000000)
[    6.037392] core: _opp_supported_by_regulators: OPP minuV: 1400000 maxuV: 1400000, not supported by regulator
[    6.037400] cpu cpu0: _opp_add: OPP not supported by regulators (1368000000)

Maybe something missing in the dts?