ramips: add support for 6.6 kernel version

[namiltd]

Initial support for 6.6 kernel version.

Build system: x86/64
Build-tested: ramips/mt7620/tplink_archer-c5-v4
Run-tested: ramips/mt7620/tplink_archer-c5-v4

[msandber]

Building fails with message:

make[2]: Leaving directory '/home/ubuntu/git/openwrt/scripts/config'
/home/ubuntu/git/openwrt/include/kernel-version.mk:11: *** Missing kernel version/hash file for 6.6. Please create /home/ubuntu/git/openwrt/include/kernel-6.6.  Stop.
time: target/linux/prereq#0.01#0.00#0.01
    ERROR: target/linux failed to build.

Rebasing to #14751 fixes it.

[namiltd]

Compilation goes smoothly but unfortunately the compiled system does not start and this requires a more detailed investigation.
I only have:

[    0.000000] Linux 6.6.20 (builder@buildhost) (mipsel-openwrt-linux-musl-gcc (OpenWrt GCC 12.3.0 r25430-eb44f39e39) 12.3.0, GNU ld (GNU Binutils) 2.40.0) Mar 6 11:40:57 2024
[    0.000000] Board has DDR2
[    0.000000] Analog PMU set to hw control
[    0.000000] Digital PMU set to hw control
[    0.000000] SoC Type: MediaTek MT7620A ver:2 eco:6
[    0.000000] printk: bootconsole [early0] enabled
[    0.000000] CPU0 revision is: 00019650 (MIPS 24KEc)
[    0.000000] MIPS: machine is TP-Link Archer C5 v4
[    0.000000] detected buffer overflow in memcmp

It's probably a memcmp error when calling detect_memory_region which shouldn't be displayed.

[DragonBluep]

Compilation goes smoothly but unfortunately the compiled system does not start and this requires a more detailed investigation. I only have:

[    0.000000] Linux  6.6.20 (builder@buildhost) (mipsel-openwrt-linux-musl-gcc (OpenWrt GCC 12.3.0 r25430-eb44f39e39) 12.3.0, GNU ld (GNU Binutils) 2.40.0) Mar 6 11:40:57 2024
[    0.000000] Board has DDR2
[    0.000000] Analog PMU set to hw control
[    0.000000] Digital PMU set to hw control
[    0.000000] SoC Type: MediaTek MT7620A ver:2 eco:6
[    0.000000] printk: bootconsole [early0] enabled
[    0.000000] CPU0 revision is: 00019650 (MIPS 24KEc)
[    0.000000] MIPS: machine is TP-Link Archer C5 v4
[    0.000000] detected buffer overflow in memcmp

It's probably a memcmp error when calling detect_memory_region which shouldn't be displayed.

Reverting https://git.kernel.org/pub/scm/linux/kernel/git/next/linux-next.git/commit/include/linux/fortify-string.h?h=next-20240312&id=439a1bcac648fe9b59210cde8991fb2acf37bdab can workaround this issue. Not sure if GCC13 has some help.

[msandber]

Compilation goes smoothly but unfortunately the compiled system does not start and this requires a more detailed investigation. I only have:

A mt7621 build , which is based on top of this PR, boots without problems all the way. If that is of any help narrowing down the cause.

[    0.000000] Linux version 6.6.20 (ubuntu@ubuntu) (mipsel-openwrt-linux-musl-gcc (OpenWrt GCC 12.3.0 r25155-35f4442275) 12.3.0, GNU ld (GNU Binutils) 2.40.0) #0 SMP Tue Mar 12 05:27:34 2024
[    0.000000] SoC Type: MediaTek MT7621 ver:1 eco:3
[    0.000000] printk: bootconsole [early0] enabled
[    0.000000] CPU0 revision is: 0001992f (MIPS 1004Kc)
[    0.000000] MIPS: machine is DNA Valokuitu Plus EX400
[    0.000000] Initrd not found or empty - disabling initrd
[    0.000000] VPE topology {2} total 2
[    0.000000] Primary instruction cache 32kB, VIPT, 4-way, linesize 32 bytes.
[    0.000000] Primary data cache 32kB, 4-way, PIPT, no aliases, linesize 32 bytes
[    0.000000] MIPS secondary cache 256kB, 8-way, linesize 32 bytes.
...

[msandber]

@namiltd And oh yes, kernel 6.6 PR was merged yesterday in the master if you want to rebase to it. :)

[PolynomialDivision]

Please bring the PR in a state that allows to review it. Split the commits like its done on all other kernel update PRs. Thanks! :)

[timocapa]

Compilation goes smoothly but unfortunately the compiled system does not start and this requires a more detailed investigation. I only have:

A mt7621 build , which is based on top of this PR, boots without problems all the way. If that is of any help narrowing down the cause.

Made a mt7621 build, with your changes rebased on top of main, enabling “Strip unnecessary exports from the kernel image” seems to cause a build failure.

Not sure if this was introduced with 6.6 in general, or by one of the changes here.

If you haven’t checked by the time I get home, I’ll attach a build log here. :)

[PolynomialDivision]

  make[2] -C target/linux compile
    ERROR: target/linux failed to build.
make: *** [/__w/openwrt/openwrt/openwrt/include/toplevel.mk:233: target/compile] Error 1
make -r target/compile: build failed. Please re-run make with -j1 V=s or V=sc for a higher verbosity level to see what's going on
====== Make errors from logs/target/linux/compile.txt ======
      |                  ^~~~~~~~~~~~~~~~~~~~
sound/soc/ralink/ralink-i2s.c:489:18: note: (near initialization for 'ralink_i2s_dai.id')
sound/soc/ralink/ralink-i2s.c:490:10: error: 'struct snd_soc_dai_driver' has no member named 'remove'
  490 |         .remove = ralink_i2s_dai_remove,
      |          ^~~~~~
sound/soc/ralink/ralink-i2s.c:490:19: error: initialization of 'unsigned int' from 'int (*)(struct snd_soc_dai *)' makes integer from pointer without a cast [-Werror=int-conversion]
  490 |         .remove = ralink_i2s_dai_remove,
      |                   ^~~~~~~~~~~~~~~~~~~~~
sound/soc/ralink/ralink-i2s.c:490:19: note: (near initialization for 'ralink_i2s_dai.base')
cc1: all warnings being treated as errors

[PolynomialDivision]

I guess you should try .remove_new like done here:
3043bd2

[namiltd]

@DragonBluep All done. I think it's ready to merge now.

[namiltd]

Is there a chance for approval before the kernel is updated to 6.6.24 (released today) and we have to adapt the patches again? Can you help @robimarko?

[DragonBluep]

@DragonBluep Is there a chance for approval before the kernel is updated to 6.6.24 (released today) and we have to adapt the patches again?

LGTM. Reviewed-by: Shiji Yang <yangshiji66@outlook.com>.

Thank you for your contribution. I am not a repo maintainer, so I cannot provide further help.

[PolynomialDivision]

Why is the CI failing?

[namiltd]

@PolynomialDivision Errors are not the result of this PR PR but other commits that have not been adapted to the current code version (bcm53xx and ipq40xx), Patches to them are being prepared (#15042 and #15063)

[PolynomialDivision]

Thanks. Merged to main! :)