Failing to build manjaro kernel (Error 139)

Support Kernel
1 
make[2]: *** [scripts/Makefile.modfinal:77: drivers/hwmon/wm831x-hwmon.ko] Error 139
make[2]: *** Deleting file 'drivers/hwmon/wm831x-hwmon.ko'
make[2]: *** Waiting for unfinished jobs....
  BTF [M] drivers/i2c/busses/i2c-ali1563.ko
  BTF [M] drivers/i2c/busses/i2c-amd756.ko
make[1]: *** [scripts/Makefile.modpost:156: __modpost] Error 2
make: *** [Makefile:1766: modules] Error 2
==> ERROR: A failure occurred in build().
    Aborting...
2

Is that driver available for the (new) kernel?
Just off the top of my head and with a sandwich in my hand… :wink:

3

I don’t know…
I just cloned the linux514 from manjaro gitlab, added a patch and did makepkg -s

4

Does the patch have anything to do with i2c?

It seems it has errors in that case, but we can’t see it, since you only posted a small bit of the output.

5

No, it modifies drivers/net/wireless/intel/iwlwifi/pcie/drv.c

6

Is the patch based on 5.14 git source?

Because that file saw a few changes this release.

https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/log/drivers/net/wireless/intel/iwlwifi/pcie/drv.c?h=v5.14

7

It is. I cloned this Packages / Core / linux514 · GitLab

8

I was talking about your patch. What source was it created against?

What you linked is not the kernel source. Just the buildscript to build the kernel.

9

Master branch of GitHub - torvalds/linux: Linux kernel source tree
fix iwlwifi ax201 · isXander/linux@eba2632 · GitHub

10

That looks like it’s based on linux-next then. Might not be able to apply in 5.14.

11

How would you suggest I create a .patch from a github tag?

12

Anyway, I don’t see how that .patch correlates to the files associated in the error.

13

I don’t know. I can’t see what files has the errors, since the output you posted is very truncated. :slight_smile:

To create a patch from a gitub branch, simply clone the github repo, and create the patch inside that, like you did with the one you have.

14

Looking into the diffs between 5.14 and master, I see no changes that would interfere with the patch.

image
image1468×917 171 KB

15

Please do not post screenshots of text

16

That package on Packages / Core / linux514 · GitLab does not build, by the way:

In function 'memcpy',
    inlined from 'tl_to_darg.part.0' at fs/ext4/fast_commit.c:1295:2:
./include/linux/fortify-string.h:187:25: error: call to '__read_overflow2' declared with attribute error: detected read beyond size of object passed as 2nd parameter
  187 |                         __read_overflow2();
      |                         ^~~~~~~~~~~~~~~~~~
make[2]: *** [scripts/Makefile.build:271: fs/ext4/fast_commit.o] Error 1
make[1]: *** [scripts/Makefile.build:514: fs/ext4] Error 2

Consequently, not only the call to __read_overflow2 fails but the pendant __write__ovwerflow as well:

In function 'memset',
    inlined from 'init_rock_state.part.0' at fs/isofs/rock.c:74:2:
./include/linux/fortify-string.h:172:17: error: call to '__write_overflow' declared with attribute error: detected write beyond size of object passed as 1st parameter
  172 |                 __write_overflow();
      |                 ^~~~~~~~~~~~~~~~~~
make[2]: *** [scripts/Makefile.build:271: fs/isofs/rock.o] Error 1
make[2]: *** Waiting for unfinished jobs....
  CC      kernel/crash_core.o
make[1]: *** [scripts/Makefile.build:514: fs/isofs] Error 2
make: *** [Makefile:1851: fs] Error 2

According to information I got from https://askubuntu.com/questions/1116971/instal-wifi-driver-in-ubuntu, this looks like there are still problems in some parts of the code although I got the news that kernel 5.14 has been released today.
As a very dirty workaround, it was suggested to set

CONFIG_FORTIFY_SOURCE=n

before attempting to compile the kernel.