AMD Sensor Fusion Hub Laptop Driver Unlikely To Land For Linux 5.7

[schard]

What do you mean by "failed"? Do you get any error messages?
I am currently working on getting rid of the duplicated amd_mp2_pcie stuff and update and use i2c-amd-mp2* from i2c/busses instead.

[finalzone]

See the build on Fedora kernel: https://copr-dist-git.fedorainfraclo...4e644e7b92c2fa
Here is the failure: https://download.copr.fedorainfraclo...er-live.log.gz

[schard]

That's because you applying my fix before the actual patch series.

Applying: ARM: fierror: drivers/hid/amd-sfh-hid/amd_mp2_pcie.c: does not exist in index

Since my patch is patch on top of AMD's patch series, the latter must be applied first.

[finalzone]

Still failed despite your suggestion
https://copr-dist-git.fedorainfraclo...bf319efaf4b048

Failed build log: https://download.copr.fedorainfraclo...er-live.log.gz

Build system

https://copr.fedorainfracloud.org/co...build/1325792/

Maybe using the latest kernel git snapshot to set the patches solve the issue. Unfortunately, I am not familiar with kernel codes. =(

[andrei_me]

I read the review, there's no nitpicking in any way.
Reading the review, I got the feeling that whoever wrote the driver, is not used to the kernel driver development, so you have some "newbie" mistakes, like:

Not following code conventions

Not using or not knowing the existence of shared code that do the same thing, which leads to it being reviewed as NIH

Misusage of hardware bring up, like it followed a old code/cookbook/guide and copycode


Besides that, if the developer is really a newcomer, I would like to leave here a congratulations, first reviews are harsh because those silly mistakes, but please don't give up, having a code written by yourself and being used by millions of people should worth the effort 👏🏻👏🏻

[schard]

For those among the readers here, that are testing the latest patch series v4 by AMD from February and still experiencing

AMD-Vi: Event logged [IO_PAGE_FAULT domain=0x000c address=0xcc201000 flags=0x0020]

errors when not booted with amd_iommu=off, I have a refactored the patch which does not have this issue and implements the HID LL driver more cleanly: https://github.com/conqp/linux/tree/...id/amd-sfh-hid