A quick guide to why stand-alone checkpatch patches suck...

Nick Krause xerofoify at gmail.com
Wed Sep 17 08:29:20 EDT 2014


> Because in general we don't use asserts in the kernel. I'm sure I've used 10,000s of asserts in user space over the decades.  Zero in the kernel.
>
> Specifically, in user space when writing code we can put asserts throughout the code that will cause an immediate code explosion if unexpected things happen.  In the kernel, the better choice is printing an error message then have the code do it's best to handle it.
>
> That still begs the question of why it happened in the first place.  As long as the event itself us unexpected (ie. not routine) then the error message should remain.  Re-read the sample commit message I wrote.  The first thing I said is the "condition is well understood".  Never remove an error message unless you can explain with clarity why the "error" is happening.   Obviously in that case you should be replacing the error message with a comment that explains the condition.
>
> Greg
> --
> Sent from my Android phone with K-9 Mail. Please excuse my brevity.
Thanks Greg,
I will look into in more carefully later. In addition thanks to all
the others for the patience and help. I understand that
this is not normal in the kernel community and would like to really
thank everyone for the patience and support. I
want to help out and as I am finding out the coding is not the issue
it's my issues with the community which I hope
we can fix in order for me to help the kernel community. In addition I
do find the kernel interesting and really like
working with it, just having issues with understanding how to write patches.
Nick



More information about the Kernelnewbies mailing list