Re: [dpdk-dev] [PATCH v9 10/10] devtools: prevent new instances of rte_panic and rte_exit

[Kevin Traynor <ktraynor@redhat.com>]

On 04/26/2018 10:57 PM, Arnon Warshavsky wrote:
> 
> 
> On Thu, Apr 26, 2018 at 7:08 PM, Kevin Traynor <ktraynor@redhat.com
> <mailto:ktraynor@redhat.com>> wrote:
> 
>     I think the first line of the commit msg ^^^ should change as it no
>     longer prevents but just warns
> 
> Actually the return value did not change, only the warning text did.

So will it fail checkpatch in patchwork? I agree with Aaron and Anatoly
that patches with rte_panic/exit should flag some warning message, but
the maintainer should have final say. I don't think failing checkpatch
is the solution for that.

> I would rather keep the word 'prevents' as an intention declaration that
> puts the fact that a panic is undesired first, and the technical ability
> of a maintainer to allow it second.

It's only words, but to me 'prevents' does not indicate that it is
undesirable - it indicates it is not allowed.

> This is just in terms of how one positions the usage of panic.
> 
>  
> 
>     > 
> 
>     Seems to drop the patch name if it finds a match
> 
> 
> Not sure I understand what you meant here. Can you please elaborate?
> 

Sure, I made a patch with an rte_panic and ran checkpatch on a bunch of
patches. It gave me the name of all the patches except the one with the
rte_panic.

> Thanks
> /Arnon
>