Re: [RFC PATCH 5/9] surface_aggregator: Add error injection capabilities

[Maximilian Luz]

On 9/23/20 7:45 PM, Greg Kroah-Hartman wrote:
> On Wed, Sep 23, 2020 at 05:15:07PM +0200, Maximilian Luz wrote:
> > This commit adds error injection hooks to the Surface Serial Hub
> > communication protocol implementation, to:
> >
> >   - simulate simple serial transmission errors,
> >
> >   - drop packets, requests, and responses, simulating communication
> >     failures and potentially trigger retransmission timeouts, as well as
> >
> >   - inject invalid data into submitted and received packets.
> >
> > Together with the trace points introduced in the previous commit, these
> > facilities are intended to aid in testing, validation, and debugging of
> > the Surface Aggregator communication layer.
> >
> > Signed-off-by: Maximilian Luz <luzmaximilian@xxxxxxxxx>
>
> Ok, this is ridiculous.
>
> You are dropping a whole new subsystem on us, with full documentation,
> correct driver model integration, crazy debugfs interactions (I made fun
> of the patch, but the code did work, you just did more work than was
> needed), proper auto-loading of modules, tracing, documentation for more
> things than is ever expected, and now you are adding error injection
> support?
>
> You just made all other code submissions of new subsystems I have gotten
> in the past 2 months look like total crud.  Which, to be fair, they
> probably were, but wow, you just stepped up the level of professionalism
> to a whole new height.
>
> I can only dream that "real Linux companies" take note and try to follow
> this example.  I think I will point them all at this in the future and
> say, "go do it like this one."
>
> very very very nice work, we owe you the beverage of your choice.
>
> greg k-h


Wow, thank you very much for those kind words! That means quite a lot to
me.

To be fair, I've been working on this whole project for about two years
now and a large part of the code has been rewritten in the last half a
year, specifically to get it ready for the kernel. So I guess that might
relativize things a bit :)

Thanks again,
Max