Hi Devin, On Wednesday 18 November 2009 09:56:12 Devin Heitmueller wrote: > On Wed, Nov 18, 2009 at 2:01 AM, Hans Verkuil <hverkuil@xxxxxxxxx> wrote: > > Very nice cleanup! > > The last time I saw one of these relatively innocent-looking changes > being done across all drivers without testing, it introduced a rather > nasty and hard to find OOPS into one of my drivers and I had to fix > it: > > http://linuxtv.org/hg/v4l-dvb/rev/5a54038a66c9 > > Is there some reason this is one massive patch instead of individual > patches for each driver? It was just easier to do so in a single patch, there's no other particular reason. The patch can be split. > How confident are we that this *really* isn't going to break some bridge > without anyone realizing it? Is this going to be some situation where it > just "goes in" and then the maintainers of individual bridges are going to > have to clean up the mess when users start complaining? Hopefully not. I haven't changed the drivers blindly but I've tried to understand the logic behind every piece of code I changed. Obviously a bug can still slip in, regardless of how careful we are. So to answer your question, no, the patch will not blindly introduce a mess that will need to be cleaned by driver maintainers, but a bug could still get in. > If there are going to be a series of cleanups such as this, perhaps it > makes sense for Laurent to setup a tree with all the proposed fixes, > and put out a call for testers so we can be more confident that it > doesn't screw anything up. Good idea, I'll do that. I'll incorporate the review comments and I'll send a link to the tree to the mailing list. > Don't get me wrong, I'm all for seeing these things cleaned up, and > the more functionality in the core the better. But I am admittedly a > bit nervous to see huge patches touching all the drivers where I am > pretty sure that the developer probably only tested it on a couple of > drivers and is assuming it works across all. I share your concern. Unfortunately I can't test all the changes myself (unless people start sending me lots of hardware samples, but in that case I'll probably have to move to a bigger house :-)). By the way, how would splitting the patches help solve (or at least mitigate) the problem ? -- Regards, Laurent Pinchart -- To unsubscribe from this list: send the line "unsubscribe linux-media" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
