On Thu, 9 Dec 2010, Tejun Heo wrote: > On 12/09/2010 02:49 PM, Thomas Gleixner wrote: > >> * Patches should be split according to logical steps of changes, not > >> per-file. > >> > >> * Patches should be bisectable. IOW, after applying upto any patch in > >> the series, the tree should be buildable and working. > > > > That does not work for a new architecture. There is nothing to bisect. > > Sure, but at least it shouldn't introduce build scripts first which > wouldn't work at all. > > >> * When posting a patch series, especially one as large as 211, please > >> make the mails for the actual patches replies to the head message. > >> Don't post it as 212 separate messages or replies to the immediate > >> previous patch. > >> > >> So, in short, if you're adding a whole new arch, just post it as a > >> single patch or a series of several patches if it requires changes > >> outside of the specific arch subtree. > > > > Crap. a single patch is a major PITA for review. It's even worse than > > 211 per file patches. > > Cut the crap. A single patch may not be perfect for reviewing but > archs are often merged as a single giant patch as bisection is > meaningless anyway. It's not a question of merging. It's a question of reviewing and I've done quite a bunch of reviews on new archs, so I know what I'm talking about. Reviewing a single patch with everything and the world included is just not workable. Thanks, tglx -- To unsubscribe from this list: send the line "unsubscribe linux-arch" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
