On Mon, 2007-02-19 at 10:50 +0000, Christoph Hellwig wrote: > I think this is the wrong approach. For one thing the unit terms is > rather foregin in Linux
I would rather disagree. Subjective. Unit is a generic word, just like subsystem. Unit-tests for example is a widespread word it refer to internal units of a big system. > , and second this one header file "unit" is > definitly too much. It is .c + .h. Why definitely, where is the definition? I appreciate your opinion, but do not want any endless discussion about "how the life has to go", though. > Normally you'd have as little as possible and at > most four: > > (1) internal definitions of the driver > (2) external kernel interface of the driver if nessecary > (3) user interface of the driver if nessecary > (4) ondisk structure if nessecary Where does this rather restrictive model come from? What does it mean in practice for UBI? Please, be more specific. > Also please make sure you submit patches at this driver level. You really > want one git commit per driver that is totally self-contained. If nessecary > you'd split this into multiple patches if a single one would be bigger then > the message size limit. Please, refine what does this exactly mean. I do not see how I should have sent it, sorry. OK, I've separated external headers, JFFS2 support, build stuff. What next? And I sent it just for review, as you requested, I assumed that for the pull we have git, and I specified the URL. Thanks, Artem. -- Best regards, Artem Bityutskiy (Битюцкий Артём) - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/