Arnd > -----Original Message----- > From: Arnd Bergmann [mailto:a...@arndb.de] > Sent: 2013年7月15日 3:30 > To: Neil Zhang > Cc: Jason Cooper; Matt Sealey; grant.lik...@linaro.org; > haojian.zhu...@gmail.com; devicetree-disc...@lists.ozlabs.org; > linux-kernel@vger.kernel.org; linux-arm-ker...@lists.infradead.org > Subject: Re: [PATCH V3 1/3] dts: change Marvell prefix to 'marvell' > > On Thursday 11 July 2013, Neil Zhang wrote: > > > > > To do this properly, the drivers are going to have to be > > > > > compatible with the old and the new names, and the binding docs > > > > > updated to reflect the legacy name and the preferred name. > > > > > > > > Properly would be as above. You can stop using stock tickers for > > > > new company names, but anything that has been defined in a device > > > > tree before has to stay that way, and all the manufacturer > > > > prefixes to device names should be the same. What you're proposing > > > > is purely driver bloat and increasing the size of kernel. > > > > > > *I'm* not proposing to change the name, Neil is. I'm proposing that > > > iff they chose to do that, don't break sh*t along the way. > > > > > > > What's your opinion? > > We discussed the topic of fixing bad bindings vs. keeping backwards > compatibility > during last week's Linaro Connect. > > The main outcome was that we need to have a better review for new bindings > getting merged to avoid this situation in the future, but the general > consensus > seems that bindings that are already in use but were never reviewed properly > (which should have caught this) should be changed.
So would you mind if I leave it there (discard this patch) and still use the current 'mvrl' prefix for this patch set. > We will likely establish an annotation in the binding soon to mark those that > can > not be changed as opposed to those that are not considered stable yet. > > Arnd Best Regards, Neil Zhang
